1. health and fitness
2. disease

From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
Human Bone Marrow Microvascular Endothelial Cells Support Long-Term
Proliferation and Differentiation of Myeloid
and Megakaryocytic Progenitors
By Shahin Rafii, Fred Shapiro, Ruth Pettengell, Barbara Ferris, Ralph L. Nachman, Malcolm A.S. Moore,
and Adam S. Asch
19% were megakaryocytic,expressingGPllb/llla orGPlb.
Endothelial cellsare a major component
of the bone marrow
CD34+ cells derived from umbilical cord blood, cultured in
(BM) microenvironment that regulate the traffickingand
the upper chamber of transwell culture plates, as well as
homing of hematopoietic progenitor and stem cells. In this
the cells grown in direct contact with BMEC monolayers,
paper, we provide evidence thatBM
endothelial cells
generatedprogenitorsforup
to 70 days. Unstimulated
(BMECs) also supportmultilineage hematopoiesis by elaboBMEC
monolayers
constitutively
produce
interleukin-6,
ration of soluble cytokines. Hematopoietic progenitor cells
incubated in direct contactwith BMEC monolayers, or physi- Kit-ligand, granulocyte colony-stimulating factor, and granulocytemacrophagecolony-stimulatingfactor.Thesedata
cally separated by microporous membrane, expanded fivesuggest that BMEC regulate proliferation of hematopoietic
fold to sevenfold at 7 days, in the absence of exogenous
long-termculture initiating cells by elabcytokines. Plow cytometric analysis of proliferating progeni-progenitor cells and
tor cells grown in the presence of BMEC monolayers showed oration of lineage-specific cytokines.
0 1995 by The American Societyof Hematology.
that by day 14 of coculture, 70% to 80% of hematopoietic
cells were myeloid, expressing CD15 or CD14, and 14% to
T
HE BONE MARROW (BM) microenvironment is a
complex three-dimensional structure where hematopoietic cells proliferate, mature, migrate into the sinusoidal
space and enter the circulation in an exquisitely regulated
fashion. Stromal cells within the BM microenvironment provide a suitable environment for self-renewal, proliferation,
and differentiation of hematopoietic stem cells."' BM stromal cells consisting of endothelial cells, fibroblasts, reticular
cells, monocytes, osteoclasts, osteoblasts, and adipocytes secrete cytokines and produce extracellular matrix in addition
to providing direct cellular contact that regulates hematopoiesis."6
Our understanding of stromal regulation of hematopoiesis
derives largely from investigations of long-term human BM
culture (LTBMC).2.3.7In LTBMC systems, the irradiated
stromal elements, which are comprised ofBM fibroblasts,
adipocytes, occasional endothelial cells, and macrophages,
are seeded with hematopoietic stem and progenitor cells in
the presence of standard cell culture medium supplemented
with horse serum, fetal calf serum (FCS), and hydrocortisone.' BM stroma used in LTBMC system support the proliferation of long-term culture initiating cells (LTCICs),' and
proliferation of hematopoietic progenitors for several
BM stromal cells and fibroblast cell lines support
the proliferation of predominantly myeloid progenitors.
Murine lymphoid progenitors can only be expanded in selected culture conditions originally described by Whitlock
et
Although stromal cells used in LTBMC occasionally contain a small percentage of contaminating endothelial cells,
the relative contribution of this cell type in the regulation
of hematopoiesis is not well established.'0*'' The anatomic
location and complex network ofBM
endothelial cells
(BMECs) within the BM microenvironment and its close
association with mature hematopoietic elements suggests
that it may regulate both early and late proliferation of hematopoietic elements. Much of the study of regulation of hematopoiesis by endothelial cells derives from studies of human
umbilical vein endothelial cells (HUVECs). Other than macrophage colony-stimulating factor (M-CSF) and Kit-ligand,
resting vascular endothelial cells such as HUVECs have not
been reported to constitutively produce significant quantities
Blood, Vol86, No 9 (November l), 1995: pp 3353-3363
of hematopoietic cytokines.'*." However, stimulation of HUVEC monolayers with interleukin-1 (IL- 1) or tumor necrosis
factor results in production of cytokines which include IL1, IL-4, IL-6, granulocyte-macrophage CSF (GM-CSF), and
granulocyte CSF (G-CSF).'2s'4-20
Phorbol ester-stimulated
HUVEC monolayers express messenger RNA for IL- 1 1, but
IL-11 protein has notbeen detected either in resting or
phorbol myristate acetate (PMA)-activated HUVECs by
IL-11 bioassays or enzyme-linked immunosorbent assay
(ELISA)." In addition, resting vascular endothelium constitutively produces factors that may regulate hematopoiesis,
such as basic fibroblast growth factor (FGF),22-25transforming growth factor p (TGF-p),22 and platelet-derived
growth factor.26Human endothelial cells have recently been
reported to express receptors for IL-3, kit-ligand, erythropoietin, thrombopoietin, and basic FGF."~.27~29
Finally, expression of adhesion molecules by resting and activated endothelium may play a critical role in regulating the transit of
CD34+ progenitor and mature blood cells in and out of the
BM.
We have recently reported the isolation and characterization ofBMECs." In this report we provide evidence that
resting BMEC monolayers support long-term hematopoiesis,
specifically myelopoiesis and megakaryocytopoiesis, by the
From the Department of Hematology-Oncology, New York Hospital-Come11 Medical Center: and The Laboratory of Developmental
Hematopoiesis, Sloan-Kettering Institute, New York, NY.
Submitted March 24, 1995: accepted June 28, 1995.
Supported by Grants No. KO8-HL-02926, ROI-HL-44389,K08HL-02541, P50-HL-18828,ROI-DK-42693,ROI-CA-59461,
and
ROI-HL-46546 from the National Institutes of Health. A.S.A. and
S.R. received support from the Dorothy Rodbell Cohen Foundation
for Sarcoma Research and The Rich Foundation.
Address reprint requests to Shahin Rafii, MD, Department of Hematology-Oncology, NewYork Hospital-Cornel1 Medical Center,
Room C-648, 1300 York Ave, New York, NY 10021.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
"advertisement" in accordance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 1995 by The American Society of Hematology.
OOO6-4971/95/8609-0036$3.00/0
3353
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
3354
elaboration of lineage-specific cytokines. BMEC monolayers
maintain their cellular integrity during long-term coculture
experiments, and do not require immortalization or growth
inhibition with irradiation, allowing the study of the physiologic role of these cells in regulation of hematopoietic cell
trafficking, proliferation, and differentiation in vitro. In addition, we provide evidence thatdirect
contact between
BMECs and CD34+ progenitor cells may augment megakaryocyte proliferation and enhance hematopoietic progenitor
cell expansion.
RAFll ET AL
at 4°C. The majority of MoAb used in these experiments (Table I
wereconjugated with fluorescein(FITC),phycoerythrin(PE).
or
rhodamine (RDI j. Even though the endothelial cell monolayers re-
mained intact during the experimental period reported
here. a few
endothelial cells could have detached during demidepopulation
of
hematopoieticcellsgrown in directcontact with BMECs or HUVECs. As a result. the demidepopulated cells were initially stained
with CD45, which stains all hematopoieticcells. All subsequent
staining was gated on CD45" cells.
Megakaryocyteswere quantified by MoAbstoGPIlb/IlIaand
GPlb. Three different conjugated MoAbs that were specific for different epitopes of GPllb/lIla, were 10E5, PLTI-FITC (Coulter), and
MATERIALS AND METHODS
P2-FITC(Immunotech).UnconjugatedMoAb
lOE5 gavesimilar
results to P2 and PLTI. CD61 (GPIlIa) was another marker used to
Isolation of endothelial cells from BM aspirate. After obtaining
consent, BM aspirates were obtained with a standard Jamshidi needle confirm the presence of megakaryocytes. Ulex europeaus is a lectin
that binds to human endothelial cells as well as subtypes of human
in preservative free heparin (SO UlmL) from posterior or anterior
erythroid precursors and megakaryocytes (unpublished
data. Novemiliac crests of normal volunteer donors undergoing BM harvest at
ber 1994). Because no glycophorin A+ cells were detected in coculMemorial Sloan-Kettering Cancer Center (New York, NY). BMECs
ture experiments, we used Ulex europeaus conjugated with PE o r
were isolated as described recently."' Briefly, the BM aspirate was
FITC to confirm the presence of megakaryocytes proliferating on
passaged through a 40-pm nylon filter (Falcon, Lincoln Park, NJ),
the transwell plates.
and the retained stromal elements were digested withS mL of 0.I %
A relatively high number of CD34' cells isolated from umbilical
collagenase for 20 to30 minutes at 37°C. Subsequently, the digested
cord blood (CB)orperipheral
blood (PB) mononuclearcells by
material was passaged througha 22-gauge needle, and thenrefiltered
immunoadsorption technique stain positively with GPllb/lIla antithrough another 40-pm filter to obtain microvessel fragments. The
body. This false-positive result is caused by nonspecific binding o f
microvessel explants were cultured in endothelial cell growth meplatelets or platelet dust to CD34+ cells. To circumvent this problem,
dium (ECGM) containing M199 medium (MA Bioproducts, WalkCD34' cells were washed three timeswith phosphate-buffered saline
ersville, MD), heparin 90 pg/mL (Sigma Chemical CO, St Louis,
(PBS) and 1 mmol/L EDTA and resuspendedin PBS for immunohisMO), 20% FCS, endothelial cell growth factor (ECGF) 20 &mL
tochemicalstaining or flow cytometryanalysis with antibodies to
(OrganonTeknikaCorp,Durham,NC)
or l0 ng/mL of vascular
GPIIb/IlIa complex.By this method. only 2% of cytospun cells were
ECGF (VEGF) (PeproTech, Rocky Hill, NJ), L-glutamine 2 mmoll
GPIIb/lIIa+.
L (Sigma) penicillin (80 U/mL), and streptomycin (80 ,ug/mL), for
CD2-RDI (Leu-%), CD3-RD1, CD4-RDI (T4), CD8-RDI
(TXj.
5 to 7days.Subsequently,theendothelialcellswereselectively
CDI9-FITC (B4), and CD20-FITC (B 1 ) were used for detection of
isolatedby Ulex europeausselectionasdescribed
previously'" or
lymphoid series. CD14 (Leu"3-FITC.
and M02-RDI clone) were
alternatively,thefibroblastscontaminatingtheculturesweremeused for detection of monocytes. CD33-PE (Leu-MY), which binds
chanically weeded out." The purity of the BMEC monolayers were
to myeloidcells in additiontoearlymegakaryocyticprecursors,
assessed by metabolic labeling with Dil-ac-LDL, and immunohistowas used as a marker for determination of differentiation
state of
chemical staining with monoclonal antibody (MoAb) to factor VIW
hematopoietic
progenitor
cells.
CDI5-FTTC
(Leu-M1)
was
used
for
von Willebrand factor (vWF). On average, 9.5% to 98% of BMEC
detection of early and late myeloid precursors. Although
CD I3 (Leumonolayers had characteristic staining with factor VIII/vWF antiM7) is a suitable marker for detection of myeloid precursors,
we
body, or were metabolically labeled with Dil-ac-LDL.
could not use this marker in coculture experiments because it is also
The pure BMEC collected from each wash were pooled, centriexpressed and released by endothelial cells. HPCA-2-PE. QBEndfuged, then resuspended in ECGM and plated on gelatin-coated sixIO-FITC, and 1 1, I .6 were three MoAbs used for detection of CD34
well Costar tissue culture dishes (Costar, Cambridge, MA). For cocells during coculture experiments.
culture experiments confluent monolayers
of BMEC, HUVEC, or
Controls were isotype-matched nonimmune lgGs and
FITC. PE.
BM, stroma were incubatedwith CD34' cells in a moditied coculture
or
RDl-conjugated
antimouse
IgG
F(ab)2.
Cell-associated
immunomedium containing Iscove's modified Dulbeccos medium (IMDM;
fluorescencewasassayed
by quantitative flow cytometryCoulter
MABioproducts),20%FCS,
penicillin (80 U/mLj,streptomycin
EPICS, available at Cornell Medical College (New York, NY).
(80 pg/mL), with or without 0.5 pg/mL of heparin and 0.5 ng/mL
Immunohistochemit.altechnique.\.
MoAbs to factorVIIIIvWF
of ECGF. HUVEC monolayers were cultivated in ECGM,
by the
(Dako,Glostrup,Denmark),CD34(HPCA-2,BectonDickinson.
technique described by Jaffe et al." BM stroma were obtained from
MountainView,
CA), intercellularadhesionmolecule
(ICAMj
plating BM aspirate, as previously described.x The BM stroma were
(AMAC).vascular
cell adhesionmolecule(VCAM)(AMAC).
maintained in IMDM with 10% FCS, 10% horse serum, 2 mmol/L
CDIS-FITC (Leu-MI), CD14-PE (Mo2-RDI), GPIb (SZ2, AMAC).
L-glutamine, and 10.' mol/L hydrocortisone. Less than 2% of the
GPlIb/llta (10E5, a gift from
Dr B.Coller, Mount Sinai Medical
VIIVvWF, suggesting
established BM stromastainedwithfactor
or polyelonal antibodiestofactor
VI111
that the predominant cell type
in these monolayers are BM fibroblasts Center,NewYork,NYj,
vWF (Sigma) at different dilutions were incubated with fixed cells
and reticular cells. All monolayers used in these experiments were
for 1 hour. BMEC monolayers with bound hematopoietic cells from
between passages 2 and 4.
different weekly time points of coculture experiments were washed
Immunofluorescent flow cyrometry. Hematopoieticcells
( I O6
threetimes in PBSsupplemented with 2 mmol/L of calciumand
cells/mL) from the upper chamber of the transwell cultures as well
then
magnesium, fixed in 3% formalin in PBSfor30minutes,
as those proliferating in direct contact with BMECs were obtained
quenched with 0.1 glycine for I O minutes, and blocked with 1.5%'
by weekly demidepopulation and were placed in Hanks' balanced
horseserum,Afterwashing
with PBS,biotinylatedantimouse or
salt solution (HBSS, Sigma),0.02% azide, 1% bovine serum albumin
antirabbit Ig diluted 1 :200 in PBS containing 1 % BSA was incubated
(Sigma)containing I mmol/Ladenosine,2
mrnoIL theophylline,
with the cells for 30 minutes at room temperature. After 30 minutes
and 2 mmoL/L EDTA (buffer A) to inhibit megakaryocyte activation
or alkaline phosphatase,
of incubation with avidin-labeled peroxidase
or aggregation. Subsequently, the cells were incubated with saturatslides were rinsed and incubated with peroxidase substrate amino30 minutes
ing doses of conjugated and nonconjugated MoAb for
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
OW
BONE
3355
Table l.MoAbs Used in This Study
Cluster
CD2-RD1
CD3-RD1
CD4-FITC
CD8-RD1
CD13-PE
CDl4-RD1
CD14-PE
CDl5-FITC
CD19-FITC
CDZO-FITC
CD33-PE
CD34-PE
CD34-FITC
CD34
CD36-FITC
CD4la-FITC
IgG1
CD4la
CD42b-PE
IgM
CD4la-FITC
CD45-FITC
CD61-FITC
ND
ND
17
CD1
Ulex europeaus-FITC
Ulex europeaus-PE
MoAb
lsotype
Cellular Distribution
lgGl
T cells, thymocytes
lgGl
T lymphocytes
Helper/inducer
lgGl
cells
IgG1
Cytotoxic, suppressor T cells
T8
IgG1
Myeloid
progenitors, monocytes, endothelial cells
Leu-M7
IgM
Monocytes
M02
IgGm
Monocytes, neutrophils, eosinophils
Leu-M3
Leu-MI
IgM
Myeloid precursor cells, granulocytes
All B cells including pre-B cells
lgGl
B4
IgGza
All B cells
B1
lgGl
Myeloid cells, BFU-Meg,
CFU-Meg
Leu-M9
lgGl
Hematopoietic
progenitor cells, endothelial cells
HPCA-2
IgG1
Hematopoietic progenitor cells, endothelial cells
QBEnd-IO
IgG1
Hematopoietic progenitor cells, endothelial cells
11.1.6
IOP36
IgG1
erythroid
Monocytes, early
progenitors
P2
lgGl
GPllb Meg, platelets
10E5
GPllb/llla Meg, platelets
S22
Meg, lgGl
GPlb
platelets
PLT-1
GPllb/llla Meg, platelets
All hematopoietic cells
KC56
lgGl
Meg,lgGl
GPllla
platelets
IOP61
13
Class
lgGl
II (HM-DR, DP, DO)
Glycophorin
Erythroid
A
lgGl
progenitors, erythrocytes
95c3
lgGl
C-kit receptor, hematopoietic progenitor cells
Lectin
Endothelial
cells,
subtype of erythroid cells, Meg
Lectin
Endothelial
cells,
subtype of erythroid cells,
Biomeda
Meg
TI 1
T3
T4
Manufacturer
C
C
C
C
BD
C
BD
BD
C
C
BD
BD
IM
Dr M.A.S Moore
IM
IM
Dr B. Coller
IM
C
C
BD
C
IM
IM
Sigma
Abbreviations: IM, Immunotech; BD, Becton Dickenson; C, Coulter; ND, not defined; Meg, megakaryocyte.
ethyl carbazole (red stain) or alkaline phosphatase substrate New
Fuchsin (Dako) (red stain) for 10 minutes. After a final rinse, the
cells were counter stained with 1% hematoxylin.
Samples of BM aspirate or confluent monolayers of BMEC, HUVEC, and BM stroma were grown in chamber slides, air dried, and
fixed with acetonelalcohol. Primary mouse MoAbs to factor VIIY
vWF (Dako) were incubated on the slide for 1 hour or at 4°C overnight, then washed three times with PBS, and counter stained with
goat-antimouse IgGl-FITCIPE. Photographs were taken with a Nikon fluorescence microscope (Tokyo, Japan) on Kodak Ektachrome
160T ASA color film (Kodak Ltd, Liverpool, UK). FITC or PE was
visualized using standard FITC/PE excitation/emission filter combinations.
Metabolic labeling withDil-Ac-LDL.
Near confluent monolayers of HUVEC and BMEC or mixed populations of endothelial
cells and fibroblasts were incubated with 1 pg/mL of Dil-Ac-LDL
(acetylated low-density lipoprotein labeled with dioctadecyl 1,3,3,3,tetramethyl-indocarbocyanine perchlorate), (Biomedical Technologies Inc, Stoughton, MA) for 4 hours at 37°C.The cells were washed
with HBSS supplemented with 2 mmol/L calcium and magnesium
chloride for 10 minutes, and examined with a Nikon epifluorescence
microscope with phase-contrast optics. Dil-Ac-LDL uptake was visualized using standard rhodamine excitation/emission filter combinations.
Isolation of CD34' cells. After informed consent, peripheral
blood mononuclear cells (PBMC) mobilized by cyclophosphamide
plus G-CSF were obtained from patients with previously treated
stage IV breast cancer or untreated ovarian cancer. PBMCs underwent low-density mononuclear cell (<1.077 g/mL) separation over
Ficoll-Paque (Pharmacia, Upsala, Sweden). Cells obtained from human umbilical CB were first platelet-depleted using Percoll sepmtion, before Ficoll separation. CD34+ cells were enriched using the
CellPro Ceprate System (kindly provided by CellPro, Bothell, WA)
following standard instructions provided, or using immunomagnetic
beads (Dynal AS, Oslo, Norway). For the CellPro Ceprate System,
low-density mononuclear cells were washed twice with 1% BSA in
PBS, and resuspended in 1% BSA to a concentration of l to 2 X
10' cells/mL and incubated for 25 minutes with an IgM biotinylated
mouse antihuman antibody to CD34 (12.8) at room temperature.
The cells were washed with 1 % BSA to remove excess antibody,
then were resuspended in 5% BSA at a concentration of 1 to 2 X
108/mL, and passedthrough an avidin column. Cells collected in the
CD34+ fraction were resuspended inmodified coculture medium.
For separation using immunomagnetic beads, low-density mononuclear cells were washed twice with 0.1% BSA in PBS and were
resuspended to 1 X lo8 cells/mL. A mouse IgGl antihuman CD34'
antibody developed in our laboratory ( 11.1.6; licensed to Oncogene
Science, Uniondale, NY) was added to the cells at a concentration
of 50 pg/mL for 30 minutes at 4°C. The cells were washed twice
with 0.1%BSA in PBS and resuspended to a concentration of 1 X
IO' cells/mL. Sheep-antimouse IgGl(Fc) immunomagnetic beads
(30 pg/mL), providing a 16:l bead-to-cell ratio, were added for 30
minutes at 4°C. The bead-positive fraction was selected with a magnetic separator, resuspended in 20% FCS, and kept overnight at 37°C
in 100% humidified air with 5% CO2. The following day, the cells
in the bead-negative fraction were recovered.
Coculture model. CD34+ cells (4 X 104/mL)isolated from CB
or PB were plated in triplicate in six-well plates (Costar) in direct
contact with HUVEC, BMEC, or BM stromal monolayer. CD34'
cells (4 X I@/mL) were plated on transwell plates (Costar, 0.4 pm
porosity) on the upper chamber with HUVEC, BMEC, or BM stromal layers grown as confluent monolayers separately on the lower
chamber. Each week, hematopoietic cells were demidepopulated and
replaced with fresh modified coculture medium. For hematopoietic
cells grown in contact with BMECs, HUVECs, or stromal monolayers, one half of the coculture medium was removed and replaced
with fresh medium. Hematopoietic cells grown on the transwells
were demidepopulated as follows: one half of the volumecontaining
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
3356
RAFll ET AL
the cells growing on the transwell were demidepopulated each week
and replaced with fresh coculture medium. In addition, one half of
the medium on the lower chamber of the transwells was removed
and replaced with fresh coculture medium. To determine the phenotype of hematopoietic cells adherent to BMECs,HUVECs, or stromal
monolayers, each week the monolayers were washed three times
with HBSS supplemented with 2 mmol/L magnesium and 2 m o l /
L calcium chloride to remove the nonadherent cells, and fixed with
formalin or acetone for immunohistochemical staining.
Days In Liquid Culture
B
=a,
10B-
I
-Contact
--+-
Transwell
Days In Liquid Culture
C
For BMEC activation studies, BMEC monolayers were incubated
with 10 UlmL of IL-lP for 18 hours at 37°C. Subsequently, BMEC
monolayers were washed five times with ECGM to remove IL-I@
and resuspended in modified coculture medium.
Long-tern culrures. CD34' cells (4 X 104/mL)from CB or PB
were suspended in modified coculture medium and were placed in
either direct contact with BMECs, HUVECs, or stromal monolayers
(contact) or separated by a 0.4 pm transwell membrane (noncontact).
Eachweekthe
wells were demidepopulated, cells counted on a
Coulter Counter ZM (Hialeah, FL), and replaced with an equal volume of fresh modified coculture medium. Cultures were continued
untilno detectable progenitors were generated, though cells were
still present.
After complete removal of nonadherent cells from the stroma, the
adherent cells were removed with 1 mmollL EDTA and 0.1 % trypsin. Trypsin was deactivated with FCS and the cells were washed
three times with PBS. Cells were quantified by hemocytometer and
were evaluated for progenitor content using agarose assays and cell
surface markers using either flow cytometry or immunocytochemistry as described. All of the trypsinized cells were plated for the
agarose assays.
Progenitor assay (agarose assay). Starting at day zero, and then
weekly thereafter, escalating numbers of cells (1 X lo3 to 4 X IO4)
collected after demidepopulation were placed in triplicate in 35-mm
tissue culture dishes (Coming, Coming, NY), containing 20% FCS
in IMDM supplemented with 50 pg/mL gentamicin (GIBCO),
monothioglycerol (7.3 X IO-' mom; Sigma), 0.36% agarose (FMC
Bioproducts, Rockland, ME), and a combination of five cytokines:
20 nglmL human kit-ligand (kindly provided by Immunex, Seattle,
WA); 50 ng/mL human IL-3 (Immunex); 20 ng/mL mutein IL-6
(kindly provided by Imclone Systems Inc, New York, NY); 6 U/
mL erythropoietin (Amgen, Thousand Oaks, CA); and 100 ng/mL
human G-CSF (Amgen). This combination is abbreviated as K36EG.
Cultures were incubated at 37°C in 100% humidified 5% CO2 in air
and removed from the incubator at 14 days. Colonies definedas
greater than 40 cells were counted using an inverted microscope.
Burst-forming unit erythroid (BFU-E) and granulocyte-macrophageCFU (CFU-GM) were observed using this assay. To determine total
progenitors generated over time, the area under the curve (AUC)
was calculated as the sum of the parallelograms created from the
data points (Fig 1, A and B). A simplified formula is: AUC = nA
+ nB + . . . . . + nX, where nA is the number of progenitors at
day zero and nX is the number of progenitors on the last day of
culture.
cDNA-PCR technique. Hematopoietic cells proliferating in direct contact with BMECs or on the transwell plates separated from
BMEC monolayers were removed for determination of GPIIb expression each week. Total RNAwas isolated from 2 X 10' cells
proliferating on days 7, 14, and 21 of coculture by RNAzol reagent
(Biotecx Lab, Houston, TX). Total RNA was also isolated from
intact resting BMEC monolayers. Total RNA was precipitated with
30000
Transwell
0 NA+A
2 20000
p m Non-Adherent
~
.-0c
tsm Adherent
CI
P)
m
2 10000
P
0
Days In Llquid CultUt'e
Fig1.Long-termcultureof
CDW+ progenitor cdb: CD%+ cdls
from PB (A) or CB (B) were placed either in contact with BMEC or
separated by a 0.4-Fm microporoustramell membrane. Each week
the nonadherent d h were domidepopulated and progenitors (mean
f SEM, n = 3) were determined in an agarose assay. On days 14,
28, and 42, the adherent hematopoietic cells were removed with
enzymatic digestionof the adherent monolayer, and the progenitors
were quantified by agarose assay (C). The number of progenitors
(mean 2 SEM, n = 4) adherent to the BMECs during the coculture
experiment compriseda small perwntage of ovorall expanding progenitors. Thereis no d h e n c e betweanthe total number of progenitors (NA + A) that were generated from contact cultures and those
from the transwell plates( P > .05 for all time points).
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
BONEMARROWENDOTHELIUMSUPPORTSHEMATOPOIESIS
0.2 volume of sodiumacetateand 2.5 volume of 1 0 0 % ethanol,
dried,andresuspendedinDEPC-treated dH20. RNAwasreverse
transcribed using the Moloney murine leukemia virus reverse transcriptase, and oligo dT as the primer for 15 minutes at 42°C. DNA
amplicationwasperformedinpolymerasechainreaction(PCR)
bufferusingtheGPIIbprimerpair(5”CTCGCTGCTCTITGACCTCC-3’) and (5’-AAGTCGCGGCTCAGGGTGTTCC-3’).33
For
thrombopoietindetermination,theprimerpair5“AGAAATTGCTCCTCGTGGTCATGCTTCT-3’andthereverseprimer 5’-CAGTCTGCCGTGAAGGACATGG-3’34
were used for DNA amplification.Sampleswere cycled 35timesandthePCRproductswere
analyzed on a 2% NuSieve agarose gel. Theexpectedproductat
265forGPIIbbasepairwascutandsequencedtoconfirmthe
identity of the product.
Quant$cation of cytokines elaborated by BMEC and HUVEC
monlayers. BMEC or HUVEC-conditioned medium were collected
from 3-day-old cultures from equally confluent culture flasks (5.56
X lo6 cells/flask), filtered, and 200 microliter samples were used in
asandwichELISA(R
& D Systems, Minneapolis,MN)for
the
quantitative determinationof IL-6, G-CSF, GM-CSF, and kit-ligand.
Controlmedium,whichcontained M199 medium, 20% FCS,was
used to sustain endothelialcell monolayers during these experiments.
Statistical analysis. Theresultsarepresentedasthemean
?
SEM of the data obtainedfrom three or more experiments performed
in triplicates. Statistical significance was determined using the Student’s r-test.
RESULTS
BMEC monolayers induce proliferation of CD34+ cells.
CD34+ cells isolated by immunoadsorption from PB or CB
were placed in liquid culture in direct cellular contact with
BMEC monolayer, or on 0.4-pm microporous transwell
plates to separate them from BMEC monolayers. Each week,
the nonadherent hematopoietic cells were demidepopulated
and placed in agarose assays to evaluate progenitor content.
Progenitors grown in direct contact with BMEC expanded
fivefold by 7 days and were detectable for up to 56 days in
experiments with PB-derived CD34+ cells, and 70 days for
CB-derived CD34+ cells (Fig 1, A and B). CB CD34+ cells,
grown physically separated from BMEC monolayers in the
transwell plates, generated progenitors for up to 63 days,
whereas PB CD34+ cells generated progenitors for 49 days.
Total progenitors generated from CB CD34+ cells over time
as calculated by integrating the AUC was 88,300 progenitors
in experiments in which cells were in BMEC contact and
80,200 progenitors in BMEC noncontact experiments. For
the first 28 days, a time when the majority of progenitors
were generated, AUC for the two conditions was almost
identical. However, after 28 days, BMEC contact generated
approximately 1 log greater number of progenitors per week.
For PB, AUC for BMEC contact was 39,500 progenitors
and for BMEC noncontact, 17,500 progenitors. The contact
cultures show a slight but consistent advantage in generating
progenitors over those of noncontact cultures. However, this
difference does not reach statistical significance.
Fig 1C shows the number of progenitors from CB that
are adherent to the BMEC monolayers during the coculture
period. On day 14 of coculture, approximately 10% of expanding progenitors are adherent to BMEC monolayers. On
day 28 and 42 of coculture, a comparable number of progenitors are attached to the underlying stroma. However, the
relative number of adherent to nonadherent progenitors are
3357
significantly higher because the nonadherent progenitors are
decreased. There was no statistical difference between the
total number of progenitors (NA + A) generated from contact cultures as compared with those generated from
transwell plates.
Phenotype ofproliferating hematopoietic cells. Flow cytometric and immunohistochemical staining of the proliferating hematopoietic cells grown in direct contact with BMEC
showed that by day 7, 3% 5 2%of hematopoietic cells
remained CD34+, 88% 5 2% of cells expressed CD33,40%
5 5% expressed CD15, and 12% ? 1% were monocytic,
expressing CD14 antigen (Table 2). By day 14, the number
of mature myeloid precursors CD15 were increased to 47%
? 2%, and monocytic cells expressing CD14 increased to
22% L 5%. Similar results were obtained in experiments
performed in transwell plates (Table 2).
Compared withBM stroma and HUVEC monolayers,
BMEC supported greater megakaryocyte proliferation as determined by expression of GPIIb/IIIa and GPIb in Table 2
and Fig 2. Flow cytometric analysis of cells grown on the
transwell plates showed that by day 14, 14% 5 3% of proliferating cells were GPIIb/IIIa+, and CD34+ cells grown in
direct contact with BMEC gave rise to 17% ? 5% GPIIb/
IIIa+ cells. Because a significant number of megakaryocytes
were attached to BMEC and HUVEC monolayers, the number of attached megakaryocytes were quantified each week
by direct immunohistochemistry of monolayers with GPIIb/
IIIa antibodies. Figure 3 displays the total number of adherent and nonadherent megakaryocytes on day 14 of coculture,
which comprised =19% of hematopoietic cells grown in
contact with BMECs. No adherent GPIIb/IIIa+ cells were
detected on the surface of transwell plates. Even though
HUVEC and BM stromal monolayers support expansion of
myeloid series (85% ? 8% of expanding cells), they support
the expansion of only a small percentage of megakaryocytic
GPIIb/IIIa+ cells (Table 2), and Fig 2. The presence of megakaryocytes was also confirmed by reverse transcriptase-PCR(RT-PCR) (using oligonucleotide primers specific for GPIIb
cDNA) on total RNA that was isolated from the hematopoietic cells proliferating in transwells as well as cells grown
in direct contact with BMECs.
In parallel studies, hematopoietic progenitors proliferating
in the coculture experiments were also stained with MoAbs
to lymphoid and erythroid progenitors. Only a small number
of CD2 or CD19 cells were detected. No staining was detected with MoAbs to glycophorin A, CD3, CD20, CD4, or
CD8.
IL-Ip stimulationofBMEC
monolayers. To study the
effect of stimulated endothelium on hematopoiesis, endothelial cells were stimulated with IL-lp (10 U/mL) for 18 hours.
Incubation of CD34+ cells on the transwell plates with ILlp-activated endothelium resulted in 10-fold expansion of
progenitor cells over 7 days, and induced a relatively rapid
maturation of myeloid cells as evidenced by expression of
CD15 (53 5 3, day 7), and CD14 (25% ? 5%, day 7) (Table
3). By day 14, 70% 5 6% of cells were myeloid, expressing
CD33, CD15, or CD14 markers; 6% t 1% were megakaryocytic, expressing GPIlb/IIIa or GPIb markers.
Morphologicanalysis.
Figure 2A shows a typical day
1 coculture experiment where CD34+ cells were added to
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
3358
RAFll ET AL
Table 2. Flow Cytometric Analysis of Hematopoietic Cells Proliferating in the Presence of BMECs, HUVECs, and BM Stroma
Myeloid/Monocytic
CD34'
1
1
Phenotype of proliferating hematopoietic cells
On direct cellular contact with BMEC
Day 1
Day 7
Day 14
Day 21
Day 28
Day 35
On Transwell plates in the presence of BMEC
Day 1
Day 7
Day 14
Day 21
Day 28
Day 35
On direct cellular contact withHUVEC
Day
Day 7
Day 14
Day 21
Day 28
Day 35
On direct contact with BM stroma
Day
Day 7
Day 14
Day 2 1
Day 28
Day 35
(CD33/CD15/CD14)
Megakaryocytic
(CD41iCD42b)
92 t 6
3 2 2
ND
ND
ND
ND
54
88
85
82
86
80
i 615 t 2/ND
92 i 6
ND
ND
ND
ND
ND
54
92
80
83
86
80
t 615 t 2/ND
t 5/47 t 6/12
t 2/54 t 2/24
t 8/62 t 1/29
t 4/50 t 3/28
t 6/49 t 9/30
92 t 6
ND
ND
ND
ND
ND
54
90
91
95
95
92
t 615 t 2/ND
t 4/47 f 10115 t 7
t 6/56 t 6/30 f 9
t 9/59 c 3/35 c 7
t 4/60 t 7/32 t 2
t 3/40 i 9/34 i 8
2
5
8
4
3
92 t 6
ND
ND
54
80
85
88
90
84
t 615 t 2/ND
2 2 1/ND
3 t 2/ND
2 i 1/ND
2 i 1/ND
ND
ND
t 2/40 c 5/12 t 1
t 5/47 t 2/22 t 5
-t
6/52
?
5/26 i 2
2 4/60 t 6/29 c 5
t 6/56 t 4/28 t 6
t 4/30
t 2/45
t 9/60
t 9/56
i- 6/58
t8
t4
t5
24
i8
2 2/10 i- 2
2 7/26 t 2
2 4/30 t 9
t 3/38 +- 7
2 9/34 t 8
2 t 1/ND
14 t 415 t 3
17 2 519 c 4
10 t 316 t 2
LCNErythrotd
(CD4WGlyco A)
9i3/7f4
4 i 213 t 1
95
91
87
92
95
92
t 5/ND
L 3/ND
i 2/ND
t 4/ND
f 2/ND
t 3/ND
2 t 1/ND
10 t 2/ND
14 t 315 2 2
6 t 313 t 1
3 t 214 t 2
2 i l/ND
95
92
90
94
96
96
t 5/ND
t 6/ND
rt 6/ND
t 4/ND
2 5/ND
2 2/ND
95
91
93
92
95
94
2 5/ND
1/ND
t 3/ND
t 214 t 1
t 3/ND
t 2IND
N DIND
-t
ND/ND
NDIND
95
92
94
90
92
94
t 4/ND
t 3/ND
t 5/ND
t 3/ND
t 2/ND
t 5/ND
r 2/ND
t 3/ND
t 4/ND
t 6/ND
I
5/ND
Phenotype of proliferating hematopoietic cells were determined by flow cytometry.Each week CB CD34' cells proliferating on direct contact
with either BMEC, HUVEC, BM stroma, or from transwell plates were washed three times with buffer A and incubated with conjugated MoAbs.
The values reported above are the percent positive cells from IO4 labeled hematopoietic cells (mean t SEM, n = 6).
Abbreviations: LCA, leukocyte common antigen; ND. none detected; Glyco A, Glycophorin A.
confluent monolayers of BMEC. By day 14, large numbers
of hematopoietic cell colonies were detected over the underlying intact BMEC monolayer (Fig 2, B and C). Figure 2D
shows proliferation of hematopoietic colonies of different
size and shape including relatively numerous refractile large
cells resembling polyploid megakaryocytes. Wright-Giemsa
staining of the cytospins of hematopoietic cells proliferating
cells on the transwell, as well as those in direct contact with
BMEC, showed cells with morphologic features consistent
with myeloid series including granule-rich promyelocytes,
myelocytes, metamyelocytes, bands, neutrophils, and large
monocytes (Fig 4, A and B). Cytospins performed on day 14
of coculture experiments showed abundant large polyploid
megakaryocytes that stained with GPIIbnIIa antibody (Fig
4, C and D). Cytospins performed on day 42 and day 56
showed predominantly mature myeloid cells such as monocytes and neutrophils. Quantification of GPIIb/IIIa' cells
detected on the cytospins are consistent with the results obtained from flow cytometry.
Continuous monitoring of BMECandHUVEC
monolayers during the coculture with hematopoietic cells with
phase-contrast microscopyshowedthat endothelial monolayers maintain their morphologic integrity for the first 5
weeks of coculture. However, beyond day 42, the BMEC
and HUVECmonolayers grown in direct or indirect conctact
with hematopoietic cells form characteristic tube-like structures. Endothelial cells maintained in coculture medium in
the absence of hematopoietic cells underwent no major morphologic changes.
Determination of cytokines produced by BMEC. ELISA
assays performed on the conditioned medium obtained from
BMEC and HUVECmonolayers are shown in Table 4. Resting BMEC elaborated high levels of kit-ligand, IL-6, GMCSF, and G-CSF. Resting BMEC monolayers used in these
experiments did not express activation-dependent adhesion
molecules such as VCAM, ICAM, ELAM, suggesting that
constitutive elaboration of cytokines by BMECwasnot
caused by activation of BMEC, but rather was caused by a
physiologic programming of these cells.
DISCUSSION
We show that unstimulated human microvascular endothelium derived from BM microenvironment support proliferation and differentiation of CD34' cells by constitutive
production of G-CSF, GM-CSF, IL-6, and kit-ligand. Constitutive production of cytokines such as G-CSF, GM-CSF,
and IL-6 is a unique feature of microvascular endothelium
because vascular endothelium derived from lining of aorta,
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
BONE MARROW ENDOTHELIUM SUPPORTSHEMATOPOIESIS
3359
Fig 2. Proliferation of CD34' cells. Phase-contrast micrographs of proliferating hematopoietic progenitors in direct contact with resting
BMEC monolayers on day l (A), day 14 (B and C), and day 21 (D).The small 0 3 4 ' cells shown in A (arrows), proliferate to give rise to floating
colonies (arrows) consistent with CFU seen in agarose assays (Band C). D shows proliferation of numerous large refractile cells resembling
megakaryocytes (short arrows). Only a few colonies of hematopoietic cells were attached to the underlying BMEC monolayers. As shown
here, the BMEC monolayers maintain their cobblestone appearance during coculture experiments (original magnification x 2001.
or umbilicalveincanonly
produce these cytokines upon
stimulation with IL-l or TNF.'','""X
Long-term culture studies have shown that BM stroma or
fibroblasts support predominantly myelopoiesis.3 BMECs
not only support myelopoiesis, but also megakaryocytopoiesis, a feature not seen in BM stroma thatweremaintained in similar culture conditions. Elaboration of cytokines
such as IL-6 and kit-ligand by BMECs induce proliferation
of megakaryocytic precursors. CD34' cells that were cocultured in direct contact with BMEC monolayers gave rise to
a larger number of megakaryocytes than those thatwere
cocultured on the transwells. This finding suggests that membrane-bound cytokines such as kit-ligand may influence
megakaryocyte precursor committment and proliferation.
Previously, it has also been shown that endothelial cell extracellular matrix (ECM), which is rich in heparin sulfate and
glycosaminoglycans, functions as a cytokine sink for heparin
binding factors. Among these ECM-bound cytokines, bFGF, kit-ligand, and GM-CSF are positive regulatory factors
that promote megakaryocytopoiesis,"5.25
whereas TGF-/3 is
a potent negative regulatory factor." Thus, it is possible that
interaction of CD34' cells with these positive and negative
regulatory factors released by BMECs have an overall selective advantage for megakaryocyte lineage committment and
proliferation. The critical role of these negative and positive
regulatory factors wererecentlyconfirmed by Waegellet
aI,j7 who have shown that neutralization of TGF-/3 in Dextertype cultures results in significant proliferation of megakaryocytes. This finding suggests that even though BM stroma
produces large amounts of megakaryopoietins such as IL6,'8.3') production of negative regulatory factors such as TGF/3 haspredominant effect in suppressing megakaryocyte
production. Although IL-6 and kit-ligandj5may promote proliferation of megakaryocytic precursors, the presence of
polyploid megakaryocytes in our coculture experiments suggests that BMECs may also elaborate other cytokines such
as thrombopoietin.3'.4"Availability of sensitive ELlSA or
bioassays may facilitate detection of thrombopoietin in coculture conditioned medium.
We have previously shown that a large number of CD34'
cells and megakaryocytes derived from freshly isolated BM
mononuclear cells adhere to BMEC monolayers.'" However.
during the first 4 weeks of coculture period, only a small
fraction of overall proliferating CD34' derived from CB or
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
3360
RAFll ET AL
Table 3. Phenotype of Proliferating Hematopoietic Cellson
Transwell Platesin the Presence of IL-l-Activated BMECs
..V
Day 1
Day 7
Day 14
Day 21
Day 28
c)
aP
.
cp
M
W
W
a
M
E
D
". -
"
0
I
I
I
7
14
21
'.
l
T
I
28
35
42
Myeloid/Monocytic
Megakaryocytic
CD34'
(CD33/CD15/CD14l
(CD41/CD42bl
92 2 6
ND
ND
ND
ND
54 2 615 t 2/ND
95 5 2/53 t 3/10 f 4
91 t 6/70 2 6/25t 5
90 f 4/68 2 4/40 2 7
91 5 2/60 f 7/35 _f 3
2 f 1/ND
4 t 2IND
6 t I/ND
ND/ND
N D/ND
Flow cytometric analysisof hematopoietic cells proliferatingon the
transwell plates in the presence of
IL-10 activated cells. CBCD34'
cells (4 x IO4)were incubated on transwell plates physically separated
from IL-10-stimulated BMEC monolayers. Under this condition, stimulated BMEC resulted in tenfold expansion of CD34' cells with predominance of myeloidand, t o a smaller extent, megakaryocyticcells.
As compared with resting
BMECs, activated BMECs induce differentiaof CD14' cells
tion of CD34' progenitors into a greater percentage
(mean 2 SEM, n = 6).
Abbreviation: ND, none detected.
Days in coculture
Fig 3. Megakaryocyte expansion. The percentageof megakaryocytes proliferating on transwell plates was determined by flow cytomedry by using three different M o b s t oGPllb/llla and GPlb. The
percentage of proliferating megakaryocytes on direct contact with
either BMECs (0).HUMCs (A), or BM stroma (0)was determined
by flow cytometry as well as immunohistochemistryto account for
adherent megakaryocytes. This figure shows the total percentages
of megakaryocytes (meanf SEM, n = 4) that were presentin suspension as well as adherent to BMEC, HUVEC, and BM stromal monolayers at any given week. Compared with BM stroma, BMEC monolayers support expansion of a relatively large number of GPllb/llla+
cells ( P i.05).
PB cells were attached to the BMEC monolayers. The low
affinity of both megakaryocytes and myeloid precursors for
BMEC monolayers under coculture experiments described
in this paper is particularly striking. Continued coculturing
of proliferating hematopoietic cells beyond day 56 to 63,
results in adhesion of relatively large numbers of monocytes,
neutrophils, dendrtitic as well as polyploid megakaryocytes
to BMEC and HUVEC monolayers (data not shown). This
enhanced binding probably reflects complex interaction between maturing hematopoietic cells and endothelial cells.
Both HUVEC and BMEC monolayers support proliferation of a large number of myeloid cells. Production of GCSF, GM-CSF, b-FGF, and kit-ligand by resting BMEC
account for proliferation and differentiation of myeloid prog e n i t o r ~Although
.~~
cellular contact may be critical for megakaryocyte cell proliferation and maturation, direct cellular
contact between CD34+ progenitor cells and BMEC did not
offer any advantage with respect to myeloid cell prolifera-
tion. The majority of proliferating myeloid cells, particularly
early promyelocytes, were highly granular. This finding is
striking because myeloid cells proliferating in the presence
of cytokines or BM stroma were significantly less granular
than those that were grown in the presence of endothelial
cell monolayers.
The number of progenitors generated by BMECswere
estimated by calculation of AUC in Fig 1, A and B. AUC
reflects the dynamics of proliferation, differentiation, and
depletion of hematopoietic stem cells and indicates the net
production of progenitors over several weeks in the face of
weekly demidepopulation. Figure 1, A and B, shows that in
the first few weeks of coculture, BMEC monolayers were
generating progenitors, whereas during the remaining period
of coculture, there were preservation of progenitors despite
weekly demidepopulation. Verfaille et a14' have shown that
direct contact between hematopoietic progenitor cells is not
critical for long-term hematopoiesis. Our data show that direct contact cultures consistently show better preservation
of progenitors, although this did not reach statistical significance. Adherence of a small subset of proliferating progenitors to BMEC monolayers may preserve or prevent the exhaustion of progenitors during the coculture periodand
weekly demidepopulation.
IL-ID activation of BMEC results in significant adhesion
of CD34' cells to these monolayers. These progenitors remained attached to BMEC monolayers for several days,
forming large adherent colonies. Thus, itwasdifficultto
assess the progenitor generated from CD34' cells that were
grown in contact with IL-IP-activated BMEC or HUVEC
Fig 4. Morphologic characteristics of hematopoieticcells proliferating in the presence of BMEC monolayers. Figure4 is the Wright-Giemsa
stain of day 14cytospun hematopoieticcells proliferating in the transwell plates separated from BMECs (A) and in direct contact with BMECs
(B), ahowingthe presenceof polyploid megakaryocytes aswell as early andlate myeloid precursors. Highly granular promyelocytes, myelocytes.
metamyelocytes, and foamy monocytescould be detected as early as day 14of coculture. (Cl is a Wright-Giemsa stain of polyploid megakaryocytes from day 21 of cocultureof hematopoietic cells in direct contact with BMEC. (D) is an immunoh*tochemical stain using MoAb to GPllbl
llla (10E5) of day 21 cytospun
hematopoietic cells demipopulated from direct contact with BMEC monolayem showing the presence of
megakaryocytes (original magnification x 1,OOO).
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
HEMATOPOIESIS
BONE
SUPPORTS
MARROW ENDOTHELIUM
3361
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
3362
RAFil ET A L
Table 4. Cytokines Elaborated by BMEC and HUVEC Monolayers
~~~
Control Medium
Sample (MDC)
IL-6 (3.13)
SCF (kit-ligand)110)
G-CSF (10)
GM-CSF (1.51
~~
(pg/mL)
HUVECs
(pg/mL)
0
150 i 32
105 2 19
1141 t 89
0
0
0
26 i 12
BMECs
(pglmL)
292
3000
345
373
2 21
2 478
t 32
i 36
BMEC-, or HUVEC-conditionedmedium were collected from 3-dayold cultures from equally confluent culture flasks (5.56 x lo6 cells/
flask), filtered, and 200 microliter samples were used in a sandwich
ELlSA for the quantitative determination of IL-6. G-CSF, GM-CSF, and
kit-ligand. Control medium, which contains M199 medium, 20% FCS,
and 90 pg/mL of heparin, wasused to sustain endothelial monolayers
during these experiments.
Abbreviation: MDC, minimal detectable concentration.
monolayers with BM stroma, which would influence proliferation and maturation of hematopoietic cells.
In summary, BMEC monolayers support multilineage hematopoiesis by elaboration of soluble cytokines. Because
BMEC monolayers canmaintain their cellular integrity over
long periods of coculture with different types of hematopoietic cells, they provide anideal feeder layer for LTBMC
studies. The reciprocalinteractionbetween
BMEC monolayers and proliferating hematopoietic cells may be crucial
for self-renewal of pluripotent stem cells, as well as for
regulation of differentiation of committed progenitor cells.
The coculture model described in this report can be used to
studythis complex interactionaswell
asto characterize
the cytokines released by BMECs that are responsible for
of pluripotent
multilineage hematopoiesis andexpansion
progenitor cells.The studyof surface glycoproteins expressed on the luminal and subliminal surface of BMECs in
resting and activated state is critical to the characterization
of the factors that regulate proliferation of CD34' stem cells
as well as the characterization of mature and malignant hematopoietic cells.
monolayers. However, CD34'
cells incubated on transwell
plates, physically separated from IL- 10-activated endothelium, show very rapid proliferation. Flowcytometric analysis
of the proliferating cells show a predominance of myeloid
lineage,suggestingthat
BMEC activation may resultin
ACKNOWLEDGMENT
upregulation of G-CSF and GM-CSFproduction. This result
is consistent with studiesthat have shown that I L - l p activaWe are grateful to Janice Godwin, MT, Mike Querijero, and
Koji
Shido for providing samples and outstanding technical support.
tion of HUVEC cells results in upregulation of G-CSF and
GM-CSF, which would favor myeloid cell proliferation and
differentiation.'*.''.''
REFERENCES
The coculture experiments were performed in an IMDM1. Litchman MA: The ultrastructure of the hematopoietic microbased medium to supply adequate nutrition for proliferating
environment: A review. Exp Hematol 9:391,
1981
2. Dexter TM, Allen TD, Lajtha LG: Conditions controlling the
hematopoietic cells. Very low concentration of heparin and
proliferation of hematopoietic cells in vitl-o. J Cell Physiol 91:344.
endothelial cell growth supplementwere added to IMDM to
1976
maintain endothelial cellmonolayers. Under these conditions
3. Deryugina El, Muller-Sieburg CE: Stromal cells in long-term
BMEC monolayers remained relatively intact and retained
cultures: Keys to the elucidation of hematopoietic development? Crit
their cobblestone appearance till day 42 to 49 of coculture.
Rev Immunol 13:1 l S , 1993
Becauseendothelialcell monolayers,particularly BMEC,
4. Roberts R, Gallagher J , Spooncer E, Allen TD, Blommfield F,
are contact inhibited and can maintain their integrity over
Dexter TM: Heparin sulphate bound growth factors: A mechanism
long periods of time, irradiation or growth inhibition with
for stromal cell mediated hemopoiesis. Nature 322:376, 1988
mitomycin C was not required.Furthermore,addition of
5. Taichman RS, Emerson SG: Human osteoblasts support hemalow concentrationsof ECGF and heparin had no measurable topoiesis through the production of granulocyte colony-stimulating
factor. J Exp Med 179:1677, 1994
effect on proliferation or phenotype of proliferating hema6. Gordon MY, Riley GO, Watt SM, Greaves M F Compartmentopoietic cells. However, prolonged coculture beyond day
tilization
of a hematopoietic growth factor (GM-CSF)
by glycosami42 of BMEC and HUVEC monolayers with hematopoietic
noglycans in the bone marrow microenvironment. Nature 326:403.
cells resulted in reproducible shape changes in the endothe1987
lial cell monolayers. These changes consisted of reorgani7. Quesenberry PJ: Stromal cells in long-term bone marrow culzation of monolayers into tubular cord-like structures remitures, in Tavassoli M (ed): Handbook of the Hematopoietic Microenniscent of in vivo sinusoidal capillaries and microvessels.
vironment. Clifton, NJ, Humana, 1989, p 253
The morphologic change in endothelial cells was seen not
8. Sutherland HJ, LandsdorfPM,HenkelmanDH,EavesAC,
onlyin directcontactcocultures, but alsoinendothelial
Eaves CJ: Functional characterization of individual human hematocell monolayers that were cocultured with hematopoietic
poietic stem cells cultured at limiting dilution of supporting marrow
cells on the transwell plates. BMEC monolayers maintainedstromal layers. Proc Natl Acad Sci USA 87:2S84, 1990
9. Whitlock CA, Witte ON: Long-term culture of B lymphocytes
in coculture medium in the absence of hematopoietic cells
and their precursors from murine bone marrow. Proc Natl Acad Sci
undergonomajormorphologicchanges.These
findings
USA 79:3608, 1982
suggest that soluble factors releasedby hematopoietic cells
10. MonroyRL,DavisTA,NielsenTB,Staton
AJ: Peripheral
may contribute to the changes observed
in these experiBlood hematopoietic progenitorktem cells proliferate to form coloments.
nies in liquid culture but require contact with vascular endothelial
CD34' cells isolated from BM contain CFU-F cells that
cells and GM-CSF. Int J Cell Cloning 10:105, 1992
may give rise to BM stroma during the cocultureconditions
I 1. Bagby GC, Heinrich MC: Vascular endothelial cells and hedescribed in this study. As a result, in experiments designed
matopoiesis: Regulation of gene expression
in human vascular endoin thisreport only CD34' cells derived from CB and PB
thelial cells. Hematol Path 5:93, 1991
12. Sieff CA. Niemeyer CM, FallerDV: The productionof hemawere used to avoid the repopulationof the BMEC or HUVEC
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
BONE MARROWENDOTHELIUM SUPPORTS HEMATOPOIESIS
topoietic growth factors by endothelial accessory cells. Blood Cells
I3:65, 1987
13. Broudy VC, Kovach NL, Bennett LC, Lin N, Jacobsen F W ,
Kidd PG: Human umbilical vein endothelial cells display high-affinity c-kit receptors and produce a soluble form of c-kit receptor. Blood
83:2145, 1994
14. Libby P, Ordovas JM, Auger KR, Robbins AH, Birinyi LK,
Dinarello CA: Endotoxin and tumor necrosis factor induce interleukin-l gene expression in adult human vascular endothelial cells. Am
J Pathol 124:179, 1986
15. Shalaby MR, Waage A, Espevik T: Cytokine-regulation of
interleukin-6 production by human endothelial cells. Cell Immunol
121:372, 1989
16. Segal CA, McCall E, Bagby GC: Erythroid burst promoting
activity produced by interleukin-l stimulated endothelial cells is
granulocyte-macrophage colony-stimulating factor. Blood 72: 1364,
1988
17. Broudy VC, Kaushansky K, Segal GM, Harlan JM, Adamson
JW: Tumor necrosis factor-alpha stimulates human endothelial cells
to produce multilineage hematopoietic growth factor@). Proc Natl
Acad Sci USA 83:7467, 1986
18. Pober SS: Cytokine-mediated activation of vascular endothelium. Am J Pathol 133:426, 1988
19. Zsebo KM, Yuschenkoff VN, Schiffer S, Chang D, McCall
E, Dinarello CA, Brown MA, Altrock B, Bagby GC Jr: Vascular
endothelial cells and granulopoiesis: Interleukin-l stimulates release
of G-CSF and GM-CSF. Blood 71:99, 1988
20. Broudy VC, Kaushansky K, Harlan JM, et al: Interleukin-l
stimulates human endothelial cells to produce granulocyte-macrophage colony-stimulating factor and granulocyte colony-stimulating
factor. J Immunol 139:464, 1987
21. Suen Y,Chang M, Lee SM, Buzby SJ, Cairo MS: Regulation
of interleukin-l 1 protein andmRNA expression in neonatal and
adult fibroblasts and endothelial cells. Blood 84:4125, 1994
22. Hannan RL, Kourembanas S, Flanders KC, Rogelj SJ, Roberts
AB, Faller DV, Klagsbrun M: Endothelial cells synthesize basic
fibroblasts growth factor and transforming growth factors beta.
Growth Factors 1:7, 1988
23. Wilson EL, Rifkin DB, Kelly F, Hannocks MJ, Gabrilove JL:
Basic fibroblast growth factor stimulates myelopoiesis in long-term
human bone marrow cultures. Blood 77:954, 1991
24. Avraham H,BanuN, Scadden DT, Abraham J, Groopman
JE: Modulation of megakaryocytopoiesis by human basic fibroblast
growth factor. Blood 83:2126, 1994
25. Bikfalvi A, Han ZC, Fuhrmann G: Interaction of fibroblast
growth factor (FGF) with megakaryocytopoiesis and demonstration
of FGF receptor expression in megakaryocytes and megakaryocyticlike cells. Blood 801905, 1992
26. Collins T, Ginsberg D, Boss JM, Orkin SH, Pober JS: Cultured human endothelial cells express platelet-derived growth factor
B chain: cDNA cloning and structural analysis. Nature 316:748,
1985
27. Anagnostou A, Liu Z, Steiner M, Chin K, Lee ES, Kessimian
3363
N, Noguchi CT: Erythropoietin receptor mRNA expression in human
endothelial cells. Proc Natl Acad Sci USA 913974, 1994
28. Korpelainen EI, Gamble JR, Smith WB, Goodall GJ, Qiyu
S, Woodcock JM, Dottore M, Vadas MA, Lopez A F The receptor
for interleukin 3 is selectively induced in human endothelial cells
by tumor necrosis factor alpha and potentiates interleukin 8 secretion
and neutrophil transmigration. Proc Natl Acad Sci USA 90:11137,
1993
29. Methia N, Louache F, Vainchenker W, Wendling F: Oligodeoxynucleotides antisense to the proto-oncogene c-mpl specifically
inhibit in vitro megakaryocytopoiesis. Blood 82:1395, 1993
30.Rafii S, Shapiro F, Rimarachin J, Nachman RL, Fems B,
Weksler B, Moore MAS, Asch AS: Isolation and characterization of
human bone marrow microvascular endothelial cells: Hematopoietic
progenitor cell adhesion. Blood 84:10, 1994
31. Davison PM, Bensch K, Karasek U: Isolation and growth of
endothelial cells from microvessels of the newborn human foreskin
in cell culture. J Invest Dermatol 75:316, 1980
32. Jaffe EA, Nachman RL, Becker CG, Minick CR: Culture of
human endothelial cells derived from umbilical veins. J Clin Invest
52:2745, 1973
33. Thoma SJ, Lamping CP, Ziegler BL: Phenotype analysis of
hematopoietic CD34' cell populations derived from human umbilical blood using flow cytometry and cDNA-polymerase chain reaction. Blood 83:2103, 1994
34. De Sauvage FJ, Hass PE, Spencer SD, Malloy BE, Gurney
AL, Spencer SA, Darbonne WC, H e m 1 WJ, Wong SC, Kuang
WJ, Oles KJ, Hultgren B, Solberg LA, Goeddel DV, Eaton DL:
Stimulation of megakaryocytopoiesis by the c-mpl ligand. Nature
369:533, 1994
35. Avraham H, Scadden DT, Chi S, Broudy VC, Krisztina MZ,
Groopman JE: Interaction of human bone marrow fibroblasts with
megakaryocytes: Role of the c-kit ligand. Blood 80:1679, 1992
36. Kuter JK, Gminski DM, Rosenberg RD: Transforming growth
factor beta inhibits megakaryocyte growth and endomitosis. Blood
79:619, 1992
37. Waegell WO, Higley HR. Kincade PW, Dasch JR: Growth
acceleration and stem cell expansion in Dexter-type cultures by neutralization of TGF-beta. Exp Hematol 22:1051, 1994
38. Mei RL, Burstein AS: Megakaryocytic maturation in murine
long-term bone marrow culture: Role of interleukin-6. Blood
78:1438, 1991
39. Stahl PC, Zucker-Franklin D, Evatt LB, Winton FE: Effects
of human interleukin-6 on megakaryocyte development and thrombocytopoiesis in primates. Blood 78:1467, 1991
40. Kaushansky K, Lok Si, Holly RD, Broudy VC, Lin N, Bailey
MC, Forstrom J W , Buddle MM, Oort PJ, Hagen FS, RothGJ, Papayannopoulou T, Foster DC: Promotion of megakaryocyte progenitor
expansion and differentiation by the c-Mpl ligand thrombopoietin.
Nature 369:568, 1994
41. Verfaillie CM: Direct contact between human primitive hematopoietic progenitors and bone marrow stroma is not required for
long-term in vitro hematopoiesis. Blood 79:2821, 1992
From www.bloodjournal.org by guest on February 16, 2015. For personal use only.
1995 86: 3353-3363
Human bone marrow microvascular endothelial cells support
long-term proliferation and differentiation of myeloid and
megakaryocytic progenitors
S Rafii, F Shapiro, R Pettengell, B Ferris, RL Nachman, MA Moore and AS Asch
Updated information and services can be found at:
http://www.bloodjournal.org/content/86/9/3353.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.