T
RANSLATIONAL AND
C
LINICAL
R
ESEARCH
Human Alternatives to Fetal Bovine Serum for the Expansion of
Mesenchymal Stromal Cells from Bone Marrow
K
AREN
B
IEBACK
,
a
A
NDREA
H
ECKER
,
a
A
SLI
K
OCAO
¨ MER
,
a
H
EINRICH
L
ANNERT
,
b
K
ATHARINA
S
CHALLMOSER
,
c,d
D
IRK
S
TRUNK
,
c,e
H
ARALD
K
LU
¨ TER
a
a
Institute of Transfusion Medicine and Immunology, German Red Cross Blood Service of Baden-Wu¨rttemberg-
Hessen, Faculty of Clinical Medicine Mannheim, University of Heidelberg, Heidelberg, Germany;
b
Department
Hematology, Oncology and Rheumatology, Medical Clinic of the University Heidelberg, Heidelberg, Germany;
c
Stem Cell Research Unit Graz, Medical University of Graz, Graz, Austria;
d
University Clinic of Blood Group
Serology and Transfusion Medicine, Medical University of Graz, Graz, Austria;
e
University Clinic of Internal
Medicine, Department of Hematology, Medical University of Graz, Graz, Austria
Key Words. Mesenchymal stromal cells
•
Fetal bovine serum
•
Platelet-derived factors
•
Pooled platelet lysate
•
Human serum
•
Bone
marrow
A
BSTRACT
Mesenchymal stromal cells (MSCs) are promising candi-
dates for novel cell therapeutic applications. For clinical
scale manufacturing, human factors from serum or plate-
lets have been suggested as alternatives to fetal bovine se-
rum (FBS). We have previously shown that pooled human
serum (HS) and thrombin-activated platelet releasate in
plasma (tPRP) support the expansion of adipose tissue-
derived MSCs. Contradictory results with bone marrow
(BM)-derived MSCs have initiated a comprehensive com-
parison of HS, tPRP, and pooled human platelet lysate
(pHPL) and FBS in terms of their impact on MSC isola-
tion, expansion, differentiation, and immunomodulatory
activity. In addition to conventional Ficoll density gradient
centrifugation, depletion of lineage marker expressing cells
(RosetteSep) and CD271
1
sorting were used for BM-MSC
enrichment. Cells were cultured in medium containing ei-
ther 10% FBS, HS, tPRP, or pHPL. Colony-forming units
and cumulative population doublings were determined,
and MSCs were maximally expanded. Although both HS
and tPRP comparable to FBS supported isolation and
expansion, pHPL significantly accelerated BM-MSC prolif-
eration to yield clinically relevant numbers within the first
two passages. MSC quality and functionality including cell
surface marker expression, adipogenic and osteogenic dif-
ferentiation, and immunosuppressive action were similar
in MSCs from all culture conditions. Importantly, sponta-
neous cell transformation was not observed in any of the
culture conditions. Telomerase activity was not detected in
any of the cultures at any passage. In contrast to previous
data from adipose tissue-derived MSCs, pHPL was found
to be the most suitable FBS substitute in clinical scale
BM-MSC expansion. S
TEM
C
ELLS
2009;27:2331–2341
Disclosure of potential conflicts of interest is found at the end of this article.
I
NTRODUCTION
Bone marrow (BM) is a complex tissue harboring hematopoi-
etic stem and progenitor cells, endothelial cells, adipocytes,
osteocytes, and fibroblastoid stromal cells. On cell culture
expansion, BM can yield a multipotent precursor population.
These mesenchymal stromal cells (MSCs) have been assessed
in a variety of preclinical and clinical settings ranging from
regenerative medicine to immunological or hematopoietic
support [1]. With MSCs becoming established in the clinical
setting, issues have been raised regarding how to expand these
cells in large-scale good-manufacturing practice (GMP)-com-
pliant protocols [2–4]. Most expansion protocols use a me-
dium supplemented with fetal bovine serum (FBS). Serum
supplementation is practical because it provides the cells with
vital nutrients, attachment factors, and growth factors. How-
ever, the use of xenogenic serum is complicated because of
high lot-to-lot variability and is associated with a risk of
transmitting infectious agents and immunizing effects [5–7].
Regulatory guidelines aiming to minimize the use of FBS
have further reinforced an intensive search for possible
Author contributions: K.B.: Conception and design, financial support, administrative support, collection and assembly of data, data
analysis, manuscript writing; A.H.: Conception and design, collection and assembly of data, data analysis, manuscript writing, final
approval of the manuscript; A.K.: Conception and design, provision of study material; H.L.: provision of study material, collection of
data; K.S.: provision of study material, manuscript writing; D.S.: financial support, data interpretation, manuscript editing; H.K.:
financial and administrative support, final approval of the manuscript. K.B. and A.H. contributed equally to this work.
Correspondence: Karen Bieback, Ph.D., Institute of Transfusion Medicine and Immunology, German Red Cross Blood Service of
Baden-Wu¨rttemberg-Hessen, Faculty of Clinical Medicine Mannheim, University of Heidelberg, Ludolf-Krehl-Str. 13-17, 68167
Mannheim, Germany. Telephone: 49-621-383-9720; Fax: 49-621-383-9720;; e-mail: karen.bieback@medma.uni-heidelberg.de
Received
February 11, 2009; accepted for publication May 21, 2009; first published online in
S
TEM
C
ELLS
E
XPRESS
June 4, 2009.
V
C
AlphaMed
Press 1066-5099/2009/$30.00/0 doi: 10.1002/stem.139
S
TEM
C
ELLS
2009;27:2331–2341 www.StemCells.com
alternatives [8–10]. Most current clinical data have been
accomplished with MSCs having been expanded in FBS sup-
plemented media without the appearance of major side
effects. In some cases, however, immunological reactions and
anti-FBS antibodies have been observed and considered as
having possibly affected the therapeutic outcome [7, 11].
A chemically defined standardized, xenogeneic antigen-
and serum-free media composition would be the preferential
solution for pharmaceutical scale manufacturing. Such a for-
mulation allowing for both isolation and expansion has not
been achieved thus far [12]. Based on extensive demand, FBS
may also become scarce and expensive.
In the development of a cell-based medicinal product, any
change in the manufacturing process that impacts final prod-
uct quality must show comparability or superiority [13].
Human blood products are already considered to represent
drugs and are produced accordingly, thus offering certain
advantages as potential FBS substitutes. Accordingly, a vari-
ety of human supplements have been postulated as alterna-
tives to FBS to provide nutrients, attachment factors, and
especially growth factors. These include autologous or alloge-
neic human serum, human plasma, cord blood serum, human
platelet derivatives including platelet lysate, and platelet
released factors [3, 4, 14-24]. Analysis of platelet releasates,
lysates, and subcellular fractions has shown that numerous
bioactive molecules are stored within distinct platelet organ-
elles including adhesive proteins, coagulation factors, mito-
gens, protease inhibitors, and proteoglycans [25]. Compared
with serum, buffy coat-derived platelet preparations are of
particular interest because they do not compete with erythro-
cyte and plasma preparation for the limited available blood
donations [26].
In a previous study, we evaluated a variety of platelet
activation protocols to obtain biologically active proteins to
isolate and expand MSCs. Thrombin-activated platelet relea-
sate in plasma (tPRP) and human blood type AB serum (HS)
were found to be superior adjuvants in isolating and expand-
ing human adipose tissue-derived MSCs (AT-MSCs) [27].
The efficiency of both HS and tPRP, but not of pooled human
platelet lysate (pHPL), in expanding AT-MSCs was notable in
contrast to previous reports on BM-MSCs [20, 28]. Conse-
quently, in a recent study we compared the effects of these
three human alternatives on the isolation, expansion, differen-
tiation, and immunomodulatory capacities, as well as the
immunophenotype of BM-MSCs using FBS as the standard
substitute. The experimental setup was expanded by additional
analysis of product purity, because our previous observations
showed reduced depletion of contaminating hematopoietic
cells in AT-MSCs cultured in human supplements. In this
new study, the standard Ficoll gradient density centrifugation
method was therefore compared with the enrichment of MSCs
by either depleting mature hematopoietic cells or by purifying
MSCs expressing CD271 (low affinity nerve growth factor re-
ceptor [LNGFR]) and cultivating the obtained mononuclear
cells in the four different supplements.
M
ATERIALS AND
M
ETHODS
Media and Supplements
Dulbecco modified Eagle’s medium low glucose (Lonza Group
Ltd., Basel, Switzerland, http://www.lonza.com), supplemented
with
4
mM
L
-glutamine
(PAA,
Coelbe,
Germany,
http://
www.paa.at), 50,000 units (U) penicillin/50,000
lg streptomycin
(PAA) served as basal medium in all instances. It was completed
with (a) 10% FBS (MSCGM Single Quots; Lonza Group Ltd.),
(b) 10% HS, (c) 10% tPRP, or (d) 10% pHPL.
Human AB Serum
HS was derived from whole blood donations of prescreened AB
blood group-typed donors. From each donor, whole blood was
drained into blood bags without anticoagulants and allowed to clot
overnight at 4
C. The serum was aliquoted and separated by cen-
trifugation at 2,000
g for 15 minutes. Subsequently, the supernatant
was aliquoted into 15-ml sterile tubes (Greiner Bio-One, Fricken-
hausen, Germany, http://www.gbo.com/en) and frozen at –30
C.
After thawing aliquots from at least five donors, HS was pooled
and sterilely filtered through 0.2-
lm pore filters (Nalgene filtration
device; Nalgene Nunc International, Rochester, NY, USA, http://
www.nuncbrand.com). HS-supplemented medium was pretested to
maintain its mitogenic capacity over a period of at least 4 weeks.
Thus, HS medium was not freshly made for each individual use.
At least 10 different pools were checked to verify reproducibility.
Thrombin-Activated Platelet Releasate Plasma
Four whole blood donations of AB or O blood group-typed donors
were used to prepare one pooled platelet concentrate derived from
buffy coats. Instead of using an additive solution like T-Sol, the
pooled platelet concentrate was suspended in AB plasma of one
donor. Platelet counts ranged between 20
 10
11
and 30
 10
11
platelets per liter determined by CellDyn 3,200 (Abbott, Wiesba-
den, Germany, http://www.abbott.de). Subsequently, the platelet
concentrate was activated by 1 U of human thrombin (Sigma
Aldrich, Hamburg, Germany, http://www.sigmaaldrich.com) [27].
The released factors were separated from the cellular debris by
centrifugation at 3,000
g, followed by filtration through 0.2-lm
pores. By pooling two pooled platelet concentrates, tPRP finally
represented eight donors. Five-milliliter aliquots were stored at
–80
C. After thawing, the aliquot was centrifuged again for 5
minutes at 1,500
g to remove any developing clots. To prevent in
vitro gel formation, 2 U of heparin (Heparin-Natrium-5000-ratio-
pharm; Ratiopharm, Ulm, Germany, http://www.ratiopharm.de)/ml
of medium was added before the tPRP. tPRP was shown to rapidly
lose mitogenic activity. A storage time exceeding 48 hours
resulted in extensive loss of mitogenic activity; thus, the medium
was prepared freshly for each individual use. To verify reproduci-
bility, at least 11 different pools were applied.
Pooled Human Platelet Lysate
pHPL was prepared in Graz as previously described [3]. Briefly,
four buffy coat units of blood group O-typed donors were pooled
in AB plasma and centrifuged (340
g, 6 minutes, 22
C). The pla-
telet rich plasma (PRP) was leukocyte depleted by inline filtration
and was frozen at –30
C. After thawing at 37
C, at least 10 units
of freeze-thaw lysed human platelets were further pooled result-
ing in approximately 40-50 donations per batch to minimize do-
nor variations. pHPL was aliquoted and stored at –30
C. Before
use in cell culture, pHPL was thawed and centrifuged at 4,000
g
for 15 minutes, whereas only the supernatant was added to the
culture medium containing 2 U/ml of preservative-free heparin.
Reproducibility of pHPL effects was verified by using at least
seven different batches of pHPL identically prepared in Mann-
heim by pooling platelet concentrates from eight donors. Where
specified, tPRP and pHPL were prepared from one platelet con-
centrate split in two halves to directly compare both.
Isolation and Culture of BM-Derived MSCs
BM aspirates were harvested using an optimized bone marrow
harvesting technique [29]. Illiac crest bone marrow aspirates were
derived from 14 young healthy donors (median age 22) after hav-
ing received informed consent. Mononuclear cells (MNCs) were
isolated from all heparinized BM aspirates by density gradient
centrifugation (Ficoll Paque, GE Healthcare, Uppsala, Sweden,
http://www.gehealthcare.com) as described elsewhere [30]. Inde-
pendent of the cell number, the MNCs were split into equal sub-
fractions and cultured within the respective basal medium
2332
Human Alternatives to FBS for BM-MSC Expansion
supplemented with either FBS (
n ¼ 14), HS (n ¼ 12), tPRP (n ¼
12), or pHPL (
n ¼ 6) (Fig. 1B).
In
n ¼ 6 BM samples, MSC enrichment using RosetteSep
(StemCell Technologies Inc, St. Katharinen, Germany, http://
www.cellsystems.de) was compared with Ficoll-only isolation.
The RosetteSep antibody cocktail (CD3, CD11b, CD14, CD16,
CD19, CD56, CD66b, and glycophorin A) crosslinks undesirable
cells and forms immunorosettes with red blood cells. These are
pelleted after Ficoll gradient centrifugation. In this case, the BM
aspirate was split into two equal aliquots before MNC isolation.
Resulting cells were cultured in media supplemented with FBS,
HS, or tPRP. On four other samples, CD271 (LNGFR) enrich-
ment was performed in one half of the BM sample, whereas the
other half was split to perform Ficoll and RosetteSep separation.
CD271 sorting involved a magnetic bead-assisted preselection
(AutoMACS device; program ‘‘Possel D’’ [2 columns] and ‘‘Possel
S’’ [sensitive]) using CD271 microbeads (Miltenyi Biotec GmbH,
Bergisch Gladbach, Germany, http://www.miltenyibiotec.com).
Because purity reached, at best, 80%, a flow cytometric sorting fol-
lowed the enrichment (BD FACS Vantage TM SE: sorter used for
flow cytometric cell sorting). This yielded a
>99% CD271
þ
cell
population as assured by flow cytometric analysis (CD 271-FITC
and CD 271-PE; Miltenyi Biotec GmbH).
All cell cultures were incubated with the respective supple-
ments at 37
C, 5% CO
2
in a humidified atmosphere. In a standar-
dized fashion, all nonadherent cells were removed 24 hours after
initial plating by media changes. The cells were cultured with
media changed twice weekly until reaching confluence of 70-
80%. At this time, cells were passaged using 1
 trypsin-EDTA
(PAA). At each passage (p), cells were replated at a standard
density of 200 cells per cm
2
at any subsequent passage.
Proliferation Kinetics
Cells were passaged and counted once they reached a subconflu-
ence of 70-80%. The population doubling (PD) rate was deter-
mined using the following formula [31]:
X ¼
½log10ðN
H
Þ À log10ðN
1
Þ
log10
ð2Þ
N
H
is the harvested cell number and
N
1
is the plated cell
number. The PD for each passage was calculated and added to
the PD of the previous passages to generate data for cumulative
population doublings (CPD).
In addition, the generation time (average time between two
cells doublings) of four BM within all media conditions was cal-
culated at passage 1 (p1) and p4 using the following formula:
X ¼
log2
 Dt
log
ðN
H
Þ À logðN
1
Þ
The effects of heparin and thrombin, which are present in
tPRP and pHPL, were checked separately. BM-MSCs of two
donors were cultured in (a) 10% FBS, (b) 10% FBS
þ 2 U hepa-
rin/ml, or (c) 10% FBS
þ 1 U thrombin/ml for three passages.
No impact on MSC growth kinetics was observed within the
three passages.
Colony-Forming Unit-Fibroblast Assays
The colony-forming unit-fibroblast (CFU-F) assay in primary cul-
ture was determined for six donor BM, and colonies were
Figure 1.
Morphology of bone marrow
(BM)-mesenchymal stromal cells (MSCs)
and BM-MSC allocation to specific tests.
(A): Photomicrographs of one representa-
tive donor at primary culture at day 10 for
fetal bovine serum human serum, thrombin-
activated platelet releasate in plasma, and
pooled human platelet lysate are shown in
rows. Columns reflect cells either isolated
using Ficoll density centrifugation, deple-
tion of lineage positive cells by RosetteSep,
or CD271 selection followed by plastic ad-
hesion.
Magnification,
Â100. (B): The
scheme shows the total number BM sam-
ples and those used for the respective paral-
lel tests. Numbers of BM samples were
reduced in subsequent passages because of
replicative senescence-induced growth re-
tardation. Abbreviations: FBS, fetal bovine
serum; HS, pooled human serum; tPRP,
pooled
thrombin-activated
platelet-rich-
plasma;
pHPL,
pooled
human
platelet
lysate.
Bieback, Hecker, Kocao¨mer et al.
2333
www.StemCells.com
counted after 10 days. Freshly isolated BM-MNCs derived from
the three different isolation methods and cultured within the four
different media were plated in duplicate in 6-well plates at den-
sities of 1
 10
4
, 5
 10
4
, and 1
 10
5
per well. On day 10, the
cell layer was fixed with methanol and stained with Giemsa solu-
tion (Merck, Darmstadt, Germany, http://www.merck.de). Individ-
ual colonies composed of at least 50 cells were counted. CFU-F
frequency was calculated based on the respective input cell num-
ber as CFU-F per 1
 10
4
MNCs.
In Vitro Differentiation Potential
The adipogenic and osteogenic differentiation capacity of MSCs
was assessed at p2/p3 for all BM donors and for all culture con-
ditions [27]. To detect the osteogenic differentiation, cells were
stained for calcium deposition using von Kossa stain. Adipogenic
differentiation was indicated by the morphological appearance of
lipid droplets stained with Oil Red O.
Flow Cytometry Analysis
Immunophenotypic analyses were performed on three BM-MSC
batches for all supplements and selected MSC samples derived
from the different isolation methods at p3.
The following mouse anti-human antibodies were used in
multiplexed flow cytometric analysis: CD105-FITC (clone 8E11;
Chemicon/Millipore, Schwalbach/TS, Germany, http://www.milli-
pore.com), CD144-PE (TEA1/31; Beckman Coulter GmbH,
Krefeld, Germany, http://www.beckman.com), CD90-APC (5E10;
Becton Dickinson GmbH, Heidelberg, Germany, http://www.
bdeurope.com),
CD106-FITC
(51-10C9;
Becton
Dickinson
GmbH), CD146-PE (TEA-1/34; Beckman Coulter), CD34-PerCP-
Cy5.5 (8G12; Becton Dickinson GmbH), CD133/1-APC (AC133;
Miltenyi), CD44-APC-Alexa750 (IM7; NatuTec, Frankfurt/Main,
Germany, http://www.natutec.de), CD15-FITC (HI98; Becton
Dickinson
GmbH),
CD45-FITC
(HI30;
Becton
Dickinson
GmbH),
CD3-FITC
(UCHT1;
Becton
Dickinson
GmbH),
CD235a-FITC (GA-R2; Becton Dickinson GmbH), CD14-FITC
(M5E2; Becton Dickinson GmbH), CD19-FITC (AE1; Diatec/
Dianova Hamburg, Germany, http://www.dianova.de), CD117-PE
(104D2; Becton Dickinson GmbH), CD33-PerCP-Cy5.5 (P67.6;
Becton Dickinson GmbH), CD31-APC (WM59; NatuTec), CD29-
APC-Cy7 (TS2/16 BioLegend/Biozol Eching b, Mu¨nchen, Ger-
many, http://www.biozol.de), CD73-PE (AD2; Becton Dickinson
GmbH), HLA-ABC-APC (G46-2.6; Becton Dickinson GmbH),
HLA-DR-PE-Cy7 (L243; Becton Dickinson GmbH), and 7-AAD
(Beckman Coulter) for dead cell exclusion. The samples were an-
alyzed using the BD FACS-Canto II and DIVA software. Com-
parative analysis was performed with FlowJo Version 7.2.5 (Tree
Star, Inc., Ashland, OR, USA, http://www.treestar.com).
Inhibition of Phytohemagglutin-Induced
T-Cell Proliferation
Dostları ilə paylaş: |