To test the allostimulatory or inhibitory effect of MSCs on T-cell
proliferation, we co-cultured phytohemagglutin (PHA)-stimulated
MNCs derived from healthy human buffy coat preparations with
decreasing numbers of MSCs. Specifically, 1
 10
4
, 5
 10
3
, and
2.5
 10
3
MSCs were preseeded as quadruplicates into wells of a
flat-bottom 96-well plate in Roswell Park Memorial Institute
1640 (PAA) medium containing 10% FBS. On the following day,
the medium was discarded, and 1
 10
5
MNCs were added to each
well in RPMI medium containing 10% FBS and interleukin-2 (IL-
2; 20 U/ml
¼ 0.01 lg/ml; Roche Applied Science, Mannheim,
Germany, http://www.roche-applied-science.com). One half of the
wells were stimulated with PHA (2.5
lg/ml PHA-L; Roche
Applied Science) to induce T-cell proliferation. Controls included
nonstimulated, co-cultured MNCs, nonstimulated MNCs, and
stimulated MNCs without MSC co-culture. To simultaneously
quantify the cell number, viability, phenotype, and activation level
of T-cell subsets, we used a modified method established by
Nguyen et al. [32]. CD3, CD4, and CD8 antibodies were used to
distinguish T-cell subsets, whereas T-cell activation was measured
by the assessment of CD71 expression. For T-cell quantification,
fluorescent microparticles with defined concentrations were used.
The absolute count of target cells was calculated on the basis of
the known bead concentrations using the following equation:
Total cells per microliter
¼
Numbers of cells measured
Number of fluorespheres measured
 Flow count concentration:
After 5 days, the MNCs were harvested and stained with a
cocktail of the following reagents: flow-count fluorospheres
(Beckmann Coulter) to directly determine absolute cell counts,
anti-CD3-PE-Cy7 (Becton Dickinson GmbH), anti-CD4-PE (Bec-
ton
Dickinson
GmbH),
anti-CD8-FITC
(Becton
Dickinson
GmbH), anti-CD71-APC (Becton Dickinson GmbH), and 7-AAD
(Beckmann Coulter) to exclude dead cells. The samples were an-
alyzed using the FACS-Canto II and DIVA software. The per-
centage of inhibition of T cells was calculated by comparing con-
trol cultures stimulated with PHA in the absence of MSCs (
¼0%
inhibition) to those in the presence of MSCs.
We used BM-MSC samples from three donors cultured in the
differing supplements. Because of the fact that, in two HS-MSC
cultures, T cells showed reduced viability (
<70%) in unstimu-
lated co-cultures because of unknown reasons, data for HS-MSCs
were not statistically evaluated.
Detection of Telomerase Activity
To detect potential telomerase activity, we used the Telo TAGGG
Telomerase PCR ELISA (Roche Applied Science) following the
manufacturer’s instructions. Samples of BM aspirates and BM-
MSCs at different passages cultured in the various supplements
were analyzed. Samples were regarded as telomerase negative if
the difference in absorbance after subtraction of the negative con-
trol was
<0.2.
Human Cytokine Expression Profile
The cytokine profile of culture medium supplemented with 10%
FBS, HS, tPRP, or pHPL and that of 3-day MSC-conditioned me-
dium was analyzed with a semiquantitative human cytokine anti-
body array that can detect 174 cytokines per experiment (RayBio
Human Cytokine Antibody Array G series 2000; Tebu-bio
GmbH, Berlin, Germany, http://www.tebu-bio.com). To minimize
variances, tPRP and pHPL were derived from the same initial
platelet concentrates.
Despite the human specificity of the array, we also tested
FBS, but finally interpreted only the conditioned medium. All
sample measurements were performed in duplicate according to
the manufacturer’s instructions. The signals were detected using a
laser scanner (GMS 418 array scanner; Affymetrix, Santa Clara,
CA, USA, http://www.affymetrix.com) and analyzed with array
vision version 7 (Imaging Research, Inc., St. Catharines, Canada,
http://www.imagingresearch.com). Signals were normalized using
positive, negative and internal controls included on the array. For
analysis, the internal negative controls were used to determine
the cut-off rate for a positive signal as 2
 SD. Thus, signal in-
tensity values of
>2,000 were regarded as positive.
Statistical Analysis
Statistical tests were performed using SPSS 12.0 (SPSS, Inc.,
Chicago, IL, USA, http://www.spss.com) or SigmaPlot 11.0
(Systat Software, Inc., San Jose, CA, USA, http://www.systat.
com) statistical software. Data are represented as arithmetic mean
Æ SD. Data were tested for normality and equal variance before
analysis. Statistical differences were calculated using analysis of
variance (ANOVA; or ANOVA on ranks if equal variance testing
failed) and
t test (paired t test where applicable). Differences
were considered significant at *,
p < .05 or **, p < .01.
2334
Human Alternatives to FBS for BM-MSC Expansion
R
ESULTS
MSCs Isolation and Expansion
Effect of Isolation Strategies. Because our previous experi-
ments with AT-MSCs resulted in a transient contamination with
hematopoietic cells in human supplement cultures, we applied
two MSC enrichment strategies. First, we depleted mature line-
age marker expressing cells by rosetting to erythrocytes (Roset-
teSep). Second, we used magnetic combined with flow cytomet-
ric sorting of CD271-expressing cells to enrich MSCs.
Both enrichment strategies reduced contaminating round
and loosely adherent cells in the primary passage (Fig. 1).
Ficoll gradient-derived cultures supplemented with FBS had
experientially few contaminating cells indicated by the pres-
ence of small loosely adherent round cells reactive with anti-
CD45 (data not shown). These cells were easily depleted by
repetitive media changes that occurred in the primary culture.
RosetteSep very efficiently depleted the round contaminat-
ing cells in all culture conditions and yielded MSCs to be pas-
saged 11.2
Æ 1.48 days after seeding compared with 15.42 Æ
4.46 days after seeding for Ficoll/FBS, respectively (
p ¼ .01;
Fig. 2). A more rapid proliferation was observed by p3. How-
ever, this did not correlate with higher cumulative population
doublings. Up to p3 expansion kinetics of RosetteSep-
enriched cells showed a higher proliferation. Interestingly,
immunodepleted cells from some donors showed an earlier
onset of replicative senescence compared with Ficoll-isolated
cells from p4 on, indicated by reduced proliferation and mor-
phologic changes. We observed differences in the effects of
HS and tPRP on RosetteSep-enriched cultures. The cell incre-
ment in HS exceeded that of ficolled cells up to p4 (Fig. 2).
Proliferation rates were accelerated in HS, with a significant
increase only in p1.
FBS-supplemented cultures sorted for CD271
þ
cells
showed bacterial contamination in three of four cases. This
necessitated the abandonment of said cultures. Despite the
addition of the same concentration of penicillin/streptomycin,
parallel cultures using the human alternatives displayed no
bacterial outgrowth. This may suggest that human FBS alter-
natives might have intrinsic antibacterial components.
In summary, the experiments with CD271
þ
cells cultured
in FBS, HS, or tPRP showed that CD271
þ
cells tended to
grow in colonies during the entire culture period, never form-
ing confluent monolayers. Expansion kinetics were delayed in
FBS-driven cultures of CD271
þ
cells after p3 (Fig. 2). This
corresponded to the low cell numbers yielded within each
passage. Cells stopped proliferation in p5 yielding maximum
24.91 CPD for FBS (
n ¼ 1; 17.47 for HS, p4, n ¼ 1 and
21.92 for tPRP, p5,
n ¼ 1).
Effect of Supplements. As indicated above, the culturing of
MNCs after density gradient centrifugation in FBS-supple-
mented medium yielded few contaminating hematopoietic
cells. In contrast, supplementing MNCs with HS or tPRP
resulted in variably high numbers of hematopoietic cells.
Interestingly, and unlike HS and tPRP, pHPL-supplemented
cultures were devoid of contaminating cells (Fig. 1).
Calculating the number of CFU-Fs showed a precursor
frequency of 1:25,000 MSCs/MNCs, which was not affected
Figure 2.
Effect of isolation strategies. (Top) Mean cumulative population doublings of bone marrow-mesenchymal stromal cells isolated using
Ficoll gradient centrifugation, RosetteSep, or CD271 sorting followed by plastic adhesion in medium supplemented with fetal bovine serum,
human serum, or pooled thrombin-activated platelet-rich-plasma. (Bottom) Cumulative days needed for each MSC culture to be passaged. Low
numerical values indicate high proliferative activity (for initial
n, see Fig. 1B). *In comparison to Ficoll, p < .05 using analysis of variance and
paired
t test. Abbreviations: FBS, fetal bovine serum; HS, pooled human serum; tPRP, pooled thrombin-activated platelet-rich-plasma.
Bieback, Hecker, Kocao¨mer et al.
2335
www.StemCells.com
by the use of different culture substitutes. However, the num-
ber of cells composing the colonies was larger in all of the
human supplements. Colonies in pHPL were densely packed
with very small spindle-shaped cells compared with only
loosely connected cells in FBS cultures (Fig. 3).
The comparison of the expansion rates of MSCs in FBS to
either HS- or tPRP-supplemented culture conditions showed no
significant differences (Fig. 4). However, cells cultured in HS
and tPRP decelerated proliferation from p4, reaching 14.46
Æ
3.46 (HS,
n ¼ 3) and 18.47 Æ 2.92 CPD (tPRP, n ¼ 7) com-
pared with 18.73
Æ 1.96 CPD (FBS, n ¼ 13). This correlated
well with an increased population doubling time. Beginning in
p1, the generation time of ficolled MSCs cultured in HS was
significantly prolonged with 3.41
Æ 1.23 days compared with
2.51
Æ 0.87 days in the FBS cultures. In p4, both HS and tPRP
showed an extended generation time (12.53
Æ 6.54 days for HS
and 12.72
Æ 7.39 days for tPRP) compared with FBS (3.72 Æ
0.44 days). Cells from one donor (HS) or two donors (tPRP)
expanded until p5. In contrast, 11 samples from a total of 14
donors cultured in FBS reached p5.
Consistently, cultures supplemented with pHPL yielded sig-
nificantly higher expansion rates than cells in FBS, reaching a
maximum of 52.82 CPD (in p8;
n ¼ 1 from initially six donors)
compared with 31.43
Æ 3.13 in FBS (in p7; n ¼ 8 from initially
14 donors,
p ¼ .004). Calculating the generation time at p1 and
p4 yielded, in both cases, significantly reduced population dou-
bling times: 1.27
Æ 0.23 days in p1 and 1.9 Æ 0.32 days in p4
compared with FBS (2.51
Æ 0.87 days in p1 and 3.72 Æ 0.44
days in p4;
p ¼ .012 and p ¼ .004, respectively).
MSC Quality and Functionality
Immune Phenotype. Typical CD44, CD73, CD90, CD105,
CD146, and HLA-ABC surface marker expression was
detected in all MSC cultures at p3 despite a measurable donor
variance. CD29 was expressed on 44.46
Æ 9.49% of BM-
MSCs cultured in tPRP, whereas the other supplement yielded
significantly higher positivity; for example, in HS, 98.12
Æ
0.76%. CD29 mean fluorescence intensity was significantly
higher in HS (FBS, 967.38
Æ 476.55; HS, 2,176.07 Æ 416.98;
tPRP, 134.57
Æ 19.18; pHPL, 438.88 Æ 306.06). CD15,
CD33, lineage (CD45, CD3, CD235a, CD14, and CD19),
CD117, CD144, and HLA-DR showed less than 5% positiv-
ity. Selected antigens representing one donor are depicted in
Figure 5. No further statistically significant differences
between FBS and the other supplements were detected.
Figure 3.
Colony-forming unit-fibroblast of bone marrow (BM)-mes-
enchymal stromal cells (MSCs). Photomicrographs represent BM-
MSCs from one donor assessed after 10 days cultivated in fetal bovine
serum, pooled human serum, pooled thrombin-activated platelet-rich-
plasma, or pooled human platelet lysate (magnification,
Â100). Abbre-
viations: FBS, fetal bovine serum; HS, pooled human serum; tPRP,
pooled thrombin-activated platelet-rich-plasma; pHPL, pooled human
platelet lysate.
Figure 4.
Effect of supplements. (Top) Mean cumulative population
doublings of bone marrow-mesenchymal stromal cells (MSCs) iso-
lated using Ficoll gradient centrifugation and cultivation either in fetal
bovine serum (FBS), pooled human serum, pooled thrombin-activated
platelet-rich-plasma, or pooled human platelet lysate. (Middle) Cumu-
lative days needed for each MSC culture to be passaged (for initial
n,
see Fig. 1B). (Bottom) Generation time of MSCs at p1 (black) and p4
(white) (
n ¼ 4). *p < .05 and **p < .01 in comparison to FBS using
analysis of variance (ANOVA) or ANOVA on the ranks, respectively.
Abbreviations: FBS, fetal bovine serum; HS, pooled human serum;
tPRP, pooled thrombin-activated platelet-rich-plasma; pHPL, pooled
human platelet lysate; CPD, cumulative population doublings.
2336
Human Alternatives to FBS for BM-MSC Expansion
Other markers showed donor-dependent variable reactivity
perhaps influenced by the supplement used and/or the degree
of hematopoietic cell contamination: CD31 (FBS, 0.95
Æ 0.74;
HS, 0.71
Æ 0.31; tPRP, 5.5 Æ 4.48; pHPL, 19.12 Æ 10.78),
CD133 (FBS, 6.45
Æ 11.23; HS, 5.14 Æ 7.21; tPRP, 6.04 Æ
5.94; pHPL, 0.77
Æ 0. 89), and CD106 (FBS, 6.18 Æ 7.91; HS,
8.35
Æ 9.7; tPRP, 13.54 Æ 14.34; pHPL, 18.28 Æ 9.41).
Differentiation Potential. MSCs derived from all conditions
demonstrated differentiation toward the osteogenic and adipo-
genic lineage as assessed by von Kossa and Oil Red O stain-
ing (supporting information Fig. 1).
Inhibition of PHA-Induced T-Cell Proliferation
We used a flow cytometric method to simultaneously quantify
mitogen-driven T-cell proliferation, subtypes, activation level,
and viability. T-cell stimulation by PHA led to strong prolif-
eration and activation. To study the impact of culture condi-
tions on MSC inhibitory activity, the same donor MNCs were
used for all MSC samples. In co-culture controls without add-
ing PHA, MSCs did not induce an alloreaction of the T cells
but rather a loss of T cells in the range of 10-20% compared
with the control. All MSCs independent of the culture condi-
tions inhibited the PHA-driven T-cell proliferation and activa-
tion dose dependently (Fig. 6). Both CD4
þ
and CD8
þ
T-cell
subsets were similarly affected. MSCs cultured in the human
platelet-derived substitutes showed a tendency toward aggra-
vated inhibitory activity at ratios of 1:10 and 1:20 that was
not statistically significant.
Telomerase Activity
Telomerase activity was analyzed in MSCs at different pas-
sages to control the onset of spontaneous immortalization. We
never detected telomerase activity except for the primary BM.
Figure 5.
Flow cytometric char-
acterization of bone marrow (BM)-
mesenchymal stromal cells (MSCs).
Comparison of the expression of
surface proteins of ficolled BM-
MSCs
cultured
in
fetal
bovine
serum,
pooled
human
serum,
pooled thrombin-activated platelet-
rich-plasma, or pooled human plate-
let lysate analyzed by flow cytome-
try. One representative donor and
typical
MSC
marker
expression
are depicted in the overlay to the
unstained/control. For statistical an-
alyses,
n ¼ 3 BM-MSC donors
were paired assessed at passage 3.
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.
2337
www.StemCells.com
Here we could attribute the low telomerase activity to the
CD34
þ
proportion (data not shown).
Cytokine Content in Supplements
and Conditioned Medium
The cytokine content in the 10% supplemented media and the
conditioned media (CM) after 3 days of culture was eval-
uated. For these analyses, tPRP and pHPL were derived from
the same pools to eliminate donor-specific differences.
Because the cytokine array is human specific, data for the
FBS-containing medium were not interpreted (Fig. 7; support-
ing information Table 1).
Overall growth factor levels in FBS-CM were lower than
in any of the human supplemented cultures, indicating that
detectable levels are continuously present in the human sup-
plements and remain unchanged including Acrp30 (adiponec-
tin), angiogenin, CD14, glucocorticoid-induced tumor necrosis
factor receptor (GITR), platelet-derived growth factor (PDGF)
AB, and sgp130 (soluble gp130). Other cytokine levels
dropped during culture (because of consumption or degrada-
tion) such as epidermal growth factor, macrophage-derived
chemokine/CCL22, pulmonary and activation-regulated che-
mokine, PDGF-AA, and PDGF-BB. Insulin-like growth factor
binding protein (IGFBP)-3, interleukin 6, monocyte chemoat-
tractant protein-1, macrophage stimulating protein (MSP)
a,
osteoprotegerin, thrombopoietin, and tissue inhibitor of metal-
loproteinases-1 and -2 levels increased during culture, pre-
sumably because of production by MSCs.
tPRP differed from HS and pHPL with regard to a variety
of cytokines. In tPRP-CM, hepatocyte growth factor/scatter
factor, IGFBP-2, and vascular endothelial growth factor
(VEGF) D were elevated. Unfortunately, no obvious candidate
for the strong proliferative support from pHPL could be iden-
tified: basic fibroblast growth factor (bFGF), GITR, macro-
phage migration inhibitory factor (MIF), macrophage inflam-
matory protein-1
b (MIP-1b), MSP-a, regulated on activation,
normal T-cell expressed, and secreted (RANTES; CCL-5),
and VEGF were differentially regulated in pHPL/pHPL-CM
compared with HS/HS-CM and tPRP/tPRP-CM.
Figure
6.
Immunomodulatory
capacity of bone marrow (BM)-
mesenchymal
stromal
cells
(MSCs). BM-MSCs, irrespective
of the supplement, mediated a
dose-dependent
inhibition
of
phytohemagglutin-induced T-cell/
CD3 stimulation. Proliferation of
the CD4 (T-helper) and CD8
(cytotoxic) subsets were simi-
larly
affected.
Simultaneously,
we quantified the proportion of
activated T cells by means of
CD71 expression. Like T-cell
proliferation,
T-cell
activation
was inhibited dose dependently.
The
overlay
depicts
CD71
expression of CD3
þ
cells. The
same buffy coat mononuclear
cells were used for all experi-
ments. Three MSC batches were
paired assessed at passage 3.
Dose dependent differences were
observed compared to ratio 1:10
with
p < .05 (* ¼ fetal bovine
serum (FBS), #
¼ thrombin-acti-
vated platelet releasate in plasma
(
þPRP) and $ ¼ pooled human
platelety lysate (pHPL)). tPRP-
and pHPL MSCs-induced CD3
(plus CD4 for tPRP MSCs and
CD8 for pHPL MSCs) inhibition
did not differ significantly from
that at 1:10. Statistically reduced
inhibition was found at the ratio
1:40. A paired
t test was used to
compare dose dependency and
analysis of variance to compare
culture supplements.
2338
Human Alternatives to FBS for BM-MSC Expansion
D
ISCUSSION
Currently, the ex vivo expansion of MSCs seems to be inevi-
tably to get the common therapeutic dose of
>2 Â 10
6
/kg
body weight for infusion (e.g., in treatment of graft vs. host
disease). Also, for other indications, there exists a need to
study the applicability of MSCs with dose escalation, indicat-
ing the need to propagate MSCs in sufficient quantity.
In a recent concise review in this
Journal, Manello and
Tonti underlined that elaboration of a culture medium for the
production of MSCs for clinical application still remains a
crucial matter [12]. We and others [2-4, 14-24, 27, 33-38]
have since developed various protocols for the clinical scale
propagation of human MSCs. Most of these protocols actually
avoid the use of animal serum and some get rid off antibiotics
and density gradient separation of the culture initiating cells.
The major limitation of these studies relates to the fact that
they compare FBS-based media to only selected FBS-free cul-
ture conditions (only HS or only pHPL). There are some data
indicating that autologous serum in general supports greater
amplification of MSCs than FBS [14, 39]. Limited availability
and high variability regarding MSC growth clearly hamper
the clinical applicability of autologous serum for large-scale
MSCs production. Pooled preparations of allogeneic human
serum can be produced in large amounts for pharmaceutical
manufacturing and are easily controlled for quality according
to blood banking standards (one batch
¼ 25 blood donors;
produced in the Institute of Clinical and Experimental Trans-
fusion Medicine, University Hospital Tu¨bingen, Tu¨bingen,
Dostları ilə paylaş: |