Mesenchymal stem cells impair in vivo T-cell priming
by dendritic cells
Sabrina Chiesaa,b,c,1, Silvia Morbellid, Sara Morandoa,b, Michela Massollod, Cecilia Marinie, Arinna Bertonia,b,
Francesco Frassonif, Soraya Tabera Bartolomég,h, Gianmario Sambucetid,e, Elisabetta Traggiaic,i,2,
and Antonio Uccellia,b,j,1,2
aDepartment of Neurosciences, Ophthalmology and Genetics,dDepartment of Nuclear Medicine andjCenter of Excellence for Biomedical Research,
University of Genoa, 16132 Genoa, Italy;bAdvanced Biotechnology Center, 16132 Genoa, Italy;c2nd Division of Pediatrics, Laboratory of Immunology
and Rheumatic Diseases, Istituto di Ricovero e Cura a Carattere Scientifico, Institute G. Gaslini Hospital, 16147 Genoa, Italy;eConsiglio Nazionale delle
Ricerche, Institute of Bioimages and Molecular Physiology, Milan, Section of Genoa, 16132 Genoa Italy;fStem Cell Centre and Cellular Therapy, San Martino
Hospital, 16132 Genoa, Italy;gInstitute of Neuroscience of Castilla y León, University of Salamanca, 37007 Salamanca, Spain;hCentro en Red de Medicina
Regenerativa y Terapia Celular de Castilla y Leon, University of Salamanca, 37007 Salamanca, Spain; andiNovartis Institute for Biomedical Research,
Translational Medicine, CH4057 Basel, Switzerland
Edited* by Darwin J. Prockop, Texas A & M Health Science Center, Temple, TX, and approved September 1, 2011 (received for review March 9, 2011)
Dendritic cells (DC) are highly specialized antigen-presenting cells
characterized by the ability to prime T-cell responses. Mesenchymal
stem cells (MSC) are adult stromal progenitor cells displaying
immunomodulatory activities including inhibition of DC maturation
in vitro. However, the specific impact of MSC on DC functions, upon
in vivo administration, has never been elucidated. Here we show
that murineMSC impairToll-likereceptor-4induced activation ofDC
molecules involved in the migration to the lymph nodes, antigen
presentation to CD4+T cells, and cross-presentation to CD8+T cells.
These effects are associated with the inhibition of phosphorylation
of intracellular mitogen-activated protein kinases. Intravenous ad-
ministration of MSC decreased the number of CCR7 and CD49dβ1
local antigen priming of DO11.10 ovalbumin-specific CD4+T cells.
Upon labeling of DC with technetium-99m hexamethylpropylene
amine oxime tofollow their invivobiodistribution, we demonstrated
that intravenous injection of MSC blocks, almost instantaneously,
the migration of subcutaneously administered ovalbumin-pulsed DC
to the draining lymph nodes. These findings indicate that MSC signif-
icantly affect DC ability to prime T cells in vivo because of their in-
ability to home tothe draininglymph nodes andfurther confirm MSC
potentiality as therapy for immune-mediated diseases.
enchymal stem cells (MSC) or multipotent mesenchymal
stromal cells, area heterogeneous population ofself-renewingand
multipotent cells isolated from the bone marrow (BM) (1). MSC
raised hopes for their clinical exploitation for tissue-repair strat-
egies and increasing experimental evidence supports their use also
for immune-mediated diseases (2). In fact, MSC display a striking
capacity of modulating the immune response (3). Despite a large
on immune cells, little is known about the mechanisms of MSC-
mediated inhibition of the in vivo immune response. Dendritic
cells (DC) are unique antigen-presenting cells (APCs) endowed
with the ability of acquiring and processing antigens, up-regulating
costimulatory molecules and therefore priming naive T cells. To
present antigens to naive T cells, CCR7-expressing DC must mi-
grate through lymphatic vessels from sites of inflammation to
the closest draining lymph node (4). Activation of DC via Toll-like
receptors (TLRs) up-regulates the expression of chemokine
receptors involved in DC migration to the lymph nodes and
enhances their in vivo mobilization properties (5). As a conse-
quence, the total number of DC migrating in the draining lymph
nodes deeply affects naive T-cell priming (6). Here we show that
murine MSC inhibit in vitro DC effector properties, including
antigen processing and presentation to T cells through the in-
hibition of the activation of mitogen-activated protein kinases
tromal progenitors of mesodermal cells, referred to as mes-
(MAPKs) occurring upon TLR4 stimulation. Most important, we
report that in vitro exposure of DC to MSC as well as in vivo in-
travenous administration of MSC results in a significant down-
regulation of CCR7 and CD49dβ1, two molecules involved in DC
homing to lymphoid organs. This event leads to inhibition of mi-
priming of antigen-specific naive T cells.
MSC Prevent Maturation of BM-Derived DC. To evaluatetheeffectof
MSConTLR4-mediated activation ofDC, wecoculturedDCwith
MSC at a 1:3 ratio during LPS-induced maturation (MSC-condi-
tioned DC). MSC-conditioned DC showed decreased levels of the
pan-DC and activation surface markers CD11c, CD86, CD80,
CD40, MHC class I, and MHC class II compared with DC acti-
vated without MSC (ctr-DC), suggesting a significant impairment
of DC maturation (Fig. S1). Moreover, MSC impaired the ability
of LPS to induce the up-regulation of both CCR7 and CD49dβ1,
two molecules crucial for DC migration to lymph nodes (Fig. S1).
We also observed a consistent impairment of IL-12 intracellular
production by CD83+MSC-conditioned DC compared with ctr-
DC (Fig. 1A). Moreover, MSC exposure reduced levels of TNF-α
and IL-10in thesupernatantsofLPS-activated DC comparedwith
controls (Fig. 1B). The significant inhibition of cytokine pro-
duction by DC activated by LPS in the presence of MSC was
confirmed by real-time PCR (Fig. S2). We also observed a signifi-
cant decrease, in MSC-conditioned DC, of the expression of
indoleamine 2,3-dioxygenase (IDO) (Fig. S2).
Because other groups have demonstrated that mouse MSC
express the full array of TLRs (7, 8), we sought to analyze by
quantitative real-time PCR the expression of TLRs from 1 to 9
on the MSC used in our experimental conditions. Similar to what
has been previously reported, MSC express all TLRs at com-
parable levels, except for TLR8 and -9, which display much
lower expression (Fig. 2A, Left). Interestingly, the exposure of
MSC to LPS for 12 h did not significantly affect MSC TLR ex-
pression (Fig. 2A). Next, we addressed the possibility that LPS
may affect the behavior of TLR4-bearing MSC during the co-
culture with DC. We repeated the experiments exposing DC to
the supernatants from MSC cultured with or without LPS, thus
Author contributions: S.C., F.F., G.S., E.T., and A.U. designed research; S.C., S. Morbelli,
S. Morando, M.M., C.M., A.B., and S.T.B. performed research; S.C., S. Morbelli, C.M., G.S.,
E.T., and A.U. analyzed data; and S.C., E.T., and A.U. wrote the paper.
The authors declare no conflict of interest.
*This Direct Submission article had a prearranged editor.
1To whom correspondence may be addressed. E-mail: firstname.lastname@example.org or
2E.T. and A.U. contributed equally to this work.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| October 18, 2011
| vol. 108
| no. 42 www.pnas.org/cgi/doi/10.1073/pnas.1103650108
avoiding direct contact of the two cell types. We observed that
supernatants of LPS-stimulated MSC were able to inhibit IL-
12p70, TNF-α, and IL-10 secretion by DC (Fig. 2B), similar to
what we observed in DC cultured in the presence of MSC, thus
suggesting that the inhibition of the release of cytokines by DC is
not because of competition for LPS by the two cell types. These
findings strongly support the role of soluble factors as mediators
of such inhibition. Of note, supernatants from MSC not stimu-
lated in vitro with LPS exhibited a lower ability to inhibit cy-
tokines release by activated DC (Fig. 2B), confirming that
inflammatory cues license MSC immunomodulatory activity
(9, 10). Taken together, these findings demonstrate that MSC
impair LPS-induced activation of DC through the release of
MSC Down-Modulate Key Molecules Involved in DC TLR4 Signaling.
To further investigate the effect of MSC on DC activation by
LPS, we analyzed gene-expression profiles of DC activated by
LPS in the presence or absence of MSC through the use of a
mouse Toll-Like Receptor Signaling Pathway PCR Array (see
the Gene Table list at SABiosciences, available at http://sabio-
sciences.com). Among the genes analyzed, 5 of 84 genes were
significantly modified in MSC-conditioned DC (Table S1). Re-
markably, myeloid differentiation primary response gene 88
(MyD88), MAP2K3, and NF-κB1 genes were significantly down-
modulated (>4-fold) in MSC-conditioned DC compared with
ctr-DC (Fig. 1C). We next studied their effect on two members
of the MAPK superfamily, the extracellular signal-regulated
kinases (ERK1/2) and the p38 MAPKinase (p38). As shown in
Fig. 1D, we observed a time-dependent decrease in phosphory-
lated ERK1/2 and p38 being higher 30 min after LPS activation.
Accordingly, the ratio between phosporylated p38 and unphos-
phorylated p38 in control DC increases over time. In contrast, in
MSC-conditioned DC this ratio did not change upon prolonged
exposure to LPS as the result of the impairment of p38 phos-
phorylation (Fig. 1E). We addressed the effect of MSC on Akt,
a serine/threonine protein kinase involved in the inhibition of
apoptosis. MSC increased phophorylation of Akt compared with
ctr-DC mainly at 60 min (Fig. 1D). These findings confirm that
MSC inhibit the MAPK cascade involved in the pathway leading
to the phosphorylation of cytoplasmic components and nuclear
transcription factors that promote cytokines production by
mouse DC following TLR4 stimulation. On the other side, they
can promote cell survival acting on the Akt pathway.
MSC Impair Class I-Associated Antigen Processing in DC and Cross-
Presentation to CD8+T Cells. As MSC induced down-regulation of
class I molecules, we investigated their effects on class I re-
stricted cross-presentation of ovalbumin (OVA) to CD8+T cells
by DC. After LPS exposure, ctr-DC and MSC-conditioned DC
were pulsed with OVA and cultured with 5,6-carboxyfluorescein
diacetate succinimidyl ester (CFSE)-labeled OT-I CD8+T cells
(ratio 1:5). MSC-conditioned DC failed to induce the expression
of the early activation marker CD69 on CD8+T cells compared
with control DC (Fig. 3A). Furthermore, T-cell proliferation was
strongly inhibited by OVA-pulsed MSC-conditioned DC com-
pared with OVA-pulsed ctr-DC as depicted by cell division
profile on CFSE labeled cells analyzed at 48 and 72 h (Fig. 3B).
We also measured cell death in the CD8+subset by propidium
iodide staining and we did not observe any significant increase of
dead cells. Thus, MSC impair antigen cross-presentation by DC.
The processing and presentation of HLA class I antigen-
derived peptides is accomplished through a complex series of
intracellular events involving constituents of the constitutive
proteasome, as well of the IFN-γ–inducible proteasome (im-
munoproteasome). Thus, we asked whether the inhibition of DC
cross-presentation by MSC to CD8+T cells associated with an
impairment of some of the constituents of the intracellular an-
tigen processing machinery (APM). As shown in Fig. 3C, DC
activated in the presence of MSC showed a striking down-reg-
ulation of MB-1 and LMP-10 but not of Delta. These results
suggest that the impaired capacity of MSC-treated DC to prime
CD8+antigen-specific T cells leads to a deranged antigen pro-
cessing involving both proteasome- and immunoproteasome-
associated molecules in the cytosol compartment.
MSC Inhibit DC Ability to Prime CD4+T Cells. To evaluate the in-
fluence of MSC on MHC class II restricted antigen presentation,
we used naive DO11.10 CD4+T cells carrying a transgenic T-cell
receptor (TCR) specific for the OVA peptide323-339. LPS-treated
ctr-DC and MSC-conditioned DC, pulsed with increasing doses
of OVA peptide, were cocultured with CFSE-labeled DO11.10
CD4+T cells. We observed that DC activated in the presence of
MSC showed an impaired capacity to prime CD4+pOVA323–339-
specific T cells in a dose-dependent manner (Fig. 3D). There-
fore, we established a mouse model based on the adoptive
expression and phosphorylation of kinases involved in the TLR4 pathway. (A)
Levels of intracellular IL-12 produced by CD83+control DC (gray bars) and DC
exposed to MSC during activation with 10 μg/mL of LPS overnight (black
bars; P = 0.033). Results are representative of five independent experiments
(mean ± SD). (B) TNF-α and IL-10 levels by ctr-DC (gray bars) and DC exposed
to MSC during LPS activation (black bars), were quantified by ELISA (**P =
0.007 and *P = 0.03, respectively). Data are expressed as picogram per mil-
liliter. Results are representative of four and seven independent experi-
ments, respectively (mean ± SD). (C) After TLR4 stimulation, DC exposed to
MSC (black bars) significantly down-modulated MyD88, MAP2k3, and NFkB1
compared with control DC (gray bars), as demonstrated using the RT2Pro-
filer PCR Array mouse Toll-Like Pathway. One of three independent
experiments is shown. Gene expression is represented as relative mRNA
amount (fold-induction) compared with the control sample. (D) Western
blots of intracellular Akt, ERK1/2, p38 proteins from ctr-DC and MSC-
conditioned DC, stimulated and not stimulated with LPS for 30 and 60 min,
are shown. Anti–β-tubulin was used as control. One of three independent
experiments are shown. (E) Histograms show the ratio between the acti-
vated phosphorylated p38 MAPK (Thr180/Tyr182) and total unphosphory-
lated p38 MAPK in ctr-DC (gray bars) and in MSC-conditioned DC (black bars)
after 0, 15, 30, and 45 min of LPS stimulation. Data shown are representative
of one of three independent experiments.
MSC inhibit LPS-induced cytokine production and modulate gene
Chiesa et al.PNAS
| October 18, 2011
| vol. 108
| no. 42
transfer of purified CFSE-labeled, OVA-specific naive T cells
from DO11.10 transgenic mice (H-2d) into syngeneic BALB/c
recipient mice (H-2d) followed, 24 h later, by subcutaneous im-
munization with LPS-activated, OVA-pulsed DC. Three hours
later, one group of animals was injected intravenously with PBS
and another group with 1.25 × 106MSC. On day 3, draining
lymph nodes of adoptively transferred mice from the two groups
were isolated and the size, cell number, phenotype, CFSE di-
lution, and cytokine secretion of injected transgenic CD4+T
cells were analyzed. Upon macroscopic inspection, we observed
a remarkable decrease in size of draining lymph node from mice
injected with MSC compared with the draining lymph node of
control mice, where a more prominent swelling was observed
(Fig. 4A). Notably, we observed a significant difference in size
between the draining and the controlateral lymph nodes (not
draining) in the control group, which was not observed in the
MSC-injected mice. Such difference in size was not exclusively a
result of the inhibition of local swelling, but mainly because of a
significant decrease of the draining lymph node cells number in
the MSC-injected mice compared with controls (cellular num-
ber: ctr-DC mice: 25 ± 3.49 vs. MSC-treated mice: 16 ± 3.29,
*P = 0.016) (Fig. 4B). Analysis of the lymphocyte cellular
component showed a significant decrease of absolute numbers
of CD4+T cells (*P < 0.05) (Fig. 4C). CD8+T cells,
CD4+CD25brightT cells, B220+B cells, B220−CD138+plasma
cells, and CD3−DX5+NK cells were decreased compared with
control mice, albeit not reaching a statistical significance (Fig.
4C). When we compared, by CFSE dilution, priming of OVA-
specific CD4+DO11.10 cells from draining lymph nodes, we
observed an impairment of proliferation in the treated mice
compared with controls (Fig. 4D). OVA-specific T cells from
MSC-treated mice displayed a remarkably decreased IL-2 and
TNF-α production with respect to control mice, but IFN-γ pro-
duction was unaltered (Fig. 4E). Taken together, these results
indicate that MSC intravenous administration leads to an im-
pairment of priming of antigen-specific naive T cells in the
draining lymph nodes.
MSC Inhibit DC Migration to Lymph Nodes. To study the effects of
MSC on the migratory properties of DC, we labeled LPS-acti-
vated OVA-pulsed DC with CFSE and injected them sub-
cutaneously in naive animals that previously received transgenic
DO11.10 naive T cells, as previously described. Following in-
travenous administration of MSC, we observed a significant de-
crease of the absolute number of CD11chighCFSE+DC in the
lymph nodes of MSC-treated animals compared with PBS-
injected controls, suggesting a reduced recruitment of activated
DC to the draining lymph nodes (Fig. 4F). More importantly, the
percentage, as well as the absolute number of CCR7highCFSE+
CD11c+DC, was significantly decreased in the lymph nodes
of MSC-injected animals compared with PBS-treated controls.
Similarly, the absolute number of CD49dβ1highDC was signifi-
cantly lower in MSC-treated animals compared with controls,
but the percentage was unaltered (Fig. 4G and Fig. S3). Overall,
these results provide strong experimental evidence that MSC are
able to in vivo affect the expression of surface molecules impli-
cated in DC migration to the lymph nodes.
To further confirm these data, we labeled DC with techne-
tium-99m hexamethylpropylene amine oxime (99mTc-HMPAO)
and followed their in vivo bio-distribution and trafficking by
scintigraphic imaging of their migratory pattern in the 60 min
following their administration. In the negative control group,
naive mice were subcutaneously immunized with 2 × 106LPS-
activated, OVA-pulsed DC. In the positive control group, in-
travenous injection of purified pOVA323–339-specific naive
DO11.10 T cells was followed, 24 h later, by subcutaneous im-
munization with LPS-activated, OVA-pulsed DC. We immu-
nized BALB/c mice, previously injected with CD4+DO11.10
duction by DC with MSC supernatants. (A) TLRs expression was assessed by
quantitative RT-PCR on unstimulated (white bar) and LPS-stimulated MSC
(10 μg/mL) (black bar). Mean ± SD from three independent experiments with
unstimulated and LPS-stimulated MSC is shown. To magnify the differences
among the expression of different TLRs, particularly in the case of TLR1,
TLR8, and TLR9, two different scales have been depicted in the left and right
graphs. (B) Secreted levels of IL12p70, IL-10, and TNF-α by ctr-DC (white bars),
DC exposed to MSC (black bars), DC exposed to supernatants of MSC (light
gray bar), and DC exposed to supernatants of LPS-stimulated MSC (dark gray
bar), were quantified by ELISA (IL-12p70: ***P = 2.30931E-05, ***P =
4.53805E-05, ***P = 2.18932E-05, respectively; TNF-α: **P = 0.004, **P =
0.008, ***P = 9,42621E-05, respectively; IL-10: ***P = 0.0003, ***P = 0.0005,
**P = 0.005, respectively). Data are expressed as picogram per milliliter.
Results are the mean of three independent experiments (mean ± SD).
Gene expression of TLRs in MSC and inhibition of cytokines pro-
(A) Percentage, at 48 h, of CD69+CFSE-gated CD8+T cells primed by ctr-DC
(Left) or MSC-conditioned DC (Right) (DC:CD8 = 1:5). (B) Proliferating CD8+T
cells primed by ctr-DC (Left) or MSC-conditioned DC (Right), as depicted by
CFSE dilution at 48 (Upper) and at 72 h (Lower); numbers indicate the ab-
solute values of cells recovered in standardized acquisitions. One of three
independent experiments is shown. (C) The expression of three APM-associ-
ated molecules, Delta, MB1, and LMP-10 on Raji cells (Top), control DC
(Middle), and MSC-conditioned DC (Lower) is depicted. One three in-
dependent experiments is shown. (D) Inhibition of naive CD4+OVA-specific
transgenic DO11.10 T cells cultured with ctr-DC (Left) or MSC-conditioned DC
(Right) pulsed with increasing concentrations of pOVA323–339(0.02 μM; 0.2
μM; 2 μM) (DC:CD4 = 1:8). Absolute numbers indicate the absolute values of
cells recovered in standardized acquisitions (shown at the top of the histo-
grams) as described in Materials and Methods. One of three independent
experiments is shown.
MSC interfere with antigen presentation to CD8+and CD4+T cells.
| www.pnas.org/cgi/doi/10.1073/pnas.1103650108 Chiesa et al.
transgenic T cells, either with LPS-activated, OVA-pulsed DC
in vitro exposed to MSC (in vitro MSC-conditioned DC group)
or with LPS-activated, OVA-pulsed DC followed by the in-
jection, 15 min later, with 1.25 × 106MSC (in vivo MSC-
conditioned DC group). As depicted in Fig. 5A, the number of
DC released by the site of injection in mice pretreated with
OVA-specific transgenic T cells (positive control) was signifi-
cantly higher than in the negative-control mice lacking Ag-spe-
cific T cells (**P < 0.004). Strikingly, in mice immunized with
in vitro MSC-conditioned DC, we observed a significantly de-
creased migration of DC from the site of injection (**P < 0.004
compared with the positive-control group, respectively). In-
travenous MSC administration virtually blocked the escape of
OVA-pulsed DC (in vivo MSC-conditioned DC group) from the
injection site demonstrated by the sudden change of the slope of
the curve within few minutes from MSC injection (Fig. 5A,
closed square line). The relevance and the rapidity of this effect
were confirmed by the fact that the number of DC still stuck in
the injection site was significantly higher in these animals with
respect to the positive-control group (**P < 0.001).
The analysis of the draining lymph node uptake showed
complementary data to these results. Positive-control mice, un-
dergoing active immunization with OVA-pulsed DC after in-
travenous injection of DO11.10 transgenic T cells, showed the
largest and fastest recruitment of DC in the draining lymph
nodes (Fig. 5B and Fig. S4) (*P < 0.03 with respect to negative
control). In vitro MSC-conditioned mice groups displayed a sig-
nificant reduction in the uptake of labeled OVA-pulsed DC
similar to that one of negative controls (*P < 0.01 with respect to
positive control). Remarkably, we noticed a sudden change in
the slope of the curve of in vivo MSC-conditioned DC group
(closed square line) occurring immediately after MSC intra-
venous injection, suggesting that in vivo MSC administration can
completely abrogate migration of DC to the draining lymph
nodes (Fig. 5B). These results indicate that MSC alter signifi-
cantly the migration of antigen-bearing DC in the draining lymph
nodes where T-cell priming occurs. Of relevance, injection of
MSC affects DC migration patterns within few minutes, sug-
gesting that in vivo MSC-mediated inhibition of DC–T-cell
interactions is an extremely rapid event.
In this study, we are unique in providing evidence that the in vivo
immunomodulatory activity of MSC is recapitulated by their
ability to directly inhibit DC effector functions. Several reports
have demonstrated that MSC have a profound impact on im-
mune functions both in humans and mice (3). For example, MSC
have been shown to inhibit T-cell proliferation through the in-
duction of cell division arrest (11), as well as to inhibit the dif-
ferentiation, maturation, and functions of monocyte-derived DC
(12–18). However, most studies reporting on the immunomod-
ulatory features of MSC have been carried out in vitro. Despite
some reports addressing the effect of MSC in experimental
models of immune-mediated diseases, little information about
the in vivo mechanisms has been provided. In particular MSC
in vivo administration resulted mainly in the inhibition of path-
ogenic antigen-specific T cells, as in a mouse model of allograft
rejection (19), experimental autoimmune encephalomyelitis
(20), graft-versus-host disease (21), and collagen-induced ar-
thritis (22). Despite these findings, it is not known whether T-cell
tolerance induced in vivo by MSC is the consequence of the
direct interaction between MSC and T cells or the result of an
impaired T-cell priming by APC. Here we confirm that MSC halt
maturation of BM-derived DC, leading to an inefficient antigen
presentation. MSC affect LPS-induced activation, leading to a
decreased production of cytokines, such as TNF-α and IL-12,
and this effect is caused by soluble factors, as demonstrated by
experiments with MSC supernatants. Interestingly we confirmed
that MSC expressed all sets of TLRs, including TLR4, the trig-
gering of which by LPS enhances the inhibitory activity of MSC
supernatants. In contrast to reports by others (16), we could not
detect an increased production of IL-10, suggesting that the DC-
mediated tolerogenic effect is mainly because of inefficient T-cell
priming. We also observed that MSC inhibit LPS-activated DC
production of IDO, a key enzyme involved in NF-κB activation,
leading to DC maturation (23).
To evaluate at the molecular level the effect of MSC on TLR-
induced DC activation, we investigated the effect of MSC on
MAPKs involved in the phosphorylation of cytoplasmic compo-
nents and nuclear transcription factors promoting cytokine pro-
duction. We demonstrated that MSC inhibited phosphorylation
of MAPKs downstream of MyD88, whose activation following
TLR4 stimulation leads to the NF-κB–mediated production of
IL-12, a key cytokine driving the Th1 response (24). We also
observed an enhanced activation of Akt on DC exposed to MSC.
These findings indicate that activation of Akt and impairment of
NF-κB are responsible for the impaired activation/maturation
nodes from a control (Left) and a MSC-treated mouse (Right) are shown.
(Scale bar, 0.5 cm.) One of three representative experiments is depicted. (B)
Histograms show the difference in the absolute cell number (×106) of the
draining lymph node from control mice (gray bars) and MSC-treated mice
(black bars); *P = 0.016. Results are representative of three independent
experiments (mean ± SD). (C) The histograms show the percentage of posi-
tive cells (CD4+, CD4+/CD25+, CD8+, B220+, B220+/CD138+, CD3−/DX5+) ex-
pressed as absolute cell number for each cell subset in the draining lymph
node of control mice (gray bars) and MSC-treated mice (black bars), *P < 0.05
for CD4+T cells. Results are representative of three independent experiments
(mean ± SD). (D) CFSE profiles of KJ1-26-gated transgenic DO11.10 CD4+T
cells demonstrate the absolute numbers of proliferating cells in the draining
lymph node of control mice (Left) and MSC-treated mice (Right). Absolute
numbers indicate the absolute values of cells recovered in standardized
acquisitions (shown at the top of the histograms). One of three representa-
tive experiments is depicted. (E) Intracellular staining for IFN-γ, IL-2, and TNF-
α on proliferating CD4+DO11.10 T cells from the draining lymph node of
control mice (Left) and MSC-treated mice (Right) are depicted. The percent-
age of transgenic DO11.10 CD4+T cells producing cytokines is shown at the
top of plots. One of three representative experiments is displayed. (F) The
percentage (Left) and the absolute number (Right) of CFSE+CD11c+DC re-
covered from draining lymph nodes are shown. *P = 0.0232 (Left); *P =
0.0353 (Right) (G) The percentage (Left) and the absolute number (Right) of
CFSE+CD11c+DC expressing CCR7 and CD49dβ1 recovered from draining
lymph nodes are depicted. Control mice (white bars) and MSC-treated mice
(black bars). **P = 0.0013 (Left); *P = 0.0195, **P = 0.008 (Right).
MSC impair in vivo CD4+T-cell priming by DC. (A) The draining lymph
Chiesa et al. PNAS
| October 18, 2011
| vol. 108
| no. 42
state of LPS-activated MSC-exposed DC, a molecular mecha-
nism recently associated with the immunogenicity subversion of
Leishmania infantum-infected DC (25).
We have also shown that MSC affect DC ability to process
soluble OVA and subsequently present them in the context of
MHC class I molecules to OVA-specific CD8+T cells, a process
know as “cross-priming” (26). These observations have focused
our attention on modulation of the APM, the defects in expres-
sion and function of which may lead to abnormalities in the
formation of complexes transported to the cell surface. The
processing and presentation of HLA class I antigen-derived
peptides is accomplished through a complex series of intracellular
events involving multiple molecular species. These species include
the constitutive proteasome subunits δ, MB1, and ζ, as well as the
immunoproteasome β-type subunits LMP2, LMP7, and LMP10.
Down-regulation of key components of the constitutive protea-
some, as well as of the immunoproteasome, has been reported by
several studies following exposure to different stimuli. For ex-
ample, it has been demonstrated that tumor cells can inhibit the
expression of several APM components, including MB1 and
LMP10, and that such diminished expression occurs also in im-
mature DC (27). A similar profile of change in catalytic subunit
profiles has been recently reported to occur in plasma cells from
patients with myeloma (28). The results reported in our study are
in lines with these observations concerning a paired down-regu-
lation of MB1 a-catalytic β-subunits located in the inner rings of
the constitutive proteasome (29) and LMP10, an IFN-γ–inducible
proteasomal subunit (30), and are consistent with the proposed
effect of MSC on DC maturation, resulting in an accumulation of
immature DC and an impaired capability to present MHC class
I-associated antigens to CD8 T cells.
The most striking result of this study arises from the demon-
T cells are primed in the lymph nodes by antigen-presenting DC,
the number and activation state of which affects the efficiency and
quality of the developing T-cell response (6). We first demon-
strated that activated CD11chighCFSE+DC are significantly re-
duced in the draining lymph nodes of MSC-treated animals
compared with controls, and that this effect is the result of an
in vivo decrease of CCR7 and CD49dβ1 expression on recruited
DC. These results are unique in suggesting that MSC are able to
affect in vivo theexpressionof surfacemoleculesimplicated in DC
migration to the lymph nodes. To further confirm these findings,
we used an experimental model where subcutaneous injection of
OVA-pulsed DC and subsequent migration to the draining lymph
nodes results in the activation of OVA-specific DO.11.10 trans-
genic naive CD4+T cells (31).
We showed that the intravenous injection of MSC in BALB/c
mice previously immunized with LPS-activated DC pulsed with
OVA results in a decrease in cell number of the draining lymph
nodes compared with that of the control mice, suggestive of an
antiproliferative effect on immune cells, probably caused by in-
hibition of cell division (11). Of relevance, MSC administration
led in vivo to an impaired priming of OVA-specific transgenic
CD4+naive T cells. In line with these findings, we demonstrated
that impairment of T-cell priming is because of the inability of
in vitro DC-conditioned MSC to efficiently leave the site of in-
jection and reach the draining lymph nodes. Even more impor-
tant, we observed that MSC intravenous injection induces an
almost immediate arrest of the otherwise rapid migration of LPS-
activated OVA-pulsed DC toward the draining lymph nodes. This
inhibition of migration is likely to be the result of an impaired up-
regulation of molecules involved in DC trafficking, such as CCR7
and CD49dβ1, as suggested by a significant decrease of the ex-
pression of these molecules on CFSE-labeled activated DC
detected in the draining lymph nodes. If such inhibition occurring
in vivo is because of a direct interaction between DC and MSC at
the site of DC injection, or is mediated by the release of soluble
factors, has yet to be determined. The data provided by experi-
ments performed with MSC supernatants strongly suggests that
soluble factors play a major role, at least in vitro. Moreover, the
large sequestration of MSC in the lungs following intravenous
administration (32) and the minimal time required—less than
20 min—for transplanted MSC to arrest DC migration to the
draining lymph nodes, makes very unlikely that intravenously
injected MSC could reach the subcutaneous tissue where DC
were implanted. Indeed the potential role of soluble factors re-
leased in vivo by MSC at a distance from the target organ has
been recently suggested by the demonstration that the interaction
of intravenously injected MSC with the lung microenvironment
leads to the release of the anti-inflammatory molecule TSG-6,
resulting in the improvement of the myocardial function after
global ischemia (33). It has been also recently demonstrated that
MSC-derived metalloproteases could paracrinally cleave MSC-
derived CCL2, leading to the generation of an antagonistic form
of CCL2 with the ability of inhibiting encephalitogenic T cells
(34). Moreover MSC-derived soluble factors have been shown
to significantly modify the migration pattern of immune cells,
resulting in the reversal of fulminant hepatic failure (35). Thus,
injected MSC appear to quickly modify the recipient immune
response upon interaction with the host tissues (36). Taken to-
gether, these observations suggest that both soluble factors, as
In vitro MSC
In vivo MSC
(iv after 15 minutes)
Nr of dendritic cells
DCs recruited by the draining lymph node
0510152025303540 4550 5560
DCs remaining in the injection site
DC released by the site of injection (A) and recruited to the draining lymph
nodes (B) as measured by the detection of radioactivity released by 99mTc-
HMPAO–labeled DC. Labeled DC was followed for 60 min after injection by
scintigraphic imaging. Four different experimental conditions were studied:
(i) negative control group represents mice exclusively subcutaneously im-
munized with OVA-pulsed DC (○, 6 mice); (ii) positive control group (•, 7
mice) indicates mice intravenously injected with DO11.10 T cells and sub-
sequently immunized with OVA-pulsed DC; (iii) in vitro MSC-conditioned DC
group (□, 6 mice) represents mice intravenously injected with DO11.10 T
cells and, subsequently immunized with OVA-pulsed in vitro MSC condi-
tioned-DC; (iv) in vivo MSC–conditioned DC group (■, 4 mice) designate mice
intravenously injected with DO11.10 T cells, subsequently immunized with
OVA-pulsed DC and intravenously injected with MSC 15 min later (arrow).
MSC impair DC migratory features. Graphs represent the number of
| www.pnas.org/cgi/doi/10.1073/pnas.1103650108Chiesa et al.
well as cell-to-cell contact mechanism (11), could be involved in Download full-text
the modulation of DC effectors functions by MSC.
The results reported in this study are unique in providing a de-
tailed analysis of MSC effects on DC functions in vitro as well as
in vivo, and link this impairment to the inability of priming T cells
andmounting an efficient antigen-specific immuneresponsein the
secondary lymphoid organs. Although the possible exploitation of
MSC for the treatment of immune-mediated diseases is currently
under scrutiny (37), these results confirm that MSC have a pro-
found and immediate effect on immune cells in vivo.
Materials and Methods
A summary of experimental techniques is given here, with full methods and
associated references presented in SI Materials and Methods.
BM-Derived MSC and DC. MSC were isolated as previously described (20) and
BM derived DC were generated as described elsewhere (38, 39).
In Vitro CD4 and CD8 Cell Cultures. After LPS activation, mature ctr-DC and
MSC-conditioned DC were pulsed with pOVA323-339or with OVA antigen and
cultured with pure naive CD4+and CD8+T cells labeled with 2.5 mM of
5-(and 6)-carboxyfluorescein diacetate, succinimidyl ester (CFSE, Molecular
Probes). For detailed methods, see SI Materials and Methods.
DC Transplantation. CFSE–labeled naive OVA-specific T cells (3 × 106) were
injected into syngeneic Balb/c mice (iv). LPS-activated DC were pulsed with
2 μM of pOVA323-339peptide and subcutaneously (sc) injected (2 × 106per
mouse) into the footpad 24 h after the adoptive transfer of DO11.10 cells.
Three h later, Balb/c mice were iv injected with MSC (1.25 × 106/mouse). On
day 3, priming of CFSE-labeled KJ1-26 positive DO11.10-specific T cells was
evaluated. Lymph node T cells were stimulated with PMA and Ionomycin
(Sigma-Aldrich) for 4 h and Brefeldin A for the last 2 h. Anti-TNF-α, IL-2, and
IFN-γ (Pharmingen BD) were used for intracellular staining.
DC Migration Study. For the in vivo analysis of CCR7+CD49dβ1+DC in the
draining lymph nodes, we sc injected 2 × 106CFSE-labeled DC in naive animals
followed by the iv administration of 1.25 × 106MSC or PBS (controls). On day
3 post DC injection, we isolated the draining lymph nodes from the two an-
imal groups (three animals per group) and analyzed CCR7 and CD49dβ1 ex-
pression on CD11c high CFSE+DC. DC radioactive labeling was performed
with 99mTc-exametazime (HMPAO, Ceretec, GE Healthcare) according to a
procedure described in SI Materials and Methods and elsewhere (40, 41).
ACKNOWLEDGMENTS. We thank Prof. S. Ferrone for kindly supplying
antibodies against the antigen processing machinery and Fabio Grassi for
kindly providing us with DO.11.10 and OT.1 transgenic mice. This research
was supported by grants from the Fondazione Italiana Sclerosi Multipla (to
A.U. and E.T.), the Italian Ministry of Health (Ricerca Finalizzata) (to A.U.),
the Italian Ministry of the University and Scientific Research (to A.U.), the
“Progetto LIMONTE” (to A.U.), and the Fondazione CARIGE (to A.U.).
1. Prockop DJ (1997) Marrow stromal cells as stem cells for nonhematopoietic tissues.
2. Uccelli A, Pistoia V, Moretta L (2007) Mesenchymal stem cells: A new strategy for
immunosuppression? Trends Immunol 28:219–226.
3. Uccelli A, Moretta L, Pistoia V (2008) Mesenchymal stem cells in health and disease.
Nat Rev Immunol 8:726–736.
4. Banchereau J, Steinman RM (1998) Dendritic cells and the control of immunity. Nature
5. De Smedt T, et al. (1996) Regulation of dendritic cell numbers and maturation by li-
popolysaccharide in vivo. J Exp Med 184:1413–1424.
6. MartIn-Fontecha A, et al. (2003) Regulation of dendritic cell migration to the
draining lymph node: Impact on T lymphocyte traffic and priming. J Exp Med 198:
7. Pevsner-Fischer M, et al. (2007) Toll-like receptors and their ligands control mesen-
chymal stem cell functions. Blood 109:1422–1432.
8. Tomchuck SL, et al. (2008) Toll-like receptors on human mesenchymal stem cells drive
their migration and immunomodulating responses. Stem Cells 26(1):99–107.
9. Krampera M, et al. (2006) Role for interferon-gamma in the immunomodulatory ac-
tivity of human bone marrow mesenchymal stem cells. Stem Cells 24:386–398.
10. Ren G, et al. (2008) Mesenchymal stem cell-mediated immunosuppression occurs via
concerted action of chemokines and nitric oxide. Cell Stem Cell 2(1):141–150.
11. Glennie S, Soeiro I, Dyson PJ, Lam EW, Dazzi F (2005) Bone marrow mesenchymal stem
cells induce division arrest anergy of activated T cells. Blood 105:2821–2827.
12. Jiang XX, et al. (2005) Human mesenchymal stem cells inhibit differentiation and
function of monocyte-derived dendritic cells. Blood 105:4120–4126.
13. Beyth S, et al. (2005) Human mesenchymal stem cells alter antigen-presenting cell
maturation and induce T-cell unresponsiveness. Blood 105:2214–2219.
14. Nauta AJ, Kruisselbrink AB, Lurvink E, Willemze R, Fibbe WE (2006) Mesenchymal
stem cells inhibit generation and function of both CD34+-derived and monocyte-
derived dendritic cells. J Immunol 177:2080–2087.
15. Aldinucci A, et al. (2010) Inhibition of immune synapse by altered dendritic cell actin
distribution: A new pathway of mesenchymal stem cell immune regulation. J Im-
16. Aggarwal S, Pittenger MF (2005) Human mesenchymal stem cells modulate allogeneic
immune cell responses. Blood 105:1815–1822.
17. Spaggiari GM, Abdelrazik H, Becchetti F, Moretta L (2009) MSCs inhibit monocyte-
derived DC maturation and function by selectively interfering with the generation of
immature DCs: Central role of MSC-derived prostaglandin E2. Blood 113:6576–6583.
18. Zhang B, et al. (2009) Mesenchymal stem cells induce mature dendritic cells into
a novel Jagged-2-dependent regulatory dendritic cell population. Blood 113(1):46–57.
19. Bartholomew A, et al. (2002) Mesenchymal stem cells suppress lymphocyte pro-
liferation in vitro and prolong skin graft survival in vivo. Exp Hematol 30(1):42–48.
20. Zappia E, et al. (2005) Mesenchymal stem cells ameliorate experimental autoimmune
encephalomyelitis inducing T-cell anergy. Blood 106:1755–1761.
21. Sudres M, et al. (2006) Bone marrow mesenchymal stem cells suppress lymphocyte
proliferation in vitro but fail to prevent graft-versus-host disease in mice. J Immunol
22. Augello A, Tasso R, Negrini SM, Cancedda R, Pennesi G (2007) Cell therapy using al-
logeneic bone marrow mesenchymal stem cells prevents tissue damage in collagen-
induced arthritis. Arthritis Rheum 56:1175–1186.
23. Hill M, et al. (2007) IDO expands human CD4+CD25high regulatory T cells by pro-
moting maturation of LPS-treated dendritic cells. Eur J Immunol 37:3054–3062.
24. Manetti R, et al. (1994) Interleukin 12 induces stable priming for interferon gamma
(IFN-gamma) production during differentiation of human T helper (Th) cells and
transient IFN-gamma production in established Th2 cell clones. J Exp Med 179:
25. Neves BM, et al. (2010) Activation of phosphatidylinositol 3-kinase/Akt and impair-
ment of nuclear factor-kappaB: Molecular mechanisms behind the arrested matura-
tion/activation state of Leishmania infantum-infected dendritic cells. Am J Pathol 177:
26. Bevan MJ (2006) Cross-priming. Nat Immunol 7:363–365.
27. Whiteside TL, Stanson J, Shurin MR, Ferrone S (2004) Antigen-processing machinery in
human dendritic cells: Up-regulation by maturation and down-regulation by tumor
cells. J Immunol 173:1526–1534.
28. Racanelli V, et al. (2010) Alterations in the antigen processing-presenting machinery
of transformed plasma cells are associated with reduced recognition by CD8+ T cells
and characterize the progression of MGUS to multiple myeloma. Blood 115:
29. Groettrup M, Soza A, Kuckelkorn U, Kloetzel PM (1996) Peptide antigen production
by the proteasome: Complexity provides efficiency. Immunol Today 17:429–435.
30. Nandi D, Jiang H, Monaco JJ (1996) Identification of MECL-1 (LMP-10) as the third IFN-
gamma-inducible proteasome subunit. J Immunol 156:2361–2364.
31. Martín-Fontecha A, et al. (2004) Induced recruitment of NK cells to lymph nodes
provides IFN-gamma for T(H)1 priming. Nat Immunol 5:1260–1265.
32. Barbash IM, et al. (2003) Systemic delivery of bone marrow-derived mesenchymal
stem cells to the infarcted myocardium: Feasibility, cell migration, and body distri-
bution. Circulation 108:863–868.
33. Lee RH, et al. (2009) Intravenous hMSCs improve myocardial infarction in mice be-
cause cells embolized in lung are activated to secrete the anti-inflammatory protein
TSG-6. Cell Stem Cell 5(1):54–63.
34. Rafei M, et al. (2009) Mesenchymal stromal cells ameliorate experimental autoim-
mune encephalomyelitis by inhibiting CD4 Th17 T cells in a CC chemokine ligand 2-
dependent manner. J Immunol 182:5994–6002.
35. Parekkadan B, et al. (2007) Mesenchymal stem cell-derived molecules reverse fulmi-
nant hepatic failure. PLoS ONE 2:e941.
36. Uccelli A, Prockop DJ (2010) Why should mesenchymal stem cells (MSCs) cure auto-
immune diseases? Curr Opin Immunol 22:768–774.
37. Uccelli A, Laroni A, Freedman MS (2011) Mesenchymal stem cells for the treatment of
multiple sclerosis and other neurological diseases. Lancet Neurol 10:649–656.
38. Inaba K, et al. (1992) Generation of large numbers of dendritic cells from mouse bone
marrow cultures supplemented with granulocyte/macrophage colony-stimulating
factor. J Exp Med 176:1693–1702.
39. Lutz MB, et al. (1999) An advanced culture method for generating large quantities of
highly pure dendritic cells from mouse bone marrow. J Immunol Methods 223:77–92.
40. Botti C, et al. (1997) Comparison of three different methods for radiolabelling human
activated T lymphocytes. Eur J Nucl Med 24:497–504.
41. Blocklet D, et al. (2003) 111In-oxine and 99mTc-HMPAO labelling of antigen-loaded
dendritic cells: In vivo imaging and influence on motility and actin content. Eur J Nucl
Med Mol Imaging 30:440–447.
Chiesa et al. PNAS
| October 18, 2011
| vol. 108
| no. 42