Christine Schirmer,1,2Claudia Klein,1Martin von Bergen,2*Jan C. Simon,1and *Anja Saalbach1
1Department of Dermatology, Venerology andAllergology, Medical Faculty of the Leipzig University, Leipzig; and2Helmholtz Centre for Environmental Research
UFZ, Department of Proteomics, Leipzig, Germany
The initiation of immune responses is
associated with the maturation of den-
dritic cells (DCs) and their migration to
draining lymph nodes. En route activated
DCs encounter cells of the tissue micro-
environment, such as fibroblasts. Be-
cause we have shown that DCs interact
sponses, we studied the impact of skin
fibroblasts on human monocyte-derived
DC function and subsequent human T-
cell (TC) differentiation. We show that
fibroblasts support interleukin-23 (IL-23)
secretion from DCs preactivated by lipo-
polysaccharide (DCact) compared with
lipopolysaccharide-activated DCs alone.
The underlying complex feedback-loop
mechanism involves IL-1?/tumor necro-
sis factor-? (from DCact), which stimulate
fibroblasts prostaglandin E2production.
Prostaglandin E2, in turn, acts on DCact
and increases their IL-23 release. Further-
more, fibroblast-stimulated DCactare far
superior to DCactalone, in promoting the
expansion of Th17 cells in a Cox-2-, IL-23-
dependent manner. Using CD4?CD45RO?
memory TCs and CD4?CD45RA?naive
TCs, we showed that fibroblasts induce a
phenotype of DCact that promotes the
expansion of Th17 cells. Moreover, in
psoriasis, a prototypic immune response
in which the importance of IL-23/Th17 is
known, high expression of Cox-2 in fibro-
blasts was observed. In conclusion, skin
fibroblasts are involved in regulation of
IL-23 production in DCs and, as a result,
of Th17 expansion. (Blood. 2010;116(10):
Dendritic cells (DCs) are in permanent contact with their tissue
microenvironment. For example, niches created by the stromal
microenvironment in the bone marrow enable DC precursors to
differentiate into DC subpopulations.1,2In peripheral tissues, DCs
interact with a microenvironment composed of the extracellular
matrix and stromal cells, such as fibroblasts, macrophages, and
endothelial cells. Recently, we demonstrated an interaction of DCs
and fibroblasts, resulting in enhanced migration of DCs.3Several
groups reported the effect of stromal cells on the regulation of DC
functions under steady-state as well as inflammatory conditions.4,5
Renkl et al6and Termeer et al7showed that maturation and
differentiation of DCs are regulated by the extracellular matrix
components hyaluronic acid and osteopontin. Zhang et al reported
splenic stromal cells to drive DCs toward a regulatory phenotype.8
Furthermore, emigration of Langerhans cells from the skin to the
lymph node is influenced by a lack of matrix-associated SPARC
(secreted protein, acidic and rich in cysteine), resulting in enhanced
cutaneous contact hypersensitivity.9
Fibroblasts are thought to be involved in the regulation of DC
functions because they synthesize factors known to influence DC
function, such as cytokines, chemokines, prostanoids, matrix
components, and matrix-degrading enzymes.10-12Moreover, tran-
scriptome analysis of fibroblasts showed significant differences
between fibroblasts from different anatomic regions.12,13Thereby,
fibroblasts could transfer tissue-specific information to DCs.12,13In
skin, fibroblasts regulate the emigration of Langerhans cells from
epidermis to dermis via CXCL12 secretion.14Briard et al showed
that fibroblasts can mediate the maturation of DC and natural killer
cells in spleen.15
In response to pathogens and under the influence of the
microenvironment, DCs secrete specific cytokines, involved in
polarization of T-cell (TC) subsets mediating distinct types of
inflammation to combat pathogens.16Interleukin-12 (IL-12) pro-
motes Th1 differentiation, IL-4 is critical for the Th2 lineage.17
IL-23 promotes the development of the novel Th17 population,
characterized by the production of IL-17A,17,18IL-22,18-20and
expression of ROR?t.17,21Th17 cells are important for defense
17A axis plays a role in different autoimmune diseases, including
Here, we show that skin fibroblasts actively participate in the
regulation of an immune response by affecting the IL-23 produc-
tion of DCs and thus supporting the Th17 cell expansion.
Antibodies include anti–CD83-fluorescein isothiocyanate (FITC), anti–
CD86-FITC, anti–CD80-FITC, anti–CD1a-phycoerythrin (PE), anti–CD3-
PE, and anti–Cox-2 (BD Biosciences); anti–HLA-DR-FITC, anti–CD11c-
anti-CD1c, anti–CD45RA-FITC, and anti–IL-17A-PE; anti–ING?-PE
(Miltenyi Biotec); anti–ROR?t-PE (eBioscience); and anti-iNOS (Di-
anova). Secondary goat anti–rabbit FITC was purchased from Santa Cruz
Submitted January 11, 2010; accepted May 22, 2010. Prepublished online as
Blood First Edition paper, June 10, 2010; DOI 10.1182/blood-2010-01-263509.
*J.C.S. and A.S. contributed equally to this study and share the senior
The online version of this article contains a data supplement.
The publication costs of this article were defrayed in part by page charge
payment. Therefore, and solely to indicate this fact, this article is hereby
marked ‘‘advertisement’’ in accordance with 18 USC section 1734.
© 2010 by TheAmerican Society of Hematology
1715 BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
Biotechnology, and neutralizing anti–human IL-23p19 antibody, anti–
human IL-6, and normal IgG (isotype control) from R&D Systems.
Cell preparation and culture conditions
Human DCs were generated from CD14-positive monocytes and cultured
in DC-medium (RPMI 1640, Invitrogen) supplemented with 2% fetal calf
(Invitrogen), 0.1mM nonessential amino acids (Invitrogen), 10mM HEPES
(Invitrogen), 100 U/mL IL-4 (PeproTech), and 1000 U/mL granulocyte-
macrophage colony-stimulating factor (Leukine, Berlex).3Morphologic
analysis and high expression of CD1a and CD11c were parameters for
quality and purity of DC preparations.An immature phenotype of DCs was
verified by low expression of CD80, CD83, CD86, and HLA-DR.
Human Pan TC, CD4?TCs and memory CD4?TCs were isolated from
PBMC using the Pan T Cell Isolation Kit II, CD4?T Cell Isolation Kit II
and Memory CD4?T Cell Isolation Kit (Miltenyi Biotec), respectively,
according to manufacturer’s instructions. Naive CD4?TCs were gained by
2-step isolation with first using the CD4?TCell Isolation Kit II followed by
the Naive CD4?T Cell Isolation Kit (Miltenyi Biotec). Purity of TC
populations was verified by flow cytometry with anti-CD3 for Pan TC,
anti-CD4 for CD4?TC as well as anti-CD4, anti-CD45RO, and anti-
CD45RAfor memory and naiveTCs. Cell culture medium was a mixture of
DC medium and proliferation medium (DC-RPMI supplemented with 10%
fetal calf serum).
Human fibroblasts were isolated from skin biopsies by enzymatic
digestion as described and cultured in Dulbecco modified Eagle medium
(Biochrom) containing 10% fetal calf serum and 1% penicillin/streptomy-
cin (Biochrom) up to the fourth passage.3Morphologic analysis and high
expression of Thy-128controlled for purity of fibroblasts. The study was
conducted in accordance with the Declaration of Helsinki and was approved
by the ethics committee of the University of Leipzig (065-2009).
Coculture experiments and supernatant transfer
DC-fibroblast coculture. Dermal fibroblasts were cultured in 24-well-
plates to confluence. For preactivation, DCs were stimulated with a final
concentration of 1 ?g/mL lipopolysaccharide (LPS). After 3 hours, DCs
were harvested and washed with phosphate-buffered saline (PBS) to
remove LPS. For coculture, 4 ? 105/mL DCs were added to washed,
allogeneic fibroblasts in DC-RPMI and incubated for 18 hours. As control,
LPS-stimulated DCs and fibroblasts were cultured alone under identical
conditions. Supernatants were collected after 18 hours. For RNA prepara-
tion, cells were harvested after 3 hours of coculture.As indicated, 1 ?g/mL
bioactivity-neutralizing antibodies or 2?M indomethacin (Sigma-Aldrich)
DC-TC coculture. A total of 3 ? 105TCs and 1.5 ? 104DCs were
cocultured in 96-well F-bottom plates in DC-medium containing 10%
fetal calf serum. Before DC-TC coculture, fibroblasts were removed
from the DC-fibroblast coculture using anti–Thy-1–coupled magnetic
beads (Invitrogen).28,29For control, LPS-stimulated DCs, which had
been cultured alone, were used under the same conditions. After 3 days,
cells were harvested and used for intracellular flow cytometric analysis.
After 6 days, cell culture supernatants were collected.As indicated, 1 ?g
anti–IL-23p19, anti–IL-6, or isotype control was added to DC-TC
For experiments with supernatants of DC-fibroblast cocultures, 2 ? 105
Pan, CD4?, memory, or naive TCs per well were resuspended in a 1:1
in the presence of 1.5 aCD3/CD28 beads (Invitrogen) per TCs. After a
proliferation time of 5 days, TC supernatants were analyzed by enzyme-
linked immunosorbent assay (ELISA).
Transfection of fibroblasts and DCs with siRNA
Subconfluent fibroblasts were transfected with either 25nM siRNA (primer
1: UCCAGACAAGCAGGCUAAUACUGAU, primer 2: AUCAGUAUU-
AGCCUGCUUGUCUGGA) specific for human cyclooxygenase (Cox)
2 or 25nM scrambled siRNA (Invitrogen) using Lipofectamine RNAiMax
Transfection Reagent (Invitrogen) overnight according to manufacturer’s
DCs were transfected with same siRNA specific for human Cox-2 or
scrambled siRNA (Invitrogen) using Amaxa-transfection system (Lonza
Germany). A total of 2 ? 106DCs were transfected with 2?M Cox-2
specific siRNA or scrambled siRNA using program X001. Subsequently,
5 ? 105transfected DCs were cultured in DC medium without IL-4 and
granulocyte-macrophage colony-stimulating factor overnight. For transfec-
tion, control 1 ?g/mL LPS was added, prostaglandin E2(PGE2) secretion
was detected after 24 hours by ELISA. For coculture experiments,
transfected DCs were stimulated for 3 hours with 1 ?g/mL LPS and
coculture was performed as described in “Coculture experiments and
For analysis of extracellular epitopes, cells were incubated with indicated
labeled antibody for 45 minutes at 4°C. Cells were washed twice with
PBS/Gelafusal (Serumwerke Bernburg)/sodium acid. Antibody binding
was analyzed by flow cytometry (FC 500, Beckman Coulter). For intracel-
lular flow cytometric staining, TCs were restimulated using 5 ng/mL
phorbol myristate acetate (Sigma-Aldrich) and 500 ng/mL ionomycin
4 hours. Cells were harvested, washed twice with PBS/Gelafusal/sodium
acid, and incubated with anti-CD3 for 45 minutes at 4°C for selective
recognition of TCs. Afterward, cells were fixed with 4% paraformaldeh-
hyde in PBS and permeabilized with 0.1% saponin. Then, indicated labeled
antibody was added for 45 minutes at 4°C, and cells were washed twice
with PBS/Gelafusal/sodium acid.
RNA preparation and real-time PCR
Fibroblasts and DCs from coculture experiments were separated using
anti–Thy-129coupled magnetic beads (Invitrogen). Success of the separa-
tion was controlled by FACS staining with anti–Thy-1 and anti-CD11c29
and real-time polymerase chain reaction (PCR) analysis of Thy-1 and
CD11c. FACS analysis revealed complete depletion of DC from fibroblasts,
whereas the more sensitive PCR showed that approximately 1% of DCs
could not be removed from the fibroblasts.
After separation of DCs and fibroblasts, total RNA was then isolated
with the innuPREPRNAMini Kit (Analytik Jena), and 0.5 ?g of total RNA
was used for first-strand cDNAsynthesis with M-MLVreverse transcriptase
(Promega) according to the manufacturer’s protocol. Real-time PCR was
performed. The following primers (metabion International) were used:
TCTG-3?, reverse: 5?-TTCACATACAGCTTGGGAAGCA-3?, IL-23p19:
GATTTATCTTGG-3?, COX2: forward: 5?-TGCTGTGGAGCTGTATC-
CTG-3?, reverse: 5?-TCATCTAGTCCGGAGCGG-3?, CD11c: forward: 5?-
TTCTGACAGCCAATGTGAGC-3?, reverse: 5?-TCCTGGTTCAGCTC-
CACAG-3?, TNF-?: forward: 5?-GCAATGATCCCAAAGTAGACCTGC-
CCAGACT-3?, reverse: 5?- GAGTGACAAGCCTGTAGCCCATGTTG-
TAGCA-3?, IL-1?: forward: 5;-GACACATGGGATAACGAGGC-3?,
reverse: 5?-ACGCAGGACAGGTACAGATT-3?. Genes were quantified
through a standard curve, normalized to the unregulated housekeeping
gene RPS26 and computed as percentage of mRNA expression in
DC-LPS. For TNF-?, the relative level of each mRNAwas calculated on
the basis of ?Ct values. Genes were normalized to the unregulated
housekeeping gene RPS26 and computed as percentage of mRNA
expression in DC-LPS.
IL-23, IL-17, TNF-?, IL-12, IL-6, IL-1?, transforming growth factor-?
(TGF-?), interferon-? (IFN-?), and PGE2 were detected by ELISA
(eBioscience, BD Biosciences, and R&D Systems) according to the
1716SCHIRMER et alBLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
Frozen tissue sections of healthy skin and lesional psoriatic skin (n ? 6)
were fixed with 4% paraformaldehyde and permeabilized with PBS/0.5%
Tween20. Cox-2 expression was detected with an anti–Cox-2 antibody,
avidin-biotin-alkaline phosphatase complex, according to the manufactur-
er’s protocol (Biogenix) and visualized colorimetrically using the New
Fuchsin substrate system (Dako). For quantification of Cox-2 expression,
skin sections were scanned with a digital camera (Carl Zeiss; 5 shots per
skin) and analyzed with HistoClick software based on morphometric image
analysis developed in our laboratory.30The degree of Cox-2 expression is
expressed by the number of pixels.
For double labeling, Cox-2 expression was detected in the same way.
Subsequently, sections were washed and blocked with an avidin/biotin
blocking kit (Vector Laboratories).After washing, Thy-1 was detected by a
biotinylated anti–Thy-1 antibody and strepavidin-peroxidase (Vector Labo-
ratories). Bound antibodies were detected by ImmuPACT substrate SG
resulting in a gray-blue color.Tissue sections were extensively washed with
PBS/0.1% Tween20 after every antibody incubation step. In control
sections, primary antibodies were replaced by appropriate isotype control
Distribution of data was assessed by Shapiro-Wilk test; and depending on
the normality of the data, analysis was performed using Mann-Whitney
rank-sum test or Student t test. Values of P less than .05 were considered to
Human dermal fibroblasts increase IL-23 production of
Monocyte-derived DCs were preactivated with LPS (DCact) for
3 hours to imitate DCs within an inflammatory environment.
Subsequently, DCactwere cocultured with skin-derived fibroblasts,
and IL-23 was measured in the coculture supernatants. DCactalone
produced intermediate amounts of IL-23, whereas fibroblasts
markedly increased IL-23 secretion by DCact(Figure 1A). Because
fibroblasts stimulated with LPS did not express or secrete IL-23,
we can exclude a direct effect of residual amounts of LPS carried
over with DCs into the coculture on IL-23 secretion from fibro-
blasts (data not shown). In supernatants of unstimulated DCs or
fibroblasts alone as well as unstimulated DCs in coculture with
fibroblasts, no IL-23 could be detected (Figure 1A). To identify
the cellular source of IL-23, we separated DCactand fibroblasts
after cocultivation and measured the amount of IL-23p19 tran-
scripts. Figure 1B shows that IL-23p19 mRNA was solely gener-
ated by DCact.
Characterization of DCs cocultured with fibroblasts
Next, we investigated the attributes of DCactthat were cocultured
with fibroblasts compared with DCs preactivated by LPS alone
(Table 1). Flow cytometry revealed both DCact and DCact in
coculture with fibroblasts expressed similar amounts of HLA-DR,
CD1c, and inducible nitric oxide synthase (iNOS) (Table 1). To
further characterize fibroblast-stimulated DCact, cytokine secretion
was analyzed by ELISA. As summarized in Table 1, supernatants
of fibroblast-stimulated DCactcontained higher amounts of IL-23
and IL-6. IL-12 was expressed at low levels, but fibroblasts
stimulated IL-12 production 2-fold in DCactcompared with DCact
alone. TGF-? and IL-1? amounts were similar in DCactand in
fibroblast-DCactcocultures. Level of TNF-? was less in fibroblast-
DCactcoculture than in supernatants of DCact.
Figure 1. Fibroblasts support IL-23 release from DCs. (A) Monocyte-derived DCs were preactivated with LPS for 3 hours. Subsequently, LPS-stimulated DCs (termed DCact
in the entire manuscript) were washed extensively and then cocultured with fibroblasts (DCact ? Fb) overnight.As control, LPS-stimulated DCs were cultured alone (DCact).
Furthermore, we cultured immature DCs without (DC) and with fibroblasts (DC ? Fb) or fibroblasts (Fb) alone. IL-23 protein levels were measured by ELISA. *P ? .001
compared with DCact(n ? 5 independent experiments). (B) DCactand fibroblasts were cocultured for 3 hours; afterward, DCs (DCact separated after coculture) and fibroblasts
(Fb separated after coculture) were separated from the coculture by anti–Thy-1–coupled magnetic beads. For comparison, DCactand fibroblasts (Fb) were cultured alone.
Subsequently, RNApreparation and real-time PCR of IL-23p19 mRNAwere performed. IL-23p19 mRNAexpression values were normalized to the unregulated housekeeping
gene RPS26 and are given as percentage of IL-23p19 mRNAexpression in DCact. *P ? .001 compared with DCact(n ? 3 independent experiments).
Table 1. Phenotype of DCactand DCact-Fb on coculture with
97.8 ? 1.3
102 ? 39
48.6 ? 16
36 ? 3
3020 ? 1608
210 ? 85
38 ? 15
24 250 ? 6339
178 ? 113
168 ? 184
96.8 ? 3
92 ? 28
46.6 ? 15
27 ? 6.9
1188 ? 502*
722 ? 160†
67 ? 26†
120 092 ? 38 586†
107 ? 74
247 ? 350
DCs were preactivated with LPS and afterward cocultured without (DCact) and
with fibroblasts (DCact-Fb) overnight. HLA-DR (percentage positive cells and mean),
CD11c, and iNOS (percentage positive cells) were measured by flow cytometry.
IL-23, IL-12, IL-6, TGF-?, IL-1?, and TNF-? protein levels were measured by ELISA.
*P ? .05 vs DCact(n ? 3 independent experiments).
†P ? .001 vs DCact(n ? 3 independent experiments).
HUMAN FIBROBLASTS 1717BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
Blocking of TNF-? and IL-1? prevents fibroblast-stimulated
DCs from increased IL-23 secretion
We next investigated the mechanisms of the fibroblast-stimulated
IL-23 secretion in DCact. Supernatants of fibroblasts did not
stimulate IL-23 release from DCact(Figure 2A). Moreover, fixed
fibroblasts allowing exclusively cell-cell interactions did not in-
crease IL-23 secretion in DCact(Figure 2A). These results sug-
gested a more complex feedback loop mechanism between DCact
and fibroblasts. Thus, we first investigated potential mediators
produced by DCs in coculture. The proinflammatory cytokines
TNF-? and IL-1? are known to be present abundantly in inflamed
stimulated DCact(Table 1; Figure 2C-D). In addition, supernatants
of fibroblasts activated with TNF-? and IL-1? also increased IL-23
release of DCactcompared with DCact(supplemental Figure 1,
available on the Blood Web site; see the Supplemental Materials
link at the top of the online article).
To study the role ofTNF-? and IL-1? in the up-regulation of IL-23
secretion in DCacton coculture with fibroblasts, neutralizing antibodies
were used. Anti–TNF-? reduced IL-23 production of fibroblast-
whereas anti–IL-1? alone led to a significant decrease (P?.05) of
IL-23 at least to the level of IL-23 in DCact. The combination of
identify the cellular source of TNF-? and IL-1?, DCactand fibroblasts
were separated from the coculture and quantitative reverse-transcribed
cocultured with DCact, TNF-?– and IL-1?–mRNA could be detected;
by contrast, DCactexpressed high levels of TNF-?– and IL-1?–mRNA
Figure 2. DC-derived TNF-? and IL-1? are involved in a feedback loop mechanism that causes increased IL-23 production in cocultures of DCs and fibroblasts.
(A) DCactwere cultured in the presence of fibroblast supernatants (DCact ? [Fb-sn]) or PFA-fixated fibroblasts (DCact ? Fb-fix). As controls, DCacteither cultured alone
(DCact) or cocultured with fibroblasts (DCact ? Fb) were used. IL-23 was detected by ELISA. *P ? .001 compared with DCact (n ? 3 independent experiments).
(B) Neutralizing anti–TNF-? or anti–IL-1? antibodies were added to coculture of DCactand fibroblasts separately (DCact ? Fb ? aTNF-? and DCact ? Fb ?aIL-1beta) or in
combination (DCact ? Fb ? aTNFalpha/aIL-1beta). IL-23 levels were compared with coculture with isotype control (DCact ? Fb ? iso) and preactivated DCs cultured alone
(DCact). #,*P ? .05 (n ? 4 independent experiments). (C-D) DCactand fibroblasts were cocultured for 3 hours and then separated (DCact separated after coculture and Fb
separated after coculture). For comparison, preactivated DCs (DCact) and fibroblasts (Fb) were cultured alone. Subsequently, RNA preparations and real-time PCR were
performed. (C) Relative level ofTNF-? mRNAwas calculated on the basis of ?Ct values. mRNAexpression was normalized to the unregulated housekeeping gene RPS26 and
computed as percentage of TNF-? mRNA expression in DCact. *P ? .01 compared with DCact(n ? 3 independent experiments). (D) IL-1? mRNA was quantified through a
standard curve, normalized to the unregulated housekeeping gene RPS26, and computed as percentage of mRNAexpression in DCact(n ? 3 independent experiments).
1718SCHIRMER et al BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
PGE2derived from fibroblasts supports IL-23 production of
Activation of fibroblasts with TNF-? and IL-1? results in PGE2
secretion.3In addition, we showed that PGE2in concentrations mea-
sured in DCact-fibroblast cocultures drives DCactto secrete increased
amounts of IL-23 compared with DCactalone (supplemental Figure 2).
Therefore, we added indomethacin to DCact-fibroblast cocultures to
inhibit Cox-1 and Cox-2. Coculture of DCactand fibroblasts led to an
increase in PGE2compared with DCactalone, which was completely
prevented by indomethacin (Figure 3A). Importantly, indomethacin
reduced IL-23 secretion in DCact-fibroblast cocultures to the level of
DCactalone (Figure 3B), indicating that PGE2is an important mediator
To differentiate between DCs and fibroblasts as PGE2produc-
ers, DCactand fibroblasts were separated after coculture to perform
quantitative RT-PCR. Both DCactand fibroblasts showed increases
in Cox-2 mRNA levels compared with their counterparts cultured
alone (Figure 3C). Specifically, Cox-2 mRNA increased 3-fold in
DCact, whereas it increased more than 10-fold in fibroblasts on
coculture with DCact.
To discriminate whether DCact- or fibroblast-derived PGE2
promotes IL-23 secretion in fibroblast-stimulated DCact, Cox-2
expression in fibroblasts or DCs was inhibited by siRNA silencing
before setting up DCact-fibroblast coculture. To evaluate efficiency
of Cox-2 silencing, transfected fibroblasts were incubated with
TNF-?/IL-1? and transfected DCs were activated by LPS over-
night, both known stimulators of PGE2production.3Cox-2 siRNA
transfection dramatically reduced PGE2secretion of stimulated
fibroblasts as well as stimulated DCs (P ? .05; Figure 4A)
compared with the scrambled siRNA control. However, PGE2
decline was not complete because approximately 10% to 20% of
PGE2could still be detected in Cox-2–silenced cells (Figure 4A-B).
Interestingly,IL-23secretionwassignificantlydecreased(P ? .001)
in DCactcocultured with Cox-2–silenced fibroblasts compared with
cocultures with scrambled-transfected fibroblasts (Figure 4B). In
Figure 3. PGE2, which is produced because of coculturing of activated DCs and fibroblasts, is also involved in IL-23 up-regulation. (A) Indomethacin’s PGE2inhibiting
efficiency was verified by detection of PGE2levels in DCactcultured alone (DCact) and in coculture of DCactand fibroblasts with and without 2?M indomethacin (DCact ? Fb
and DCact ? Fb ? indomethacin, respectively). *P ? .01 (n ? 7 independent experiments). (B) DCs were preactivated with LPS and subsequently cocultured with fibroblasts
(DCact ? Fb) or without fibroblasts (DCact). Indomethacin (2?M) was added to the cocultures (DCact ? Fb ? indomethacin) and IL-23 was measured by ELISA. *P ? .001
(n ? 7 independent experiments). (C) DCs and fibroblasts were separated using anti–Thy-1 coated microbeads after a coculture time of 3 hours (DCact separated after
coculture and Fb separated after coculture). Controls were again LPS-stimulated DC (DCact) and fibroblasts (Fb) cultured alone. Then, RNA was prepared and quantitative
RT-PCR was performed. Cox-2 mRNA expression values were normalized to the unregulated housekeeping gene RPS26 and are given as percentage of Cox-2 mRNA
expression in DCact. *P ? .01 (n ? 3 independent experiments).
HUMAN FIBROBLASTS1719BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
contrast, silencing of Cox-2 in DCs did not affect their IL-23
secretion on coculture with fibroblasts. Thus, fibroblast-derived
PGE2is essential to stimulate IL-23 production of DCact.
However, knockdown of Cox-2 in fibroblasts did not com-
pletely inhibit the stimulation of IL-23 secretion in DCactobserved
after coculture with fibroblasts, supposedly because of the incom-
plete Cox-2 knockdown in fibroblasts. Indeed, stimulation of DCact
with PGE2concentrations present in cocultures on transfection of
fibroblasts with Cox-2 siRNA still caused a duplication of IL-23
compared with DCactalone, which was also seen in coculture of
DCactwith Cox-2–silenced fibroblasts (data not shown).
These results support the notion that mainly fibroblast-derived
PGE2is responsible for enhanced IL-23 secretion in DCacton
interaction with fibroblasts.
Fibroblasts support Th17 development via stimulation of IL-23
secretion from DCs
To analyze the consequences of the up-regulated IL-23 secretion of
fibroblast-stimulated DCactfor TC function, DCactand fibroblast-
stimulated DCact were harvested after 18 hours of coculture.
Fibroblasts were removed from the coculture using anti–Thy-1–
coupled magnetic beads as described.29Thereafter, CD3?(Pan) or
CD4?TCs were cocultured with the harvested DCactor DCact
purified from DCact-fibroblast coculture in the presence of respec-
tive supernatants of the DCactculture or DCact-fibroblast coculture.
As control, TCs were cultured alone. CD3?as well as CD4?TCs
cocultured with fibroblast-stimulated DCactproduced significantly
(P ? .01) higher amounts of IL-17Acompared with TCs that were
stimulated with DCact(Figure 5A). Intracellular flow cytometric
staining of CD3?TCs revealed that augmented IL-17 secretion was
associated with an increased number of IL-17A–producing TCs
(Figure 5B). ROR?t-positive TCs were also elevated through
fibroblast-stimulated DCactcompared with DCactalone (Figure 5C).
In addition, we showed that fibroblast-conditioned DCact also
enhanced IL-22 secretion of CD3?TCs compared with TCs
cultured with DCact(P ? .05; Figure 5D). Interaction of CD3?TCs
with fibroblast-stimulated DCact also enhanced IFN-? release
Next, we were interested in the mechanism of the IL-17Aboost
in TCs mediated through the action of fibroblasts on DCs.Analysis
of the cytokine secretion in DCactand fibroblast-stimulated DCact
revealed an enhanced IL-6 and IL-23 release in the DCact-fibroblast
Figure 4. Fibroblast-derived PGE2is the critical PGE2for stimulation of IL-23 secretion from DCact. (A) Fibroblasts or DCs were transfected with Cox-2 siRNA or
scrambled siRNA. To induce PGE2 expression, Cox-2– and scrambled-transfected fibroblasts were stimulated with TNF-?/IL-1? (Fb[Cox-2si]?TNFalpha/IL-1beta;
Fb[scr] ? TNFalpha/IL-1beta). Cox-2– and scrambled-transfected DCs (DC[Cox-2si] ? LPS; DC[scr] ? LPS) were activated with LPS. PGE2release was measured by
ELISA. *P ? .05 compared with scrambled-transfected cells (n ? 3 independent experiments). (B) LPS-preactivated DCs (DCact) were cocultured with scrambled-siRNA-
transfected fibroblasts (DCact ? Fb[scr]) or with Cox-2-siRNA transfected fibroblasts (DCact ? Fb[Cox-2si]). As control, DCactwere cultured alone. IL-23 production was
measured by ELISA. *P ? .001 (n ? 5 independent experiments). (C) Cox-2– or scrambled siRNA–transfected DCs were preactivated with LPS and subsequently cocultured
with fibroblasts (DCact[scr] ? Fb; DCact[Cox-2si] ? Fb).As control, transfected DCactwere cultured alone (DCact[scr]; DCact[Cox-2si]). *P ? .001.
1720 SCHIRMER et al BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
coculture, whereas IL-1? and TGF-? were not differentially
secreted (Table 1). Therefore, we studied whether IL-6 or IL-23
are the responsible factors mediating the enhanced IL-17
secretion of TCs on interaction with fibroblast-stimulated DCact.
Coculture of CD3?TCs and fibroblast-stimulated DCactwas
performed in the presence of neutralizing antibodies. Neutraliz-
ing of IL-23 by an anti–IL-23p19 antibody almost completely
abrogated, whereas an anti–IL-6 antibody only slightly reduced,
the boost of IL-17A achieved by fibroblast-primed DCact
(P ? .05; Figure 5F). Thus, IL-23 seems to be the most
important factor responsible for the increase of IL-17A in TCs
cultured with fibroblast-stimulated DCact.
Next, we were interested whether fibroblast-stimulated DCact
support the differentiation or expansion of Th17 cells. There-
fore, CD4?TC, CD4?CD45RO?memory, or CD4?CD45RA?
naive TCs were cultured with aCD3/CD28 beads in the presence
of supernatants of either DCactor fibroblast-stimulated DCact.
Supernatants of fibroblast-stimulated DCact induced signifi-
cantly higher IL-17Aproduction in CD4?and memory TCs than
DCact supernatants did (Figure 5G). In contrast, naive TCs
released only minimal amounts of IL-17A, and no significant
difference between DCact- or DCact-fibroblast supernatants could
be observed (Figure 5G). Thus, under our experimental condi-
tions, differentiation of Th17 cells from naive TCs did not take
Figure 5. Fibroblasts promote Th17 development from TCs via the stimulation of IL-23 from activated DCs. LPS-preactivated DCs were cocultured with fibroblasts.After
18 hours, fibroblasts were removed from the coculture using anti–Thy-1–coupled magnetic beads. As control, LPS-preactivated DCs (DCact) were cultured alone.
Subsequently, CD3?TCs or CD4?were cultured alone (TC) or with either LPS-preactivated DCs (TC ? DCact) or LPS-preactivated DCs isolated from the coculture with
fibroblasts (TC ? [DCact ? Fb]) in the presence of respective coculture supernatants. (A,D-E) After 6 days of coculture, IL-17A (A), IL-22 (D), and IFN-? (E) production was
measured by ELISA. *P ? .01, compared with TC ? DCact(n ? 5 independent experiments). (B-C) After 3 days of coculture, TCs were restimulated with phorbol myristate
acetate/ionomycin in the presence of brefeldin A and IL-17A expression (B) and transcription factor ROR?t expression (C) were measured by intracellular flow cytometric
staining. One representative experiment of 3 is shown. (F) Coculture of CD3?TCs and DCact(TC ? DCact) or with fibroblast-stimulated DCact(TC ? [DCact ? Fb]) were
performed without antibody (TC ? [DCact ? Fb]) in the presence of a control antibody (TC ? [DCact ? Fb ? ctr ab]), a neutralizing anti–IL-6 (TC ? [DCact ? Fb ? aIL-6]), or
anti–IL-23 antibody (TC ? [DCact ? Fb ? aIL-23]). After 5 days, IL-17A was quantified by ELISA. *#P ? .05 (n ? 7 independent experiments). (G) IL-17A secretion of CD3?
Pan TCs, CD4?TCs, CD4?CD45RO?memory TCs, and CD4?CD45RA?naive TCs was compared by culturing the different TC types with medium (aTC), supernatants of
LPS-preactivated DC (aTC ? [DCact]sn), or supernatants of DCact-fibroblast coculture (aTC ? [DCact ? Fb]sn) in the presence of CD3/CD28 beads for 5 days. *P ? .05,
compared with aTC. #P ? .01, compared with aTC[DCact]sn (n ? 4 independent experiments).
HUMAN FIBROBLASTS1721BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
place, whereas expansion of Th17 cells from memory TCs could
Cox-2 expression in fibroblasts is enhanced in psoriatic skin
To demonstrate whether the fibroblast-mediated stimulation of
IL-23 in DC via PGE2may play a role during inflammation, we
analyzed Cox-2 expression by immunohistochemistry in healthy
skin and in psoriatic skin, a chronic inflammatory skin disease
where the importance of the IL-23/Th17 axis was already shown.18
In healthy skin samples, no or only a very rare Cox-2 expression
was detectable in the epidermis (Figure 6A). In contrast, in lesional
psoriatic skin, considerable Cox-2 expression was seen in the
epidermis as well as in the dermal compartment (Figure 6C-D).
Quantitative analysis revealed a significantly higher (P ? .001)
Cox-2 expression in psoriatic skin (Figure 6E). To explore whether
fibroblasts are involved in PGE2secretion in psoriatic skin, tissue
sections were stained with an anti–Thy-1 antibody to identify
fibroblasts (blue-gray) and an anti–Cox-2 antibody (red; Figure
6F-G). We observed single Cox-2- and Thy-1-positive cells under-
lining the specificity of the double-labeling procedure. In addition,
we observed double-positive cells (arrows) indicating that fibro-
blasts are one source of PGE2in psoriatic skin.
In recent years, the influence of the stromal microenvironment on
function and cytokine production of DCs was documented in
several systems.6-9Regulation of these processes is carried out
through tissue stromal cells, such as fibroblasts, macrophages, and
endothelial cells.4Recently, we demonstrated a close colocalization
of fibroblasts and DCs during cutaneous immune responses in
situ,33and several effects of fibroblasts on the function of DCs were
The influence of the stromal microenvironment on DC function
is exerted under both steady-state and inflammatory conditions.4To
imitate inflammatory conditions, DCs were preactivated with LPS
for 3 hours (DCact) in our study and subsequently cocultured with
skin-derived fibroblasts. Interestingly, fibroblasts augmented IL-23
release of preactivated DCs. In contrast, immature DCs alone or in
coculture with fibroblasts did not secrete IL-23, suggesting that
Figure 6. Cox-2 expression is enhanced in lesional psoriatic skin. Cox-2 expression (red) was detected by immunohistochemistry in (A) healthy skin and (C-D) in psoriatic
skin. Nuclei were stained by hematoxylin (blue). (B)An isotype control antibody in lesional psoriatic skin served as negative control. (E) For quantification of Cox-2 expression,
morphometric image analysis was performed by scanning 5 images per sections.The sum of pixels was evaluated, and the mean plus or minus SD of the number of pixels from
6 different psoriatic skin samples and 6 healthy skin samples is shown. (F-G) In lesional psoriatic skin, fibroblasts were detected by an anti–Thy-1 antibody (blue-gray) and
Cox-2 expression by the anti–Cox-2 antibody (red).Arrows indicate Thy-1/Cox-2 double-positive cells.
1722 SCHIRMER et al BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
under steady-state conditions IL-23 production of DCs may not be
affected by fibroblasts.Although knowing that fibroblasts per se do
not secrete any IL-23,26we confirmed that also in our cocultures
IL-23 expression was restricted to DCactand was not found in
fibroblasts. Fibroblasts also stimulated IL-23 mRNA expression in
DCact, indicating that the boost of IL-23 secretion in DCacton
fibroblast-stimulation is transcriptionally regulated.
Fixation of fibroblasts revealed that cell-cell contacts were not
responsible for fibroblast-stimulated IL-23 release. However, we
cannot completely exclude an involvement of cellular interactions
because fixation could destroy functional epitopes on fibroblast
surface. In addition, supernatants from fibroblasts were not able to
substitute for the effect of fibroblasts in the coculture with DCact.
Thus, a more complex network seems to be responsible for
fibroblast-stimulated IL-23 release from DCact, including the
secretion of factors from DCs resulting in a subsequent stimulation
of fibroblasts, which then in turn boost IL-23-release in DCs.
Because the proinflammatory cytokines TNF-? and IL-1? are
known to be present in inflamed skin31,32and are produced by
activated DCs,16we studied the role of these cytokines in the
stimulation of IL-23 secretion from DCact on coculture with
fibroblasts. PCR analyses of coculture DCactand coculture fibro-
blasts demonstrated that both cytokines are exclusively produced
by DCact. We could show that IL-1? and, to a lesser extent, TNF-?
secreted by DCactaccount for the first step of the IL-23-promoting
mechanism because anti–IL-1? led to a significant decrease of
IL-23 secretion of fibroblast-stimulated DCact. Notably, blocking of
TNF-? alone led to a small reduction of IL-23 production, whereas
neutralization of IL-1? alone was sufficient to inhibit IL-23
secretion completely. Because the combination of both blocking
antibodies had no significant additive effect, we assume that IL-1?
is the major mediator to trigger IL-23 release in DCact. It is however
possible that TNF-? stimulates IL-1? production of DCactin an
Next, we explored which mediator released by fibroblasts on
stimulation by DCactin turn up-regulates IL-23 secretion in DCact.
Previously, we had shown that stimulation of fibroblasts with
TNF-? and IL-1? results in PGE2secretion.3Furthermore, stimula-
tion of DCactwith PGE2resulted in increased IL-23 secretion. This
is in accordance with Sheibanie et al who suggested that PGE2
induces IL-23 production in DCs.35Indeed, blocking of PGE2
secretion by indomethacin completely blocked IL-23 release in the
coculture of DCactand fibroblasts to IL-23 levels produced by DCact
alone. Moreover, Cox-2 mRNA was increased 10-fold in fibro-
blasts cocultured with DCact.
Although both DCactas well as fibroblasts in our cocultures
expressed Cox-2, we identified fibroblast-derived PGE2 as the
responsible PGE2for stimulation of IL-23 from DCact. Specifically,
a marked suppression of IL-23 secretion in DCactcocultured with
Cox-2–silenced fibroblasts was observed. In contrast, inhibition of
PGE2secretion from DCactby siRNAsilencing did not affect IL-23
secretion from DCacton interaction with fibroblasts. The control of
transfection efficiency confirmed a similar effectiveness of inhibi-
tion of PGE2secretion in DCactand fibroblasts. We realize that
IL-23 release in DCactcocultured with Cox-2 siRNA transfected
fibroblasts could not be blocked completely to levels observed in
DCact. We suppose that this is because of the incomplete Cox-2
mRNA knockdown in fibroblasts. This notion is supported by the
finding that addition of indomethacin to DCact-fibroblast coculture
was sufficient to completely block PGE2and decreasing IL-23
levels to that of DCactcultured alone. In summary, our data indicate
that preactivation of DC by a 3-hour pulse with LPS induces IL-1?
in DCs, which (possibly in combination with TNF-?) promotes
PGE2release in fibroblasts resulting in an up-regulation of IL-23 in
Finally, we looked at the consequences of the enhanced IL-23
secretion by fibroblast-stimulated DCact on the differentiation/
polarization of TCs. In accordance with Acosta-Rodriguez et al36
and Zenaro et al,37DCactalone induced a remarkable Th1 response
characterized by IFN-? secretion but a weak Th17 response as
shown by IL-17 and IL-22 secretion. In sharp contrast, fibroblast-
stimulated DCactsupported the emergence of IL-17A–producing
TCs as shown by an increase in IL-17A secretion, an expansion of
IL-17A–secreting cells, as well as an enhanced secretion of IL-22
and up-regulated expression of ROR?t. Because TGF-? and IL-1?
were not differentially expressed in fibroblast-stimulated DCact
versus DCactcultured alone, we excluded these mediators as the
responsible factors for the fibroblast-driven enhancement of the
Th17 response. Romagnani et al21described that IL-23 and IL-1?
seem to drive development of human naive TCs from cord blood or
memory TC into Th17 cells, but the function of TGF-? and IL-6 in
this process is discussed controversially.21,27,36The central role of
fibroblast-stimulated IL-23 secretion of DCactin the support of
Th17 expansion was confirmed by specific antibody blocking of
IL-23 in the coculture of fibroblasts and DCact, which almost
completely prevented the boost of IL-17A release in TCs. In
addition, blocking IL-6 only slightly reduced IL-17A secretion
fromTCs cocultured with fibroblast-stimulated DCact.We conclude
that, in our system, IL-23 secreted by fibroblast-stimulated DCactis
the major cytokine-promoting elevated IL-17A production in TCs.
However, fibroblast-stimulated DCactalso promoted an increased
IFN-? secretion inTCs cocultured with fibroblast-stimulated DCact.
IL-12 is the most important factor known to stimulate IFN-?
expression in TCs. Indeed, in our experiments, IL-12 release was
also up-regulated in fibroblast-stimulated DCact. The simultaneous
stimulation of a Th1- and Th17-type response was also described
by other investigators.37
In accordance with our data, Sheibanie et al showed that PGE2
induced IL-23 up-regulation in DCs, resulting in increased produc-
tion of IL-17 from activated CD4?T cells.35Khayrullina et al
reported that bone marrow-derived DCs generated in the presence
of PGE2promote Th17 differentiation in vitro and in vivo via
IL-23.38On the other hand, several reports show that PGE2directly
increases IL-17A production in activated TCs.39-41It should be
noted, however, that the direct effect of PGE2on Th17 differentia-
tion was only achieved by concentrations of PGE210- to 1000-fold
higher than those detected in our cocultures of fibroblasts and
DCact.Another piece of evidence also indicates that, in our system,
Figure 7. Model how fibroblasts could stimulate IL-23 secretion from preacti-
vated DCs resulting in expansion of Th17 cells. (1) Preactivation of DC (LPS).
(2) Then, DCs secrete IL-1? and TNF-?, which stimulate fibroblasts. (3) Stimulated
fibroblasts secrete PGE2, which acts on DCs and therewith increases their IL-23
production. (4) Through the action of elevated IL-23, CD3?(Pan), CD4?, and
CD4?CD45RO?(memory) TCs are supported to produce IL-17A, IL-22, and to
express ROR?t as well as to secrete IFN-?.
HUMAN FIBROBLASTS1723BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
IL-17Ais not the result of a direct effect of PGE2on TCs. Notably,
anti–IL-23p19 blocked almost completely IL-17A release of TCs
cocultured with fibroblast-stimulated DCact, although not affecting
Whether IL-23 is able to induce IL-17A secretion in both
CD4?CD45RO?(memory) and CD4?CD45RA?(naive) TCs or
only in memory TCs is still debated.27,36We could show that our
cell culture supernatants of fibroblast-stimulated DCactor DCact
alone containing IL-23, IL-1?, IL-6, and TGF-? support the
emergence of Th17 cells from Pan TC, CD4?TCs as well as
memory TCs. In naive TCs, only minute amounts of IL-17A could
be detected. Evidently, under our experimental conditions, DCs
activated by LPS with and without fibroblasts do not drive Th17
polarization from naive TCs. This is in accordance with Acosta-
Rodriguez et al36who failed to induce Th17 differentiation from
naive TCs with LPS-stimulated monocyte-derived DCs.
Finally, we could demonstrate that this mechanism may be
involved in psoriasis, a prototypic immune response for which the
importance of IL-23/Th17 immune axis has been shown.18Specifi-
cally, in psoriasis, overexpression of IL-23p19 mainly by mono-
cytes and DCs and expansion of Th17 cells have been reported.42,43
The IL-23/Th17 axis is involved in the development of the psoriatic
sia, acanthosis, and hyperkerathosis.18,44We observed a signifi-
cantly higher Cox-2 expression in psoriatic skin. By double-
labeling, fibroblasts were identified as one source of PGE2 in
psoriasis. Taken together with our previous finding of close
apposition of fibroblasts and DCs in lesional psoriatic skin,33we
speculate that fibroblasts, via the secretion of PGE2, are involved in
the perpetuation of the immune response in psoriasis by promoting
IL-23 secretion of DCs and a subsequent expansion of Th17 cells.
In addition, keratinocytes and other yet unidentified dermal cells
also express Cox-2 and may act via similar mechanisms in
stimulation of IL-23 secretion from inflammatory DCs. As of yet,
little information about the effectiveness of Cox inhibition as an
antipsoriatic treatment exist (ie, one study reported no improve-
ment of skin lesions); on the other hand, Cox inhibitors are used
widely and successfully to treat psoriatic arthritis.45-47Bearing in
mind that Cox-blocking agents drive arachidonate metabolism into
the lipoxygenase pathway resulting in enhanced production of
leukotrienes, including potent chemoattractants for inflammatory
cells in psoriasis,48it is possible that blocking of just the Cox
pathway is not sufficient for successful treatment.
In conclusion, our results support a model (Figure 7) in which
under inflammatory conditions DCs produce TNF-? and Il-1?,
which in turn activate resident fibroblasts. Via the secretion of
different factors, these fibroblasts modulate DC functions and
finally the polarization of the TC response. In particular, we
demonstrated that PGE2released by activated fibroblasts stimulates
IL-23 secretion from activated DCs, which results ultimately in a
prominent expansion of Th17 cells from the memory pool of TCs.
The authors thank Ms Heidi Gedicke for technical assistance.
This work was kindly supported by Helmholtz Impulse and
Networking Fund through the Helmholtz Interdisciplinary Gradu-
ate School for Environmental Research (HIGRADE) and in part by
the German Research Council (TRR 67 Si 397/15-1).
Contribution: C.S., C.K., and A.S designed and performed experi-
ments and analyzed data; C.S., A.S., and J.C.S. wrote the paper;
and C.S., C.K.,A.S., J.C.S, and M.V.B. discussed the data and read
and edited the paper.
Conflict-of-interest disclosure: The authors declare no compet-
ing financial interests.
Correspondence: Anja Saalbach, Department of Dermatology,
Venerology and Allergology, Medical Faculty of the University
Leipzig, Johannisallee 30, 04103 Leipzig, Germany; e-mail:
1. Cheng P, Nefedova Y, Corzo CA, Gabrilovich DI.
Regulation of dendritic-cell differentiation by bone
marrow stroma via different Notch ligands. Blood.
2. Despars G, Tan J, Periasamy P, O’Neill HC. The
role of stroma in hematopoiesis and dendritic cell
development. Curr Stem Cell Res Ther. 2007;
3. SaalbachA, Klein C, Schirmer C, et al. Dermal
fibroblasts promote the migration of dendritic
cells. J Invest Dermatol. 2010;130(2):444-454.
4. Svensson M, Kaye PM. Stromal-cell regulation of
dendritic-cell differentiation and function. Trends
5. Tang H, Guo Z, Zhang M, et al. Endothelial
stroma programs hematopoietic stem cells to dif-
ferentiate into regulatory dendritic cells through
IL-10. Blood. 2006;108(4):1189-1197.
6. RenklAC, Wussler J,Ahrens T, et al. Osteopontin
functionally activates dendritic cells and induces
their differentiation toward a Th1-polarizing phe-
notype. Blood. 2005;106(3):946-955.
7. Termeer C, Sleeman JP, Simon JC. Hyaluronan:
magic glue for the regulation of the immune re-
sponse? Trends Immunol. 2003;24(3):112-114.
8. Zhang M, Tang H, Guo Z, et al. Splenic stroma
drives mature dendritic cells to differentiate into
regulatory dendritic cells. Nat Immunol. 2004;
9. Sangaletti S, Gioiosa L, Guiducci C, et al.Accel-
erated dendritic-cell migration and T-cell priming
in SPARC-deficient mice. J Cell Sci. 2005;
10. Buckley CD, Pilling D, Lord JM, et al. Fibroblasts
regulate the switch from acute resolving to
chronic persistent inflammation. Trends Immunol.
11. Buckley CD, FilerA, Haworth O, Parsonage G,
Salmon M. Defining a role for fibroblasts in the
persistence of chronic inflammatory joint disease.
Ann Rheum Dis. 2004;63(suppl 2):ii92-ii95.
12. Parsonage G, Falciani F, BurmanA, et al. Global
gene expression profiles in fibroblasts from syno-
vial, skin and lymphoid tissue reveals distinct cy-
tokine and chemokine expression patterns.
Thromb Haemost. 2003;90(4):688-697.
13. Chang HY, Chi JT, Dudoit S, et al. Diversity, topo-
graphic differentiation, and positional memory in
human fibroblasts. Proc Natl Acad Sci U S A.
14. Ouwehand K, Santegoets SJ, Bruynzeel DP, et
al. CXCL12 is essential for migration of activated
Langerhans cells from epidermis to dermis. Eur
J Immunol. 2008;38(11):3050-3059.
15. Briard D,Azzarone B, Brouty-Boye D. Importance
of stromal determinants in the generation of den-
dritic and natural killer cells in the human spleen.
Clin Exp Immunol. 2005;140(2):265-273.
16. Merad M, Manz MG. Dendritic cell homeostasis.
17. Dong C. TH17 cells in development: an updated
view of their molecular identity and genetic pro-
gramming. Nat Rev Immunol. 2008;8(5):337-348.
18. Di CesareA, Di Meglio P, Nestle FO. The IL-23/
Th17 axis in the immunopathogenesis of psoria-
sis. J Invest Dermatol. 2009;129(6):1339-1350.
19. Boniface K, Bernard FX, Garcia M, et al. IL-22
inhibits epidermal differentiation and induces
proinflammatory gene expression and migration
of human keratinocytes. J Immunol. 2005;174(6):
20. Wolk K, Witte E, Wallace E, et al. IL-22 regulates
the expression of genes responsible for antimi-
crobial defense, cellular differentiation, and mobil-
ity in keratinocytes: a potential role in psoriasis.
Eur J Immunol. 2006;36(5):1309-1323.
21. Romagnani S, Maggi E, Liotta F, Cosmi L,
Annunziato F. Properties and origin of human
Th17 cells. Mol Immunol. 2009;47(1):3-7.
22. Happel KI, Dubin PJ, Zheng M, et al. Divergent
roles of IL-23 and IL-12 in host defense against
Klebsiella pneumoniae. J Exp Med. 2005;202(6):
23. Kleinschek MA, Muller U, Brodie SJ, et al. IL-23
enhances the inflammatory cell response in Cryp-
tococcus neoformans infection and induces a cy-
tokine pattern distinct from IL-12. J Immunol.
1724 SCHIRMER et al BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10
24. Annunziato F, Cosmi L, Santarlasci V, et al. Phe-
notypic and functional features of human Th17
cells. J Exp Med. 2007;204(8):1849-1861.
25. Cua DJ, Sherlock J, Chen Y, et al. Interleukin-23
rather than interleukin-12 is the critical cytokine
for autoimmune inflammation of the brain. Nature.
26. Langrish CL, McKenzie BS, Wilson NJ, et al.
IL-12 and IL-23: master regulators of innate and
adaptive immunity. Immunol Rev. 2004;202:96-
27. Wilson NJ, Boniface K, Chan JR, et al. Develop-
ment, cytokine profile and function of human in-
terleukin 17-producing helper T cells. Nat Immu-
28. SaalbachA, Kraft R, Herrmann K, Haustein UF,
Anderegg U. The monoclonal antibodyAS02 rec-
ognizes a protein on human fibroblasts being
highly homologous to Thy-1. Arch Dermatol Res.
29. SaalbachA,Aust G, Haustein UF, Herrmann K,
Anderegg U. The fibroblast-specific MAbAS02: a
novel tool for detection and elimination of human
fibroblasts. Cell Tissue Res. 1997;290(3):593-
30. Polte T, BehrendtAK, Hansen G. Direct evidence
for a critical role of CD30 in the development of
allergic asthma. J Allergy Clin Immunol. 2006;
31. Aggarwal BB, Shishodia S, Takada Y, et al. TNF
blockade: an inflammatory issue. Ernst Schering
Res Found Workshop. 2006;56:161-186.
32. Taylor PC, Williams RO, Feldmann M. Tumour
necrosis factor alpha as a therapeutic target for
immune-mediated inflammatory diseases. Curr
Opin Biotechnol. 2004;15(6):557-563.
33. SaalbachA, Klein C, Sleeman J, et al. Dermal
fibroblasts induce maturation of dendritic cells.
J Immunol. 2007;178(8):4966-4974.
34. Blanco P, PaluckaAK, Pascual V, Banchereau J.
Dendritic cells and cytokines in human inflamma-
tory and autoimmune diseases. Cytokine Growth
Factor Rev. 2008;19(1):41-52.
35. SheibanieAF, Tadmori I, Jing H, Vassiliou E,
Ganea D. Prostaglandin E2 induces IL-23 pro-
duction in bone marrow-derived dendritic cells.
FASEB J. 2004;18(11):1318-1320.
Sallusto F. Interleukins 1beta and 6 but not trans-
forming growth factor-beta are essential for the
differentiation of interleukin 17-producing human
T helper cells. Nat Immunol. 2007;8(9):942-949.
immune response by Mycobacterium tuberculo-
sis: role of dectin-1, mannose receptor, and DC-
SIGN. J Leukoc Biol. 2009;86(6):1393-1401.
38. Khayrullina T, Yen JH, Jing H, Ganea D. In vitro
differentiation of dendritic cells in the presence of
prostaglandin E2 alters the IL-12/IL-23 balance
and promotes differentiation of Th17 cells. J Im-
39. Boniface K, Bak-Jensen KS, Li Y, et al. Prosta-
glandin E2 regulates Th17 cell differentiation and
function through cyclicAMP and EP2/EP4 recep-
tor signaling. J Exp Med. 2009;206(3):535-548.
40. Chizzolini C, Chicheportiche R,Alvarez M, et al.
Prostaglandin E2 synergistically with interleu-
kin-23 favors human Th17 expansion. Blood.
Sallusto F. Prostaglandin E2 enhances Th17 re-
sponses via modulation of IL-17 and IFN-gamma
production by memory CD4? T cells. Eur J Im-
42. Lee E, Trepicchio WL, Oestreicher JL, et al. In-
creased expression of interleukin 23 p19 and p40
in lesional skin of patients with psoriasis vulgaris.
J Exp Med. 2004;199(1):125-130.
43. Lowes MA, Kikuchi T, Fuentes-Duculan J, et al.
Psoriasis vulgaris lesions contain discrete popu-
lations of Th1 and Th17 T cells. J Invest Derma-
44. Boniface K, Guignouard E, Pedretti N, et al.Arole
for T cell-derived interleukin 22 in psoriatic skin
inflammation. Clin Exp Immunol. 2007;150(3):
45. Gordon KB, Ruderman EM. The treatment of pso-
riasis and psoriatic arthritis: an interdisciplinary
approach. J Am Acad Dermatol. 2006;54(3 suppl
46. Sheehan-Dare RA, Goodfield MJ, Rowell NR.
The effect of oral indomethacin on psoriasis
treated with the Ingram regime. Br J Dermatol.
47. Soriano ER, Rosa J. Update on the treatment of
peripheral arthritis in psoriatic arthritis. Curr
Rheumatol Rep. 2009;11(4):270-277.
48. Friedman ES, LaNatra N, Stiller MJ. NSAIDs in
dermatologic therapy: review and preview. J Cu-
tan Med Surg. 2002;6(5):449-459.
HUMAN FIBROBLASTS1725 BLOOD, 9 SEPTEMBER 2010?VOLUME 116, NUMBER 10