Download Bronchial Epithelial Cell-Derived Prostaglandin E2 Dampens the

Bronchial Epithelial Cell-Derived
Prostaglandin E 2 Dampens the Reactivity of
Dendritic Cells
This information is current as
of August 9, 2017.
Lotte M. Schmidt, Maria G. Belvisi, Konrad A. Bode, Judith
Bauer, Claudia Schmidt, Maria-Theresia Suchy, Dimitrios
Tsikas, Jutta Scheuerer, Felix Lasitschka, Herman-Josef
Gröne and Alexander H. Dalpke
References
Subscription
Permissions
Email Alerts
This article cites 55 articles, 14 of which you can access for free at:
http://www.jimmunol.org/content/186/4/2095.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2011 by The American Association of
Immunologists, Inc. All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
J Immunol 2011; 186:2095-2105; Prepublished online 12
January 2011;
doi: 10.4049/jimmunol.1002414
http://www.jimmunol.org/content/186/4/2095
The Journal of Immunology
Bronchial Epithelial Cell-Derived Prostaglandin E2 Dampens
the Reactivity of Dendritic Cells
Lotte M. Schmidt,* Maria G. Belvisi,† Konrad A. Bode,* Judith Bauer,*
Claudia Schmidt,‡ Maria-Theresia Suchy,x Dimitrios Tsikas,x Jutta Scheuerer,{
Felix Lasitschka,{ Herman-Josef Gröne,‡ and Alexander H. Dalpke*
B
ecause airways are frequently exposed to a variety of
inhaled Ags and microbes, local immune reactivity has to
be tightly regulated. Recently, the concept emerged that
epithelial cells exert simple immune functions (1, 2). Airway
epithelial cells express different pattern recognition receptors (3)
and respond to microbes by induction of immunological mediators. It has been proposed that epithelial cells create an anti-inflammatory microenvironment that modulates the phenotype of
local APCs (4). Similar observations within the intestine support
the concept that the local microenvironment of individual organs
induces organ-specific immune response, with the degree of microbial contact being an important restriction variable (5).
Respiratory tract dendritic cells (DCs) represent only a small cell
population, and different DC subtypes have been identified that
can be separated based on their location as well as their function
(6). APCs can be activated under different conditions, and at least
for macrophages, two seemingly opposing activation modes have
been identified. IFN-g induces classically activated macrophages
*Department of Infectious Diseases—Medical Microbiology and Hygiene, University of Heidelberg, 69120 Heidelberg, Germany; †Respiratory Pharmacology Group,
Pharmacology and Toxicology Section, National Heart and Lung Institute, Imperial
College London, SW7 2AZ London, United Kingdom; ‡Cellular and Molecular
Pathology, German Cancer Research Center, 69120 Heidelberg, Germany; xInstitute
of Clinical Pharmacology, Hannover Medical School, 30625 Hannover, Germany;
and {Institute of Pathology, University of Heidelberg, 69120 Heidelberg, Germany
Received for publication July 20, 2010. Accepted for publication December 5, 2010.
This work was supported by German Research Foundation Grants Da 592/3 and Da
592/4 (to A.H.D.) and by Medical Research Council (U.K.) Grant G0800195 (to M.B.).
Address correspondence and reprint requests to Dr. Alexander H. Dalpke, Department of Infectious Diseases—Medical Microbiology and Hygiene, University of
Heidelberg, Im Neuenheimer Feld 324, 69120 Heidelberg, Germany. E-mail address:
[email protected]
Abbreviations used in this article: BMDC, bone marrow-derived dendritic cell; Cox,
cyclooxygenase; Ct, cycle threshold; d4-PGE2, [3,39,4,49-2H4]-PGE2; DC, dendritic
cell; ECM, epithelial cell-conditioned medium; EP, E-prostanoid; GC-MS/MS, gas
chromatography-mass spectrometry/mass spectrometry; PGEM, PGE metabolite.
Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00
www.jimmunol.org/cgi/doi/10.4049/jimmunol.1002414
that are characterized by a vigorous inflammatory response; in
contrast, IL-10 and PGE2 induce alternatively activated macrophages (7). The latter cells express several marker genes, such as
chitinase-like lectin Ym1, resistin-like secreted protein Fizz-1,
and arginase-1, as well as the anti-inflammatory cytokine IL-10
(8). A similar phenotype exists in tolerogenic DCs, which are
immature DCs that are modulated by suppressive factors (9).
Prostanoids are soluble lipid mediators that are produced in two
enzymatic steps from C20-unsaturated fatty acids. Cyclooxygenase
(Cox)-1 and Cox-2 form the unstable intermediates PGG2 and
PGH2. PGH2 is then converted by specific enzymes to the different
PGs and thromboxane A2 (10). PGE2 has a variety of biological
effects, with DCs exposed to PGE2 having reduced secretion capacity of proinflammatory cytokines (11, 12). Prostanoids are
potent immune modulators and their production is tightly controlled to avoid damage of neighboring cells (10). E-prostanoid
(EP) receptors mediate the biological effects of PGE2, and knockout studies have revealed that PGE2 exerts not only proinflammatory, but also anti-inflammatory actions depending on the specific context (13–15). EP1 and EP3 receptors signal through Ca2+
levels, whereas EP2 and EP4 receptors elevate the level of cAMP
in cells (10, 13).
In this study, we investigated how airway epithelial cells induce
an inhibitory DC phenotype. Specifically, we set out to identify
and characterize soluble factors that are secreted by epithelial cells
and that downregulate the secretion of proinflammatory cytokines
by DCs. We were able to identify PGE2 as one of those factors
and characterized its effects in the interplay of airway epithelial
cells with DCs.
Materials and Methods
Reagents and Abs
RPMI 1640 and a 1:1 mix of DMEM/Ham’s F12 medium were obtained
from Biochrom (Berlin, Germany). FBS was from BioWest (Nuaillé,
France). PBS, penicillin, and streptomycin were obtained from PAA (Coelbe,
Germany). LPS from Salmonella minnesota was provided by U. Seydel
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
Airway epithelial cells regulate immune reactivity of local dendritic cells (DCs), thus contributing to microenvironment homeostasis.
In this study, we set out to identify factors that mediate this regulatory interaction. We show that tracheal epithelial cells secrete
soluble factors that downregulate TNF-a and IL-12p40 secretion by bone marrow-derived DCs but upregulate IL-10 and arginase1. Size exclusion chromatography identified small secreted molecules having high modulatory activity on DCs. We observed that
airway tracheal epithelial cells constitutively release the lipid mediator PGE2. Blocking the synthesis of PGs within airway
epithelial cells relieved DCs from inhibition. Cyclooxygenase-2 was found to be expressed in primary tracheal epithelial cell
cultures in vitro and in vivo as shown by microdissection of epithelial cells followed by real-time PCR. Paralleling these findings we
observed that DCs treated with an antagonist for E-prostanoid 4 receptor as well as DCs lacking E-prostanoid 4 receptor showed
reduced inhibition by airway epithelial cells with respect to secretion of proinflammatory cytokines measured by ELISA. Furthermore, PGE2 mimicked the effects of epithelial cells on DCs. The results indicate that airway epithelial cell-derived PGE2
contributes to the modulation of DCs under homeostatic conditions. The Journal of Immunology, 2011, 186: 2095–2105.
2096
(Research Center Borstel, Borstel, Germany). DNAse I was from Roche
(Mannheim, Germany) and protein A/G Plus-agarose beads were from
Santa Cruz Biotechnology (Heidelberg, Germany). Murine Cox-1 and
Cox-2 polyclonal Abs and EP antagonists GW 627368X and AH6890
were purchased from Cayman Chemicals (Ann Arbor, MI). The polyclonal
Cox-2 Ab used for immunoenzyme staining was from Abcam. IL-10 was
from ImmunoTools (Friesoythe, Germany), and the IL-10 blocking Ab
as well as CD3 and CD28 Abs were from eBioscience (Frankfurt, Germany).
PGE2 was from Sigma-Aldrich (Taufkirchen, Germany), and the Cox
inhibitors NS-398 and SC-560 were obtained from Merck (Mannheim,
Germany). CD4 T cell isolation kit II was from MACS Miltenyi Biotec
(Bergisch-Gladbach, Germany).
Preparation of primary airway epithelial cells
The isolation and culture of tracheal epithelial cells were performed with
small adaptations as previously described (4). Four- to 10-wk-old female
C57BL/6 mice (Charles River Laboratories) were sacrificed by CO2
treatment, and tracheae were prepared and digested with pronase E and
DNAse I overnight. Cell suspensions were allowed to adhere for 2–3 h in
a petri dish at 37˚C. Nonadherent cells were grown for 4–7 d until confluence was reached in a Transwell system on collagen-precoated membranes (Costar). They were further cultured as air–liquid interface cultures.
DCs were differentiated from bone marrow from 4- 10-wk-old female
BALB/c or C57BL/6 mice as previously described (16). Culture supernatant
of a GM-CSF–transfected cell line was used as a source of GM-CSF. EP
knockout cells were obtained as frozen bone marrow from Maria Belvisi
(London). DCs from mice devoid of one of the following genes—Ptger1
(EP1), Ptger2 (EP2), Ptger3 (EP3), or Ptgdr (DP)—had been backcrossed
at least eight times onto the C57BL/6 background. Ptger42/2 (EP4) mice
do not survive on the C57BL/6 background due to patent ductus arteriosus,
and so they were backcrossed on a mixed background of 129/Ola 3
C57BL/6 mice. Mice were provided by Dr. Shuh Narumiya, Kyoto University, and breeding colonies were maintained at Imperial College, London.
At day 8, 2 3 105 bone marrow-derived DCs (BMDCs) were seeded into
96-well plates and incubated with 50% (v/v) epithelial cell-conditioned
medium (ECM). ECM derived of cells pretreated with Cox inhibitors NS398 (1mM), SC-560 (1mM), or with EP antagonists (EP4, GW627368X or
EP2, AH6890, 10 mM) was used equally. BMDCs were stimulated with LPS
(100 ng/ml) overnight, and the supernatants were analyzed by ELISA
(OptEIA ELISA; BD Biosciences). Where indicated, cells were treated
with IL-10 (40 ng/ml) or an IL-10–blocking Ab (10 mg/ml).
Measurement of PGE metabolites (PGE2)
PGE2 metabolites were measured in 48-h supernatants of airway epithelial
cells with a commercially available PGE metabolite (PGEM) ELISA from
Cayman Chemicals. This competitive assay converts the intermediates 13,
14-dihydro-15-keto PGA2 and 13,14-dihydro-15-keto PGE2 into stable
derivates.
Gas chromatography-mass spectrometry/mass spectrometry
identification and quantification of PGE2 in cultured epithelial
cells
PGE2 from cultured epithelial cells was analyzed by gas chromatographymass spectrometry/mass spectrometry (GC-MS/MS) as described elsewhere (17, 18). The internal standard [3,39,4,49-2H4]-PGE2 (d4-PGE2; 98
atom% 2H) was added to 0.9-ml aliquots of samples, resulting in a final
concentration of 500 pg/ml. Samples were applied directly to 4 ml PGE2immunoaffinity columns. d4-PGE2 and PGE2-immunoaffinity columns
were obtained from Cayman Chemicals. Columns were washed with 2 ml
column buffer (0.1 M potassium phosphate buffer [pH 7.4], 7.7 mM NaN3,
0.5 M NaCl), followed by 2 ml distilled water. Compounds were eluted by
allowing 2 ml elution solution, consisting of absolute ethanol/distilled
water (95:5, v/v) to pass. Solvents were removed under a stream of nitrogen, and the pentafluorobenzyl ester methoxyamine trimethylsilyl ether
derivatives were prepared. GC-MS/MS analyses were performed on a
ThermoQuest TSQ 7000 triple-stage quadrupole mass spectrometer interfaced with a ThermoQuest gas chromatograph model Trace 2000 (ThermoQuest, Egelsbach, Germany). Fused silica capillary columns (Optima
35MS; 30 3 30.25 mm interior diameter, 0.25 mm film thickness) from
Macherey–Nagel (Düren, Germany) were used. The product ions at m/z
268 for PGE2 and m/z 272 for d4-PGE2, which were generated by
collision-induced dissociation of the parent ions [M-pentafluorobenzyl]2
at m/z 524 and m/z 528, respectively, were monitored in the selectedreaction monitoring mode. The dwell time was 400 ms for each ion.
Peak areas were used for calculation of concentrations.
Quantitative RT-PCR
Total RNA was isolated using a High Pure RNA isolation kit (Roche) and
was reverse transcribed with a cDNA synthesis kit (Verso cDNA kit;
Thermo Fisher Scientific). Two microliters of cDNA (diluted 1:4) was used
as template in a total reaction volume of 20 ml (quantitative PCR mix;
Abgene) and analyzed on ABI Prism 7700 (Applied Biosystems). Quantifications were made using SYBR Green, including no-template and noRT controls. Gene expression was measured in duplicates, and automatic
detection of baseline and threshold values was used. Cycle threshold (Ct)
values were subtracted from the Ct values of housekeeping genes (b-actin,
GAPDH), resulting in DCt for each target gene, which was then used to
calculate the relative expression (22DCt). All primer sequences are available on request.
Size exclusion chromatography
Forty-eight hour epithelial supernatant was analyzed on an ÄKTAprime
(Amersham Pharmacia, Uppsala, Sweden) using a 200-kDa Sephacryl S100 column (GE Healthcare) with a pressure of 0.24 MPa and a speed of
1.6 ml/min. To compare and to get a size for the different fractions, we also
used the Gel Filtration Calibration Kit LMW (GE Healthcare). The
resulting fractions were collected and 50% (v/v) were added to BMDCs,
which were then stimulated with LPS.
Western blotting
Protein concentrations of cell lysates were measured by Pierce assay
(Thermo Scientific), and equal amounts were fractionated by 12% SDSPAGE and electrotransferred to polyvinylidene difluoride membranes.
Unspecific binding was blocked by TBS (pH 7.8) with 5% nonfat dry milk
and 0.05% Tween 20. After blocking, proteins were visualized using
specific Abs, HRP-marked secondary Abs, and ECL (Amersham Pharmacia).
Immunoprecipitation of Cox-2
In vitro-cultured epithelial cells were stimulated with 100 ng/ml LPS for 4 h.
Cells were harvested, pelleted at 4˚C and lysed in RIPA buffer containing 1
mg/ml proteinase inhibitors (aprotinin, leupeptin, and pepstatin), 1 mM
Na3VO4, NaF, and PMSF for 20 min on ice. The lysates were treated with
30 ml protein A/G Plus-agarose beads and 5 ml Cox-2 Ab and incubated at
4˚C in an overhead shaker for 4 h. Beads were washed three times (13,000
rpm, 4˚C, 1 min) with RIPA buffer containing 2% SDS. Samples were
incubated at 95˚C for 5 min, centrifuged again, and the supernatant was
taken off and used for Western blot analysis.
Immunoenzyme staining
Immunoenzyme staining of Cox-2 was performed on 4-mm cryostat
sections of fresh-frozen tissues, which were postfixed in 2% paraformaldehyde and further processed by the paraformaldehyde-saponin procedure in combination with the standard alkaline phosphatase anti-alkaline
phosphatase technique (Dako), as previously published (19). As primary
Ab, anti-mouse Cox-2 (1:100) (Abcam) was added overnight at room
temperature. Normal rabbit Ig (Dianova) served as negative control. A
mouse anti-rabbit mAb (1:50) (Dako), was used as secondary reagent (30
min at room temperature). Naphthol AS/BI phosphate (Sigma-Aldrich)
with new fuchsin (Merck, Darmstadt, Germany) served as substrate.
Immunofluorescence
Tracheae of female C57BL/6 mice were isolated and immediately fixed in
formalin. Paraffin sections (5 mm) were stained after Ag retrival in steamer
with citrate puffer (pH 6.0) for 20 min, cooled down to room temperature,
and then incubated with anti-cytokeratin K8 (1:50) mouse Ab (Progen
Biotechnik, Heidelberg, Germany). K8 was detected with TSA kit number
23 (Alexa Fluor 546 tyramide; Invitrogen/Molecular Probes, Eugene, OR).
This was followed by incubation with polyclonal Ab against Cox-1 (1:20)
or Cox-2 (1:10) (Cayman Chemicals), detected with the secondary Ab
Alexa Flour 488 anti-rabbit (1:200) (Invitrogen). Nuclei were depicted with
DRAQ5. Sections were mounted with fluoromount-G (SouthernBiotech).
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
Generation and stimulation of bone marrow-derived DCs
INHIBITION OF DENDRITIC CELLS BY PGE2
The Journal of Immunology
2097
Laser microdissection and pressure catapulting
Statistics
Fresh-frozen tracheal samples were cut into 18-mm-thick sections using
a cryostat (Leica CM1850; Leica Microsystems, Wetzlar, Germany) and
processed as follows. First, the sections were mounted on membrane
slides (polyethylene naphthalate membrane, 1 mm glass; Carl Zeiss MicroImaging, Bernried, Germany). For further preservation, samples were
fixed in ethanol and stained in cresyl violet acetate (1% [w/v] in American
Chemical Society-grade ethanol; Sigma-Aldrich, Munich, Germany) for
15 s. Subsequently, the slides were washed in ethanol and incubated for 5
min in xylene. After air-drying the slides were mounted on the stage of an
inverse microscope, which is a component of a MicroBeam LMPC system
(Carl Zeiss MicroImaging). We employed the RoboLPC method to microdissect and capture the tracheal epithelium (∼7 mm2 epithelium,
∼70,000–200,000 cells). For sample collection, we applied 0.5 ml AdhesiveCaps (opaque) (Carl Zeiss MicroImaging).
All experiments were repeated at least twice unless indicated otherwise.
Means + SD are shown. Significant differences were evaluated by the
unpaired Student t test with two-tailed distributions; p , 0.05 (*), p , 0.01
(**), and p , 0.001 (***) were considered significant.
RNA isolation of microdissected samples
In vitro proliferation assay
BMDCs (4 3 104) from C57BL/6 mice were incubated with or without
ECM 50% (v/v) and stimulated with LPS (100 ng/ml) or left unstimulated
overnight. Subsequently, cells were washed three times and pulsed with
100 mg/well EndoGrade OVA (Hyglos) for 3 h and washed three times
again. CD4+ T cells were isolated via MACS beads from OT-II mice, and
2 3 105 T cells were cocultured for 72 h with the pretreated DCs. After
3 d, proliferation of OT-II cells was measured by [3H]thymidine incorporation.
Epithelial cell-conditioned DCs have reduced proinflammatory
capacity and show reduced induction of T cell proliferation
Previous work of our own has shown that epithelial cells modulate
reactivity of professional immune cells (4). To further characterize
this interplay of airway epithelial cells with APCs, we used a
model system in which we preincubated BMDCs with supernatants derived from primary tracheal epithelial cells (ECM).
Prepared tracheal cells showed the typical characteristics of epithelial cells, including tight junction formation, high transepithelial resistance, tight barrier for high-molecular mass FITCdextran, and expression of specific marker genes, including defB1
and cftr; we observed no significant contamination with fibroblasts
or myeloid cells as judged by immunofluorescence and RT-PCR
(data not shown). BMDCs were stimulated with LPS to mirror
a standard innate immune response and analyzed for induction of
various inflammatory genes. We observed that ECM incubation
reduced the secretion of IL-12p40 by DCs from 70 to 10 ng/ml
(Fig. 1A). Similarly, secretion of TNF-a was significantly suppressed, whereas in contrast the secretion of anti-inflammatory
IL-10 was increased. The same results were obtained when transcriptional activity was examined by quantitative RT-PCR (data
not shown). Epithelial cell-conditioned DCs retained this pheno-
FIGURE 1. Epithelial cell conditioning inhibits DC reactivity and function. A and B, BMDCs (2 3 105) from BALB/c (A) or C57BL/6 (B) mice were
incubated overnight with LPS (100ng/ml), 50% (v/v) of 48 h supernatant of epithelial cells (ECM), or a combination thereof. Supernatants were analyzed
for the secretion of IL-12p40, TNF-a, or IL-10 by ELISA. C and D, BMDCs (8 3 105) from BALB/c (C) or C57BL/6 (D) mice were treated as above for 4
h. Cells were lysed and gene expression of arginase-1 and Ym1 was measured by quantitative RT-PCR. Depicted are expression levels normalized to the
expression of the housekeeping gene b-actin. rE, relative expression. E, C57BL/6 DCs (4 3 104) were treated as in A and B. The next day DCs were washed
intensively and loaded with OVA for 3 h. Cells were washed again, incubated with 2 3 105 OT-II T cells for 72 h, and pulsed overnight with [3H]thymidine.
T cells were also stimulated with plate-bound CD3 and CD28 Abs. F, C57BL/6 DCs (2 3 105) were treated with ECM as above, and additionally 10 mg/ml
IL-10 blocking Ab was added where indicated. The supernatant was analyzed for IL-12p40 secretion. For A, B, and F, results shown are means + SD from
three to five experiments. C and D show representative data for three independent experiments, and results in E are representative of two experiments.
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
Total RNA was isolated from each sample using the RNeasy Micro kit
(Qiagen, Hilden, Germany) according to the manufacturer’s instructions by
applying 300 ml RLT lysis buffer supplemented with 1% 2-ME (SigmaAldrich, Steinheim, Germany). For precipitation, 20 mg glycogen (from
mussels; Sigma-Aldrich) per sample was used. The quality of the RNA was
assessed on RNA 6000 Nano microfluidics chips (Agilent Technologies,
Waldbronn, Germany) on an Agilent 2100 bioanalyzer.
Results
2098
in this context. Similar results were observed with C57BL/6 DCs
(data not shown).
The inhibitory capacity of epithelial supernatant is mediated
by multiple soluble factors
Because DCs could be modulated by epithelial cell-conditioned
medium, we concluded that soluble factors contribute to the modulation of DC reactivity. Because epithelial cells did not produce
IL-10 (data not shown), and because we already excluded IL-10
as a mediator (Fig. 1F), we next wanted to investigate the influence of other soluble factors. We first analyzed the supernatant
via size exclusion chromatography, thus separating molecules in
molecular mass range of 1–200 kDa. Fractions were tested for
modulation of IL-12p40 secretion (and TNF-a, data not shown)
from LPS-stimulated BMDCs. We identified at least four different
fractions that contained inhibitory activity (Fig. 2A). Of note,
small molecules ,6.5 kDa showed considerable if not highest
inhibitory activity. In fact, targeted separation of molecules that
were .5000 Da showed reduced inhibition of DC reactivity as
compared with the nonseparated ECM (Fig. 2B). Proteinase K
digestion did not significantly reduce the modulation by ECM
(data not shown), which prompted us to conclude that nonproteinaceous small compounds are constitutively released by airway
epithelial cells and modulate DCs.
FIGURE 2. Epithelial cell supernatant contains different soluble factors with DC inhibitory capacity. A, FPLC analysis via Sephacryl S-100 column from
48 h supernatant derived from primary tracheal epithelia cells. BMDCs (2 3 105) were incubated with 50% (v/v) supernatant of the different fractions and
stimulated with LPS (100 ng/ml) overnight. IL-12p40 production was measured by ELISA. The gray shadowed region indicates the mean + SD of LPS
control stimulation without ECM. B, Forty-eight hour supernatant was concentrated with a 5000 Da cut-off column. Concentrated fractions were incubated
with BMDCs as indicated above and stimulated with LPS (100 ng/ml) overnight. Supernatants were analyzed by IL-12p40 ELISA. C and D, Forty-eight
hour supernatants were qualitatively analyzed via GC-MS/MS chromatography against an internal standard (d4-PGE2) on a Thermo TSQ 7000 triple-stage
quadrupole mass spectrometer interfaced with a ThermoQuest gas chromatograph model Trace 2000. Results of A show representative data for three
independent experiments, and results in B show representative data for two independent experiments.
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
type for 48 h after removal of the epithelial supernatant (still up to
40% inhibition of IL-12p40). Using DCs from C57BL/6 mice
instead of BALB/c mice we observed the same suppressive activity of the ECM on TNF and IL-12p40 secretion (Fig. 1B).
Moreover, we observed that ECM-conditioned cells showed enhanced IL-10 secretion in LPS-stimulated BALB/c and to a lesser
extent C57BL/6 DCs (Fig. 1A, 1B).
Furthermore, transcription of some genes, including arginase-1
and the lectin Ym1 (Fig. 1C, 1D), was specifically induced by the
treatment with ECM alone. The latter genes have been identified
as markers for alternative activation of macrophages (8). The
results of the phenotypic analysis of preconditioning of DCs with
epithelial cell-derived soluble mediators indicate that DCs are not
suppressed globally but instead adopt a phenotype that is characterized by increased anti-inflammatory properties. An important
function of DCs is activation of T cells. Whereas LPS-matured,
OVA-loaded DCs showed high induction of OT-II T cell proliferation (Fig. 1E), this capacity was reduced for ECM DCs.
Because ECM-treated DCs showed increased IL-10 secretion, we
excluded a role of DC-derived autocrine IL-10 by adding a
blocking IL-10 Ab in BALB/c DCs (Fig. 1F). Exogenous added
IL-10 was entirely blocked in its inhibitory actions on IL-12p40
secretion. However, ECM conditioning was still effective in suppressing IL-12p40, showing that IL-10 does not play a major role
INHIBITION OF DENDRITIC CELLS BY PGE2
The Journal of Immunology
PGs are bioactive lipid mediators capable of inhibiting the secretion of proinflammatory cytokines from BMDCs (11, 12). We
therefore decided to directly test ECM for the presence of PGs by
GC-MS/MS. After specific extraction by immunoaffinity chromatography, presence of PGE2 was identified in epithelial supernatant. Because of the keto group of PGE2, methoximation yields
two methoxyamine isomers that have distinct chromatographic
and MS/MS properties. The GC-MS/MS spectra of both peaks
(Fig. 2C, 2D) obtained from the PGE2 in the supernatant were
identical with those obtained from synthetic PGE2 (17). Thus,
PGE2 was identified in epithelial cell-conditioned medium from
nonstimulated ex vivo cultured tracheal epithelial cells.
Resting primary tracheal epithelial cells produce PGE2 in vitro
Previous publications showed that intestinal epithelial cells (20)
and human airway epithelial cells are capable of producing PGE2
2099
in vitro and in vivo (21, 22). Having observed that murine primary
epithelial cells cultured in air–liquid interface cultures are capable
of producing PGE2 in vitro (Fig. 2C, 2D), we quantified the
amounts of PGE2 that are released by those cells (Fig. 3A, 3B).
Using GC-MS/MS chromatography and analyzing 48 h epithelial
supernatants, airway epithelial cells secreted between 950 and
1600 pg/ml bioactive PGE2 in our experimental setting. In contrast, no significant PGE2 was detected in normal culture medium.
Further experiments revealed that there was a time-dependent
accumulation of PGE2 in the supernatant with higher production
after 48 h (1500 pg/ml) compared with 16 h supernatant (500 pg/
ml) (Fig. 3C). This parallels the observation that ECM collected
after 48 h has superior inhibitory capacity compared with earlier
or later time points (data not shown). Next, we compared dosedependency of ECM conditioning (Fig. 3D) to direct administration of PGE2 (Fig. 3E) on LPS activation of DCs. ECM con-
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
FIGURE 3. Tracheal epithelial cells produce PGE2 in vitro. A, Forty-eight hour supernatant from different epithelial cell preparations was tested for PGE2
production via GC-MS/MS after immunoaffinity chromatography using an internal standard (d4-PGE2). B, One representative result of A is shown as a chromatogram. C, PGE2 concentrations of supernatants from different ages were analyzed as in A. D, BMDCs (2 3 105) were incubated with different amounts of 48 h
supernatant of epithelial cells (70–12.5% [v/v]) or (E) PGE2 (0.05–500 ng/ml) and stimulated overnight with LPS (100 ng/ml). Supernatants were analyzed for
secretion of IL-12p40 ELISA. F, BMDCs (8 3 105) were seeded in a 24-well plate and incubated with ECM (50% [v/v]) or PGE2 (50 ng/ml) overnight and stimulated
with LPS (100 ng/ml) or left unstimulated for 4 h. Cells were lysed and analyzed for gene expression of arginase-1 by quantitative RT-PCR. rE, relative expression. G,
Four different preparations of ECM were measured by PGEM ELISA. BMDCs (2 3 105) were incubated with 50% (v/v) of these supernatants and stimulated with
LPS (100 ng/ml) overnight. IL-12p40 production was measured and normalized against the LPS control (100%). pEp, primary epithelial cells. Results in A, C, F, and
G show data representative for two independent experiments, and results in D and E show data representative for four independent experiments.
2100
ditioning with 12.5–70% (v/v) dose-dependently inhibited IL12p40 secretion (and TNF-a, data not shown) by DCs. Similarly,
PGE2 inhibited the activation of DCs as shown by diminished
TNF-a and IL-12p40 secretion in concentrations between 0.05
and 500 ng/ml. Thus, the inhibitory capacity of ECM when calculated in comparison with the administration of pure PGE2 was
slightly more effective, confirming our observation that additional
factors are operative. To further examine the role of PGE2 in the
interplay of epithelial cells with DCs, we analyzed genes that had
been shown to be specifically regulated in DCs when conditioned
with ECM. We observed that similar to the activity of ECM, PGE2
also upregulated expression of arginase-1 (Fig. 3F). Furthermore,
we examined different epithelial cell supernatants that varied in
content of PGE2 and analyzed their inhibitory capacity on DCmediated IL-12p40 secretion. We observed a strict correlation
between inhibition of IL-12p40 secretion and concentration of
stable PGEMs within ECM. These results confirm our notion that
PGE2 or its metabolites contribute to the modulation of DC activation by airway epithelial cells.
To further investigate and confirm the role of epithelial cell-derived
PGE2, we blocked the synthesis of PGs by inhibition of the two
isoforms of Cox, Cox-1 and Cox-2, with specific inhibitors. Pretreatment of epithelial cells with SC-560 or NS-398, specific
inhibitors for Cox-1 and Cox-2, respectively, resulted in a marked
reduction of PGE2 in the epithelial supernatant (Fig. 4A) as determined by GC-MS/MS (Fig. 4B). Applying both inhibitors resulted in a near complete inhibition of PGE2 secretion. We then
incubated DCs with supernatants of epithelial cells that had been
blocked by SC-560 or NS-398 and stimulated the DCs with LPS.
Whereas untreated ECM inhibited IL-12p40 secretion to ∼22% of
LPS control, ECM derived from NS-398–treated cells showed
a marked reduction in inhibitory capacity (90%) (Fig. 4C). In
contrast to Cox-2 inhibition, inhibition of Cox-1 by SC-560 was
less effective with respect to de-repression of DC activity (55%).
Applying higher concentrations of the Cox-1 inhibitor did not
result in increased relief of repression (data not shown). Treatment
of the epithelial cells with both inhibitors did not result in a more
pronounced de-repression of DC activation (data not shown).
Moreover, none of the inhibitors affected reactivity of DCs alone
(Fig. 4C and data not shown), allowing us to conclude that the
observed effects were due to the action of Cox inhibitors on the
epithelial cells. Again, we also examined markers that were specifically induced in DCs upon ECM conditioning. Confirming the
above findings, ECM treatment increased LPS-mediated IL-10
secretion in DCs, and this increase was inhibited or reduced
when using ECM derived from NS-398– or SC-560–pretreated
epithelial cells (Fig. 4D). The data allow for the conclusion that
functional Cox isoforms (resulting in production of PGE2) within
epithelial cells are necessary to induce the modulating effects of
ECM on LPS-activated DCs.
FIGURE 4. Blocking cyclooxygenases diminishes the inhibitory capacity of epithelial cell supernatant. A and B, Epithelial cells were treated with either
1 mM NS398 (Cox-2 inhibitor) or 1 mM SC560 (Cox-1 inhibitor) overnight. Supernatants were harvested and concentrations of PGE2 were measured via
GC-MS/MS chromatography using an internal standard (d4-PGE2). pEp, primary epithelial cells. C and D, BMDCs (2 3 105) were incubated with 50% (v/
v) supernatants after epithelial cell pretreatment as in A and stimulated overnight with LPS (100 ng/ml). BMDCs were analyzed for secretion of IL-12p40
(C) and IL-10 (D) by ELISA. Results in A and B show data representative for two independent experiments, and results in C and D show data representative
for three independent experiments.
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
Epithelial cell-mediated inhibition of DCs depends on Cox
activity
INHIBITION OF DENDRITIC CELLS BY PGE2
The Journal of Immunology
Cox isoforms are constitutively expressed in resting tracheal
epithelial cells
ferent Ab for Cox-2 and applying immunoenzyme staining we
detected constitutive Cox-2 protein expression in tracheal epithelial cells (Fig. 5E, control with IgG in Fig. 5F). Tissue of
a colon carcinoma stained with the Cox-2 Ab served as positive
control (data not shown). To further confirm our findings we decided to use an additional approach for detecting Cox-2 expression
in untreated epithelial cells. We prepared tracheal samples from
untreated mice and microdissected epithelial tissue using laser
capture microscopy. Frozen tracheae were cut into pieces and epithelial cells were stained, successfully cut out with the RoboLPC
method (Fig. 5G), and total RNA was isolated. RNA integrity was
sufficient (RNA integrity number 5.7) to analyze the samples by
quantitative RT-PCR. By measuring gene expression levels of the
different isoforms of Cox, we detected high expression levels of
Cox-1 in resting epithelial cells comparable to those from in vitrocultured epithelial cells (Fig. 5H). Purity of the samples was proven
by lack of CD11b expression and by detection of cystic fibrosis
transmembrane conductance regulator. Moreover, we also detected
high gene expression of the microsomal PGE synthase 1, which is
coexpressed with Cox-2 (27). With respect to expression of Cox-2,
we confirm that this isoform was expressed in microdissected, pure
epithelial samples from murine tracheae in similar amounts as
observed for in vitro-cultivated cells. Overall, the data strongly
indicate that murine tracheal epithelial cells constitutively express
Cox-1 and Cox-2 as well as the further enzyme machinery to
produce PGE2 without any specific stimulation.
Airway epithelial cells modulate DCs via the PGE2/EP4 axis
PGE2 signals via the E-PG receptors EP1–4. The receptors are
widely distributed in different tissues, including hematopoietic
FIGURE 5. Cox-2 is expressed in primary epithelial cells under unstimulated conditions. A, In vitro-cultured, unstimulated epithelial cells (2 3 105/
Transwell) of two different cell preparations were assessed for the expression of Cox-1and Cox-2 by quantitative RT-PCR. rE, relative expression. B and C,
Epithelial cells were stimulated with LPS (100 ng/ml) for 4 h or left unstimulated and analyzed by immunoprecipitation for expression of Cox-2 (B) and
Cox-1 (C). D, Murine tracheae of C57BL/6 mice were prepared, fixed in formalin, and Cox-1 was detected by immunofluorescence. Scale bar, 75 mm. E
and F, Cryosections of mouse trachea were fixed with 2% paraformaldehyde and stained with a Cox-2 Ab (E) as well as Ig control (F). Freshly isolated
frozen tracheae were cut into 18-mm pieces, mounted on membrane slides, and stained with crystal violet. Original magnification 3200. G, Epithelial cells
were cut out with the MicroBeam LMPC system using the RoboLPC method (original magnification 3200); 7-mm2 epithelium (∼70,000–200,000 cells)
were captured. H, Cells were lysed, transcribed into cDNA, and the levels of Cox-1, Cox-2, CD11b, and cystic fibrosis transmembrane conductance
regulator were measured (gray bars) in comparison with cultivated primary epithelial cells (black bars). pEp, primary epithelial cells; rE, relative expression. Results are representative of three (A–C) or two (H) independent experiments.
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
In most tissues Cox-1 is the constitutive form of the cyclooxygenases expressed under homeostatic conditions, and Cox-2 is
only induced during infections or after stimulation of the cells (23).
However, our data obtained so far, as well as single reports in the
literature (24, 25), suggested that Cox-2 might also be expressed
constitutively within airway epithelial cells. Therefore, we investigated constitutive expression of Cox-2 within the used primary tracheal epithelial cells in more detail. We first analyzed the
gene expression levels of Cox-1 and Cox-2 in primary cultures of
epithelial cells by quantitative RT-PCR. We observed expression
of Cox-1 as well as Cox-2 mRNA in different epithelial cell
preparations (Fig. 5A). Expression levels of Cox-2 were even
slightly higher as compared with Cox-1. None of these epithelial
cell cultures showed contamination with myeloid cells as determined by absence of CD11b expression (data not shown). Furthermore, we detected expression of Cox-1 and Cox-2 protein by
immunoprecipitation in untreated cultures of tracheal epithelial
cells (Fig. 5B, 5C). Stimulation of epithelial cells with LPS, although being effective with respect to increased IL-8 secretion
(data not shown), did not affect protein expression of Cox-2.
We next wanted to verify that constitutive in vitro expression
of Cox-2 is not a consequence of culture conditions but in fact
reflects the physiological in vivo situation. We first tried to confirm
Cox-2 expression by immunofluorescence in murine tracheae of
C57BL/6 mice. Although we observed strong expression of Cox-1
in epithelial cells (marked by K8 cytokeratin, Fig. 5D), as well as
in macrophages, we hardly detected any constitutive expression of
Cox-2 in epithelial cells (data not shown). However, using a dif-
2101
2102
To confirm these findings, we made use of DCs from knockout
mice for all four individual EP receptors (Fig. 6D). In comparison
with wild-type DCs, LPS-stimulated DCs from EP1 and EP4
knockout mice secreted slightly more TNF-a; however, those
differences were not significant. ECM conditioning resulted in
a strong reduction of TNF-a secretion in wild-type mice as observed before. In EP1, EP2, and EP3 knockout DCs, ECM similarly and significantly inhibited TNF-a secretion. This inhibition
was reduced only in EP4 knockout mice. However, even EP4
knockout mice still showed residual inhibition, again confirming
that additional inhibitory factors are operative. Interestingly, we
observed no significant increase of IL-12p40 in the EP4 receptor
knockout cells (data not shown). Again, we also analyzed expression of the ECM-regulated marker arginase-1 in the EP
knockout DCs. In DCs of wild-type mice, arginase-1 expression
was increased by ECM conditioning but not by LPS stimulation.
Differences were only observed in EP4 knockout mice: in those
DCs, ECM did not upregulate arginase-1 expression anymore
(Fig. 6E).
Taken together, these results confirm that tracheal epithelial
cell-derived PGE2 acts on DCs via the EP4 receptor and that
PGE2–EP4 signaling contributes to downregulation of proinflammatory cytokines and upregulation of arginase-1 in DCs.
Discussion
Over the years, the prevailing view of the immunological function
of airway epithelial cells had been that they build up a tight barrier and physically exclude inhaled microbes. However, it has now
been convincingly shown that epithelial cells express pattern rec-
FIGURE 6. Epithelial cell-derived PGE2 inhibits DCs via EP4 receptor signaling. A, BMDCs (8 3 105) were treated with 50% (v/v) ECM or PGE2 (50
ng/ml) overnight and stimulated with LPS (100 ng/ml) or left unstimulated for 4 h. Expression of the different EP receptors was detected by quantitative
RT-PCR. rE, relative expression. B and C, BMDCs (2 3 105) were cultured overnight with or without 50% (v/v) ECM and treated with 10 mM different EP
antagonists (EP2 or EP4) as indicated. Cells were stimulated afterwards with or without LPS (100 ng/ml) overnight. Supernatants were analyzed for
secretion of IL-12p40 or IL-10 by ELISA. D, BMDCs (2 3 105) from different EP knockout mice were cultured with 50% (v/v) ECM and stimulated with
LPS (100 ng/ml) overnight. Supernatants were measured by TNF-a ELISA. E, Cells as in D were taken and treated as indicated overnight. Afterwards cells
were stimulated with LPS (100 ng/ml) for 4 h, and gene expression levels of arginase-1 were measured by quantitative RT-PCR. Values represent the
relative induction of arginase-1 normalized to the housekeeping gene b-actin. rE, relative expression. Results from A and D are representative for two
independent experiments. Results for B are representative for three experiments, and results from C are representative for two experiments. Wt, wild-type.
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
cells (13). Assuming that epithelial cell-derived PGE2 drives DCs
to adopt a noninflammatory phenotype, we hypothesized that EP
receptors should be activated in the ECM DCs. We first verified
expression of EP1–4 within DCs by quantitative RT-PCR (Fig.
6A). We observed that all four receptors were expressed, with EP2
and EP4 showing the highest mRNA expression levels, followed
by EP3 and EP1. Expression did not change much upon stimulation of the cells with LPS (despite a small decrease for EP2
expression), preconditioning with ECM, or application of PGE2.
To block the signaling of epithelial cell-derived PGE2 in DCs,
we next used specific antagonists for EP2 and EP4 receptors,
which had been linked to PGE2-mediated modulation of DC
function previously (27). Preincubating DCs with the specific
antagonists GW 627368X (EP4) or AH 6809 (EP2) and subsequent conditioning with epithelial supernatant diminished the
inhibitory capacity of the supernatant on DCs when blocking EP4
but not EP2 (Fig. 6B). EP4 antagonist increased the production of
IL-12p40 from 13 ng/ml in ECM-blocked DCs to nearly 50 ng/ml.
However, release of inhibition did not result in complete derepression (LPS, 120 ng/ml IL-12p40). The little increase of IL12p40 secretion in DCs observed after the treatment with the EP2
antagonist together with ECM was not significant. Importantly,
treatment of nonconditioned DCs with either of the antagonists did
not affect IL-12p40 secretion by LPS, thus ruling out a role of DCderived PGs. Combinations of the antagonist were not more effective
(data not shown). When investigating the role of EP receptors for
IL-10 production, which was increased by ECM conditioning (Fig.
4D), we observed that the increasing effect of ECM was reduced
when blocking the EP4 receptor but not EP2 (Fig. 6C).
INHIBITION OF DENDRITIC CELLS BY PGE2
The Journal of Immunology
ments we showed that PGE2 is an important player in airway
epithelial cell-mediated modulation of DCs. PGE2 has an accepted
role in inflammation (37) but is also known to possess an inhibitory potential on proinflammatory cytokines. PGE2 diminishes
TNF-a and IL-12p40 secretion by DCs and favors upregulation of
IL-10 (11, 12), the same phenotype that we observed for ECM
DCs. PGE2 also influences Th17 differentiation (38), which we
also noted previously for epithelial cell-conditioned DCs (4).
It is known that airway epithelial cells are capable of producing
PGE2 (22, 23), and we found bioactive PGE2 in epithelial supernatant that mimicked the dose- and time-dependent effects of epithelial cell-conditioned medium on DCs. Moreover, PGE2 (metabolites) content in ECM correlated with suppression of IL-12p40
secretion. Whereas proinflammatory functions of PGs are well
known, it had been suggested that PGE2 contributes to the generation of anti-inflammatory microenvironments (15) in a manner dependent on activation of the different EP receptors (38). In other
mucosal microenvironments (e.g., the intestine) it is well documented that epithelial cells secrete PGE2 under injured conditions,
thereby regulating intestinal homeostasis (20, 39). Furthermore,
a role of PGE2 in resolution of autoimmune arthritis has been noted (40). PGE2 has also been linked to suppression of LPS-induced
IDO and maturation of DCs as analyzed by inhibition of CD80
and CD86 expression (also observed in ECM DCs) (4), which
supports a suppressive immunomodulatory action for PGE2 (41).
The isoenzyme Cox-1 is constitutively expressed within airway
epithelial cells, but Cox-2 is considered to be inducibly expressed
during inflammation and infection (23). By blocking the activity of
Cox enzymes in epithelial cells, we observed that mainly the Cox2 inhibitor affected epithelial modulation of DCs. In line with this,
some reports claimed a “quasiconstitutive” expression of Cox-2 in
resting lung epithelial cells (42). Our own results support the notion that there is basal expression of Cox-2 under resting conditions
in airway epithelial cells. Although inducible Cox-2 is implicated
to play a role in cytokine-induced inflammation, other reports indicate that Cox-2 also affects conditions beyond inflammation.
Thus, specific inhibition of Cox-2 resulted in increased airway
inflammation and airway hyperresponsiveness (43). In the gastrointestinal tract Cox-2 is suggested to play a homeostatic role because Cox-2 inhibition resulted in enhanced colonic mucosal injury
(15, 44). Cox-2 has been suggested to be preferred over Cox-1 in
PGE2 synthesis under conditions of low arachidonic acid concentrations (45).
We show that Cox-2–mediated PGE2 derived from airway epithelial cells affects DC reactivity. PGE2 signals mainly via four
different G protein-coupled EP receptors, that is, EP1–4 (10, 13).
BMDCs express all four different receptors, and EP2 and EP4 are
important receptors in PGE2-mediated DC signaling (27). Inhibiting EP2 and EP4 receptors on DCs and using EP receptor
knockout mice, we identify EP4 to play an exclusive role in epithelial cell-mediated inhibition of DCs. In line with our findings,
the PGE2/EP4 axis is important for gut mucosal integrity (15)
and homeostatic regulation of vascular remodeling in airway inflammation (46). It was also shown recently that PGE2 inhibits
IFN-a secretion and Th1 costimulation in plasmacytoid DCs via
EP2 and EP4 engagement (47). However, in a model of murine
contact hypersensitivity, PGE2–EP3 signaling exerted an antiinflammatory signal on keratinocytes, which argues for organspecific adaptions (48). In T cells PGE2–EP2/EP4 signaling was
shown to promote Th1 and TH17 differentiation, and therefore
cell-specific adaptations might add further complexity (49). Indeed, previous work of our own showed increased production of
IL-17 by T cells directly treated with ECM (4). EP4 signaling is
coupled to cAMP production, which has inhibitory effects in DCs.
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
ognition receptors found otherwise in professional innate immune
cells. Thus, airway epithelial cells are able to sense infectious danger and actively contribute to the induction of immune responses
(1, 2). As the airways encounter dangerous as well as harmless
microbes, mucosal immune recognition has to be tightly regulated.
Indeed, a number of reports show that intestinal epithelial cells
(28, 29) but also bronchial epithelial cells (4, 30) modulate the
activation of local professional immune cells. However, the molecular mechanisms that mediate the interplay between epithelial
and hematopoietic cells are only beginning to be understood.
Analyzing epithelial cell-derived factors that contribute to the
generation of a specific immunological microenvironment in the
airways was aim of this study. We used an established experimental
approach (30) that examined the effects of soluble epithelial cellderived factors on BMDCs. We can definitively show that epithelial cells, through the constitutive secretion of PGE2, drive
DCs to adopt an anti-inflammatory phenotype. DC populations in
general show tremendous heterogeneity and exist in many phenotypes that adapt to the local microenvironment (5, 30). Intestinal
epithelial cells are known to inhibit DC activation, resulting in low
MHC class II and CD86 expression as well as reduced secretion
of proinflammatory cytokines (28, 29). In a similar manner, we
previously observed that airway epithelial cells regulate DCs to
adopt an inhibitory phenotype indicated by diminished IL-12p40
and TNF-a production and downregulation of costimulatory molecules (CD40, CD86) (4). In this study, we show that soluble
factors that are secreted constitutively by epithelial cells are sufficient to achieve this specific DC phenotype. Moreover, we report
that epithelial cell-conditioned DCs are not globally inhibited but
instead modulate their gene expression profile with upregulation
of certain genes (arginase-1, IL-10, Ym1). ECM DCs therefore
resemble alternatively activated macrophages (8). Alternatively
activated DCs produce high amounts of IL-10 and expand regulatory T cells (31, 32).
Our phenotypic characterization of ECM DCs allows for the
speculation that airway epithelial cells modulate local DCs toward
an alternatively activated, tolerogenic state similar to a report on
intestinal cells (9). In support of such a model, preliminary results
show that purified primary airway DCs have high expression of
Ym1 and arginase-1 and secrete less IL-12 (data not shown).
Moreover, ECM DCs were less effective in inducing T cell proliferation. Previously it was shown that ECM drives induction of
regulatory T cells (4). However, alternative activation has also
been coupled to Th2 differentiation, and thus a more in-depth
analysis will be necessary.
Our results indicate that this modulation is induced by soluble
factors that modify DC biology instead of being primarily ascribed
to different DC populations. Whether the presence of local subpopulations of DCs, for example, airway CD103+ DCs that have
regulatory function and inhibit permanent leukocyte recruitment
(33), is influenced by epithelial cells is currently unknown.
Soluble factors secreted by intestinal epithelial cells that inhibit DCs have been identified before. TGF-b or retinoic acid was
shown to suppress the activation of APCs in a cell contactindependent manner (28, 34). Additionally, thymic stromal lymphopoietin secreted by intestinal epithelial cells promotes tolerogenic DCs (29), but in the airways thymic stromal lymphopoietin
is associated with an inflammatory DC phenotype (35). Negative
regulatory cell surface molecules that could interact with DCs
have also been identified on airway epithelial cells (36). Our
experiments show that epithelial cells constitutively release soluble factors, among which small molecules (,6.5 kDa) showed
strong modulatory impact. We identified the small lipid mediator
PGE2 being secreted by epithelial cells, and in the set of experi-
2103
2104
Also, cAMP is currently seen as a universal regulator for innate
immune functions by enhancing the production of IL-10 or the
negative regulator suppressor of cytokine signaling 3 (50, 51). For
the factors that were shown to be inhibited in this study, it has
been reported that IL-12p40 contains a PGE2-responsive repressor
element (52).
PGE2 is not the only prostanoid found in the microenvironment
of the airways. Under inflammatory conditions, leukotriene B4 and
under asthmatic conditions PGD2 have been reported (53, 54). It is
also known that the crosstalk between PGE2 and leukotriene B4
regulates the phagocytotic activity of alveolar macrophages, and
leukotriene B4 opposes the immunosuppressive effect of PGE2
(55). Thus, it might well be that during infectious conditions the
functions of PGE2 within the airway microenvironment change.
Taken together, we show that airway epithelial cell-derived
PGE2 modulates DCs. As a result, PGE2–EP4 receptor signaling
generates DCs with reduced proinflammatory properties, thereby
limiting DC activation. Local immunity might therefore be shaped
and fine-tuned by epithelial cells to serve the organ-specific needs.
We thank Prof. S. Narumiya (Kyoto, Japan) and ONO Pharmaceuticals for
help with EP knockout cells. We thank F.M. Gutzki (Institute of Clinical
Pharmacology, Hannover, Germany) for performing GC-MS/MS analyses
of PGE2. We also appreciate the help of B. Lambrecht (Ghent, Belgium)
and especially of M. Willart.
Disclosures
The authors have no financial conflicts of interest.
References
1. Kato, A., and R. P. Schleimer. 2007. Beyond inflammation: airway epithelial
cells are at the interface of innate and adaptive immunity. Curr. Opin. Immunol.
19: 711–720.
2. Fritz, J. H., L. Le Bourhis, J. G. Magalhaes, and D. J. Philpott. 2008. Innate
immune recognition at the epithelial barrier drives adaptive immunity: APCs
take the back seat. Trends Immunol. 29: 41–49.
3. Greene, C. M., and N. G. McElvaney. 2005. Toll-like receptor expression and
function in airway epithelial cells. Arch. Immunol. Ther. Exp. (Warsz.) 53: 418–
427.
4. Mayer, A. K., H. Bartz, F. Fey, L. M. Schmidt, and A. H. Dalpke. 2008. Airway
epithelial cell modify immune responses by inducing an anti-inflammatory microenvironment. Eur. J. Immunol. 38: 1689–1699.
5. Raz, E. Organ-specific regulation of innate immunity. 2007. Nat. Immunol. 8: 3-4
6. Stumbles, P. A., J. W. Upham, and P. G. Holt. 2003. Airway dendritic cells: coordinators of immunological homeostasis and immunity in the respiratory tract.
APMIS 111: 741–755.
7. Gratchev, A., K. Schledzewski, P. Guillot, and S. Goerdt. 2001. Alternatively
activated antigen-presenting cells: molecular repertoire, immune regulation, and
healing. Skin Pharmacol. Appl. Skin Physiol. 14: 272–279.
8. Raes, G., W. Noël, A. Beschin, L. Brys, P. de Baetselier, and G. H. Hassanzadeh.
2002. FIZZ1 and Ym as tools to discriminate between differentially activated
macrophages. Dev. Immunol. 9: 151–159.
9. Thomson, A. W. 2010. Tolerogenic dendritic cells: all present and correct? Am.
J. Transplant. 10: 214–219.
10. Harris, S. G., J. Padilla, L. Koumas, D. Ray, and R. P. Phipps. 2002. Prostaglandins as modulators of immunity. Trends Immunol. 23: 144–150.
11. Vassiliou, E., H. Jing, and D. Ganea. 2003. Prostaglandin E2 inhibits TNF
production in murine bone marrow-derived dendritic cells. Cell. Immunol. 223:
120–132.
12. Harizi, H., and N. Gualde. 2006. Pivotal role of PGE2 and IL-10 in the crossregulation of dendritic cell-derived inflammatory mediators. Cell. Mol. Immunol.
3: 271–277.
13. Sugimoto, Y., and S. Narumiya. 2007. Prostaglandin E receptors. J. Biol. Chem.
282: 11613–11617.
14. Honda, T., E. Segi-Nishida, Y. Miyachi, and S. Narumiya. 2006. Prostacyclin-IP
signaling and prostaglandin E2-EP2/EP4 signaling both mediate joint inflammation in mouse collagen-induced arthritis. J. Exp. Med. 203: 325–335.
15. Kabashima, K., T. Saji, T. Murata, M. Nagamachi, T. Matsuoka, E. Segi,
K. Tsuboi, Y. Sugimoto, T. Kobayashi, Y. Miyachi, et al. 2002. The prostaglandin receptor EP4 suppresses colitis, mucosal damage and CD4 cell activation in the gut. J. Clin. Invest. 109: 883–893.
16. Bode, K. A., F. Schmitz, L. Vargas, K. Heeg, and A. H. Dalpke. 2009. Kinetic of
RelA activation controls magnitude of TLR-mediated IL-12p40 induction. J.
Immunol. 182: 2176–2184.
17. Tsikas, D., E. Schwedhelm, M. T. Suchy, J. Niemann, F. M. Gutzki,
V. J. Erpenbeck, J. M. Hohlfeld, A. Surdacki, and J. C. Frölich. 2003. Divergence
in urinary 8-iso-PGF2a (iPF2a-III, 15-F2t-IsoP) levels from gas chromatographytandem mass spectrometry quantification after thin-layer chromatography and
immunoaffinity column chromatography reveals heterogeneity of 8-iso-PGF
(2a): possible methodological, mechanistic and clinical implications. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 794: 237–255.
18. Tsikas, D. 1998. Application of gas chromatography-mass spectrometry and gas
chromatography-tandem mass spectrometry to assess in vivo synthesis of prostaglandins, thromboxane, leukotrienes, isoprostanes and related compounds in
humans. J. Chromatogr. B Biomed. Sci. Appl. 717: 201–245.
19. Autschbach, F., T. Giese, N. Gassler, B. Sido, G. Heuschen, U. Heuschen,
I. Zuna, P. Schulz, H. Weckauf, I. Berger, et al. 2002. Cytokine/chemokine
messenger-RNA expression profiles in ulcerative colitis and Crohn’s disease.
Virchows Arch. 441: 500–513.
20. Stenson, W. F. 2007. Prostaglandins and epithelial response to injury. Curr. Opin.
Gastroenterol. 23: 107–110.
21. Holtzman, M. J. 1992. Arachidonic acid metabolism in airway epithelial cells.
Annu. Rev. Physiol. 54: 303–329.
22. Goldie, R. G., L. B. Fernandes, S. G. Farmer, and D. W. Hay. 1990. Airway
epithelium-derived inhibitory factor. Trends Pharmacol. Sci. 11: 67–70.
23. Radi, Z. A., D. K. Meyerholz, and M. R. Ackermann. 2010. Pulmonary
cyclooxygenase-1 (COX-1) and COX-2 cellular expression and distribution after
respiratory syncytial virus and parainfluenza virus infection. Viral Immunol. 23:
43–48.
24. Mitchell, J. A., M. G. Belvisi, P. Akarasereenont, R. A. Robbins, O. J. Kwon,
J. Croxtall, P. J. Barnes, and J. R. Vane. 1994. Induction of cyclo-oxygenase-2 by
cytokines in human pulmonary epithelial cells: regulation by dexamethasone. Br.
J. Pharmacol. 113: 1008–1014.
25. Ermert, L., M. Ermert, M. Goppelt-Struebe, D. Walmrath, F. Grimminger,
W. Steudel, H. A. Ghofrani, C. Homberger, H. R. Duncker, and W. Seeger. 1998.
Cyclooxygenase isoenzyme localization and mRNA expression in rat lungs. Am.
J. Respir. Cell Mol. Biol. 18: 479–488.
26. Kudo, I., and M. Murakami. 2005. Prostaglandin E synthase, a terminal enzyme
for prostaglandin E2 biosynthesis. J. Biochem. Mol. Biol. 38: 633–638.
27. Harizi, H., C. Grosset, and N. Gualde. 2003. Prostaglandin E2 modulates dendritic cell function via EP2 and EP4 receptor subtypes. J. Leukoc. Biol. 73: 756–
763.
28. Butler, M., C. Y. Ng, D. A. van Heel, G. Lombardi, R. Lechler, R. J. Playford,
and S. Ghosh. 2006. Modulation of dendritic cell phenotype and function in an
in vitro model of the intestinal epithelium. Eur. J. Immunol. 36: 864–874.
29. Rimoldi, M., M. Chieppa, V. Salucci, F. Avogadri, A. Sonzogni,
G. M. Sampietro, A. Nespoli, G. Viale, P. Allavena, and M. Rescigno. 2005.
Intestinal immune homeostasis is regulated by the crosstalk between epithelial
cells and dendritic cells. Nat. Immunol. 6: 507–514.
30. Rate, A., J. W. Upham, A. Bosco, K. L. McKenna, and P. G. Holt. 2009. Airway
epithelial cells regulate the functional phenotype of locally differentiating dendritic cells: implications for the pathogenesis of infectious and allergic airway
disease. J. Immunol. 182: 72–83.
31. Lan, Y. Y., Z. Wang, G. Raimondi, W. Wu, B. L. Colvin, A. de Creus, and
A. W. Thomson. 2006. “Alternatively activated” dendritic cells preferentially
secrete IL-10, expand Foxp3+CD4+ T cells, and induce long-term organ allograft
survival in combination with CTLA4-Ig. J. Immunol. 177: 5868–5877.
32. Schreiber, T., S. Ehlers, L. Heitmann, A. Rausch, J. Mages, P. J. Murray,
R. Lang, and C. Hölscher. 2009. Autocrine IL-10 induces hallmarks of alternative activation in macrophages and suppresses antituberculosis effector
mechanisms without compromising T cell immunity. J. Immunol. 183: 1301–
1312.
33. del Rio, M. L., G. Bernhardt, J. I. Rodriguez-Barbosa, and R. Förster. 2010.
Development and functional specialization of CD103+ dendritic cells. Immunol.
Rev. 234: 268–281.
34. Wada, Y., T. Hisamatsu, N. Kamada, S. Okamoto, and T. Hibi. 2009. Retinoic
acid contributes to the induction of IL-12-hypoproducing dendritic cells.
Inflamm. Bowel Dis. 15: 1548–1556.
35. Ziegler, S. F., and D. Artis. 2010. Sensing the outside world: TSLP regulates
barrier immunity. Nat. Immunol. 11: 289–293.
36. Stanciu, L. A., C. M. Bellettato, V. Laza-Stanca, A. J. Coyle, A. Papi, and
S. L. Johnston. 2006. Expression of programmed death-1 ligand (PD-L) 1, PDL2, B7-H3, and inducible costimulator ligand on human respiratory tract epithelial cells and regulation by respiratory syncytial virus and type 1 and 2
cytokines. J. Infect. Dis. 193: 404–412.
37. Oka, T. 2004. Prostaglandin E2 as a mediator of fever: the role of prostaglandin E
(EP) receptors. Front. Biosci. 9: 3046–3057.
38. Boniface, K., K. S. Bak-Jensen, Y. Li, W. M. Blumenschein, M. J. McGeachy,
T. K. McClanahan, B. S. McKenzie, R. A. Kastelein, D. J. Cua, and R. de Waal
Malefyt. 2009. Prostaglandin E2 regulates Th17 cell differentiation and function
through cyclic AMP and EP2/EP4 receptor signaling. J. Exp. Med. 206: 535–
548.
39. Dey, I., M. Lejeune, and K. Chadee. 2006. Prostaglandin E2 receptor distribution
and function in the gastrointestinal tract. Br. J. Pharmacol. 149: 611–623.
40. Chan, M. M., and A. R. Moore. 2010. Resolution of inflammation in murine
autoimmune arthritis is disrupted by cyclooxygenase-2 inhibition and restored
by prostaglandin E2-mediated lipoxin A4 production. J. Immunol. 184: 6418–
6426.
41. Jung, I. D., Y. I. Jeong, C. M. Lee, K. T. Noh, S. K. Jeong, S. H. Chun,
O. H. Choi, W. S. Park, J. Han, Y. K. Shin, et al. 2010. COX-2 and PGE2
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
Acknowledgments
INHIBITION OF DENDRITIC CELLS BY PGE2
The Journal of Immunology
42.
43.
44.
45.
46.
47.
48.
49. Yao, C., D. Sakata, Y. Esaki, Y. Li, T. Matsuoka, K. Kuroiwa, Y. Sugimoto, and
S. Narumiya. 2009. Prostaglandin E2-EP4 signaling promotes immune inflammation through TH1 cell differentiation and TH17 cell expansion. Nat. Med.
15: 633–640.
50. Koga, K., G. Takaesu, R. Yoshida, M. Nakaya, T. Kobayashi, I. Kinjyo, and
A. Yoshimura. 2009. Cyclic adenosine monophosphate suppresses the transcription of proinflammatory cytokines via the phosphorylated c-Fos protein.
Immunity 30: 372–383.
51. Serezani, C. H., M. N. Ballinger, D. M. Aronoff, and M. Peters-Golden. 2008.
Cyclic AMP: master regulator of innate immune cell function. Am. J. Respir.
Cell Mol. Biol. 39: 127–132.
52. Becker, C., S. Wirtz, X. Ma, M. Blessing, P. R. Galle, and M. F. Neurath. 2001.
Regulation of IL-12 p40 promoter activity in primary human monocytes: roles of
NF-kB, CCAAT/enhancer-binding protein beta, and PU.1 and identification of
a novel repressor element (GA-12) that responds to IL-4 and prostaglandin E2. J.
Immunol. 167: 2608–2618.
53. Rao, N. L., J. P. Riley, H. Banie, X. Xue, B. Sun, S. Crawford, K. A. Lundeen,
F. Yu, L. Karlsson, A. M. Fourie, and P. J. Dunford. 2010. Leukotriene A4 hydrolase inhibition attenuates allergic airway inflammation and hyperresponsiveness. Am. J. Respir. Crit. Care Med. 181: 899–907.
54. Matsuoka, T., M. Hirata, H. Tanaka, Y. Takahashi, T. Murata, K. Kabashima,
Y. Sugimoto, T. Kobayashi, F. Ushikubi, Y. Aze, et al. 2000. Prostaglandin D2 as
a mediator of allergic asthma. Science 287: 2013–2017.
55. Lee, S. P., C. H. Serezani, A. I. Medeiros, M. N. Ballinger, and M. PetersGolden. 2009. Crosstalk between PGE2 and leukotriene B4 regulates phacocytosis in alveolar macrophages via combinatorial effects on cAMP. J. Immunol.
182: 530–537.
Downloaded from http://www.jimmunol.org/ by guest on August 9, 2017
signaling is essential for the regulation of IDO expression by curcumin in murine
bone marrow-derived dendritic cells. Int. Immunopharmacol. 10: 760–768.
Park, G. Y., N. Hu, X. Wang, R. T. Sadikot, F. E. Yull, M. Joo, R. S. Peebles, Jr.,
T. S. Blackwell, and J. W. Christman. 2007. Conditional regulation of
cyclooxygenase-2 in tracheobronchial epithelial cells modulates pulmonary
immunity. Clin. Exp. Immunol. 150: 245–254.
Peebles, R. S., Jr., K. Hashimoto, J. D. Morrow, R. Dworski, R. D. Collins,
Y. Hashimoto, J. W. Christman, K. H. Kang, K. Jarzecka, J. Furlong, et al. 2002.
Selective cyclooxygenase-1 and -2 inhibitors each increase allergic inflammation
and airway hyperresponsiveness in mice. Am. J. Respir. Crit. Care Med. 165:
1154–1160.
Gretzer, B., K. Ehrlich, N. Maricic, N. Lambrecht, M. Respondek, and
B. M. Peskar. 1998. Selective cyclo-oxygenase-2 inhibitors and their influence
on the protective effect of a mild irritant in the rat stomach. Br. J. Pharmacol.
123: 927–935.
Smith, W. L., and R. Langenbach. 2001. Why there are two cyclooxygenase
isozymes. J. Clin. Invest. 107: 1491–1495.
Lundequist, A., S. N. Nallamshetty, W. Xing, C. Feng, T. M. Laidlaw,
S. Uematsu, S. Akira, and J. A. Boyce. 2010. Prostaglandin E2 exerts homeostatic regulation of pulmonary vascular remodeling in allergic airway inflammation. J. Immunol. 184: 433–441.
Fabricius, D., M. Neubauer, B. Mandel, C. Schütz, A. Viardot, A. Vollmer,
B. Jahrsdörfer, and K. M. Debatin. 2010. Prostaglandin E2 inhibits IFN-a secretion
and Th1 costimulation by human plasmacytoid dendritic cells via E-prostanoid 2
and E-prostanoid 4 receptor engagement. J. Immunol. 184: 677–684.
Honda, T., T. Matsuoka, M. Ueta, K. Kabashima, Y. Myachi, and S. Narumiya.
2009. Prostaglandin E2-EP3 signalling suppresses skin inflammation in murine
contact hypersensitivity. J. Allergy Clin. Immunol. 124: 809–818.e2.
2105