| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of
* Pathology and
Medical Oncology, Vrije Universiteit Medical Center, Amsterdam, The Netherlands
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Detrimental effects of several soluble immunosuppressive factors on the
differentiation and maturation of DC from monocytes or
CD34+ myeloid stem cell progenitors have been
demonstrated (15). Among these factors are a number of
cytokines that are often produced by malignant cells, i.e., IL-10
(16), TGF-
1 (15),
IL-6 (17), and vascular endothelial growth factor (VEGF)
(18). Recent studies have revealed the involvement of
tumor-derived IL-6 and VEGF in the inhibition of DC differentiation
from both CD34+ and monocytic precursors.
However, these studies used supernatants from tumor-derived cell lines
rather than supernatants from primary tumor cell cultures (11, 17, 18, 19). To more accurately assess the putative involvement of
tumor-derived immunosuppressive factors in the hampered DC
differentiation in vivo we have generated supernatants from primary
tumor cell cultures. To this end, single-cell suspensions were prepared
from colon carcinoma samples that were collected for tumor cell-based
vaccination purposes (20). Supernatants were collected
after 24 h of culture. The effects of these tumor-derived
supernatants (TDSN) were tested on the in vitro differentiation of DC
from monocytes and from CD34+ stem cells.
Although our studies mostly concentrated on colon tumors, their results
were also extended to other solid tumors, i.e., breast and renal cell
cancer and melanoma. In contrast to previously published data on
tumor-derived cell lines (11, 17), our data on primary
tumors reveal a predominant role of cyclooxygenase (COX)-1- and
-2-regulated prostanoids (i.e., arachidonic acid metabolites) in the
tumor-associated inhibition of DC differentiation. As COX-2
dysregulation, resulting in, notably, an increased production of
PGE2, is a common feature of many tumors
(21, 22), it may well be the primary cause of the
widespread down-modulation of DC development and functions observed
among cancer patients.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following mAbs and recombinant cytokines were purchased from
R&D Systems (Minneapolis, MN): anti-human IL-6 (MAB206),
anti-human IL-10 Ab (MAB217), biotinylated anti-human
TGF-1 Ab (BAF240), anti-human
TGF-1 capture Ab (MAB240), anti-human
M-CSF Ab (MAB216), recombinant human (rh)VEGF165, and rhTGF-1
(240-B). rhIL-6 was purchased from ITK Diagnostics (Uithoorn,
The Netherlands). Mouse mAbs with the following (anti-human)
specificities were purchased from BD Biosciences (Mountain View, CA):
CD14, CD80 (B7.1), CD86 (B7.2), CD54 (ICAM), and anti-human HLA-DR.
Anti-human CD1a mAbs were obtained from Monosan (Uden, The
Netherlands). Directly PE-conjugated mouse mAbs against human CD1a,
CD80, CD86, CD40, and CD54 were obtained from BD PharMingen (San Diego,
CA). PGE2 and recombinant human stem cell factor
were purchased from Sigma-Aldrich (St. Louis, MO). rhM-CSF and rhIL-10
were purchased from Strathmann Biotec (Hanover, Germany).
Cell lines
Colon carcinoma cell lines WiDr, A2233, HT29, Colo320, and SW620 were obtained from the American Type Culture Collection (Manassas, VA). The CD40 ligand (CD40L)-transfected J558 plasmacytoma cell line (J558-CD40L) was a kind gift of Dr. M. Kapsenberg (AMC, Amsterdam, The Netherlands).
Tumor dissociation
Cell suspensions prepared from human primary tumors (colon carcinoma Duke’s B or Duke’s C (n = 30); mamma (n = 2); renal cell carcinoma (n = 2), or melanoma (n = 2)) or from nontumor tissue, taken from colon cancer resection material, were handled within 24 h of surgical removal as described by Vermorken et al. (20). In short, viable tumor was minced with a scalpel and dissociated for 1–3 h with 0.02% DNase (Boehringer Mannheim, Mannheim, Germany) and 0.14% collagenase type IV (Boehringer Mannheim) in HBSS (BioWhittaker, Walkersville, MD). The cell suspension was filtered and washed to remove tissue debris and immediately used for preparation of TDSN or cryopreserved using a control rate freezing system and stored in liquid nitrogen. Before use, tumor cell suspensions were rapidly thawed and diluted in HBSS containing 0.02% DNase (Boehringer Mannheim). After three washes cell numbers and cell viability were determined by trypan blue exclusion and used to prepare TDSN.
Generation of TDSN
TDSN were prepared from colon carcinoma cell lines and freshly dissociated primary tumors. Cells were seeded at a density of 1 x 106 cells/ml in complete medium (IMDM containing 10% FCS, 50 U/ml penicillin-streptomycin, 1.6 mM L-glutamine, and 0.01 mM 2-ME). After a 24-h culture, tumor cells were centrifuged (at 1500 rpm (530 x g) for 5 min), and supernatants were run over a 0.22-µm Millex-GP filter (Millipore, Bedford, MA) and stored at -80°C.
Neutralization of cytokines in the TDSN
Supernatants were preincubated with mAbs for 30 min at room temperature. Abs against IL-10 (clone 23738.11), IL-6 (clone 6708.111), and M-CSF (clone 26730.11) were used at optimal neutralizing concentrations. A mouse IgG1 mAb (ICN Biomedicals, Eschwege, Germany) was used as isotype control. mAbs were added at the start of DC culture and were present during the 7- or 14-day differentiation period.
Separation of CD45+ tumor infiltrate
Indirect panning to separate infiltrate and colon cells was performed as described by Wysocki and Sato (23).
Briefly, cells were incubated at a concentration of 5 x 107 cells/ml with mAbs against CD45 (15D9; CLB, Amsterdam, The Netherlands), 1/1000 diluted in PBS containing 5% FCS (Life Technologies, Paisley, U.K.). To remove excess Ab, cells were washed twice with PBS/5% FCS. A total of 2 x 106 labeled cells in 1 ml PBS/5% FCS were layered in six-well plates precoated with a 1/100 dilution of affinity isolated goat anti-mouse Ig Abs (DAKO, Glostrup, Denmark). Cells were allowed to bind the dishes for 70 min at 4°C. Unattached cells were removed by gentle rinsing and used to prepare supernatant at a concentration of 1 x 106 cells/ml. CD45+ leukocytes were cultured at a concentration corresponding with 1 x 106 colon cells/ml.
Inhibition of prostanoid production
Production of prostanoids by tumor or nontumor cells was blocked by the use of COX inhibitors. TDSN were prepared, as described above, in the presence or absence of 10 µM indomethacin (Merck, Darmstadt, Germany), a nonselective COX-1/COX-2 inhibitor. The selective COX-1 inhibitor SC560 (10 µM; Cayman Chemicals, Ann Arbor, MI) was similarly used.
PBMC and CD34+ stem cell isolation
PBMC were prepared from buffy coats obtained from healthy donors by density gradient centrifugation over Hypaque Lymphoprep (Nycomed, Oslo, Norway) and cryopreserved as previously described (24). Cells bearing CD34 were isolated from peripheral blood mononuclear fractions of G-CSF-mobilized patients through positive selection by MACS (Miltenyi Biotec, Bergisch Gladbach, Germany), using anti-CD34+ mAbs and goat anti-mouse IgG-coated microbeads (88–95% purity).
Generation of DC
DC were prepared from PBMC or CD34+ progenitor cells as described (17, 25).
MoDC. PBMC were thawed and resuspended in IMDM containing 10% FCS, 50 U/ml penicillin-streptomycin, 1.6 mM L-glutamine, and 0.01 mM 2-ME, and allowed to adhere to six-well tissue culture plates (3 x 106 cells/ml) for 2 h at 37°C. Nonadherent cells were removed and adherent cells were cultured in complete medium supplemented with 100 ng/ml GM-CSF (sp. act. 1.11 x 106 IU/mg; Schering-Plough, Madison, NJ) and 1000 U/ml IL-4 (sp. act. 108 U/mg; CLB). Monocyte-derived DC (MoDC) were collected at day 7.
CD34+ progenitor-derived DC.
CD34+ cells were seeded at 5 x
104 cells/ml in 24-well plates and cultured in
complete medium supplemented with 100 ng/ml GM-CSF, 2.5 ng/ml rhTNF-
(sp. act. 5 x 106 U/mg; CLB), and 20 ng/ml
recombinant human stem cell factor (PeproTech, Rocky Hill, NJ). At day
9 of culture 1000 U/ml IL-4 was added. Cells were collected at days
12–14. Semiadherent DC were harvested with 0.05 mM EDTA. TDSN were
added at the onset (day 0) of DC culture and were present during the
entire differentiation period (7 days for MoDC and 11–14 days for
CD34-derived DC).
Immunocytochemistry
Cytospins were prepared in a cytocentrifuge, loading 1 x 104 cells per cytospot. Cells were centrifuged for 5 min at 500 rpm. Cytospins were dried overnight at room temperature and fixed with acetone for 10 min. Cells were blocked with normal rabbit serum (1:50; DAKO) for 10 min. Subsequently, slides were incubated with mAbs to CD1a (1:2000) or CD14 (1:100) for 1 h. Slides were washed with PBS and incubated with biotin-labeled rabbit anti-mouse (1:150; Zymed Laboratories, San Francisco, CA) for 1 h, washed again, and stained with streptavidin-HRP (1:500; Zymed Laboratories). The peroxidase activity was visualized by aminoethyl carbazol (0.4 g/L in sodium acetate buffer (0.1 M, pH 5) containing 0.05% H2O2). Staining was stopped with tap water and cells were counterstained with hematoxylin.
Cytokine and PGE2 detection
Commercial kits were used to measure human IL-10, IL-6
(detection limits of 10 and 12 pg/ml, respectively; CLB) VEGF, M-CSF
(detection limits of 16 and 31 pg/ml, respectively; Quantikine; R&D
Systems), and PGE2 (detection limit 1 pg/ml;
Biotrak; Amersham Pharmacia Biotech, Piscataway, NJ).
TGF-1 was quantified using a combination of
capture and secondary Abs that were paired for this application (clone
9016.2 (MAB240) as capture Ab and
anti-TGF-1 chicken IgY (BAF240) as
detection Ab; detection limit of 16 pg/ml; R&D Systems). TGF- was
activated before ELISA by acidification as previously described
(26). Human IL-12 was detected in a capture ELISA as
previously described (27).
FACS analysis
Immunophenotypic analysis was performed using FACS. In short, cells (0.25 x 106) were washed in PBS supplemented with 1% BSA and 0.02% NaN3 (PBA) and incubated for 30 min at room temperature with PE- or FITC-conjugated specific mAbs or with the corresponding isotype-matched mAb. Excess mAb was removed by washing in PBA. Results were expressed as either mean fluorescence or the percentage of positive cells. Fluorescence analysis was performed on a FACStar flow cytometer (BD Biosciences) after acquisition of 5000 events and analyzed with BD Biosciences CellQuest software.
Mixed lymphocyte reaction
A total of 5 x 104 responder PBMC per well, isolated from buffy coats as described above, were incubated with titrated amounts of allogeneic stimulator DC. Cells were cultured in 96-well U-bottom plates (Costar, Cambridge, MA) in complete medium for 3–5 days at 37°C and 5% CO2 in a humidified atmosphere. Cells were pulsed with [3H]thymidine (0.5 µCi/well (18.5 x 103 Bq/well); Amersham, Braunschweig, Germany) during the last 4 h of culture. [3H]Thymidine incorporation was measured using a liquid scintillation counter (Wallac, Turku, Finland). Responses are shown as mean cpm from triplicate wells.
IL-10 and IL-12 (p70) release
Differentiated MoDC were analyzed for functional (p70) IL-12 and
IL-10 release as described previously (27). Briefly,
4 x 104 DC were incubated with 4 x
104 J558-CD40L cells in the presence or absence
of 1000 U/ml rhIFN-
(CLB) in 200 µl complete medium. After 24
h, the supernatants were collected and stored at -20°C. IL-12 and
IL-10 concentrations were determined by capture ELISA (see
Cytokine and PGE2 detection).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To study the effect of tumor-derived soluble factors on DC
development, TDSN were prepared from 24-h cultures of colon tumor cell
lines and single cell suspensions of primary colon tumors. During the
generation of DC from monocytes in the presence of GM-CSF and IL-4,
TDSN were added to the culture medium in a concentration range of
0.3–30%. Using de novo CD1a expression as a measure for DC
development, primary TDSN effected a total inhibition at concentrations
of 3–30% (n = 3), whereas this was not yet achieved
at a concentration of 30% using colon tumor cell line-derived
supernatants (tested cell lines: WiDr, A2233, HT29, Colo320, and
SW620). In all subsequent experiments, 10–30% of TDSN were added to
culture medium to achieve an optimal suppression of DC differentiation.
Besides CD1a neo-expression, the down-regulation of CD14 and the
acquisition of a veiled morphology were also used throughout as
hallmarks of DC differentiation. The addition of primary tumor cell
supernatants to the MoDC cultures led to an inhibition of CD1a
expression, resulted in a maintained CD14 expression, and hampered the
development of a typical DC morphology. In contrast, inclusion of cell
line-derived supernatants could not completely prevent the monocytes
from adopting DC properties. The effects of primary tumor- and cell
line-derived supernatants on CD1a expression and DC morphology are
shown in Fig. 1
A. The superior
suppressive effect on MoDC development of primary tumor supernatants,
as compared with tumor cell line supernatants, is further demonstrated
by FACS analysis of CD1a neo-expression, CD14 expression, and DC
morphology (with the typical veiled appearance of DC translating into
high side light scatter (SSC) values) (Fig. 1B).
|
Identification of DC development-inhibitory factors in the colon TDSN: a dominant role for prostanoids in the suppression of MoDC differentiation
Dose response relationships for tumor-associated factors, which
were previously reported to hamper the differentiation of DC, i.e.,
IL-10, TGF-1, VEGF, IL-6, M-CSF (all in a
range of 0–125 ng/ml), and PGE2 (in a range of
0–350 ng/ml), were determined in the MoDC system (results for
PGE2 and IL-6 are shown in Fig. 1
C).
Indeed, the expected inhibitory effect on DC development (as judged on
the basis of CD1a expression) was found for all these agents. However,
only IL-10, IL-6, and PGE2 were found to be
present in active concentrations in the tested TDSN (exceeding the
IC20; Table I).
However, the maintained expression of CD14 (as observed with the
primary tumor supernatants) was found only for
PGE2 and IL-6, while the inhibitory effect on the
acquisition of DC morphology (based on SSC) was observed only for
PGE2 (data not shown). Thus,
PGE2 appeared to most accurately mimic the
observed DC-inhibitory effects of the primary TDSN.
|
1, VEGF, M-CSF, and IL-6 before addition
to the MoDC cultures. While it proved possible to neutralize
recombinant cytokines (at equivalent concentrations as present in the
used supernatants) and thus abrogate their suppressive effects, the
addition of the blocking Abs did not prevent the inhibition caused by
the addition of TDSN (based on CD1a and CD14 expression patterns and
SSC characteristics; data not shown). Not even a partial abrogation of
inhibition of DC development was observed. In contrast, the addition of
indomethacin (a combined COX-1 and -2 inhibitor) during the culture of
primary colon tumor cells to obtain conditioned supernatants led to the
elimination of the primary tumor-induced inhibitory effects on the
differentiation of DC from monocytes, signified by the presence of
CD1a, the absence of CD14, and high SSC values (Fig. 2). Treatment with indomethacin did not
affect the IL-6 and IL-10 concentrations in the TDSN, nor did the
addition of IL-6 and IL-10 neutralizing mAbs to the TDSN, generated in
the presence of indomethacin, further abrogate any inhibitory effects
on MoDC development (data not shown). Thus, prostanoids (the production
of which is dependent on COX-1/-2 activity and of which
PGE2 is the main constituent) were identified as
the factors directly responsible for the observed inhibitory effect of
primary TDSN in the MoDC differentiation model.
|
1 were detected
(Table I), thus confirming the involvement of different mechanisms in
the inhibition of DC development by primary tumors and the tested
tumor-derived cell lines. Both prostanoids and IL-6 contribute to tumor-induced inhibition of DC differentiation from CD34+ progenitor cells
To extend our findings to other pathways of DC differentiation, we next tested the effect of primary colon TDSN on the development of DC from peripheral blood CD34+ stem cells (G-CSF mobilized and isolated to 88–95% purity).
Similar to the MoDC model, differentiation of
CD1a+ DC from the CD34+
cells could be prevented by the addition of primary colon
tumor-conditioned supernatants (in equivalent concentrations). The
percentage of CD1a-positive cells was reduced from 28.7% in the medium
control to 6.4% (range, 5–8%) in the supernatant-containing
conditions, while a simultaneous increase in the percentage of
CD14+ cells was observed from 8.8 to 47.0%
(range, 46–49%) (n = 4, p < 0.00001
for both markers). Again, we did not observe this effect on DC
development by the cell line-derived supernatants (see Fig. 3A). In contrast to our
findings in the MoDC model, addition of indomethacin during the
generation of primary TDSN did not result in a total abrogation of the
inhibitory effect on DC differentiation from
CD34+ precursors (Fig. 3B). Further
neutralization experiments revealed an additional role in the
inhibition of DC differentiation for IL-6. While the use of
indomethacin and neutralizing IL-6 Abs separately hardly had any
effect, their combined use led to a near-complete abrogation of the
primary tumor-induced suppression of DC differentiation from
CD34+ progenitor cells (Fig. 3B).
|
The colon carcinoma single-cell suspensions used to generate TDSN
contained leukocyte infiltrate (10% on average). To address the
possible role of this infiltrate in the observed DC suppression, the
CD45+ tumor-infiltrating leukocytes were depleted
from the single-cell suspensions by indirect panning with anti-CD45
mAbs. Unbound colon cells and bound leukocytes were used separately to
obtain conditioned supernatants at a concentration of 1 x
106 cells/ml over a period of 24 h. FACS
analysis showed that the inhibition of MoDC development was associated
with the addition of tumor cell-conditioned supernatants and not with
the addition of leukocyte-conditioned supernatants (Fig. 4). Therefore, it can be concluded that
the infiltrating immune effector cells were not responsible for the
observed suppression. This corresponds with the presence of
PGE2 in the CD45-negative supernatants (at
similar levels to the supernatants of the unseparated cell suspensions)
and its absence (i.e., below detection level) in the
CD45+ leukocyte-conditioned supernatants (data
not shown; n = 3). Similarly, IL-6 release, partly
responsible for suppression in the CD34+ model
system, was most likely tumor derived, as IL-6 could not be detected in
the infiltrate-conditioned supernatant, while the CD45-negative
cell-conditioned supernatants contained high concentrations of IL-6
(data not shown).
|
To ascertain any restriction of prostanoid-induced DC suppression
to tumor tissues, normal tissue samples were collected from colon tumor
resection specimens. Both normal and tumor single-cell suspensions were
thus generated from the same surgical samples. Supernatants derived
from normal colon single-cell suspensions contained
PGE2 in active, but significantly lower,
concentrations than their corresponding malignant counterparts (146
ng/ml (range, 57.6–319 ng/ml) vs 282 ng/ml (range, 142–424 ng/ml),
respectively; n = 9; p = 0.009 in a
paired t test). Accordingly, suppressive effects of normal
colon cell-derived supernatants (NDSN) were clearly detectable, but
were less profound than those of TDSN (Fig. 5). Testing the effect of coculture of
the normal or tumor cells with the specific COX-1 inhibitor SC560
revealed the suppressive effect exerted by the NDSN to be fully COX-1
dependent (Fig. 5A). In contrast, the inhibitory effect of
the TDSN could only partially be eliminated through coculture with
SC560. Complete elimination of the suppressive effect required
simultaneous COX-1 and COX-2 inhibition by the indiscriminate COX
inhibitor indomethacin, indicating the additional involvement of COX-2
in the production of tumor-derived prostanoids (Fig. 5A). In
Fig. 5B the mean values from four NDSN/TDSN couples are
shown, revealing a significant difference in CD1a and CD14 expression
between MoDC generated in the presence of NDSN or TDSN, which were
derived from colon cultures with the COX-1 inhibitor SC560
(p = 0.037 and p = 0.004,
respectively).
|
Although the DC differentiation-inhibitory effect of IL-6 was
previously reported for tumor cell lines derived from various tissue
types, a similar effect for primary tumor-derived prostanoids is a
novel finding. To exclude that the inhibitory effect of the
tumor-derived prostanoids on the DC development was colon-restricted,
supernatants conditioned by mamma (n = 2) and renal
cell (n = 2) carcinomas, and by melanomas
(n = 2), in the absence or presence of indomethacin,
were similarly tested in the MoDC and CD34+ model
systems. Similar inhibitory effects were observed, regardless of the
tissue origin of the tested tumors (Fig. 6). CD1a expression was inhibited, CD14
expression was maintained (in the MoDC model) or acquired (in the
CD34+ model), and SSC values remained low when
TDSN conditioned in the absence of indomethacin were used (consistent
with the conservation or acquisition of the monocyte phenotype and
morphology). The opposite was observed after the addition of TDSN
conditioned in the presence of indomethacin (evidence of the
acquisition of DC phenotype and morphology). As before, prostanoids
appeared to only partially contribute to the observed inhibition of
differentiation in the CD34-derived DC model (Fig. 6B).
|
Having established the inhibitory effect of tumor-derived soluble
factors on DC development, we next tested their effect on the ability
of DC, generated in their presence, to stimulate T cells in an
allogeneic MLR. When colon tumor-conditioned supernatants were added
during MoDC generation, a slight but consistent stimulatory effect was
observed in the MLR (Fig. 7A).
In contrast, a considerable inhibitory effect was observed in the
CD34-derived DC system (Fig. 7A). To assess the involvement
of IL-6 and prostanoids in these observed effects, neutralization
experiments were performed in which TDSN, generated with or without
indomethacin, were incubated with IL-6 blocking Abs before their
addition to the DC cultures. Supernatants generated in the presence of
indomethacin no longer increased the T cell-stimulatory ability of
MoDC, while preincubation with IL-6 blocking Abs had no effect (Fig. 7B). In contrast, the use of IL-6 Abs led to a
partial abrogation of the tumor-induced inhibition of allogeneic T cell
stimulation by CD34-derived DC; this effect was further enhanced in
indomethacin-generated TDSN (Fig. 7B). The observed T
cell-stimulatory capacities of the MoDC and CD34-derived DC, generated
in the presence or absence of TDSN, corresponded to levels of CD86 and
HLA-DR expression: addition of tumor supernatants during DC
differentiation resulted in higher levels of CD86 and HLA-DR on MoDC
and lower levels on CD34-DC (Fig. 8).
While no pronounced effects of the TDSN were observed on the expression
of CD40 and CD80 (data not shown), CD54 expression appeared to be
affected by the TDSN in an opposite way from CD86 and HLA-DR
expression, both on MoDC and on CD34-derived DC (Fig. 8). In
correspondence with the observed effects in the MLR (Fig. 7),
generation of TDSN in the presence of indomethacin completely reversed
the stimulatory effect of the TDSN on the expression levels of HLA-DR
and CD86 on MoDC (Fig. 8A). While the use of indomethacin
similarly reversed the effects of TDSN on CD86 and HLA-DR levels on
CD34-derived DC, this reversal was even more complete after
preincubation of the indomethacin-generated TDSN with IL-6-neutralizing
Abs (Fig. 8B). Indeed, the inclusion of anti-IL-6 even
resulted in higher DR and CD86 levels than in the medium control
condition, possibly due to neutralization of autocrine IL-6 production
by the DC precursors (Fig. 8B).
|
|
To determine the effect of tumor-derived prostanoids during DC
differentiation on subsequent IL-12 and IL-10 production,
differentiated MoDC were stimulated by CD40L-transfected cells in the
presence or absence of IFN-. Fig. 8A shows a profound
inhibition of IL-12 production by MoDC differentiated in the presence
of TDSN, after stimulation by CD40L and IFN-. However, this effect
could only be partially prevented by indomethacin (complete prevention
was observed in one of the three TDSN tested). In contrast, IL-10
release by MoDC was increased as a result of the addition of TDSN
during differentiation, and in all cases this could be prevented by
indomethacin during the generation of TDSN (Fig. 9A). IL-12:IL-10 ratios
clearly demonstrate the favored development of DC that are more skewed
toward the induction of type-2 T cell responses under the influence of
tumor-derived prostanoids (Fig. 9B). CD40L stimulation
without IFN- resulted in the production of considerably lower
amounts of IL-12, which dropped below detection levels in the TDSN
conditions, and higher amounts of IL-10, but the same effects of the
(indomethacin-conditioned) TDSN were observed in both cases (data not
shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Prostanoids are derived from membrane-associated arachidonic acids. COX
proteins control the rate of prostanoid synthesis by catalyzing the
conversion of arachidonic acid to PGH2, the
common PG precursor. COX-1 is constitutively expressed, while COX-2
is induced by a variety of proinflammatory signals (28).
In various tumor types, including colon, breast, lung, and
melanoma, overexpression of COX-2 has been observed (21, 22, 29, 30, 31). Moreover, high levels of COX-2 expression correlate to
poor clinical outcome. In keeping with these findings, we found a
significant contribution of COX-2-regulated prostanoids to primary
colon tumor-induced inhibition of DC development. In contrast, the
significantly less pronounced inhibitory effect of soluble factors
derived from adjacent nonmalignant tissue could be completely blocked
by the COX-1 inhibitor SC560. Numerous reports have drawn attention to
a role for prostanoids (most notably for PGE2) in
the carcinogenesis of solid tumors. Nonsteroidal anti-inflammatory
drugs, which inhibit their synthesis and of which indomethacin is an
example, have been reported to reduce the risk of developing colorectal
cancer (32). Our results indicate that prostanoids may
also contribute to carcinogenesis through the inhibition of DC
development, thus seriously crippling immune surveillance. This
certainly ties in with numerous reports on hampered DC development in
cancer patients and a negative prognostic value of reduced numbers of
tumor-infiltrating DC (12, 14, 33). Previous reports
indicate that DC are most sensitive to the modulatory effects of
PGE2 (34). Because this is the most
prominently expressed prostanoid and its concentration in the tested
supernatants correlated well with the observed effects (data not
shown), it seems most likely that this is the agent responsible for the
observed inhibition of DC differentiation in our model systems.
Although PGE2 production by monocytes has also
been observed, we clearly established that the intratumoral leukocyte
infiltrate was not involved in the DC suppression. By the same token,
PGE2-induced IL-10 release by tumor-infiltrating
lymphocytes could not be responsible, in keeping with the finding that
IL-10 neutralization could not abrogate the inhibition of DC
development. A direct inhibition of DC differentiation by
PGE2 was previously described (35).
In contrast with this inhibitory ability, PGE2
can actually stimulate DC maturation. This may account for the observed
increased allogeneic T cell stimulation by MoDC generated in the
presence of TDSN, even though they were obviously stinted in their
differentiation. This increased stimulatory ability was found to be
related to a prostanoid-induced increase in expression levels of CD86
and HLA-DR. Recent studies also point to autocrine stimulation of DC by
PGE2 and its importance in reaching full
phenotypic and functional maturity (36). However, these
autocrine production levels are 
1000-fold lower than the
PGE2 concentration in TDSN. Moreover, we found no
effect on DC differentiation of direct coculture with indomethacin
(data not shown). A tumor-associated increase in phenotypic MoDC
activation, characterized by an up-regulation of, among others, CD86
and HLA-DR, was also previously reported by Kiertscher et al.
(19) and Menetrier-Caux et al. (37), both
using tumor cell line-derived supernatants. Maturation induction of
MoDC thus appears to be a common feature of soluble tumor-derived
products from different sources, as reported by different research
groups, and is in marked contrast to effects observed on CD34-derived
DC (17, 37). Although tumor-derived soluble factors may
increase maturation and T cell-stimulatory capacity of MoDC, the
simultaneous reduction in IL-12 production and increase in IL-10
production will most likely lead to the generation of a deleterious
type-2 antitumor T cell response. We found this tumor-induced shift in
IL-12:IL10 ratio to be mostly prostanoid dependent. This is in keeping
by findings of Kalinski et al. (38), who showed that
PGE2-induced maturation skewed DC toward the
preferential induction of a type-2 T cell response. In addition,
Kiertscher et al. (19) reported the induction of apoptosis
in MoDC through coculture with tumor cell line-derived supernatants. We
were unable to reproduce this finding in our DC differentiation model
systems (apoptosis determined by annexin V/propidium iodide staining;
data not shown) by coculture of cell line supernatants or of
primary TDSN.
The maturational effect of TDSN appears to depend on the hematopoietic precursor origin of the DC, because TDSN inhibited rather than increased the allogeneic T cell-stimulatory ability of CD34-derived DC. Again, this effect correlated to CD86 and HLA-DR expression levels, which remained low in the presence of TDSN, in keeping with previous reports by Menetrier-Caux and colleagues (17, 37).The finding that CD54 showed opposite expression patterns from CD86 and did not correlate to T cell-stimulatory capacity of the DC is in keeping with previous reports that CD86 plays a more important role in the induction phase of the immune response (39).
The properties of prostanoids, and more specifically PGE2, with regard to the inhibition of DC differentiation and the skewing of type-1 to type-2 T cell responses, have been reported as essential to the maintenance of tolerance against food-derived Ags in the gut mucosa (40). Indeed, although significantly elevated in supernatants from malignant colon tumors, we also found high levels of PGE2 in supernatants from normal colon tissue cultures. To establish that the dominant effect of tumor-derived prostanoids on DC development was not merely a colon-associated trait, we also studied the effects of TDSN from breast, renal, and melanoma origin. A similarly overriding influence of prostanoids was found for these TDSN, in accordance with previous reports of elevated levels of PGE2 in these tumor types. Thus, prostanoid-induced inhibition of DC differentiation appears to be a generalized tumor-associated phenomenon.
The inability of the tested tumor cell lines to suppress DC development may be due to their inability to produce sufficient levels of prostanoids. Indeed, in one early passage tumor cell line generated from a primary colon tumor in our lab, we saw a decline in PGE2 levels with relatively low levels still detectable at passage number 4 (886 pg/ml). By passage number 6 PGE2 levels had dropped below the detection limit (data not shown). These observations indeed suggest that differences in PGE2 production between primary tumors and cell lines derive from in vitro propagation, which may result in methylation of the COX-2 gene, as previously described (21, 22). However, that this may not always be the case is indicated by previous reports of high-level prostanoid production by melanoma cell lines (21). Similarly, we observed only low levels of suppressive IL-6 in the five colon tumor cell lines we tested, while more heterogeneous levels were previously reported (17). Therefore, we conclude that expression of the various DC-inhibitory factors can be very heterogeneous among cell lines but appears to be more consistently high in primary tumors.
Previous studies using tumor cell lines implicated IL-6, M-CSF (17), VEGF (18), and recently gangliosides (41) as tumor-derived factors inhibiting the development of DC. This was observed with both MoDC and CD34-derived DC. However, our results clearly show that in primary TDSN prostanoids play a dominant, overriding role. Although present at potentially active concentrations, no role for either IL-10 or IL-6, determined by incubation with neutralizing mAbs, could be established in the inhibition of MoDC differentiation by primary TDSN, not even after the release of prostanoids was blocked by indomethacin. The presence of high concentrations of IL-4 (1000 U/ml) in our MoDC cultures may largely explain this. Indeed, Menetrier-Caux and colleagues (17, 37) demonstrated that the inhibition of differentiation of MoDC and CD34-derived DC by supernatants of renal cell carcinoma lines was directly mediated by IL-6, but that IL-4 protected against this inhibition. The inhibitory effect described by Menetrier-Caux and colleagues (17, 37) was related to up-regulation of the M-CSF receptor CD115 and down-regulation of the GM-CSF receptor CD116 on the developing DC. Again, our use of high concentrations of IL-4 during DC differentiation may explain why we did not observe these effects in our system, neither with the use of cell line supernatants nor using primary TDSN (data not shown). We did observe a clear IL-6-mediated effect in the CD34-derived DC model, even in the presence of IL-4. However, to completely abrogate inhibition of DC differentiation in this model, both prostanoids and IL-6 had to be eliminated from the primary TDSN. In contrast, we found an overriding role of primary tumor-derived IL-6 in the inhibition of DC development from CD34+ precursors in the absence of IL-4 in comparison to tumor-derived prostanoids, which is in keeping with findings by Menetrier-Caux and colleagues (17, 37) (C. Sombroek, F. van den Eertwegh, R. Scheper, and T. de Gruijl, manuscript in preparation).
The relative contribution of the different suppressive factors to the cancer-associated inhibition of DC development in vivo remains to be assessed. Almand et al. (42) showed that decreased numbers of functional DC in the blood and regional lymph nodes of cancer patients correlated well with tumor stage and duration of the disease, as well as with serum levels of VEGF, which is indicative of a suppressive systemic effect of VEGF. We were unable to establish an inhibitory effect of primary tumor-derived VEGF, which was present in the TDSN at subactive levels in vitro. However, the in vivo DC-inhibitory activity of VEGF has been reported at considerably lower concentrations than in vitro (11). Therefore, it is conceivable that the relatively low VEGF levels in the primary TDSN were not sufficient for an in vitro effect but might still be able to affect DC development in vivo. Similar studies correlating PGE2 serum levels to DC number and activation status in vivo need to be performed to establish a possible systemic effect of tumor-related prostanoid overproduction and hampered DC differentiation. Alternatively, prostanoids may exert a profound influence on local DC functions. In this regard, the contrasting phenotypic effects of prostanoids on the processes of DC differentiation and maturation of monocytic or CD34+ progenitors deserve further study.
This study, as well as previous studies, determined the effects of soluble tumor factors on DC development. However, a recent report by Chomarat et al. (43) revealed that physical contact with tumor cells and/or fibroblasts skewed the differentiation of monocytes from T cell-stimulatory DC to scavenging macrophages without T cell-stimulatory capacities. This phenomenon was found to be IL-6 dependent. It is clear that interactions between soluble and contact-dependent factors will require further study to unravel the intricate processes leading up to the tumor-induced inhibition of DC development; however, it is equally clear that, besides tumor-derived cell lines, such studies should make use of primary tumor cells and TDSN to draw conclusions that more reliably reflect the in vivo situation. Based on our findings, an important role for primary tumor-derived prostanoids is certainly indicated. Clearly, the correct identification of the factors responsible for hampered DC development in cancer patients is of utmost importance for the selection of appropriate immunotherapeutic approaches to overcome this defect (1, 44, 45).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Abbreviations used in this paper: DC, dendritic cell; CD40L, CD40 ligand; rh, recombinant human; COX, cyclooxygenase; MFI, mean fluorescence intensity; SSC, side light scatter; FSC, forward light scatter; MoDC, monocyte-derived DC; NDSN, normal colon cell-derived supernatant; TDSN, tumor-derived supernatant; VEGF, vascular endothelial growth factor.
Received for publication March 21, 2001. Accepted for publication February 27, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
activation in irradiated murine mammary gland. J. Clin. Invest. 93:892.
This article has been cited by other articles:
|
|
 |
|
  E. Eruslanov, S. Kaliberov, I. Daurkin, L. Kaliberova, D. Buchsbaum, J. Vieweg, and S. Kusmartsev Altered Expression of 15-Hydroxyprostaglandin Dehydrogenase in Tumor-Infiltrated CD11b Myeloid Cells: A Mechanism for Immune Evasion in Cancer J. Immunol., June 15, 2009; 182(12): 7548 - 7557. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D.-M. Kuang, Q. Zhao, J. Xu, J.-P. Yun, C. Wu, and L. Zheng Tumor-Educated Tolerogenic Dendritic Cells Induce CD3{epsilon} Down-Regulation and Apoptosis of T Cells through Oxygen-Dependent Pathways J. Immunol., September 1, 2008; 181(5): 3089 - 3098. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. J. J. van der Vliet, R. Wang, S. C. Yue, H. B. Koon, S. P. Balk, and M. A. Exley Circulating Myeloid Dendritic Cells of Advanced Cancer Patients Result in Reduced Activation and a Biased Cytokine Profile in Invariant NKT Cells J. Immunol., June 1, 2008; 180(11): 7287 - 7293. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Tonello and G. Poli Tubal ectopic pregnancy: macrophages under the microscope Hum. Reprod., October 1, 2007; 22(10): 2577 - 2584. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. D. de Gruijl, C. C. Sombroek, S. M. Lougheed, D. Oosterhoff, J. Buter, A. J. M. van den Eertwegh, R. J. Scheper, and H. M. Pinedo A Postmigrational Switch among Skin-Derived Dendritic Cells to a Macrophage-Like Phenotype Is Predetermined by the Intracutaneous Cytokine Balance. J. Immunol., June 15, 2006; 176(12): 7232 - 7242. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. van de Ven, M. C. de Jong, A. W. Reurs, A. J. N. Schoonderwoerd, G. Jansen, J. H. Hooijberg, G. L. Scheffer, T. D. de Gruijl, and R. J. Scheper Dendritic Cells Require Multidrug Resistance Protein 1 (ABCC1) Transporter Activity for Differentiation J. Immunol., May 1, 2006; 176(9): 5191 - 5198. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Gottfried, L. A. Kunz-Schughart, S. Ebner, W. Mueller-Klieser, S. Hoves, R. Andreesen, A. Mackensen, and M. Kreutz Tumor-derived lactic acid modulates dendritic cell activation and antigen expression Blood, March 1, 2006; 107(5): 2013 - 2021. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. P.J.M. van Gisbergen, C. A. Aarnoudse, G. A. Meijer, T. B.H. Geijtenbeek, and Y. van Kooyk Dendritic Cells Recognize Tumor-Specific Glycosylation of Carcinoembryonic Antigen on Colorectal Cancer Cells through Dendritic Cell-Specific Intercellular Adhesion Molecule-3-Grabbing Nonintegrin Cancer Res., July 1, 2005; 65(13): 5935 - 5944. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. de Jong, M. I. E. van Poelgeest, J. M. van der Hulst, J. W. Drijfhout, G. J. Fleuren, C. J. M. Melief, G. Kenter, R. Offringa, and S. H. van der Burg Human Papillomavirus Type 16-Positive Cervical Cancer Is Associated with Impaired CD4+ T-Cell Immunity against Early Antigens E2 and E6 Cancer Res., August 1, 2004; 64(15): 5449 - 5455. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Brown and R. N. DuBois Cyclooxygenase-2 in Lung Carcinogenesis and Chemoprevention: Roger S. Mitchell Lecture Chest, May 1, 2004; 125(5_suppl): 134S - 140S. [Full Text] [PDF]
|
|
|
|
|
|
D. Giordano, D. M. Magaletti, E. A. Clark, and J. A. Beavo Cyclic Nucleotides Promote Monocyte Differentiation Toward a DC-SIGN+ (CD209) Intermediate Cell and Impair Differentiation into Dendritic Cells J. Immunol., December 15, 2003; 171(12): 6421 - 6430. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Rivoltini, C. Castelli, M. Carrabba, V. Mazzaferro, L. Pilla, V. Huber, J. Coppa, G. Gallino, C. Scheibenbogen, P. Squarcina, et al. Human Tumor-Derived Heat Shock Protein 96 Mediates In Vitro Activation and In Vivo Expansion of Melanoma- and Colon Carcinoma-Specific T Cells J. Immunol., October 1, 2003; 171(7): 3467 - 3474. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Hayashi, T. Hideshima, M. Akiyama, N. Raje, P. Richardson, D. Chauhan, and K. C. Anderson Ex vivo induction of multiple myeloma-specific cytotoxic T lymphocytes Blood, August 15, 2003; 102(4): 1435 - 1442. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
