[Cancer Research 62, 2583-2591, May 1, 2002]
© 2002 American Association for Cancer Research
Glioma-associated Hyaluronan Induces Apoptosis in Dendritic Cells via Inducible Nitric Oxide Synthase
Implications for the Use of Dendritic Cells for Therapy of Gliomas1
Tianbing Yang,
Timothy F. Witham,
Lorissa Villa,
Melanie Erff,
Jason Attanucci,
Simon Watkins,
Douglas Kondziolka,
Hideho Okada,
Ian F. Pollack and
William H. Chambers2
Brain Tumor Center, University of Pittsburgh Cancer Institute [T. Y., T. F. W., L. V., M. E., J. A., H. O., I. F. P., W. H. C.], and Departments of Neurological Surgery [T. Y., T. F. W., M. E., J. A., D. K., H. O., I. F. P.], Pathology [L. V., W. H. C.], and Cell Biology and Physiology [S. W.], University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15213
 |
ABSTRACT
|
|---|
As a means of enhancing immunity to gliomas, we investigated local deliveryof rat, bone marrow-derived dendritic cells (DCs) into rat 9L gliosarcoma tumors and into 9L tumors induced to undergo apoptosis by gamma knife radiosurgery. Contrary to other tumors, local delivery of DCs had no therapeutic effect on 9L gliomas, even when tumor apoptosis was induced via radiosurgery, which leads to efficient "loading" of the DCs with tumor antigen. To determine whether antigen-presenting cells, such as DCs, were viable in tumors, we carried out multiparametric staining of 9L tumors, using phycoerythrin-conjugated OX6 (MHC class II) or OX62 (DC specific) and FITC-labeled Val-Ala-Asp-fluoromethyl ketone (FITC-VAD-FMK; activated caspases). It was determined that DCs were undergoing apoptosis in these tumors. We therefore sought to determine which glioma cell surface receptors or components of the extracellular matrix in gliomas influenced DC viability. Hyaluronan (HA) is a major component of glioma extracellular matrix and has been found to support tumor cell migration and metastasis. However, its influence on the immune system, and particularly on DCs, via its receptor CD44 is not well documented. Using reverse transcription-PCR, Northern blot, and Western blot analyses, we determined that HA stimulated production of inducible nitric oxide synthase (iNOS) in DCs. NO production by HA-stimulated DCs was then verified biochemically. NO production was dependent on the size of HA; intermediate HA fragments had the greatest capacity to induce NO production in DC, whereas completely digested HA oligosaccharides failed to induce NO. Furthermore, N-monomethyl-L-arginine, an inhibitor of iNOS, completely blocked HA-induced NO production by DCs. Because induction of NO results in the induction of apoptosis in macrophages as well as other cells, DCs treated with HA were examined for apoptosis in terminal deoxynucleotidyl transferase (TdT)-mediated dUTP biotin nick-end labeling assays. It was demonstrated that HA induced apoptosis in DCs and that induction of apoptosis was dependent on the production of NO because it was entirely inhibited by N-monomethyl-L-arginine. Using flow cytometric analyses with FITC-VAD-FMK, which is specific for activated caspases, we also determined that induction of apoptosis in DCs with HA could be titrated. Coincubation of 9L tumor cells with DCs was found to induce apoptosis in DCs as indicated by fluorescent staining with FITC-VAD-FMK. Specificity of this reaction for CD44-HA interactions was determined by pretreatment of DCs with anti-CD44 or pretreatment of 9L tumor cells with hyaluronidase, which blocked the induction of apoptosis in DCs. These data indicate that HA expressed by gliomas may contribute to their immunosuppressive effects by promoting apoptosis among professional antigen-presenting cells such as DCs via iNOS induction after CD44-HA interactions.
|
INTRODUCTION
|
|---|
Although significant technical advances in surgery and radiotherapy for malignant brain tumors have emerged in recent years, they have not had a significant impact on clinical outcomes (1)
. This has led to the emergence of alternative approaches to therapy, including immunotherapy. Recent insights into the immunobiology of the central nervous system and of gliomas have provided rational avenues to the development of immunotherapies for gliomas, and numerous approaches have been attempted, including local delivery of immunotoxins (2
, 3)
, adoptive transfer of activated natural killer cells (4
, 5)
, adoptive transfer of activated T cells (6, 7, 8, 9)
, and active immunization to promote specific reactivity (10
, 11)
. Active immunization protocols have included the use of genetically modified tumors (e.g., cytokine gene-transduced gliomas; Refs. 12, 13, 14, 15, 16
) or APCs3
loaded with tumor lysates (17
, 18)
or MHC class I-associated peptides (19, 20, 21)
. In fact, numerous approaches to inducing specific responses with APCs have been attempted. In the past 10 years, it has been unequivocally established that DCs serve as the most potent APCs for induction of immunity in naive individuals (22
, 23)
.
Potent immune responses have been induced when DCs were used as adjuvants for immunization, and there are numerous reports of enhanced induction of immunity to various tumors (24, 25, 26, 27, 28)
, including gliomas (17, 18, 19, 20, 21)
. In fact, several approaches have been forwarded as being of significant benefit for cancer therapy. These include systemic immunization with DCs loaded with tumor cell lysates (17
, 18
, 29, 30, 31, 32, 33)
, MHC class I-derived peptides (34
, 35)
, or tumor cell apoptotic bodies (36)
, which have been demonstrated to induce potent immunity and/or therapeutic benefit (27
, 28)
. In those cases where TRAs are not defined, DCs have been loaded with tumor cell lysates, tumor-derived RNA, or peptides extracted from MHC-peptide complexes. Each of these approaches has also been reported to induce potent immunity and/or therapeutic benefits (37
, 38)
.
Because it is frequently difficult to obtain sufficient quantities of tumor cells for loading of DCs ex vivo for therapy, alternative approaches have been proposed that might be more broadly applicable. In those cases where specific TRAs have been identified, DCs pulsed with MHC class I or II antigen-restricted peptides have been used for therapeutic immunizations (34
, 35
, 39
, 40)
. This approach has the advantage of being uniquely specific for defined antigens known to be important for tumor rejection. However, it is not applicable in many individuals because there are restrictions for expression of peptide antigens in specific MHC class I or II alleles (34
, 35
, 39
, 40)
.
Because of the limitations of the use of TRAs or of obtaining sufficient tumor material for isolation of antigenic peptides, or even tumor lysates, for efficient loading of DCs, alternative approaches have been used. In particular, it has been suggested that intratumoral delivery of DCs might serve as an efficient means of generating antigen-loaded DCs (41)
. In fact, several reports from both preclinical and clinical studies have indicated that this approach has potential therapeutic benefit for treatment of melanoma and breast cancer (41, 42, 43, 44)
. Furthermore, it has been suggested that combining local delivery of DCs with either chemo- or radiotherapy, which induces apoptosis in tumor cells, might be an even more effective therapy.
On the basis of these reports, we investigated whether local delivery of DCs intratumorally into established 9L gliosarcomas would provide therapeutic benefit. In addition, because DCs load efficiently with tumor apoptotic bodies (36
, 45, 46, 47, 48, 49)
, we attempted to improve therapeutic responses by inducing apoptosis in established 9L tumors, using gamma knife RS. In these experiments, we determined that intratumoral delivery of DCs into established, i.c. gliomas did not provide therapeutic benefit even when coupled with induction of RS to generate tumor apoptotic bodies. Furthermore, we determined that HA, which is produced by glioma cells (50, 51, 52)
and is a major component of ECM in gliomas (53
, 54)
, can induce apoptosis in DCs by stimulating production of iNOS. These data represent the first description of the induction of iNOS in DCs by tumors and indicate that gliomas may modulate immune responses via elimination of intratumoral APCs such as DCs.
|
MATERIALS AND METHODS
|
|---|
Animals
Fischer 344 rats (male; 100–125 g) were obtained from Taconic Farms (Germantown, NY) and housed in a specific pathogen-free barrier facility at the University of Pittsburgh Cancer Institute Central Animal Facility. All experimental procedures were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh.
Cell Culture Procedures
Tumor Cells.
9L cells were cultured as stationary suspension cultures in a humidified incubator at 37°C under 5% CO2 tension in air in CM composed of RPMI 1640 supplemented with 5% (v/v) heat-inactivated fetal bovine serum, 1 mM sodium pyruvate, 0.1 mM nonessential amino acids, 10 mM HEPES buffer, 5 x 10-5 M 2-mercaptoethanol, 2 mM L-glutamine, 50 units/ml penicillin, and 50 µg/ml streptomycin (all from Atlanta Biologicals).
DCs.
BM was isolated from femurs and tibias of rats and passed through wire mesh to remove pieces of bone and debris as described previously (55)
. DCs were derived from day 7 or 8 cultures of BM in CM (6 x 106 cells/well) supplemented with 1000 units/ml rat granulocyte macrophage colony-stimulating factor (PharMingen) and 1000 units/ml rat interleukin-4 (PharMingen). On day 7 or 8, cells were dislodged by pipetting and subcultured at 8 x 105 cells/ml in 100-mm dishes for an additional 24 h. DC cultures were characterized phenotypically and were typically OX6+ (MHC class II+; 92%), B7.1+ (CD80+; 64%), B7.2+ (CD86+; 75%), OX62+ (62%), CD44+ (98%; data not shown), which is representative of rat BM-derived DCs (19
, 55, 56, 57)
.
|
Stereotactic Surgery and Tumor Implantation
|
|---|
i.c. tumors were established as described previously (58
, 59)
. Rats were anesthetized with ketamine (100 mg/kg) and acepromazine (0.02 mg/kg) and immobilized in a stereotactic frame. An opening was made with an 18-gauge needle tip in the frontal bone via a small right paramedial scalp incision. The site for craniectomy was 2 mm lateral and 2 mm anterior to the bregma. 9L cells (1 x 104) in 5 µl of HBSS were injected stereotactically into the right frontal lobe with a Hamilton syringe, at a depth of 3 mm. Immediately after tumor implantation, a 1-mm section of 25-gauge needle was placed in the craniectomy site for later RS targeting. At 24 h post-RS, DCs (2 x 106 cells/5 µl) were introduced intratumorally, by a Hamilton syringe, through the craniectomy site.
|
Gamma Knife RS
|
|---|
Gamma knife RS was performed as described previously (58
, 59)
. Animals treated with RS were targeted 4 mm perpendicular to the extradural needle fragment in the right frontal lobe, which corresponded to the center of the site of tumor injection. The 4-mm collimator of the 201-source Cobalt 60 Model U Gamma unit (Elektra Instruments, Atlanta, GA) was used for RS. The 50% isodose line (35 Gy) matched the approximate tumor size on the day of RS treatment (day 7) based on prior studies.4
All animals treated with RS received a maximum dose of 70 Gy. Previous data have shown that this dose does not cause necrosis in the normal rat brain at 90 days post-treatment (59)
.
|
Flow Cytometric Analyses
|
|---|
Cell Surface Phenotype.
Phenotypic analyses of DCs were performed as described previously (55, 56, 57)
. Fluorochrome-conjugated anti-MHC class II (OX6), anti-B7.1 (CD80), anti-B7.2 (CD86), anti-CD44 (all from PharMingen, Mountain View, CA), and antirat DC (OX62; Serotec, Oxford, United Kingdom) were purchased from commercial vendors. Cells were then analyzed in a FACScan Plus cytometer (Becton Dickinson, Mountain View, CA). Electronic gates for DCs were set using forward and orthogonal light scatter, and markers were set based on background signal after incubation of cells with appropriate isotype-matched antibodies. A minimum of 10,000 events was collected for analysis of each cell population. FACScan data analyses were performed using REPROMAN software (TRUEFACTS Software, Seattle, WA).
VAD-FMK Staining of Activated Caspases in DCs.
To determine whether DCs were undergoing apoptosis after incubation with 9L cells, we incubated 9L and DCs (1:1) for 48 h in CM. Cells were then analyzed using anti-MHC class II (OX6-PE; PharMingen) and fluorochrome-conjugated VAD-FMK (FITC; Promega, Madison, WI), which binds activated caspases (60)
. DCs were discriminated from 9L cells by gating on MHC class II+ cells. One-parameter staining was also carried out with FITC-VAD-FMK for analysis of DCs treated with various concentrations of HA (Sigma) for 48 h. A minimum of 10,000 events was collected for analysis of each cell population. FACScan data analyses were performed using REPROMAN software.
|
Detection of Apoptosis by TUNEL Assays
|
|---|
TUNEL assays were performed using a commercially available in situ apoptosis detection kit (Intergen, Purchase, NY) according to the manufacturer’s instructions. DCs were treated with HA (100 µg/ml) for 48 h, at which point the DCs (5 x 107 cells/ml) were fixed in 1% paraformaldehyde in PBS for 10 min at room temperature. Fifty µl of the fixed cell suspension were then centrifuged onto a siliconized microscope slide, followed by incubation for 60 min with TUNEL reaction mixture in a humidified incubator at 37°C. The reaction was terminated with stop/wash buffer, and cells were further stained with an antidigoxigenin-peroxidase conjugate. Color was developed using 3,3'-diaminobenzidine, and each specimen was counterstained in Mayer’s hematoxylin solution.
|
Fluorescence Microscopy
|
|---|
9L tumors were established i.c. as described previously (58
, 59) . At day 12, animals were euthanized and perfused with 1% paraformaldehyde; brains bearing tumors were removed, suspended in 30% sucrose overnight, and then snap-frozen. Sections were then sliced on a cryostat, placed on slides, and stained with PE-conjugated OX6 (MHC class II; 10 µg/ml) or PE-OX62 (DC marker; 10 µg/ml; Ref. 56
) and FITC-VAD-FMK (10 µM) for 15 min on ice. Controls consisted of isotype-matched control antibody staining and PBS. Nuclei of cells were identified by staining with Hoescht 33258 (2 µg/ml; Sigma).
|
RT-PCR and Northern Blot Analyses
|
|---|
RNA was extracted from DCs using RNAzol (Biotecx Laboratories Inc., Houston, TX) according to the manufacturer’s protocol. RNA from control and HA-treated DCs was analyzed for iNOS expression by semiquantitative RT-PCR. The constitutively expressed GAPDH gene was used as a measure of the amount of input RNA. The primer sequences for rat iNOS were synthesized according to the published sequences (GenBank accession no. NM_012611): forward primer, 5'-CATTCTGAAGCCCCGCTACTA-3' (nucleotides 2715–2735); reverse primer, 5'-TTCTGCAGGATGTCTTGAACG-3'(nucleotides 3177–3197). The predicted product size is 483 bp. For GAPDH, the forward primer was 5'-TCCCTCAAGATTGTCAGCAA-3', and the reverse primer was 5'-AGATCCACAACGGATACATT-3'. The predicted product size is 308 bp. The sequence of the amplified fragment of iNOS was compared with the published sequence for iNOS. To verify the identity of the RT-PCR products, we performed Northern blot analyses using labeled PCR products for iNOS.
|
Western Blot Analyses
|
|---|
DCs were washed three times with PBS and resuspended in extraction buffer [100 mM Tris (pH 8.1), protease inhibitor cocktail (Sigma), and 0.2% Triton X-100] and homogenized. The protein concentration was determined by a commercial kit (Bio-Rad). Two hundred µg of total cell lysate were fractionated by SDS-PAGE (10% acrylamide), transferred to an Immobilon-P transfer membrane (Millipore, Bedford, MA), blocked, washed, incubated with polyclonal anti-iNOS antibody (Calbiochem, San Diego, CA) or anti-
-tubulin monoclonal antibody (Oncogene Science Inc.), and then developed with an enhanced chemiluminescence system according to the manufacturer’s instructions (Amersham).
|
Measurement of Nitrite
|
|---|
The nitrite concentration in DC culture supernatants, which is a reflection of NO production, was measured using the Griess reagent (1% sulfanilamide, 0.1% naphthylethylenediamine dihydrochloride, 2.5% H3PO4; Sigma). DCs (4 x 105/200 µl) were cultured with 100 µg/ml HA. After 48 h, 100 µl of the supernatants from DC cultures were mixed with 100 µl of Griess reagent. After a 10-min reaction at room temperature, the absorbance (540 nm) was measured using an automatic plate reader. The nitrite concentration was determined by comparison to a sodium nitrite standard curve in CM.
|
Enzymatic Digestion of HA
|
|---|
High molecular weight rooster comb HA (4 mg/ml in H2O; Sigma) was mixed with 10x digestion buffer [100 mM CH3COONa (pH 4.0)] and HAse from bovine testes (1 unit/10 µg of HA; Sigma). The reaction was carried out at 37°C, and at 20-min intervals, one-fifth of the mixture was put into a boiling water bath to terminate the reaction. The last fraction was digested completely by the addition of 500 units/10 µg to produce mainly tetrasaccharide fragments (61)
. The product sizes were determined by agarose gel electrophoresis for analysis of HA molecular weight distribution, as described previously (62)
. Volumes corresponding to 10 µg of HA from each digestion mixture were loaded into wells of a 0.5% agarose gel in TAE buffer [40 mM Tris, 5 mM CH3COONa, 0.9 mM disodium EDTA (pH 7.9)]. Electrophoresis was carried out at room temperature for 6 h with a constant voltage of 50 V. Immediately after the run, the gel was placed in a solution containing 0.005% Stains-All (Sigma) in 50% ethanol. The gel was stained over night under a light-protective cover at room temperature. Gels were photographed over a translucent light box.
|
Enzymatic Treatment of 9L Tumor Cells
|
|---|
9L cells were treated with HAse (50 units/106 cells) for 2 h at 37°C. After treatment, cells were washed twice in PBS and resuspended in CM.
|
Statistical Analyses
|
|---|
The log-rank test was used to analyze differences in survival data between control animals and animals receiving RS and/or DCs for significance. Statistical significance in experiments investigating in vitro effects of HA on induction of NO were determined using a Student’s t test.
|
RESULTS
|
|---|
Local Delivery of DCs into i.c. 9L Tumors Does Not Provide a Therapeutic Advantage.
To determine whether local delivery of DCs into established, i.c. 9L tumors provided any survival advantage, we implanted 9L tumor cells (1 x 104) in the right frontal lobe of rats, and on days 7 and 8, we injected 2 x 106 DCs into each rat. These rats were monitored daily for neurological symptoms and for survival. As illustrated in Fig. 1
, there was no survival advantage for i.c. delivery of DCs into 9L tumors. As a positive control for these experiments, we included a cohort of rats given RS on day 7 after tumor implantation, which resulted in long-term survival (>100 days) of 25% of treated rats (4 of 16). This is comparable to previously reported data (59)
. The experiment shown is representative of seven experiments that included a comparison of control 9L tumors and 9L tumors into which DCs were injected. Furthermore, because RS induces apoptosis in 9L and other tumors (59
, 63)
and because apoptotic bodies of tumors efficiently load DCs (36
, 45, 46, 47, 48, 49
, 59)
, we included a cohort of rats undergoing RS and receiving intratumorally delivered DCs. There was no survival advantage in rats undergoing RS and receiving DCs compared with rats that only underwent RS; 
20% of treated rats were long-term survivors. These data are in contrast to previous reports suggesting a therapeutic benefit of local delivery of DCs in breast cancer and metastatic melanoma (41, 42, 43, 44)
. Therefore, we began to investigate whether interactions of DCs with tumor cells or with components of the local tumor microenvironment resulted in alterations in DC viability.
Induction of Apoptosis in DCs Infiltrating 9L Tumors.
To determine whether the viability of potential APCs infiltrating 9L tumors was altered in the tumor microenvironment, we set up s.c. and i.c. 9L tumors. On day 12, tumors or brains bearing tumors were isolated after perfusion of rats. Frozen tumors were then sectioned and stained with Hoescht dye to indicate nuclei and with PE-conjugated OX6 (MHC class II) or OX62 (pan-DC marker; Ref. 56
) and FITC-VAD-FMK to indicate cells undergoing apoptosis. As illustrated in Fig. 2
, numerous APCs in i.c. 9L tumors were undergoing apoptosis as indicated by coexpression of OX6 and activated caspases (white arrows). To verify that DCs in 9L tumors were undergoing apoptosis, we carried out similar analyses using PE-OX62 and FITC-VAD-FMK. As illustrated in Table 1
, we determined that 21% of OX62+ DCs were apoptotic (n = 8 tumor samples; range, 4.5–41.6%).
View larger version (165K):
[in this window]
[in a new window]
|
Fig. 2. Induction of apoptosis in OX6+ cells infiltrating i.c. 9L tumors. Established (day 12) i.c. 9L tumors were sectioned and stained with Hoescht dye (blue), PE-conjugated OX6 (red), and FITC-VAD-FMK (green). OX6+ cells undergoing apoptosis are indicated by white arrows. Punctate staining of cells with FITC-VAD-FMK is illustrated in the inset (bottom right). The white bar corresponds to 100 µm in the tumor section and 20 µm in the inset.
|
|
HA Induces iNOS Expression in DCs.
NO is generated by the enzyme NOS, of which there are three related, but distinct forms (64)
. Types I and III NOS are constitutively expressed in cells of neural and endothelial origin, respectively (64)
. In contrast, the inducible type II NOS (iNOS) is expressed in cells with immunoregulatory function (64)
. Therefore, we investigated the ability of HA to regulate iNOS expression in rat BM-derived DCs. HA-induced iNOSmRNA expression was determined by RT-PCR (Fig. 3A
, and Northern blot analyses (Fig. 3B)
. To confirm production of iNOS, Western blot analyses were carried out. It was determined that HA-induced iNOS mRNA expression in DCs was accompanied by production of iNOS protein (Fig. 3C)
.
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 3. HA induces iNOS expression in DCs. A, DCs were cultured for 12 h with HA (100 or 10 µg/ml) or without HA (0 µg/ml), and total RNA was isolated using RNAzol. RT-PCR was performed using primers specific for iNOS or GAPDH. PCR products were visualized in ethidium bromide-containing gels after electrophoresis in agarose gels. B, RT-PCR products were used as probes for iNOS and GAPDH, using total RNA samples from treated and untreated DC cultures. C, Western blot analyses were performed on lysates of HA-treated and untreated DCs. Blots were probed with polyclonal antibody to iNOS. Consistency in protein loading was determined by probing blots with antitubulin.
|
|
HA Induces NO Production by DCs.
It is well established that HA is a major component of ECM in gliomas (53
, 54)
and is made by glioma cells (50, 51, 52
, 65) . HA has been reported to induce NO production in macrophages and in endothelial and Kupffer cells (66
, 67)
. Furthermore, it has been shown that macrophages and BM-derived DCs undergo apoptosis via induction of NO in culture (68, 69, 70, 71)
. We therefore investigated whether soluble and glioma-associated HA could induce rat BM-derived DCs to produce NO. In our initial experiments, DCs were exposed to various concentrations of soluble HA (Fig. 4a)
or to 100 µg/ml HA for various intervals of time (Fig. 4b)
. NO production was determined by measuring the accumulation of the stable end product nitrite in culture supernatants. As illustrated in Fig. 4a
, it was determined that HA induced NO release by cultured DCs in a dose-dependent manner over a range of 0.1–1000 µg/ml and that significantly increased levels of nitrite were detected when as little as 1 µg/ml HA was used compared with control cultures. Furthermore, HA-induced NO production was inhibited by addition of the specific NOS inhibitor NMMA (Fig. 4a)
. To determine the kinetics of NO release, nitrite levels were determined over a 4-day culture period after stimulation of DCs with HA. As shown in Fig. 4b
, NO biosynthesis was first evident at 24 h and reached the maximum level 48 h after stimulation.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 4. HA induces NO production by DCs. DCs were harvested after 7–8 days of culture, washed, and resuspended in CM containing the LPS inhibitor polymyxin B (10 µg/ml). A, 1 x 105 DCs/well in 0.2 ml were cultured in 96-well plates. Twenty µl of CM containing HA were added to each well to achieve final concentrations of 1000, 100, 10, 1, and 0.1 µg/ml. Twenty µl of CM were added to control wells. For inhibition, NMMA (1.0 µM) was added to wells containing 1000 µg/ml HA. All cells were then incubated for 48 h, and the supernatants were collected for nitrite determination by the Griess reaction. Data are the mean ± SD (bars; n = 3), and a dose-dependent increase nitrite levels was detected in cultures treated with HA. Statistically significant differences in NO production were apparent at 1000, 100, 10, and 1 µg/ml HA. Nitrite production was inhibited by the addition of NMMA (1 µM). B, 6 x 106 cells in 6 ml of CM were cultured in 6-well plates at a final concentration of 100 µg/ml HA. A 250-µl aliquot of each cell supernatants was collected every 24 h, and the nitrite concentration was determined by the Griess reaction. Significantly elevated nitrite production was detectable at 24, 48, 72, and 96 h, but peak production was achieved by 48 h. Data are the mean ± SD (bars; n = 3).
|
|
Enzymatic Digestion of HA Ablates NO Production by DCs.
It has been reported that different molecule-weight fragments of HA display differential capacities for induction of iNOS expression compared with intact, high molecular weight HA or completely digested products of HA (66)
. To determine whether partial digestion of HA altered the induction of NO production by DCs, we prepared HA fragments by partial digestion of high molecular weight HA with limited amounts of testicular HAse (1 unit/10 µg of HA) for various time intervals. The sizes of HA fragments were characterized by agarose gel electrophoresis as described previously (62
, 72)
. As shown in Fig. 5A
(Lanes 3–10), various HA fragments were produced by partial digestion of high molecular weight HA with low levels of HAse (1 unit/10 µg). The HA fragments generated using a large amount of HAse (500 units/10 µg) are shown in Fig. 5A
, Lanes 11 and 12. As illustrated in Fig. 5B
, NO induction by HA was abrogated by enzymatic digestion of HA, but only after digestion with 500 units/10 µg for 24 h. There was a trend toward greater NO production by DCs treated with partially digested HA, but these did not achieve statistical significance.
View larger version (55K):
[in this window]
[in a new window]
|
Fig. 5. Enzymatic digestion of HA ablates NO production in DCs. A, after cleavage of HA by HAse, fragments were separated by electrophoresis on a 0.5% agarose gel and visualized with Stains-All. Lanes 1 and 2, untreated HA; Lanes 3 and 4, 20-min digestion (1 unit of HAse/10 µg of HA); Lanes 5 and 6, 40-min digestion (1 unit of HAse/10 µg of HA); Lanes 7 and 8, 60-min digestion (1 unit of HAse/10 µg of HA); Lanes 9 and 10, 80-min digestion (1 unit of HAse/10 µg of HA); Lanes 11 and 12, 24-h digestion (500 units of HAse/10 µg of HA). B, concentrations of digested HA from the various enzyme concentrations and times of digestion, corresponding to 100 µg of intact HA, were used to stimulate DCs. Nitrite production by DCs was determined and compared with nitrite in untreated cultures of DCs. Data are the mean ± SD (bars; n = 3).
|
|
HA-induced Apoptosis in DCs.
As indicated above, it has been reported that macrophages and BM-derived DCs undergo apoptosis via induction of NO in culture when activated by IFN-
and LPS (68, 69, 70, 71)
. Because our data indicate that HA is an efficient NO inducer in DCs and that NO is correlated with induction of apoptosis in macrophages and DCs (68, 69, 70, 71)
, we examined the ability of HA to induce apoptosis in cultured rat DCs and whether this correlated with NO production. BM-derived DCs were cultured with HA (100 µg/ml) in CM supplemented with granulocyte macrophage colony-stimulating factor and interleukin-4 for 48 h and then analyzed for induction of apoptosis by TUNEL staining. As illustrated in Fig. 6A
, DCs stimulated with HA were determined to be undergoing apoptosis; whereas no evidence of apoptosis was found in the DCs cultured without HA (Fig. 6C)
. Quantitation of DCs undergoing apoptosis indicated that HA induced apoptosis in 30% of DCs under these experimental conditions (data not shown). Furthermore, addition of NMMA (1 µM) to cultures of DCs stimulated with HA (100 µg/ml) inhibited apoptosis (Fig. 6B)
, thereby confirming the correlation between iNOS and induction of DC apoptosis. We also assessed the induction of apoptosis in DCs by HA, using a flow cytometric assay with FITC-VAD-FMK. As illustrated in Fig. 7
, DCs treated with HA underwent apoptosis, and the percentage of apoptosis in DCs could be titrated with HA. The data shown are representative of two identical assays.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 6. HA-induced apoptosis in DCs. A, DCs cultured with HA (100 µg/ml) for 48 h were assayed by TUNEL staining for apoptotic cells, using in situ nick-labeling of DNA strand breaks (brown staining). B, addition of NMMA (1 µM) to cultures of DCs treated with HA (100 µg/ml) inhibited apoptosis. C, TUNEL staining of control DC cultures, incubated in the absence of HA (negative control). Magnification, x400.
|
|
View larger version (11K):
[in this window]
[in a new window]
|
Fig. 7. Induction of apoptosis in DCs by HA is concentration dependent. DCs (2 x 106 cells/ml) were cultured in the presence or absence of various concentrations of HA (1000, 100, 10, 1, or 0 µg/ml). Apoptosis was determined by flow cytometric analyses of cells stained with FITC-VAD-FMK.
|
|
9L Gliosarcoma Cells Induce DC Death via CD44-HA Interactions.
To verify the induction of apoptosis in DCs after interaction with tumor-associated HA, we coincubated 9L tumor cells with DCs and then carried out two-parameter flow cytometric analyses using PE-conjugated anti-MHC class II (OX6; FL2) and FITC-VAD-FMK (FL1), which binds activated caspases. DCs were segregated from tumor cells based on gating of MHC class II+ cells because the tumor cells do not express MHC class II (data not shown). As illustrated in Fig. 8A
, DCs cultured without 9L cells were not labeled by FITC-VAD-FMK, whereas 30% of DCs cocultured with 9L were positive for FITC-VAD-FMK staining. Pretreatment of DCs with anti-CD44 (10 µg/ml) before coculture with 9L resulted in a loss of FITC-VAD-FMK incorporation by DCs (Fig. 8B)
. We also pretreated DCs with isotype control antibody (IgG2a; 10 µg/ml); this antibody did not affect induction of apoptosis in DCs (data not shown). Similarly, pretreatment of 9L with HAse before culture with DCs resulted in a loss of induction of apoptosis in DCs (Fig. 8B)
. These data indicate that 9L cells induced apoptosis in DCs and that this induction was dependent on stimulation via CD44 interactions with HA.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 8. 9L gliosarcoma cells induce apoptosis in DCs via CD44-HA Interactions. A, DCs were cultured without (thin line) or with 9L tumor cells (thick line) and were stained with PE-OX6 (anti-MHC class II; FL2) and FITC-VAD-FMK (FL1). DCs were segregated from 9L cells by gating on MHC class II+ cells. B, DCs were cultured with 9L cells (thin solid line), pretreated with anti-CD44 and incubated with 9L cells (dashed line), or cultured with 9L cells pretreated with HAse (thick solid line) and then stained with PE-OX6 (anti-MHC class II; FL2) and FITC-VAD-FMK (FL1). DCs were segregated from 9L cells by gating on MHC class II+ cells.
|
|
|
DISCUSSION
|
|---|
It has been established that several inherent features of the brain and of the tumors themselves are critical determinants in limiting therapy for gliomas (73)
. One of the most interesting biological features of malignant gliomas is a striking state of immunosuppression (74)
. The generation of an effective antitumor response is a multistep process that requires presentation of tumor antigens by professional APCs, e.g., DCs, to CD4+ T-helper cells. Stimulated CD4+ T-helper cells then activate CD8+ cytotoxic T cells to recognize tumor antigen and to destroy tumor cells (75)
. Although the immunodeficient state of glioma patients can be attributed to several different factors, such as production of transforming growth factor-ß and expression of FasL (76)
, effective antitumor responses in glioma patients may also be absent because of a lack of, or reduced function of professional APCs in the brain (77)
. Interestingly, the presence of increased numbers of DCs within solid tumor masses has been correlated in some studies with improved prognosis (78)
. By extension, it has been demonstrated that intratumoral delivery of DCs has therapeutic benefit for various tumors (41, 42, 43, 44)
. In contrast, the data reported here demonstrate that local delivery of DCs into established, i.c. 9L tumors did not provide any survival advantage, even when tumor cells were induced to undergo apoptosis, which enhances DC loading for antigen presentation (Fig. 1)
. These results suggest that the interaction of DCs with tumor cells or with components of the local 9L tumor microenvironment resulted in a loss of DC function or viability.
There are three known isoforms of NOS. The constitutive isoforms in neurons (nNOS) and endothelium (eNOS) are Ca2+/calmodulin dependent and rapidly release small amounts of NO in response to increases in intracellular calcium. In contrast, iNOS, which is expressed in many cell types, is Ca2+/calmodulin independent and produces large and sustained quantities of NO in cells stimulated by bacterial endotoxins or by cytokines (79)
. NO is known to induce apoptosis in various cell types, including cardiomyocytes (80)
, mesangial cells (81)
, trophoblasts (82)
, tumor cells (83
, 84)
, and DCs (68, 69, 70, 71)
. Various stimuli have been reported to induce NO; and LPS and IFN- have been shown to lead to induction of apoptosis in mouse DCs via NO. As an extension of those studies and our data indicating a lack of therapeutic effects of intratumor delivery of DCs, we investigated whether 9L tumor cells or their products could induce apoptosis in DCs and whether this was a result of the activation of NO production via iNOS.
In the studies reported here, we demonstrated that rat BM-derived DCs express iNOS and that these DCs produce NO in response to tumor cells via tumor-derived HA. NO generation was induced after incubation of DCs with HA, a nonsulfated glycosaminoglycan of high molecular mass, which consists of repeated glucuronic acid-N-acetylglucosamine disaccharide units, is ubiquitous in the ECM of many tissues (85)
, and is a major component of the ECM of gliomas (50, 51, 52, 53, 54)
. Compared with the ECM in other tissues, mature brain has very low levels of collagen, fibronectin, and laminin, but has high levels of HA (86)
. HA is distributed mainly in white matter fiber tract, a frequent route for glioma invasion into the surrounding brain (54
, 87)
. Glioma cells also produce substantial amounts of HA (50, 51, 52, 53)
. By interacting with CD44, a receptor of HA, or other HA-binding proteins, HA plays an important role in numerous biological processes, such as cell migration, differentiation, and tumorigenesis (88)
. In fact, increased production of HA by tumor cells or tumor-associated fibroblasts has been correlated with tumor cell migration and metastasis (89, 90, 91, 92)
.
Although it has been reported that NOS is induced in murine macrophages after interaction with HA (69)
, there are no reports demonstrating NO production in DCs stimulated with tumor cells or tumor-derived components of ECM. In this report, we demonstrate that induction of NO synthesis by DCs in response to HA is dose dependent (Fig. 3)
. A time course study of NO release indicated that nitrite accumulated in culture supernatants over 48 h. Furthermore, nitrite production by DCs was significantly inhibited by the NOS inhibitor NMMA. We also verified that NO production by DCs was associated with iNOS activation, using both biochemical and molecular means: the induction of mRNA encoding iNOS was confirmed by RT-PCR and Northern blotting; and protein production was confirmed by Western blotting (Fig. 4)
. Partial degradation of HA resulted in increased induction of NO compared with undigested HA, but complete degradation of HA resulted in a loss of induction of NO (Fig. 5)
. Similar to these data, it has been demonstrated that HA fragments of intermediate molecular mass stimulated iNOS expression in murine macrophages; whereas intact, high molecular mass HA and HA disaccharides did not (66)
. As a functional consequence of iNOS production, we determined that a substantial number of DCs were induced to apoptose after interaction with HA (Figs. 6
and 7
). DC apoptosis was inhibited by NMMA, suggesting that the observed accumulation of NO in these cultures accounted for the activation of DC apoptosis. Coculture of DCs and 9L tumor cells resulted in the induction of apoptosis in DCs as indicated by the presence of activated caspases by two-parameter flow cytometric analyses (Fig. 8)
. Similar results were obtained in assays using DC interactions with RG2 and F98, indicating that the induction of apoptosis in DCs is generalizable at least to these three syngeneic tumor lines (data not shown). Pretreatment of DCs with anti-CD44 inhibited induction of apoptosis in DCs, and pretreatment of 9L cells with HAse partially blocked the induction of apoptosis in DCs. These data indicate that CD44-HA interactions induce apoptosis in DCs, but that other receptor-ligand interactions may also contribute. Given that 9L tumor cells express FasL (data not shown) and that gliomas are clearly capable of mediating lysis of immune cells via Fas-FasL interactions (10)
, it is likely that this also contributes to apoptosis in DCs. Because the only antibodies currently available for rat Fas and FasL bind cytoplasmic portions of the molecules, it is not yet possible to verify, by blocking studies, the contribution of such interactions to DC apoptosis in the rat model.
DCs are critical to the function of the immune system because they are the primary APCs for the initiation of T-lymphocyte responses (22, 23, 24, 25, 26, 27, 28)
. Our findings suggest a role for HA in the dysregulation of induction of immunity to gliomas by stimulating NO production in DCs, thus leading to their elimination. In fact, it has been reported that NO production can also cause apoptosis in microglial cells (93)
, which function as APCs in the central nervous system (94)
. Similarly, because production of NO by DCs has been reported to induce apoptosis in T lymphocytes (70)
, it is likely that gliomas can indirectly affect adaptive responses as well by eliminating effector T cells via NO produced by DCs. Interestingly, HAse has been used to therapeutic advantage in combination with standard therapy in pediatric glioblastomas (95, 96, 97)
. This clinical effect has been attributed to a role for HAse in disruption of the migration and infiltration of gliomas by digestion of HA. Data presented here suggest that treatment of gliomas with HAse may enhance immune reactivity by providing a protective effect against induction of apoptosis in immune cells mediated via CD44-HA interactions as well.
|
FOOTNOTES
|
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This project supported in part by Grants CA68550, CA68067, and NS01810 from the NIH, and by a grant from the James S. McDonnell Foundation. 
2 To whom requests for reprints should be addressed, at W948 Biomedical Science Tower, DeSoto at O’Hara Street, Pittsburgh, PA 15213. Phone: (412) 624-0370; E-mail: chamberswh{at}msx.upmc.edu
3 The abbreviations used are: APC, antigen-presenting cell; DC, dendritic cell; TRA, tumor rejection antigen; RS, radiosurgery; i.c., intracranial; HA, hyaluronan; ECM, extracellular matrix; iNOS, inducible nitric oxide synthase; CM, complete medium; BM, bone marrow; VAD-FMK, Val-Ala-Asp-fluoromethyl ketone; TUNEL, TdT-mediated dUTP-biotin nick-end labeling; PE, phycoerythrin; RT-PCR, reverse transcription-PCR; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HAse, hyaluronidase; NMMA, N-monomethyl-L-arginine; LPS, lipopolysaccharide.
4 T. F. Witham and D. Kondziolka, unpublished data.
Received 7/20/01.
Accepted 3/ 1/02.
|
REFERENCES
|
|---|
Top
ABSTRACT
INTRODUCTION
) or untreated (9L; 
) was compared. A similar comparison was made in rats bearing day 7 i.c. 9L tumors and treated with gamma knife RS (9L + Gamma; 
) or gamma knife RS and DCs (9L + Gamma/DC; 
). All rats in the 9L (17 of 17) and 9L + DC (20 of 20) groups were dead or sacrificed by day 30. In contrast, 4 of 16 rats in the 9L + Gamma group survived longer than 150 days. Similarly, 3 of 16 rats in the 9L + Gamma/DC group survived longer than 150 days.