Published online before print June 20, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Institute for Medical Immunology, Université Libre de Bruxelles, Gosselies, Belgium;
Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands; and
Laboratory of Molecular and Cellular Therapy, Department of Physiology and Immunology, Medical School of Vrije Universiteit Brussel, Brussels, Belgium
1 Correspondence: Institute for Medical Immunology, Université Libre de Bruxelles, Rue Adrienne Bolland 8, 6041 Gosselies, Belgium. E-mail: vflamand{at}ulb.ac.be
ABSTRACT
Overexpression of CD95 (Fas/Apo-1) ligand (CD95L) has been shown to induce T cell tolerance but also, neutrophilic inflammation and rejection of allogeneic tissue. We explored the capacity of dendritic cells (DCs) genetically engineered to overexpress CD95L to induce an antitumor response. We first found that DCs overexpressing CD95L, in addition to MHC class I-restricted OVA peptides (CD95L-OVA-DCs), induced increased antigen-specific CD8+ T cell responses as compared with DCs overexpressing OVA peptides alone. The enhanced T cell responses were associated with improved regression of a tumor expressing OVA, allowing survival of all animals. When DCs overexpressing CD95L (CD95L-DCs) were injected with the tumor expressing OVA, in vivo tumor proliferation was strikingly inhibited. A strong cellular apoptosis and a massive neutrophilic infiltrate developed in this setting. Neutrophil depletion prevented tumor regression as well as enhanced IFN-
production induced by CD95L-OVA-DCs. Furthermore, the CD8+ T cell response induced by the coadministration of tumor cells and CD95L-DCs led to rejection of a tumor implanted at a distance from the DC injection site. In summary, DCs expressing CD95L promote tumor rejection involving neutrophil-mediated innate immunity and CD8+ T cell-dependent adaptative immune responses.
Key Words: murine model • immunotherapy • CD95 ligand
INTRODUCTION
Severe lymphoproliferative disorders and autoimmunity observed in mouse strains lacking CD95 (lpr) or expressing a defective CD95 ligand (CD95L) molecule (gld) have highlighted the critical role of CD95/CD95L interaction in the activation-induced cell death homeostatic mechanism [1 ]. CD95L causes apoptosis of CD95+ T cells and is constitutively expressed in immunologically privileged sites including the testis and the eye [2 ]. CD95L expression in tumors may contribute to a tumor immune escape mechanism by inducing infiltrating T cell apoptosis [3 ]. Immunosuppressive properties of CD95L have therefore led to the use of several CD95/CD95L expression systems to induce transplantation tolerance and autoimmunity prevention. A major controversy about the effective use of CD95L as an immunomodulator in vivo emerged in the late 1990s and is still active today [4 ]. Despite encouraging primary results [5 ], it appears that overexpression of CD95L by transplants or tumors promotes neutrophilic inflammation, ultimately leading to tissue or graft destruction [6 , 7 ].
The role of inflammatory cells in tumor progression is highly controversial. Polymorphonuclear leukocytes (PMN) can trigger tumor progression by degrading the extracellular matrix and enhancing angiogenesis [8
]. Activation of circulating granulocytes was reported in advanced cancer patients and was correlated with the inhibition of cytokine production and TCR
-chain expression by T cells through granulocyte-derived H2O2 [9
]. Conversely, PMN can act as direct effector cells in the immune surveillance against cancer by releasing cytotoxic mediators such as oxygen radicals and proteolytic enzymes or through antibody-dependent, cell-mediated cytotoxicity [10
, 11
]. In various studies, neutrophils were made primarily responsible for the rejection of CD95L+ tumors [12
]. An indirect, CD95L-dependent mechanism for the recruitment of neutrophils has also been demonstrated. Indeed, a CD95 ligation on surrounding cells such as dendritic cells (DCs) may cause the secretion of proinflammatory cytokines (such as IL-1β, TNF-
) and chemokines (such as IL-8, keratinocyte-derived chemokine, MCP-2, MIP-2), resulting in enhanced chemoattraction of neutrophils [13
, 14
]. Kidoya et al. [15
] have recently shown that CD95 ligation induced IL-23 production in DCs, which resulted in IL-17-mediated neutrophil infiltration.
In a previous study, we observed that mice injected with allogeneic CD95L overexpressing DCs elicit a strong, neutrophil-dependent, Th1-type response sufficient to prime for acute skin allograft rejection [16 ]. Similarly, neutrophil presence in IL-6-, lymphotactin-, and CD95L-expressing tumors has also been associated with a type 1 cytokine response and CD8+ T cell-mediated, protective immunity against subsequent challenge with parental tumor cells [17 18 19 ].
DCs represent potent vectors for tumor immunotherapy because of their unique properties of antigen capture and (cross-) presenting capacity, migration, and naïve T cell priming. Herein, we hypothesized that CD95L expressed in a DC vaccine could reinforce an antitumoral Th1 and CTL response and could lead to long-term tumor protection. To address this question, CD95L was overexpressed by DCs in addition to MHC I OVA peptides and used as a vaccine or overexpressed by DCs that were administered separately from the EG7-OVA thymoma cells. We found that CD95L has an adjuvant effect on Th1 and a CTL response and that they contribute to the establishment of a protective immune response against tumors. This observation led us to investigate the role of neutrophils as well as other mechanisms implicated in the CD95L-triggered, antitumoral T cell response.
MATERIALS AND METHODS
Mice, cell lines, and reagents
C57BL/6 (H-2b) breeders were purchased from Harlan Nederland (Horst, The Netherlands). C57BL/6-lpr/lpr Fas-deficient (lpr/lpr) mice were originally obtained from The Jackson Laboratory (Bar Harbor, ME, USA) and bred in our animal facility. Animals were maintained and treated according to institutional guidelines. The tumor cell lines used were EL-4 (C57BL/6, H-2b, thymoma) and EG7-OVA (EL-4 cells transfected with chicken albumin cDNA, American Type Culture Collection, Manassas, VA, USA). The T cell hybridoma RF33.70, recognizing a H-2-Kb OVA peptide (amino acids 257–264), the murine (m)CD95-transfected P815 target cells, and the PhoenixECO ecotropic packaging cell line were cultured as described previously [16
, 20
]. The recombinant (r)mGM-CSF used for the DC generation was produced as described previously [21
]. The RB6-8C5 cell line was used to produce anti-Gr-1 rat IgG2b mAb ascitic preparation and was kindly provided by Dr. Robert Coffman (Schering-Plough Biopharma, Palo Alto, CA, USA). The endotoxin level of ascites was evaluated as <5 endotoxin units/ml, as determined by the Coatest Limulus amoebocyte lysate assay (Lonza, Switzerland). The anti-DNP rat IgG2a mAb LO-DNP-16 was purchased from LO-IMEX (Université Catholique de Louvain, Belgium). CFSE was diluted for use to a final concentration of 10 mM in DMSO (Molecular Probes, Eugene, OR, USA).
Generation and transduction of DCs with recombinant retroviruses
DCs were propagated from bone marrow progenitors in GM-CSF-containing medium and harvested after 10 days as described previously [16
]. For retrovirus production, the retroviral vector MFG, derived from the Moloney murine leukemia virus, was used. This vector does not contain a drug-resistance marker nor does it express any potential antigenic protein other than the inserted cDNA. All of the cDNAs were obtained by PCR. The amplification products were sequenced before insertion into the MFG vector. For amplification of OVA cDNA, we used the pAc-neo-OVA plasmid (provided by Dr. Michael Bevan, University of Washington, Seattle, Washington, USA) as a template. Retroviral vectors pMFG-mock, pMFG-CD95L, and pMFG-OVA, containing the cDNA encoding a truncated form of the OVA protein sequence (amino acids 40–386), were generated as described previously [16
, 20
]. Recombinant retroviruses’ supernatants were produced by PhoenixECO packaging cell line transfection with pMFG-mock, pMFG-CD95L, or pMFG-OVA following the method described previously [16
]. In that previous study [16
], we made the observation that suicidal or fratricidal death occurred when DCs were generated from C57BL/6 wild-type bone marrow progenitors and transduced with CD95L retrovirus. This CD95L transduction resulted in massive cell death, as more than 90% of bone marrow cells were annexin V- and propidium iodide-positive 48 h post-transduction. Approximately 85% viability is retained when CD95L transduction is performed in lpr/lpr bone marrow progenitors. The transduction of lpr/lpr progenitors was realized on Days 1–3 after the start of the bone marrow cell culture. The medium was removed and replaced with 2 ml viral supernatant containing 8 µg/ml polybrene (Sigma-Aldrich, St. Louis, MO, USA). The cells were transduced by centrifugation of the six-well plates for 2 h at 2400 rpm at room temperature. To overexpress OVA as well as CD95L in DCs, progenitors were cotransduced with 2 ml CD95L and 2 ml OVA rMFG retroviruses. The transduced DCs were used on Day 10. When OVA-MFG virus was used for progenitor transduction, the medium was replaced on Day 8 by the X-VIVO 20 free-serum medium (Lonza). The transduction of lpr/lpr bone marrow progenitors with mock-, CD95L-, and/or OVA-MFG retrovirus did not modify DC phenotypes. At the end of the culture in GM-CSF, approximately 85% of the cells were CD11c+GR1low DCs with immature phenotype as indicated by low expression of MHC class II, CD80, CD86, and CD40.
The H-2-Kb OVA peptide presentation by OVA-MFG-transduced DCs was assessed by the production of IL-2 by H-2-Kb-OVA (257–264)-restricted T hybridoma cell lines RF33.70 during a coculture. The amount of IL-2 released in the supernatant was monitored by ELISA (Duoset, R&D Systems, Minneapolis, MN, USA). CD95L-MFG retroviral transduction efficiency was evaluated by an in vitro apoptosis assay with CD95+ target cells in the presence of CD95L ± OVA-MFG-transduced DCs compared with mock-MFG-transduced DCs.
In vivo CTL assay
C57BL/6 mice received one s.c. injection into the footpad with 1 x 105 transduced DCs. Seven to 10 days after immunization, mice were i.v.-injected with 107 OVA-pulsed CFSE+ (10 µM) and 107 unpulsed CFSE+ (1 µM) spleen cells. The CFSE labeling of spleen cells took place for 10 min at 37°C. Labeling was blocked by adding DMEM medium with 5% FCS. Spleen cells were then washed two times in PBS. The antigen loading of CFSE+ (10 µM) spleen cells was performed using the class I-restricted SIINFEKL OVA peptide (257–264, NeoMPS). Spleen cells diluted to 107 cells/ml in HBSS were incubated for 2 h at 37°C, 5% CO2, with 5 µl OVA peptide (10 mM) to reach a final peptide concentration of 5 µM. After pulsing, spleen cells were washed twice and resuspended in HBSS for i.v. injection. Twenty-four hours after i.v. immunization, mice were bled, circulating cells were collected from the blood, and the percentage of OVA-specific CTL activity was measured by the relative quantification of the two CFSE+ populations by flow cytometry.
CD95-mediated apoptosis assay
EG7-OVA or mCD95-transfected P815 target cells were submitted to an apoptosis assay as described previously [16
]. Briefly, target cells were labeled with 5 µCi/ml 3H-thymidine (ICN, Asse-Relegem, Belgium) during an overnight incubation at 37°C and 5% CO2. Labeled target cells were seeded at a density of 104 cells/well. Effector lpr/lpr control (Ctrl)- or CD95L-DCs were added to the target at the indicated ratios. After 18 h of incubation at 37°C and 5% CO2, radioactivity was measured, and the data were expressed as percentages of cytotoxicity calculated by the following formula: [1–(cpm with effector/cpm without effector)] x 100.
Coinjection and challenge protocol
C57BL/6 mice were s.c.-coinjected in the flank with 50 µl PBS containing 6 x 105 Ctrl-DC or CD95L-DCs and 1.5 x 105 EG7-OVA tumor cells. Tumor growth was measured for 4 weeks. Mice were monitored on a regular basis for mice survival and tumor size (smallest diameterxlargest diameterxheight). The role of neutrophils in CD95L-DC-coinjected mice was tested by i.p. and s.c. injection of anti-Gr-1-depleting antibodies (RB6-8C5) 4 h before coinjection and on Days 4 and 11. Neutrophil immunostaining using anti-CD11b and anti-Gr1-specific mAb in the peritoneal lavage from wild-type mice or Gr1-depleted mice immunized i.p. with Ctrl-DCs or CD95L-DCs revealed more than 90% of neutrophil (CD11b+Gr1high) depletion in anti-Gr1 mAb-treated mice.
The generation of a tumor-specific, protective response in coinjected mice was assessed by s.c. injection with 1.5 x 105 EL-4 or EG7-OVA tumor cells in the contralateral flank. To demonstrate the involvement of CD4+ or CD8+ T lymphocytes in the protective response, CD4+ T cells were depleted after challenge by repeated i.p. injections of a tested, depleting amount of ascitic preparations of anti-CD4 mAb (clone GK1.5) or anti-CD8 mAb (clone H35) 11, 15, 19, and 24 days after challenge.
Tumor vaccine
C57BL/6 mice were s.c.-immunized in the footpad with 2 x 104-transduced DCs on Day 0. Ten days later, 2 x 105 EG7-OVA were s.c.-injected into the flank of the mice. Four days after tumor implantation, a second immunization with 3 x 104-transduced DCs occurred. Mouse survival was monitored frequently for 3 months.
In vitro cytokine production
Seven to 10 days after immunization with 105-transduced DCs, 2 x 106 cells isolated from draining lymph nodes (LNs) of C57BL/6 mice were restimulated in vitro with 4 x 105 -irradiated EL-4 or EG7-OVA tumor cells. Supernatants were harvested after 72 h of culture for determination of IFN- levels. Quantification of cytokines was performed using a commercially available ELISA (Duoset, R&D Systems).
Histology
Tumor-implanted skin sites were harvested 4 days after injection. Skin from the injection site was fixed in 10% neutral-buffered formalin and embedded in paraffin for immunostaining or for staining with H&E for histological analysis. To detect apoptotic cells, deparaffinized sections were boiled two times for 5 min each in citrate buffer (pH 6.0). Endogenous peroxidase activity was quenched with a solution of methanol/0.03% H2O2 (Merck, Darmstadt, Germany), and nonspecific binding was blocked by 10% normal goat serum (Dako, Glostrup, Denmark). The sections were incubated with a rabbit anti-human, active caspase 3 polyclonal antibody (Cell Signaling, Beverly MA, USA), followed by further incubation with poly-HRP goat anti-rabbit IgG (Powervision, Immunovision Technologies, Duiven, the Netherlands) and developed using a solution of 1% H2O2 and 3.3'-diaminobenzidin-tetra-hydrochloride (Sigma-Aldrich) in Tris-HCl. The sections were mounted in glycerin gelatin after rapid counter-staining with methyl green. To stain granulocytes, the same procedure was followed. Slides were incubated with anti-Ly6-G mAb (BD PharMingen, San Jose, CA, USA).
Statistical analysis
Statistical analysis was performed using the two-tailed Mann-Whitney nonparametric test and when indicated, the two-tailed Student’s t-test. Mouse survival curves were compared using the log-rank test.
RESULTS
Immunization with DCs expressing OVA peptides in the context of MHC class I molecules in addition to CD95L enhances T cell responses and improves protection against OVA+ tumor
We have shown previously that administration of allogeneic lpr/lpr CD95L-transduced DCs significantly increases the allospecific Th1 response and induces allograft rejection [16
]. We wanted to evaluate the ability of DCs that overexpress CD95L to improve a T cell-mediated, protective immunity against tumor using the EG7-OVA cell line model. DCs retrovirally transduced with a cytosolic form of OVA were described previously to present OVA peptides only in the context of MHC class I and to be poor stimulators of an OVA-specific CTL response [20
]. We first decided to test the impact of CD95L expression in these MHC class I-restricted OVA peptides, presenting DC on the antigen-specific T cell response. As described already [16
, 20
], DCs were generated from lpr/lpr bone marrow progenitors during a 10-day culture in the presence of GM-CSF and were submitted to transduction with mFasL.2-MFG (CD95L-MFG) retrovirus and/or OVA-MFG retrovirus containing the cDNA encoding a truncated form of the OVA protein sequence (amino acids 40–386). The phenotype of DCs was evaluated after transduction at the end of the culture. Between 80% and 85% of the cells were CD11c+ GR1low DCs and displayed an immature phenotype, as assessed by low expression of MHC class II, CD80, CD86, and CD40 molecules (not shown). Presentation of OVA-derived peptides in the context of MHC class I complexes by OVA-MFG-transduced DCs was confirmed by the exclusive activation of OVA-specific T cell hybridoma RF33.70 recognizing an H-2Kb OVA peptide (amino acids 257–264; not shown). CD95L expression by CD95L-MFG-transduced DC was confirmed by an apoptosis assay on CD95-transfected target cells, as illustrated by a representative experiment (see below).
C57BL/6 mice were immunized with OVA-transduced DCs (OVA-DCs), CD95L-transduced DCs (CD95L-DCs), or OVA-DCs cotransduced with CD95L (CD95L-OVA-DCs). As seen in Figure 1A
, OVA-DCs and CD95L-DCs failed to elicit IFN- production in response to OVA-positive tumor cells (EG7-OVA). In this setting, the injection of CD95L-OVA-DCs strongly enhanced the production of IFN- by OVA-specific T cells. This higher IFN- production remained specific, as no significant cytokine production was detected against EL-4 when mice were immunized with CD95L-OVA-DCs or OVA-DCs (Fig. 1A)
. An in vivo CTL assay revealed that mice immunized with OVA-DCs were able to kill OVA-expressing target cells compared with mice immunized with CD95L-DCs. This OVA-specific CTL response increased slightly in mice injected with CD95L-OVA-DCs (Fig. 1B)
. To assess whether vaccination with CD95L-OVA-DCs could lead to a better antitumor protection, mice were immunized with 2 x 104 CD95L-, OVA-, or CD95L-OVA-DCs 10 days before s.c. injection of 1.5 x 105 EG7-OVA tumor cells. As expected, administration of CD95L-DCs did not prevent tumor growth, and mice died within 35–55 days (Fig. 1C)
with the same kinetics as naive mice inoculated with the same amount of tumor cells (not shown). OVA-DC inoculation improved mice survival significantly, but only mice immunized with CD95L-OVA-DCs were protected completely against tumor growth (Fig. 1C)
.
 View larger version (12K): [in a new window] |
Figure 1. Overexpression of CD95L by OVA-DCs induces a strong Th1 response and improves CTL response and protection against tumor growth. (A) Popliteal and inguinal LN cells isolated from nontreated (NT) C57BL/6 mice or C57BL/6 mice immunized with 1 x 105 lpr/lpr CD95L-DCs, OVA-DCs, or CD95L-OVA-DCs 7–10 days before were stimulated or not (NS) in vitro with -irradiated EL-4 or EG7-OVA cells. Supernatants were collected 72 h later, and concentrations of IFN- were assessed by ELISA. Results were expressed as mean ± SEM of four mice per group (*, P<0.03). (B) In vivo OVA-specific CTL response was performed on C57BL/6 mice 7–10 days after footpad s.c. immunization with 1 x 105 lpr/lpr CD95L-DCs, OVA-DCs, or CD95L-OVA-DCs. Results are mean percentage of OVA-specific lysis ± SEM of eight to 10 mice per group (*, P<0.005, as compared with lpr/lpr OVA-DC immunization or with lpr/lpr CD95L-DC immunization; **, P<0.005, as compared with the lpr/lpr CD95L-DC-injected group). (C) C57BL/6 mice were immunized in the footpads with 2 x 104 lpr/lpr CD95L (n=9)-, OVA (n=11)-, or OVA-CD95L-DCs (n=10) 10 days before s.c. injection of 1.5 x 105 EG7-OVA tumor cells in the flank of the mice (Day 0). A footpad recall was performed 4 days later with 3 x 104 lpr/lpr DCs from each type. Mice survival was monitored daily (*, P<0.005, as compared with lpr/lpr CD95L-DC-immunized mice; **, P<0.01, as compared with lpr/lpr OVA-DC-immunized mice, and P<0.005 as compared with lpr/lpr CD95L-DC-immunized mice).
|
Coinjection of CD95L-DCs and EG7-OVA cells induced neutrophil-dependent tumor growth inhibition
As CD95L expression on DCs improves the protective immune response induced by DCs that present MHC class I-restricted tumor peptides, we wanted to evaluate the ability of CD95L-expressing DCs to cross-present tumor-derived peptides in vivo. Therefore, we performed intradermal coadministration of CD95L-DCs and EG7-OVA tumor cells and monitored the impact on tumor growth. In contrast with Ctrl-DCs/EG7-OVA-coinjected mice, no tumor progression was observed in mice 30 days after the coadministration of CD95L-DCs and tumor cells (Fig. 2A
). As EG7-OVA expresses detectable amounts of CD95 (not shown), we checked the ability of CD95L-DCs to directly kill the tumor in vitro. Compared with CD95-transfected P815 cells used as positive control targets, a weak lytic activity could be observed against EG7-OVA at different ratios with CD95L-DCs (Fig. 2B)
. We therefore excluded a predominant, direct role of CD95L-DC on the mortality of tumor cells in vivo. Histologic examination of tumor-implanted skin sites 4 days after the coinjection revealed a complete destruction of the tumor in CD95L-DCs from coinjected mice (Fig. 3B
) compared with Ctrl-DCs from coinjected mice (Fig. 3A)
. Anticaspase-3 staining revealed a strong apoptotic activity only in the dermis of mice receiving CD95L-DCs/EG7-OVA injection (Fig. 3D)
compared with Ctrl-DCs/EG7-OVA-coinjected mice (Fig. 3C)
. Furthermore, anti-Ly-6G staining demonstrated a dense infiltration of granulocytes (Fig. 3F
compared with Fig. 3E
). To investigate the role of neutrophils in the inhibition of tumor growth after CD95L-DC coinjection, we performed depletion of neutrophils by administration of anti-Gr-1 mAb. Interestingly, we observed that tumor growth was restored in CD95L-DCs/EG7-OVA-coinjected mice that were submitted to anti-Gr-1 mAb treatment, suggesting a predominant role of neutrophils in tumor destruction (Fig. 4A
). As bone marrow-derived DCs express low levels of the Gr-1 marker at their surface (data not shown), it was important to exclude that the amount of anti-Gr-1 mAb used for the neutrophil depletion affected the viability of the inoculated DCs. For this purpose, we labeled DCs with CFSE and injected them i.p. 3 h after the anti-Gr1 mAb treatment. The results presented in Figure 4B
demonstrate that the percentage of CFSE+ DCs detected in the peritoneal lavage 16 h after their inoculation was not affected by the anti-Gr-1 mAb treatment compared with the isotype control mAb treatment.
 View larger version (20K): [in a new window] |
Figure 2. CD95L-DCs inhibit EG7-OVA tumor growth in vivo but not in vitro. (A) C57BL/6 mice were s.c.-coinjected with 6 x 105 lpr/lpr Ctrl- or CD95L-DCs and 1.5 x 105 EG7-OVA tumor cells on Day 0. The tumor growth was measured for 28 days. Results were expressed as mean ± SD of 10 mice per group (*, P<0.002, as compared with lpr/lpr Ctrl-DC-coinjected mice). (B) EG7-OVA or mCD95-transfected P815 cells were incubated with lpr/lpr CD95L-DCs (•) or Ctrl-DCs ( ) at several ratios. Percentages of lysis are representative of four experiments.
|
 View larger version (159K): [in a new window] |
Figure 3. Tumor granulocytic infiltration and apoptosis in mice coinjected with CD95L-DCs and EG7-OVA tumor cells. Representative H&E stainings (original magnification, x4) of tumor-implanted skin sites with 6 x 105 lpr/lpr Ctrl (A)- or CD95L-DCs (B), showing a clear tumor progression in A compared with B. Representative immunostainings for active caspase-3 (original magnification, x2; C and D) and Ly6-G (original magnification, x20; E and F) of tumor-implanted skin sites with 6 x 105 lpr/lpr Ctrl (C and E) or CD95L-DCs (D and F) showing a strong, apoptotic activity in D compared with C as well as a dense, granulocytic infiltration in F compared with E.
|
 View larger version (23K): [in a new window] |
Figure 4. Neutrophil depletion facilitates the EG7-OVA tumor growth in CD95L-DC-coinjected mice and reduces the Th1 response induced by OVA-CD95L-DCs. (A) C57BL/6 mice were s.c.-coinjected with 6 x 105 lpr/lpr CD95L-DCs and 1.5 x 105 EG7-OVA cells or EG7-OVA tumor cells alone. Depletion of neutrophils was realized by i.p. and s.c. injection of 600 µg anti ()-Gr-1 antibodies (RB6-8C5 clone) or Ctrl isotype antibodies (LODNP clone) 4 h before coinjection and at Days 4 and 11. The tumor growth of coinjected mice was measured at Days 15, 22, 28, 32, 36, and 41. Results were expressed as mean ± SD of four to eight mice per group [*, P<0.02, as compared with the anti-Gr-1-treated lpr/lpr CD95L-DC-coinjected group, and P<0.005 as compared with EG7-OVA-inoculated mice; **, not significant compared with EG7-OVA-inoculated mice; symbols ( ,  , ) correspond to individual experiment points; , individual dead mouse]. (B) Anti-Gr-1 antibody treatment does not deplete DCs in vivo. Two mice per group were i.p.-injected with 5 x 106 CFSE+ lpr/lpr Ctrl-DCs and treated i.p. with anti-Gr-1 or Ctrl isotype antibodies (1 mg/mouse) 3 h before the injection. Sixteen hours later, peritoneal exudate cells were analyzed for CFSE+ cell quantification by flow cytometry. Experiments were repeated at least three times with similar results. FL1-H, Fluorescence 1-height. (C) C57BL/6 mice received 3 x 105 lpr/lpr (continued) Ctrl-DCs, OVA-DCs, or OVA-CD95L-DCs in the footpad. RB6-8C5 (600 µg; anti-GR1 Ab) or control isotype mAb were i.p.-inoculated 3 h before and 3 days after the immunization. Popliteal and inguinal LN cells were restimulated in vitro with -irradiated EL-4 or EG7-OVA cells. Supernatants were collected after 72 h, and concentrations of IFN- were assessed by ELISA. Results were expressed as mean ± SD of four mice per group (*, P<0.03, as compared with all experimental groups).
|
elicited by EG7-OVA in culture with draining LN cells isolated only from mice immunized with CD95L-OVA-DCs is strongly reduced when mice undergo neutrophil depletion. Taken together, these results demonstrated that local neutrophil recruitment induced by DC overexpressing CD95L may contribute to tumor destruction and may amplify tumor antigen-specific T cell response.
Coinjection of CD95L DCs and EG7-OVA cells induces a protective CD8+ T cell-mediated response
To determine whether inhibition of tumor growth in CD95L-DCs/EG7-OVA-coinjected mice was associated with the development of a specific antitumoral-protective immune response, mice were s.c.-challenged in the contralateral flank with EG7-OVA or EL-4 tumor cells 20 days after the first injection. We observed that mice, in which primary rejection of EG7-OVA occurred through the first coadministration with CD95L-DCs, were protected only against EG7-OVA and not EL-4 tumor growth, indicating the development of an OVA-specific, protective response in those mice (Fig. 5
). We further demonstrated that this protective immune response is mostly mediated by CD8+ T cells, as depletion of those cells with anti-CD8 mAb during the EG7-OVA challenge favored the secondarily implanted tumor progression (Fig. 5)
. This CD8+ T cell-mediated, protective response does not seem to depend on CD4+ T cells, as anti-CD4 mAb treatment alone or in CD8-depleted, EG7-OVA-challenged mice did not modify the outcome of tumor growth (Fig. 5)
.
 View larger version (20K): [in a new window] |
Figure 5. OVA-specific, CD8+ T cell-protective response in mice coinjected with CD95L-DCs and EG7-OVA. C57BL/6 mice were s.c.-coinjected with 6 x 105 lpr/lpr CD95L-DCs and 1.5 x 105 EG7-OVA tumor cells and then challenged with 1.5 x 105 EL-4 (n=7) or 1.5 x 105 EG7-OVA 20 days after the first coinjection. Depletion of particular T cell subsets was realized after i.p. injection of anti-CD4 mAb or anti-CD8 mAb on Days 11, 15, 19, and 24 after challenge, and the tumor growth was monitored at the same time-points. Results were expressed as mean ± SD of four to seven mice per group [*, P<0.03, as compared with the EL-4-challenged group, and P<0.02 as compared with the anti-CD8 mAb-treated (n=5) and anti-CD4+anti-CD8 mAb-treated (n=6), challenged group; **, P<0.01, anti-CD4 mAb-treated mice (n=4) compared with the EL-4-challenged group; ***, not significant as compared with the EL-4-challenged group at all time-points; symbols (•) correspond to individual experimental points; , individual dead mouse].
|
It has long been recognized that the ultimate goal of tumor immunotherapy, which is to induce tumor-specific, T cell-mediated immunity that controls tumor growth in cancer patients, is not achieved when there is low tumor immunogenicity and tumor-induced immunosuppression. Among the strategies designed to activate the helper function at the tumor site, neutrophil-induced inflammation based on the CD95-CD95L interaction was evaluated in this study. It is well recognized that CD95L is a potent chemoattractant of neutrophils in vitro and in vivo. Indeed, CD95 ligation on neutrophils results in their attraction and activation [22 , 23 ]. Several studies have reported that CD95L overexpression by tumors is associated with their in vivo eradication involving neutrophils, which was followed by a T cell-dependent, tumor-specific, protective immunity [17 18 19 ]. The present study was conducted to test antitumor vaccination based on DCs coexpressing OVA tumor-associated antigens and CD95L as an adjuvant molecule. We demonstrated that CD95L-OVA-DCs, compared with OVA-DCs, have a significantly higher capacity to elicit a Th1 and CTL response in vivo against OVA-expressing tumor cell lines and to protect against their development in vivo. This adjuvant effect was highlighted with a DC vaccine that is unable to elicit an efficient CD8+ T cell-mediated, protective antitumor response because of a lack of CD4+ T cell activation as a result of OVA peptides that presented only in the context of MHC class I molecules [20 ]. CD95L-mediated signaling by DCs was therefore able to replace the CD4+ T cell help requirement in this setting. In another set of experiments allowing the activation of the CD4+ T cell, CD95L-OVA-DCs did not induce a better antitumor protection. Experiments were performed with a retrovirus encoding for OVA and the invariant chain (invchova) allowing OVA peptide presentation in the class I and class II MHC complexes (data not shown). Mice immunized with CD95L-invchOVA-DC generated a higher OVA-specific CTL response compared with invchOVA-DC but did not significantly improve the prevention of tumor growth, suggesting that CD95L-expressing DC are particularly effective for inducing tumor protection when CD8+ T cells do not benefit from CD4+ T cell help.
We further highlighted the capacity of CD95L-DCs to induce a protective tumor antigen-specific T cell response in mice implanted with the tumors. We clearly demonstrated that CD95L expression on DCs made them able to recruit neutrophils at the tumor site in which strong cellular apoptosis is taking place, and no tumor progression occurred. We identified neutrophils as the cells able to participate in tumor eradication and to boost the antitumoral-adaptive immune response in this setting. Neutropenic mice (i.e., GR1high cell-depleted mice), indeed, failed to reject the tumor and to prime the increased CD95L-OVA-DC-dependent, specific Th1 response.
The manner in which tumor antigens are presented to the immune system is a long-standing matter of debate. Whether antigens are presented directly by tumor cells in tumor-draining LNs or are cross-presented by host APC through the class I MHC pathway [24 ], they are considered insufficient to elicit a protective antitumor immunity [25 , 26 ]. DCs are able to internalize and process dying as well as apoptotic cells in vivo. Under steady-state conditions, this cross-presentation would lead to cross-tolerance [27 , 28 ], but inflammatory signals may switch the outcome of cross-presentation to cross-priming [29 ]. A previous study reported that DCs cocultured with CD95L-expressing tumors were able to elicit a tumor-specific immune response in vivo [30 ]. In this setting, the uptake of tumor antigens by DCs during a CD95/CD95L-mediated DC-tumor contact was suggested. The authors proposed to use CD95L-expressing tumor cells to generate in vitro tumor-specific DC vaccines but could not demonstrate clear, CD95-mediated DC maturation in terms of costimulatory molecule up-regulation. The approach used here, based on vaccination with host-derived CD95L-DCs, allows in vivo tumor-antigen uptake and presentation without ex vivo manipulation of a panel of tumor cell lines. This alternative approach for cancer immunotherapy based on administration of CD95L-DCs would better induce DC maturation and/or cross-presentation of tumor-derived, MHC class I-restricted peptides to CD8+ T cells as a result of a neutrophil-dependent, inflammatory response, as we demonstrated that in vivo tumor apoptosis caused by CD95L-DCs depends on neutrophil recruitment. Although clinical translation of this strategy is hypothetical, it could be facilitated by the fact that monocyte-derived DCs used in human cancer therapy trials are resistant to CD95L-mediated apoptosis [31 ].
Several reports supported the benefits of tumor apoptosis in the induction of a protective immune response. Indeed, encouraging clinical studies revealed that a multidrug regimen inducing a high level of tumor cell apoptosis promotes a significantly higher specific CTL response than tumor patients without treatment [32 ]. Likewise, particular DC vaccines loaded with killed, allogeneic tumor cells have been used in patients with metastatic melanoma. This vaccine induced melanoma antigen recognized by T cell 1-specific CD8+ T cell immunity, in which cross-presentation of tumor antigens has been strongly suggested [33 ]. The role of neutrophils in these settings was not evaluated. In our own experiments, the rapid tumoricidal action, which occurs in association with neutrophil infiltration within 96 h in naïve mice (Fig. 3) , strongly suggests that a neutrophil-dependent, innate-immune response is responsible for tumor apoptosis.
Various reports have supported a role for neutrophils in Th1 polarization. As it is still unclear how neutrophils might transmit these Th1-inducing signals to naïve T cells in the LNs, it was suggested that the role of neutrophils in Th1 polarization is indirect and requires cellular communication of neutrophils with DCs that potently interact with naïve T cells and instruct T cell polarization [34
]. Indeed, neutrophils can generate chemotactic signals that attract host DCs. They can proteolytically activate prochemerin to generate chemerin that attracts immature and plasmacytoid DCs [35
] or antimicrobial compounds such as -defensins implicated in the recruitment of immature DCs [36
]. In humans, it has been shown that the cellular interaction between neutrophils and DCs through membrane-activated complex-1 expressed on neutrophils and DC-SIGN expressed on immature DCs leads to DC maturation [37
]. During Toxoplasma gondii infection, neutrophils were described to induce DC maturation, as well as IL-12 and TNF- production by DCs [38
]. Taken together, the data reported herein demonstrate that DCs overexpressing CD95L can exert antitumoral activities in vivo by causing an inflammatory response involving neutrophils and by harnessing a tumoricidal CD8+ T cell-mediated response.
ACKNOWLEDGEMENTS
The Institute for Medical Immunology is sponsored by the government of the Walloon Region and GlaxoSmithKline Biologicals. This study was also supported by the Fonds National de la Recherche Scientifique (FNRS; Belgium) and an Interuniversity Attraction Pole of the Belgian Federal Science Policy. S. B. and F. M. were supported by FNRS grants, and V. F. was a research associate at the FNRS. S. B. performed research and wrote the paper, N. O. H. and F. M. contributed to performing research, S. F. performed the histological work, K. T. contributed vital reagents, M. G. contributed to design research, and V. F. designed research and wrote the paper. All authors checked the final version of the manuscript. We thank Pierre Coulie and André Tonon from the Cellular Genetics Unit of the Université de Louvain (Brussels, Belgium) for help in cell sorting. We thank Nike Claessen for technical assistance with the immunohistochemistry and Valérie Lebon for maintenance assistance.
Received January 31, 2008; revised May 26, 2008; accepted May 26, 2008.
REFERENCES
in the absence of IL-12 during DC-T cell cognate interaction: a new role for Fas ligand in inflammatory responses J. Exp. Med. 192,1661-1668 coupled to membrane of apoptotic cells favors the cross-priming to melanoma antigens J. Immunol. 172,2643-2650
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
