Abstract
Dendritic cell (DC)-based cancer immunotherapy has been paid much attention as a new and cancer cell-specific therapeutic in the last decade; however, little clinical outcome has been reported. Current limitations of DC-based cancer immunotherapy include sparse information about which DC phenotype should be administered. We here report a unique, representative, and powerful method to activate DCs, namely recombinant Sendai virus-modified DCs (SeV/DC), for cancer immunotherapy. In vitro treatment of SeV without any bioactive gene solely led DCs to a mature phenotype. Even though the expression of surface markers for DC activation ex vivo did not always reach the level attained by an optimized amount of LPS, superior antitumor effects to B16F1 melanoma, namely tumor elimination and survival, were obtained with use of SeV-GFP/DC as compared with those seen with LPS/DC in vivo, and the effect was enhanced by SeV/DC-expressing IFN-β (SeV-murine IFN-β (mIFN-β)/DC). In case of the treatment of an established tumor of B16F10 (7–9 mm in diameter), a highly malignant subline of B16 melanoma, SeV-modified DCs (both SeV-GFP/DC and SeV-mIFN-β/DC), but not immature DC and LPS/DC, dramatically improved the survival of animals. Furthermore, SeV-mIFN-β/DC but not other DCs could lead B16F10 tumor to the dormancy, associated with strongly enhanced CD8+ CTL responses. These results indicate that rSeV is a new and powerful tool as an immune booster for DC-based cancer immunotherapy that can be significantly modified by IFN-β, and SeV/DC, therefore, warrants further investigation as a promising alternative for cancer immunotherapy.
Cancer vaccines have focused on the induction of CTLs that specifically attack tumor cells in an Ag-restricted manner without exerting a significantly harmful effect on nontumor cells. The induction of tumor-specific CD4+ T cells is also important not only in helping the CD8+ CTL response but also in mediating antitumor effector functions through the induction of eosinophils and macrophages (1). To boost these immune responses, several substances have been used as cancer vaccines, including gene-modified autologous tumor cells, peptide vaccine, plasmid DNA, and Ag-loaded dendritic cells (DCs)3 (2). Despite current extensive efforts by physicians and scientists in clinical studies of cancer immunotherapy, very little clinical outcome has been reported (3).
DCs are the most potent and professional APCs that determine either Th1 or Th2 polarization of naive T cells, and they have been a promising tool for cancer immunotherapy. The immature state of DCs is known to be appropriate for Ag processing, and in turn, they must be matured to fully activated DCs, which express high levels of cell surface MHC-Ag complex and costimulatory molecules, for a sufficiently productive CTL response (4). Since the first promising clinical study of a DC-based cancer vaccine was reported in 1996 (5), similar studies have been performed using this type of vaccine against several cancers, including metastatic melanoma, all over the world (6, 7). These early clinical trials suggested the potential of DC-based immunotherapy, but the extensive follow-up studies concluded that the current strategy remains immature as a standard therapeutic for cancer treatment (2).
Current issues of DC-based cancer immunotherapy include a lack of information regarding the following points: 1) the most effective DC subtypes, 2) the optimal conditions and activation stimuli to generate activated DCs showing optimal antitumor effect in vivo, 3) the optimal route for administration, and 4) the optimal dose and frequency of DC vaccinations (2). Because no one knows what the optimal performance and precise clinical indications of DC-based cancer immunotherapy are at present, preclinical assessment regarding these points in detail is absolutely required to obtain more data about DC therapy in clinical settings.
Currently, “virotherapy” is an indicative term for tumor-selective oncolytic virus therapy for cancer, namely “oncolytic virotherapy” (8). However, the first use of the term virotherapy appeared in a Japanese article in 1960, demonstrating the modest antitumor effect of a direct intratumor (i.t.) and/or intradermal injection of bovine vaccinia virus for patients with skin cancers (9). This use of the term virotherapy, therefore, included the concept of “immunostimulatory virotherapy.”
In the last decade, we extensively examined the use of recombinant Sendai virus (SeV) as a novel and powerful gene transfer agent as the cytoplasmic gene expression system (10, 11, 12). SeV, a member of family Paramyxoviridae, has a nonsegmented negative-strand RNA genome and makes use of sialic acid residue on surface glycoprotein or asialoglycoprotein present on most cell types as a receptor (13, 14). As SeV uses a cytoplasmic transcription system, it can mediate gene transfer to a cytoplasmic location, avoiding possible malignant transformation due to genetic alteration of host cells (10, 15), that is a safety advantage of SeV. Furthermore, there are technical advantages in the use of rSeV as a gene therapy vector; first, the infectious activity of SeV particles is stable to be easily concentrated to high titers by ultracentrifugation, which is in clear contrast to the features of retroviral vectors. Second, and most importantly, the modalities of target cell processing and viral transduction are technically nondemanding and feasible in clinical situations that require transduction into large numbers of target cells, including hemopoietic stem cells (16). Despite these advantageous features of SeV in clinical gene therapy strategies over the other vector systems (10, 11, 12), the related immune responses due to virus administration in vivo have been hazardous to expand the use of this mode of vector in the clinical setting, similar to other viral vectors including adenoviruses. During extensive assessment of the mechanisms of immune responses against SeV, we found that ex vivo infection of SeV to immature DCs resulted in their maturation and activation spontaneously, suggesting their possible use for cancer immunotherapy as an immunostimulatory virotherapy.
In the present study, for the first time, we show that i.t. administration of replication-competent SeV-modified DCs (SeV/DCs) induces a dramatically efficient antitumor effect on established tumors in vivo, an effect comparable to that seen with DCs treated with LPS that is well-known as a strong DC stimulator irrelevant to clinical use. Furthermore, we here show that antitumor immunity against an IFN-β-sensitive tumor, a B16 melanoma, is strongly enhanced by the use of SeV/DCs expressing a foreign IFN-β gene.
Materials and Methods
Mice and tumor cell lines
Female 6- to 8-wk-old C57BL/6 mice (H-2b, for B16 melanomas) and C3H/HeN mice (H-2k, for MH134) of Charles River grade were obtained from KBT Orientals and kept under specific pathogen-free and humane conditions. Murine malignant melanoma B16F1 and B16F10 cells were purchased from American Culture Collections (ATCC). MH134, a murine hepatocellular carcinoma cell line, and X5563, a plasmacytoma cell line derived from C3H/HeN mice, were maintained as described (17). An NK-sensitive lymphoma cell line, YAC-1, and a T cell lymphoma cell line of C57BL/6 mice origin, EL-4, were also purchased from ATCC. These cell lines were maintained in complete medium (RPMI 1640 medium; Sigma-Aldrich) supplemented with 10% FCS (BioWest), penicillin, and streptomycin under a humidified atmosphere containing 5% CO2 at 37°C.
Recombinant SeVs
rSeVs were constructed as described previously (10). In brief, the entire cDNA-coding jelly fish enhanced GFP (for SeV-GFP), luciferase (for SeV-luciferase), and murine IFN-β (for SeV-mIFN-β) were amplified by PCR, using primers with a NotI site and new sets of SeV E and S signal sequence tags for an exogenous gene, and then inserted into the NotI site of the cloned genome. Template SeV genomes with an exogenous gene and plasmids encoding N, P, and L proteins (plasmids pGEM-N, pGEM-P, and pGEM-L, respectively) were conjugated with commercially available cationic lipids, then cotransfected with UV-inactivated vaccinia virus vT7–3 into LLMCK2 cells. Forty hours later, the cells were disrupted by three cycles of freezing and thawing and injected into the chorioallantoic cavity of 10-day-old embryonated chicken eggs. Subsequently, the virus was recovered and the vaccinia virus was eliminated by a second propagation in eggs. Virus titer was determined using chicken RBC in a hemagglutination assay, and viruses were kept frozen at −80°C until use. Expression of mIFN-β was confirmed by Western blotting in the culture medium of COS7 cells and DCs transfected with SeV-mIFN-β (data not shown).
Generation of DCs and transfection with SeVs
When preparing murine bone marrow-derived DCs (mBM-DCs), we paid serious attention to maintaining an endotoxin-free condition using endotoxin-free reagents throughout this study. mBM-DCs were generated as previously described with minor modification (18, 19). Briefly, bone marrow cells from C57BL/6 or C3H/HeN mice were collected and passed through a nylon mesh, and RBC and lineage-positive (B220, CD5, CD11b, Gr-1, TER119, 7/4) cells were depleted by using the SpinSep mouse hemopoietic progenitor enrichment kit (StemCell Technologies). These lineage-negative cells (5–10 × 104/5 ml/well) were cultured in 50 ng/ml GM-CSF (PeproTech) and 25 ng/ml IL-4 (PeproTech) in endotoxin-free complete medium in 6-well plates. On day 4, half of the culture medium was replaced by fresh medium supplemented with GM-CSF and IL-4 at the same concentration. On day 7, DCs were collected and used for subsequent experiments. For SeV-mediated transduction, DCs (1 × 106 cells/ml) were simply incubated with SeVs at an indicated multiplicity of infection (MOI) without any supplementation.
In vitro cytotoxic assay with IFN-β
B16F1, B16F10, and MH134 cells were seeded in 96-well plates at 5000 cells/well, and 24 h later, recombinant murine IFN-β (rmIFN-β; PBL Biomedical Laboratories) was added to each well at various concentrations. Forty-eight hours later, cell viability was assessed by a modified MTT assay using a Cell Counting Kit-8 (Dojin Laboratories). Results were calculated as the percentage of viability = (OD of sample − OD of blank)/(OD of A − OD of blank) × 100, where OD corresponds to A wells without rmIFN-β.
Influence of MHC class I expression on tumor cells by IFN-β
B16F1, B16F10, and MH134 cells (1 × 105/ml) were incubated in the presence or absence of rmIFN-β (1000 U/ml) at 37°C for 48 h. B16 or MH134 cells were collected and stained with FITC-conjugated anti-mouse H-2Kb or H-2Kk (BD Pharmingen), respectively, and were analyzed using a FACSCalibur (BD Biosciences). Dead cells were excluded by staining with propidium iodide.
Luciferase assay
The collected mBM-DCs were treated with lysis buffer (Promega) with a protease inhibitor mixture (10), centrifuged, and 20 μl of the supernatant was subjected to luciferase assay. Light intensity was measured after 10 s of preincubation at room temperature using a luminometer (model LB9507; EG&G Berthold) with 10 s integration. Protein concentrations were measured by Bradford’s method using a commercially available protein assay system (Bio-Rad) (11). The data were expressed as relative light units per milligram of protein, and each sample was measured more than twice.
Flow cytometric analysis for costimulation-related molecules on DCs
DCs were plated in fresh medium (1 × 106 cells/ml) and were incubated with SeV-GFP or SeV-mIFN-β, each at a MOI of 40, or LPS (2 μg/L) for 48 h. Biotinylated anti-mouse I-Ab, CD40, CD80, CD86, CCR7, ICAM-1, and allophycocyanin-conjugated anti-CD11c (BD Pharmingen) mAbs were used for each primary Ab. The collected DCs were centrifuged and incubated with 100 μl of the supernatant from cultured hybridoma-producing anti-mouse CD16/32 mAb (2.4G2; from ATCC) for 30 min at 4°C. The cells were incubated with primary Abs for 30 min at 4°C, and biotinylated Abs were detected by subsequent staining with streptavidin-PE (BD Pharmingen). Just before application to the cytometer, we added 125 ng of propidium iodide to cell suspension to exclude dead cells. Cells were analyzed using a FACSCalibur with the CellQuest software (BD Biosciences Japan). Data analysis was performed using FlowJo 4.5 software (Tree Star).
Cytokine production of cultured DCs
The cultured DCs were plated in fresh medium (1 × 106
DC-based immunotherapy of the established tumor
B16F1 melanoma: early treatment regimen (see Fig. 2⇓).
For tumor lysate preparation, B16 melanoma cells were harvested and processed by three rapid cycles of freezing and thawing. As a control, mBM-DCs with neither tumor lysate nor stimulator were used. The other mBM-DCs were pulsed with tumor lysate (ratio of DC number to number of tumor cells for lysate = 1:3) for 18 h and then were incubated with SeV-GFP (MOI = 40; SeV-GFP/DCs), SeV-mIFN-β (MOI = 40; SeV-mIFN-β/DCs), or LPS (2 μg/L; LPS/DCs) for 8 h. Then all DCs were added with 50 μg/ml polymyxin B (Sigma-Aldrich) and were carefully washed two times before injection. Intradermal implantation (C57BL/6 for 1 × 105 B16F1 cells) was done onto the abdomen on day 0, and 1 × 106 DCs were injected i.t. on days 3, 10, and 17. For all injections, materials were suspended in a 100-μl volume of PBS. The size of tumors was assessed using microcalipers three times a week, and the volume was calculated by the following formula: (tumor volume; mm3) = 0.5236 × (long axis) × (short axis) × (height) (20).
B16F10 melanoma and MH134 hepatocellular carcinoma: later treatment regimen (see Fig. 4⇓).
To examine the potentials of cancer vaccines tested here to treat highly malignant phenotypes in vivo, we further assessed “later treatment regimen” when the tumors were well-established and vascularized (7–9 mm in diameter) (3).
DCs were collected as described above, except for control DCs which were pulsed with tumor lysate but not stimulated. Intradermal implantation (C57BL/6 for 1 × 105 B16F10 cells and C3H/HeN for 1 × 106 MH134 cells) was done onto the abdomen on day 0, and 1 × 106 DCs were injected i.t. on days 10, 17, and 24. The size of tumors was assessed as described above.
51Cr release assay for cytolytic activity of NK cells and CTLs
Prepared DCs were i.t. administered three times into tumor-bearing C57BL/6 mice (B16F10) or C3H/HeN mice (MH134) at 1 × 106 cells/100 μl on days 10, 17, and 24. One week after the last immunization, splenocytes were obtained and contaminated erythrocytes were depleted by 0.83% ammonium chloride. For NK cell-lysis assay, the splenocytes were directly used as NK effector cells. For CTL assay, 4 × 106 splenocytes were cultured with 1 μM TRP-2 peptide (H-2b-restricted peptide = SVYDFFVWL) (21) for B16 melanoma model, or with 3 × 105 inactivated MH134 cells treated with 100 μg/ml mitomycin for the MH134 model in 1 ml of complete medium in a 24-well culture plate. Two days later, 30 IU/ml human rIL-2 was added to the medium. After 5 days, the cultured cells were collected and used as CTL effector cells. Target cells (YAC-1, TRP2-peptide-pulsed EL-4, lymphocytic choriomeningitis virus (LCMV) peptide (H-2b-restricted peptide = AVYNFATCGI) pulsed EL-4 (for third party of B16), MH134 cells, and X5633 (for third party of MH134)) were labeled with 100 μCi Na251CrO4 for 1.5 h, and Cr release assay was performed as previously described (22). The percentage of specific 51Cr release of triplicates was calculated as follows: ((experimental cpm − spontaneous cpm)/(maximum cpm − spontaneous cpm)) × 100. Spontaneous release was always <10% of maximal Cr release (target cells in 1% Triton X-100).
In vivo depletion of immune cell subsets
Anti-CD4 and anti-CD8 mAbs (250 μg/dose) were derived from GK1.5 and 53-6.72 hybridoma cells, respectively (23, 24). Anti-asialo GM1 (Wako) was given i.p. (50 μg/dose) for NK cell depletion. Elimination of CD4+ or CD8+ cells in tumor-bearing mice (n = 4–5 in each group) was done by i.p. injection of mAbs on days 5, 6, 7, 10, 13, 16, 19, 21, 24, 27, and 30 after the primary tumor inoculation. Flow cytometry confirmed >98% depletion of the target cells for at least 7 days after injection in all animals.
Histopathological analysis
B16F10 tumor was treated twice (days 10 and 17), and freshly excised tumor tissues on day 20 were divided into the longitudinal two sections; the half was embedded in Tissue-Tec OCT compound (Sakura) and the other was embedded in paraffin. Paraffin sections were stained with H&E. The cryostat sections were subjected to immunochemical examinations with mAbs specific to CD4 (L3T4; BD Pharmingen) or CD8 (Ly-2; BD Pharmingen). Tumor area was measured by Macscope (Mitani). CD4- or CD8-positive cells were counted in total viable tumor area and peripheral stromal tissue within 0.5 mm from the margins of tumor tissue under the optical measure-assisted microscope.
Statistical analysis
All data were expressed as the mean ± SEM, and were analyzed by one-way ANOVA with Fisher’s adjustment, except for animal survival. Survival was plotted using Kaplan-Meier curves and statistical relevance was determined using log-rank comparison. A probability value of p < 0.05 was considered significant.
Results
Transfection efficiency of SeV into mBM-DCs and their spontaneous activation
Immature mBM-DCs from BL/6 mice, propagated in the presence of GM-CSF and IL-4 for 7 days, were collected and transfected by SeV-luciferase or SeV-GFP for investigating gene transduction efficiency. As shown in Fig. 1⇓a, left panel, dose-dependent luciferase expression was shown and the optimized expression was found around MOI = 40–100. Repeated FACS analyses for DCs transfected by SeV-GFP demonstrated that >90% of GFP-positive DCs were detected at more than MOI = 40, a finding representatively shown with forward scatter/side scatter (fsc/ssc) gating at MOI = 40 (Fig. 1⇓b), therefore, all of the following experiments were performed at this titer as an optimal dose.
Gene transfer efficiency of SeV to bone marrow-derived immature DCs and their spontaneous activation. Seven days after cultivation to generate immature DCs under IL-4 and GM-CSF, DCs were treated by SeV-luciferase or SeV-GFP. Forty-eight hours later, DCs were subjected to each analysis. Each experiment was done in triplicate or more and showed similar results. a, Quantitative optimization of dose-dependent gene expression efficiency using SeV-luciferase (left graph) and SeV-GFP (upper right two graphs: positive cell ratio, bottom: MFI, mean fluorescent intensity). Optimized expression was seen around MOI = 40–100. b, Scattered plots for fsc/ssc gating of DCs transfected with SeV-GFP at MOI = 40. Note 2 major populations of GFP-expressing DCs, GFPhigh (66.0%) and GFPlow (26.6%) (bottom right panel). c, Expression level of activation markers (CD40, CD80, CD86, MHC class II) and Ags related to trafficking (CCR7) and to adhesion (ICAM-1) of DCs, treated with SeV-GFP, SeV-mIFN-β, or LPS (2 μg/ml). MFIs of each analysis were given below. d, Typical cytokine secretion from DCs treated with LPS, SeV-GFP, or SeV-mIFN-β. Immature DCs were used as negative control.
We next assessed the surface markers of DCs treated with SeV-expressing GFP or mIFN-β without any other stimulant, directly compared with a well-known strong but clinically irrelevant DC activator, LPS, at 2 μg/ml for 48 h of exposure, which had been shown to be the optimal dose for mDC activation in our preliminary data (data not shown). As shown in Fig. 1⇑c, repeated FACS analyses showed that DCs treated with SeV-GFP or SeV-mIFN-β resulted in the high-level expression of the costimulatory molecules tested here, namely MHC class II, CD80, and CD86 molecules, which did not reach the level seen in the DCs treated with LPS. In comparison to their sharp expression patterns on LPS/DC, those seen on SeV-GFP or SeV-mIFN-β showed broad expression, suggesting the result of the broad expression of transgene seen in Fig. 1⇑b. Other surface markers related to trafficking (CCR7) and adhesion (ICAM-1) were also up-regulated on DCs treated with SeV-GFP or SeV-mIFN-β which were nearly comparable to the level seen on LPS/DC. These results thus demonstrated that SeV could not only effectively transfer exogenous genes into DCs, but also spontaneously transform immature DCs to near fully activated mature DCs without other manipulation irrespective of the exogenous mIFN-β expression. In turn, LPS-activated DCs were seriously resistant to SeV-mediated gene transfer (usually <5%, data not shown).
To assess further phenotype of DCs activated by SeV, release of typical cytokines was examined by ELISA. As shown in Fig. 1⇑d, up-regulation of type I IFN, e.g., IFN-β, was seen in DCs treated with SeV-GFP or SeV-mIFN-β, but not in immature DCs and LPS/DCs. In contrast, strongest expression of other cytokines tested, including Th1 cytokines (p70 subunit of IL-12 and IFN-γ), was seen in LPS/DCs.
Together with these results, SeV induces spontaneous maturation and activation of mBM-DCs, however, their phenotype is not equal to those seen in the treatment with LPS.
SeV/DC therapy induces complete elimination of B16F1 melanoma in vivo
Next, we asked whether DCs activated by SeV might have therapeutic potentials against an immune-competent murine melanoma model. We tested this by an early treatment regimen as follows.
First, to assess the effective route for antitumor activity of DC therapy, i.t., distant s.c., and i.v. (via tail vein, i.v.) injections of tumor lysate-pulsed DCs activated by SeV-GFP was started and repeated three times every week 3 days after B16F1 (low malignant subline) cell inoculation. Efficient antitumor effect was only seen in the case of i.t. injection (Fig. 2⇓a), a finding similar to the previous report by the other group in use of naive DCs (25). Therefore, the following experiments were done via the i.t. route.
Assessment of antitumor activity of DCs modified by SeV against low-malignant murine melanoma B16F1. Three days after tumor cell inoculation, DCs were injected i.t. according to the indicated regimen (a, bottom scheme: early treatment regimen). Tumor lysate was pulsed to immature DCs which were subsequently treated with LPS, SeV-GFP, or SeV-mIFN-β, and these DCs were used for DC immunotherapy 8 h later. DC treatment was done three times. The data demonstrated were the total of three independent experiments. a, Time course of the tumor volume treated with SeV-GFP/DC/lysate via different administration routes (distant s.c.: s.c. injection, i.v.: i.v. injection via tail vein, and i.t.: intratumor injection). Numbers indicate tumor-bearing animals that survived over 28 days. b, Experimental design to assess the stimulator dependent antitumor effect of DCs. c, Dot plots indicating the tumor size of B16F1 melanoma following DC therapy. Numbers below indicate the dead animals within day 28. d, Survival curve of the mice bearing B16F1 melanoma treated with various DCs. Significant prolongation of survival was seen in SeV/DC groups. e, Typical and representative gross observation of mice with B16F1 tumors treated with or without SeV-mIFN-β/DC/lysate 26 days after tumor cell inoculation. Note complete rejection of tumors by SeV-mIFN-β/DC/lysate treatment (arrow), which was confirmed by histopathological examination (data not shown). Tumor rejection rate of mice at day 75 was also presented in the right column.
Next, we directly compared the antitumor effect of i.t. administration of immature DC, LPS/DC/lysate, SeV-GFP/DC/lysate, and SeV-mIFN-β/DC/lysate by early treatment regimen indicated in Fig. 2⇑b. In this study, we used immature DCs without tumor lysate as a control, because it has been known that ex vivo uptake of tumor Ag itself led DCs to activated state (26). Because B16 melanoma-burden mice were well-known to start to die around 2 wk after inoculation irrespective of tumor size, we here evaluated two more parameters, namely survival and the number of mice with eliminated tumor, to assess the beneficial effects of cancer vaccines.
Tumor size.
As shown in Fig. 2⇑c, six animals of the untreated group and 1 of the LPS/DC/lysate group were dead within day 28. The tumor size on day 28 was efficiently disturbed by DC immunotherapy modified by SeV-GFP, SeV-mIFN-β, or LPS. However, a similar finding was also seen in the use of immature DCs, suggesting the results of spontaneous activation of DCs in vivo via i.t. route, as previously described by the other group (25).
In contrast, other parameters, survival, and tumor elimination ratio showed interesting results.
Survival.
Significant prolongation of survival of animals was found only in groups of SeV-GFP/DC/lysate and SeV-mIFN-β/DC/lysate (Fig. 2⇑d).
Number of mice with tumor eliminated.
Ten of 14 mice (71.4%) treated with SeV-GFP/DC and 8 of 9 (88.9%) treated with SeV-mIFN-β/DC completely eliminated the tumor at day 75 (Fig. 2⇑e), findings that were also confirmed microscopically (data not shown).
These results indicated that DC immunotherapy was beneficial for treating B16F1 melanoma in immune-competent mice, in views of survival and tumor elimination, and the therapeutic effects of DCs modified with SeV were equal or possibly more to that with strong DC activator LPS, an effect that was significantly improved by exogenous IFN-β expression.
Limited responses of mIFN-β protein as well as mIFN-β gene therapies on highly malignant and less immunogenic B16F10 melanoma in vivo
Direct efficacy evaluation demonstrated in Fig. 2⇑ suggests the potential use of SeV-mediated modification of DC functions for cancer immunotherapy; however, significant and clear improvement over LPS/DC was not found. Furthermore, immature DCs showed significant reduction of tumor size, suggesting that the tumor model of low malignant B16F1 melanoma was not appropriate to evaluate the potentials of SeV/DCs. Therefore, we next tried to treat a highly malignant and less immunogenic subtype of B16 melanoma, namely B16F10, using SeV/DCs. Before this trial, we first assessed the biological effects of mIFN-β in vitro and of protein and gene therapy with mIFN-β in vivo.
As shown in Fig. 3⇓a, mild and dose-dependent growth inhibition was seen in B16F1 cells in vitro, and the effect was more pronounced in B16F10 cells, indicating that B16 melanomas are sensitive to IFN-β, similar to findings in clinical settings (27). In contrast, murine hepatocellular carcinoma, MH134, was not apparently sensitive to IFN-β (Fig. 3⇓a).
Antitumor activity of murine IFN-β against murine tumor cells in vitro (a) and in vivo (b). a, In vitro cytotoxicity of mIFN-β to murine melanomas (lower malignancy, B16F1; higher malignancy, B16F10) and a hepatocellular carcinoma (MH134). After 48-h culturing in the presence of mIFN-β protein, cell viability was assessed. The viability of MH134 (□) was not affected by mIFN-β, and in contrast, melanomas were sensitive to mIFN-β, suggested by the observed dose-dependent effects. Note that B16F10 (○) was more sensitive than B16F1(▴) to mIFN-β. b, Antitumor activity of protein (▴) and gene (•) therapies of mIFN-β in vivo. Ten days after tumor cell inoculation, when the established tumor was well-vascularized, protein or gene therapy was started according to the indicated regimen (b, bottom scheme: late treatment regimen).
Next, we treated B16F10 tumors in vivo with mIFN-β protein, which has been a standard clinical therapeutic, as well as with SeV-mIFNβ as a gene therapy, by a treatment regimen beginning at a later stage of tumor development (tumor diameter = 7–9 mm). As shown in Fig. 3⇑b, both treatments tended to delay the growth of the B16F10 tumors; however, the tumor volume had relapsed by about day 30. The relapse of tumor growth was likely due to withdrawal of the local concentration of mIFN-β, because in vivo expression of SeV-mediated gene transfer has been shown to be transient (11).
These findings suggest that protein and gene therapies by mIFN-β contributed to the suppression of tumor growth of B16F10 tumors in vitro and in vivo; however, the effect was not sufficient to control a highly malignant cell type, even though the tumor cells are sensitive to mIFN-β. This could be explained because the duration of local concentration of mIFN-β might not be sufficient to show the long-lasting antitumor effect.
Modulation of antitumor effects and immune responses to IFN-β-sensitive, established tumors by SeV/DCs expressing mIFN-β
Antitumor effect.
IFN-β is known to be an antitumor cytokine via multiple mechanisms, including a direct antiproliferative effect, enhancement of NK cell activity, and up-regulation of tumor Ag and MHC class I and II (28). Therefore, we next asked whether DC therapy modulated by exogenous mIFN-β genes might affect the antitumor effect as well as immune responses against B16F10 melanoma.
First, we assessed MHC class I expression and its modulation by mIFN-β by flow cytometry analyses, because it is well-known that the expression of this molecule on tumor cells is required to be recognized by antitumor CTLs induced by cancer immunotherapy (29). As shown in Fig. 4⇓a, the baseline expression of MHC class I (H-2k) was high, and the level was not significantly changed by the treatment with mIFN-β. In contrast, both B16F1 and B16F10 cells expressed a very low level of MHC class I (H-2b) at baseline, and after 48 h of culturing in the presence of 1000 U/ml mIFN-β protein, the level of MHC class I was dramatically increased. Together with Fig. 3⇑a, these results might suggest that SeV/DCs expressing mIFN-β possibly enhanced the antitumor effect seen in the use of sole SeV/DCs.
Tumor cell-dependent, divergent effects of SeV/DC immunotherapy expressing mIFN-β (late treatment regimen). ∗, p < 0.01. a, Flow cytometry analysis for MHC class I expression level of tumors treated with or without mIFN-β. After 48-h cultivation of tumor cells (MH134; H-2k H-2k, and B16F1, B16F10 melanoma; H-2b) with or without mIFN-β, cells were subjected to the analysis. Note that the expression of MHC class I in both melanoma cells was very low, findings that were strongly up-regulated by mIFN-β treatment. b, Antitumor effect of SeV/DCs on a MH134 tumor, which was resistant to mIFN-β, with or without expression of mIFN-β in vivo. C3H/HeN mice were inoculated intradermally with 1 × 106 MH134 cells, which were resistant to mIFN-β, on day 0, and DC treatment was started at day 10 using 106 cells of LPS/DC (×), SeV-IFN-β/DC (•) or SeV-GFP/DC (□), as indicated in the bottom scheme. The data contains the results of all animals in two independent experiments. c, Time course of the volume of individual B16F10 tumors, a IFN-β-sensitive malignancy, treated with DC/lysate, LPS/DC/lysate, SeV-GFP/DC/lysate, or SeV-mIFN-β/DC/lysate. Untreated animals were used as a control group. DW20 indicates the number of death animals within day 20. +, Individual animals with death after day 21. d, Survival curve of the mice bearing B16F1 melanoma treated with various DCs. Significant prolongation of survival of mice bearing B16F10 melanoma was seen in SeV/DC groups, but not others. Data demonstrated were the total of four independent experiments. e, A graph indicating the effect of exogenous mIFN-β on tumor size of B16F10 melanoma. SeV/DC-treated animals surviving over 36 days were evaluated.
To test this possibility, we next evaluated the antitumor effect SeV-mIFN-β/DCs, directly compared with those of LPS/DCs and SeV-GFP/DCs using established MH134 and B16F10 melanoma in vivo via later treatment regimen.
MH134 tumor
Tumor size (Fig. 4⇑b).
As expected, the tumor growth of established MH134 tumors was markedly inhibited irrespective of the types of DCs, namely, LPS/DCs, SeV-GFP/DCs, and SeV-mIFN-β/DCs, by the later treatment regimen (Fig. 4⇑b).
Survival.
No animal bearing MH134 was dead during experimental course.
Number of mice with tumor eliminated.
No animal showed elimination of tumor irrespective of treatments.
B16F10 tumor
Tumor size (Fig. 4c).
Evaluation of tumor size was relatively difficult, because 40% and all of untreated animals were dead within days 20 and 36, respectively (Fig. 4⇑c, upper panel). In addition, death of animals was not corresponded to the size of tumor.
Survival.
Although DC/lysate and LPS/DCs were likely to inhibit the growth of B16F10 tumors in vivo (Fig. 4⇑c), both treatments did not contribute to the significant prolongation of the survival (Fig. 4⇑d). In contrast, significant prolongation of survival over day 50 and tumor dormancy of animals bearing B16F10 melanoma were observed only in SeV/DC groups. A beneficial effect of the IFN-β transgene was seen in tumor size which was evaluated on day 50 (Fig. 4⇑e).
Number of mice with tumor eliminated.
No animal showed elimination of tumor irrespective of treatments.
Together, these results indicate that modification of the function of DCs by SeV expressing mIFN-β dramatically enhances the antitumor effect to tumor cells that are sensitive to IFN-β. Furthermore, these results strongly suggested that the modification of DC functions by SeV was beneficial on the survival, and exogenous IFN-β mainly contributes to the reduction of tumor volume.
Immunological assessments
To explore the immunological effectors for the antitumor effects in DC therapy as well as gene therapy, we next focused on NK cells and CTL activities in mice bearing B16F10 and MH134 tumors treated by each method.
NK cells obtained from mice bearing B16F10 tumors treated with DCs, including LPS/DCs, SeV-GFP/DCs, and SeV-mIFN-β/DCs, did not show significant cell lysis activity. In contrast, NK cells from mice treated with direct i.t. vector injection of SeV-mIFN-β, aforementioned in Fig. 3⇑b, showed apparently strong cell lysis activity which was comparable to a positive control level (poly I:C) (Fig. 5⇓a). These findings indicate that NK cells do not significantly contribute to the antitumor effect of DC therapy.
Effectors contributing to the antitumor effect of SeV/DC immunotherapy. a, Assessment of NK cell activity in B16F10 melanoma. Thirty-one days after B16F10 cell inoculation, splenocytes were isolated from the mice treated with SeV-GFP/DC/lysate (□), SeV-mIFN-β/DC/lysate (•), LPS/DC/lysate (▪), or intratumoral direct injection of SeV-mIFN-β vector (X). The cytolytic function against 51Cr-labeled YAC-1 targets was assessed by 51Cr release. Splenocytes from the mouse, which were treated with 150 μl of poly I:C 24 h before the assay, were used as a positive control. b, Assessment for CTL activity of MH134. Induction of tumor-specific CTLs after i.t. administration of SeV-GFP/DC/lysate (□), SeV-mIFNβ/DC/lysate (•), and LPS/DC/lysate (▪), which was repeated three times every week according to the late treatment regimen. Controls included tumor-bearing mice without any treatment (○). X5563 was also used as a target of a third party. Seven days after the last treatment, splenocytes were isolated and restimulated in vitro for 5 days with mitomycin C-treated MH134 cells, and cytolytic activity against 51Cr-labeled targets was measured. The figure shows results from one of three similar experiments. c, Assessment for CTL activity of B16F10 cells. Induction of tumor-specific CTLs after i.t. administration of SeV-GFP/DC/lysate (□), SeV-mIFN-β/DC/lysate (•), and LPS/DC/lysate (▪), which was repeated three times every week according to the late treatment regimen. Controls included tumor-bearing mice without any treatment (○). LCMV peptide was also used as a target of a third party. The method was same as above. d, Determination of immune cell subsets against MH134 or B16F10 responsible for the protective immunity induced by SeV-mIFN-β/DCs was Ab-mediated via an in vivo depletion analysis, as described in Materials and Methods. These bar graphs show the tumor volume on day 30 after inoculation of tumor cells. Anti-CD4 (GK1.5), anti-CD8 (53-6.72), or anti-asialo GM1 was i.p. injected according to the indicated schedule. In all animals, >98% of specific depletion of target cells in the spleen and lymph nodes was confirmed by flow cytometry (data not shown). The data contains all animals of two or three separate experiments.
In contrast, CTLs from the splenocytes of mice showed opposite results, and demonstrated the comparable findings suggested in Fig. 4⇑. In the case of CTLs against MH134 tumor cells, the expression of mIFN-β did not significantly affect the CTL activities induced by SeV/DC therapies (Fig. 5⇑b, left graph). As a control experiment, X5563 tumor cells were used as a target, and showed negative result (Fig. 5⇑b, right panel). In the case of TRP-2 peptide, a tumor-specific Ag of B16 melanoma as a target of CTLs obtained from spleens from mice bearing B16F10 melanoma, SeV/DCs showed a similar level of CTL activity compared with that seen in LPS/DCs, a finding that was markedly enhanced by SeV/DCs expressing mIFN-β (Fig. 5⇑c, left panel). In this case, no significant CTL activity was detected in mice treated with direct vector injection of SeV-mIFN-β. In contrast, a relatively low level of background release was observed when third-party target (LCMV) was used (Fig. 5⇑d, right panel).
These results thus indicate the significant contribution of IFN-β to the CTL activity for an IFN-β-sensitive B16F10 tumor but not for an insensitive MH134 tumor during DC therapy. Furthermore, these data suggest the distinct mechanisms of the antitumor effect of DC therapy compared with that of gene therapy, even when the same therapeutic gene, IFN-β, was used.
Effector cell subsets in DC therapies
To make it clear which subset(s) of cells is important for the antitumor effect of SeV-mIFN-β/DC therapy, we next conducted the effector cell depletion experiment by administration of each depletion Ab against CD8, CD4, and NK (by asialo-GM1) in mice with MH134 or B16F10 tumors treated with or without SeV-mIFN-β/DCs. The dose of each Ab was determined by our repeated preliminary experiments of FACS analyses that showed >98% of the target subject in some lymph nodes and the spleen (data not shown). The same lot of each Ab was used throughout the experiments. The tumor size was evaluated every day, and the data on day 30 are presented.
As shown in Fig. 5⇑c, in the case of MH134 tumors, as expected, depletion either of CD4+ T cells or CD8+ T cells almost totally abrogated the antitumor effect induced by SeV-mIFN-β/DCs, and the depletion of NK cells partly canceled the effect (Fig. 5⇑c, left graph). These findings indicate that the tumor-specific CTL as well as the helper function of CD4 are absolutely necessary to the therapeutic effect of SeV-mIFN-β/DCs, and further, NK cell activity is also involved in the effect.
In the case of B16F10 melanoma, in contrast, such depletion studies demonstrated unexpected results. Depletion of CD8+ T cells during therapy partially cancelled the protective immunity induced by SeV-mIFN-β/DCs; however, mice depleted of NK cells during therapy showed no significant effect (Fig. 5⇑c, right graph), data supported by the cell lysis analysis shown in Fig. 5⇑a. When CD4+ T cells were depleted, however, the antitumor effect of SeV-mIFN-β/DCs was markedly and significantly enhanced, indicating that the net effect of CD4+ T cells was to function as a helper for tumors rather than providing antitumor immunity. These data might possibly reflect the function of regulatory T cells (CD4+ CD25+ T cells) in the CD4+ T cell subset (30).
These findings confirm that CD8+ T cells were the ubiquitous and predominant effector cells in antitumor immunity for both MH134 and B16 melanoma, even though their quantitative contribution may vary depending on the tumor type.
Histological and immunohistochemical assessments for modification of tumor environment
Next, we histopathologically examined the B16F10 melanoma tumor environment, including tumor area (viable tumor and ratio of necrosis) and infiltration of CD4 and CD8 T lymphocytes, using another set of DC therapy experiments. At that time, DC therapy was done on days 10 and 17, and the tumor was harvested on day 20. Four of 10 animals of the untreated group and 3 of 10 of the LPS/DC group but not all SeV/DC groups were dead within day 20.
Histological examination suggested that the infiltration of inflammatory cells around tumor was more pronounced in DC therapy groups (Fig. 6⇓a), and interestingly, more reduction of tumor area and viable tumor area was seen in the SeV/DC groups, but not in the LPS group (Fig. 6⇓b, left and middle graphs). In contrast, the ratio of tumor necrosis was more pronounced in the SeV/DC groups (Fig. 6⇓b, right graph), and these findings were observed irrespective of use of SeV/DCs expressing IFN-β.
Histopathological and immunohistochemical examination of B16F10 tumors on day 20. Dot lines indicate the margin of tumors. a, Typical representative findings of the tumor environment on 20 days after inoculation of B16F10 treated with or without LPS/DC/lysate, SeV-GFP/DC/lysate, and SeV-mIFN-β/DC/lysate on days 10 and 17. Panels of H&E staining indicated as “high powered” are high powered view of their corresponding left panels. Arrowheads indicate the infiltration of chronic inflammatory cells, including lymphocytes and macrophages. Immunohistochemical positive reaction (red cytoplasm) of CD4+ and CD8+ cells are also demonstrated (arrows). These are representative of two separate studies using three to five mice per group. b, Bar graphs indicating the squares of total tumor area (left), viable tumor area (middle), and the ratio of necrotic area (right). H&E-stained sections were subjected to the computer-assisted square measurement. c, Bar graphs indicating the number of immunohistochemically CD4+ and CD8+ cells per mm2 in the viable tumor area and peripheral tumor tissue.
Immunohistochemistry demonstrated more frequent infiltration of CD8 T lymphocytes into tumor and surrounding s.c. tissue by treatment of SeV-mIFN-β/DCs compared with other groups (Fig. 6⇑c, right graph), a comparable finding obtained in Fig. 5⇑c, while no significant difference was determined in CD4 T lymphocyte infiltration among all groups tested (Fig. 6⇑c, left panel).
These results thus suggest the split mechanisms of SeV-mediated modulation of DC function and expression of IFN-β on the antitumor effect of B16F10 melanoma.
No requirement of Ag loading during ex vivo culture of SeV-mIFN-β/DCs
Finally, we investigated the requirement of ex vivo Ag loading during the culturing of SeV-mIFN-β/DCs, because ex vivo Ag loading has been considered to be an essential step in inducing tumor-specific immunity (31, 32). To assess this, we used a late treatment regimen against B16F10 melanoma via i.t. injection of SeV-mIFN-β/DC with or without tumor lysate.
As shown in Fig. 7⇓, the therapeutic effect of SeV-mIFN-β/DCs was not affected by ex vivo Ag-loading when DCs were injected i.t., indicating that the antitumor effect can be led by direct i.t. injection of SeV-mIFN-β/DCs without any supplementation of ex vivo Ags.
a, Requirement of ex vivo pulsation of tumor lysate during SeV-modified DC immunotherapy. The late treatment regimen against B16F10 melanoma was done via intratumor injection of SeV-mIFN-β/DC with or without tumor lysate. Ten days after B16F10 cell inoculation, i.t. injection of SeV-mIFN-β/DC pulsed with (♦) or without tumor lysate (▵) was performed according to the protocol as indicated below. ∗, p < 0.01. b, Schematic representation of the comparison of conventional DC-based immunotherapy and that using SeV-modified DCs. The greatest advantages of the SeV/DC method are that: 1) there is no requirement of a specific stimulus to induce strong DC activation representatively, and 2) Ag-loading is not critical when SeV/DC is administered directly to tumors.
Discussion
During the last decade, clinical trials of DC-based immunotherapy have revealed a relatively limited clinical outcome against intractable malignancies (3). These early results do not always suggest the limited potential of DC-based immunotherapy, however, because there are still a number of issues to be clarified, including how to monitor the activation of DCs, which route and with what frequency DCs should be administered, how Ags should be targeted, and which malignancy and what stages of malignancy should be selected (2). Physicians and scientists should clarify these points and apply the answers to the basic mechanisms of tumor biology and DC therapy, because no one knows, at present, the optimized potentials of DC-based cancer immunotherapy in clinical settings.
We here assessed the antitumor potential of a novel DC-activating modality, rSeV, and directly compared it to that of LPS, which is not relevant clinically but experimentally shows the high performance in terms of DC activation. Key observations obtained in this study were as follows: 1) SeV not only transduced foreign genes to immature DCs but also led them a highly activated state that was comparable to that induced by LPS, 2) the bioactive potential of SeV/DCs as an antitumor agent was seen in tumor-bearing mice in vivo, 3) complete elimination of an established low malignancy tumor, IFN-β-sensitive B16F1, could be found via an early treatment regimen using SeV-GFP/DCs, findings enhanced by SeV/DCs expressing mIFN-β, 4) SeV/DC therapy was likely to be more effective than protein and gene therapies, 5) distinct mechanisms of antitumor effects were suggested between gene therapy and SeV/DC therapy, even though both used the same therapeutic gene, mIFN-β, and 6) the process of ex vivo Ag loading was not important in the therapeutic effect of SeV-mIFN-β/DCs when they were injected i.t.
Because this is the first demonstration of a negative strand RNA virus-mediated DC activation for cancer immunotherapy, there are still a number of unanswered questions regarding this system. However, it should be true that SeV is a strong and important modality for activating DCs which may be useful for cancer immunotherapy.
At present, the precise molecular and cellular mechanisms of activation of DCs by SeV are not well-understood. In this study, we demonstrated that DCs treated by SeV secrete several inflammatory cytokines, including IL-1β, IL-6, and TNF-α, that are strong activators to DCs in mixture (33); however, no knowledge of how DCs exposed to SeV express these cytokines is available at present. A recent important report indicated that DCs treated with SeV produce multiple proinflammatory cytokines independent of TLRs and their adaptors, including MyD88 (34), unlike LPS, CpG (35), or OK-432 (36), which stimulate DCs through a TLR-dependent mechanism.
An important finding obtained in the current study is that stronger antitumor effect on B16 melanomas was seen in SeV-modified DCs compared with that seen in LPS/DCs, even though the expression of costimulatory molecules and typical Th1 cytokines of SeV/DCs did not always reach the levels of LPS/DCs. These results strongly suggest the complex mechanism of antitumor effect; namely, antitumor activity of DC immunotherapy cannot generally be predicted by the expression of these molecules. We are now assessing this point extensively.
An essential finding in the current study is that SeV-mIFN-β/DC therapy was considerably effective for treating IFN-β-sensitive tumors, B16 melanomas, even in the case of the highly malignant B16F10 subtype. More importantly, this effect was shown by the later treatment regimen, on day 10, when the tumor was 7–9 mm in diameter and well-vascularized as revealed by histopathological assessment (data not shown). Because “there are no cancer vaccine models that reproducibly demonstrate that vascularized tumors can be rejected” at present, as noted by Rosenburg et al. (3), SeV-mIFN-β/DC therapy may be an important candidate for overcoming the current limitation in immunotherapy for melanoma in the clinical setting.
An important question raised in this study is what is the precise mechanism(s) of SeV-mIFN-β/DC therapy against B16F10 melanoma? Fig. 5⇑ suggested that NK cells seemed not to be important in this case, while they did actually show the antitumor effect seen in MH134, at least in part. A CD8+ cell depletion study demonstrated CD8+ CTLs as major effectors for SeV-mIFN-β/DC therapy for B16F10 melanoma, and mIFN-β expression in DCs enhanced the CTL activity (Fig. 5⇑b). These findings show that the direct cytotoxicity of mIFN-β and the induction of CD8+ CTLs may be major players in the antitumor effect of SeV-mIFN-β/DC therapy against B16F10 melanoma, which may be modified by the expression of MHC class I in tumor cells. This explains why the antitumor effect against MH134, which was insensitive to mIFN-β, was not modified by the use of SeV-mIFN-β/DCs. Modulation of MHC class I expression is an important part of antitumor immunity; loss or down-regulation of MHC class I is an important mechanism for the tumor’s escape from the host immune system, resulting in the peripheral tolerance (37, 38, 39).
In addition, it is of interest that our current study demonstrated that distinct beneficial effect of SeV/DCs and those expressing exogenous IFN-β; namely, modification of DC function by SeV contributed to the survival and increase of the tumor necrotic area, and exogenous expression of mIFN-β significantly enhanced CTL activity and increased recruitment of CD8+ T cells. These results thus suggest the modulation of host immune response to malignancies via DC immunotherapy by mIFN-β; however, exact mechanisms of apparent improvement of survival and increase of necrotic area via SeV/DCs without exogenous gene, that was not seen by LPS/DCs, are still unknown. One possible explanation is the effect caused by DC-derived endogenous type I IFNs, which were not detected by LPS treatment, caused by viral transduction. According to our careful observation, death of mice bearing B16F10 within day 20 was caused by extensive hemorrhage from ulceration of tumors. It has been shown that low-dose type I IFNs, including IFN-β, reduced tumor vasculature, a possible mechanism of reduced rate of hemorrhage. This is likely because some previous studies demonstrated that the antiangiogenic activity of mIFN-β did not show apparent dose response (40), a comparable finding obtained in the current study. Such multiple functions of IFN-β would be favorable for cancer immunotherapy, further studies should be done to assess the precise mechanism of action of IFN-β in clinical settings.
One more question raised in this study involves the role of CD4+ T cells in antitumor immunity, which showed conflicting results between two different tumors, namely, depletion of CD4+ cells completely canceled the antitumor effect of SeV-mIFN-β/DCs against MH134 tumors and, inversely, markedly enhanced that against B16F10 melanoma. It is still premature, at present, to draw a conclusion, but we believe a possible contribution of CD4+/CD25high regulatory T lymphocytes (T-reg) may be involved. Recent progress in knowledge about how tumors escape from host immune surveillance shows an essential contribution of T-reg in tumor immunity (41). The distinct effects of CD4 depletion between two different tumor types, however, may imply that the findings depend on the cell sources, so further study is called for to determine the tumor and/or DC factors contributing to the function of T-reg.
Can SeV/DC show a significant anticancer effect over that seen using current DC immunotherapy in clinical use? It may be premature to suggest this, however, the present study implies some important advantages of SeV/DC-based cancer immunotherapy. In this study, we showed that SeV/DC therapy revealed a strong antitumor effect over that seen with the use of LPS/DCs, which can produce one of the strongest anticancer effects in experimental conditions, and these findings suggest that the therapeutic potential of SeV/DCs warrants further studies, including clinical trials. Because mass production of good manufacturing practice grade SeV is now available, it is not a long way to move SeV/DC-mediated cancer immunotherapy to clinical practice.
In summary, we demonstrated that SeV-modified DCs showed antitumor effects on established tumors in vivo; and we would therefore like to propose a concept for tumor immunotherapy using SeV-modified DCs known as “immunostimulatory virotherapy.” The current study results strongly suggest that SeV/DC-based cancer immunotherapy may be an important alternative therapy and that the technique warrants further investigation in research as well as in clinical trials.
Acknowledgments
We thank Hiroshi Fujii for tissue processing and immunohistochemistry, and Chie Arimatsu for her help on animal experiments.
Disclosures
The authors have no financial conflict of interest.
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 work was supported in part by a Grant-in-Aid (to Y.Y. and K.S.) from the Japanese Ministry of Education, Culture, Sports, Science, and Technology; and by the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (to Y.Y. and K.S.; Project MF-21).
↵2 Address correspondence and reprint requests to Dr. Yoshikazu Yonemitsu, Division of Pathophysiological and Experimental Pathology, Department of Pathology, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan. E-mail address: yonemitu{at}med.kyushu-u.ac.jp
↵3 Abbreviations used in this paper: DC, dendritic cell; SeV, recombinant Sendai virus; SeV/DC, SeV-modified DC; mIFN-β, murine IFN-β; mBM-DC, bone marrow-derived DC; i.t., intratumoral; T-reg, regulatory T; MOI, multiplicity of infection; LCMV, lymphocytic choriomeningitis virus; fsc/ssc, forward scatter/side scatter.
- Received August 11, 2005.
- Accepted June 23, 2006.
- Copyright © 2006 by The American Association of Immunologists
References
In this issue
