Abstract
Antiangiogenic immunotherapy, which targets molecules critical to tumor angiogenesis, is expected to counteract the negative effect of tumor cell genetic instability on the outcome of immunotherapy targeting tumor antigens. Previously, targeting of individual angiogenic molecules has been shown to inhibit tumor angiogenesis and limit tumor growth. Nevertheless, this approach may be bypassed by redundant angiogenic pathways. To overcome this limitation, we have developed an immunization strategy targeting multiple molecules critical to angiogenesis. To this end, hybrids of dendritic cells (DC) and syngeneic endothelial cells (EC) were used as immunogens, because (a) whole EC express multiple molecules involved in angiogenesis and (b) DC tumor cell hybrids are effective in generating self-antigen–specific immune responses. The immunization strategy included the administration of an agonist 4-1BB–specific monoclonal antibody (mAb), because it augments self-antigen–specific immune responses elicited by DC hybrids. Immunization of mice with DC-EC hybrids and 4-1BB–specific mAb inhibited the growth of B16.F10 melanoma and MC38 colon adenocarcinoma tumors. This effect is mediated by EC-specific CD4+ and CD8+ T-cell responses, which markedly inhibited tumor angiogenesis. No therapy-related side effects, except minor and transient hematologic changes, were observed. Our findings represent a useful background for the design of antiangiogenic immunotherapeutic strategies to control tumor growth in a clinical setting. [Cancer Res 2007;67(16):7875–84]
Introduction
The limited clinical responses in patients with malignant diseases, despite the induction of tumor antigen-specific immune responses (1), are likely to reflect escape mechanisms that enable tumor cells to avoid immune recognition and destruction (2). To counteract this limitation, caused at least in part by tumor cell genetic instability (3), stromal components of the tumor have been targeted, in most cases with cell-mediated immunity and in a few with antibody-mediated immunity. The rationale underlying this strategy is provided by the relative genetic stability of stromal cells, which minimizes the development of immune escape mechanisms (4, 5), and by the crucial contribution of these cells to tumor cell survival and proliferation (6, 7). In this light, we have selected tumor-associated endothelial cell (EC) as targets, because they are critical to tumor angiogenesis and targeting individual EC antigens, such as vascular endothelial growth factor receptor (VEGFR)-2 (8–10), Tie-2 (11, 12), matrix metalloproteinase (MMP)-2 (13), and fibroblast growth factor (FGF)-2 (14), have resulted in inhibition of tumor angiogenesis and control of tumor growth in multiple murine models. However, EC may be provided with alternative pathways to achieve similar growth and survival signals by the redundancy in EC growth factor signaling pathways (15–18). The latter include the highly promiscuous interactions between VEGF and VEGFR family members (19). In addition, the pathways regulating EC growth and survival are not restricted to those that are VEGFR dependent, because tumor angiogenesis can be inhibited by targeting Tie-2, MMP-2, FGF-2, or other angiogenic molecules. These potential limitations have prompted us to develop an immunization strategy that targets multiple molecules critical to the angiogenesis process.
Whole EC, which express multiple molecules involved in tumor angiogenesis, represent useful immunogens, because animals immunized with xenogeneic EC exhibited inhibition of tumor angiogenesis and tumor growth (20). However, the use of xenogeneic EC may be difficult to translate into the clinical setting. Therefore, we have tested whether syngeneic EC, which represent a more clinically relevant immunogen, can induce an antiangiogenic immune response that can control tumor growth. We have used the SVR EC cell line as a source of EC, because they proliferate rapidly in vitro and in vivo and express high VEGF and MMP bioactivity levels (21). Consequently, their expression profile of angiogenesis-associated molecules is expected to closely mimic that found on highly proliferative EC in vivo, such as tumor-associated EC in the tumor microvasculature, rather than on resting EC. To enhance the immunogenicity of syngeneic EC, we have hybridized them with dendritic cell (DC), because hybrids of DC and tumor cells are effective in generating specific immune responses (22–26), leading to antitumor effects in murine models (22–25) and in some patients with malignant diseases (26). In addition, we have included in our immunization strategy the administration of 4-1BB–specific mAb 2A (27), because this and other agonist mAb to the T-cell costimulatory molecule 4-1BB augment self-tumor antigen-specific immune responses elicited by DC hybrids (25). The potential clinical relevance of our strategy has been tested by using B16.F10 melanoma and MC38 colon adenocarcinoma tumor models, and assessing the safety level in the context of the tumor-bearing host.
Materials and Methods
Mice. Six- to 8-week-old female C57BL/6N (B6) mice were purchased from the National Cancer Institute (Bethesda, MD). They were maintained under specific pathogen-free conditions in the Roswell Park Cancer Institute (RPCI) animal core facility. Experiments were done under an animal protocol approved by the Institutional Animal Care and Use Committee at RPCI.
Cell lines. The murine EC line SVR (21) was purchased from the American Type Culture Collection. This cell line, the murine colon adenocarcinoma cell line MC38 (28), and the melanoma cell line B16.F10 (29), which are all B6 in origin, were maintained in RPMI 1640 supplemented with 10% heat-inactivated FCS, 0.1 mmol/L nonessential amino acids, 1 μmol/L sodium pyruvate, 2 mmol/L l-glutamine, streptomycin (100 μg/mL), penicillin (100 units/mL), gentamicin (50 μg/mL), Fungizone (0.5 μg/mL; Life Technologies), and 50 μmol/L 2-mercaptoethanol (Sigma-Aldrich; complete medium). All cell lines were maintained at 37°C in a humidified 5% CO2 atmosphere.
Antibodies. CD3-specific mAb 145-2C11 (30), CD4-specific mAb GK1.5 (31), CD8-specific mAb 2.43 (32), and 4-1BB–specific mAb 2A (27) have been previously described. mAbs were purified from ascitic fluid by sequential ammonium sulfate and caprylic acid precipitation (33). Mouse CD16/CD32–specific unconjugated mAb; mouse CD3-, CD4-, CD8-, CD11b-, CD11c-, CD31, CD54-, CD80-, CD86-, VEGFR-2, H-2 Kb–, H-2 I-Ab–, IFN-γ–specific FITC, phycoerythrin, or phycoerythrin-Cy5–conjugated mAb; and isotype-matched FITC, phycoerythrin, or phycoerythrin-Cy5–conjugated control antibody were purchased from BD Biosciences. Human/mouse CD3-specific rat mAb CD3-12 was purchased from Serotec, and mouse CD31-specific rat mAb MEC13.3 and rat isotype-matched control antibody were purchased from BD Biosciences.
Flow cytometry. Flow cytometry was done as described (34). Briefly, cells were incubated with CD16/CD32–specific mAb to block surface Fc receptors, washed with 0.5% bovine serum albumin (BSA)-PBS, and then incubated with fluorochrome-conjugated mAb. Intracellular staining was done by incubating cells with GolgiStop (0.7 μL/mL; BD Biosciences) for the final 6 h before harvest, fixing cells with 2% paraformaldehyde after staining of surface antigens, permeabilizing cells with 0.1% saponin/1% BSA-PBS, and incubating them with fluorochrome-conjugated mAb. Cells were then washed, fixed in 2% paraformaldehyde, and analyzed with a FACScan flow cytometer (BD Biosciences) using CellQuest.
DC generation. DC were generated as described (35). Briefly, bone marrow cells were harvested from the tibias and femurs of 6- to 8-week-old female B6 mice and then cultured at 37°C in a 5% CO2 atmosphere for 6 days in complete medium supplemented with granulocyte macrophage colony-stimulating factor (GM-CSF; 10 ng/mL). The culture medium was replenished every 2 to 3 days. On day 6, most of the nonadherent cells had acquired typical DC morphology and expressed moderate to high levels of MHC class I and II antigens, CD11b, CD11c, CD54, CD80, and CD86, as determined by flow cytometry (Supplementary Fig. S1).
Polyethylene glycol fusion. DC (1 × 107) were mixed with cultured and irradiated (7,000 rad) SVR EC (1 × 107). The cell mixture was centrifuged (6 min, 250 × g), the supernatant was aspirated, and 1 mL 50% polyethylene glycol 1500 (PEG1500; Roche), prewarmed to 37°C, was added dropwise over 1 min with continuous shaking while the tube was maintained at 37°C in a water bath. After an additional 1-min incubation with shaking, serum-free RPMI 1640 (10 mL) was added to the mixture at a rate of 1 mL/min. Cells were then centrifuged (6 min, 250 × g) and cultured in 24-well plates in complete medium at 37°C overnight. On the following day, cells were washed with PBS before injection or analysis.
Tumor and Matrigel implantation. For tumor implantation, mice were injected with B16.F10 (1 × 105 or 5 × 105 cells per mouse) or MC38 (5 × 105 cells per mouse) cells s.c. in the abdomen. For Matrigel implantation, each mouse was injected s.c. with Matrigel (500 μL; BD Biosciences) in one flank, and with Matrigel (500 μL) mixed with 2 × 105 MC38 cells in the opposite flank.
Treatment regimen. On the specified day after tumor or Matrigel implantation, mice were randomized immediately before treatment. Mice were anesthetized by i.p. injection of ketamine (2 mg)/xylazine (200 μg). The spleen was exteriorized via a small subcostal incision, and was injected with DC-EC hybrids (5 × 105/100 μL PBS). The surgical incision was then closed with 9-mm MikRon Autoclips (Becton Dickinson). 4-1BB–specific mAb or rat IgG control antibody (Sigma-Aldrich) was administered to mice i.p. (150 μg/1 mL PBS/injection) 1 and 3 days after DC-EC hybrid immunization, based on an established protocol (25) with minor modifications.
Cytotoxicity assay. Cytotoxicity assays were done as described (36). Briefly, splenocytes were harvested, pooled (three mice per group), and stimulated in vitro for 5 days with irradiated (7,000 rad) SVR cells and interleukin 2 (IL-2; 15 units/mL) before their use in a 6 h 51Cr-release assay. Assays were done in triplicates. The percentage of specific cytotoxicity was calculated using the formula: % cytotoxicity = (cpmexperimental − cpmspontaneous) / (cpmmaximum − cpmspontaneous) × 100. The spontaneous 51Cr release was <25% of maximum 51Cr release.
In vivo cell depletion. CD4+ or CD8+ cells were depleted in vivo by injecting mice i.v. with CD4-specific mAb GK1.5 or CD8-specific mAb 2.43 (500 μg/mouse), respectively, 1 day before immunization, as described (36). Splenocytes from mice injected with CD4- or CD8-specific mAb were analyzed by flow cytometry to confirm depletion.
Immunohistochemistry. Samples were placed in zinc fixative (BD Biosciences) overnight. They were then dehydrated, paraffin-embedded, sectioned at 5-μm thickness, placed on charged slides, and dried. Endogenous peroxidase was quenched with aqueous 3% H2O2. Slides were then processed in the DAKO Autostainer as follows: 0.05% Tween 20 in PBS (PBS/T) wash, followed by blocking with 0.03% casein in PBS/T, and incubation with primary mAb or rat IgG control antibody (10 μg/mL). Following a 0.05% PBS/T wash, slides were incubated sequentially with biotinylated rat IgG-specific xenoantibodies (BD Biosciences), streptavidin complex (Invitrogen), and 3,3′-diaminobenzidine (DAKO). The slides were counterstained with hematoxylin. CD31+ microvessels were counted from high-power fields chosen in the areas of highest vascularization, as described (37).
Light and confocal microscopy. Light photomicrographs of cells in culture plates were obtained via a Zeiss Axiovert 25 inverted microscope, whereas light photomicrographs of stained tissues were obtained as still images from a Hitachi HV-C20 3-CCD color camera coupled to an Olympus BX40 microscope. Images were analyzed using Image-Pro Plus (Media Cybernetics). For confocal microscopy, cells labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE) and/or tetramethyl rhodamine isothiocyanate (TRITC; Molecular Probes) were harvested from culture plates, washed, resuspended in PBS, and spotted onto a glass coverslip. Cells were then rinsed with PBS, fixed with 2% paraformaldehyde, mounted with Polymount (Polysciences), and analyzed using a Bio-Rad MRC-600 laser scanning confocal microscope.
Hemoglobin quantitation. The hemoglobin concentration of Matrigel plugs was determined by Drabkin's method (38). Briefly, Matrigel plugs were individually weighed, resuspended in Drabkin's reagent (Sigma-Aldrich), and homogenized using a Dounce homogenizer (Kimble). The hemoglobin concentration was determined by measuring optical absorbance at 540 nm using a Bio-Rad SmartSpec 3000 spectrophotometer, and comparing the values to a standard curve generated from serially diluted hemoglobin standards (Sigma-Aldrich).
Assessment of side effects. At the indicated time points, mice were weighed and then were injected i.p. with heparin (50 units; APP). After 10 min, ∼40 μL of blood were obtained from each mouse via the retro-orbital venous plexus and resuspended in 200 μL Haema-Line 2 reagent (Serono-Baker) in an EDTA-containing Microtainer tube (Becton Dickinson). Samples were submitted to the RPCI Hematology Laboratory for complete and differential blood cell counts, using the ADVIA 120 Hematology System (Bayer).
Statistical analysis. Statistical analysis was done using SigmaPlot (Systat). Differences between groups were analyzed by the two-tailed, unpaired Student's t test, with P ≤ 0.05 considered significant.
Results
Generation of DC-EC hybrids by PEG fusion. DC, prepared by a 6-day in vitro culture of bone marrow cells from naïve B6 mice in the presence of GM-CSF (35), expressed high levels of H-2 Kb, CD11b, CD11c, and CD54 and moderate levels of H-2 I-Ab, CD80, and CD86, as determined by flow cytometry (Supplementary Fig. S1). These cells were mixed with SVR cells (21), which express CD31, VEGFR-2, and other antigens associated with proliferative EC, based on flow cytometric analysis (data not shown). The cell mixture was pelleted and incubated in the presence of PEG1500, washed, and cultured overnight. Using phycoerythrin-conjugated CD11b-specific mAb to identify the DC population and CFSE-labeled EC, ∼25% of the total cell population was both CD11b+ and CFSE+ after PEG1500 fusion and overnight incubation (Fig. 1A). The CD11b+CFSE+ cells, but not the CD11b+CFSE− or CD11b−CFSE+ cells, expressed both CD11c and CD31 (Supplementary Fig. S2A). Additionally, the CD11b+CFSE+ cells expressed VEGFR-2 (Supplementary Fig. S2B), an EC-derived antigen not found on the parental DC (Supplementary Fig. S1A). Across the DC/EC cell ratios of 1:1, 2:1, and 3:1 and total cell numbers of 1 × 107 or 2 × 107, 15% to 25% of the total cell population were consistently CD11b+CFSE+ (data not shown). The generation of DC-EC hybrids was proven by two additional lines of evidence. First, large and multinucleated cells were identified by light microscopy after PEG1500 fusion of DC and EC (Fig. 1B), suggesting that membrane fusion had occurred. Second, hybrids generated by PEG1500 fusion of TRITC-labeled DC and CFSE-labeled EC expressed both TRITC-labeled and CFSE-labeled components by confocal microscopy (Fig. 1C).
Phenotypic and morphologic analysis of DC-EC hybrids. Six-day bone marrow–derived DC (1 × 107) were mixed with SVR EC (1 × 107) and incubated with PEG1500. As controls, DC and EC were mixed (1:1 ratio) but not incubated with PEG1500. Cells were pelleted, washed, and cultured overnight before analysis. For flow cytometry (A), EC were labeled with 2 μmol/L CFSE before PEG1500 fusion, and after overnight incubation, the harvested cells were labeled with phycoerythrin-conjugated CD11b-specific mAb. For light microscopy (B), PEG1500 fusion was done with unlabeled cells, and photomicrographs of low-power (base images) and high-power (insets) fields are shown. For confocal microscopy (C), DC were labeled with 2 μmol/L TRITC and EC were labeled with 2 μmol/L CFSE before PEG1500 fusion. Results are representative of three independent experiments.
Phenotypic and morphologic analysis of DC-EC hybrids. Six-day bone marrow–derived DC (1 × 107) were mixed with SVR EC (1 × 107) and incubated with PEG1500. As controls, DC and EC were mixed (1:1 ratio) but not incubated with PEG1500. Cells were pelleted, washed, and cultured overnight before analysis. For flow cytometry (A), EC were labeled with 2 μmol/L CFSE before PEG1500 fusion, and after overnight incubation, the harvested cells were labeled with phycoerythrin-conjugated CD11b-specific mAb. For light microscopy (B), PEG1500 fusion was done with unlabeled cells, and photomicrographs of low-power (base images) and high-power (insets) fields are shown. For confocal microscopy (C), DC were labeled with 2 μmol/L TRITC and EC were labeled with 2 μmol/L CFSE before PEG1500 fusion. Results are representative of three independent experiments.
Abrogation of growth of B16.F10 melanoma tumors after immunization with DC-EC hybrids and 4-1BB–specific mAb. Mice bearing 3-day s.c. B16.F10 melanoma tumors (1 × 105 cells per mouse) were injected with DC-EC hybrids, followed 1 and 3 days later by administration of 4-1BB–specific mAb. Remarkably, all the treated mice exhibited no tumor growth and survived over the course of 60 days after tumor inoculation (Fig. 2A and B). To determine if this immunization strategy can control B16.F10 tumors in a more aggressive setting, 3-day tumors were established after s.c. injection of 5 × 105 cells per mouse. Treatment with DC-EC hybrids and 4-1BB–specific mAb was done as in the previous experiment. Similar to that experiment, all the treated mice exhibited no tumor growth and survived over the course of >65 days (Fig. 2C and D). In each study, both DC-EC hybrids and 4-1BB–specific mAb were required, as immunization with DC-EC hybrids and control antibody (Fig. 2C and D) or administration of 4-1BB–specific mAb alone (Fig. 2A–D) neither inhibited the growth of B16.F10 tumors nor prolonged the survival of tumor-bearing mice, when compared with untreated or control antibody-treated mice (Fig. 2A–D). Enhancement of the antitumor response by 4-1BB–specific mAb required the provision of EC antigens in the form of DC-EC hybrids, because immunization with a DC + EC mixture and 4-1BB–specific mAb also did not significantly (P > 0.05) inhibit the growth of B16.F10 tumors or prolong the survival of tumor-bearing mice (Fig. 2A–D).
Antitumor effects against 3 d B16.F10 melanoma in mice immunized with DC-EC hybrids and 4-1BB–specific mAb. B6 mice (n = 5 per group) were injected with B16.F10 cells s.c. (1 × 105, A and B; 5 × 105, C and D) on day 0, and immunized with DC-EC hybrids (5 × 105 intrasplenically, day 3) and 4-1BB–specific mAb (150 μg i.p., days 4 and 6; •). As controls in (A) and (B), additional groups of B16.F10 tumor-bearing B6 mice (n = 4–5 per group) were untreated (▿); treated with 4-1BB–specific mAb only (150 μg i.p., days 4 and 6; ▾); or immunized with a DC + EC mixture (1:1 ratio, 5 × 105 total cells intrasplenically, day 3) and 4-1BB–specific mAb (150 μg i.p., days 4 and 6; ○). As controls in (C) and (D), additional groups of B16.F10 tumor-bearing B6 mice (n = 4–5 per group) were treated with 4-1BB–specific mAb (150 μg i.p., days 4 and 6; ▾) or rat IgG (150 μg i.p., days 4 and 6; ▿) only; immunized with DC-EC hybrids (5 × 105 intrasplenically, day 3) and rat IgG (150 μg i.p., days 4 and 6; □); or immunized with a DC + EC mixture (1:1 ratio, 5 × 105 total cells intrasplenically, day 3) and 4-1BB–specific mAb (150 μg i.p., days 4 and 6; ○). Mice were monitored for tumor growth (A and C) and survival (B and D). Individual tumor sizes were estimated as the product of the maximum and perpendicular diameters. Points, mean tumor size; bars, SE. Results are representative of three independent experiments. *, P < 0.05 versus all other groups.
Antitumor effects against 3 d B16.F10 melanoma in mice immunized with DC-EC hybrids and 4-1BB–specific mAb. B6 mice (n = 5 per group) were injected with B16.F10 cells s.c. (1 × 105, A and B; 5 × 105, C and D) on day 0, and immunized with DC-EC hybrids (5 × 105 intrasplenically, day 3) and 4-1BB–specific mAb (150 μg i.p., days 4 and 6; •). As controls in (A) and (B), additional groups of B16.F10 tumor-bearing B6 mice (n = 4–5 per group) were untreated (▿); treated with 4-1BB–specific mAb only (150 μg i.p., days 4 and 6; ▾); or immunized with a DC + EC mixture (1:1 ratio, 5 × 105 total cells intrasplenically, day 3) and 4-1BB–specific mAb (150 μg i.p., days 4 and 6; ○). As controls in (C) and (D), additional groups of B16.F10 tumor-bearing B6 mice (n = 4–5 per group) were treated with 4-1BB–specific mAb (150 μg i.p., days 4 and 6; ▾) or rat IgG (150 μg i.p., days 4 and 6; ▿) only; immunized with DC-EC hybrids (5 × 105 intrasplenically, day 3) and rat IgG (150 μg i.p., days 4 and 6; □); or immunized with a DC + EC mixture (1:1 ratio, 5 × 105 total cells intrasplenically, day 3) and 4-1BB–specific mAb (150 μg i.p., days 4 and 6; ○). Mice were monitored for tumor growth (A and C) and survival (B and D). Individual tumor sizes were estimated as the product of the maximum and perpendicular diameters. Points, mean tumor size; bars, SE. Results are representative of three independent experiments. *, P < 0.05 versus all other groups.
Inhibition of tumor growth in MC38 tumor-bearing mice after immunization with DC-EC hybrids and 4-1BB–specific mAb. To determine if our immunization strategy effectively inhibited the growth of an unrelated tumor, mice bearing 1-day s.c. MC38 colon adenocarcinoma tumors (5 × 105 cells per mouse) were injected with DC-EC hybrids, followed 1 and 3 days later by administration of 4-1BB–specific mAb. The growth of MC38 tumors was significantly (P < 0.05) inhibited in mice immunized with DC-EC hybrids and 4-1BB–specific mAb, compared with untreated mice (Fig. 3A and B). In agreement with the results obtained in the B16.F10 tumor model, both components of this immunization regimen were required, because immunization with DC-EC hybrids or administration of 4-1BB–specific mAb alone did not significantly (P > 0.05) inhibit MC38 tumor growth (Fig. 3A). Enhancement of the antitumor response by 4-1BB–specific mAb required immunization with DC-EC hybrids, because immunization with DC only, EC only, or a DC + EC mixture followed by 4-1BB–specific mAb administration did not significantly (P > 0.05) inhibit MC38 tumor growth, compared with administration of 4-1BB–specific mAb alone (Fig. 3B). In addition, EC and DC that were pretreated with PEG1500 separately, washed, and then mixed before injection elicited no antitumor effects even when administered with 4-1BB–specific mAb (data not shown).
Antitumor effects against MC38 tumors in mice immunized with DC-EC hybrids and 4-1BB–specific mAb. A and B, B6 mice (n = 5 per group) were injected with MC38 cells (5 × 105 s.c.) on day 0, and immunized with DC-EC hybrids (5 × 105 intrasplenically, day 1) and 4-1BB–specific mAb (150 μg i.p., days 2,4; •). As controls in (A), additional groups of MC38 tumor-bearing B6 mice (n = 5 per group) were untreated (▿) or treated with only DC-EC hybrids (5 × 105 intrasplenically, day 1; □) or 4-1BB–specific mAb (150 μg i.p., days 2,4; ▾). As controls in (B), additional groups of MC38 tumor-bearing B6 mice (n = 5 per group) were untreated (▿); treated with 4-1BB–specific mAb only (150 μg i.p., days 2,4; ▾); or immunized with EC (5 × 105 intrasplenically, day 1; ◊), DC (5 × 105 intrasplenically, day 1; ⧫), or a DC + EC mixture (1:1 ratio, 5 × 105 total cells intrasplenically, day 1; ○) and 4-1BB–specific mAb (150 μg i.p., days 2 and 4). Mice were monitored for tumor growth, and individual tumor sizes were estimated as the product of the maximum and perpendicular diameters. Points, mean tumor sizes; bars, SE. Results are representative of three independent experiments. *, P < 0.05 versus all other immunization groups. C and D, B6 mice were injected with MC38 cells (5 × 105 s.c.) on day 0. Eight to 12 h later, mice were randomized into groups (n = 5 per group), and were depleted of CD4 (C; ▪) or CD8 (D; □) cells with CD4-specific mAb GK1.5 or CD8-specific mAb 2.43, respectively (500 μg i.v.). The mice were immunized with DC-EC hybrids on day 1 and 4-1BB–specific mAb on days 2 and 4 after tumor inoculation. As controls, additional groups of mice were not depleted of either CD4 or CD8 cells but were immunized with DC-EC hybrids and 4-1BB–specific mAb (•), or were not depleted of either CD4 or CD8 cells and were untreated (▿). Mice were monitored for tumor growth, and individual tumor sizes were estimated as the product of the maximum and perpendicular diameters. Points, mean tumor size; bars, SE. Results are representative of three independent experiments. *, P < 0.05 versus nondepleted, immunized mice.
Antitumor effects against MC38 tumors in mice immunized with DC-EC hybrids and 4-1BB–specific mAb. A and B, B6 mice (n = 5 per group) were injected with MC38 cells (5 × 105 s.c.) on day 0, and immunized with DC-EC hybrids (5 × 105 intrasplenically, day 1) and 4-1BB–specific mAb (150 μg i.p., days 2,4; •). As controls in (A), additional groups of MC38 tumor-bearing B6 mice (n = 5 per group) were untreated (▿) or treated with only DC-EC hybrids (5 × 105 intrasplenically, day 1; □) or 4-1BB–specific mAb (150 μg i.p., days 2,4; ▾). As controls in (B), additional groups of MC38 tumor-bearing B6 mice (n = 5 per group) were untreated (▿); treated with 4-1BB–specific mAb only (150 μg i.p., days 2,4; ▾); or immunized with EC (5 × 105 intrasplenically, day 1; ◊), DC (5 × 105 intrasplenically, day 1; ⧫), or a DC + EC mixture (1:1 ratio, 5 × 105 total cells intrasplenically, day 1; ○) and 4-1BB–specific mAb (150 μg i.p., days 2 and 4). Mice were monitored for tumor growth, and individual tumor sizes were estimated as the product of the maximum and perpendicular diameters. Points, mean tumor sizes; bars, SE. Results are representative of three independent experiments. *, P < 0.05 versus all other immunization groups. C and D, B6 mice were injected with MC38 cells (5 × 105 s.c.) on day 0. Eight to 12 h later, mice were randomized into groups (n = 5 per group), and were depleted of CD4 (C; ▪) or CD8 (D; □) cells with CD4-specific mAb GK1.5 or CD8-specific mAb 2.43, respectively (500 μg i.v.). The mice were immunized with DC-EC hybrids on day 1 and 4-1BB–specific mAb on days 2 and 4 after tumor inoculation. As controls, additional groups of mice were not depleted of either CD4 or CD8 cells but were immunized with DC-EC hybrids and 4-1BB–specific mAb (•), or were not depleted of either CD4 or CD8 cells and were untreated (▿). Mice were monitored for tumor growth, and individual tumor sizes were estimated as the product of the maximum and perpendicular diameters. Points, mean tumor size; bars, SE. Results are representative of three independent experiments. *, P < 0.05 versus nondepleted, immunized mice.
In a parallel experiment, MC38 tumor-bearing mice were depleted of CD4+ or CD8+ cells 24 h before immunization with DC-EC hybrids and 4-1BB–specific mAb. Flow cytometric analysis of splenocytes from mice injected with CD4-specific mAb GK1.5 or CD8-specific mAb 2.43 consistently revealed >95% depletion of CD4+ T cells or CD8+ T cells, respectively, at 24 h to 7 days after mAb injection (data not shown). The induction of an immune response by DC-EC hybrids and 4-1BB–specific mAb to effectively inhibit MC38 tumor growth was dependent on both CD4+ and CD8+ cells, because depletion of either cell subset before immunization significantly (P < 0.05) limited the ability of immunized mice to control tumor growth (Fig. 3C and D).
Specificity of CD8+ and CD4+ T-cell responses after immunization with DC-EC hybrids and 4-1BB–specific mAb. To determine the specificity of the CD8+ and CD4+ T-cell responses, splenocytes from tumor-bearing mice injected with DC-EC hybrids and 4-1BB–specific mAb were cocultured with MC38, B16.F10, and SVR cells as syngeneic stimulator cells. Following a 24-h incubation, an immune response was detected only in CD8+ T cells stimulated with SVR cells, as indicated by the detection of intracellular IFN-γ staining (Fig. 4A). Additional proof of the specificity of the immune response was provided by the lack of intracellular IFN-γ staining in CD8+ T cells from tumor-bearing, control-immunized mice and tumor-bearing, untreated mice (Fig. 4A). This lack of response does not reflect functional abnormalities in the CD8+ T-cell subset, because a similar percentage of intracellular IFN-γ+ cells was detected among CD8+ T cells from all the groups of mice after stimulation with CD3-specific mAb (data not shown). In parallel, splenic CD4+ T cells from mice immunized with DC-EC hybrids and 4-1BB–specific mAb exhibited an IFN-γ response to SVR cells, but not to MC38 and B16.F10 cells (Fig. 4B). In tumor-bearing mice immunized with DC-EC hybrids and 4-1BB–specific mAb, IFN-γ responses to SVR cells by CD8+ and CD4+ T cells were detected at both early (day 6) and late (day 25) time points. These results are consistent with the role of these immune responses in the control of MC38 tumor growth during this time span.
EC-specific reactivity and cytotoxicity by splenocytes from MC38 tumor-bearing mice immunized with DC-EC hybrids and 4-1BB–specific mAb. A and B, MC38 tumor-bearing B6 mice (n = 6 per group) were immunized with DC-EC hybrids and 4-1BB–specific mAb (•), or were untreated (▿), administered 4-1BB–specific mAb only (▾), or immunized with DC + EC mixture (1:1 ratio) and 4-1BB–specific mAb (○), as described in Fig. 3. Mice were sacrificed on the indicated days, and pooled splenocytes (n = 2 per group) were cocultured with the indicated stimulator cells at a 5:1 (effector/stimulator) ratio for 24 h, with the addition of GolgiStop in the final 6 h. Following incubation, IFN-γ production by CD8+ (A) and CD4+ (B) T cells was enumerated by flow cytometric analysis using surface staining with FITC-conjugated CD3-specific mAb and phycoerythrin-Cy5–conjugated CD4- or CD8-specific mAb, and intracellular staining with phycoerythrin-conjugated IFN-γ–specific mAb after permeabilization with 0.1% saponin/1% BSA-PBS. Gates for CD3-, CD4/CD8–, and IFN-γ–positive cells were assigned based on FITC-, phycoerythrin-Cy5–, and phycoerythrin-conjugated isotype control antibody staining, respectively. The values are expressed as % IFN-γ+CD8+ T cells among total CD8+ T cells (A), or as % IFN-γ+CD4+ T cells among total CD4+ T cells (B). All values are subtracted for values obtained in medium-only controls. Results are representative of two independent experiments. C, MC38 tumor-bearing B6 mice were immunized with DC-EC hybrids and 4-1BB–specific mAb (•), or were untreated (▿), administered 4-1BB–specific mAb only (▾), or immunized with DC + EC mixture (1:1 ratio) and 4-1BB–specific mAb (○), as described in Fig. 3. Mice were sacrificed on day 6, and pooled splenocytes (n = 3 per group) were cocultured with irradiated SVR cells (5:1 effector/stimulator ratio) and IL-2 (15 units/mL) for 5 d. The in vitro–stimulated splenocytes were then cocultured with the indicated 51Cr-labeled syngeneic cells in a 6-h 51Cr-release cytotoxicity assay. Points, mean percentage lysis values of triplicate wells; bars, SE. Results are representative of three independent experiments. *, P < 0.05 versus all other immunization groups.
EC-specific reactivity and cytotoxicity by splenocytes from MC38 tumor-bearing mice immunized with DC-EC hybrids and 4-1BB–specific mAb. A and B, MC38 tumor-bearing B6 mice (n = 6 per group) were immunized with DC-EC hybrids and 4-1BB–specific mAb (•), or were untreated (▿), administered 4-1BB–specific mAb only (▾), or immunized with DC + EC mixture (1:1 ratio) and 4-1BB–specific mAb (○), as described in Fig. 3. Mice were sacrificed on the indicated days, and pooled splenocytes (n = 2 per group) were cocultured with the indicated stimulator cells at a 5:1 (effector/stimulator) ratio for 24 h, with the addition of GolgiStop in the final 6 h. Following incubation, IFN-γ production by CD8+ (A) and CD4+ (B) T cells was enumerated by flow cytometric analysis using surface staining with FITC-conjugated CD3-specific mAb and phycoerythrin-Cy5–conjugated CD4- or CD8-specific mAb, and intracellular staining with phycoerythrin-conjugated IFN-γ–specific mAb after permeabilization with 0.1% saponin/1% BSA-PBS. Gates for CD3-, CD4/CD8–, and IFN-γ–positive cells were assigned based on FITC-, phycoerythrin-Cy5–, and phycoerythrin-conjugated isotype control antibody staining, respectively. The values are expressed as % IFN-γ+CD8+ T cells among total CD8+ T cells (A), or as % IFN-γ+CD4+ T cells among total CD4+ T cells (B). All values are subtracted for values obtained in medium-only controls. Results are representative of two independent experiments. C, MC38 tumor-bearing B6 mice were immunized with DC-EC hybrids and 4-1BB–specific mAb (•), or were untreated (▿), administered 4-1BB–specific mAb only (▾), or immunized with DC + EC mixture (1:1 ratio) and 4-1BB–specific mAb (○), as described in Fig. 3. Mice were sacrificed on day 6, and pooled splenocytes (n = 3 per group) were cocultured with irradiated SVR cells (5:1 effector/stimulator ratio) and IL-2 (15 units/mL) for 5 d. The in vitro–stimulated splenocytes were then cocultured with the indicated 51Cr-labeled syngeneic cells in a 6-h 51Cr-release cytotoxicity assay. Points, mean percentage lysis values of triplicate wells; bars, SE. Results are representative of three independent experiments. *, P < 0.05 versus all other immunization groups.
In addition, after a 5-day in vitro stimulation with irradiated SVR cells, splenocytes from tumor-bearing mice injected with DC-EC hybrids and 4-1BB–specific mAb lysed SVR cells in a standard 6-h 51Cr-release cytotoxicity assay, but lysed neither MC38 nor B16.F10 cells (Fig. 4C). This lysis effect was specific, because in vitro–stimulated splenocytes from tumor-bearing mice that were untreated, treated with 4-1BB–specific mAb only, or immunized with DC + EC mixture and 4-1BB–specific mAb lysed none of the target cells (Fig. 4C).
Collectively, our findings indicate that immunization with DC-EC hybrids and 4-1BB–specific mAb induces EC-specific T-cell responses that do not cross-react with other syngeneic cells.
Inhibition of tumor angiogenesis after immunization with DC-EC hybrids and 4-1BB–specific mAb. To determine if our immunization strategy inhibited tumor angiogenesis, naïve B6 mice were injected with MC38-containing Matrigel plugs, and were immunized with DC-EC hybrids on day 1, followed 1 and 3 days later by administration of 4-1BB–specific mAb. At day 13 postimplantation, the immunized mice exhibited a markedly lower degree of tumor-induced angiogenesis than untreated mice, as determined by the significantly (P = 0.05) lower hemoglobin concentration (Fig. 5A) in their MC38-containing Matrigel plugs. At day 20 postimplantation, bulk MC38 tumor outgrowth was observed in MC38-containing Matrigel plugs from untreated mice, but not in those from immunized mice (Fig. 5B). Immunohistochemical analysis of day 20 Matrigel plugs revealed a uniformly high density of blood vessels surrounded by a layer of CD31+ cells in plugs harvested from untreated mice (Fig. 5C). In contrast, Matrigel plugs from immunized mice exhibited fewer intratumoral CD31+ cells, which formed a significantly (P < 0.05) lower number of CD31+ microvessels (Fig. 5C). In the same Matrigel plugs from immunized mice, immunohistochemical analysis with a CD3-specific mAb to identify infiltrating T cells revealed a significantly (P < 0.05) higher number of CD3+ cells, compared with plugs from untreated mice (Fig. 5D). These results indicate that growth inhibition of MC38 tumors in mice immunized with DC-EC hybrids and 4-1BB–specific mAb is due to the induction of antiangiogenic immunity.
Inhibition of tumor angiogenesis after immunization with DC-EC hybrids and 4-1BB–specific mAb. B6 mice (n = 5–7 per group) were implanted s.c. with 500 μL Matrigel + 2 × 105 MC38 cells, or 500 μL Matrigel alone, and immunized with DC-EC hybrids on day 1 and 4-1BB–specific mAb on days 2 and 4, as described in Fig. 3. As controls, additional groups of mice received the same Matrigel plugs but were untreated. Matrigel plugs were harvested on day 13 (A) or day 20 (B–D). A, hemoglobin content ([Hb]) in MC38-containing Matrigel plugs harvested from untreated and treated mice on day 13 was determined by Drabkin's spectrophotometric method. Columns, mean hemoglobin concentration per gram of Matrigel; bars, SE. B, MC38-containing Matrigel plugs harvested from untreated and treated mice on day 20, in comparison with non–tumor cell-containing Matrigel plugs. C, immunohistochemical staining of tumor vasculature in zinc-fixed, MC38-containing Matrigel plugs from untreated (a) and treated (b) mice on day 20, using CD31-specific mAb and hematoxylin counterstain. Quantitation of microvessel density (c) was done by counting CD31+ microvessels in 5 high-power fields (h.p.f.; area = 0.30 mm2), chosen at the areas of highest vascularization from four to five samples per group. Columns, microvessel density expressed as mean counts per high-power field; bars, SE. D, low (a–b base images) and high (a–b insets) power fields of immunohistochemical staining of zinc-fixed, MC38-containing Matrigel plugs from untreated (a) and treated (b) mice on day 20, using CD3-specific mAb and hematoxylin counterstain. Quantitation of T cells (c) was done by counting CD3+ cells (white arrowheads, a–b insets) in six high-power fields, chosen from three samples per group for their proximity to the tumor vasculature (thin lines, a base image). Columns, T-cell numbers expressed as mean counts per high-power field; bars, SE. *, P ≤ 0.05 versus MC38-containing Matrigel plugs from untreated mice. Bars, 1 cm (B), 100 μm (C–D base images), and 10 μm (D insets).
Inhibition of tumor angiogenesis after immunization with DC-EC hybrids and 4-1BB–specific mAb. B6 mice (n = 5–7 per group) were implanted s.c. with 500 μL Matrigel + 2 × 105 MC38 cells, or 500 μL Matrigel alone, and immunized with DC-EC hybrids on day 1 and 4-1BB–specific mAb on days 2 and 4, as described in Fig. 3. As controls, additional groups of mice received the same Matrigel plugs but were untreated. Matrigel plugs were harvested on day 13 (A) or day 20 (B–D). A, hemoglobin content ([Hb]) in MC38-containing Matrigel plugs harvested from untreated and treated mice on day 13 was determined by Drabkin's spectrophotometric method. Columns, mean hemoglobin concentration per gram of Matrigel; bars, SE. B, MC38-containing Matrigel plugs harvested from untreated and treated mice on day 20, in comparison with non–tumor cell-containing Matrigel plugs. C, immunohistochemical staining of tumor vasculature in zinc-fixed, MC38-containing Matrigel plugs from untreated (a) and treated (b) mice on day 20, using CD31-specific mAb and hematoxylin counterstain. Quantitation of microvessel density (c) was done by counting CD31+ microvessels in 5 high-power fields (h.p.f.; area = 0.30 mm2), chosen at the areas of highest vascularization from four to five samples per group. Columns, microvessel density expressed as mean counts per high-power field; bars, SE. D, low (a–b base images) and high (a–b insets) power fields of immunohistochemical staining of zinc-fixed, MC38-containing Matrigel plugs from untreated (a) and treated (b) mice on day 20, using CD3-specific mAb and hematoxylin counterstain. Quantitation of T cells (c) was done by counting CD3+ cells (white arrowheads, a–b insets) in six high-power fields, chosen from three samples per group for their proximity to the tumor vasculature (thin lines, a base image). Columns, T-cell numbers expressed as mean counts per high-power field; bars, SE. *, P ≤ 0.05 versus MC38-containing Matrigel plugs from untreated mice. Bars, 1 cm (B), 100 μm (C–D base images), and 10 μm (D insets).
Lack of significant side effects after immunization with DC-EC hybrids and 4-1BB–specific mAb. The antiangiogenic immune response elicited by our immunization strategy caused no weight loss, because no differences in body weight were detected among naïve mice, immunized mice, and immunized MC38 tumor-bearing mice at days 13, 27, and 154 postimmunization (Fig. 6A). During this time frame, the mice in each group gained weight at the same rate. In addition, mice that were immunized with DC-EC hybrids and 4-1BB–specific mAb exhibited no overt delays in wound healing. As with control-immunized mice that have also undergone surgery (DC + mAb, EC + mAb, or DC + EC mixture + mAb), the subcostal surgical incision site was completely healed by 7 days postsurgery (data not shown).
Limited and transient side effects after immunization with DC-EC hybrids and 4-1BB–specific mAb. B6 mice (n = 5 per group) were immunized with DC-EC hybrids and 4-1BB–specific mAb in the absence () or presence (
) of MC38 tumors, as described in Fig. 3. As a control, an additional group of mice was neither injected with MC38 cells nor treated (▪). Mice were weighed on days −14, 13, 27, and 154 relative to the start of immunization (A). Immediately after weighing on days −14, 13, and 27, the mice were injected with heparin before obtaining blood from the retro-orbital venous plexus. Blood samples were analyzed to determine the absolute counts of WBC (B), platelets (C), and RBC (D). Columns, mean; bars, SE. Representative of three independent experiments. *, P < 0.05 for the indicated comparisons.
Limited and transient side effects after immunization with DC-EC hybrids and 4-1BB–specific mAb. B6 mice (n = 5 per group) were immunized with DC-EC hybrids and 4-1BB–specific mAb in the absence () or presence (
) of MC38 tumors, as described in Fig. 3. As a control, an additional group of mice was neither injected with MC38 cells nor treated (▪). Mice were weighed on days −14, 13, 27, and 154 relative to the start of immunization (A). Immediately after weighing on days −14, 13, and 27, the mice were injected with heparin before obtaining blood from the retro-orbital venous plexus. Blood samples were analyzed to determine the absolute counts of WBC (B), platelets (C), and RBC (D). Columns, mean; bars, SE. Representative of three independent experiments. *, P < 0.05 for the indicated comparisons.
In the hematopoietic compartment, no differences in RBC counts were observed in immunized mice (Fig. 6D), corroborating our observations that there was no major impairment of wound healing and no abnormal bleeding. On the other hand, WBC and platelet counts were significantly (P < 0.05) decreased at day 13 postimmunization in immunized tumor-bearing mice (Fig. 6B and C), and the platelet count was significantly (P < 0.05) decreased at the same time point in immunized non–tumor-bearing mice (Fig. 6C). Although statistically significant, these effects were not associated with major toxicities (e.g., increased rates of infection or bleeding abnormalities) and were transient, as the absolute WBC and platelet counts had normalized to the same levels as those found at baseline (day −14) and in day-matched, untreated mice by day 27 postimmunization.
Discussion
This study has shown that tumor growth can be controlled in therapeutic mouse models by antiangiogenic immunity elicited by immunization with hybrids of DC and syngeneic EC. Our immunization strategy differs in three aspects from previously described strategies that control tumor growth via induction of antiangiogenic immunity. First, we have used DC-EC hybrids as immunogens, because somatic fusion is simple, practical, and effective in loading DC with antigens, particularly when the source of immunizing cellular antigens is not well characterized. Second, we have selected syngeneic SVR EC as the fusion partner, because their high proliferative potential enables them to resemble more closely tumor-associated EC than resting EC in their phenotype, including their expression of several molecules important in angiogenic pathways. Furthermore, from a practical viewpoint, these cells can be easily obtained in large numbers, thus facilitating the preparation of hybrid cells. Third, we have combined DC-EC hybrid immunization with the administration of 4-1BB–specific mAb 2A, an agonist of the 4-1BB costimulatory molecule on activated T cells (27), because this and other agonist mAb to 4-1BB have been shown to enhance the cellular immune response to self-antigens (25, 27, 39–41).
Immunization of mice with DC-EC hybrids and 4-1BB–specific mAb elicited antiangiogenic immunity, because splenic T cells obtained from the immunized mice produced IFN-γ in response to syngeneic EC, and lysed EC. This immune response seems to be restricted to EC, because no cellular immunity to other syngeneic cells was detected. The induction of antiangiogenic immunity to control tumor growth was both CD4+ and CD8+ cell dependent, because depletion of either cell subset before immunization limited the ability of immunized mice to control tumor growth. The immune response elicited by our immunization strategy shares some features with those described in the literature, but also displays distinct characteristics. The crucial role played by CD8+ T cells in our system parallels the results described by Hicklin (8), Reisfeld (9), and Gilboa (12), and their collaborators, but is at variance with those described by Wei et al. (10, 13, 20), who ascribed the antitumor effects in their system to humoral immunity. The requirement for CD4+ T cells in the antitumor effects seems to depend on the strategy used, because Wei and his collaborators (10, 13, 20) showed that their antitumor effects were CD4-dependent, whereas Hicklin (8) and Reisfeld (9) and their collaborators showed that their antitumor effects were largely CD4 independent. Our system differs from the published studies, because the antitumor effects seem to be mediated not only by CD4+ T cells, but also by CD8+ T cells. The induction of a subset of CD4+ T cells that recognize and respond to syngeneic EC is likely to enhance the antitumor effects mediated by EC-specific CD8+ T cells, because CD4+ T cells play an important role in the maintenance of effective and durable CTL responses (42–44), and in certain cases are required for effective antitumor CTL immunity (42, 45, 46).
The lack of detection of a tumor antigen-specific cellular immune response in tumor-bearing mice after administration of 4-1BB–specific mAb is in agreement with the results described by Shu (25) and Chen (27) and their collaborators. These results altogether are at variance with those of Chen and his collaborators in the P815 mastocytoma and Ag104A sarcoma tumor models (39). The conflicting findings may reflect differences in the immunogenicity of the tumor antigens expressed by the various types of tumors tested, although a controlled experiment is required to prove this possibility.
Both DC-EC hybrids and 4-1BB–specific mAb were required to elicit antiangiogenic immunity to control tumor growth, because no significant EC-specific cellular immune response or tumor growth inhibition was observed in mice immunized with (a) DC-EC hybrids in the absence of 4-1BB–specific mAb and (b) DC, EC, or a DC + EC mixture in the presence of 4-1BB–specific mAb. The requirement for 4-1BB–specific mAb in our strategy is consistent with the ability of 4-1BB–specific mAb to enhance cellular immune responses to other self-antigens (25, 27, 39–41). The inability of DC hybrids to elicit self-antigen–specific immune responses for tumor control in the absence of 4-1BB–specific mAb parallels the findings of Shu and his collaborators (25), who noted that DC-tumor cell hybrids were unable to elicit tumor antigen–specific immune responses for tumor control in the absence of 4-1BB–specific mAb. The requirement for DC-EC hybrids in our strategy suggests that DC-EC hybrids were uniquely able to present EC antigens in an immunogenic context. In addition, the contaminating parental DC and EC populations are not likely to play a major role in the induction of antiangiogenic immunity by DC-EC hybrids, because mixtures of DC and EC were not effective in eliciting antiangiogenic immunity.
The fine specificity of EC antigens recognized by the CD4+ and CD8+ T cells elicited by our immunization strategy was not investigated, as it was beyond the scope of the current study. Nevertheless, one can anticipate that the CD4+ and CD8+ T-cell responses are likely polyclonal in nature, both with respect to their recognition of different antigens as well as different epitopes on any given antigen. The recognition of multiple targets has the advantage that it may overcome limitations associated with the selection of antigen-loss variants, after antigen-specific targeting of a population of cells that are heterogeneous in their antigenic profiles. This phenomenon has been frequently reported for strategies that target tumor cells (2), and less frequently for strategies targeting the tumor vasculature. Although EC are generally considered more genetically stable than tumor cells, epigenetic changes have been recently documented in tumor-associated EC, which may contribute to alterations in their antigenic profiles (47). An additional advantage of our approach is that the T-cell responses may be restricted by multiple MHC class I and II antigens, therefore minimizing the need to select patients to be treated on the basis of their HLA phenotype. However, this immunization strategy requires the identification of the immune response(s) that correlates with the clinical response to implement an informative protocol to immunomonitor the immunized patients.
The antiangiogenic immunity that we have elicited was associated with the inhibition of tumor angiogenesis, and control of B16.F10 melanoma and MC38 colon adenocarcinoma tumors. Our observations parallel the information in the literature, which describe vastly different immunotherapeutic approaches to inhibit angiogenesis for tumor control (8–14, 20). The immunogens used range from individual molecules crucial for angiogenesis, such as VEGFR-2 (8–10), Tie-2 (11, 12), MMP-2 (13), and FGF-2 (14), to a complex mixture of EC antigens, such as xenogeneic EC (20). In addition, the strategies used to enhance the immune response include (a) administration of adjuvant (10); (b) utilization of DC loaded with antigen by protein pulsing (8) or RNA transfection (12); (c) delivery of antigen in liposomes (14); (d) delivery of antigen via bacteria (e.g., Salmonella; ref. 9) or virus (e.g., adenovirus; ref. 11); and (e) use of xenoantigens as immunogens (10, 13, 20). Most, if not all, of these strategies seem to control tumor growth therapeutically and to cause minor, if any, side effects on wound healing (9), fertility (8, 12), and/or hematopoiesis (9, 10). Consistent with the latter findings, we observed only minor therapy-related side effects, such as transient decreases in platelet and WBC counts, in our system. The lack of severe side effects in all of these studies may be accounted for by several possibilities that are not mutually exclusive. In one scenario, the elicited antiangiogenic immunity may be directed specifically to the tumor vasculature rather than the normal vasculature, due to differences in markers expressed by the tumor vessels (48). Alternately, EC associated with the normal vasculature may be less sensitive to the antiangiogenic immune response than EC associated with the tumor vasculature, because of constitutive differences and/or changes induced by the microenvironment.
An obvious question is how our strategy compares with the existing strategies in terms of efficacy and feasibility. An answer to this question is hampered by the difficulties in comparing results reported in the literature, because of multiple differences in the tumor models and the therapeutic modalities (e.g., immunogen and adjuvant used, and the dosage, route, and timing of immunizations). The need for this information emphasizes the importance of comparative studies under well-controlled conditions to select the approach(es) to be optimized.
Ultimately, the promising clinical responses achieved in clinical trials of DC-tumor cell hybrids (26), as well as recent progress in the clinical development of antiangiogenic therapies targeting tumor-associated EC (49), suggest that an analogous strategy using tumor-derived EC for the generation of DC hybrids has a high degree of clinical translational potential. Optimization must be carried out for (a) the recovery of autologous, tumor-derived EC; (b) the efficiency of the hybridization process; and (c) the requisite quality control of DC-EC hybrids before immunization. Significant progress has already been achieved for the first of these steps (50), and the latter steps are likely to parallel the optimization steps for DC-tumor cell hybrids.
In summary, we have shown that immunization with DC-EC hybrids and 4-1BB–specific mAb elicits antiangiogenic immunity that inhibits tumor growth but does not generate significant side effects. These findings provide a background for the design of an immunization strategy to be tested in a clinical setting, with the expectation that it may counteract many of the limitations currently encountered in the immunotherapy of malignant diseases.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
Acknowledgments
Grant support: PHS grants P01 CA89480 and R01 CA105500 awarded by the National Cancer Institute, Department of Health and Human Services.
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.
We gratefully acknowledge Dr. Peter Kanter's assistance with our side effects studies, and Mary Vaughan's assistance with the immunohistochemical studies.