Immunity against Tumor Angiogenesis Induced by a Fusion Vaccine with Murine β-Defensin 2 and mFlk-1

Angiogenesis plays a central role in the process of growth and metastasis of primary solid tumors. Antiangiogenesis has proved to be an effective strategy of antitumor treatment (1). Vascular endothelial growth factor (VEGF) and its receptors (VEGFR) are principal regulators of blood vessel formation (2). VEGF functions through specific binding to three different cell membrane receptors, VEGFR1 (Flt-1), VEGFR2 (Flk/KDR), and VEGFR3 (Flt-4); of these, VEGFR2 has a strong tyrosine kinase activity and involves the transduction of major signals for angiogenesis (3). Consequently, Flk is recognized as a direct signal transducer for pathologic angiogenesis, especially in cancer. Moreover, overexpression of Flk and the VEGF/Flk autocrine signaling loop have been reported in a variety of tumor types. The autocrine loop enables cancer cells to promote their own malignant biological actions (4–8). Thus, Flk itself and the associated signal pathway were thought to be critical targets for the suppression of angiogenesis (9, 10). Flk consists of an extracellular domain that binds specific VEGF ligands, a transmembrane domain, and an intracellular region. There is increasing evidence that the antibodies against extracellular region of murine VEGFR2 (mFlk-1) could inhibit tumor growth by hindering VEGF-stimulated proliferation of endothelial cells both in solid tumors and leukemia (11–13). Antibodies, but not CTL targeting the extracellular region, were generated for antiangiogenesis in these studies. In addition, other studies suggested that a specific cellular immune response to mFlk-1 was induced through a minigene vaccine encoding a peptide from self VEGFR2 extracellular region (14). It is obvious that the targets of both the humoral and cellular immune responses lie in the extracellular region of mFlk-1, and that both humoral and cellular immunity against mFlk-1 are effective for inhibiting angiogenesis. Thus, there is a potential for a two-pronged inhibition of tumor growth if both cellular and humoral immunity against mFlk-1 could be evoked simultaneously in the inhibition of angiogenesis.

DNA-based immunization may offer significant advantages compared with other forms of immunization for cancer immunotherapy applications (15). Fusion DNA vaccines have proved a particularly effective strategy in inducing autoimmunity against tumor and have been cited as the second-generation vaccine of such formulations (16, 17). Increasing evidence suggested that vaccines based on tumor-associated antigens could evoke specific active immunity and effectively induce both humoral and cellular immunity when fused with cytokines, chemokines, or some molecular chaperones, which usually interact with their receptors on surface of antigen-presenting cells (APC); however, these tumor-associated antigens usually do not work in an independent manner (18, 19). As known, dendritic cells, the most important APCs, play a critical role in inducing immune responses, but their potential is not fully used in the DNA vaccine setting because they take up only a minor fraction of the injected DNA. Increasing evidence shows that genetic fusion vaccines targeting dendritic cells could significantly amplify immunologic recognition and further enhance specific immune response (18, 20, 21). The present authors therefore considered the hypothesis that the immune tolerance of mFlk-1 could be broken down through genetic fusion with appropriate molecules by which the fusion protein could target to APCs and, in particular, to dendritic cells.

Some antigens are recognized by APCs through pattern recognition receptors such as Toll-like receptors. Immature dendritic cells respond to most of signals from the pattern recognition receptors by undergoing internalization, maturation, and turning on the production of specific sets of cytokines and by promoting antigen presentation. There is evidence that pattern recognition receptors also recognize some self molecules, suggesting that they may have evolved to alert the immune system to change in the body (22). A fusion vaccine with one ligand of pattern recognition receptors and self antigens may be postulated to target antigen delivery and possibly recruit APCs and then deliver an activation signal to promote adaptive immune responses. A recent investigation also showed that exogenous antigens might use the MHC class I processing pathway when they presented to the APCs as fusion proteins with chemokines (23).

Murine β-defensin 2 (MBD2) is a small antimicrobial peptide involved in both the innate and adaptive immunity (24). MBD2 can also modulate the adaptive immune response not only by recruiting immature dendritic cells through chemokine receptor CCR6 but also by activating signaling for dendritic cell maturation through an important pattern recognition receptor, Toll-like receptor 4 (TLR4; refs. 24–26). A receptor-mediated process has also been shown for MBD2-mediated antigen cross-presentation, which induces protective, T-cell–dependent antitumor immunity (19).

Here, we presented a fusion DNA vaccine with MBD2 and murine VEGFR2 (mFlk-1) through a link peptide [(G₄S) × 3] sequence. In this study, we found that the activity of MBD2 remained in spite of the fusion. Inhibition of tumor growth was observed in mice immunized with the fusion DNA pMBD2-mFlk-1 encapsulated with cationic nanoliposome. The fusion vaccine functions depended on the impairment of formation of new tumor blood vessels through humoral and cellular immunity against angiogenesis.

Animals and cell lines. C57BL/6J (CD4^−/−, CD8^−/−, IgH^−/−, and CD1^−/−) knockout mice were obtained from The Jackson Laboratory. C57BL/6J and BALB/c were purchased from the West China Experimental Animal Center. Murine Lewis lung carcinoma cell line LL/2, colon carcinoma cell line CT26, and endothelial cell line MS1 were purchased from the American Type Culture Collection. The MS1, COS7, and Lewis lung carcinoma LL/2 cells were cultured in DMEM, and colon carcinoma CT26 cells were cultured in RPMI 1640, each supplemented with 10% (vol/vol) fetal bovine serum.

Fusion gene cloning and plasmid constructions. All constructs were cloned into the pSecTag/CMV vector (Invitrogen), with a gene placed under the control of the cytomegalovirus (CMV) promoter and behind an Igκ leader sequence. The mature sequences MBD2 and VEGFR2 were cloned by PCR techniques from plasmids constructed previously. The following PCR primers were used: MBD2-forward, 5′-GCGGCCCAGCCGGCCGAACTTGACCACTGCCAC-3′; MBD2-reverse, 5′-CCGCTCGAGTTTCATGTACTTGCAACAGG-3′; MBD2-mFlk-1-linker-reverse, 5′-GCCAGAGCCACCTCCGCCTGAACCGCCTCCACCTTTCATGTACTTGCAACAG-3′; MBD2-mFlk-1-linker-forward, 5′-GTTCAGGCGGAGGTGGCTCTGGCGGTGGCGGATCGGAGAGCAAGGCGCTGCTC-3′; Flk-1-reverse, 5′-CCGCTCGAGCTAGGCACCTTCTATTATG-3′; and mFlk-1-forward, 5′-GCGGCCCAGCCGGCCGAGAGCAAGGCGCTGCTC-3′. MBD2 and mFlk-1 were fused by overlap PCR containing a 15-amino-acid [(G₄S) × 3] linker sequence. All constructs were verified by DNA sequencing (Invitrogen). DNA for vaccination and in vitro transfection of mammalian cells was prepared with EndoFree Kits from Qiagen, Inc.

Expression and biofunctional assay of MBD2-mFlk-1. The reconstructed plasmid or empty plasmid vector was transfected into COS7 cells using LipofectAMINE 2000 reagent (Invitrogen) according to the manufacturer's instruction. Expression of plasmid DNA was detected in transfected COS7 cells by reverse transcription-PCR and with an anti–mFlk-1 rat monoclonal antibody (mAb; BD Bioscience). These culture supernatants were harvested 48 h after transfection and then concentrated by super filter (5 kDa; Minipore). The proteins were analyzed by Western blot.

The bioactivity of these protein products produced by COS7 was determined by the ability of cell supernatants to chemoattract immature dendritic cells. Murine immature dendritic cells were isolated from bone marrow as described previously (27). The migration of immature dendritic cells was assessed using a 96-well microchamber chemotaxis plate (5 μm; Neuro Probe) following published methods (28). Briefly, immature dendritic cells were added on top of the membrane, and transfectant supernatants (30 μL) were concentrated 4-fold using the super filter (5 kDa; Minipore) before being added to the lower compartment. The number of cells that migrated to the lower surface was microscopically counted at six randomly chosen high-power fields. IFN-γ inducible protein 10 (IP10), previously described, was used as a control, as immature dendritic cells exhibit significant chemotaxis through CXCR3 (28, 29). Three replicates were done for each treatment and the experiment was done thrice.

Immunization and tumor models. Mice were immunized with 100 μg per mouse per injection of DNA vaccine encapsulated with cationic liposome (plasmid/nanoliposome, 1:3) by i.m. injection in both later quadriceps once a week for 4 weeks with pMBD2-mFlk-1 alone, pMBD2 alone, pmFlk-1 alone, mixture of pMBD2 and mFlk-1, and pSecTagB vector or normal saline (n.s.; nonimmunized mice) as described previously (30). Seven days after the last immunization, the mice were challenged with 3 × 10⁵ tumor cells s.c. in the right flank. Tumor dimensions were measured every 3 days with calipers, and tumor volumes were calculated according to the following formula: width² × length × 0.52. Due to the death of mice, the tumor volume on day 28 in control groups only came from the surviving mice.

Therapy of established LL/2 lung metastasis models with DNA vaccine. To investigate the therapeutic effects of the fusion DNA, LL/2 lung metastasis models were established. Briefly, C57BL/6J mice were injected with 2 × 10⁵ LL/2 cells i.v. On days 3, 6, 10, and 17, these mice were immunized with DNA plasmid. Mice were sacrificed on day 28 and lung was harvested for measurement of lung weight.

Immunohistochemistry and alginate-encapsulate tumor cell assay. To explore whether the antitumor immunity involved the inhibition of angiogenesis, detection of vessel density in tumor tissue and angiogenesis in vivo was done as described previously (30). Frozen sections were used to determine vessel density with an anti-CD31 antibody. In addition, an alginate-encapsulate tumor cell assay was done. Mice were immunized as above. Alginate beads containing ∼1 × 10⁵ tumor cells per bead were implanted s.c. into both dorsal sides of the immunized mice. After 12 days, mice were injected i.v. with 0.1 mL of a 100 mg/kg FITC-dextran (Sigma) solution. Alginate beads were photographed after being exposed surgically and then rapidly removed 20 min after FITC-dextran injection. The uptake of FITC-dextran was measured as described (30, 31).

Antibody and CTL assays. Anti–Flk-1 antibodies were identified with frozen section of tumor tissue and Western blot. The tumor from untreated mice was used to detect the autoantibodies. The frozen sections were incubated with mouse sera at 1:500. An FITC-labeled secondary antibody (goat anti-mouse IgG) was used to detect the deposition of immunoglobulins in the tumor tissues. In addition, the concentrated culture supernatants from transfected COS7 cells were separated by SDS-PAGE. Gels were electroblotted onto a polyvinylidene difluoride membrane. Subsequently, the membrane was probed with mouse sera at 1:500 and an anti–mFlk-1 rat mAb (BD Bioscience) as a positive control.

The possible mFlk-1–specific cytotoxicity mediated by CTLs was determined by ⁵¹Cr release assay as described previously (12, 32). Briefly, C57BL/6J mice were immunized with 100-μg DNA as described above. Spleens were collected on day 7 after the last vaccination. T lymphocytes were isolated from single-cell suspensions with Nylon Fiber Column T (L-Type, WAKO) as CTL effector cells; MS1 murine endothelial cells, which express mFlk-1 (14), were used as target cells. Effector and target cells were seeded into the 96-well microtiter plate at various effector/target ratios. The CTL activity was calculated by the following formula: %lysis = [(experimental release − spontaneous release) / (maximum release − spontaneous release)] × 100.

Adoptive transfer in vivo. To assess the efficacy of humoral and cellular immunity against tumor through antiangiogenesis in vivo, BALB/c mice were immunized and then T cells were isolated as above. Freshly isolated T lymphocytes (1 × 10⁷) were injected into recipient BALB/c mice through the tail vein on the second day after CT26 challenge.

Immunoglobulins were purified from the pooled sera collected from the immunized or control mice as previously described (12, 33). Purified immunoglobulins were adoptively and i.v. transferred at 50 mg/kg per mouse 1 day before mice were challenged with 3 × 10⁵ CT26 cells and then were administered twice per week for 3 weeks. Tumor growth and survival were monitored in both of the experiments.

Depletion of immune cell subsets in vivo. Immune cell subsets were depleted for BALB/c mice as described (31, 34). Briefly, BALB/c mice were injected i.p. with anti-CD4 (clone GK1.5, rat IgG), anti-CD8 (clone 2.43, rat IgG), anti–natural killer (clone PK136) mAb, or isotype controls at 500 μg/kg per mouse 1 day before the immunization and then immunized with 100 μg of plasmid. Mice were challenged with CT26 cells (3 × 10⁵) on day 7 after the fourth immunization. Tumor size was measured 25 days after injection.

C57BL/6J knockout mice (CD4^−/−, CD8^−/−, IgH^−/−, and CD1^−/−) and wild-type C57BL/6J were immunized. LL/2 cells (2 × 10⁵) were inoculated on day 7 after the fourth immunization. Mice were sacrificed on day 20 and sera were harvested for antibody test. Tumor size was measured.

Statistical analysis. SPSS 11.5 was used for statistical analysis. The statistical significance of results in all of the experiments was determined by Student's t test and ANOVA. Survival curves were compared by the log-rank test. The findings were regarded as significant if P < 0.05.

Characterization and biofunctional assay of MBD2-mFlk-1. The plasmid DNA constructs were generated as described in Materials and Methods. COS7 cells were transfected with the following expression vectors: pMBD2-mFlk-1, pMBD2, mFlk-1, or the empty vector. The supernatant of each transfectant cultured for 48 h was analyzed by Western blot with mFlk-1–specific mAb. Secreted MBD2-mFlk-1 and mFlk-1 were detected in culture supernatants; this indicated that MBD2 was also expressed in a secreted form (Fig. 1A).

To investigate the functional activities of secreted MBD2 and MBD2-mFlk-1, the culture supernatants were tested for their activity to chemoattract murine bone marrow–derived immature dendritic cells, which are known to express CCR6 (26, 35). The results showed that the supernatant from COS7 transfected with pMBD2-mFlk-1 chemoattracted the immature dendritic cells as did the supernatant from pMBD2 or IP10, whereas the supernatants from pmFlk-1, pSecTagB vector, and n.s did not exhibit chemotaxis of the immature dendritic cells (Fig. 1B).

pMBD2-mFlk-1 fusion vaccine elicits protective and therapeutic antitumor immunity. We first investigated whether DNA vaccination with pMBD2-mFlk-1 could induce a specific immune response in vivo. Mice were immunized as described above and then challenged with CT26 or LL/2 tumor cells. Tumors grew progressively in nonimmunized mice and in mice immunized with pmFlk-1, pMBD2, mix, and blank vector. In contrast, there was an obvious inhibition of tumor growth in mice immunized with pMBD2-mFlk-1; this was observed in both BALB/c and C57BL/6J mice (Fig. 2). In contrast, no significant antitumor immunity was detected in mice immunized with the constructs expressing MBD2, mFlk-1, and the mixture of MBD2 and mFlk-1. Therefore, these data suggest that only the fusion vaccine pMBD2-mFlk-1, targeting immature dendritic cells, can elicit a potent protective antitumor immunity. A previous study showed that an oral DNA minigene vaccine encoding a peptide from the extracellular region of mFlk-1 can elicit cellular immunity against mFlk-1–positive endothelial cells (14), whereas in this study, pmFlk-1 failed to induce protective immunity. One possible explanation was that the APCs failed to recognize mFlk-1, which resulted in the default of antigen presentation.

To test whether the fusion constructs can inhibit or abrogate the growth of an established tumor, we used the LL/2 lung metastasis model, which grows relatively slowly at the early stage; because microtumor growth could initiate angiogenic sprouting with simultaneous expression of VEGF and VEGFR2 (36), we also tested the therapeutic potency of the fusion vaccine. The mice were treated with four booster vaccinations at 3, 6, 10, and 17 days after tumor injection. Significantly fewer metastases (data not shown) and lower lung weight were observed in pMBD2-mFlk-1 compared with the other groups (P < 0.05; Fig. 3). Furthermore, mild but not statistically significant inhibitory effects on lung metastases were observed in pmFlk-1 (P > 0.05, compared with pSecTag, pMBD2, and N.S; Fig. 3). It is probable that mFlk-1 was secreted as a soluble receptor that could neutralize a fraction of the VEGF in blood.

Adaptive immunity through T-cell or immunoglobulin adoptive transfer. To explore the possible mechanism by which the antitumor activity was elicited by pMBD2-mFlk-1, T cells were isolated from immunized mice, as described in Materials and Methods, and then transferred i.v. into BALB/c mice, which had been inoculated with CT26 cells 1 day before. Tumor growth was significantly restrained and increased survival was observed in the group that received T cells from pMBD2-mFlk-1–immunized mice (Fig. 4A and C).

The purified immunoglobulins had no direct effect on the proliferation of CT26 or LL/2 tumor cells in vitro (data not shown). However, adoptive transfer with purified immunoglobulins isolated from pMBD2-mFlk-1–immunized mice resulted in apparent inhibition of tumor growth and significantly increased survival compared with control groups (Fig. 4B and D).

Inhibition of angiogenesis. Immunohistochemical staining of the tumor tissue from pMBD2-mFlk-1–immunized mice with anti-CD31 showed significantly decreased microvessel density compared with control groups (Fig. 5A, a-f). The number of microvessels also showed a significant difference (Fig. 5A, g). In addition, inhibition of angiogenesis could also be detected in the alginate-encapsulate tumor cell assay (Fig. 5B, a-f). New blood vessels in alginate beads from pMBD2-mFlk-1–immunized mice were apparently sparse. Besides, FITC-dextran uptake was significantly decreased from mice treated with pMBD2-mFlk-1 when compared with control groups (Fig. 5B, g). These results suggested that tumor angiogenesis was inhibited in pMBD2-mFlk-1–immunized mice, which resulted in suppression of tumor growth.

H&E staining was done to investigate whether the morphologic changes resulted from the inhibition of angiogenesis. Both decreased density of vessels and obvious necrosis were observed in center of tumor from pMBD2-mFlk-1–immunized mice (data not shown). In addition, the increased apoptotic tumor cells were observed, which distributed locally in tumor (Fig. 6A). Tumor cells grew well and no obvious necrosis or apoptosis was observed in other groups including pmFlk-1 (Fig. 6B), pSecTag (Fig. 6C), and other control groups (results not shown).

Characterization of autoantibodies and assay of CTL-mediated cytotoxicity. To investigate the mechanism by which the antitumor activity was elicited by pMBD2-mFlk-1, we identified autoantibodies against Flk-1 in the immunized mice. Sera obtained from pMBD2-mFlk-1–immunized mice were tested with frozen tumor tissue sections from untreated mice and the supernatant from pmFlk-1–transfected COS7 as mentioned above. Antibodies were detected both by immunohistochemical staining and Western blot analysis (Fig. 6D-G). The immunofluorescence was observed only in the sections treated with the sera from pMBD2-mFlk-1–immunized mice (Fig. 6D). Antibodies also were detected by Western blot analysis (Fig. 6G, lane 2), which showed a transparent band of the same size that was revealed by an anti–mFlk-1 rat mAb (BD Bioscience). No anti–mFlk-1 antibody was found from sera obtained from pmFlk-1–immunized mice (Fig. 6G, lane 3). Similarly, the sera isolated from other groups also showed negative staining (results not shown). In addition, specific CTL activity was tested by ⁵¹Cr release assay; these CTL assays showed that T lymphocytes from the mice immunized with pMBD2-mFlk-1 were more cytotoxic to mFlk⁺ MS1 cells than control groups (Fig. 6H). These findings indicate the evocation of both humoral and cellular immunity against mFlk-1.

Function of T-cell subsets in antitumor activity. To further confirm the roles of immune cell subsets in the antitumor activity elicited by pMBD2-mFlk-1, CD4⁺ or CD8⁺ T lymphocytes or natural killer cells were depleted as described above. Mice depleted of CD8⁺ T lymphocytes and vaccinated with pMBD2-mFlk-1 showed decreased protection from CT26 challenge compared with wild-type BALB/c mice. Depletion of CD4⁺ T lymphocytes almost completely removed the protective effects from pMBD2-mFlk-1–immunized mice. In contrast, treatment with mAb against natural killer cells had no discernable effect on the antitumor activity (Fig. 7A).

Furthermore, C57BL/6J knockout mice (CD4^−/−, CD8^−/−, IgH^−/−, and CD1^−/−) and wild-type C57BL/6J were also immunized and challenged. Protective immune responses were dramatically impaired in CD4^−/−, CD8^−/−, and IgH^−/− mice versus wild-type C57BL/6J (P < 0.05); however, significant differences (reduced tumor growth) remained in pMBD2-mFlk-1–immunized mice compared with N.S group (P < 0.05; Fig. 7B). No changes were observed in CD1^−/− mice. Sera obtained from knockout mice immunized with pMBD2-mFlk-1 were also tested. No anti–mFlk-1 antibody was detected with sera from CD4^−/− and IgH^−/− mice. Traces of anti–mFlk-1 antibodies were found in CD8^−/− and CD1^−/− mice (data not shown).

These findings indicated that the function of the fusion vaccine partially depended on CD8⁺ T lymphocytes and B cells (plasma cells) and, to a greater extent, on CD4⁺ T lymphocytes.

Endothelial cells play a critical role in a large number of physiologic and pathologic processes, especially in tumorigenesis (37). Proliferation and growth of endothelial cells in tumor-derived vessels is activated in response to many growth factors produced by tumor cells or overexpression of their receptors on the surfaces of endothelial cells (38, 39). The overexpression or coexpression of these growth factors and their receptors shows relative specificity, which suggests their potential as targets of immunotherapy. Currently, the development of antiangiogenesis drugs focuses on mAbs and molecular target drugs through which these growth factors and their signal pathways are restrained. Some of these have proved effective in clinical application (e.g., Avastin and Iressa). However, tumor angiogenesis is a complicated process involving numerous factors and signal pathways and, as such, is not conducive for a long-term control by mAbs and molecular target drugs. Considering the low effective ratio, short-duration effectiveness, and high cost of the mAbs and molecular target drugs, vaccine-based immunotherapy targeting angiogenesis shows potential advantages, which can target activated endothelial cells and generate lasting effects at a relatively low cost. Because vascular endothelium serves as the key barrier between the intravascular compartment and extravascular tissues, and thus their location at the blood-extravascular tissue interface, endothelial cells from tumor-derived vessels as target cells are constantly and readily exposed to circulating antibodies and lymphocytes. Evocation of both humoral and cellular immunity against endothelial cells through breakdown of the immune tolerance of the self antigen VEGFR2 should be an effective strategy of immunotherapy for antiangiogenesis.

Our study indicates that effective immunity, including humoral and cellular immunity, against mFlk-1 is elicited by targeting antigens to APCs and especially to dendritic cells when the fusion protein MBD2-mFlk-1 is secreted into the blood. In our design, the efficacy of the fusion vaccine pMBD2-mFlk-1 depends on the functional expression of MBD2, the ligand of CCR6, and TLR4. Both CCR6 and TLR4 are internalization receptors that show ligand-induced receptor signaling and internalization. CCR6 is expressed in many APCs including B cells and memory T cells, whereas immature dendritic cells express both CCR6 and TLR4. Dendritic cells are widely distributed as immature cells in blood and all tissues. Fusion protein MBD2-mFlk-1 could obtain an increased chance of being recruited and targeted to immature dendritic cells by both CCR6 and TLR4 when secreted into the blood. Internalization of MBD2-mFlk-1 could subsequently trigger the maturation and migration of immature dendritic cells from peripheral tissues to lymphoid organs, which could contribute to antigen presentation. It is obvious that, regardless of which of these receptors or cells expressing CCR6 or TLR4 is mostly responsible for MBD2-mediated antigen-presentation, it is most likely that other APCs, in addition to immature dendritic cells, also contribute to the augmentation of vaccine potency by targeting CCR6.

Subsequently, several observations have been made in the present study about the vaccine. The fusion protein MBD2-mFlk-1 showed a similar chemotactic capacity with that of MBD2 and IP10; this ensured that the fusion protein can be recruited and targeted to APCs and deliver an activation signal to promote adaptive immune responses. Inhibition of tumor growth was found in mouse models immunized with pMBD2-mFlk-1, thus indicating that protective immunity was evoked. Furthermore, inhibition of angiogenesis was observed in mouse tumor tissue and in an alginate-encapsulate tumor cell assay. Autoantibodies against mFlk-1, but not MBD2, from pMBD2-mFlk-1–immunized mice were identified by Western blot analysis. T cells isolated from mice immunized with pMBD2-mFlk-1 showed increased cytotoxicity against mFlk-1–positive target cells, MS1. The antitumor activity and the inhibition of angiogenesis were also acquired by the adoptive transfer of the purified immunoglobulins or isolated T lymphocytes.

It was also noted that the effects of the fusion vaccine were restrained by the depletion of CD4⁺ or CD8⁺ T lymphocytes and impaired in CD4^−/−, CD8^−/−, and IgH^−/− mice. CD4⁺ helper T lymphocytes play a key role in both humoral and cellular immune responses. Knockout or depletion of CD4 almost completely abrogated protective immune responses; this indicated that CD4⁺ helper T lymphocytes were required for the activity of the fusion vaccine. IgH^−/− mice could not produce the immunoglobulin heavy chain, and this results in the absence of antibodies. Knockout of IgH decreased the effects of the pMBD2-mFlk-1 vaccine, which indicates that the action of this vaccine is, in part, dependent on humoral immune response. Likewise, the degradation of pMBD2-mFlk-1 function in CD8⁺ T lymphocyte–depleted mice indicated that CD8⁺ T lymphocyte immune response also played an important role. Based on these findings, we can conclude that both CD4⁺ and CD8⁺ T lymphocytes and immunoglobulin-secreting plasma cells are involved in the immune response against mFlk-1 and exclude nonspecifically augmented immune responses against tumor growth in host mice.

Tumor vaccines are primarily aimed on the treatment of established tumor; however, in most mouse tumor models, tumors grow fast, which results in a short life span of the host animal. In this study, the mouse lung metastasis model was used to investigate the therapeutic potential of the fusion vaccine. The results showed that lung metastases were significantly restrained by the fusion vaccine. Because the multicellular aggregates (≪1 mm³) initiate vascular growth by angiogenic sprouting through the expression of VEGFR2 by host and tumor endothelium (36), it is theoretically possible for the fusion vaccine to be applied in the prevention or treatment of early tumor metastases.

In fact, a previous study also showed that both humoral and cellular immunity against mFlk-1 were elicited through dendritic cell vaccine pulsed with murine flk protein (40). As an antigen, mFlk-1 should be more effectively processed and presented by dendritic cells. However, extensive application of a dendritic cell vaccine is difficult because of its limitation of resource and MHC restriction. Because fusion DNA vaccines do not have these problems and also evocate both similar humoral and cellular immunity against endothelial cells, the fusion vaccine with MBD2 and mFlk-1 may be a preferable choice.

In summary, robust antibody and CTL responses indicated that secreted MBD2-mFlk-1 fusion proteins stimulate both T-cell and B-cell responses. In addition, immune tolerance to self antigen mFlk-1 was consequently broken down through the genetic fusion vaccine. These results indicated that functional linkage of MBD2 and mFlk-1 was necessary to elicit an adequate immune response to mFlk-1. Our study also provided a primary investigation through which the fusion vaccine pMBD2-mFlk-1 could be further developed for clinical study.

We thank Dr. Stephen Attwood (State Key Laboratory of Biotherapy, Huaxi hospital, Sichuan University) for review of the manuscript.