H-2K^b–Restricted CTL Epitopes from Mouse Heparanase Elicit an Antitumor Immune Response In vivo Free

The identification of CTL epitopes from tumor antigens is very important for the development of peptide-based, cancer-specific immunotherapy. Heparanase is broadly expressed in various advanced tumors and can serve as a universal tumor-associated antigen. Although several epitopes of heparanase antigen are known in humans, the corresponding knowledge in mice is still rather limited. The present study was designed to predict and identify the CTL epitopes in the mouse heparanase protein. For this purpose, H-2K^b–restricted CTL epitopes were identified by using the following four-step procedure: (a) a computer-based epitope prediction from the amino acid sequence of mouse heparanase, (b) a peptide-binding assay to determine the affinity of the predicted epitopes with the H-2K^b molecule, (c) the testing of the induction of CTLs toward various carcinoma cells expressing heparanase antigens and H-2K^b, and (d) the induction of immunoprotection and immunotherapy in vivo. The results showed that, of the tested peptides, effectors induced by peptides of mouse heparanase at residue positions 398 to 405 (LSLLFKKL; mHpa398) and 519 to 526 (FSYGFFVI; mHpa519) lysed three kinds of carcinoma cells expressing both heparanase and H-2K^b (B16 melanoma cells, EL-4 lymphoma cells, and Lewis lung cancer cells). In vivo experiments indicated that mHpa398 and mHpa519 peptides offered the possibility of not only immunizing against tumors but also treating tumor-bearing hosts successfully. Our results suggest that the mHpa398 and mHpa519 peptides are novel H-2K^b–restricted CTL epitopes capable of inducing heparanase-specific CTLs in vitro and in vivo. These epitopes may serve as valuable tools for the preclinical evaluation of vaccination strategies. [Cancer Res 2008;68(5):1529–37]

Heparanase is an endo-β-d-glucuronidase that can cleave heparan sulfate proteoglycans (HSPG) and has been implicated in tumor angiogenesis and metastasis (1, 2). In the process of metastasis, the basement membrane and extracellular matrix act as barriers to tumor cell invasion and spread (3). Heparanase produced by malignant tumor cells can mediate the degradation of HSPGs in the extracellular matrix and basement membrane, and it is one of the key enzymes involved in the invasion and metastasis of malignant tumors. A large number of publications now clearly link heparanase expression to the process of tumorigenesis and invasion in a wide number of cancers, including gastric, liver, colon, pancreatic, esophageal, breast, bladder, prostate, brain, thyroid, ovary and lung cancer, and acute myeloid leukemia (4). Overexpression of the heparanase cDNA in low-metastatic tumor cells confers high-metastatic potential in experimental animals, resulting in an increased rate of mortality (5). In addition, the enzyme also releases extracellular matrix–resident angiogenic factors in vitro, and its overexpression induces an angiogenic response in vivo (6, 7). Therefore, heparanase may facilitate both tumor cell invasion and neovascularization, which are critical steps in cancer progression. The broad expression of heparanase in advanced tumors indicates that heparanase can serve as a tumor-associated antigen (TAA) in the immunotherapy of advanced tumors.

Cellular adoptive immunotherapy with specific CTLs has been used to treat malignant tumors. Antigen presentation is critical for the initiation of an adaptive immune response. Dendritic cells are the most potent professional antigen-presenting cells and have the most powerful antigen-presenting capacity (8, 9). They can present a tumor antigen to the immune system and initiate a specific immune response. The presentation of TAAs by dendritic cells, the recognition by CTLs, and the induction of a specific antitumor immune response have been focal areas for research on tumor therapy. Dendritic cells pulsed with various TAAs have proven to be effective in producing specific antitumor effects in vitro and in vivo (10–13).

CTLs are considered chief mediators of tumor immunosurveillance via the recognition of TAAs as cognate peptides bound to MHC molecules expressed on the surface of tumor cells. A major achievement in the field of tumor immunology over the last 20 years has been the clear demonstration that CTL epitopes binding to MHC, rather than integral TAA, induce CTL reactions. These epitope peptides are usually 8 to 10 amino acid long with two to three primary anchor residues that interact with the MHC class I molecules and two to three amino acid residues that bind to the T-cell receptor (14). Therefore, the identification of CTL epitopes from TAAs has become a critical step in the development of peptide-based immunotherapy for cancer (15, 16).

Our previous study showed denditric cells loading full-length heparanase cDNA can induce heparanase-specific CTLs, which showed potent lysis of gastric cancer cells in a MHC-restricted fashion (17). This result indicates that heparanase can serve as a TAA in the immunotherapy of tumors. CTL epitopes must exist to induce the specific CTLs of the heparanase protein. Recently, Sommerfeldt (18) successfully predicted three epitopes derived from the human heparanase amino acid sequence. Their results showed that these three epitopes could elicit heparanase-specific CTLs to lyse breast cancer cells in vitro. We also predicted and identified three HLA-A2–restricted heparanase epitopes, which were different from those in Sommerfeldt's report. In vitro experiments showed that these heparanase epitopes could induce heparanase-specific CTLs to lyse various tumor cells in an HLA-A2–restricted fashion (data not shown).

To investigate the in vivo immune response elicited by heparanase CTL epitopes, the present study was undertaken to identify candidate CTL epitopes derived from the mouse heparanase protein (mHpa). For this purpose, we first predicted candidate epitopes of mHpa using bioinformatic methods. These predicted epitopes were then identified by in vitro and in vivo experiments. We hoped to find one or more mHpa epitopes that could induce heparanase-specific antitumor immune responses both in vitro and in vivo.

Mice and cell lines. C57BL/6(H-2K^b) mice were purchased from Chongqing Medical University and were used at ages 8 to 12 weeks. Animal studies were performed in agreement with the local ethics committee of the Third Military Medical University. Syngeneic tumor B16 melanoma cells (H-2K^b), EL-4 lymphoma cells (H-2K^b), and Lewis lung cancer cells (H-2K^b) were purchased from American Type Culture Collection. The mouse P815 mastocytoma cell line (H-2K^d) was presented by Dr. Wan Y (Institute of Immunology of People's Liberation Army, Third Military Medical University, Chongqing, P.R. China). RMA-S cells that were transporter-associated with antigen processing (TAP) deficiency and were derived from the T-lymphoma cell line RMA, were also donated by Dr. Wan Y. All cell lines mentioned above were maintained as monolayers in RPMI 1640 containing 10% heat-inactivated FCS, penicillin (200 units/mL), and streptomycin (100 μg/mL) and were kept at 37°C in a humidified atmosphere of 5% CO₂ in air.

Epitope prediction. There are several algorithms used to predict H-2K^b–restricted CTL epitopes. In the present study, we used the matrix-based algorithm, BioInformatics and Molecular Analysis Section (BIMAS), SYFPEITHI, and IMTECH to identify the candidate H-2K^b–restricted CTL epitopes from the mHpa antigen. Octapeptides derived from the mHpa amino acid sequence and the influenza virus epitope octapeptide (NYKHCFEI), which served as a negative control, were synthesized by Shanghai C-Strong Co. Ltd with a purity of >90% (as determined by high performance liquid chromatography and amino acid analysis). The amino acid sequences were dissolved in DMSO (Sigma) and stored at −20°C.

Peptide-binding assay with flow cytometry. The binding activities of selected peptides to the H-2K^b molecule were determined semiquantitatively by measuring peptide-induced expression of H-2K^b molecules on RMA-S cells with flow cytometry. The RMA-S peptide-H-2K^b stabilization assay was performed as previously described (19). Briefly, RMA-S cells were cultured at 26°C overnight and washed with serum-free medium. Then, the cells were seeded into a 96-well plate at 5 × 10⁵ cells per well. The predicted peptides were added at a concentration of 10, 20, 30, or 40 μmol per well. Assays were performed in triplicate for each peptide at each concentration in a 96-well round-bottomed plate. The cells were incubated at 37°C for 18 hours and washed twice with cold (4°C) PBS. Next, the cells were immunostained with 1 μL of FITC-conjugated H-2K^b antibody (Biolegend) for 30 minutes. The cells were then washed thrice with cold PBS, after which fluorescence intensity was measured with a flow cytometer (Becton Dickinson). Fluorescence index, the relative binding affinity of the respective peptides, was calculated from the mean fluorescence intensities (MFI) as follows: [MFI (peptide) − MFI (unloaded cells)]/MFI (unloaded cells). Relative binding affinities of >1.5 were considered strong; 1.0 to 1.5, intermediate; and <1.0, low (20).

Denditric cell generation from mouse bone marrow. Denditric cells from mouse bone marrow (mDC) were generated according to the protocol described previously (21). In brief, bone marrow was flushed from the tibias and femurs of C57BL/6 mice and depleted of erythrocytes with commercial lysis buffer (Sigma). The cells were washed twice in serum free RPMI 1640 and cultured in a six-well plate at 5 × 10⁶ cells per well with RPMI 1640 containing 10 ng/mL recombinant murine granulocyte macrophage colony-stimulating factor (mGM-CSF; Endogene) and 10 ng/mL recombinant murine interleukin (IL)-4 (mIL-4; Endogene). On days 3 and 5, half of the medium was refreshed without discarding any cells, and fresh cytokine (mGM-CSF and mIL-4) was added. On days 7 and 8 of culture, murine tumor necrosis factor α was added to the medium. On day 10, nonadherent cells obtained from these cultures were considered to be mature bone marrow–derived mDCs. The phenotypic markers of mDCs were confirmed by flow cytometry.

Pulsing of denditric cells with the immunodominant peptide. Denditric cells generated from mouse bone marrow were cultured in 1 mL of RPMI 1640 supplemented with 10% FCS, 2 mmol/L l-glutamine, 50 μmol/L 2-mercaptoethanol, 100 U/mL penicillin, and 100 μg/mL streptomycin containing 100 μg/mL of the different octapeptides. Culturing occurred at 37°C for 3 hours (with gentle shaking every 30 minutes). Cells were then washed twice in PBS and used for further experiments.

Mouse immunization. For mHpa peptide immunizations, C57BL/6 mice were immunized thrice (with a 7-day interval) by s.c. injection of 5 × 10⁵ of the above peptide-pulsed denditric cells. As a control, mice were immunized with the same procedure using negative peptide–pulsed denditric cells replacing mHpa peptide–pulsed denditric cells.

Isolation of splenic lymphocytes and preparation of peptide-specific CTLs. Mice were sacrificed, and spleens were removed 21 days after denditric cell immunization. Then, the in vivo primed splenocytes were cocultured (4 × 10⁶/mL) with different mHpa peptides in a 6-well plate in complete medium containing IL-2 at 50 units/mL. After 5 days of coculture, the in vivo restimulated splenocytes were assayed in a 4-hour ⁵¹Cr-release assay.

Reverse transcription-PCR. Total RNA of target cells (B16 melanoma cells, EL-4 lymphoma cells, Lewis lung cancer cells, and P815 mastocytoma cells) was isolated from cell extracts using the Tripure RNA isolation kit (Roche) following the manufacturer's instructions. One microgram of total RNA was reverse transcribed and the resulting cDNA was amplified with Taq polymerase (Takara) using mHpa-specific primers (sense primer, 5′-CGCCCAGCTTCTCCTTGACTAC-3′; antisense primer, 5′-GCTCCTGTCTGGGCCTTTCAC-3′) or β-actin(sense primer, 5′- GTGGGCGCCCCAGGCACCA-3′; antisense primer, 5′-CTTCCTTAATGTCACGCACGATTTC-3′). After an initial denaturing for 5 minutes at 94°C, the PCR thermal cycle profile was 45 seconds at 94°C, 60 seconds at 66°C, and 60 seconds at 72°C for 30 cycles, with a final 5 minutes extension at 72°C. The products of reverse transcription-PCR (RT-PCR) were electrophoresed through 1.5% agarose and visualized by UVP after ethidium bromide staining.

Immunohistochemical staining. The above target cells were cultured in six-well culture dishes and fixed with 3.7% paraformaldehyde for 7 minutes at room temperature. The dishes were then washed with buffer containing 0.1 mol/L Tris (pH 7.5), 1.5 mol/L NaCl, and 1% bovine serum albumin and permeabilized in the presence of 2% Triton X-100. Next, cells were incubated with a heparanase antibody (0.5 μg/mL; Santa Cruz Biotechnology) for 1 hour at room temperature, washed, and further incubated with secondary antibody. Finally, the cells were incubated for 15 minutes with an avidin-biotin enzyme reagent. Slides were immersed in 3,3′-diaminobenzidine/H₂O₂ solution to develop the staining. PBS was used as a negative control in place of the primary antibody.

Chromium release assays. To evaluate levels of CTL activity, a standard 4-hour ⁵¹Cr-release assay was used as was previously described (22). Briefly, target cells (B16 melanoma cells, EL-4 lymphoma cells, Lewis lung cancer cells, and P815 mastocytoma cells) were pulsed with peptides (100 μg per 2 × 10⁶ cells) and were then incubated with ⁵¹Cr (100 μ Ci per 1 × 10⁶ cells; Amersham Biosciences Corp) for 2 hours in a 37°C water bath. After incubation with ⁵¹Cr, target cells were washed thrice with PBS, resuspended in RPMI 1640, and mixed with effector cells at effector-to-target (E/T) ratios of 5:1, 10:1, 20:1, or 40:1. Assays were performed in triplicate for each sample at each ratio in a 96-well round-bottomed plate. After a 4-hour incubation, the supernatants were harvested and the amount of ⁵¹Cr released was measured with a γ counter. The percent specific lysis was calculated according to the following formula:

Enzyme-linked immunospot assay. Heparanase epitope–specific T cells were enumerated using enzyme-linked immunospot (ELISPOT; ref 23). Splenocytes obtained from vaccinated animals, at 1 × 10⁶, 2 × 10⁶, and 4 × 10⁶ in a volume of 100 μL, were added to ImmunoSpot plates (Dakewe) precoated with an anti–IFN-γ monoclonal antibody (mAb). Then, the different peptides were added to each well at a final concentration of 30 μmol/L. A positive control (15 μg/mL, phytohemagglutinin) and a “no peptide” negative control were included in all assays. The plates were incubated overnight at 37°C in a 5% humidity incubator. After incubation, the wells were washed twice with deionized water and thrice with wash buffer (PBS + 0.05% Tween 20). The wells were then incubated with biotin-conjugated anti–IFN-γ mAb at 37°C for 1 hour. After three washes, streptavidin-alkaline phosphatase was added to each well, and the plates were incubated for 1 hour at room temperature. The plates were again washed thrice. Thirty microliters of activator solution (Dakewe) were added to develop spots. After 10 to 30 minutes, the plates were washed with distilled water to stop the reaction. Experiments were performed in triplicate. After being air dried, the number of spots in each well was counted using the Bioreader 4000 PRO-X (Bio-Sys).

Vaccination and tumor challenge experiments. All animal protocols were approved under guidelines of the animal protection act. For the protective experiment, 30 C57BL/6 mice (6–8 weeks old) were randomly divided into 6 groups and injected thrice s.c. with either mHpa519- or mHpa398-pulsed denditric cells at 1 week intervals. One week after the third injection, each mouse was challenged for tumor growth by s.c. injection of 1 × 10⁶ B16 cells in 100 μL PBS; after 4 weeks, mice were sacrificed and the tumor size was measured by caliper ruler. As controls, 100 μL PBS or human influenza virus octapeptide (NYKHCFEI)-pulsed denditric cells were injected (24).

For therapeutic experiments, 30 C57BL/6 mice (6–8 weeks old) were randomly divided into 6 groups and injected s.c. with 1 × 10⁶ B16 cells in the right flank. After 1 week, when the tumor growth was first seen (diameter, 1 mm), mHpa519- or mHpa398-pulsed denditric cells were injected s.c. The injection was repeated thrice at 1-week intervals. After 4 weeks, mice were sacrificed and the tumor size was measured with a caliper ruler. Control mice were injected with 100 μL PBS or human influenza virus octapeptide (NYKHCFEI)-pulsed denditric cells.

Statistics. All the experiments were run in triplicate, and the results are given as means ± SD of triplicate determinations. Statistical analyses were performed using the Student's t test. The difference was considered statistically significant when the P value was <0.05. All statistical analyses were carried out with SPSS 11.5 software.

Epitope prediction of H-2K^b–binding peptides derived from mHpa. The amino acid sequence of the mouse heparanase transcript was predicted for octapeptides with anchor residues according to the peptide-binding motif of H2-K^b 4

⁵⁶. Five peptides were selected containing tyrosine or phenylalanine in position 5 and one of the hydrophobic amino acids—leucine, isoleucine, valine, or methionine—at the COOH terminus (25). These peptides were derived from positions 234 to 271 (mHpa234; LGEDFVEL), 351 to 358 (mHpa351; FAAGFMWL), 386 to 393 (mHpa386; VDENFEPL), 398 to 405 (mHpa398; LSLLFKKL), and 519 to 526 (mHpa519; FSYGFFVI) of the amino acid sequence of the molecule.

MHC peptide–binding assay. To investigate whether the predicted peptide could up-regulate MHC class I expression on H-2K^b TAP-deficient cells, we measured H-2K^b expression on RMA-S cells in the absence or presence of individual predicted peptides using flow cytometry. A human HLA-A2–restricted influenza virus epitope (NYKHCFEI) served as a negative control. As shown in Fig. 1, we found that all the predicted peptides could up-regulate expression of H-2K^b on RMA-S cells. The peptides increased the expression of H-2K^b molecules on RMA-S cells, accompanied by an increase in the concentration of peptides (Fig. 1A). When the concentration of peptides was 30 μmol/L, the relative binding affinity was >1.5 (Fig. 1B). These results indicate that all the predicted peptides bound to and up-regulated H-2K^b molecules on RMA-S cells.

Expression of heparanase in tumor cell lines. The expression of heparanase mRNA in all cell lines in this study was analyzed with RT-PCR. The results show that heparanase mRNA was detected in B16 melanoma cells (H-2K^b), EL-4 lymphoma cells (H-2K^b), Lewis lung cancer cells (H-2K^b), and P815 mastocytoma cells (H-2K^d; Fig. 2A). Moreover, immunohistochemistry indicated the expression of heparanase protein in the cytoplasm of all target cells regardless of their MHC phenotype (Fig. 2B).

Morphologic and phenotypic characteristics of mouse bone marrow–derived denditric cells. On day 9 of the cell culture, the mature denditric cells displaying typical characteristics of morphology were harvested from monocytes cultured in the medium containing GM-CSF, IL-4, and tumor necrosis factor-α. When viewed by a phase-contrast microscopy, these mature cells were loosely suspended, exhibited irregular cell shape, and displayed many fine processes at their edges. The mature denditric cells were analyzed for phenotype by flow cytometry. The results showed that these mature denditric cells expressed high level of CD11c (94.7%), CD86 (93.8%), H-2K^b (98.5%) and MHC-I (96.3%).

Induction of CTLs specific to mouse heparanase antigen by synthetic peptide. To detect whether the predicted peptides could generate heparanase-specific cytotoxic activity in vivo, C57BL/6 mice were immunized s.c. thrice at a 7-day interval with peptide-pulsed denditric cells. Splenocytes from the immunized mice served as effectors. As shown in Fig. 3A, of the five peptides tested, peptides mHpa398 and mHpa519 were able to elicit mHpa-specific CTLs, which could lyse B16 cells expressing both heparanase and H-2K^b at an E/T ratio from 20:1 to 80:1. However, the induced effectors generated from mHpa234, mHpa351, and mHpa386 could not lyse B16 cells. Even at the highest E/T ratio, the lysis rate was ∼20%. Further study indicated that, in addition to B16 cells, CTLs generated from mHpa398 or mHpa519 could lyse Lewis and EL-4 cells, which also expressed both heparanase and H-2K^b. However, these induced effectors could not lyse P815 cells, which were heparanase positive but H-2K^b negative (Fig. 3B and C).

Peptide-specific T cells were enumerated by measuring IFN-γ–producing cells by ELISPOT assay. As shown in Table 1, a substantial number of T cells harvested from mice immunized with the selected mHpa peptides responded by producing IFN-γ. The responses observed indicate that mHpa519 and mHpa398 peptides were found to generate strong peptide-specific T-cell responses by virtue of their ability to induce increased frequencies of IFN-γ–producing T cells compared with negative peptides (P < 0.01). The frequency of IFN-γ–producing cells in the mHpa519 and mHpa398 primed splenocyte population was found to increase, accompanied by an increase of splenocytes (Table 1). Taken together, these results indicate that the predicted peptides mHpa519 and mHpa398 could induce heparanase-specific CTL responses.

Inhibition of the recognition of effectors by anti–H-2K^b antibody. To determine whether the predicted peptide mHpa519- and mHpa398-induced effectors recognized the heparanase-positive target tumor cells in an H-2K^b–restricted manner, the mAbs against H-2K^b were used to block recognition by effectors. The results show that anti–H-2K^b antibody could significantly eliminate the cytotoxicity of the effectors against B16 melanoma cells, Lewis lung cancer cells, and EL-4 lymphoma cells (Fig. 4), which indicates that the induced effectors lysed the tumor cells in an H-2K^b–restricted manner.

Killing effect of mouse heparanase-specific CTLs on autologous lymphocytes and denditric cells. Although expressed in malignant tissues, heparanase could also be detected, to a lower degree, in some normal tissues and cells. It was previously reported that heparanase could be expressed in immunologically competent cells, natural killer cells, and inflammatory cells, such as neutrophils, granulocytes and activated T cells and B cells (4, 17). Theoretically, immunotherapy aimed at heparanase may elicit adverse effects on the immune system. To investigate the effect of heparanase-specific CTLs on immunologically activated lymphocytes, CTLs induced by mouse heparanase-specific peptides were also used to lyse autologous lymphocytes and denditric cells. The results indicate that heparanase vaccination had no detectable lysis effect on these cells (Fig. 5).

Induction of immunoprotection and immunotherapy in mice. To investigate the effects of mHpa epitope peptide immunization in an in vivo tumor model, we immunized mice with mHpa519, mHpa398, or the respective controls, pulsed denditric cells thrice in weekly intervals, and attempted to promote tumor growth by s.c. injection of 1 × 10⁶ viable B16 melanoma cells. After 4 weeks, mice were sacrificed, tumors were stripped completely, and the tumor size was measured with a caliper ruler. In these experiments, we could find that all mice immunized with mHpa519- or mHpa398 peptide–pulsed denditric cells were protected from tumor growth (Fig. 6A). The tumor size in the mHpa519 or mHpa398 protective group was significantly smaller than that in control groups (P < 0.05; Fig. 6A).

In clinical settings, a therapeutic approach (i.e., the treatment of tumor-bearing individuals) is desirable. Therefore, we investigated the effect of mHpa519- or mHpa398 peptide–pulsed denditric cell injection on preexisting tumors. We injected B16 melanoma cells and waited until palpable tumors (1-mm diameter) had developed. Thereafter, mice received three injections of mHpa519- or mHpa398 peptide–pulsed denditric cells at 1-week intervals, and the tumor size was measured with a caliper ruler. Here, we showed that the mHpa519 or mHpa398 peptide could slow the growth of tumors (Fig. 6B). The tumor size in the mHpa519 or mHpa398 therapeutic group was significantly smaller than that in the control groups (P < 0.05; Fig. 6B). Thus, these data indicate that mouse heparanase epitope peptide offers the possibility not only to immunize against tumors but also to treat tumor-bearing hosts successfully.

The malignant tumor is one of the most lethal diseases that threaten human beings. Therapy of malignant tumors still relies mainly on surgical exeresis, radiotherapy, and chemotherapy. Surgical exeresis can be a better choice for early stage parenchyma tumors, but for advanced stage tumors and liquid tumors, invasion and metastasis more often occur, and operation therapy is not suitable. Due to the intolerable side effects of radiotherapy and chemical therapy, these treatments are not desirable either. Recently, denditric cell–based immunotherapy, which has the advantages of strong immunogenicity, weak side effects, and broad applicability, is becoming a topic of intense interest in the study of malignant tumor therapy (26–29).

The major mechanism of tumor immunotherapy is through TAA-loaded denditric cells. At present, identification of TAAs in cancer patients has been limited to a few cancers, such as MART-1 (melanoma antigen recognized by T cell-1), which is specific for melanoma (30), and carcinoembryonic antigen, which is specific for gastrointestinal tumors (22, 31). These tumor-specific antigens, also called autoantigens, are shared among patients with the same type of tumor. Therefore, immunization with these autoantigens can only induce an immune response for the identical type of tumor that expresses the self-antigens but cannot induce an immune response against other types of tumors that do not express these autoantigens. It has thus been a challenge for clinicians to identify universal TAAs for immunotherapy of multiple types of tumors.

An ideal universal TAA should have the following characteristics: (a) be expressed by the vast majority of human cancers but rarely be expressed in normal tissues, (b) be indispensable in the process of tumorigenesis to avoid antigen variation or depletion, (c) be able to include peptide sequences that bind to MHC molecules, and (d) be recognized by the T-cell repertoire in an MHC-restricted fashion to elicit a specific T-cell response (32).

Heparanase is the only endogenous endoglycosidase found to date that can degrade the HSPG in extracellular matrix and basement membrane. Unlike most other TAAs, the expression of heparanase in tumor cells has been linked to tumor invasion and metastasis. Heparanase can be found in almost all of the metastatic malignant tumor cells. With the exception of neutrophils, granulocytes, and activated T cells and B cells, it is not expressed in normal cells (4, 17). Inhibition of heparanase can obviously inhibit the proliferation and metastasis of tumor cells. Activation of heparanase is a determinant factor for the occurrence of metastasis that allows tumor cells to break through the extracellular matrix and basement membrane barriers, release multiple types of cytokines, and cause the formation of new vessels and local permanent planting. Heparanase is a potential universal TAA for the treatment of advanced stage tumors. Our previous study showed that the denditric cell–loading full-length heparanase cDNA can induce a heparanase-specific CTL, which showed a potent lysis of gastric carcinoma cells with matching MHC in positive expression of heparanase, while having no effects on cells without heparanase-matching MHC, indicating that CTL epitopes must exist that can induce the specific CTL in the structure of heparanase (17). Recently, Sommerfeldt et al. (18) identified three HLA-A2–restricted heparanase epitopes. They found that CTLs elicited by these three epitopes could lyse heparanase-transfected and heparanase-expressing untransfected breast cancer cells in vitro. According to these results, they concluded that heparanase could be an attractive new TAA, and that its HLA-A2–restricted peptides ought to be good candidates for peptide vaccination through reactivation of memory immune responses to invasive and metastatic cancer cells.

Although several epitopes of heparanase antigen are known in humans, the corresponding knowledge in the mouse is still rather limited. Here, we first predicted candidate epitopes from mouse heparanase antigen based on computer algorithms. Computational methods combined with in vitro/in vivo studies have proven to be very useful in the identification of immunogenic T-cell epitopes from defined antigens and pathogens (33, 34). Many groups have shown the utility of the matrix-based algorithms, such as BIMAS and SYFPEITHI, in predicting viral- and tumor antigen–specific CTL epitopes. Using the matrix-based algorithm, BIMAS, SYFPEITHI, and IMTECH, we predicted five H-2K^b–restricted peptides within the structural proteins of mouse heparanase.

Second, to validate the ability of these predicted epitopes to ligate H-2K^b, we used a quantitative flow cytometric assay to measure surface induction of class I molecules on RMA-S TAP-deficient cells. This commonly used method measures the relative binding affinity of different peptides to the same class I allele (35). We found that five candidates had high affinity to H-2K^b when the concentration of peptides was >30 μmol/L.

Third, to verify whether these five candidates could induce CTL responses, denditric cells generated from mouse bone marrow were pulsed with the one of the five peptides and were then used to immunize the mouse s.c. thrice. Splenocytes from mouse spleen served as effectors. We found that effectors induced by mHpa398 and mHpa519 could lyse B16 melanoma cells effectively, whereas effectors induced by mHpa234, mHpa351, and mHpa386 could not lyse B16 melanoma cells. Further study showed that effectors induced by mHpa398 and mHpa519 could also lyse EL-4 lymphoma cells and Lewis lung cancer cells; conversely, these heparanase peptide-specific CTLs could not lyse P815 mastocytoma cells, which were heparanase positive but H-2K^b negative. To determine whether the effector CTLs induced by mHpa398 and mHpa519 peptides could recognize the heparanase-positive target tumor cells in an H-2K^b–restricted manner, we examined the inhibition of cytotoxicity with anti–H-2K^b mAbs. These results showed that the cytotoxic activity of effectors against targets could be eliminated by anti–H-2K^b mAbs.

Finally, we observed the immunoprotection and immunotherapy generated by s.c. injection of mHpa398- and mHpa519 peptide–pulsed denditric cells. The results showed that mHpa398 and mHpa519 peptides not only offered the possibility of immunizing against tumors but also of treating tumor-bearing hosts successfully. Thus, our present study indicates that peptide mHpa398 (LSLLFKKL) and mHpa519 (FSYGFFVI) are novel H-2K^b–restricted CTL epitopes capable of inducing heparanase CTLs in vitro and in vivo.

For the development of cancer vaccines, safety concerns still limit the use of these epitopes in the near future. Although heparanase expression is largely restricted to cancer cells, it has been detected in activated immune cells, including T and B cells, dendritic cells, macrophages, neutrophils, and mast cells mediating extravasation and traffic to inflammatory sites (4, 17, 36, 37). Consequently, any heparanase-based cancer vaccine therapy will require thorough assessment of the potential adverse effects associated with autoimmunity to cells and organs that are heparanase positive. Our previous study showed that CTLs induced by denditric cell–loaded heparanase full-length cDNA could not lyse autologous lymphocytes in vitro. In this study, to investigate the effects of heparanase peptide–specific CTLs on immunologically activated lymphocytes and denditric cells, heparanase peptide–specific CTLs were also used to lyse autologous lymphocytes and denditric cells. The results revealed that heparanase peptide vaccination did not markedly lyse these lymphocytes and denditric cells. Moreover, no CTL-mediated toxicity against normal tissues was observed in the immunoprotection and immunotherapy experiments (data not shown). It may be conceivable that the level of heparanase expression in normal cells is below the threshold needed for recognition by these heparanase peptide–specific CTL populations (17).

In summary, our results suggest that the mHpa398 (LSLLFKKL) and mHpa519 (FSYGFFVI) peptides derived from the mouse heparanase protein might be capable of inducing an H-2K^b–restricted CTL reaction, which would be lethal for tumor cells expressing heparanase and H-2K^b in vitro and in vivo. Moreover, these heparanase peptide–specific CTLs could not lyse autologous lymphocytes and denditric cells that also express heparanase. To our knowledge, this is the first report describing mouse heparanase epitopes. As a counterpart to human heparanase, our systematic studies in mouse heparanase indirectly show that heparanase CTL epitopes can serve as a new target for the immunotherapy of advanced tumors.

Grant support: National Nature Science Foundation of China (30570841 and 30200123).

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 thank Dr. Wan Y (Institute of Immunology of People's Liberation Army, Third Military Medical University, China) for many constructive suggestions on epitope prediction.