Tumor vaccine is a useful strategy for cancer therapy. However, priming of the immune system requires the relevant antigen to be presented by antigen-presenting cells (APCs). Here, we employed telomerase reverse transcriptase as a model antigen to explore the feasibility of using mannan-modified adenovirus as a tumor vaccine. We found that tumor immunogene therapy with the vaccine was effective at protective antitumor immunity in mice. The antigen-specific cytotoxic T lymphocytes were found in in vitro cytotoxicity assay. The elevation of the killing activity could be abrogated by anti-CD8 or anti-major histocompatibility complex-I antibodies. Adoptive transfer of purified CD8+ cells, and CD4+ cells to a less extent, was effective at antitumor activity. In vivo antitumor activity could be abrogated by depleting CD4+ T lymphocytes. A possible explanation for the antitumor effects may be the antigen was transfered to APCs in the presence of mannan. These observations provide insights into the design of novel vaccine strategies and might be important for the future application of antigens identified in other diseases.
Cytotoxic T lymphocyte (CTL) is a critical component of host immunity against tumors.1, 2 Priming of CTL responses requires the relevant antigen presented by antigen-presenting cells (APCs). Tumor vaccines have been proposed to enhance the antigen presentation by APCs, especially dendritic cells (DCs). These include DC pulsed with tumor-associated antigens (TAAs), DC fused with whole tumor cell and DC transduced with mRNA or viral vectors.3, 4, 5 However, preparation of patient-specific DC vaccine ex vivo is both time-consuming and expensive, which poses a great challenge for up-scale production in clinical setting. Efforts are therefore continuing to develop a new strategy for tumor vaccine.
The observation that telomerase reverse transcriptase (TERT), the catalytic subunit of telomerase, is elevated in most tumors of human and murine origin6, 7 suggests its potential value for immunotherapy against cancer.8 It was demonstrated that CTLs can recognize peptides from TERT and kill TERT-positive tumor cells, both in human9, 10, 11 and murine systems.12 Induction of anti-TERT immunity in mice vaccinated with DCs transduced with mouse TERT mRNA was reported.12 After vaccination, mice were partially protected from tumor challenge. In another study, CTLs against human TERT could be elicited following stimulation ex vivo with autologous DCs transduced with whole tumor mRNA in cancer patients.10, 11 The breaking of immune tolerance to autologous TERT molecule should be a useful approach for cancer therapy.
The replication-deficient adenovirus has been widely used as vectors for gene delivery in preclinical models and clinical indications. It is easy to produce in high titers, readily transduces non-proliferating cells and does not integrate into the host genome. Recently, a series of studies have demonstrated the efficiency of recombinant adenovirus encoding TAAs as tumor vaccine.3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 However, the lack of cell specificity appears to be one of the key problems in adenovirus-based therapies.
The limitation of broad native tropism could be solved by conferring the targeting capability to adenovirus. Strategies currently are under development, with use of heterologous retargeting complexes such as chimeric fusion proteins and bispecific antibodies, or genetic capsid modifications.16, 17, 18 APC including DC or macrophage expresses mannose receptors (MRs) on its surface.19 Previous study showed that conjugation of a recombinant antigen (mucin-1 (MUC1)) to mannan under appropriate oxidizing condition could stimulate cellular immune responses.20 The possible mechanism might be that the presence of mannan serves to bind the antigen to DC through MRs. Also, MR was implicated as an important receptor for the recognition of enveloped viruses such as herpes simplex virus by DC and the subsequent stimulation of interferon-α production.21 We may reason that the recombinant adenovirus conjugated with mannan could load the encoded antigen into DC through MRs in vivo, thus it would greatly alleviate the problem of time-consuming and costly process of DC preparation ex vivo. To explore the feasibility of redirecting the adenovirus in this manner, our strategy was to couple the polysaccharide mannan (polymannose) to the surface of the adenovirus chemically.
Here, an adenovirus expressing TERT was constructed. It was then either conjugated with (AdT-m) or without mannan (AdT) as vaccines. The vaccines were tested for the ability to induce antitumor immunity in mouse tumor models. In addition, mice treated with adenovirus harboring no foreign expression cassette (Ad-null) and phosphate-buffered saline (PBS) were also included as controls.
Mannan conjugation efficiently delivers antigen to DC in vivo
We tested our hypothesis that the adenovirus coupled with mannan could deliver its antigen to DC in vivo. The adenovirus encoding green fluorescent protein (GFP) (AdG) was conjugated with mannan (AdG-m). The efficacy of the conjugation in vitro was confirmed by periodic acid schiff (PAS) staining. The PAS stain is a histochemical reaction commonly used to detect glycogen in biological specimen. We found it stained positive for the conjugation group, whereas negative for the unconjugated control (Figure 1). In addition, the mean particle sizes of the mannan-conjugated adenovirus and unconjugated adenovirus were 170–210 and 180–230 nm, respectively. This implied that there might be no much difference in physicochemical properties between the two groups.
Intraperitoneal (i.p.) immunization with AdG-m led to in vivo expression of GFP, which might be used as a surrogate antigen. We detected GFP expression in the spleen of mouse 3 days post i.p. immunization with AdG-m. Spleen from mock-infected (immunized with PBS) group showed no green signal (Figure 2a). In the AdG immunized control group, the distribution of green signal was nonspecific (Figure 2b). However, in the AdG-m immunized spleen, green cells were arranged in small clusters (Figure 2c) and were detected up to 7 days after inoculation (data not shown). The GFP-expressing cells display morphology consistent with DC and also express the DC surface marker CD11c (Figure 2d).
The adenovirus conjugated with mannan inhibits tumor growth
To further test the efficiency of the vaccine, we constructed replication-deficient adenovirus encoding TERT (AdT). Then the adenovirus was linked to mannan (AdT-m). We immunized mice i.p. twice at a 2-week interval with different doses of AdT-m, AdT and Ad-null or treated them with PBS alone (mock infection), and then challenged with 1 × 106 live tumor cells at day 7 after the second immunization. There was apparent protection from tumor growth in mice vaccinated with AdT-m (Figure 3a). The treatment with AdT also showed inhibitory effects on tumor growth, compared with PBS or Ad-null alone, but was less effective than the AdT-m. Furthermore, the survival of the tumor-bearing mice treated with AdT-m was also longer than that of the mock-infected mice or mice immunized with AdT or Ad-null (Figure 3b).
CD8+ T cells were responsible for the antitumor immunity
T cells isolated from the spleens of AdT-m immunized mice recognized and killed TERT-positive syngeneic tumor cells. However, the T cells failed to lyse TERT-positive allogeneic tumor cells, nor TERT-positive normal mesenchymal stem cells (MSCs) (Figure 4a). T cells from the AdT-m vaccine group exhibited higher cytotoxicity against TERT-positive syngeneic tumor cells than those from the control groups (Figure 4b). This cytotoxicity could be blocked by anti-CD8 or anti-major histocompatibility complex (MHC) class I monoclonal antibody (mAb), but not by anti-CD4 or anti-MHC-II in vitro, suggesting that the killing activity observed may result from MHC class I-dependent CD8+ CTL activity (Figure 4c).
Mice depleted of CD8+ or CD4+ T lymphocytes by the injection of mAb against CD8 or CD4 and vaccinated with AdT-m were not protected from tumor challenge. Treatment with mAb against natural killer (NK) cells showed partial abrogation, whereas treatment with normal rat immunoglobulin (Ig) failed to abrogate the antitumor activity (Figure 5). Mice depleted of CD4+ T lymphocytes did not develop detectable CTL response against tumor cells (data not shown).
Adoptive transfer of CD4-depleted (CD8+), or CD8-depleted (CD4+) T lymphocytes isolated from mice immunized with AdT-m vaccine showed the antitumor activity against TERT-positive tumor cells. The transfer of T lymphocyte subsets from mice immunized with AdT or mock-infected mice had no effect (Figure 6).
Observation of potential toxicity from the immunization
Vaccinated animals without tumor were particularly investigated for potential toxicity for more than 10 months. No adverse consequences were indicated in gross measures such as weight loss, ruffling of fur, life span, behavior or feeding. No pathologic changes of liver, lung, kidney, spleen, brain, heart or bone marrow were found by microscopic examination.
For evaluation of the effects of AdT-m on hematopoiesis, animals were investigated for more than 10 months and were subjected to complete peripheral blood counts and differentials. In addition, liver function was assayed and no abnormality was found (data not shown).
Among many different immunotherapeutic strategies currently evaluated, cancer vaccines based on tumor antigens represent promising approaches for cancer therapy. Priming of CTL responses frequently, if not always, requires the foreign antigen presented by APC, especially DC. However, to date, manipulation of DC ex vivo is still complex and expensive. As expected, tumor vaccines based on DC transfusion have limited application in clinical setting. Therefore, we seek to improve the efficacy of tumor vaccine through promotion of antigen presentation.
The present study, to our knowledge, was the first description of the use of mannan-modified adenovirus as an APC-targeting tumor vaccine. Our findings suggest that MHC class I-dependent CD8+ CTL-mediated immune response may be responsible for the observed antitumor activity. In our preliminary experiments, mice treated with mixture of non-conjugated mannan and AdT, Ad-null conjugated with mannan or mannan alone showed no protection (data not shown). Based on the findings mentioned, we may rule out the possibility that the antitumor activity may result from a nonspecifically augmented immune response against tumor growth in host mice.
In this work, we found that a specific CTL response against telomerase-positive tumors could be induced in mice immunized with mannan-modified adenovirus. The effects were comparable with those of a previous study.12 However, in the latter study, DCs were bone marrow derived, and generated in vitro in the presence of granulocyte macrophage colony-stimulating factor and interleukin-4. This process is both difficult and expensive. Compared with the current complex approach of isolating and pulsing DCs with tumor antigens, our method of targeting and activating DCs was preformed in vivo instead of in vitro. Recently, DC-based immunotherapy achieved some success in treating cancer patients.22, 23, 24 Our method of manipulating APCs in vivo would have potential for future vaccine design.
In this study, we found that mice depleted of CD4+ T lymphocytes by the injection of anti-CD4 antibody and vaccinated with AdT-m were not protected from tumor inoculation. At the same time, mice depleted of CD4+ T lymphocytes did not develop detectable CTL activity. These findings suggested that the induction of the CTL response to TERT, which is responsible for the observed antitumor activity, may involve CD4+ T lymphocytes. These suggestions were further supported by the important roles of CD4+ T lymphocytes in the antitumor immunity.25, 26 These findings may help explain the requirement for CD4+ T lymphocytes in the induction of CTL response against tumor cells in the present study. It is known that CD4+ T lymphocytes can steer and amplify immune responses through the secretion of cytokines and the expression of surface molecules.27, 28 However, antitumor CTL response mainly depends on CD8+ T lymphocytes in some mouse models, whereas CD4+ lymphocytes often have little, if any, function.1, 2, 28 In fact, CD8+ T lymphocytes have been the focus of recent efforts in the development of a therapeutic antitumor vaccine. We speculated that this might be the reason why we failed to find the roles played by CD4+ T lymphocytes in our in vitro cytotoxicity assay.
Several strategies have been proposed for preparation of DC vaccines in vivo, including co-expressing Fas and the immunogen in the same cell, subcutaneous implantation of ethylene-vinyl-acetate polymer rods releasing macrophage inflammatory protein-3 to entrap migratory Langerhans cells (LCs), and co-administration of immunostimulatory DNA and antigen in mice pretreated with Flt3 ligand.29, 30, 31 However, our approach is different in several ways compared with the reports above. First, our vaccine is based on a replication-deficient adenovirus encoding a tumor antigen without use of any adjuvant. Second, mice in our study only received two times of i.p. injections, compared with the complex immune protocols. Third, our study was based on an immunocryptic antigen but not an immunodominant model antigen. Our approach is more similar to a clinical setting.
It has been reported that conjugation of a tumor antigen (MUC1) with mannan can stimulate cellular immune responses.20 Here, we coupled the adenovirus with mannan, and found the conjugation could efficiently target the antigen to DC. The MRs on the surface of DCs may play a pivotal role, as GFP could be presented by DC only in the conjunction of mannan. In immune system, protein–glycan interactions may contribute to immune recognition. For example, in rheumatoid arthritis, specific glycoforms may induce association with mannose-binding lectin and contribute to the pathology.32 Some kinds of viruses use the host glycosylation machinery to assemble their own envelope glycoproteins and in such manner they interact with immune recognition.33
We paid special attention to investigating potential toxicity in the mice immunized and have not found marked adverse effects. In mice, TERT is found in germ cells, hematopoietic progenitors, activated lymphocytes, and some other organs.34, 35, 36 In our hands, the T lymphocytes from immunized mice failed to lyse MSCs, whose TERT expression is positive. In accordance to this, in vivo findings also showed that the treatment was well tolerated in mice, no signs of toxicity were observed and no organ abnormalities were seen at autopsy. These data were also supported by the previous reports.12 The possible explanation was the expression level of TERT in normal tissue might not reach to some threshold.22 However, it is unwise to conclude that telomerase-based therapy will be safe in human, although some pivotal clinical trials have confirmed the safety in some patients.22, 23, 24
Taken together, our findings may provide a new tumor vaccine strategy through targeting APC in vivo with the mannan-modified adenovirus. Because of the simplicity, the method may be used to target other antigens as well, including those of infectious pathogens. The mannan-modified adenovirus as vaccines may be of importance for the future exploration.
Materials and methods
Construction of adenovirus vectors
The adenovirus vectors were constructed in a bacteria homologous recombination system.37 The mouse TERT full-length complementary DNA was amplified from total RNA isolated from the mouse liver by reverse transcriptase-polymerase chain reaction using primers derived from the published sequence of mouse TERT. When the product of the expected size was obtained, it was inserted into pCR2.1 plasmids (Invitrogen, San Diego, CA, USA) and then subcloned into pShuttle-CMV. The shuttle vector was linearized with PmeI, and cotransfected with AdEasy-1 backbone plasmid into a recombinant permissive Escherichia coli strain BJ5183. The structure of the resultant recombinant vectors was confirmed by restriction enzyme digest. The correct construct was digested with PacI, transfected 293 cells by Lipofectamine (Life Technologies Inc., Rockville, MA, USA) and cultures were monitored for cytopathic effect. The expression was confirmed by Western blot. The recombinant adenovirus was purified by CsCl gradient ultracentrifugation and titered by plaque assay (PFU/ml).
Adenovirus and mannan conjugation
Mannan (Sigma, St Louis, MO, USA), at 14 mg/ml in 0.1 M phosphate buffer (pH 6.0), was oxidized with sodium periodate (0.01 M) for 60 min at 4°C. Ethanediol (10 μl) was added and incubated for a further 30 min at 4°C and the mixture was passed through a Sephadex-G25 column equilibrated in bicarbonate buffer (pH 6.0–9.0). The oxidized mannan that eluted was mixed with recombinant adenovirus, incubated overnight at room temperature and used without further purification.
We used PAS stain to confirm the efficacy of the in vitro conjugation. Briefly, the conjugated adenovirus was subjected to the round of ultracentrifugation purification to separate the free (unconjugated) mannan from the conjugated adenovirus. Then the adenovirus fraction was collected and incubated with PAS agent under room temperature. The positive stain showed a magenta color. In addition, the mean particle size of the conjugated or unconjugated adenovirus was measured by the Zetasizer Nano ZS (Malvern, PA, USA).
Immunofluorescence assays on spleens
Mice received 1 × 108 PFU/100 μl Ad-GFP with (AdG-m) or without conjunction to mannan (AdG). Three days later, the spleens were removed and frozen in optimal cutting temperature (OCT) compound. Frozen sections (5 μm) were made using a Leica 1800 cryostat (leica Inc., Deerfield, IL, USA). Following sectioning, samples were fixed in cold acetone for 10 min, and then incubated 1 h at 37°C with phycoerythrin-conjugated mAb to CD11c (BD PharMingen, San Diego, CA, USA). The samples were viewed with an Olympus fluorescing microscope.
Tumor models and immunization
B16 melanoma was established in C57BL/6 (H-2b). CT26 colon cancer, and Meth A fibrosarcoma models were established in BALB/c mice (H-2d). The adenovirus conjugated with mannan (AdT-m) was used as a vaccine. The adenovirus without mannan (AdT), and a ‘null’ adenovirus (Ad-null) were used as controls. Administration route is an important factor for transfection to DCs. I.p. administration has advantages as far as transfection efficacy to DCs is concerned.38, 39 The adenovirus vaccine was given to mice at 6–8 weeks of age via i.p. route at an interval of 2 weeks with different doses (1 × 106–1010 PFU/mouse). Additional control animals were injected with PBS (mock infection). Mice were then challenged with 1 × 106 live tumor cells 1 week after the second immunization. All studies involving mice were approved by the institute's Animal Care and Use Committee.
In vivo depletion of immune cell subsets
Immune cell subsets were depleted as described previously.40 Mice received i.p. injections of 500 μg of either the anti-CD4 (clone GK1.5, rat IgG), anti-CD8 (clone 2.43, rat IgG), anti-NK (clone PK136) mAb or isotype controls 1 day before the immunization and then twice per week for 3 week. Mice were challenged with tumor cells (LLc and Meth A, 2 × 106) after the boost immunization. These hybridomas were obtained from American Type Culture Collection (ATCC). The depletion of CD4+, CD8+ and NK cells was consistently >98%, as determined by flow cytometry (Coulter Elite Esp, Coulter, Hialeah, FL, USA).
In vitro cytotoxicity assay
A 4-h 51Cr release assay was performed as we described and previously described by other reports.40 Briefly, splenocytes obtained from the immunized or control mice were treated with ammonium chloride lysing buffer to deplete erythrocytes. T-enriched cell fraction was prepared by nylon wool purification. NK cells were also depleted using NK1.1 mAb (clone DX5, BD PharMingen) as described previously. A total of 100 μl of effector cells and 51Cr-labeled target cells were assigned at different E:T ratios to each well of microtiter plates and incubated for 4 h at 37°C. Samples were then harvested, and the activity was calculated by the formula: %cytotoxicity=((experimental release–spontaneous release)/(maximum release–spontaneous release)) × 100. To evaluate NK activity, splenocytes treated with ammonium chloride lysing buffer to deplete erythrocytes were used as effectors against YAC-1 NK target cells. The spontaneous 51Cr release usually did not exceed 20% of the total isotope count.
In the cytotoxicity inhibition assays, effectors cells or 51Cr-labeled tumor cells were pretreated with mAb at room temperature for 30 min, washed and tested. mAbs used included anti-CD4 (10 μg/ml), anti-CD8 (10 μg/ml), anti-H-2Kb/H-2Db (50 μg/ml) (BD PharMingen).The above concentrations of mAb were effective in mediating their activity in preliminary experiments. Control cytotoxicity assay was performed in the presence of control mAb (anti-H-2Dd) or isotype IgG.
Purification of CD4+ or CD8+ T cells in vitro and their adoptive transfer in vivo
Briefly, nylon wool-purified splenic T cells (107 cells/ml) in complete medium were incubated on ice for 45 min with previously determined optimal amounts of purified mAb to mouse CD4 (10 μg/ml, clone GK1.5; ATCC) or anti-CD8 (10 μg/ml, clone TIB 211; ATCC). The cells were then washed with Roswell's Park Memorial Institute 1640 medium supplemented with 5% fetal bovine serum. The lymphoctytes were then incubated at 37°C for 1 h in the presence of complement (rabbit Low-Tox-M; Cedarlane Laboratories, Hornby, ON, Canada). Viable cells were recovered by centrifugation through a 1.119 g/cm3 Ficoll gradient (Sigma-Aldrich, St Louis, MO, USA), and were washed twice in complete medium before counting (>95% viable by trypan blue exclusion). γδ T cells and NK cells were also depleted using anti-TCR-γδ (clone UC7-13D5; BD PharMingen) and anti-NK1.1 (clone DX5; BD PharMingen) plus complement as described previously. Furthermore, Abs plus complement-treated cells were removed of B cells and adherent cells by panning on anti-mouse Ig-coated dishes, as described previously. Depletion of immune cell subsets was confirmed by flow cytometry (Coulter Elite ESP; Coulter, Miami, FL, USA) using fluorescein isothiocyanate-labled anti-CD4, anti-CD8, anti-NK and anti-γδ T (BD PharMingen). In all cases, the specific cellular depletions were 99% effective. One day after the adoptive transfer of 2 × 106–2 × 107 cells, the mice were challenged with 1 × 106 tumor cells.
For comparison of individual time points, analysis of variance and an unpaired Student's t-test were used. Survival curves were constructed according to the Kaplan–Meier method. Statistical significance was determined by the log-rank test.
This work was supported by National Key Basic Research Program of China (2004CD518800 and 2001CB510001), Project of National Natural Sciences Foundation of China, National 863 projects.
About this article
Trends in Biotechnology (2009)