Memory and cellular immunity induced by a DNA vaccine encoding self antigen TPD52 administered with soluble GM-CSF

Abstract

Tumor protein D52 (TPD52) is involved in cellular transformation, proliferation and metastasis. TPD52 over expression has been demonstrated in several cancers including prostate, breast, and ovarian carcinomas. Murine TPD52 (mD52) has been shown to induce anchorage independent growth in vitro and metastasis in vivo, and mirrors the function and normal tissue expression patterns of the human orthologue of TPD52. We believe TPD52 represents a self, non-mutated tumor associated antigen (TAA) important for maintaining a transformed and metastatic cellular phenotype. The transgenic adeno-carcinoma of the mouse prostate (TRAMP) model was employed to study mD52 as a vaccine antigen. Naïve mice were immunized with either recombinant mD52 protein or plasmid DNA encoding the full-length cDNA of mD52. Following immunization, mice were challenged with a subcutaneous, tumorigenic dose of mD52 positive, autochthonous TRAMP-C1 tumor cells. Sixty percent of mice were tumor free 85 days post challenge with TRAMP-C1 when immunized with mD52 as a DNA-based vaccine admixed with soluble granulocyte-macrophage colony stimulating factor (GM-CSF). Survivors of the initial tumor challenge rejected a second tumor challenge given in the opposite flank approximately 150 days after the first challenge, and remained tumor free for more than an additional 100 days. The T cell cytokine secretion patterns from tumor challenge survivors indicated that a T_H1-type cellular immune response was involved in tumor protection. These data suggest that mD52 vaccination induced a memory, cellular immune response that resulted in protection from murine prostate tumors that naturally over express mD52 protein.

Introduction

Chromosome 8q gain is one of the most frequently identified cytogenetic aberrations in human cancer [1]. Amplification mapping studies have identified several candidate cancer associated genes at chromosome 8q21 [2, 3] the most definitive of which is tumor protein D52 (TPD52 or D52) at 8q21.13 [4]. All D52-like sequences either identified or predicted to date include coiled-coil domains, suggesting that protein–protein interactions are integral to their function [5]. Human D52 (hD52) protein is over expressed in breast [4, 6, 7], prostate [8–10] and ovarian cancers [11], and is likely a result of increased gene copy number. Expression microarray analyses predict hD52 over expression in many other cancer types, including multiple myeloma [12, 13], Burkitt’s lymphoma [14, 15], pancreatic cancer [16], testicular germ cell tumors [17–19], and melanoma [20, 21].

The murine orthologue of TPD52 (mD52) naturally mirrors hD52 with respect to known function and over expression in tumor cells, and shares 86% protein identity with the human orthologue [22]. Recently we demonstrated that transfection and stable expression of mD52 cDNA in mouse 3T3 fibroblasts (3T3.mD52) induced increased proliferation, anchorage independent cell growth, and the ability to form subcutaneous tumors and spontaneous lethal lung metastases in vivo when 3T3.mD52 cells were inoculated subcutaneously into naïve, syngeneic, immuno-competent mice [23]. Together these data strongly suggest that D52 expression may be important for initiating and perhaps maintaining a tumorigenic and metastatic phenotype and thus may be important for tumor cell survival.

Scanlan and colleagues identified hD52 as a candidate breast cancer tumor associated antigen (TAA) by using sera from breast cancer patients to screen a library of expressed genes from breast cancer tissue, demonstrating that hD52 protein in tumor cells is capable of inducing IgG antibodies [24]. This report suggests that hD52 protein may be immunogenic and capable of inducing a cellular immune response, thus warranting study of TPD52 protein as an anti-cancer vaccine. In a recent study, we demonstrated for the first time that mD52 is immunogenic when administered as recombinant protein-based vaccine admixed with CpG/ODN 1826 in mice. The immune response generated was capable of rejecting tumor cells that naturally over express mD52 protein without inducing harmful autoimmunity [25]. These data suggest that hD52 protein may also be a potent vaccine antigen that could be administered to patients to treat or prevent cancers that over express hD52.

In the present study, we sought to determine whether mD52 DNA vaccination would induce an immune response capable of rejecting tumors in the transgenic adenocarcinoma of the mouse prostate (TRAMP) model of prostate cancer [26]. Approximately 60% of mice were tumor free 85 days post challenge with autochthonous TRAMP-C1 tumor cells when immunized with mD52 as a DNA vaccine admixed with soluble granulocyte-macrophage colony stimulating factor (GM-CSF). Survivors of initial tumor challenge rejected a second tumor challenge given in the opposite flank more that 150 days after the initial challenge was given. These mice remained tumor free for more than an additional 100 days. The T cell cytokine secretion patterns from tumor challenge survivors indicated that a T helper 1-type cellular immune response was involved in tumor protection. Together, these data suggest that mD52 DNA-based vaccination induced a cell-mediated, memory immune response that resulted in protection from prostate tumors that naturally over express mD52 without the induction of discernable harmful side effects.

Materials and methods

Mice and tumor cell lines

Male 6- to 8-week-old C57BL/6 mice were purchased from Jackson Labs (Bar Harbor, ME). All animals were cared for and treated according to Institutional Animal Care and Use Committee guidelines at Texas Tech University Health Sciences Center (Lubbock, TX, USA). The tumorigenic, autochthonous C57BL/6 cell lines TRAMP-C1 and TRAMP-C2 [27] were used for tumor challenge and/or immunoassays, and the tumorigenic SV40-transformed Balb/c murine kidney cell line designated mKSA was used as an mD52 positive MHC mis-matched control target for immunoassays. The mKSA cell line was cultured in RPMI 1640 (Fisher Scientific, Pittsburgh, PA, USA) supplemented with 10% heat-inactivated fetal bovine serum, 2 mmol/l l-glutamine, 250 ng/ml fungizone, 50 IU/ml penicillin, 50 μg/ml streptomycin, 50 μg/ml gentamicin sulfate, and 10 mmol/l HEPES. The autochthonous TRAMP-C1 and TRAMP-C2 cell lines were cultured as previously reported [27].

Immunization and tumor cell challenge

Individual mice were immunized (s.c.) with 100 μg of mD52 DNA admixed with 5 μg of recombinant murine GMCSF for a total of 2 injections given at 14 day intervals followed by a booster immunization given 35 days after the second immunization. Empty vector DNA (pCDNA 3.1 vector minus mD52 cDNA) served as a control immunization. Mice in all groups were bled from the dorsal tail vein prior to immunization and 2 weeks following each immunization. Two weeks following the final immunization, mice in all groups were challenged with a tumorigenic dose of autochthonous TRAMP-C1 (5 × 10⁶) tumor cells. Tumor cells were harvested, counted and re-suspended in PBS (Fisher Scientific, Pittsburg, PA, USA) and 100 μl of viable cell suspension was injected subcutaneously in the right flank of each mouse. Tumor size was determined by taking perpendicular measurements with calipers every 2–3 days and tumor volume (mm³) was calculated using the following formula: (a × b ²)/2, where b is the smaller of the two measurements. Mice that survived the primary challenge were re-challenged in the opposite flank with TRAMP-C1 (5 × 10⁶) approximately 150 days after the initial challenge.

To compare mD52 DNA with recombinant mD52 protein as active vaccines, individual mice were immunized either i.m. or s.c. four times at 14 day intervals with 10 μg of recombinant mD52 protein admixed with 10 μg of CpG oligonucleotide (ODN 1826: TCCATGACGTTCCTGACGTT) [25]. The protein vaccines were administered as an alum precipitate and a booster of the same dose was given approximately 2 weeks following the third immunization for a total of four injections. mD52 DNA immunizations were administered as described above. Two weeks after the booster immunizations mice were challenged s.c. with TRAMP-C1 tumor cells (5 × 10⁶) and tumor size was monitored as described above.

Analysis of cytotoxic T lymphocyte (CTL)-mediated tumor cell lysis

T cells from immunized mice were stimulated in vitro by culturing ficol separated spleen cells from immunized mice that survived tumor challenge and subjecting them to standard CTL-mediated tumor cell lysis analysis. CTLs were generated by culturing spleen cells with irradiated TRAMP-C1 tumor cells (using the same tumor cell line as was used for the in vivo challenge) in the presence of IL-2 (10 ng/ml), IL-7 (5 ng/ml), and IL-12 (5 ng/ml) at 37°C for 5–7 days. Specificity was evaluated by mixing various numbers of CTLs with a constant number target cells (5 × 10³ cells per well) in 96 well round bottom plates. Specific lysis was determined using a Europium time-resolved fluorescence based 2 h method and measured using a Victor3™ plate reader (Perkin Elmer, Boston, MA, USA). Percent lysis was calculated as: % specific lysis = 1 − (E − S)/(M − S) × 100, where E represents Eu release in the presence of effector cells, S is spontaneous Eu release in medium alone and M represents maximum Eu released in the lysis buffer [25, 28]. To confirm H-2^b-restricted tumor recognition tumor targets included TRAMP-C1 tumor cells (H-2^b+, mD52⁺; cell line used for tumor challenge), TRAMP-C2 tumor cells (H-2^b+, mD52⁺), and mKSA (Balb/c) tumor cells which served as a control MHC mis-matched target (H-2^d+, mD52⁺).

T cell culture and ELISAs for cytokine production

T cells from mD52 DNA immunized mice were stimulated in vitro by culturing ficol separated spleen cells with irradiated TRAMP-C1 tumor cells (the same tumor cell line used for the in vivo challenge) in the presence of IL-2 (10 ng/ml), IL-7 (5 ng/ml), and IL-12 (5 ng/ml) at 37°C for 5–7 days. Assessment of cytokine secretion by tumor-specific T cell cultures was accomplished by applying culture supernatants to commercially available sandwich ELISA kits for IFN-γ, IL-10, IL-4 and IL-17 detection (R&D Systems, Minneapolis, MN, USA). Culture supernatants were harvested from 24 h cultures of T cells (1 × 10⁶ cells/ml in 200 μl of medium in 96 well plates) in medium alone, compared to T cells cultured with various tumor cell targets (1:1 ratio). TRAMP-C1 tumor cells (H-2^b+, mD52⁺; cell line used for tumor challenge), TRAMP-C2 tumor cells (H-2^b+, mD52⁺), mKSA (Balb/c) tumor cells which served as a control MHC mis-matched target (H-2^d+, mD52⁺), and Yac-1 as a control for the presence of non-specific cells. To confirm MHC-I restricted tumor recognition, blocking assays were performed by incubating TRAMP-C1 tumor cells with anti-H-2^b or anti-H-2^d (negative control) mAb prior to incubation with T cells. Briefly, 10 μl of mAb in PBS (final concentration of 30 μg/ml) was added to individual wells of 96 well round bottom plates in duplicate. Next, 100 μl TRAMP-C1 tumor targets was added to each well and incubated for 30 min at room temperature. Finally, 100 μl of T cell effectors was added to the appropriate wells and the plates were incubated for 24 h at 37°C. Assays were analysed using the Victor3™ plate reader (Perkin Elmer, Boston, MA, USA). We preformed all assays with the internal controls provided by the manufacturer and the standards from which standard curves were generated in order to determine concentration of cytokines produced in experimental sets. All the controls provided by the manufacturers worked indicating that the assays were able to detect the cytokines in question.

Elispot assay for IFN-γ production

Following immunization with mD52 DNA and subsequent TRAMP-C1 tumor cell protection, T cell responses specific for TRAMP-C1 tumor cells were assessed using a murine IFN-γ Elispot assay following the manufacturer’s instructions (R&D Systems, Minneapolis, MN, USA). Briefly, splenic mononuclear cells from immunized mice that successfully rejected in vivo TRAMP-C1 tumor challenge were isolated and T cells were cultured either alone as a negative control, with 5 μg/ml of the T cell mitogen CON-A as a positive control, with mKSA control tumor cell target (Balb/c H-2^d, MHC mismatched), or with TRAMP-C1 tumor cells and incubated overnight at 37°C and 5% CO₂. T cells were plated at 100,000 cells (100 μl)/well with irradiated tumor cells at 1,000 cells (20 μl)/well. Following T cell/tumor cell incubation, wells were washed a total of four times. Next, 100 μl of detection antibody was added per well and incubated overnight at 4°C. Following incubation, plates were washed four times as above and 100 μl of streptaviden-AP was added per well and incubated for 2 h at room temperature. After four plate washes, 100 μl of developing solution was added per well and incubated 1 h at room temperature in the dark. Finally, plates were rinsed with water and dried completely at room temperature. The AID Viruspot automated ELISpot reader (CTI Technologies Inc., Columbia, MD, USA) was used to analyze spot counts. Data are represented as the number of IFN-γ spots per 100,000 cells.

Flow cytometry

Lymphocytes from spleens cultured in vitro with TRAMP-C1 as described above were stained with monoclonal antibodies specific for CD3, CD4, CD19, NK marker and CD8. MHC class I expression was assessed on tumors cell lines. Antibodies were purchased from BD-Bioscience (San Jose, CA, USA). Cells were fixed in 1% paraformaldehyde at 4°C for 1 h and then analyzed by flow cytometry using a BD-FacsVantage™.

Statistical analysis

When necessary, tumor challenge data were analyzed with a t test to determine whether significant differences existed between mean tumor volume for mD52 immunized and control immunized mice. Elispot and ELISA data were analyzed using an unpaired t test for independent samples with equal variances. A P value of less than 0.05 was taken as being of statistical significance (GraphPad Prism 5.0).

Results

Superior CTL induction following immunization with mD52 DNA

Since mD52 is an intracellular protein we were interested in determining if MHC class I restricted CTLs were induced in mice following immunization with mD52 DNA and GM-CSF or mD52 protein and CpG/ODN. Previously, we reported that TRAMP-C1, TRAMP-C2 (C57BL/6, H-2^b) and mKSA (Balb/c H-2^d) all over express mD52 making these cell lines suitable targets and controls with which to address mD52 vaccine induced tumor immunity [25]. Groups of mice were immunized subcutaneously (s.c.) with either plasmid DNA encoding the full-length cDNA for mD52 admixed with recombinant murine GM-CSF (PeproTech) or with recombinant mD52 protein admixed with CpG/ODN in alum every 14 days as depicted in Fig. 1a and b. In separate experiments mice were immunized intramuscularly (i.m.) with recombinant mD52 protein admixed with CpG/ODN in alum (Fig. 1a) [25]. Following immunizations, splenocytes were harvested and analyzed for tumor-specific killing as described in the methods section. Effector CTLs were generated by 5–7 days TRAMP-C1 mixed lymphocyte tumor culture (MLTC) and were determined by flow cytometry to contain approximately 30% CD4⁺ T cells and 30% CD8⁺ T cells (not shown). Targets consisted of syngeneic MHC class I matched, autochthonous TRAMP-C1 and TRAMP-C2 tumor cells, and allogeneic MHC class I mis-matched mKSA tumor cells.

Fig. 1.

mD52 DNA and mD52 protein immunization schedules. a Immunization schedule for recombinant mD52 protein with ODN in alum. Groups of mice were immunized with mD52 protein admixed with ODN as an alum precipitate either i.m. or s.c. every 14 days for a total of 4 immunizations. Controls included alum alone and ODN in alum administered either i.m. or s.c. Approximately 14 days after the final immunization mice were challenged s.c. with 5 × 10⁶ autochthonous TRAMP-C1 tumors cells. Tumor formation was monitored as described in the methods section. b Immunization schedule for mD52 plasmid DNA with soluble GM-CSF in PBS. Groups of mice were immunized with mD52 plasmid DNA admixed with GM-CSF in PBS s.c. on day 0, day 14 and day 49 for a total of 3 immunizations. Controls included empty vector DNA in PBS or PBS alone administered s.c. Approximately 14 days after the final immunization mice were challenged s.c. with 5 × 10⁶ autochthonous TRAMP-C1 tumors cells. Tumor formation was monitored as described in the methods section. See methods for concentrations of protein, CpG/ODN, DNA and GM-CSF. N = 10 for each group

Cytotoxic T lymphocytes generated from mice immunized with mD52 DNA + GM-CSF demonstrated MHC class I restricted lysis of mD52 positive tumor cells (Fig. 2). Specific lysis of the syngeneic TRAMP-C1 and TRAMP-C2 tumor cells, for an E:T ratio of 1:100, averaged nearly 90 and 60% for TRAMP-C1 and TRAMP-C2 tumor cells, respectively (Fig. 2a, b). CTLs generated from mice immunized with mD52 protein and CpG/ODN in alum demonstrated MHC class I restricted lysis of mD52 positive tumor cells whether the vaccine was administered s.c. or i.m. (Fig. 2c–f). Specific lysis of the syngeneic TRAMP-C1 and TRAMP-C2 tumor cells, for an E:T ratio of 1:100, averaged from ~50 to ~80% for TRAMP-C1 tumor cells (Fig. 2c–f) and slightly less for TRAMP-C2 tumor cells (Fig. 2c–f). Whereas, the percent specific lysis was <10% for the H-2^d, mD52+ control target mKSA (Fig. 2a–f). Of note, mice immunized with mD52 DNA s.c. without GM-CSF failed to generate a significant cellular immune response (not shown). Taken together, these data suggest that immunization with mD52 DNA generates superior MHC class I restricted CTL responses against mD52+ tumor cells, and the TRAMP-C1 autochthonous tumor cell line is somewhat more readily lysed than the TRAMP-C2 autochthonous tumor cell line. Based on these data TRAMP-C1 tumor cells were used for all subsequent in vivo tumor challenge experiments.

Fig. 2.

Cytotoxic T lymphocyte response following immunization with mD52 DNA or mD52 protein. a, b CTL responses from mice immunized with mD52 DNA + GM-CSF in PBS administered s.c. c, d CTL responses from mice immunized with mD52 protein + CpG (ODN) in alum administered i.m. e, f CTL responses from mice immunized with mD52 protein + CpG (ODN) in alum administered s.c. The autochthonous C57BL/6 TRAMP-C1 and TRAMP-C2 cell lines were derived from spontaneous tumors from the TRAMP model of prostate cancer. The Balb/c derived tumor cell line mKSA served as a control for MHC class I restriction. Shown are CTL responses from 2 representative mice from each vaccine group. Values shown are the mean ± SEM for triplicate determinations

Superior tumor protection following immunization with mD52 DNA

To determine whether it was possible to induce tumor protective immunity following vaccination with mD52, groups of C57BL/6 mice were immunized with either purified recombinant mD52 protein admixed with CpG/ODN or mD52 DNA admixed with soluble GM-CSF as depicted in Fig. 1, and protection from subsequent TRAMP-C1 tumor challenge was evaluated (Fig. 3). Immunization groups included mD52 protein admixed with CpG/ODN as an alum precipitate administered s.c. or i.m., and mD52 DNA admixed with GM-CSF administered s.c. Control immunizations included CpG/ODN in alum administered both s.c. and i.m, alum alone administered both s.c. and i.m, vector only DNA in PBS or PBS alone both administered subcutaneously. The expense of GM-CSF precluded the inclusion of it alone as a vaccine strategy, however numerous published studies have supported the wide spread belief that GM-CSF without antigen is insufficient for inducing protective tumor immunity.

Fig. 3.

Protection from TRAMP-C1 tumor cell challenge following mD52 immunization. Mice were immunized according to the schedule in Fig. 1 and as described in the methods section. Subcutaneous tumor growth was measured over time and compared for each immunization group as % tumor free mice over time (days post tumor challenge). Shown are representative results for repeated experiments comparing mice immunized with mD52 DNA + GM-CSF administered s.c. (DNA + GMCSF sc) to mice immunized with mD52 protein +CpG (ODN) administered i.m. (Protein + ODN im) or administered s.c. (Protein + ODN sc). All control immunizations are shown as a single representative line graph (controls) and include ODN in alum, alum alone, empty vector DNA in PBS and PBS alone. N = 10 for each group

Tumor inoculation and growth was determined as described in the methods section. All mice (100%) immunized with mD52 remained tumor free 50 days following s.c. challenge with a tumorigenic dose of TRAMP-C1 tumor cells, compared to 20% of control immunized mice (Fig. 3). By day 65 post tumor challenge none of the control immunized mice were tumor free. Interestingly, only mice immunized either with mD52 DNA + GM-CSF s.c. or mD52 protein + ODN i.m. remained 60% tumor free by day 65 post tumor challenge. By day 110 post tumor challenge 40% of mice immunized with mD52 DNA + GM-CSF s.c. remained tumor free compared to all other immunizations groups for which all mice had developed s.c. tumors >1 cm × 1 cm (Fig. 3). Of note, 4/10 mice immunized with mD52 DNA + GM-CSF failed to develop palpable tumors at all. These data demonstrate that immunization with mD52 DNA + GM-CSF prior to challenge with TRAMP-C1 tumor cells results in protection from tumor challenge in 40% of mice, compared to mD52 protein + ODN and control immunized animals.

mD52 DNA immunization and tumor protection induces a tumor-specific memory response

Next, we evaluated whether immunization with mD52 DNA admixed with murine GM-CSF was capable of eliciting memory immune responses against secondary challenge with TRAMP-C1 tumor cells. Groups of mice were given three injections with mD52 DNA + GMCSF every two weeks as described in Fig. 4a. A total of 14 days following the final booster immunization animals were challenged with a tumorigenic dose of TRAMP-C1 tumor cells s.c. in the right flank. Primary tumor formation was monitored for 268 days and approximately 80% of the immunized animals remained tumor free (Fig. 4b). Approximately, 150 days following the initial tumor challenge (day 192 following the first immunization), all mice that were tumor free were given a second inoculation of TRAMP-C1 tumor cells in the opposite (left) flank. All mice (100%) that received the secondary tumor challenge in the opposite flank remained tumor free for approximately 6 months following the final immunization which equals 110 days following secondary challenge (Fig. 4c). It is likely the animals would have remained tumor free for much longer however, the experiment was terminated after 10 months for the evaluation of cellular immunity. Of note, the secondary tumor challenge was given 150 days after the primary challenge, and both the primary and secondary challenge were rejected. The fact that the animals rejected two tumor challenges separated by nearly 5 months without receiving an immunization beyond day 28, along with the significant delay from primary tumor challenge before the secondary tumor challenge was given suggests induction of tumor-specific memory immunity as a result of immunization with mD52 DNA + GM-CSF.

Fig. 4.

Immunization with mD52 DNA + GM-CSF induces memory against tumor challenge. Groups of mice were immunized with mD52 DNA admixed with murine GM-CSF administered s.c. as described in the methods sections. a Mice were immunized three times s.c. every 14 days and challenged with 5 × 10⁶ autochthonous TRAMP-C1 tumor cells s.c. (right flank) 14 days after the last immunization (primary tumor challenge). On day 192, approximately 150 days after the primary tumor challenge, mice that rejected the primary tumor challenge in the right flank were given a secondary s.c. tumor challenge with 5 × 10⁶ TRAMP-C1 tumor cells in the left flank. b Primary TRAMP-C1 tumor challenge, right flank. Shown is s.c. TRAMP-C1 tumor volume over time (days post primary challenge) for mice immunized with mD52 DNA + GM-CSF s.c. Day 0 is the 14 days after the final immunization. An arrow marks day 192 when the secondary TRAMP-C1 tumor challenge was administered in the opposite flank. c Secondary TRAMP-C1 tumor challenge, left flank. Shown is s.c. TRAMP-C1 tumor volume over time (days post secondary challenge) for mice immunized with mD52 DNA + GM-CSF s.c. that rejected a s.c. primary TRAMP-C1 tumor challenge in the right flank. Day 0 is the 192 days after the final immunization, and 150 days after protection from the primary TRAMP-C1 tumor challenge. N = 5 for the primary tumor challenge. Shown are representative results for repeated experiments

Immunization with mD52 DNA does not induce autoimmunity

Since mD52 is a “self antigen”, that is it is expressed in some normal tissues as well as in tumors to include TRAMP-C1 tumor cells [25], it is possible that tolerance could prevent the induction of an immune response. Conversely, if tolerance is broken and an immune response is generated against mD52, there is the potential for the induction of harmful autoimmunity. As shown by the generation of antigen-specific CTLs to mD52 expressing tumors (Fig. 2), and by the induction of an immune response against mD52 following immunization that was capable of rejecting mD52-expressing tumor challenge in vivo (Fig. 4), it is clear that indeed tolerance to the self antigen mD52 was broken by active vaccination. Therefore, it was of interest to assess whether autoimmunity was also induced. Following methods previously reported by our lab, autoimmunity induction was assessed in the kidneys of mice that were immunized with mD52 and survived tumor challenge [25]. Individual mice that were immunized with mD52 and survived tumor challenge showed no gross morbidity and appeared healthy throughout the study. Immunohistochemical analysis of the kidneys [25] showed no T cell infiltrates and no evidence of microscopic pathology compared to kidneys from naïve mice serving as normal controls (not shown). No evidence of gross pathology was observed for livers, lungs or spleens. These data support the conclusions that immunization with mD52 DNA + GM-CSF does not result in the induction of autoimmunity.

Induction of T_H1-type cellular immunity following immunization with mD52 DNA

Since mD52 is expressed as an intracellular protein by tumor cells it was important to determine if cellular immune responses were induced in animals that were immunized with mD52 and rejected primary and secondary tumor challenges. As shown in Fig. 2, we established that vaccination with mD52 induced CTL responses that were tumor-specific. It was of interest to determine if vaccine induced cellular immunity reflected a T_H1-type response, thus supporting the generation of CTLs. Splenocytes were harvested from mice that were immunized with mD52 DNA + GM-CSF and survived both a primary and a secondary tumor challenge with TRAMP-C1 tumor cells and cultured with irradiated TRAMP-C1 cells for 5–7 days in the presence of IL-2. Following the MLTC, splenocytes were harvested ficol separated, and incubated in vitro with irradiated tumor cells for 24 h as described in the methods section. Next, supernatants were collected and assayed by ELISA for the presence of IL-4, IL-17, IFN-γ and IL-10 (Fig. 5). Neither IL-4 nor IL-17 was detected in supernatants from cultured T cells derived from vaccinated mice that survived tumor challenge (not shown). These data indicate that neither a T_H2 nor a T_H17 cellular immune response was induced by mD52 DNA vaccination nor did they play a role in tumor protection. This finding was not surprising given that we detected specific-CTL responses as a result of mD52 vaccination (Fig. 2). As expected we were able to detect IFN-γ in supernatants from cultured T cells derived from vaccinated mice that survived tumor challenge (Fig. 5a). The amount of IFN-γ present in 24 h supernatants was at least twice as much when TRAMP-C1 or TRAMP-C2 were the targets compared to control targets. For example, T cells from mouse 3 secreted 500–600 pg of IFN-γ/ml/24 h when TRAMP-C1 or TRAMP-C2 tumor cells were the targets compared to approximately 250 pg/ml/24 h or less when Yac-1 or mKSA were the targets (Fig. 5a). These data suggest that there are mD52-induced, TRAMP tumor cell-specific T cell responses in immunized mice that survived TRAMP-C1 tumor cell challenge. This was confirmed by the ability of H-2^b, class I MHC specific mAb to significantly inhibit production of IFN-γ when included in 24 h cultures of T cells and TRAMP-C1 target cells (Fig. 5a, P < 0.05). Class I MHC-specific antibody inhibition was consistently about 50% of that of T cells and TRAMP-C1 target cells without antibody. Anti-H-2^d mAb failed to inhibit IFN-γ production by T cells cultured with TRAMP-C1 tumor cells (not shown). The background observed when T cells were cultured with control targets could be attributed to non-specific NK-like or LAK-like cells, given that the MLTC was not selected for CD8⁺ T cells and contained relatively large amounts of IL-2. However, taken together these data indicate that mD52 DNA immunization induces an antigen-specific, class I MHC-restricted cellular immune response that is likely responsible for the observed tumor protection in vivo.

Fig. 5.

T lymphocyte cytokine production following mD52 DNA + GM-CSF immunization and tumor protection. Shown are results from standard antigen capture ELISAs measuring the production of cytokines in the supernatants of splenocytes harvested from mice that were immunized with mD52 DNA + GM-CSF and survived both a primary and a secondary tumor challenge with TRAMP-C1 tumor cells and cultured with irradiated TRAMP-C1 cells for 5–7 days in the presence of IL-2. Data are presented as bar graphs depicting pg/ml/24 h of cytokine for T cells cultured alone or with various targets. Targets include: TRAMP-C1 (TC1), TRAMP-C2 (TC2), Yac-1 (control for NK cells) and mKSA tumor cells (MHC class I mis-matched control). To confirm MHC class I restriction T cells were cultured with TRAMP-C1 in the presence of mAb specific for H-2^b class I MHC (T cells + TC1 + Ab). Tumor alone represents each of the tumor cell targets cultured alone. Bars numbered in the legend 1–4 represent each of 4 individual mice. a ELISA for the detection of murine IFN-γ, b ELISA for the detection of murine IL-10. Shown are representative results for repeated experiments. Values shown are the mean ± SEM for duplicate determinations. Data were analyzed using an unpaired t test for independent samples with equal variances and P < 0.05 was determined to be significant (GraphPad Prism 5.0)

Previously, we reported that mD52 expression in tumor cells was associated with increased tumor secretion of TGF-β1 [23]. It has been suggested that TGF-β1 and IL-10 may induce regulatory T cells that could dampen immunity to self antigens [29]. Interestingly, we did not observe 100% tumor protection in any of our vaccine experiments, and knowing that mD52 is a self TAA along with the fact that mD52 expressing tumors secrete TGF-β1, we were interested in determining if IL-10 was involved in the vaccine induced immune response. IL-10 secretion by T cells generated as described above was measured by specific ELISA. Unlike IL-4 and IL-17 we were able to detect IL-10 in supernatants from all immunized and tumor challenge survivor mice (Fig. 5b). However, the amount of INF-γ was as much as ten-fold more than IL-10, suggesting that IFN-γ played a more dominant role in tumor protection and that in survivor mice the vaccine overcame any possible involvement of regulatory T cells. We did not evaluate cytokine production in vaccinated mice that failed to reject primary tumor challenge. However, it is interesting to speculate that the amount of IL-10 may have been as much as ten-fold greater than INF-γ in vaccinated mice that developed tumors suggesting that in some cases regulatory T cells may have dampened the response to the self antigen mD52 following vaccination. The background observed when T cells were cultured with control targets could be attributed to non-specific NK-like or LAK-like cells generated in bulk MLTC in the presence of IL-2. For all ELISA experiments, Yac-1 served as a control for NK-like or LAK-like non-specific cell activity in the bulk MLTC and mKSA served as an mD52 positive, class I MHC mis-matched target. T cells cultured alone without tumor cells and tumor cells cultured alone without T cells failed to secrete detectable levels of any of the cytokines evaluated (Fig. 5a–d).

To examine further the induction of T_H1-type immunity we performed overnight IFN-γ Elispot assays using uncultured T cells from immunized mice that survived tumor challenge. As observed for the ELISA assays, we did not detect IFN-γ production unless T cells were cultured with tumor cell targets (Fig. 6). The number of detectable IFN-γ spots was 2–3 times greater when T cells were cultured with TRAMP-C1 cells than when T cells were cultured with control mKSA tumor cell targets and the difference in spot count was determined to be statistically significant (Fig. 6, P < 0.0001). In order to illustrate the difference in the number and intensity of spots, representative Elispot wells were included in Fig. 6. The majority of the spots in wells with T cells + mKSA target cells are uniform, pinpoint and very low in intensity, compared to spots in wells with T cells + TRAMP-C1 tumor cells. If we chose to use a more stringent spot size and intensity setting most, if not all the spots in the T cells + mKSA wells would not have been counted. In contrast most of the spots would have been counted in the T cells + TRAMP-C1 wells, indicating that vaccination induced a specific anti-TRAMP-C1 T_H1 immune response. However, the overall number of large and intense spots in the T cells + TRAMP-C1 wells is relatively few. This could reflect a low precursor frequency of mD52-specific T cells, which would be expected given that mD52 is a self TAA. Overall the Elispot data support the conclusions of the CTL assays in Fig. 2 and the ELISA assays in Fig. 5, that mD52 DNA vaccination induces a tumor protective, T_H1-type cellular immune response.

Fig. 6.

mD52 DNA + GM-CSF induces tumor reactive T_H1-type cellular immunity. Shown are results from an IFN-γ EliSpot for splenocytes harvested from mice that were immunized with mD52 DNA + GM-CSF and survived both a primary and a secondary tumor challenge with TRAMP-C1 tumor cells and cultured overnight with irradiated target cells. Data are presented as bar graphs depicting IFN-γ spots/1 × 10⁵ T cells cultured alone or with various targets. Targets include: TRAMP-C1 (TC1), and mKSA tumor cells (MHC class I mis-matched control). The mitogen conconavolin A (ConA) served as a positive control for the stimulation of cytokine secretion. Tumor alone represents each of the tumor cell targets cultured alone. Bars numbered in the legend 1–4 represent each of 4 individual mice. Shown are representative results for repeated experiments. Values shown are the mean ± SEM for duplicate determinations. Data were analyzed using an unpaired t test for independent samples with equal variances and P < 0.05 was determined to be significant (GraphPad Prism 5.0)

Discussion

This study was undertaken to determine whether vaccines targeting murine TPD52 (mD52) would induce cellular immune responses capable of rejecting TRAMP-C1 tumor cells that over express mD52 protein in a murine model of prostate cancer. For vaccine strategy comparisons, groups of male C57BL/6 mice were immunized subcutaneously (s.c.) either with plasmid DNA encoding the full-length cDNA for mD52 admixed with recombinant murine GM-CSF or with recombinant mD52 protein admixed with CpG/ODN in alum every 14 days as depicted in Fig. 1a and b. In separate experiments mice were immunized intramuscularly (i.m.) with recombinant mD52 protein admixed with CpG/ODN in alum (Fig. 1a), as previously reported by our group [25]. All three vaccinations approaches outlined in Fig. 1 were capable of inducing CTLs with specificity for mD52⁺ TRAMP-derived tumor cells, however, DNA + GM-CSF administered s.c. induced superior CTLs (Fig. 2). Subsequent experiments demonstrated that DNA + GM-CSF as a vaccine approach induced cellular immunity that resulted in tumor protection in vivo (Fig. 4b), and elicited immunologic memory (Fig. 4c) against a subsequent TRAMP-C1 tumor cell challenge (summarized in Table 1). As shown in Fig. 2, we established that vaccination with mD52 DNA + GMCSF induced CTL responses that were tumor-specific. Neither IL-4 nor IL-17 was detected in supernatants from cultured T cells derived from vaccinated mice that survived tumor challenge (Fig. 5a, b). These data suggest that neither a T_H2 nor a T_H17 cellular immune response was induced by mD52 DNA + GM-CSF vaccination and did not likely contribute to or interfere with tumor protection. We were able to detect IFN-γ in supernatants from cultured T cells derived from vaccinated mice that survived tumor challenge (Fig. 5c), suggesting that mD52-induced specific T_H1-type T cell responses in immunized mice that survived TRAMP-C1 tumor cell challenge. This was confirmed by the ability of H-2^b, class I MHC specific mAb to inhibit production of IFN-γ (Fig. 5c). Previously, we reported that mD52 expression in tumor cells was associated with increased tumor secretion of TGF-β1 [23]. It has been suggested that TGF-β1 and IL-10 may induce regulatory T cells that could dampen immunity to self antigens [29]. Interestingly, we did not observe complete tumor protection in our vaccine experiments. Knowing that mD52 is a self TAA and that mD52 expressing tumors secrete TGF-β,1 we were interested in determining if IL-10 was involved in the vaccine induced immune response. We were able to detect IL-10 in supernatants from immunized and tumor challenge survivor mice (Fig. 5d). However, the amount of INF-γ was as much as 10-fold more than IL-10, suggesting that IFN-γ played a more dominant role in tumor protection and that in survivor mice the vaccine may have overcome any possible involvement of regulatory T cells. We did not evaluate cytokine production in vaccinated mice that failed to reject primary tumor challenge. However, it is interesting to speculate that the amount of IL-10 may have been as much as 10-fold greater than INF-γ in vaccinated mice that developed tumors suggesting that in some cases regulatory T cells may have dampened the response to mD52 following vaccination. This possibility is currently under investigation. To examine further the induction of T_H1-type immunity we performed IFN-γ Elispot assays using uncultured T cells from immunized mice that survived tumor challenge. The number of detectable IFN-γ spots was 2–3 times greater when T cells were cultured with TRAMP-C1 tumor cells than when T cells were cultured with controls (Fig. 6). Overall, the Elispot data support the conclusions of the CTL assays in Fig. 2 and the ELISA assays in Fig. 5, that mD52 DNA vaccination induces a tumor protective, T_H1-type cellular immune response.

Table 1.

Summary of mD52 vaccine induced protection from TRAMP-C1 tumor cell challenge

Vaccine	T cell responsea	Tumor protectionb	Memory response
DNA + GM-CSF s.c.	CTL, IFN-γ	40–80%	100%
Protein + CpG i.m.	CTL	0	ND
Protein + CpG s.c.	CTL	0	ND
Controlsc	none	0	ND

^aIn vitro T cell response following mD52 vaccination. CTL, tumor cell lysis; IFN-γ, cytokine secretion determined by ELISA and EliSpot

^bProtection from subcutaneous TRAMP-C1 tumor challenge greater than 110 days post challenge

^cControls include: PBS alone, alum alone, alum + CpG, vector plasmid DNA

Previously, we immunized mice i.m. with recombinant mD52 protein admixed with CpG/ODN in alum which resulted in the induction of immune responses with specificity for mD52. As a test of tumor immunity, we challenged mice with syngeneic tumor cells that over expressed mD52. Approximately, one half (40–50%) of the mice immunized with recombinant mD52 protein and CpG/ODN in alum rejected s.c. tumor challenge or spontaneous lung metastasis. The observed mD52 vaccine induced tumor protection was associated with the induction of mD52-specific, MHC-restricted CTLs and no detectable evidence of autoimmunity was observed [25]. This previous report demonstrated for the first time that the “self” TAA mD52 is sufficiently immunogenic when administered i.m. as a protein-based vaccine. Moreover, the generation of anti-mD52 CTL resulted in protection from syngeneic tumor cells that naturally over expressed mD52 protein without induction of detectable autoimmunity.

Anti-cancer vaccines targeting self proteins have been applied to treat or prevent multiple cancers in preclinical studies and clinical trials [30, 31]. Tumor protein D52 (TPD52) is a novel and potentially important TAA due to its’ over expression in a number of fatal and common cancers to include prostate [8–10], breast [4, 6, 7] and ovarian [11] carcinomas. The human orthologue of TPD52 has been identified as a candidate breast cancer TAA by using sera from breast cancer patients to screen a library of expressed genes from breast cancers, demonstrating that TPD52 is capable of inducing IgG antibodies which would have required induction of T cell help [24]. This report suggests that TPD52 may be immunogenic in humans and also capable of inducing a cellular immune response, thus warranting study of TPD52 as an anti-cancer vaccine to induce cellular immunity.

It is widely accepted that inflammatory cytokines are potent mediators of adaptive immunity against solid tumors. Many cytokines have been and are being evaluated as reagents to augment vaccines as well as other modes of immunotherapy against cancer. Arguably the most studied and perhaps the most potent cytokine for enhancing immune responses against cancer is GM-CSF. GM-CSF has proven to be the most potent immunostimulatory product when applied as a transgene in whole tumor cell-based vaccine studies in murine tumor models [32, 33] and in human clinical trials for multiple cancers to include melanoma and non-small cell lung cancer [32]. Recently, Disis and colleagues [34] demonstrated that recombinant GM-CSF protein administered as a soluble cytokine in conjunction with a self antigen rat neu-DNA-based vaccine, skewed the response towards cell-mediated immunity. The idea of DNA-based tumor vaccines administered with soluble GM-CSF as an adjuvant was further explored by McNeel and colleagues in a rat model where prostatic acid phosphatase (PAP) served as the self—TAA targeted by vaccination [35]. In this study, DNA vaccination + soluble GM-CSF proved to be more effective at generating specific T_H1-type immunity than recombinant virus expressing PAP as a viral vector vaccine. In a clinical trial of DNA vaccination of prostate cancer patients it was demonstrated that DNA encoding prostate specific antigen (PSA) administered with soluble GM-CSF was safe and induced both cellular and humoral immunity against PSA protein [36].

In the present study, we have shown for the first time that the self TAA mD52 is immunogenic when administered as DNA-based vaccine admixed with the soluble cytokine GM-CSF. Further, the cellular immune response generated is capable of rejecting tumor cells that naturally over express mD52 protein without inducing harmful autoimmunity in a model of murine prostate cancer. These findings suggest that the human orthologue of TPD52 may be a vaccine antigen that could be administered to patients to treat or prevent cancers that over express TPD52. For future studies we will assess roles of immune effector cell subsets in tumor protection in vivo by depletion of CD4⁺ T cells, CD8⁺ T cells, NK cells or CD25⁺ regulatory T cells (Treg) using specific monoclonal antibodies. In addition, heterologous prime boost vaccination approaches will be evaluated by immunizing with DNA followed by protein or protein followed by DNA. Since we have in our possession the cDNA for the human orthologue of TPD52 we will also explore the efficacy of a xenogeneic antigen, DNA-based prime boost strategy.

Abbreviations

TPD52

Tumor protein D52

mD52

Murine TPD52

hD52

Human TPD52

TRAMP

Transgenic adenocarcinoma of the mouse prostate

TAA

Tumor associated antigen