Activation of Tumor-Specific CD8⁺ T Cells after Intratumoral Ad5-TRAIL/CpG Oligodeoxynucleotide Combination Therapy

Peptides derived from endogenously expressed proteins are presented by antigen-presenting cells (APC) in the context of MHC class I (MHC I) to CD8⁺ T cells, whereas peptides obtained from exogenously derived proteins are normally loaded onto MHC class II (MHC II) for presentation to CD4⁺ T cells. Exogenous antigens can be also loaded onto MHC I for “cross-presentation” to CD8⁺ T cells (1–3), which is believed to occur after the phagocytosis of apoptotic cells, dendritic cell sampling of live cells, or the shuttling of proteins to dendritic cells by heat shock proteins (4). Through cross-presentation, two outcomes can occur: (a) cross-priming and CD8⁺ T-cell activation, induction of cellular proliferation, and acquisition of effector function; or (b) there is no activation of CD8⁺ T cells with cross-tolerance, and the T cells either become anergic and/or are deleted (4).

In tumor-draining lymph nodes, both cross-priming and cross-tolerance have been detected in different tumor models. In experiments where cross-priming was observed, tumor antigen–specific T-cell proliferation was detected, but the numbers of T cells proliferating were very low, and the overall effect of CD8⁺ T-cell activation did not inhibit tumor outgrowth (5). Furthermore, when an apoptosis-inducing agent, gemcitabine, was used as a therapy in these studies, an increase in T-cell proliferation was detected and tumor clearance was observed. This indicates that cross-priming was occurring in the presence of apoptosis, and that the cross-priming could be augmented by increasing the amount of tumor antigen available for presentation. In contrast, studies with other tumor models have reported no T-cell proliferation, and the T cells become anergic and/or deleted after time (6). This study used an adoptive transfer model of hemagglutinin-specific CD8⁺ T cells (clone 4 T cells; ref. 7), and the number of cells transferred altered the outcome of the experiments. In these experiments, the greater the tumor burden, the faster the CD8⁺ T cells became tolerized and deletion occurred. The authors suggested that when greater T-cell numbers were transferred, they underwent autoactivation; whereas when lower T-cell numbers were transferred, the T cells become tolerized due to the lack of an inflammatory environment. This titration of antigen-specific T cells could be a major underlying difference in the experiments performed by various groups studying tumor antigen cross-presentation and the induction of activation or tolerance. These conflicting results also make it important to determine if apoptosis-based immunotherapy will benefit from tumor antigen cross-priming.

Tumor necrosis factor–related apoptosis inducing ligand (TRAIL) potently induces tumor cell apoptosis, but has no cytotoxic activity against normal cells and tissues (8, 9). Previously, we described the development of a recombinant adenoviral vector encoding the TRAIL cDNA (Ad5-TRAIL) to allow for prolonged TRAIL production (10, 11). However, it remains unknown whether antigens derived from Ad5-TRAIL–generated apoptotic tumor cells can be cross-presented to CD8⁺ T cells, facilitating cross-priming. The experiments described in this report examine the hypothesis that increasing the amount of tumor antigen available for dendritic cells to phagocytize via Ad5-TRAIL–induced apoptosis will allow for increased cross-priming and CD8⁺ T-cell activation, leading to increased tumor-specific CTL activity and tumor clearance for prolonged survival.

Mice and reagents. BALB/c (H-2^d) mice were purchased from The Jackson Laboratory. Clone 4 and InsHA (12) transgenic mice were obtained from Dr. Linda Sherman (The Scripps Research Institute, La Jolla, CA). β2m^−/− mice were provided by Dr. Kevin L. Legge (University of Iowa, Iowa City, IA). CpG oligodeoxynucleotide (CpG ODN) 1826 was purchased from Coley Pharmaceutical. Anti-CD25 (PC-61) and anti-GITR (DTA-1) were purchased from eBioscience.

Cell lines. The transplantable murine renal adenocarcinoma cell line, Renca (13), was obtained from Dr. Robert Wiltrout (National Cancer Institute, Frederick, MD) and maintained in RPMI 1640 supplemented with 10% fetal bovine serum, penicillin, streptomycin, sodium pyruvate, nonessential amino acids, and HEPES (hereafter called complete RPMI). Renca stably transfected to express hemagglutinin (RencaHA; ref. 14) was obtained from Dr. Hyam Levitsky (Johns Hopkins University, Baltimore, MD) and maintained in complete RPMI with 400 μg/mL G418. The mouse prostate tumor cell line, RM-11 (15), was obtained from Dr. Timothy Ratliff (University of Iowa) and maintained in complete RPMI.

Production of recombinant adenoviral vectors. Replication-deficient adenovirus encoding murine tnfsf10 (Ad5-mTRAIL; ref. 10), human Flt-3L (Ad5-Flt-3L), or β-galactosidase (Ad5-βgal) expressed from the cytomegalovirus promoter was made at the University of Iowa Gene Transfer Vector Core using the RAPAd.I system (16). Recombinant adenoviruses were screened for replication competent virus by A549 plaque assay, and virus titer was determined by plaque assay on 293 cells.

Dendritic cell phagocytosis. Immature CD11c⁺ dendritic cells were mobilized by injecting Ad5-Flt-3L [10⁹ plaque-forming units (pfu) i.v.] into naïve BALB/c mice and isolated after 7 to 10 days from spleens using a Miltenyi kit (Miltenyi Biotec, Inc.). Dendritic cells were then labeled with PKH67 membrane dye (Sigma). Renca cells were labeled with PKH26 membrane dye (Sigma) before induction of apoptosis by either UV irradiation for 5 min or Ad5-mTRAIL (300 pfu/cell) infection. Apoptosis was allowed to occur for 16 h before adding dendritic cells, and dendritic cell phagocytosis of the apoptotic debris lasted for 16 h. Phagocytosis was assessed on a FACScan (Becton Dickson) at the indicated time points. To confirm phagocytosis, irradiated Renca cells and dendritic cells were also incubated at 4°C, which inhibits phagocytosis (17). Dendritic cell maturation was also assessed by staining with an anti-CD80 phycoerythrin-Cy5 antibody (eBioscience) after 24 h.

Confocal microscopy of dendritic cell phagocytosis. Samples for phagocytosis were set up as described above. Cells were FACS (fluorescence-activated cell sorting)–sorted to obtain the double positive “phagocytosing populations,” as well as the PKH67 only (“nonphagocytosing”) population. Cells were allowed to adhere to BD BioCoat Collagen I Chamber Slides (BD Biosciences) for 4 h at 37°C, fixed in 2% paraformaldehyde, and analyzed.

In vitro cross-presentation. Ad5-Flt-3L–mobilized CD11c⁺ dendritic cells were isolated from BALB/c or β2m^−/− mice, and apoptotic cultures were set up as described above. Phagocytosis lasted for 16 h, after which the cultures were clarified with Fico/Lite (Atlanta Biologicals) to separate unphagocytosed apoptotic/dead cells from dendritic cells. Dendritic cells were then cultured with carboxyfluorescein succinimidyl ester (CFSE)–labeled (10 min at 37°C with 1 μmol/L CFSE) clone 4 T cells. The dendritic cell and clone 4 T cells were cultured in a 1:1 ratio in 96-well round-bottomed plates for 72 to 96 h. HA_518-526 peptide (IYSTVASSL; 1 μmol/L; New England Peptide, Inc.)–pulsed dendritic cells or dendritic cells infected with UV-inactivated influenza virus strain A/PR/8/34 H1N1 (Dr. Kevin Legge, University of Iowa) mixed with clone 4 T cells served as positive controls. After incubation, cells were stained with anti-CD8α and anti-CD90.2 (eBioscience), and CFSE dilution was monitored by flow cytometry.

In vivo cross-presentation. InsHA mice were injected s.c. with 2 × 10⁵ RencaHA in their hind flank. Six days after implantation, 3 × 10⁶ CFSE-labeled clone 4 T cells were injected i.v. On day 7, Ad5-βgal (10⁹ pfu), Ad5-mTRAIL (10⁹ pfu), and/or CpG ODN (100 μg) were injected intratumorally (i.t.). Five days later, inguinal lymph nodes were harvested and stained with anti-CD8α and anti-CD90.2 monoclonal antibody (mAb) to assess CFSE dilution by flow cytometry.

In vivo cytolytic assay. InsHA mice were injected s.c. with 2 × 10⁵ RencaHA cells in their hind flank. Six days after implantation, 3 × 10⁶ clone 4 T cells were transferred i.v. On day 7, Ad5-βgal (10⁹ pfu), Ad5-mTRAIL (10⁹ pfu), and/or CpG ODN (100 μg) were injected i.t. Five days later, 3 × 10⁶ syngeneic spleen cells were labeled with 5 μmol/L CFSE^high and pulsed with 1 μmol/L HA_518-526 peptide for 1 h at 37°C or 3 × 10⁶ unplused cells were labeled with 0.5 μmol/L CFSE^low. CFSE^high and CFSE^low cells were mixed in a 1:1 ratio, and 6 × 10⁶ cells in total were injected i.v. into tumor-bearing InsHA mice. Eighteen hours after transfer, spleens were harvested and the CFSE^high to CFSE^low cell ratio was analyzed by flow cytometry.

Tumor therapy experiments. BALB/c mice were implanted s.c in the hind flank with 2 × 10⁵ Renca. Ad5-βgal (10⁹ pfu), Ad5-mTRAIL (10⁹ pfu), and/or CpG ODN (100 μg) were injected i.t. on day 7. In some cases, mice were also depleted of CD8⁺ or CD4⁺ T cells with 100 μg anti-CD8 or anti-CD4 (obtained from Dr. Timothy Ratliff, University of Iowa) 3 consecutive days before tumor implantation and maintenance injections biweekly throughout the experiment, CD25⁺ cells with anti-CD25 (500 μg given 1 day before tumor implantation), or GITR⁺ cells with anti-GITR (1 mg given 1 day before and 4 and 9 days after tumor implantation). CD8⁺, CD4⁺, and CD25⁺ cell depletion was verified by flow cytometry in sentinel mice. Tumor outgrowth and animal survival were monitored over time. To determine the presence of immunologic memory after Ad5-mTRAIL/CpG ODN therapy, mice that had rejected the primary Renca challenge were rechallenged with Renca tumor cells (2 × 10⁵) implanted s.c. on the left flank and RM-11 tumor cells (2 × 10⁵) implanted s.c. on the right flank. Tumor outgrowth and animal survival were monitored over time.

Assessment of tumor-infiltrating cells. BALB/c mice were implanted s.c. in the hind flank with 2 × 10⁵ Renca. After 7 days, Ad5-βgal (10⁹ pfu), Ad5-mTRAIL (10⁹ pfu), and/or CpG ODN (100 μg) were injected i.t. After another 7 days, the tumors were harvested and physically dissociated to achieve a single-cell suspension, which was then labeled with FITC-conjugated anti-CD4 and phycoerythrin-conjugated anti-CD8 mAb (eBioscience) and analyzed by flow cytometry.

Statistics. Statistical analysis was performed using ANOVA or Student's unpaired t test to assess differences between groups. Survival data was analyzed using the Kaplan-Meier survival analysis and log-rank statistics. All reported P values are two sided, and statistical significance was determined as P < 0.05.

In vitro phagocytosis of apoptotic Renca cells. Antigen cross-presentation is necessary to generate a CD8⁺ T-cell response to exogenously derived antigens, such as those from apoptotic cells. We hypothesize that Renca cells undergoing Ad5-mTRAIL–induced apoptosis (18) are ingested by dendritic cell and the tumor antigens presented to CD8⁺ T cells. To test this hypothesis, our experiments will use both parental Renca tumor cells and Renca cells stably expressing hemagglutinin on the surface (RencaHA), making it important to first compare the responsiveness of both cell lines to Ad5-mTRAIL. Ad5-mTRAIL induced death similarly in both Renca and RencaHA, whereas Ad5-βgal did not induce cell death, indicating cell death was due to TRAIL expression and not simply the result of adenoviral infection (Fig. 1A). To test our hypothesis that apoptotic Renca cell antigens could be cross-presented, we determined the capacity of dendritic cell to phagocytose apoptotic tumor cells in the presence of adenovirus. Thus, Renca tumor cells were labeled with PKH26, induced to undergo apoptosis with Ad5-mTRAIL or UV irradiation (which yielded comparable levels of apoptotic cells; data not shown), and then mixed with splenic CD11c⁺ dendritic cell labeled with PKH67. Phagocytosis was assessed over a time course of 24 h, and the results show that the presence of an adenovirus in the culture did not alter the ability of the dendritic cells to take up apoptotic tumor cells killed by Ad5-mTRAIL compared with radiation-induced (adenovirus-free) apoptotic cultures (Fig. 1B). As predicted, Ad5-βgal–infected cells, which do not undergo apoptosis, were not taken up by the dendritic cells. To confirm that the method of uptake of the Renca apoptotic bodies was phagocytosis, a coculture of irradiated Renca cells and dendritic cells were incubated at 4°C (17). This treatment prevented the uptake of apoptotic fragments. To confirm that the dendritic cells were truly engulfing the apoptotic Renca cells, dendritic cells were sorted using the same gating scheme and then analyzed by confocal microscopy. In the representative images shown, the green phagocytic cells contained red tumor cell membranes within them (Fig. 1C and D), confirming that the double-positive cells measured in Fig. 1B were dendritic cells that had specifically phagocytosed apoptotic cells.

Based on the results in Fig. 1B where dendritic cells in the presence of adenovirus were equally capable of phagocytosing tumor cells, we next examined whether these dendritic cells were matured by the presence of adenovirus. The dendritic cells were stained for surface CD80 expression, which increased in the treatment groups where virus is present compared with either the irradiated cells/dendritic cells cultured at 4°C or irradiated cells in the absence of virus (Fig. 1E). Thus, the presence of an adenovirus in these cultures does mature the dendritic cells but does not alter their ability to phagocytose apoptotic tumor cells in vitro.

Tumor antigen cross-presentation in vitro after Ad5-mTRAIL–induced apoptosis. Delivery of trail using a recombinant adenoviral vector results in sustained local TRAIL expression and increases the local TRAIL concentration at the tumor site (11). Ad5-mTRAIL is also a potentially appealing agent for initiating cross-presentation because, in addition to inducing tumor cell apoptosis, the adenovirus stimulates dendritic cell maturation, as shown in Fig. 1. Thus, it was important to assess how effective Ad5-mTRAIL would be at stimulating dendritic cells to cross-present tumor antigen to CD8⁺ T cells. To examine this in vitro, RencaHA and clone 4 T cells were used. RencaHA cells were infected with Ad5-βgal, Ad5-mTRAIL, or UV irradiated to induce apoptosis, and 16 h later CD11c⁺ dendritic cells were added to permit phagocytosis. The dendritic cell/RencaHA cultures were then incubated with CFSE-labeled clone 4 T cells, and the percentage of proliferating cells in groups that were induced to undergo apoptosis with Ad5-mTRAIL or UV irradiation were equivalent (Fig. 2A). As a positive control, we cultured clone 4 T cells with dendritic cells pulsed with hemagglutinin peptide or UV-inactivated influenza, resulting in pronounced T-cell proliferation. Conversely, clone 4 T cells cultured with dendritic cells alone, UV-irradiated Renca, or RencaHA infected with Ad5-βgal showed minimal proliferation. We confirmed that RencaHA antigens were cross-presented to clone 4 T cells by feeding apoptotic RencaHA cells (induced by either Ad5-mTRAIL or UV irradiation) to CD11c⁺ dendritic cells isolated from β2m^−/− mice, which lack MHC I expression (19). In this setting, the clone 4 T cells failed to proliferate (Fig. 2A). Furthermore, we tested the ability of viable Renca or RencaHA cells to directly stimulate clone 4 T-cell proliferation. Just as we saw with the β2m^−/− dendritic cells, there was no clone 4 T-cell proliferation when cultured with Renca or RencaHA cells alone (Fig. 2B). These results collectively show that tumor cell antigen derived from apoptotic Renca cells are cross-presented to CD8⁺ T cells, and the presence of an adenovirus in this system does not alter the capacity of the dendritic cells to cross-present tumor antigen.

Ad5-mTRAIL therapy must be combined with CpG ODN to show a therapeutic effect and induce antigen-specific T cell proliferation and CTL activity. The previous experiments indicate that Ad5-mTRAIL can generate apoptotic tumor cells that can be phagocytosed and cross-presented in vitro to clone 4 T cells. Thus, we expected Ad5-mTRAIL administered i.t. would increase animal survival. To test this, BALB/c mice were injected s.c. with Renca in the hind flank, Ad5-βgal or Ad5-mTRAIL was injected i.t. 7 days later, and tumor size (Fig. 3A) and animal survival were monitored. Animal survival was prolonged with Ad5-mTRAIL compared with untreated or Ad5-βgal–treated mice, but all the mice eventually succumbed to their tumors (Fig. 3B). Although dendritic cells were presumably cross-presenting tumor antigens from apoptotic Renca cells killed by Ad5-mTRAIL, additional priming signals may be needed to induce a more robust CD8⁺ T-cell response. CpG ODNs are powerful immune adjuvants that activate multiple immune cell populations, including dendritic cells (20). Thus, we examined the benefits of including CpG ODN with Ad5-mTRAIL. In vivo administration of CpG ODN matures dendritic cells in draining lymph nodes when administered into the footpad of BALB/c mice (Fig. 3C), as determined by the up-regulation of CD80, CD86, CD54, MHCI, and MHCII expression levels versus untreated groups. Consequently, we determined if the combination of Ad5-mTRAIL and CpG ODN would prolong animal survival compared with single-agent therapy. Surprisingly, CpG ODN treatment alone significantly enhanced animal survival over untreated animals but still all of the animals succumbed to the tumor. The greatest increase in survival, however, was observed when Ad5-mTRAIL and CpG ODN were injected in combination i.t. (Fig. 3A). In fact, the combination therapy even resulted in some mice completely eliminating the tumor. Concurrently, we assessed the recruitment of T cells into the tumor 7 days after injection of Ad5-βgal, Ad5-mTRAIL, and/or CpG ODN. There was a dramatic increase in the percentage of both CD4⁺ and CD8⁺ cells within the tumor only after injecting both Ad5-mTRAIL and CpG ODN (Fig. 3D).

The combination of Ad5-mTRAIL and CpG ODN led to a significant survival advantage in Renca tumor-bearing animals; however, the experiments in Fig. 3 were not designed to directly assess cross-presentation in vivo. To determine the effectiveness of Ad5-mTRAIL/CpG ODN therapy at stimulating a CD8⁺ T-cell response, InsHA mice implanted with RencaHA tumors were treated with Ad5-βgal, Ad5-mTRAIL, and/or CpG ODN. One day before treatment, CFSE-labeled clone 4 T cells were adoptively transferred into the tumor-bearing InsHA mice. Five days after treatment, inguinal lymph nodes were harvested, and the proliferation of the adoptively transferred clone 4 T cells was assessed by CSFE dilution. Untreated InsHA animals had a low but detectable level of clone 4 T-cell proliferation (Fig. 4A). I.t. injection of Ad5-mTRAIL or CpG ODN alone increased clone 4 T-cell proliferation, but the greatest level of clone 4 T-cell proliferation occurred when the RencaHA-bearing InsHA mice received both Ad5-mTRAIL and CpG ODN. These data are consistent with the survival results in Fig. 3A, indicating that although there was some antigen-specific CD8⁺ T-cell activation and therapeutic effect when either Ad5-mTRAIL or CpG ODN were used alone, the greatest amount of T-cell activation and therapeutic effect was observed when Ad5-mTRAIL was combined with CpG ODN.

Enhancing the ability of dendritic cells to present antigen to, and induce proliferation of, T cells is important in tumor immunology. However, T-cell acquisition of tumor-specific cytolytic ability is also essential. To further investigate the induction of a tumor antigen–specific T-cell response after Ad5-mTRAIL/CpG ODN therapy, an in vivo CTL assay was performed to assess the cytolytic potential of the activated T cells. InsHA mice implanted with RencaHA tumors and treated with Ad5-βgal, Ad5-mTRAIL, and/or CpG ODN i.t., and the in vivo CTL assay was performed 5 days later. Just as observed with the augmentation in T-cell proliferation, CTL activity doubled in mice that were given Ad5-mTRAIL and CpG ODN compared with mice that were untreated or given Ad5-mTRAIL alone, CpG ODN alone, or Ad5-βgal and CpG ODN (Fig. 4B). Collectively, the data in Fig. 4 show that neither therapeutic agent alone (Ad5-mTRAIL or CpG ODN) is unable to induce profound antigen-specific CD8⁺ T-cell proliferation nor CTL activity, but the combination of increased apoptosis and stimulatory signals is needed to generate a robust T-cell response.

CD8⁺ T cells are necessary for prolonged survival of mice treated with Ad5-mTRAIL/CpG ODN. Our in vivo data show the activation of antigen-specific T cells and induction of CTL activity, suggesting the necessity for CD8⁺ T cells in the suppression of tumor outgrowth and prolongation of survival. To confirm that Ad5-mTRAIL/CpG ODN therapy was indeed augmenting the CD8⁺ T-cell response that prolonged survival time, CD8⁺ T cells were depleted from BALB/c mice before and throughout the duration of combination tumor therapy. Mice depleted of CD8⁺ T cells and given Ad5-mTRAIL/CpG ODN i.t. had no survival advantage over untreated mice bearing Renca tumors (Fig. 5A), providing further evidence suggesting that Ad5-mTRAIL/CpG ODN therapy activates antigen-specific CD8⁺ T cells that are critical to the generation of a successful antitumor immune response.

CD4⁺ cells suppress the therapeutic potential of Ad5-TRAIL/CpG ODN therapy. It was surprising that there was only a small percentage (∼40%) of mice treated with Ad5-mTRAIL/CpG ODN that completely eliminated their tumors, despite the significant increase in CTL activity. We considered the possibility that CD4⁺CD25⁺ regulatory T cells, which have been implicated in numerous tumor models, may be inhibiting the full therapeutic potential of Ad5-mTRAIL/CpG ODN by suppressing the CD8⁺ T-cell response. To investigate this possibility, antibody-mediated depletion of CD4⁺ or CD25⁺ cells from BALB/c mice was performed before Renca tumor implantation, and then the mice were treated with Ad5-mTRAIL/CpG ODN. Interestingly, depletion of CD4⁺ or CD25⁺ cells significantly prolonged survival compared with those mice with a full repertoire of cells (Fig. 5A). Although there was significantly enhanced survival in mice lacking CD4⁺ or CD25⁺ cells, the survival results did not indicate what effects the CD4⁺ or CD25⁺ cells were having on the CD8⁺ T cells. Thus, we assessed in vivo CTL activity in RencaHA-bearing InsHA mice that had been pretreated with the anti-CD4, CD25, or GITR antibodies 5 days before receiving Ad5-mTRAIL/CpG ODN. Depletion of CD4⁺ or CD25⁺ cells resulted in a statistically significant increase in CTL activity compared with undepleted mice that received Ad5-mTRAIL/CpG ODN (Fig. 5B). Addition of the anti-GITR antibody resulted in no alteration in CTL activity. We went on to examine clone 4 T-cell proliferation in RencaHA-bearing intact or anti-CD4, CD25, or GITR-treated InsHA mice that received Ad5-mTRAIL/CpG ODN. Despite showing increased CTL activity and survival, mice depleted of CD4⁺ cells did not show any increase in clone 4 T-cell proliferation over undepleted mice (Fig. 5C). Similarly, there was not an increase in clone 4 T-cell proliferation in the mice that received anti-CD25. However, there was increased clone 4 T-cell proliferation in the mice that received anti-GITR mAb. GITR is expressed at high levels on both CD4⁺ and CD8⁺ T cells after activation, and GITR ligation can serve as a costimulatory signal on CD8⁺ T cells to enhance proliferation and effector function (21, 22). These results suggest that CD4⁺ cells regulate the CTL activity that is induced upon Ad5-TRAIL/CpG ODN treatment. It is possible that the CD4⁺CD25⁺ regulatory T-cell subset is responsible for the inhibition of the full potential of an effective T-cell response to clear Renca tumors, as indicated by an increase in CTL activity in both groups that were depleted of either CD4⁺ or CD25⁺ cells.

Mice that reject primary Renca tumors after Ad5-mTRAIL/CpG ODN therapy also reject a secondary Renca tumor challenge. The results showing the necessity for CD8⁺ T cells to reject Renca tumors after Ad5-mTRAIL/CpG ODN therapy prompted us to test for immunologic memory. To do this, mice depleted of CD4⁺ cells were implanted with Renca tumor cells and treated with Ad5-mTRAIL/CpG ODN, as in Fig. 5. Mice that proved to be tumor-free after 100 days were then rechallenged with the same dose of Renca tumor cells on the left flank or MHC-matched RM-11 prostate tumor cells on the right flank. These mice received no additional therapy, and tumor outgrowth was monitored. Additionally, naïve mice were similarly implanted with Renca and RM-11 cells. Whereas the Renca cells grew progressively in the naïve mice, none of the mice that had previously rejected established Renca tumors after Ad5-mTRAIL/CpG ODN treatment had palpable tumors (Fig. 6A). In contrast, RM-11 grew at comparable rates in both groups (Fig. 6B), suggesting that there is tumor-specific immunologic memory established in mice that rejected the primary tumor after Ad5-mTRAIL/CpG ODN therapy.

This report outlines a novel immunotherapeutic approach to combining Ad5-mTRAIL with CpG ODN in an experimental model of renal cell carcinoma. The benefits of combining one agent that selectively kills tumor cells and has prolonged in vivo expression with another that has profound immunostimulatory capacity is extremely important, as the combination of Ad5-mTRAIL and CpG ODN facilitates the switching of the tumor microenvironment from one composed of limited antigen presentation to one with priming of tumor-specific CD8⁺ T cells having potent effector function.

There is controversy over the ability of antigens derived from apoptotic cells that are cross-presented to CD8⁺ T cells to induce T-cell immunity versus tolerance. It makes sense that apoptotic death does not allow for antigens to be cross-presented to prime CD8⁺ T cells, because this type of death is expected to quietly clear cellular debris occurring through normal cellular turnover without promoting autoimmunity (23–25). Our results are similar to those published by den Brok and colleagues (26), who combined cryoablation with CpG ODN to eradicate experimental melanoma tumors. In that study, neither cryoablation nor CpG ODN therapy alone was sufficient to induce antitumor immunity, much like our results when Ad5-mTRAIL or CpG ODN were used individually. It is important to note, however, that the mechanism and specificity of cell death in our study and the den Brok report are quite different. TRAIL only induces apoptosis in tumor cells, but cryoablation induces a necrotic-like death in both tumor and normal cells. It still remains to be conclusively determined whether apoptotic death is the means by which immunologic tolerance can be induced, but the use of a tumoricidal agent (Ad5-mTRAIL) in combination with an immunostimulatory molecule (CpG ODN) in our model system clearly pushes the immune system toward a priming environment. The addition of CpG ODN helps to mature dendritic cells and up-regulate costimulatory molecule expression, as well as stimulate cytokine production, at the time apoptosis is being induced so that T-cell activation can occur with the intention of generating a cell-mediated antitumor response (27). Induction of dendritic cell maturation before antigen uptake results in better antigen cross-presentation than when the dendritic cells are matured after they take up antigen (28), making this approach a more desirable method to further stimulate cross-presentation. It is reasonable to hypothesize that CpG ODN–mediated dendritic cell stimulation/maturation occurs before the generation of antigen from apoptotic tumor cells when Ad5-mTRAIL and CpG ODN are administered simultaneously, because Ad5-mTRAIL must first infect a cell and then have the transgene transcribed and translated before TRAIL-mediated tumor cell apoptosis can occur.

Some tumors continuously shed antigen as the result of a low level of cell death, and this either leaves the immune system tolerized or allows for a low level of T-cell proliferation, although it is not enough to inhibit tumor outgrowth (5). Our results showed that mice bearing Renca tumors do have a low level of proliferation and CTL activity, although this does not inhibit tumor outgrowth. The absence of an increase in CTL activity upon addition of Ad5-mTRAIL alone as a therapy suggests that increasing the amount of tumor antigen available for cross-presentation is not enough to stimulate a significant response. Further, CpG ODN alone does not remarkably increase CTL activity, indicating the lack of natural cross-priming is not a result of a lack of priming signals only, but likely a combination of insufficient antigen levels and potent dendritic cell stimulatory signals. This confirms the theory that although TRAIL is delivered as a recombinant adenovirus, the adenoviral signals for stimulating a priming environment are not sufficient and the addition of CpG ODN is necessary to tip the dynamics of the environment to one that supports T-cell activation.

The extent of antigen-specific T-cell proliferation and CTL activity induced when combining Ad5-mTRAIL and CpG ODN suggests this therapy is effective at stimulating endogenous T cells to respond to tumor antigen(s) and gain effector function. The in vitro killing of Renca tumor cells by immune effector cells (natural killer cells and CD8⁺ CTL) activated by immunotherapeutic protocols can occur by both granule and Fas-mediated cytotoxicity (29). However, in vivo effector cell mechanisms of killing Renca are perforin independent (30). Although our studies were not designed to investigate the molecules used by the effector cells activated by the Ad5-mTRAIL/CpG ODN therapy, this is an area of future interest. We were also surprised by the observation that CD4⁺ cell depletion before Ad5-mTRAIL/CpG ODN treatment augmented the CD8⁺ T-cell response and significantly enhanced animal survival. CD40-CD154 interactions between APC and CD4⁺ T cells are important in facilitating CD8⁺ CTL cross-priming events (31, 32). CD40 ligation on the APC increases CD80/CD86 expression (33), which is essential for proper T-cell activation. In this regard, Prilliman et al. (34) showed that CD8⁺ CTL cross-priming could still occur in the absence of CD4⁺ T-cell help, provided there was sufficient CD28 signaling with an agonistic mAb. The CpG ODN used in our system could substitute for the required CD4⁺ T-cell help. The report by Cho et al. (35) supports this concept, where they were able to induce CD8⁺ CTL activity in CD4 or MHC class II–deficient animals.

Despite the activation of tumor-specific CTL, the combination of Ad5-mTRAIL and CpG ODN did not always result in complete tumor regression. Renca is a fairly aggressive tumor, and it is possible that the slight differences in the tumor cells from experiment to experiment, as well as the variability in mice, result in the single i.t. treatment being insufficient to adequately deal with and eliminate all of the cells in the rapidly growing tumor. Additional injections of Ad5-mTRAIL would induce more apoptosis and further debulk the rapidly growing tumor, whereas additional CpG ODN could help maintain a strong CD8⁺ T cell–mediated antitumor immune response. Our data also implicates a suppressive effect from CD4⁺CD25⁺ cells that restricts the full potential of the therapeutic combination of Ad5-mTRAIL and CpG ODN. Clinically, patients with renal cell carcinoma have elevated numbers of CD4⁺CD25⁺ regulatory T cells, and decreasing their numbers with high-dose interleukin 2 (IL-2) or a recombinant IL-2 diphtheria toxin conjugate enhances antitumor immunity (36, 37). Regulatory T cells can elicit suppressive activity through contact-dependent [CTLA-4, membrane-bound transforming growth factor-β (TGF-β)] or contact-independent (soluble TGF-β, IL-10) mechanisms (38, 39). Our results show that the depletion of CD4⁺ or CD25⁺ cells resulted in increased CTL activity and prolonged animal survival after treatment with Ad5-mTRAIL and CpG ODN. Moreover, the animals that rejected the primary tumor after Ad5-mTRAIL/CpG ODN therapy showed tumor-specific immunologic memory that was sufficient to protect against tumor rechallenge. The depletion of CD25-expressing cells provides additional evidence to suggest that the cells involved in reduction of CTL activity in this experimental system are indeed CD4⁺CD25⁺ regulatory T cells. Interestingly, tumor antigen–specific (clone 4) T-cell proliferation was not increased in RencaHA-bearing mice treated with Ad5-mTRAIL/CpG ODN that also received anti-CD4 or anti-CD25 mAb, suggesting that T-cell effector function, in this case in vivo CTL activity, is not related to the ability of the cells to proliferate. One or many of the immunosuppressive mechanisms used by regulatory T cells may, therefore, be involved in this system to inhibit the full potential of Ad5-mTRAIL/CpG ODN therapy. Further characterization of the regulatory cells is under way, and it is presently unknown whether the regulation mediated by these cells is through a contact-dependent or contact-independent mechanism.

Nearly all DNA viruses have evolved to evade host immune defense mechanisms. One way this can be accomplished is by reducing the genomic content of CpG motifs, yet adenoviruses do not seem to have made such genetic alterations (40). Despite having the expected level of genomic CpG dinucleotides, DNA from adenovirus type 5 has weak immunostimulatory capabilities (41). In fact, type 5 adenoviral DNA contains sequences that actively suppress the stimulatory effects of other prokaryotic DNA sequences. Our adenoviral vectors are based on the type 5 backbone (42), raising the possibility that the immunostimulatory effects of the CpG ODN 1826 used in our experiments are being squelched by the inhibitory sequences present in the adenoviral genome. The intended immune response is then further regulated by the presence of CD4⁺ regulatory cells. Although we presently have no experimental evidence to confirm or deny this hypothesis, this could also explain the observed benefit of depleting CD4⁺ or CD25⁺ cells before Ad5-mTRAIL/CpG ODN administration.

Recombinant adenoviral vectors transduce a wide range of dividing and nondividing cell types, making this gene delivery system valuable as a tool for studying diseases, for vaccine therapy, and for potential clinical use (43). Wild-type adenovirus infections are extremely common in the general population, giving adenoviruses a well-documented safety record (44). Due to the high percentage of the population that is immune to adenoviruses, one could hypothesize two additional consequences to using an adenoviral vector in the clinical treatment of cancer. First, the presence of neutralizing mAb could render the injected vector inactive before it can infect its intended target. This would potentially affect adenoviral vectors delivered systemically greater than those delivered directly into the tumor, as in our setting. Interestingly, administration of adenoviral vectors mixed with a delivery vehicle can overcome preexisting adenoviral immunity (45), making it possible to prevail over this potential obstacle. Second, the induced primary CD8⁺ T cell–mediated antitumor response may result (in part) from the presentation of adenoviral antigens via MHC I on the tumor cells and by cross-presentation on dendritic cells, which would then be followed by the activation of adenovirus-specific CD8⁺ T cells and the killing of adenovirally infected tumor cells. Although a possibility in some experimental and clinical settings, this is not a plausible mechanism in our model because there were no adenoviral antigens present at the time of the second challenge with Renca cells in the animals that rejected the primary tumor (see Fig. 6).

In conclusion, the results outlined in this article show that Ad5-mTRAIL can induce tumor cell apoptosis, and that apoptotic tumor cell antigens can be cross-presented to generate a CD8⁺ T-cell response. For this T-cell response to be significant, however, strong priming signals need to be present, which can be provided by the addition of CpG ODN. Although the results and conclusions in this report are based on an experimental renal cell carcinoma tumor system, it is tempting to speculate that Ad5-mTRAIL/CpG ODN therapy could be used against other solid tumors. The important benefit of this type of immunotherapy is that neither TRAIL nor CpG ODN is directly toxic against normal cells, but effective against tumor cells, allowing for a therapy with little side effects. However, caution must also be exercised as systemic autoimmune responses may develop in response to the self-antigens expressed by the tumors. Obviously, further study is needed to address these concerns and investigate the activity of Ad5-mTRAIL/CpG ODN in other tumor models.

Grant support: National Cancer Institute grant CA109446.

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.