Transduction of Dendritic Cells with Bcl-x_L Increases Their Resistance to Prostate Cancer-Induced Apoptosis and Antitumor Effect in Mice¹

Prostate cancer (PCa)³ remains the most common malignancy in American men, and the second most common cause of cancer-related death (1). Although more men are diagnosed at the early stages of the disease, when radical treatment is feasible, one-third of them will recur. PCa, in general, is resistant to chemotherapy, which may be due to its intrinsic low proliferative index (2). The development of new immunotherapeutic approaches that do not require a high proliferative rate of tumor cells has a strong rationale in the treatment of PCa. For the tumor to survive and grow, it should overcome immunologic defense mechanisms of the host. In many cases tumor cells achieve this either by suppressing the key elements of the host immune response and thus disrupting the antitumor reaction, or by promoting the tolerance of the immune system toward the tumor. PCa tumors are low in immunogenicity due to the lack of MHC class I molecules in the majority of cases (3). However, tumor-induced immunosuppression was well documented in PCa patients (4, 5). It has been recently demonstrated that tumor causes apoptotic death of key immunocompetent cells, including the major APCs, dendritic cells (DC) (6, 7). DC, first discovered in 1973 (8), originate in the bone marrow and migrate to the different lymphoid and nonlymphoid tissues. They recognize, uptake, and process Ag(s), including tumor Ags, and then present it to naive T cells, stimulating their proliferation (9). Elimination of DC from the tumor environment significantly diminishes the initiation of specific immunologic responses. In fact, it has been demonstrated that PCa is almost devoid of DC, and DC number further decreases in higher grade tumors (10, 11). The importance of DC in the induction of specific antitumor immunity has been recently documented in a variety of preclinical models (9), and several DC-based clinical trials have been initiated, including those with PCa patients (12, 13). The reported limited efficacy of these trials may be due to the fact that PCa cells cause significant suppression of human DC survival (14). Therefore, effective protection of DC from PCa-induced apoptosis may significantly improve the efficacy of DC-based therapies in cancer clinical trials.

There are several things that can increase the survival of DC. Activation of DC with cytokines is the most common method. For instance, IL-12, a pleotropic proinflammatory cytokine described in 1989 (15), has been shown to enhance the survival of hemopoietic progenitor cells (16, 17) and DC (18, 19). The strong antitumor activity of IL-12, demonstrated in a variety of immunotherapy and immune gene therapy models (20, 21, 22, 23, 24), may be at least in part mediated by the activation of the DC system (25). In fact, we have recently reported that administration of IL-12 resulted in increased generation of DC in vivo (26). It is likely that the stimulation of DC survival by IL-12 is due to the increased expression of the anti-apoptotic protein Bcl-x_L (14). Bcl-x_L is a known anti-apoptotic protein of a Bcl-2 family, capable of suppressing apoptosis in various cell types (7, 27) and playing an important role in the regulation of hemopoiesis and the survival of cells of the immune system (28, 29). The first aim of this study was to characterize PCa-induced apoptosis of DC. We have demonstrated that the progression of murine PCa was accompanied by decrease in the number of tumor-infiltrating DC and that PCa-induced apoptosis of DC was not mediated by a Fas/FasL interaction. The second aim of this study was to test the hypothesis that increased survival of DC results in a more efficient induction of antitumor immunity and a higher efficacy of DC-based therapies for cancer. We have evaluated the antitumor activity of DC transduced to overexpress IL-12 or Bcl-x_L in a murine PCa model and demonstrated that intratumoral administration of DC transduced with the bcl-x_L gene resulted in a strongest inhibition of tumor growth compared with that in other groups.

Male C57BL/6 mice, 6–8 wk old, were obtained from Taconic Farms (Germantown, NY). MRL/MpJ mice carrying the lpr mutation (lpr/lpr) and their parental wild-type controls were purchased from The Jackson Laboratory (Bar Harbor, ME). Animals were maintained at the Central Animal Facility at the University of Pittsburgh according to standard guidelines. RM-1, a murine prostate cancer cell line, was a gift from Dr. Timothy C. Thompson (Baylor College of Medicine, Houston, TX). Tumor cells were maintained in RPMI 1640 medium supplemented with 5% heat-inactivated FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, and 0.1 mM nonessential amino acids (Life Technologies, Grand Island, NY).

Murine DC cultures were established as described previously (30). Briefly, bone marrow cells were depleted of RBC with lysing buffer (155 mM NH₄Cl in 10 mM Tris-HCl buffer, pH 7.5, 25°C) for 2–3 min. The single-cell suspensions were then incubated with a cocktail of Abs for 1 h at 4°C, followed by incubation with rabbit complement for 30 min at 37°C to deplete cells expressing lymphocyte Ags B220, CD4, and CD8. Treatment with Abs (partially purified supernatant of hybridoma cell cultures TIB-146, TIB-207, and TIB-105 for B220, CD4, and CD8, respectively; American Type Culture Collection, Manassas, VA) and rabbit complement (Life Technologies) removed detectable B and T lymphocytes from the cell suspensions. Cells were then incubated overnight (37°C, 5% CO₂) in six-well plates (Falcon, Franklin Lakes, NJ) at a concentration of 10⁶ cells/ml in complete medium, consisting of RPMI 1640, 2 mM l-glutamine, 50 μg/ml gentamicin sulfate, 10 mM HEPES, 10% FBS, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate (Life Technologies), 1 μg/ml indomethacin (Sigma), and 50 μM N-methyl-l-arginine (Alexis, San Diego, CA). The nonadherent cells were then collected by gentle pipetting and resuspended at a concentration of 10⁵ cells/ml in complete medium supplemented with 1000 U/ml recombinant murine GM-CSF and recombinant murine IL-4 (both provided by Schering-Plough, Kenilworth, NJ). Cells were cultured in six-well plates (4 ml/well) for 7 days at 37°C in 5% CO₂. Nonadherent DC were collected by gentle pipetting, characterized as described previously (30), and used for further studies. In some experiments cultured DC were treated with recombinant murine IL-12 (Roche, Indianapolis, IN) at a final concentration of 100 ng/ml. IL-12 was added to DC on the sixth day of culture. Cells were collected 24 h later, washed, ad lysed, and expression of proteins was assessed by Western blot as described below.

Cells were washed in FACS medium (HBSS containing 0.1% BSA and 0.1% NaN₃) and stained with appropriately diluted Abs directly conjugated with PE according to the manufacturer’s recommendations. Hamster anti-mouse Abs against CD95 (Fas) and FasL (CD95L) were used (PharMingen, San Diego, CA), and 10,000 cells were analyzed by flow cytometry with the CellQuest 1.0 software package (Becton Dickinson, San Jose, CA). For intracellular FasL detection, cells were fixed in cold 1% paraformaldehyde and methanol and permeabilized using 0.01% saponin (Sigma, St. Louis, MO) in FACS medium. The staining procedure was the same as that described above.

Tissue samples were embedded in OCT compound (Miles, Elkhart, IN), snap-frozen on dry ice, and stored at −80°C. Cryostat sections (6 μm) were air-dried and fixed in ice-cold acetone. Slides were washed in PBS and incubated for 1 h at room temperature with the following Abs: NLDC-145 (DEC 205; Serotec, Raleigh, NC; dilution, 1/5) and CD11c (N418; Serotec; dilution, 1/800). Biotinylated mouse anti-rat IgG (Jackson ImmunoResearch Laboratories, West Grove, PA; 1/500 dilution) and anti-hamster IgG (Vector, Burlingame, CA; dilution, 1/1000) were used as secondary Abs and were applied for 45 min. After developing using the peroxidase chromogen kit (3-amino-9-ethylcarbazol, Biomega, Foster City, CA) for 8 min, counterstaining was performed with hematoxylin. Two investigators analyzed the slides independently to determine the number of positive cells per area. At least 10 different areas were analyzed. Negative controls included the staining with irrelevant isotype Abs.

DC were harvested after coincubation with tumor cells and were fixed on microscope slides for morphologic evaluation using a Cytospin centrifuge (Shandon Lipshaw, Pittsburgh, PA). After drying for 5 min, cells were fixed and stained with a LeukoStat Stain Kit (Fisher Scientific, Pittsburgh, PA), and the percentage of apoptotic cells was assessed quantitatively using morphologic criteria, which include condensation of chromatin, reduction in nuclear size, shrinkage of total cell volume, increase in cell density, and formation of apoptotic bodies (31).

For annexin V binding assay and propidium iodide (PI) uptake evaluation, DC were collected and double stained with FITC-conjugated annexin V (PharMingen) and/or PI (Sigma). Annexin V was added according to the manufacturer’s recommendations. PI was used at a final concentration of 10 μg/ml. All annexin V-positive cells were considered apoptotic, and their percentage was calculated among the total number of cells. Cells taking vital dye PI were considered dead. Samples (10,000 cells) were analyzed by FACScan. When green fluorescent protein (GFP)-transfected cells were analyzed, only PI staining was used for cell death assessment.

The cell viability in the cultures after transfection with different adenovectors was evaluated using trypan blue (0.2%; Life Technologies). Trypan blue-positive cells were considered dead, and their percentage among the total cell number was calculated.

Adenovirus (Ad5) vectors were constructed through Cre-Lox recombination with reagents generously provided by Somatix (Almenda, CA). Construction and characterization of the Ad5 vector expressing mIL-12 have previously been described (23). In brief, pAdCMV-mIL-12 was constructed by transferring a BglII/BamHI fragment containing the p40 subunit of mIL-12 into the BamHI site of pAdCMV-B. A BamHI fragment containing the p35-encoding gene was subsequently cloned in the BamHI site behind the p40 subunit. This resulted in a polycistronic message expressing both subunits of mIL-12. Virus generated by recombination between the shuttle vector and the Ad DNA was propagated on 293 cells and purified from infected cells 2 days after infection by three freeze-thaw cycles followed by three successive bandings on CsCl gradients.

For the construction of Bcl-x_L encoding adenoviral vector (Ad5Bcl-x_L), the mouse bcl-x_L gene was cloned from a BALB/c mouse total spleen using the PCR technique. Primers were designed according to the published sequence (GenBank accession no. U51278), in which additional restriction enzyme target sequences were added. The forward sequence was 5′-TCA GAG CTC ATG TCT CAG AGC AAC AGG GAG-3′, and the reverse sequence was 5′-CTA GGC GGC CGC GTC TGG TCA CTT CCG ACT GAAG-3′. Mouse Bcl-x_L cDNA was subcloned into expression plasmid pCR3.1 (Invitrogen, Carlsbad, CA) and subsequently sequenced (GenBank accession no. AF060226). To prove that plasmid contained the correct insert sequence, pAdlox-mBcl-x_L was digested with EcoRI-restricted enzyme. The solution containing 5 μl of EcoRI buffer, 2 μl of EcoRI enzyme, 5 μl of pAdlox-mBcl-x_L, and 38 μl of dH₂O was incubated for 2 h at 37°C and then loaded on an 1% agarose gel at the appropriate concentration followed by electrophoresis. A SalI-NotI fragment containing Bcl-x_L cDNA derived from pCR3.1-mBcl-x_L was inserted in the shuttle vector pAdlox (Fig. 1 A). An E1-substituted recombinant adenovirus was generated by cotransfection of ψ5 helper virus DNA and SfiI-digested pAdlox-mBcl-x_L into the adenoviral packaging cell line CRE8, propagated, and purified as previously described (32).

Control adenoviral vector encoding the enhanced GFP gene (Ad5EGFP) was constructed by inserting a SnaBI-HpaI fragment containing part of the CMV promoter, the EGFP cDNA, and part of the SV40 poly(A) derived from pEGFP_NI (Clontech, Palo Alto, CA) (33) into the shuttle vector pAdlox. Remaining steps for the construction of Ad5EGP were similar to those for Ad5Bcl-x_L construction.

For the transfection, murine bone marrow-derived DC were harvested on day 5, washed twice in HBSS, and incubated at 37°C with the corresponding adenoviral vectors. Virus was used at a dose of 200 multiplicity of infection (MOI). Complete medium supplemented with mouse recombinant GM-CSF and mouse recombinant IL-4 was added 2 h later, and cells were allowed to recover for 24 h. Then cells were washed twice in HBSS and injected intratumorally at a dose of 10⁶ DC/mouse/injection.

The level of expression of the anti-apoptotic protein Bcl-x_L was examined using a Western blot technique. Briefly, cells were collected, washed in PBS, and homogenized in lysing buffer. The homogenate was centrifuged at 12,000 × g for 15 min at 4°C. The protein concentration in the supernatant was determined by the Bradford method using the Bio-Rad kit (Bio-Rad, Hercules, CA). Each sample was denatured for 5 min at 100°C in sample buffer. Equal amounts of protein were loaded for each sample in all lanes and electrophoretically separated on 16.5% SDS-PAGE followed by transfer to a nitrocellulose membrane. The membrane was blocked with 0.2% nonfat milk and 0.1% Tween-20 (Fisher, Fairlawn, NJ) in 20 mM Tris-HCl buffer, pH 7.2. Bcl-x_L was detected using specific rabbit primary Abs (Oncogene Research Products, Cambridge, MA) with a final concentration of 2.5 μg/ml, and donkey anti-rabbit secondary Abs (1/2000 dilution; Amersham Pharmacia Biotech, Piscataway, NJ). The membrane was processed and treated with chemiluminescence reagents (Tropix, Medford, MA). The bands were visualized on Kodak film (Eastman Kodak, Rochester, NY) exposed to the membrane to detect chemiluminescence signals.

Five- to 6-day-old cultured DC were harvested and coincubated with the murine prostate cancer cell line RM-1 in six-well plates. DC and tumor cells were separated using membrane inserts with 0.4-μm pore size, which excluded direct cell-to-cell contact, but allowed free exchange of soluble tumor-derived factors. Specifically, 5–10 × 10⁵ DC were placed in six-well plates in 3 ml of medium. Two million PCa cells resuspended in 2 ml of medium were placed into the inserts on the top of each well. As controls, DC were coincubated with murine splenocytes or medium alone placed in inserts. DC were harvested 48 h later, and apoptosis was assessed using the morphological criteria and an annexin V binding assay. For the positive control, DC were irradiated (25,000 rad), and apoptosis was assessed after 16 h.

To determine DC survival in cultures after transfection, cells were left in six-well plates without fresh medium and cytokines, and the number of dead cells was counted every other day using trypan blue. Each experiment was repeated twice, and combined data are presented.

RM-1 tumor cells (20,000 cells/100 μl) were inoculated s.c. in the right flank of C57BL/6 mice, and tumor establishment was determined by palpation. Measuring the perpendicular tumor diameters with a Vernier caliper (Electron Microscopy Sciences, Ft. Washington, PA) assessed tumor size. Tumor volume was calculated using the formula of rotational ellipsoid: m₁² × m₂ × 0.5236, where m₁ represents the shorter axis, and m₂ the longer axis (34). Treatment groups consisted of five mice per group. Mice were sacrificed when they exhibited signs of distress or when total tumor volume exceeded 3000 mm³. Experiments were independently repeated three times. Combined data from these experiments are presented.

A χ² analysis was performed to evaluate the significance of differences between the experimental groups in the annexin V and PI staining assays when discrete data were presented. For a single comparison of two groups, Student’s t test was used. If the data distribution was not normal, the Mann-Whitney rank-sum test was employed for the nonparametric analysis. Two-way ANOVA using the Student-Newman-Keuls method was employed for comparison of tumor size in mice after the different types of treatment. For all analyses, the level of significance was set at p < 0.05. All statistical calculations were performed using the SigmaStat statistical software package (SPSS, Chicago, IL). Data are presented as the mean ± SEM.

Murine DC were generated from bone marrow precursors, and RM-1 cells were added on day 5 in cell inserts (0.4-μm pore size) for an additional 48 h. Based on morphological characteristics, the levels of apoptotic DC were 13.48 ± 2.16 and 31.13 ± 3.43% after coincubation of DC with splenocytes or RM-1 cells, respectively (p < 0.001). The results of the annexin V binding assay revealed that the percentages of apoptotic DC were 17.71 ± 2.23 and 53.66 ± 6.06% among DC coincubated with splenocytes or RM-1 cells, respectively (p < 0.001; Fig. 1 A). In irradiated DC cultures, which served as a positive control, the level of apoptotic cells was 66.74 ± 5.47%. Thus, these data suggested that PCa cells release factors that cause apoptotic death of DC.

To elucidate the mechanism of regulation of tumor-induced DC apoptosis and to determine whether the family of Bcl-2 proteins is involved in these pathways, DC were collected after coincubation with splenocytes or RM-1 cells for 48 h, washed, and lysed, and proteins were extracted. Western blot was used to assess the expression of Bcl-x_L. Since equal amounts of protein were loaded in all lanes, the results presented in Fig. 1 B suggest that coincubation of DC with RM-1 cells, but not splenocytes, significantly decreased the expression of anti-apoptotic protein Bcl-x_L in DC. Taken together, these data demonstrated that PCa-derived soluble factors caused apoptosis of DC in vitro, and that decreased expression of Bcl-x_L in DC might mediate this effect. These data also raise the question of whether DC might undergo apoptosis within the tumor microenvironment in vivo.

To determine the infiltration of PCa tissues by DC, 10⁶ RM-1 cells were injected into the shaved right flank of C57BL/6 mice. Mice were sacrificed at different time points, tumor was measured and harvested, and immunohistochemical staining was performed using murine DC markers NLDC-145 (DEC 205) and CD11c (N418). DC were readily identifiable within the tumor tissues on day 7, but their numbers significantly decreased within 10-day-old tumors, and tumor-infiltrating DC virtually disappeared within 20-day-old tumor tissues (Fig. 2). Importantly, similar results were obtained when both DC-specific markers were used to evaluate the levels of DC infiltration within the tumor tissues. Thus, these data suggest that the progression of prostate carcinoma in vivo was accompanied by the marked decrease in DC infiltration within the tumor tissue. Although RM-1 cells are MHC class I negative (data not shown) and might be considered nonimmunogenic, the fact that DC were identified at the site of the tumor at the beginning of tumor growth and disappeared later during tumor progression taken together with our in vitro findings supports the conclusion that RM-1 cells markedly inhibit DC survival and/or accumulation within the tumor tissue.

To evaluate the role of the Fas/FasL (CD95/CD95L) interaction in apoptosis of DC induced by prostate cancer cells, MRL/MpJ mice with functional mutations of Fas (lpr/lpr) (35) were used in the next experiments. Wild-type control MRL/MpJ (+/+) and C57BL/6 mice were used as controls. FACScan analysis of FasL expression on RM-1 cells revealed moderate levels (∼30–35% of cells) of expression. Expression of Fas on DC generated in vitro from wild-type MRL/MpJ and C57BL/6 mice reached 70.66 and 61.83% of that in positive cells, respectively, while DC generated from lpr/lpr mice demonstrated a low level (2.78%) of Fas expression (Fig. 3). Next, DC obtained from different mouse strains were coincubated with RM-1 cells or splenocytes for 48 h, and apoptosis was assessed by annexin V and PI staining. Two independent experiments were performed with identical results, and combined data are presented. As expected, the level of apoptosis in control C57BL/6 mice-derived DC cultures was 16.51 ± 3.06%, while RM-1 cells caused the death of 44.47 ± 4.78% of these DC (p < 0.01). In wild-type MRL mice-derived DC cultures, RM-1 caused an elevation of the apoptotic rate from 32.27 ± 0.91% in control DC to 45.56 ± 4.41% (p < 0.05). Fas-deficient DC have demonstrated the similar sensitivity to PCa-induced apoptosis; coincubation with RM-1 cells led to an increase in the apoptotic rate from 22.96 ± 6.69 to 33.86 ± 2.45% (p < 0.05). Although the level of spontaneous apoptosis in DC cultures obtained from MRL mice was higher than that in C57BL/6-derived DC, and Fas-deficient DC appeared to be somewhat less sensitive to apoptosis, there was no significant difference between the PCa-induced apoptotic rate among DC derived from Fas-deficient or wild-type mice (1.47- and 1.41-fold increases, respectively). Furthermore, the increase in levels of RM-1-induced DC apoptosis was statistically significant in both groups. Thus, these data suggest that the Fas/FasL interaction probably does not play a crucial role in PCa-induced DC apoptosis in a murine system.

Since IL-12-encoding adenoviral vector was characterized previously(23), in this study we have evaluated transduction of DC with murine bcl-x_L gene and transgene expression (Fig. 4). For this purpose, murine cultured DC were collected on day 5 and transduced with IL-12 or the bcl-x_L gene using adenoviral vectors. Control transfection was performed with the GFP-encoding gene. In most experiments, the transfection efficacy varied between 60 and 80%, as determined by a FACScan analysis of GFP-transfected DC (data not shown). Western blot performed on Bcl-x_L-transduced DC showed significant elevation of Bcl-x_L protein expression compared with that in control cells (Fig. 4), suggesting a high transfection efficacy of the method used.

We have recently shown that the treatment of human cultured DC with human IL-12 protein in vitro results in prolonged DC survival and increased resistance to tumor-induced apoptosis (14). Since our previous studies also indicated that prolonged survival of human DC was accompanied by increased expression of Bcl-x_L, we first confirmed these results in a murine system. As shown in Fig. 5, stimulation of murine bone marrow-derived DC with mouse IL-12 protein (100 ng/ml) for 24 h resulted in increased expression of the anti-apoptotic protein Bcl-x_L as assessed by Western blot. Taken together, these data suggest that IL-12-induced increased survival of both human and murine DC is mediated by enhanced expression of the anti-apoptotic protein Bcl-x_L in these cells. This allowed us to hypothesize that the increased concentrations of IL-12 in the local DC environment or stimulation of Bcl-x_L synthesis might result in prolonged DC survival. To test this hypothesis, we have evaluated whether overexpression of IL-12 or Bcl-x_L by DC might affect their survival in long-term cultures and their sensitivity to prostate cancer-induced apoptosis in vitro.

First, we determined the survival of DC in cultures using the trypan blue excretion method (Fig. 6). Murine DC were generated in cultures with GM-CSF and IL-4 added on day 1 and day 3. Transfection was performed on day 5, and DC were further cultured without cytokines and growth factors. The analysis of cell viability demonstrated that starting from day 11, the percentage of dead cells among nontransfected or GFP-transfected DC cultures (38.33 ± 1.67 and 45.98 ± 10.69%, respectively) was significantly higher compared with that in IL-12- and Bcl-x_L-transfected DC cultures (20.10 ± 3.43 and 13.89 ± 2.78%, respectively; p < 0.05). Interestingly, starting from day 15, the survival of IL-12-transduced DC also declined, whereas the death rate of Bcl-x_L-transduced DC remained significantly lower compared with those in all other groups (p < 0.01). Almost 50% of DC overexpressing Bcl-x_L were alive in 20-day-old cultures supplemented with no growth factors (Fig. 6 b). These results suggest that DC genetically engineered to produce IL-12 performed prolonged survival for 1 wk in growth factor-deprived cultures, whereas overexpression of an anti-apoptotic protein Bcl-x_L resulted in significantly prolonged survival for at least 2 wk during growth factor withdrawal.

In the next series of experiments we examined whether transfection of DC with IL-12 or bcl-x_L genes protects them from tumor-induced apoptosis. DC were transduced on day 5, recovered for 24 h, and then coincubated with RM-1 cells on day 6 for 48 h as described above. DC were then stained with PI, and determination of the number of PI-positive (dead) cells was assessed by FACScan. Analysis of these results suggested that IL-12- or Bcl-x_L-transfected DC were characterized by increased resistance to PCa-induced apoptosis compared with nontransduced or GFP-transduced DC. For instance, the levels of PI-positive cells among nontransfected and GFP-transfected DC cultures coincubated with RM-1 cells were 30.33 ± 5.55 and 33.47 ± 3.35%, respectively, whereas the transduction of DC with IL-12 or bcl-x_L genes significantly decreased levels of PCa-cancer-induced cell death up to 20.80 ± 0.69 and 10.98 ± 1.78%, respectively (p < 0.01; Fig. 6 C). Thus, these data suggest that overexpression of IL-12 or Bcl-x_L in murine DC results in their prolonged survival in cultures and increased resistance to PCa-induced apoptotic death in vitro.

In summary, our in vitro data demonstrate that the adenoviral transfection of murine bone marrow-derived DC with different genes is a highly efficient procedure resulting in high levels of protein expression. Transfection of DC with IL-12 and bcl-x_L genes led to the enhanced survival in cultures and increased resistance to PCa-induced apoptosis. These results raise the question of whether the enhanced resistance of DC to tumor-induced cell death would be accompanied by an increased antitumor efficacy of these DC.

In the next series of experiments we examined the antitumor efficacy of IL-12- and Bcl-x_L-transfected DC in a murine PCa tumor model. On day 1, mice were injected with 2 × 10⁵ RM-1 tumor cells into the right flank. Syngeneic DC were generated from the bone marrow and transfected with adenovector encoding IL-12, bcl-x_L, or GFP gene. Cells were allowed to recover for 24 h, harvested, and injected intratumorally (10⁶ cells in 100 μl of HBSS) on day 5. Five groups of animals (five mice per group) were used in each experiment. Group 1 was injected with IL-12-transfected DC, group 2 with Bcl-x_L-transfected DC, group 3 with GFP-transfected DC (control for the adenovirus), group 4 with nontransfected DC, and group 5 with HBSS. All injections were repeated on day 12. Animals were sacrificed 20–21 days after tumor cell injection when the size of the tumor in the control group became large enough to be distressing for the animals. Experiments were repeated three times with similar results, and the combined data are presented on Fig. 7.

The mean tumor volume of the mice treated with HBSS injection alone was 2650.6 ± 506.3 mm³ on day 19. The mean tumor volumes in mice treated with DC alone or with GFP-transfected DC were 1690.3 ± 269.3 and 1566.8 ± 205.4 mm³, respectively. Both these values were significantly (p < 0.05) smaller than tumor sizes in mice treated with HBSS. Treatment of mice with IL-12- or Bcl-x_L-transfected DC resulted in further inhibition of tumor growth and a decrease in tumor size (1025.9 ± 252.2 and 816.5 ± 174.0 mm³, respectively). Tumor sizes in both these groups were significantly (p < 0.05) smaller than tumor sizes in other groups of animals. Although the mean tumor size was smallest in mice treated with Bcl-x_L-transfected DC in all independent experiments, the difference between IL-12- and Bcl-x_L-transfected DC did not reach statistical significance. Thus, these data suggest that overexpression of IL-12 or Bcl-x_L in DC significantly improved the antitumor efficacy of DC-based immunotherapy in mice.

In summary, treatment of PCa-bearing mice with DC protected from tumor-induced apoptosis appears to be an effective approach for cancer immunotherapy even in the case of nonimmunogenic, fast growing, and aggressive tumors.

During the last several years immunotherapy has been employed more frequently in both preclinical studies and human clinical trials. DC-based therapy represents a relatively new approach to treat tumor-bearing hosts. DC pulsed with isolated or synthetic tumor-associated peptides have been shown to induce an effective antitumor immune response in animal models (13, 36) and cancer patients (12, 37). DC can also be pulsed with a tumor lysate (38) or tumor RNA (39) to induce a specific immune response against tumors. For instance, it has been recently demonstrated that human DC transfected with RNA encoding prostate-specific Ag stimulate prostate-specific CTL responses in vitro (40). However, most of these approaches have certain limitations when considered for clinical use in humans, since the preparation of tumor lysates or extraction of tumor Ags requires a large amount of solid tumor, which may not be suitable in all cases. Thus, development of DC-based therapies that do not require tumor Ags is highly justified and timely.

In this study we have evaluated the delivery of genetically modified DC directly to the tumor site. The working hypothesis was that tumor-resistant DC would recognize and pick up tumor Ag(s), emigrate from the tumor mass, traffic to the regional lymph nodes to present the Ag(s) to T cells, and induce a specific antitumor response. Since we did not pulse DC with tumor Ag in this study, the important question was whether administered DC could recognize and process endogenous tumor Ags at the site of the tumor. Although synthetic or stripped Ags, tumor lysates, tumor apoptotic bodies, or tumor-derived RNA or DNA ensure relative specificity, they cannot provide protection of DC from tumor-induced apoptosis per se. Furthermore, as was demonstrated in our studies, tumor expansion was associated with the local elimination of DC within the tumor, and the coincubation of DC with RM-1 tumor cells resulted in the massive apoptotic death of DC. Importantly, PCa-induced apoptosis in DC was accompanied by a decreased expression of the anti-apoptotic protein Bcl-x_L. Therefore, we have examined whether genetic modification of DC increases their resistance to tumor-induced apoptosis and their potential to induce antitumor immune responses. Based on our previous results, IL-12 and bcl-x_L genes were chosen for this purpose.

IL-12 is a potent proinflammatory cytokine, which is produced by DC in response to CD40 ligation (41, 42, 43). IL-12, in turn, stimulates the expression of CD154 (CD40 ligand) on T cells (44) and IFN-γ production by NK and T cells (15, 45, 46) and inhibits or reverses anergy of T cells (47, 48). In addition to its immunopotentiating and antitumor effects, human IL-12 enhances resistance of human DC to prostate cancer-induced cell death (14), and as shown here, IL-12-transfected murine DC demonstrated significantly higher levels of survival in both long-term cultures and tumor microenvironment in vitro. An IL-12 cDNA-encoding adenoviral vector has been successfully used in the treatment of various tumors (22, 23, 24), including prostate cancer in mice (49). DC transfected with IL-12 using retroviral vector were also effective (50), although the antitumor properties of adenovirally transduced DC, overexpressing IL-12, have not yet been evaluated.

In this study we used an adenoviral vector approach for the genetic engineering of DC. Gene expression of an adenoviral vector is transient, since the transferred gene is not incorporated into the target cell chromosomal DNA. In our experiments this fact was not an issue, because the long-term expression of transfected genes was not required. Another concern that was considered to be a problem by some authors (51) is immunogenicity of adenoviral vectors. However, Bramson et al. recently demonstrated that the pre-existing immunity against the adenovirus reduces virus dissemination, but not the antitumor effect of therapy (52). In addition, adenovirus-mediated transfection of various cell types proven to be a highly efficient procedure. We have confirmed these results for murine DC and demonstrated here that the level of transfection in GFP-transduced DC was between 60 and 80%, as assessed by flow cytometry. Murine DC transfected with IL-12-encoding vector produced high levels of IL-12 protein, as determined by ELISA (data not shown). As demonstrated in this study, these cells caused a significant inhibition of PCa growth in mice when injected into the site of the tumor even without prior pulse with a tumor Ag. It is likely that both a direct effect of IL-12 on different immune cells and stimulation of DC survival contributed to the observed antitumor effect of IL-12-transfected DC in a murine PCa model.

It is of a great interest that the intratumoral administration of Bcl-x_L-transfected DC induced a stronger antitumor effect than the treatment with IL-12-transfected DC, although the difference was not statistically significant. It suggests that an increased survival of DC may be a key mechanism of the induction of antitumor immune responses initiated by the intratumoral administration of IL-12-transfected DC. It is possible that DC, overexpressing Bcl-x_L, also produce IL-12. Furthermore, it is likely that the overexpression of Bcl-x_L is a stronger survival factor than the overexpression of IL-12. Although we observed a significant difference between the survival of Bcl-x_L- and IL-12-transfected DC only after day 15 in cultures, it is quite possible that the viability of transfected cells in the tumor microenvironment in vivo might be different from the cell survival in vitro.

Bcl-x_L belongs to the Bcl-2 family of proteins. Members of this family with anti-apoptotic properties, including Bcl-2 and Bcl-x_L, have been shown to protect cells from NO-mediated apoptosis (53) and markedly inhibit caspase activation (54, 55, 56) and Fas/FasL (CD95/CD95L)-mediated killing of cells (57, 58).

The expression of FasL by different tumor cells and their ability to cause apoptosis of Fas-expressing T cells has been proposed as an important mechanism of tumor escape from immune surveillance (59). Since DC have been reported to express Fas and be susceptible to Fas-mediated apoptosis (58), it has been suggested that Fas/FasL-mediated tumor-induced apoptosis of DC is an additional mechanism that allows tumor cells to escape immune recognition and elimination (60). Here, we have demonstrated a high level of FasL expression by RM-1 PCa cells and evaluated the role of the Fas/FasL pathway in the induction of DC apoptosis by prostate cancer using Fas-deficient cells obtained from lpr/lpr knockout animals. Our data suggested that Fas-mediated pathway was not involved in RM-1-induced DC apoptosis, since we did not detect a significant difference in the sensitivity to apoptosis between Fas-deficient and wild-type DC. Interestingly, lpr/lpr DC appeared to be slightly less sensitive to apoptosis, although the statistical analysis did not reveal a significant difference compared with control DC. Furthermore, neutralizing anti-FasL Ab failed to block tumor-induced DC apoptosis as well. Thus, it is likely that other tumor cell-derived factors are involved in DC apoptosis in the case of the murine prostate cancer cell line RM-1. These results confirm our early findings suggesting that B16 melanoma-induced apoptosis of DC was not mediated by Fas/FasL interaction (19). Further studies are required to identify the DC-killing factors produced by different tumor cell lines and primary tumors as well as to determine mechanisms of antitumor activity of Bcl-x_L-transfected DC.

In conclusion, we have shown here that murine PCa causes significant inhibition of DC survival both in vivo and in vitro, and that DC viability can be significantly augmented in vitro by an increased expression of the anti-apoptotic protein Bcl-x_L. In vivo treatment of PCa-bearing mice with IL-12- or Bcl-x_L-transfected DC demonstrated a strong antitumor activity of intratumorally administered DC that have not been pulsed with tumor-associated Ag. Taken together, these data suggest that protection of DC from tumor-induced apoptosis and enhanced survival of DC in the tumor microenvironment are key factors that may significantly improve DC-based therapies of cancer patients.