Prostate cancer is the most common male cancer in Europe, North America and some parts of Africa, and its incidence is increasing steadily in almost all countries.1 Causes have not yet been fully clarified: both environmental factors (including diet) and genetic components are presumably responsible. Because of the long period between the initiation of the premalignant latent phase and progression to invasive prostate cancer (10–20 years), this disorder is a good candidate for prevention studies, which in men have mainly focused on chemoprevention, even though with discordant results.2 Other approaches, such as the new idea of cancer prevention through immune approaches,3, 4, 5, 6 could therefore be investigated in experimental prostate cancer models.7, 8, 9, 10
TRAMP mice are a widely studied model of gene-driven prostate carcinogenesis.11, 12 They are transgenic for and tolerant to the SV40 large T antigen (TAg), under the control of the rat probasin regulatory element. As in men, these mice develop well- or moderately-differentiated adenocarcinomas that express androgen receptor, undifferentiated carcinomas that do not express it in the prostate and phyllodes tumors in the seminal vesicles.13
Immunoprevention studies performed in TRAMP mice could substantiate the notion that prostate cancer progression, even when based on genetic factors, can be delayed by appropriate immune approaches. The best results obtained so far in TRAMP mice exploited passive approaches (such as the adoptive transfer of nontolerant, nontransgenic immune cells),14, 15 whereas overcoming tolerance with active immunization was a more challenging goal.16, 17, 18, 19, 20
This paper investigates the efficacy of an active immunoprophylactic approach combining specific antigens with strong adjuvant stimuli. Tumor-free TRAMP mice were vaccinated with allogeneic TAg-positive cells associated with the systemic administration of recombinant interleukin 12 (rIL-12), thus combining specific targets (TAg) with polyclonal stimulation (allogeneic histocompatibility) and with IL-12, which is a powerful Th1 inducer.21
TRAMP mice developed on a pure C57BL/6 background were bred in the Animal Care Facility (CeSI, Gabriele d'Annunzio University Foundation, Chieti, Italy). Experiments were authorized by the local Animal Use and Care Committee. To study the development of urogenital tumors, groups of mice were sacrificed at 5-week time intervals and when they were moribund. Animals were subjected to blind monitoring for tumor development by abdominal pelvic palpation by 2 independent researchers twice a week.
SV40 transformed cell lines used throughout the study were: C57SV (C57BL/6 origin, here referred to as TAg/Syn) and KB/cSV (BALB/c origin, TAg/Allo).22, 23 In some experiments the following cell lines derived from spontaneous tumors were used as T antigen (TAg)-negative or low-expressing controls: TUBO, BALB/c origin (here referred to as TAgneg/Allo cells)24; LAM1, a cell line derived from a TRAMP prostate carcinoma with a low TAg expression (TAglow/Syn); B78H1, C57BL/6 origin (here referred to as TAgneg/Syn cells).25 Cells were cultured in Dulbecco's MEM supplemented with 10% fetal bovine serum in 5% CO2 atmosphere.
Vaccination protocol combining cell vaccine and rIL-12 was scheduled as previously reported.26 Starting at 5–6 weeks of age, mice entered the vaccination protocol: in the first 2 weeks mice received 4 twice-weekly i.p. vaccinations with 2 × 106 mitomycin C-treated cells, followed in the third week by 5 daily i.p. administrations of rIL-12 (provided by S. Wolf, Genetics Institute, Andover, MA), 50 ng in the first course and 100 ng thereafter. Mice was given a week rest. Control group consisted in TRAMP mice receiving only the vehicle. Four-week cycles were repeated lifelong.
Histology, immunohistochemistry and electron microscopy
Histologic evaluation and immunohistochemistry were performed as reported previously.26 For electron microscopy, tissue specimens were fixed for 4 hr in 2.5% glutaraldehyde in HEPES buffer 0.2 M, pH 7.4 and processed as reported.27 A transverse section through the urethra was made to include both dorsolateral and ventral lobes of the prostate, and the trimmed sample was embedded with cut surfaces down. Hematoxylin-eosin stained slides were subjected to blind examination by 3 pathologists. In the obtained sections all the prostatic lobes were displayed and the whole prostatic area was measured (about 6 fields 100× for each section). Each section was scored and the digital images, acquired with a Leica DC500 camera, were analyzed by Adobe Photoshop. The percentage of normal gland, prostatic intraepithelial neoplasia (PIN), well or moderately differentiated adenocarcinoma (ADC) and small cell carcinoma (SCC) was calculated with respect to the total prostate area. The area corresponding to each pathological stage was evaluated by the Lasso Tool and recorded on an Excel spreadsheet. The scores were then averaged and expressed as mean ± standard error.
Mixed lymphocyte-tumor cell cultures (MLTC) were performed with spleen cells cocultured at a 50:1 ratio with proliferation-blocked restimulator cells for 6 days in RPMI 1640 supplemented with 10% fetal bovine serum and with 20 U/ml of recombinant interleukin 2 as reported previously.28 The cytotoxic ability of lymphoblasts was then tested in secondary cultures against TAg/Allo, TAg/Syn or TAgneg/Allo cells by a standard 51Cr-release assay and % of lysis was calculated as described.29 Supernatants from MLTC primary cultures performed as above with a 10:1 splenocytes:tumor cell ratio were assayed for Interferon-γ (IFN-γ) and Interleukin 4 (IL-4) by ELISA assays.28
PE-labeled Pro5 Kb/TAg-IV pentamers, carrying the modified TAg epitope IV (VVYDFLKL) as reported,30, 31 were custom synthesized by Proimmune Limited, Asnieres-sur-Seine Cedex, France. Kb-ovalbumin pentamers were used as nonspecific controls. Lymphocytes from MLTC cultures performed as above were stained according to the manufacturer's protocol with PE-labeled pentamers (diluted 1:6) for 15 min in the dark. Then cells were washed and incubated with FITC-labeled anti-CD8 (diluted 1:25) for 30 min at 4°C. After washing cells were subjected to cytofluorometric analysis with FACScan (Becton Dickinson). Dead cells were excluded by physical parameters.
Sera collected from individual mice were used as primary antibody to stain different target cells (then subjected to cytofluorometric analysis) or in complement-mediated cytotoxicity, as described previously.26
Immunoprecipitation and Western blot
Membrane protein enriched fractions of SV40 TAg-positive and -negative cell lines were obtained by the ProteoExtract Native Membrane Protein Extraction Kit (Calbiochem, Darmstadt, Germany), following the “Adherent Tissue Culture Cells” extraction protocol. Membrane protein fractions (250 μg) were first immunoprecipitated with 10 μl of TRAMP control mice sera for 2 hr at 4°C and collected by 2-hr incubation with protein A-agarose (Santa Cruz Biotechnoloy, Santa Cruz, CA). The supernatants (membrane protein fractions depleted of common recognized antigens) were then incubated for 2 hr at 4°C with 10 μl of triplex vaccinated mice sera or with 2 μg of anti-SV40 TAg mouse monoclonal antibody (Ab-2; Oncogene Research Products, Darmstadt, Germany) and immunoprecipitated proteins were eluted by overnight incubation with protein A-agarose (Santa Cruz Biotechnology). The presence of TAg among the immunoprecipitated proteins was detected by Western Blot analysis with the anti-TAg monoclonal antibody, diluted 1:150 in PBS containing 0.1% Tween 20 and 5% nonfat dry milk, followed by incubation with HRP-linked goat antirabbit antibody (Santa Cruz Biotechnology), using a colorimetric reaction (Opti-4CN Substrate kit; Bio-Rad).
Natural history of prostate carcinoma in TRAMP mice
TAg-positive cells, intensively proliferating cells and with a widely distributed positivity for Proliferating cell nuclear antigen (PCNA), appear in the prostatic epithelium of all TRAMP mice at 5 weeks of age. At 10–12 weeks many atypical hyperplasia sites (focal mPIN corresponding to human PIN)32 are found with endophytic growth and TAg-positive lined papillary projections which partly occupy the lumen of the prostatic ducts. Neoplastic proliferation is particularly evident in the lateral and posterior lobes and progressive, resulting in ducts almost totally filled by a well-differentiated adenocarcinoma at 18–20 weeks. After the 24th week, this tumor becomes moderately differentiated insofar as the cells in mitosis are more numerous and the neoplastic cells are sometimes arranged in several layers, but not usually invasive, and there are only sometimes evident signs of infiltration of the walls of the ducts. This cancer, in fact, is rarely fatal. In about one-third of these mice, however, anaplastic and highly invasive small-cell carcinomas also appeared at 19–22 weeks. These tumors originated from the basal cells, located above the basement membrane with no contact with the lumen, and expressed both TAg and PCNA. Their growth is very rapid and aggressive giving rise to substantially solid tumors that invade and replace the adenocarcinoma. Extremely voluminous masses are soon formed and these and their paraaortic, lymph node, adrenal, perirenal, lung and liver metastases result in death by the 26th week. Cells are frequently positive for synaptophysin and N-CAM, 2 neuroendocrine differentiation markers.
In the remaining two-thirds of mice, the adenocarcinoma is joined from the 27th week on by a phyllodes tumor of the seminal vesicles. This slow-growing tumor eventually gives rise to voluminous masses due to actively proliferating, cytokeratin-negative, TAg-positive fibroblast-like cells that then obstruct seminal vesicles ducts and block seminal fluid drainage. TAg-positive cells are particularly dense near the epithelial infoldings lined with cytokeratin-positive, PCNA and TAg-negative cells. Metastases are not found and death occurs as late as the 42nd week, usually due to urinary blockage.
Neoplastic proliferation is associated with angiogenesis, initially during the formation of the papillary projections within the lumen and then during the initial stages of the rapid growth of the small-cell carcinoma.
Carcinogenesis inhibition by vaccination
Prostate carcinogenesis in TRAMP mice was significantly delayed (p < 0.05, Mantel-Haenszel test) by a prophylactic vaccine combining allogeneic SV40-transformed cells with systemic rIL-12 (hereafter referred to as TAg/Allo/IL-12 vaccination) (Fig. 1). Median latency time of tumor masses observed in the controls (26 weeks) increased to 53 weeks following TAg/Allo/IL-12 vaccination. The TAg/Allo cell component of the vaccine was ineffective when administered alone, whereas rIL-12 alone was only partially effective, giving a median latency of 39 weeks. In all the experimental groups death occurred 5–6 weeks after the onset of a palpable tumor, therefore TAg/Allo/IL-12 vaccination mainly delayed early phases of prostate carcinogenesis. In mice that received the TAg/Allo/IL-12 vaccine a less diffuse prostatic hyperplasia with less numerous papillary projections and fewer TAg-positive epithelial cells were already evident at 8 weeks (Fig. 2b and Table I). At 23 weeks in vaccinated mice most prostate ducts were not yet completely filled by the adenocarcinoma and no foci of small cell carcinoma were found while they were observed in control animals (Table I). The adenocarcinoma in TAg/Allo/IL-12 vaccinated mice was well differentiated even at 27 weeks and in about 20% of mice a small-cell carcinoma appeared. Vaccination and IL-12 treatment did not modify the relative frequencies of the various tumor types.
Table I. Prostate Neoplastic Areas and Epithelial Cell TAg and PCNA Expression in Control and Vaccinated TRAMP Mice
Percentage of normal gland, prostatic intraepithelial neoplasia (PIN), well or moderately differentiated adenocarcinoma (ADC) and small cell carcinoma (SCC) with respect to the total prostate area (mean ± standard error from 4 mice).
Percentage of epithelial cells expressing TAg or PCNA (mean ± standard error from 4 mice).
p < 0.05 vs. age-matched control mice (Student's t test).
The development of all the 3 tumor histotypes was associated with cell expression of TAg and PCNA. The TAg/Allo/IL-12 vaccination was followed by an increase in CD4+ cells in the stroma below the basement membrane (Fig. 2d), and CD8+ cells in both the stroma and among the neoplastic epithelial cells beyond the membrane (Figs. 2f and 2g). CD8+ cells were more frequent at 8 weeks when the papillary projections began to form, and then decreased when the adenocarcinoma almost totally filled the ducts. They were also present in the small-cell carcinoma masses of the TAg/Allo/IL-12 vaccinated mice (Fig. 2h).
Electron microscopy also disclosed an appreciable number of lymphocytes in the stroma and between the epithelial cells of the animals that received the TAg/Allo/IL-12 vaccination. Despite the close contact between these 2 sets of cells, there was no evidence of changes due to cytotoxicity, though damage and necrosis of the epithelial cells were sometimes apparent (data not shown).
We compared the cell-mediated and humoral responses in the vaccinated mice and age-matched controls. We first examined cytokine production in MLTC cultures (Fig. 3a). Total and CD4-positive spleen cells from TAg/Allo/IL-12 vaccinated mice showed IFN-γ production, which strongly increased upon in vitro restimulation with the vaccine cells (TAg/Allo) as well as with cells expressing TAg or allogeneic histocompatibility antigens separately (see in vitro restimulation with TAg/Syn or TAgneg/Allo cells, respectively). The highest IFN-γ levels (over 20 ng/ml) were produced by CD4-enriched spleen cells restimulated by syngeneic SV40-transformed cells (TAg/Syn). A low, but significantly increased, level of IFN-γ production by CD8 cells was also found. IFN-γ production by spleen cells from control TRAMP mice and from TRAMP mice treated with rIL-12 alone did not exceed 0.2 ng/ml, even after in vitro restimulation (data not shown). IL-4 production was very low or undetectable (data not shown). On the whole, TAg/Allo/IL-12 vaccination induced a very high CD4 response biased toward the Th1 pathway.
Spleen cells from TAg/Allo/IL-12 vaccinated mice showed a very low cytotoxic activity against both syngeneic and allogeneic SV40-transformed targets (below 20% at a 100:1 effector:target ratio, data not shown). In vitro culture of splenocytes for 6 days with TAg/Allo cells led to a moderate allo-sensitization of naive splenocytes and to a strong increase of cytotoxicity by TAg/Allo/IL-12 vaccinated effectors (Fig. 3b). In fact, upon in vitro restimulation with the cell vaccine, a strong cytolytic activity against TAg/Allo cells was obtained, but no significant cytotoxicity against TAg/Syn targets (Fig. 3b), showing that cytotoxic response was mainly directed against allogeneic histocompatibility antigens. In vitro restimulation with TAg/Syn cells did not lead to significant cytotoxicity against any type of target cell. These data were mirrored by the lack of induction of SV40 specific CTL effectors, as studied with H2-Kb/epitopeIV pentamers (Fig. 3c), which were on the contrary found when nontransgenic C57BL/6 mice were vaccinated and restimulated in vitro (Fig. 3d).
Sera from vaccinated and control mice were used as primary antibody in indirect immunofluorescence experiments to detect binding activity to SV40-transformed cells. TAg/Allo/IL-12 vaccination elicited a very high level of serum binding activity against TAg/Allo cells and a lower, but significant activity against TAg/Syn cells (Fig. 4a). Control TRAMP mice and TRAMP mice which received rIL-12 treatment alone did not show significant serum binding activity against any of the cell lines tested (Fig. 4a). Sera from vaccinated mice killed both allogeneic and syngeneic SV40-transformed cells in a complement-mediated cytotoxicity test (Fig. 4b), thus showing the presence of cytotoxic antibodies directed to SV40-transformed cells. Since some SV40 TAg is associated with the cell membrane and can be recognized by antibodies,33, 34 we investigated the presence of antibodies specific for large TAg. Sera from vaccinated mice were able to immunoprecipitate TAg from TAg-positive cell membrane protein enriched fractions (Fig. 4c), thus showing the presence of specific antibodies against TAg.
TRAMP mice develop urogenital tumors of different histotypes and clinical behavior. The first is an androgen-receptor-positive, well or moderately differentiated prostate adenocarcinoma that stems from an intraepithelial neoplasia in the same way as its human equivalent. This adenocarcinoma grows slowly without forming large tumor masses and is poorly metastatic. In about one-third of mice it is subsequently associated with an androgen-receptor-negative, anaplastic, very aggressive and metastasizing small-cell prostate carcinoma, frequently with neuroendrocrine differentiation, exactly similar to the human form, whereas the remaining mice later develop a phyllodes tumor of the seminal vesicles with a stromal neoplastic component. The small cell and the phyllodes tumors give rise to large masses detectable by abdominal palpation.
Vaccination with a combination of allogeneic SV40 TAg-positive cells and systemic administration of rIL-12 significantly delayed autochthonous urogenital carcinogenesis driven by SV40 TAg in TRAMP mice, determining a more than doubled median latency time of palpable tumors (from 26 weeks to 53 weeks). Such efficacy stemmed from the combination of cell vaccine expressing both TAg and allogeneic histocompatibility antigens with systemic administration of rIL-12. The cell component alone, in fact, was totally ineffective, whereas a slight increase of latency time was achieved with rIL-12 (Fig. 1), likely due to its antiangiogenic activity. In fact, increases in the vessel network occurred in early stages of tumor growth. Treatment with rIL-12 induced transient serum levels of IL-12 in the 1–2 ng/ml range, dropped within 2–3 days, with no sign of toxicity,28 and did not elicit antibodies against the murine IL-12 cytokine even in the long term schedule here performed, as evaluated by both ELISA and Western assay (data not shown).
The rationale of the combination of cells expressing the etiological agent and allogeneic histocompatibility antigens with systemic rIL-12 was provided by our previous studies of cancer immunoprevention.4 The combined vaccine, in fact, was designed to mimic the combination found highly effective in blocking mammary carcinogenesis of HER-2/neu transgenic mice.25 With data on the delayed TRAMP carcinogenesis, the efficacy of such active immunoprophylaxis has thus been assessed against tumors with a different derivation (prostate vs. mammary gland), genetic origin and antigenic target (SV40 TAg vs. HER-2/neu) and genetic background of the host (C57BL/6 vs. BALB/c mice). Even in the profoundly different TRAMP mice, the combination delayed the development and growth of all the 3 tumor types (adenocarcinoma; small cell carcinoma; phyllodes tumor), as shown by the morphological analysis and confirmed by the tumor-free survival curve. Our histological observations performed at different times after the treatments show that TAg/Allo/IL-12 vaccination slows the development of both the adenocarcinoma and the small-cell carcinoma irrespective of the differences in their aggressiveness and hormone dependence.
Some studies sought to block TRAMP prostate carcinogenesis through a variety of active immunoprophylactic approaches.16, 17, 18, 19, 20 Reduced tumor growth or increased survivals were observed with some of these vaccines, but mechanisms are not concordant. Induction of TAg-specific CTL in TRAMP mice was observed only by some authors,18 while others suggested a major role for CD4 cells,19 but humoral immunity was never addressed. Actually, antibodies were recently suggested to be major effectors in cancer prevention approaches, while CTL activity could be more relevant for immunotherapeutic setting.6
Our vaccination protocol induced IFN-γ-producing CD4 cells, which are strongly increased by in vitro restimulation with TAg. Antisera from vaccinated mice showed a high binding activity to SV40-transformed cells and can mediate cytotoxic mechanisms. SV40 TAg mainly localizes in the nucleus, but is partly associated with the membrane.33 TAg-specific antibodies were actually found in antisera from vaccinated mice.
We also found some evidence of an increased number of CD8+ lymphocytes infiltrating prostatic tissue in vivo. Electron microscopy occasionally revealed neoplastic epithelial cell damage and necrosis close to lymphocytes, suggesting that these cells can perform some kind of cytotoxicity with a mechanism at the moment unknown because the absence of FcR for IgG excludes the recently reported CD8+ T cells ADCC.35, 36
SV40 TAg oncoprotein fulfills some of the requirements suggested for cancer immunoprevention,4 such as its role in cancer initiation, and was found to be targeted by the humoral response induced by our vaccination. However, it is likely that antisera also contain antibodies directed towards other molecules related to the SV40 transformation process. Expression of SV40 TAg leads to a deeply altered gene expression, with over 400 genes modified,37 part of which will code for membrane molecules. MUC18 antigen, for example, is up-regulated in the TRAMP model and correlates with progression.38 The best results obtained in TRAMP cancer immunoprevention (Refs.16,19 and present study) could be related to the use of cell vaccines which ensured the stimulation of the immune system by the whole SV40 TAg-induced antigenic portfolio, leading to a delay of the carcinogenic process. These data could constitute a proof of principle on the possibility to delay prostate carcinogenic process through active immunological strategies. The translational path to clinic will necessitate of further studies, dealing first with the identification of prostate-specific “oncoantigens”6 to be targeted by specific active vaccination approaches.
S. Croci is in receipt of a fellowship from University of Bologna, Italy. A. Palladini is in receipt of the “Wanda Vanini” fellowship, CIRC, Bologna.