The human PSA (hPSA) expressing plasmid pUMFG/PSA/IRES/CD25 (hereafter referred to as phPSA) was kindly supplied by Jeffrey A Medin, Division of Experimental Therapeutics, Ontario Cancer Institute, Toronto, Canada 25 . Vaccine gene free backbone plasmid (empty vector) was generated in our laboratory for use in control groups. For in vivo vaccination, plasmid DNA was prepared using an Endotoxin free mega kit (Qiagen, West Sussex, UK). Required plasmids were confirmed by enzyme digest and running on 1% agarose gel (Sigma, Dublin, Ireland). Group of mice were also treated with firefly luciferase plasmid (pCMV- luc). The firefly luciferase gene under the control of the CMV promoter was provided by Plasmid Factory GmbH (Bielefeld, Germany). Plasmid concentrations were determined with the aid of the Nano Drop 1000 Spectrophotometer (Thermo Scientific, MA, USA).

The murine recycled prostate cancer cell line TRAMPC1 was kindly provided by RP Ciavarra 26 of Eastern Virginia Medical School, Norfolk USA. TRAMPC1 cells were stably transfected with phPSA, using Fugene (Roche, West Sussex, UK) according to manufacturer's instructions. The hPSA expressing stable transfected clone was generated following selection with 200 μg/ml of Geneticin (Invivogen, Cayla, France). The clone was then isolated and purified following three rounds of single cell dilution and designated TRAMPC1/hPSA. The expression of hPSA in TRAMPC1/hPSA was analysed by isolation of RNA and reverse transcription polymerase chain reaction (RT-PCR). For RT-PCR, first strand cDNA was synthesised using omniscript reverse transcription kit (Qiagen, West Sussex, UK). The hPSA cDNA was amplified by PCR, using Pwo polymerase (Roche, West Sussex, UK), with hPSA foreword primer (5'-GCAGCATTGA ACCAGAGGAG-3') and hPSA reverse primer (5'-CGATGGTGTC CTTGATCCAC-3'). PCR reaction conditions included 15 min of initial denaturation at 95°C followed by 32 cycles of 1 min at 94°C, 1 min at 57°C, 1 min at 72°C. The wild-type and transfected TRAMPC1 cells were grown in culture at 37°C as reported previously 17 .

Male C57 BL/6 or MF1-nu/nu mice, 6 - 8 weeks old were used in the study. The mice were obtained from Harlan Laboratories (Oxfordshire, England). The animal ethics committee of University College Cork approved all experiments. Mice were kept at a constant room temperature (22°C) with a natural day/night light cycle in a conventional animal colony. Standard laboratory food and water were provided. All mice were maintained in a pathogen free animal facility for at least 2 weeks before the experiments. Subcutaneous (s.c.) tumour inoculation and the tumour growth measurements were recorded (on average every 2 days) as reported previously 17 . A mouse was considered incurable and euthanised by cervical dislocation when the tumour diameter reached 1.5 cm. From these volumes tumour growth curves were constructed. All of the immunological data reported is representative of at least two independent studies. Each study was performed with 5 or 6 mice per group.

Male C57 BL/6 mice were randomly divided into three groups; phPSA, empty vector and untreated. Mice were anaesthetised during all treatments by intra-peritoneal (i.p.) administration of 200 μg xylazine and 2 mg ketamine. For vaccine delivery, a custom-designed applicator (Cliniporator, IGEA, Modena Italy) with two needle electrode (4 mm apart) was used. Both needles were placed through the skin central to the quadriceps muscle. The muscle was injected between electrode needles with 50 μg plasmid DNA in 50 μl sterile phosphate buffer saline (PBS). After 80 s, square-wave pulses (1200 V/cm 100 ms × 1 and 120 V/cm 20 ms, eight pulses) were administered in sequence using a custom-designed pulse generator (Cliniporator). The untreated group did not receive EP. To determine the optimum vaccination protocol, three different regimens of vaccination were tested (Figure 1).

The ability to deliver plasmid DNA into the quadriceps muscles, using this electroporation technique, was demonstrated by electroporation of pCMV-luc and detection of subsequent firefly luciferase expression. The pCMV-luc was injected i.m. 50 μl (30 mg/ml)] followed by EP. After 72 h, firefly luciferin (80 l of 30 g/l conc.) (Biosynth, Basel, Switzerland) was injected intra-peritoneal as a substrate for the luciferase enzyme. Mice were anesthetised as previously outlined. Ten minutes post luciferin injection, live anesthetised mice were imaged for 1 min using an intensified CCD camera (IVIS Imaging System, Xenogen, Caliper Life Sciences, Runcorn, England). After imaging, the mice were culled and the transfected leg was separated and imaged immediately.

Specific gene expression (hPSA) by the muscles cells post EP mediated vaccination was demonstrated by RT-PCR of the transfected muscles. After 72 hours of the vaccine delivery, mice were culled by cervical dislocation and quadriceps muscle excised. The muscle tissues were homogenised in TRI Reagent^® (Molecular Research Centre, Inc.) to isolate total RNA. DNase treated total RNA from the transfected muscle was subjected to RT-PCR amplification with hPSA specific primers at same conditions (as described earlier).

Groups of the regimen 3 treated and untreated mice were challenged either with wild TRAMPC1 or with TRAMPC1/hPSA. All mice developed tumours, these tumours were surgically excised when tumours approached 5-7 mm in major dimension. The animals were observed for another 30 days without recurrence of the tumours. Tumour free mice at that stage were re-challenged, in the opposite flank, with the same tumourogenic dose of the initially challenged tumours (wild or transfected). Any re-growth of the tumours was observed and growth kinetics recorded.

The phPSA induced activation of the immune system and production of interferon gamma (IFNγ), a prototype Th1 cytokine, was tested in vaccinated and naive mice as previously reported 17 . An indirect ELISA was performed for detection of anti-human PSA antibodies. Serum from mice vaccinated with phPSA was analysed for anti-human PSA antibodies. Blood samples from the jaw veins were collected in heparin containing vials from both phPSA treated and from naive mice. To separate plasma the samples were centrifuge for 10 min at 1000 × g within 30 min of collection. Assays were performed either immediately or sample were stored at -20°C for later use. A single sample was tested from each animal at each time point. Blood samples were collected at week 2, 4, 8, and 12 after last vaccination. Plasma from the naive male C57 BL/6 mice was used as control. The 96-well plates were coated with 1 mg/ml of human PSA antigen (Europa Bioproducts, Cambridge, UK) in PBS containing 0.05% NaN3 (PBSN) and incubated at room temperature (RT) overnight. Plates were blocked for 1 hour at 37°C by the addition of 10% rabbit serum diluted in PBSN. After washing three times with PBSN, either PSA mAb (10-P20A, 1 mg/ml) (Europa Bioproducts) in blocking buffer (0.05% Tween 20 and 0.25% BSA in PBSN), as a standard, or 50 μl of mouse plasma samples in blocking buffer, as tests (1:10,1:100, 1:1000 dilutions were used) were added to the plates and incubated overnight at RT. The plates were then blocked again by 1 hour incubation at 37°C in 10% sheep serum, and further incubated with goat anti-mouse IgG conjugated to alkaline phosphatase (Sigma) for 5 hours at RT. After incubation with substrate, (pNPP) qualitative hydrolysis of NPP was detected using a microtiter plate reader (Vmax, Molecular Devices) with a 405-nm filter. Dilutions were also made using blocking buffer to re-assay samples that were beyond linearity for the initial 1000 × dilution.

For in vitro cytotoxicity assays, splenocytes were isolated from the phPSA treated and naive mice. CTL activity against the TRAMPC1/hPSA cells was analysed using same protocols reported previously 17 27 . Results of representative experiments are given as the mean +/- standard deviation and of multiple experiments as the mean +/- standard error. The development of an immune mediated anti-tumour activity following treatment was also tested by a modified Winn assay 27 28 . Splenocytes were isolated from the immunised and from the naive mice. Groups of male C57 BL/6 mice received s.c. injections of a mixture of splenocytes (from either vaccinated or naive mice) and the TRAMPC1/hPSA. For the s.c. inoculation, the splenocytes were mixed with TRAMPC1/hPSA (5 × 10⁶) in a proportion of 50:1 in serum- free Dulbecco's Modified Eagle's Medium (Gibco, Paisley, Scotland). Tumour devel-opment after inoculation was monitored on alternate days.

CpG oligodeoxynucleotides 1826, chosen according to published data 29 30 , had the following sequence TTCAT

GACGTT

GACGTT

The primary outcome variable of the statistical analyses was the tumour volume in each mouse measured at each time point. The principal explanatory variables were the different treatment groups. Tumour volume was analysed as continuous. Treatment groups were analysed as categorical variables. At each time point, a two-sampled t-test was used to compare mean tumour volume within each treatment group. One-way ANOVA was used to compare mean time of tumour appearance in various groups. Animal survival was represented by Kaplan Meyer survival curves. A p value < 0.05 was interpreted as a significant difference. Microsoft Excel 10.0 (Microsoft) was used to manage and analyse data.

To establish the growth and effects of the naive immune reactivity against the recycled TRAMPC1 cells, s.c. tumour inoculation of the TRAMPC1 was performed in both immunocompetent (C57 BL/6) and athymic nude (MF1-nu/nu) male mice with same tumourigenic dose (5 × 10⁶/mouse). Nude mice developed tumours earlier than C57 BL/6, which grew more rapidly, resulting in decreased survival of the nude mice (data not shown). These results suggested that the presence of intact immunity in C57 BL/6 mice has some inhibitory effects on the growth of TRAMPC1 tumours, indicating that the TRAMPC1 tumour can be targeted using immune base therapies.

In vivo growth of the wild TRAMPC1 and TRAMPC1/hPSA tumours in C57 BL/6 was comparable (data not shown). This showed that the presence of the human antigen (hPSA) in the TRAMPC1 did not cause significant effects on in vivo tumour growth, validating the suitability of the TRAMPC1/hPSA model.

In vivo luciferase activity was demonstrated after EP mediated transfection. The successful transfection of the quadriceps muscle group is shown in a representative image (Figure 2a). RT-PCR analysis of the muscles treated with phPSA confirmed the successful gene expression in treated mice (Figure 2b).

Three different vaccination protocols were examined (Figure 1) using the same EP parameters. All vaccination regimens had variable degrees of inhibitory effects on the tumour growth and animal survival (Figure 3). With regimen 1, mean time of tumour appearance in phPSA group was prolonged, although without statistical significance. The mean time of the tumour appearance was 23 days in hPSA group, 18 days in empty vector and 21 days in untreated group (p = 0.07). However, the rate of tumour growth was lower in the vaccinated group, which resulted in significantly prolonged survival of the phPSA immunised mice. The mean survival in phPSA group was 53 days, 32 days in empty vector and 30.5 days in untreated group (p vs empty vector = 0.04, vs untreated = 0.031) (Figure 3a).

With regimen 2, the mean time of tumour appearance in the phPSA immunised group was 35.5 days, significantly prolonged as compared to the both control groups (p < 0.01). Additionally the tumour growth was much slower in the immunised group than in untreated (p = 0.02). Although the immunisation resulted in slower tumour growth (compared to both control groups), the difference in growth rate between phPSA and empty vector group was not statistically significant (p = 0.06). The mean survival in immunised mice was 55.6 days, 45 days in empty vector and 42.5 days in untreated mice (p < 0.05) (Figure 3b).

Regimen 3 was found to be the most effective strategy resulting in delay in time of tumour appearance, retarded tumour growth, and prolonged survival of the tumour bearing mice. The mean time of the tumour appearance was 32.8 days in immunised group (p < 0.01). The tumour growth was also much slower as compared to both groups (p vs empty vector = 0.04, vs untreated = 0.01). These effects were translated into prolonged survival with mean survival after tumour inoculation in phPSA immunised group was 67.5 days, 45 days in empty vector and 40 days in untreated group (p < 0.01) (Figure 3c). An overall comparison of the immunised mice in all three regimens is shown in Figure 4. These data indicate the superior immunological and tumour inhibitory effects of the four-dose vaccination. No adverse effects related with repeated vaccination were observed. There were no immunisation related deaths and all mice remained healthy throughout the experimental period.

Production of anti hPSA antibodies in mice serum was determined at various time points after last vaccination. Higher levels of anti-hPSA antibodies were observed at all study time points in regimen 3 and these levels remained persistently higher for up to 12 weeks after last vaccination. Regimen 1 also resulted in production of the anti-hPSA antibodies, but a drop in the level was observed after 4 weeks (Figure 5a). After 8 weeks from the last vaccination, the assessment of antibodies was not possible in regimen 1 and 2 - as the mice were developing growing tumours and the tumour volumes required culling of the animals to comply with ethical committee guidelines. However in the regimen 3, the mean level of the anti hPSA antibodies was 67.83 at week 12 post last vaccination, indicating that there is persistent production of the antibodies. This factor may be responsible for the superior effects of the regimen 3. At week 2 and week 4 post final vaccinations, the levels of anti hPSA antibodies were not statistically different between the tested regimens (p > 0.05). However, at week 8 significant higher levels of the anti hPSA antibodies were recorded in the mice treated with regimen 3 (p vs regimen 1 = 0.01, vs regimen 2 = 0.02).

Histological analysis (H & E) of the tumours from the immunised mice showed abundant lymphocyte infiltration (data not shown). Immunisation with phPSA also resulted in higher production of IFNγ, indicative of Th1 immune activation. On comparison of the different vaccination protocols, variable amounts of the IFNγ was recorded in the treated mice. However, the levels in regimen 3 were much higher than in other protocols. Mean value of the IFNγ in the regimen 3 was 163.3 pg/ml, while 100.33 pg/ml and 78.33 pg/ml in the regimen 2 and 1 respectively (Figure 5b). The regimen 3 was superior to the others (p vs regimen 1 < 0.01, vs regimen 2 = 0.01). Regimen 2 also resulted in more IFNγ activity than regimen 1 (p = 0.03). These observations could explain the better tumour protective effects with regimen 2 and 3 than with regimen 1.

In vitro cytotoxicity was determined in stimulated splenocytes in all three regimens. Regimen 3 resulted in higher cytotoxicity (75%) than regimen 2 (60%) and regimen 1 (50%) (Figure 5c). In vivo cytotoxicity was also demonstrated by modified Winn assay. Splenocytes from vaccinated mice (regimen 3) with longest survival were harvested, mixed with TRAMPC1/hPSA and subsequently inoculated s.c. in groups of C57 BL/6 mice. Fifty percent of the mice receiving mixture of splenocytes from phPSA immunised group failed to develop tumours and importantly the tumour growth in tumour developing mice was significantly retarded (p < 0.01) (Figure 6a). These effects resulted in overall improvement in survival of these mice (Figure 6b).

After tumour rechallenge with TRAMPC1/hPSA, only 33% mice developed tumours with previous neo-adjuvant phPSA vaccination (Figure 6c). Additionally, the tumour protective effects were specific, as there was no tumour protection following rechallenge with wild TRAMPC1. This demonstrates that phPSA treatment induced a hPSA antigen-specific immune response giving resistance to the same tumour cell line, but not to a wild type (-ve for hPSA).

Use of a synthetic oligo CpG was examined aimed at promoting Th1 type immune responses, to augment potency of the phPSA vaccine. After four vaccines and adjuvant doses of the oligo CpG, 6 out of 11 mice remained tumour free for more than 100 days (cured). Hence, the adjuvant effects of the synthetic oligo CpG resulted in the complete tumour protection (relative risk reduction of 0.45%) as compared with single therapy alone. Time of tumour appearance was also prolonged in combined therapy groups (Figure 7). No significant increase in the levels of anti hPSA antibodies was observed in phPSA + Oligo CpG groups as compared to phPSA alone (data not shown).