Combining gene and immunotherapy for prostate cancer - Nature

Prostate Cancer and Prostatic Diseases (2008) 11, 187–193 & 2008 Nature Publishing Group All rights reserved 1365-7852/08 $30.00 www.nature.com/pcan

ORIGINAL ARTICLE

Combining gene and immunotherapy for prostate cancer JG Young, NK Green1, V Mautner, PF Searle, LS Young and ND James1 Cancer Research UK Institute for Cancer Studies, University of Birmingham, Edgbaston, Birmingham, UK

The nitroreductase (NR)/CB1954 enzyme prodrug system has given promising results in pre-clinical studies and is currently being assessed in phase I and II clinical trials in prostate cancer. Enhanced cell killing by apparent immune-mediated mechanisms has been shown in pancreatic and colorectal cancer models, by co-expressing murine granulocyte macrophage colony-stimulating factor (GM-CSF) with NR in a single replication deficient adenoviral vector. This consists of the CMV immediate early promotor driving expression of NR, with an internal ribosome entry site (IRES) and the gene for murine GM-CSF (mGM-CSF). To examine if similar enhancement of tumour cell killing could be produced in prostate cancer, the TRAMP model was chosen. Results illustrate that the combination of suicide gene therapy using NR and CB1954, with cytokine stimulation with mGM-CSF gives an improved response compared with either modality alone. The mechanism of this improved response is however likely to be non-immune based as it lacks a memory effect. Prostate Cancer and Prostatic Diseases (2008) 11, 187–193; doi:10.1038/sj.pcan.4501008; published online 28 August 2007

Keywords: adenovirus; gene therapy; GM-CSF; nitroreductase; VDEPT

Introduction Virus dependent enzyme prodrug therapy is an attractive candidate for future localized therapy of prostate cancer. The prodrug 5-(aziridin-1-yl)-2, 4 dinitrobenzamide or CB1954 are converted by the gene Escherichia coli nitroreductase (NR) into 2- and 4-hydroxylamino derivatives. The 4-hydroxylamino derivative is subsequently converted by cellular thioesters into a potent cytotoxic bifunctional alkylating agent capable of crosslinking DNA.1 Activated CB1954 acts as an alkylating agent and appears to be cell-cycle independent.2 This is a significant potential advantage, particularly in slower growing tumours such as prostate cancer, where a minority of cells is likely to be dividing at any one time. The principal disadvantages of this system are the relative toxicity of CB1954 in humans,3 although it is extremely well tolerated via the intraperitoneal route, and the relatively poor catalytic efficiency of NR for CB1954. This system has been assessed in vitro and in vivo in both retroviral4 and adenoviral vectors.5 Clinical trials of NR expressing adenoviral vectors with CB1954 have been carried out in liver,6 head and neck and prostate cancer. Whilst this system might offer effective treatment Correspondence: Professor ND James, Cancer Research UK Institute for Cancer Studies, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK. E-mail: [email protected] 1 Current address: Hybrid systems Ltd, Oxford BioBusiness Centre, Littlemore Park Littlemore, Oxford, OX4 4SS, UK. Received 15 March 2007; revised 12 June 2007; accepted 25 June 2007; published online 28 August 2007

for localized tumours, the prospect of utilizing the body’s own immune response to enhance local and add systemic response to cancer cells is obviously attractive. It could provide specificity not hitherto achieved in cancer therapy as well as possible life-long surveillance for recurrence. Many approaches have been tried to elicit such responses in prostate cancer, including the use of BCG,7 tumour vaccines,8 ex vivo dendritic cell-based approaches,9 bi-specific antibodies10 and GM-CSF.11 Autoimmunity is a hallmark of successful breakdown of immune tolerance necessary for cancer immunotherapy,12 and this would be unlikely to be a significant clinical problem in prostate cancer, compared with most solid tumours, as the prostate is inessential after reproduction. GM-CSF plays a key role in promoting anti-tumour immune responses,13 and indeed reports comparing the ability of different cytokines to enhance the immunogenicity of murine tumour cells have shown that GMCSF secreting tumour cells stimulated potent, specific and long-lasting anti-tumour immunity.14,15 A recombinant fusion protein containing prostatic acid phosphatase and GM-CSF has been used successfully to prime autologous antigen presenting cells.16 There is increasing evidence that cellular stress responses, such as may be induced by viral cell infection, are necessary to elicit an immune response.17–21 Furthermore, adenoviral vectors encoding bicistronic HSV-TK and GM-CSF were effective in generating an anti-tumour immune response.18,23 Co-expression of high levels of the stress response protein heat protein 70 enhanced antitumour effects in vivo.24,25 Previous data from work in our group suggested that cell killing by CB1954 provoked a systemic immune response, in a mouse mesothelioma model: in AB22 mice bearing murine

Therapy for prostate cancer JG Young et al

188

mesothelioma tumours that stably express NR, some tumours failed to grow after treatment with CB1954, and the surviving animals were protected from a subsequent rechallenge with unmodified AB22 cells.26 To examine these effects further and their potential enhancement by cytokines, an adenovirus vector containing CMV-NR, a picornavirus internal ribosome entry site (IRES) and the gene for murine GM-CSF (mGM-CSF) has been constructed.26 This vector has greater therapeutic efficacy in the MC26 murine colorectal tumour model, compared with viruses expressing only one of the transgenes. Historically, prostate cancer was not thought to be responsive to immunotherapy, although more recently many immunotherapeutic approaches have been attempted, including the use of cytokines to induce antitumour responses. Critical examination of the literature suggests evidence that GM-CSF may have a direct effect against prostate cancer rather than necessarily an immune-mediated effect.27 PSA responses to administration were more rapid than might be expected from a generated immune response. Furthermore, IL-12, but not GMCSF was found to enhance killing of TRAMP C2 cells using an oncolytic herpes simplex virus to deliver transgene,28 so the position of GMCSF as a key immunebased molecule in prostate cancer was far from clear. The role of GMCSF in a VDEPT model was therefore examined to find if additional anti-tumour effects could be generated. If GMCSF increased cancer cell killing it was also necessary to see if long-term anti-tumour immune memory was generated, or if GMCSF acted via other means. As an in vivo model, the TRAMP cell lines C1, C2 and C3 are well characterized.29–31 TRAMP is a transgenic line of C57BL/6 mice that develop histological prostatic intraepithelial neoplasia by 8–12 weeks of age and progress to invasive adenocarcinoma with metastasis by 24–30 weeks of age. The TRAMP model was established using the 426/ þ 28 rat probasin promoter to target expression of the simian virus 40 large T antigen to prostatic epithelium. From this model, three cell lines were established. These were designated TRAMP-C1, C2 and C3. Cell lines C1 and C2 are tumourigenic when grafted into syngeneic C57BL/6 hosts. This offers an immunocompetent model of prostate cancer that facilitates therapeutic access, necessary in VDEPT models. In this model humane end points are reached around 54 days after seeding in controls. Injection of established subcutaneous C2 tumours could not be used, as tumours were highly variable in size and form when large enough for injection reached humane end points too rapidly to establish therapeutic efficacy; therefore an ex vivo method was used to infect uniform numbers of cells at constant multiplicity of infection. The subsequent development of tumours was then used to study the effects of NR/CB1954 and mouse GMCSF in an immunocompetent prostate cancer model.

Materials and methods Cell culture TRAMP cell lines were obtained from Baylor College of Medicine, Texas Medical Centre, Houston, TX, USA. A total of 911 cells used for virus propagation were a gift Prostate Cancer and Prostatic Diseases

from Prof R Hoeben, Leiden University, The Netherlands. Cells were grown in their recommended media. CB1954 was obtained from ML Laboratories (currently Innovata PLC, Nottingham, UK).

Generation of replication defective adenoviral vectors Replication defective E1A deleted adenoviruses encoding the transgenes NR, mGM-CSF and NR/mGM-CSF were constructed as previously described.20 Recombinant replication defective viruses were subjected to two rounds of plaque purification and subjected to four rounds of amplification on 911 cells (p4 stocks). The titre of viral stocks was determined by plaque assay on 911 cells under agar.

Immunocytochemistry for nitroreductase TRAMP C2 cells were infected in vitro with the bicistronic suicide gene/immunotherapy vector Ad5CMV-NR/mGM-CSF to assess expression levels of the therapeutic transgene NR. A total of 50 000 cells per well seeded in 24-well plates were infected at varying multiplicities of infection (MOI) and the cells fixed at 48 h post-infection with 3% formaldehyde. This time point, just before peak transgene expression, was chosen as cells tended to become over confluent by 72 h postinfection, and have poor adherence to culture media during the process of immunocytochemistry. Immunocytochemistry was then performed for NR expression using a polyclonal sheep anti-NR antibody with an HRPconjugated rabbit anti-sheep secondary antibody using the Vecta ABC Horseradish peroxidase staining kit.

Cytotoxicity assays For assessing CB1954-mediated toxicity, approximately 5–10 000 cells per well were seeded into 96-well plates in a volume of 200 ml per well, before infection with the appropriate adenoviral gene therapy vector at varying MOI. After 72 h of infection, prodrug CB1954 was added in varying concentrations for a further 72 h until 40 ml CellTiter 96 reagent (Promega UK, Southampton, UK) was added, the plate incubated at 37 1C for a further hour and read at 570 nm in a Wallac Victor2 1420 Multilabel Counter and non-specific absorbance was subtracted from the values obtained by also reading optical density at 700 nm.

Animal models Syngeneic murine tumours were grown in male C57/Bl6 mice obtained from the Biomedical Services Unit, University of Birmingham. After implantation of 5 106 cells, mice were maintained in microisolator cages equipped with filter tops. All mice were pathogen free and weighed over 20 g at the commencement of experimental procedures. Procedures were in accordance with the appropriate national guidelines and were performed with the authority of Home Office Project and Personal Licences.


Statistical analysis Statistical analysis was performed using the GraphPad Prism 3 package; survival data was presented as Kaplan– Meier curves and the differences between curves were analysed using the Mantel–Haenszel log rank test.

Results Infectivity and expression of NR in TRAMP C2 cells infected with the bicistronic vector Ad5-CMV-NR/ mGMCSF in vitro Preliminary results had confirmed infectivity of the murine TRAMP C2 cells with human adenoviruses: fluorescence-activated cell sorting analysis of TRAMP cells infected with a replication defective Ad5 expressing green fluorescent protein under the control of a CMVIE promoter demonstrated an ID50 of approximately 50 plaque forming units (PFU) per cell, compared with around 10 PFU per cell for human prostate cancer cell lines such as LNCaP (data not shown). This apparently reduced infectivity may however be artifactual due to the rapid multiplication of TRAMP cells in vitro: during the

MOI 0

MOI 50

3-day experiment TRAMP cells multiply around 8-fold compared with much slower replication in vitro of androgen sensitive human prostate cancer cell lines such as LNCaP or MDA PCa 2a and 2b. Infectivity and expression of both NR and mGM-CSF were further assessed using Ad5-CMV-NR/mGM-CSF infecting TRAMP C2 cells in vitro. Cells were infected at varying MOI from 0 to 300 PFU per cell (Figure 1) and expression levels assessed by immunocytochemistry 48 h post-infection. As expected, an increasing number of cells stained positively for NR expression with increasing MOI.

189

Expression of mGM-CSF from TRAMP C2 cells infected with the bicistronic vector Ad5-CMV-NR/mGMCSF in vitro TRAMP C2 cells were infected in vitro with the monocistronic vector Ad5-CMV-mGMCSF and the bicistronic vector Ad5-CMV-NR/mGMCSF at varying MOI up to 100. Supernatants were taken at 48 h post-infection, serial dilutions performed and mGM-CSF assayed (Figure 2). The monocistronic vector Ad5-CMV-mGMCSF produced approximately 5-fold greater mGM-CSF per ml of supernatant than the bicistronic vector Ad5CMV-NR/mGM-CSF. This has been observed previously

700000 mGM -CSF pg /ml

Treatment of tumours infected ex vivo with adenoviruses expressing NR, mGM-CSF and the combination of NR and mGM-CSF TRAMP C2 cells were infected (or mock infected with vehicle only) in vitro with Ad5-CMV-NR, Ad5-CMVmGM-CSF or Ad5-CMV-NR/mGM-CSF at a MOI of 100 PFU/cell. Cells were harvested 24 h after viral infection, washed in PBS and resuspended at a concentration of 2.5 107 cells per ml. Mice were lightly anaesthetized using Halothane, fur shaved from the area of implantation (so that tumours could be visualized easily), and 5 106 cells (in 200 ml) were implanted subcutaneously into the right flank. Mice were treated with CB1954 (20 mg kg1) or vehicle (PBS) by intraperitoneal injection on the subsequent 4 days. Tumours were visible approximately 4–6 weeks after implantation; animals were monitored daily and tumour sizes measured weekly using Vernier calipers. In accordance with national recommendations, animals were culled when tumour sizes reached 12 mm in diameter (which was significantly less than 10% of total body mass), if there were signs of imminent tumour ulceration, or if animals showed any signs of illness or lost 415% of their starting body weight.

600000 500000 400000 300000 200000 100000 0 0

50 MOI

100

Figure 2 Comparison of murine GM-CSF (mGM-CSF) produced by infections of TRAMP C2 cells infected with Ad5-CMV-NR/ mGM-CSF and Ad5-CMV-mGM-CSF. A total of 50 000 TRAMP C1 and C2 cells in 24-well plates were infected at varying multiplicities of infection (MOI) and the supernatants (1 ml) collected at 48 h postinfection. Supernatants were diluted and assayed using the Quantikine mGM-CSF ELISA assay. Values shown represent the mean71 s.d. Black shaded bars represent Ad5-CMV-NR/ mGMCSF; white bars represent Ad5-CMV-mGMCSF.

MOI 100

MOI 300

Figure 1 Immunocytochemistry for NR in TRAMP-C2 cells. A total of 50 000 TRAMP C2 cells were infected with the bicistronic vector Ad5CMV-NR/mGM-CSF at varying multiplicities of infection (MOI). At 48 h, post-infection cells were fixed with formaldehyde and immunocytochemistry for NR performed using a sheep anti-NR antibody with a rabbit anti-sheep secondary antibody, followed by staining with HRP and substrate. Prostate Cancer and Prostatic Diseases


Enhancement of toxicity to CB1954 by NR encoding adenoviruses in vitro A total of 3000 TRAMP C2 cells per well were seeded in a 96-well plate and infected at varying MOI with Ad5CMV-NR or Ad5-CMV-NR/mGM-CSF. At 72 h post– infection, CB1954 was added at varying concentration. After a further 72 h, cytotoxicity was assessed with the Celltiter 96 assay. The results are shown in Figures 3a and b. There was no statistically significant difference in vitro between the Ad5-CMV-NR and the Ad5-CMV-NR/ mGM-CSF viruses in their enhancement of toxicity to CB1954, or toxicity produced by virus alone. Given that a minority of tumour cells are probably transduced to express NR even at high MOI, cell killing was remarkably effective and suggests that in this model a bystander effect is pronounced as has previously been well described in the NR/CB1954 system.33 As there is evidence that GM-CSF may have a direct effect against prostate cancer,22 a separate experiment to assess direct toxicity effects of mGM-CSF on TRAMP C2 cells in vitro was performed. TRAMP C2 cells were infected at varying MOI with Ad5-CMV-mGM-CSF, as this vector had been shown to produce approximately 5fold greater concentrations of mGM-CSF compared with Ad5-CMV-NR/mGM-CSF. A total of 3000 cells per well were seeded into a 96-well plate and infected. Six days post-infection cytotoxicity was assessed with the Celltiter 96 assay. The result is shown in Figure 4. Minimal toxicity was produced at MOIs up to 30 PFU per cell, but at 100 PFU per cell 20.5% cells produced were killed. This was similar to direct virus toxicity from previous cytotoxicity experiments with Ad5-CMV-NR and Ad5CMV-NR/mGM-CSF in cells treated with virus only (and CB1954 not added). This suggested that mGM-CSF by itself produced minimal toxicity in vitro directly on TRAMP C2 cells. Survival of C57/BL6 mice and tumour growth following ex vivo gene therapy of TRAMP C2 tumours treated with Ad5-CMV-NR/mGM-CSF, Ad5-CMV-mGM-CSF or Ad5-CMV-NR TRAMP C2 cells were infected with either Ad5-CMVNR/mGM-CSF, Ad5-CMV-mGM-CSF or Ad5-CMV-NR in vitro at an MOI of 100. A further control group of cells was uninfected. Cells were harvested after 24 h, washed, resuspended in PBS and 5 106 cells in 200 ml PBS were injected into the right flank of a shaved C57/Bl6 male mouse. Each experimental group contained eight animals. After seeding, mice in experimental groups were Prostate Cancer and Prostatic Diseases

140 120 % cell survival

to a similar extent in other murine cell lines used with these vectors in vitro.21 This difference seems to arise due to less efficient translation of the second message in the dual mRNA at the IRES compared with the first. Levels of mGM-CSF secreted from Ad5-CMV-mGM-CSF were higher than previously reported from stably transfected cell lines (13–130 ng ml1 per 106 cells per 48 h).32 Even at the lowest MOI (50) of Ad5-CMV-mGM-CSF, around 200 ng ml1 per 48 h per 106 cells was secreted into the medium, and this increased at higher MOIs (Figure 2). This may be attributed to the powerful CMV IE promoter utilized in our experiments.

100 80 60 40 20 0 0

0.1

MOI 0

1 10 50 [CB1954] µM MOI 10

100

MOI 50

MOI 100

140 120 % cell survival

190

100 80 60 40 20 0 0

MOI 0

0.1

1 10 50 [CB1954] µM MOI 10

100

MOI 50

MOI 100

Figure 3 (a) Enhancement of Toxicity to CB1954 by Ad5-CMV-NR in TRAMP C2. A total of 3000 TRAMP C2 cells were seeded per well of a 96-well plate. Cells were infected with the two adenoviruses encoding NR, and CB1954 was added at varying concentration 72 h post-infection. After a further 72 h incubation, cytotoxicity was assayed using the Celltiter 96 kit. Cytotoxicity is displayed as percentage survival compared with cells that had no exposure to either virus or CB1954. The plotted values represent the mean of triplicates71 s.d. (b) Enhancement of toxicity to CB1954 by Ad5CMV-NR/mGM-CSF in the TRAMP C1 and TRAMP C2 cell lines in vitro. A total of 3000 TRAMP C2 cells were seeded per well of a 96well plate. Cells were infected with the two adenoviruses encoding NR, and CB1954 was added at varying concentration 72 h postinfection. After a further 72 h incubation, cytotoxicity was assayed using the Celltiter 96 kit. Cytotoxicity is displayed as percentage survival compared with cells that had no exposure to either virus or CB1954. The plotted values represent the mean of triplicates71 s.d.

administered 20 mg kg1 CB1954 or vehicle (PBS) by intraperitoneal injection daily for 4 days, starting at 24 h post-tumour seeding. Figures 5a and b show the subsequent tumour growth, and the time for the tumours to reach the humane end point specified in Materials and Methods (for ease of description, this is referred to henceforth as ‘survival’). Tumour development and growth were most delayed by treatment ex vivo with NR expressing viruses followed by intraperitoneal CB1954. Only one animal developed a tumour in the Ad5-CMV-NR/mGM-CSF þ CB1954 group and this did not appear until day 63 postimplantation, compared with the vehicle þ vehicle control group, where tumours appeared from day 28. In terms of survival, differences between groups were even more marked. Survival was best in the Ad5-CMV-NR/


mean tumour size mm3

120

% cell survival

100 80 60

191

550 500 450 400 350 300 250 200 150 100 50 0 40

40

60

80

days post implantation Ad5-CMV-NR+CB

20

Veh+CB

Ad5-CMV-NR+Veh

Ad5-mGM-CSF+CB

Ad5-NR/mGM-CSF+CB

Veh+Veh

Ad5-mGM-CSF+Veh

Ad5-NR/mGM-CSF+Veh

0 100

10 30 100 Multiplicity of Infection (pfu/cell)

Figure 4 Assessment of cytotoxicity of Ad5-CMV-mGM-CSF on TRAMP C2 cells in vitro. 3000 TRAMP C2 cells per well were seeded into a 96-well plate and infected with Ad5-CMV-mGM-CSF at increasing multiplicity of infection. A total of 6 days post-infection cytotoxicity was assessed with the Celltiter 96 assay. Cell survival was calculated as a percentage of the mock-infected cells. The values illustrated represent the mean71 s.d.

% survival

0

50

0 0

25

50

75

100

125

days post implantation

mGM-CSF þ CB1954 treated group, with 7 out of 8 mice not developing tumours and therefore surviving more than the duration of the initial experiment (Po0.0001 compared with vehicle þ vehicle group). Survival of mice infected with Ad5-CMV-NR/mGMCSF þ CB1954 was superior to those treated with Ad5CMV-NR þ CB1954 (P ¼ 0.007). Enhancement of toxicity to CB1954 generated in vitro differed little between dual NR/mGM-CSF and single NR expressing vectors (Figures 3a and b). Furthermore, no evidence of direct toxicity of mGM-CSF produced from adenovirus vectors was found in vitro (Figure 4). It was initially thought therefore that the observed superiority of the dual vector in vivo might be due to enhancement of host immunity rather than direct effects of mGM-CSF on the growth of TRAMP cells. We noted the necessity of NR in addition to mGM-CSF to enhance anti-tumour effects. We speculated that this might be due to the necessity for cell killing or appropriate stress signals to help elicit an immune response. It might also have been due to reduction in tumour mass by NR/CB1954, ‘buying time’ for an antitumour response to be generated. Prostatic epithelial cells are known to exist in an environment where their growth is greatly controlled by paracrine factors, and it is likely that early cancers’ growth is still strongly influenced by locally produced growth factors.34 GMCSF might act therefore indirectly in vivo via influencing prostatic interstitial cells and their release of paracrine factors that influence the growth/apoptosis balance.

Rechallenge of long-term survivors from the Ad5-CMV-NR/mGM-CSF þ CB1954 group In order to ascertain whether the long-term survivors of the previous experiment in the Ad5-CMV-NR/mGMCSF þ CB1954 treated group had developed long-term immunity to TRAMP C2 cells they were injected on the contralateral flank with 5 106 TRAMP C2 cells. Naı¨ve,

Ad5-NR+CB1954

Veh+CB1954

Ad5-mGM-CSF+CB1954 Ad5-mGM-CSF+Veh

Veh+Veh

Ad5-NR+Veh Ad5-NR/GM+CB1954

Ad5NR/GM+Veh

Figure 5 (a) Tumour growth of ex vivo treated TRAMP C2 cells after implantation into C57/Bl6 male mice. Subcutaneous seeding of TRAMP C2 cells was performed following infection with Ad5CMV-NR/mGM-CSF, (indicated as ‘Ad5-NR/mGM-CSF’) Ad5CMV-NR, Ad5-CMV-mGM-CSF (indicated as ‘Ad5-mGM-CSF’) or vehicle (PBS). This was followed by four consecutive daily doses of 20 mg kg1 CB1954 (indicated as ‘CB’) or vehicle. Cells were allowed to grow to form tumours until end points were reached in accordance with national guidelines. Mean tumour volume is illustrated7s.e.m. Values are illustrated only until the first death within each group occurred. (b) Survival of C57/Bl6 male mice injected with ex vivo treated TRAMP C2 cells with in vivo intraperitoneal CB1954. Subcutaneous seeding of TRAMP C2 cells was performed following infection with Ad5-CMV-NR/mGM-CSF, (indicated as ‘Ad5-NR/mGM-CSF’) Ad5-CMV-NR, Ad5-CMVmGM-CSF (indicated as ‘Ad5-mGM-CSF’) or vehicle (PBS). This was followed by four consecutive daily doses of 20 mg kg1 CB1954 (indicated as ‘CB’) or vehicle. Cells were allowed to grow to form tumours until end points were reached in accordance with national guidelines. Host animal survival is illustrated.

unchallenged mice were similarly injected as a control for tumourigenicity of cells injected. This was performed on day 114 after the initial injections of TRAMP C2 cells with varying treatments. Subsequent growth of tumours and animal survival are illustrated in Figures 6a and b, respectively. There was no significant difference in growth between naı¨ve and previously challenged mice suggesting that the therapeutic effect of the dually expressing mGM-CSF and NR virus in combination with CB1954 did not confer a long-term immune benefit. Two animals in the long-term survivor group developed bilateral flank tumours very late in the experiment (day 170 post-initial injection; day 56 post-rechallenge) indicating that cells initially injected had formed very slow growing tumours, or possibly that immune mechanisms were slowly overwhelmed before tumour growth could occur. Prostate Cancer and Prostatic Diseases


192 mean tumour size mm3

400

300

200

100

0 0

10

20 30 40 days post rechallenge

50

60

Controls GM/NR treated 100

% survival

75

50

25

0 0

10

20 30 40 50 60 days since rechallenge

70

80

Control Rechallenge Ad5NR/GM+CB1954 group Figure 6 (a) Growth of tumours in C57/Bl6 male mice following rechallenge with TRAMP C2 cells. Subcutaneous seeding of TRAMP C2 cells was performed in those animals that had shown no tumour development following injection of TRAMP C2 cells previously treated with Ad5-CMV-NR/mGM-CSF and intraperitoneal CB1954. Cells were injected on the flank opposite to prior tumour injection in these animals and in control naı¨ve animals. Cells were allowed to grow to form tumours until end points were reached in accordance with national guidelines. Mean tumour volume is illustrated7s.e.m. Values are illustrated only until the first death within each group occurred. (b) Survival of C57/Bl6 male mice following rechallenge with TRAMP C2 cells. Subcutaneous seeding of TRAMP C2 cells was performed in those animals that had shown no tumour development following injection of TRAMP C2 cells previously treated with Ad5-CMV-NR/mGM-CSF and intraperitoneal CB1954. Cells were injected on the flank opposite to prior tumour injection in these animals and in control naı¨ve animals. Cells were allowed to grow to form tumours until end points were reached in accordance with national guidelines. Host animal survival is illustrated.

Similarly, there was no significant difference in median survival in the rechallenge group—69 days post-rechallenge versus 60 days in the naı¨ve group (P ¼ 0.67).

Discussion In vitro experiments in the TRAMP C2 model confirmed the cells’ infectivity with human replication deficient adenoviruses and expression of transgenes driven from a CMV IE promoter. Further in vitro experiments showed Prostate Cancer and Prostatic Diseases

enhancement of toxicity to the prodrug CB1954 following infection with NR expressing adenovirus vectors. Mouse GMCSF did not have a directly toxic effect on TRAMP C2 cells in vitro. However, in vivo, co-expression of NR and mGMCSF and administration of CB1954 was able to cure 7/8 animals of their TRAMP C2 tumours for the initial duration of the experiment (125 days). Animals treated without mGMCSF gained clinically and statistically significant tumour growth inhibition and survival benefit from ex vivo treatment of cancer cells with NR expressing vectors and administration of CB1954. This, however, generally only prolonged survival and did not prevent eventual tumour growth reaching experimental end points. In order to ascertain the mechanism of this improvement in tumour growth and survival with mGMCSF a second rechallenge experiment was performed in vivo, with long-term survivors generated from the initial experiment readministered TRAMP C2 cells on their opposite flank. No long-term tumour growth inhibition or survival advantage was conferred on these animals by their previous exposure to mGMCSF and NR expressing vectors with CB1954 and their subsequent initial survival from tumour inoculation. This does not support the proposition that mGMCSF acts via enhancement or generation of host-specific immunity to C2 tumour cells. Mouse GMCSF seems likely to act either through nonspecific inflammatory mechanisms or possibly via influencing prostatic interstitial cells’ release of paracrine factors known to influence the growth of epithelial cells. Whilst the possible therapeutic combination of NR/ CB1954 and mGM-CSF seems promising, this must be tempered by much remaining to be elucidated on this clear enhancing effect of NR/CB1954 and mGM-CSF in prostate cancer.

Acknowledgements This work was supported by the Medical Research Council, UK. We also thank Dr Chris McConkey of the University of Birmingham Clinical Trials Unit for statistical analysis of survival data.

References 1 Knox RJ, Friedlos F, Jarman M, Roberts JJ. A new cytotoxic, DNA interstrand crosslinking agent, 5-(aziridin-1-yl)-4-hydroxylamino-2-nitrobenzamide, is formed from 5-(aziridin-1-yl)-2,4-dinitrobenzamide (CB 1954) by a nitroreductase enzyme in Walker carcinoma cells. Biochem Pharmacol 1988; 37: 4661–4669. 2 Bridgewater JA, Springer CJ, Knox RJ, Minton NP, Michael NP, Collins MK et al. Expression of the bacterial nitroreductase enzyme in mammalian cells renders them selectively sensitive to killing by the prodrug CB1954. Eur J Cancer 1995; 31A: 2362–2370. 3 Chung-Faye G, Palmer D, Anderson D, Clark J, Downes M, Baddeley J et al. Virus-directed, enzyme prodrug therapy with nitroimidazole reductase: a phase I and pharmacokinetic study of its prodrug, CB1954. Clin Cancer Res 2001; 7: 2662–2668. 4 Green NK, Youngs DJ, Neoptolemos JP, Fredlos K, Knox RJ, Springer CJ et al. Sensitization of colorectal and pancreatic cancer cell lines to the prodrug 5-(aziridin-1-yl)-2,4-dinitrobenzamide (CB1954) by retroviral transduction and expression of the E. coli nitroreductase gene. Cancer Gene Ther 1997; 4: 229–238.

Therapy for prostate cancer JG Young et al 5 Djeha AH, Thomson TA, Leung H, Searle PF, Young LS, Kerr DJ et al. Combined adenovirus-mediated nitroreductase gene delivery and CB1954 treatment: a well-tolerated therapy for established solid tumors. Mol Ther 2001; 3: 233–240. 6 Palmer DH, Mautner V, Mirza D, Oliff S, Gerritsen W, van der Sijp JR et al. Virus-directed enzyme prodrug therapy: intratumoral administration of a replication-deficient adenovirus encoding nitroreductase to patients with resectable liver cancer. J Clin Oncol 2004; 22: 1546–1552. 7 Guinan PD, John T, Baumgartner G, Sundar B, Ablin RJ. Adjuvant immunotherapy (BCG) in stage D prostate cancer. Am J Clin Oncol 1982; 5: 65. 8 Hodge JW, Schlom J, Donohue SJ, Tomaszewski JE, Wheeler CW, Levine BS et al. A recombinant vaccinia virus expressing human prostate-specific antigen (PSA): safety and immunogenicity in a non-human primate. Int J Cancer 1995; 63: 231–237. 9 Tjoa B, Boynton A, Kenny G, Ragde H, Misrock SL, Murphy G. Presentation of prostate tumor antigens by dendritic cells stimulates T-cell proliferation and cytotoxicity. Prostate 1996; 28: 65–69. 10 James ND, Atherton PJ, Jones J, Howie AJ, Tchekmedyian S, Curnow RT. A phase II study of the bispecific antibody MDXH210 (anti-HER2 CD64) with GM-CSF in HER2+ advanced prostate cancer. Br J Cancer 2001; 85: 152–156. 11 Dreicer R, See WA, Klein EA. Phase II trial of GM-CSF in advanced prostate cancer. Invest New Drugs 2001; 19: 261–265. 12 Overwijk WW, Lee DS, Surman DR, Irvine KR, Touloukian CE, Chan CC et al. Vaccination with a recombinant vaccinia virus encoding a ‘self’ antigen induces autoimmune vitiligo and tumor cell destruction in mice: requirement for CD4(+) T lymphocytes. Proc Natl Acad Sci USA 1999; 96: 2982–2987. 13 Steinman RM. The dendritic cell system and its role in immunogenicity. Annu Rev Immunol 1991; 9: 271–296. 14 Dranoff G, Jaffee E, Lazenby A, Golumbek P, Levitsky H, Brose K et al. Vaccination with irradiated tumor cells engineered to secrete murine granulocyte-macrophage colony-stimulating factor stimulates potent, specific, and long-lasting anti-tumor immunity. Proc Natl Acad Sci USA 1993; 90: 3539–3543. 15 Shi FS, Weber S, Gan J, Rakhmilevich AL, Mahvi DM. Granulocyte-macrophage colony-stimulating factor (GM-CSF) secreted by cDNA-transfected tumor cells induces a more potent antitumor response than exogenous GM-CSF. Cancer Gene Ther 1999; 6: 81–88. 16 Burch PA, Croghan GA, Gastineau DA, Jones LA, Kaur JS, Kylstra JW et al. Immunotherapy (APC8015, Provenges) targeting prostatic acid phosphatase can induce durable remission of metastatic androgen-independent prostate cancer: a phase 2 trial. The Prostate 2004; 60: 197–204. 17 Todryk S, Melcher AA, Hardwick N, Linardakis E, Bateman A, Colombo MP et al. Heat shock protein 70 induced during tumor cell killing induces Th1 cytokines and targets immature dendritic cell precursors to enhance antigen uptake. J Immunol 1999; 163: 1398–1408. 18 Gough MJ, Melcher AA, Ahmed A, Crittenden MR, Riddle DSL, Linardakis E et al. Macrophages orchestrate the immune response to tumor cell death. Cancer Res 2001; 61: 7240–7247. 19 Moseley P. Stress proteins and the immune response. Immunopharmacology 2000; 48: 299–302. 20 Gallucci S, Lolkema M, Matzinger P. Natural adjuvants: endogenous activators of dendritic cells. Nat Med 1999; 5: 1249–1255.

21 Gallucci S, Matzinger P. Danger signals: SOS to the immune system. Curr Opin Immunol 2001; 13: 114–119. 22 Vile RG, Diaz RM, Castleden S, Chong H. Targeted gene therapy for cancer: herpes simplex virus thymidine kinase genemediated cell killing leads to anti-tumour immunity that can be augmented by co-expression of cytokines in the tumour cells. Biochem Soc Trans 1997; 25: 717–722. 23 Castleden SA, Chong H, Garcia-Ribas I, Melcher AA, Hutchinson G, Roberts B et al. A family of bicistronic vectors to enhance both local and systemic antitumor effects of HSVtk or cytokine expression in a murine melanoma model. Hum Gene Ther 1997; 8: 2087–2102. 24 Djeha HA, Todryk SM, Pelech S, Wrighton CJ, Irvine AS, Mountain A et al. Antitumor immune responses mediated by adenoviral GDEPT using nitroreductase/CB1954 is enhanced by high-level coexpression of heat shock protein 70. Cancer Gene Ther 2005; 12: 560–571. 25 Lipinski KS, Pelech S, Mountain A, Irvine AS, Kraaij R, Bangma CH et al. Nitroreductase-based therapy of prostate cancer, enhanced by raising expression of heat shock protein 70, acts through increased anti-tumour immunity. Cancer Immunol Immunother 2006; 5: 347–354. 26 Green NK, McNeish IA, Doshi R, Searle PF, Kerr DJ, Young LS. Immune enhancement of nitroreductase-induced cytotoxicity: studies using a bicistronic adenovirus vector. Int J Cancer 2003; 104: 104–112. 27 Small EJ, Reese DM, Um B, Whisenant S, Dixon SC, Figg WD. Therapy of advanced prostate cancer with granulocyte macrophage colony-stimulating factor. Clin Cancer Res 1999; 5: 1738–1744. 28 Varghese S, Rabkin SD, Liu R, Nielsen PG, Ipe T, Martuza RL. Enhanced therapeutic efficacy of IL-12, but not GM-CSF, expressing oncolytic herpes simplex virus for transgenic mouse derived prostate cancers. Cancer Gene Ther 2006; 13: 253–265. 29 Foster BA, Gingrich JR, Kwon ED, Madias C, Greenberg NM. Characterization of prostatic epithelial cell lines derived from transgenic adenocarcinoma of the mouse prostate (TRAMP) model. Cancer Res 1997; 57: 3325–3330. 30 Kaplan PJ, Mohan S, Cohen P, Foster BA, Greenberg NM. The insulin-like growth factor axis and prostate cancer: lessons from the transgenic adenocarcinoma of mouse prostate (TRAMP) model. Cancer Res 1999; 59: 2203–2209. 31 Kwon ED, Foster BA, Hurwitz AA, Madias C, Allison JP, Greenberg NM et al. Elimination of residual metastatic prostate cancer after surgery and adjunctive cytotoxic T lymphocyteassociated antigen 4 (CTLA-4) blockade immunotherapy. Proc Natl Acad Sci USA 1999; 96: 15074–15079. 32 Dunussi-Joannopoulos K, Dranoff G, Weinstein HJ, Ferrara JL, Bierer BE, Croop JM. Gene immunotherapy in murine acute myeloid leukemia: granulocyte–macrophage colony-stimulating factor tumor cell vaccines elicit more potent antitumor immunity compared with B7 family and other cytokine vaccines. Blood 1998; 91: 222–230. 33 Bridgewater JA, Knox RJ, Pitts JD, Collins MK, Springer CJ. The bystander effect of the nitroreductase/CB1954 enzyme/prodrug system is due to a cell-permeable metabolite. Hum Gene Ther 1997; 8: 709–717. 34 Emberton M, Mundy AR (eds). The prostate and benign prostatic hyperplasia. The Scientific Basis of Urology, 2nd edn. Taylor & Francis: London, 2004, pp 367–371.

193

Prostate Cancer and Prostatic Diseases