research reports

University of York

2002/3


Yorkshire Cancer Research Cancer Research Unit
Director: Professor N.J. Maitland

Gene therapy for prostate cancer
Prof. N.J. Maitland, Dr R.M. Sharrard, Dr L Stanbridge, Ms L Hudson, Ms R Nugent, Ms J. Hall, Ms H. Rippon, Mr H. Guan, Mr R. Birnie, Mr V. Dussupt.
Collaborators: Dr P. Berthon, Hopital Urogene SA, Evry, France
Prof C .Bangma, University of Rotherdam, Rotterdam, Netherlands
Prof. P. Mannoni and Dr Claude Bagnis, University of Marseille, Marseille, France
Prof. T. Totterman and Dr Magnus Essand, University of Uppsala, Uppsala, Sweden
Dr C. Wrighton, ML Research Laboratories
Dr Pavel Pisa, Karolinska Institute, Sweden
Prof. S. Bettuzzi, University of Parma, Italy

The aims of this program are to develop targeted and safer means of delivering therapeutic genes into human prostate cancers. Funding is provided jointly from YCR and a FP5 award from the European Union, which is coordinated from York. The European dimension is now beginning to produce the added value for which it was designed, with increasing inter-laboratory visits. The program has four main elements; considerable progress has been achieved in all areas, and highlights are summarised below:

(i) Analysis of specificity patterns of individual prostate genes, incorporation of these transcriptional control sequences into viral vectors
The group now has available a set of control sequences, cloned from genes whose expression patterns are altered in the development of prostate cancer. Most have been isolated in York, although some are a direct result of the EU program. Expression patterns of both native promoters and various deletions have been studied for prostate specificity, cancer specificity, androgen regulation and basal/luminal expression by coupling the promoters to indicator genes, specifically the GFP-CAT hybrid produced in previous studies. Expression is measured by fluorescence microscopy, FACS and if necessary western blotting and CAT activity assays. Tissue-specific promoters have restricted cellular expression patterns, thus gross measurements of activity such as homogenization and extraction of cells for enzyme activity produce false results e.g. a few strongly expressing cells result in the same overall activity as many weakly expressing cells. Helen Rippon has produced a FACS analysis program to account for this, which was a particular feature of the prostatic transglutaminase (hPTG) promoter. The hPTG promoter is one of the most tightly regulated control sequences we have studied, but we have shown that the specificity does not lie in androgen sensitivity, as previously thought, but in an absolute specificity for luminal cell differentiation. Manuscripts describing this are currently submitted for publication. We have also developed two new antibodies against pHTG, which is switched off in CaP.

Secondly, we have investigated the Prostate Stem Cell Antigen control sequences. Hong-Tao Guan has shown that this is unlikely to be either prostate-specific, or indeed a stem cell antigen, as the expression patterns are not consistent with either. However, levels of expression are elevated in prostate cancers, and as it is a cell surface molecule it has considerable potential for preferential viral vector targeting. Hong-Tao has prepared the protein in microgram quantities (expressed in bacteria) and a polyclonal antibody has been generated.

Since promoters are modular in their mechanism of action, the EU program is now concentrating on generating hybrid promoters, taken from those isolated in the YCR-funded study, which retain their activity patterns in the heterogeneous mixture of cell types in prostate cancers, including both positively and negatively androgen regulated genes.

(ii) Modification of the attachment properties of the viral vectors to preferentially attack prostate tumours
To 're-educate' viral vectors to attach specifically, their attachment can be modified, either by genetic manipulation of protein(s) in the viral capsid or membrane, or by adaptation, using peptides (to bind to prostate-specific receptors), and antibodies, to coat the viral particles and confer reactivity against molecules on the surface of the tumour cells. We are testing all 3 strategies, using our bacculovirus (BV) model.

Vincent Dussupt has engineered growth factor peptides into the gp64 membrane protein of the BV and has used these to infect human cells in culture at 4C, when the virus will attach but not penetrate. After staining with an anti-GP64 antibody, the presence of virus on the cell surface is revealed. The data suggest that specific attachment is still dominated by non-specific mechanisms, probably though heparan sulphate binding. We have identified putative binding sites and are investigating means of inactivating these, both chemically and genetically.

The baculoviral vectors are particularly tractable to such modification, and can accommodate large amounts of extra genetic information. Our analysis also indicated that they are extremely safe, and do not present the same safety hazard that retroviruses and adenoviruses may represent.

(iii) Development and exploitation of pre-clinical models of prostate
Associated with the basic cell biology of prostate and prostate cancer we have continued to develop short term organ cultures of prostate cancers as a means of pre-clinically testing the penetration and specificity of viral vectors. June Hall has shown that the model retains virtually all of the gene expression patterns seen in prostate (and cancer) in vivo, as measured by immuno-histochemistry, and produces surprising results relative to monoculture of prostate epithelium in vitro, when used to assess promoter strength and cell type specificity. In particular it emphasizes the superior ability of baculoviral vectors to penetrate and infect prostate in vivo, ahead of patient trials. A manuscript describing this model is ready for publication, and several master classes have been conducted for researchers from around Europe during the year.

(iv) Therapeutic genes
This does not form a major part of the programme. Although the signalling genes under study (Dr RM Sharrard) in prostate tumours have shown considerable potential as targets, we currently favour both GDEPT and immunotherapy as therapeutic strategies. However in collaboration with Uppsala (Prof T Totterman) where there is an active bladder cancer immunotherapy program, therapeutic constructs for prostate are being generated. In collaboration with ML Research, we are introducing their patented GDEPT enzyme into our vectors for efficacy testing in prostate.

Structure and functional studies of human papillomavirus E2 proteins
Dr J.E. Burns, Ms E. Hernandez-Ramon, Ms C. Burn, Ms H. Walker, and Prof. N.J. Maitland
Collaborators: Dr A. Antson, Ms O. Moroz, Dr I. Bronstein, Prof. K. Wilson and Prof. G. Dodson, York Structural Biology Laboratory, University of York
Prof. M Wells, Dept of Pathology, University of Sheffield
Dr D. Hicks and Dr S. Bates , Royal Hallamshire Hospital, Sheffield
Dr D. Jenkins, Dept of Pathology, University of Nottingham Medical School,
Dr I. Morgan, Institute of Comparative Medicine, University of Glasgow

Human papillomaviruses (HPV) cause hyperproliferative lesions in cutaneous and mucosal epithelia. Over 100 types of human papillomavirus have been identified; We are studying two HPVs: HPV16, which is a high risk type found in over 50% of cervical tumours, and HPV2a, a low risk type causing primarily hand warts but also implicated in oral carcinomas.

E2 proteins function as the major regulators of viral transcription and replication. They are nucleophosphoproteins, which bind to specific sequences within the viral regulatory region and control expression of the viral oncogenes E6 and E7. Malignant progression correlates with loss of functional E2 expression and dysregulated expression of E6 and E7, resulting in disruption of cell cycle controls. The E2 proteins of HPVs are highly conserved and share a similar structure: an N-terminal module of around 200 amino acids connected by a linker region of variable length and sequence to a C-terminal domain of 80 - 100 residues. Transactivation and replication functions have been mapped to the N-terminal region, while the domains responsible for DNA binding and major dimerisation are in the C-terminus.

We previously identified a second dimerisation interface within the N-terminus and noted that mutants that disrupt the E2 transactivation function without affecting replication map into this region. Dimerisation may also explain how E2 is able to induce DNA looping when bound to distant E2 binding sites. We have used the 3D structure to generate mutations that we predict to prevent N-terminal dimerisation without causing gross structural alterations. These are being tested in vivo by functional assays for replication (using cooperation with cloned E1 proteins produced in previous YCR work) and transactivation of E2 controlled promoter sequences. These are complemented by in vitro studies of E2 DNA binding by electrophoretic mobility shift assays and DNA looping assays, by electron and atomic force microscopy. The mutants have also been cloned into the intact HPV genome and will be examined in organotypic culture for effects on cell growth and differentiation.

Alterations in the activation pathways of Protein Kinase B and Erk in the
progression to metastasis of human prostate cancer
Dr R.M. Sharrard, Ms J. Spalton, and Prof. N.J. Maitland
Collaborators: Dr D. Alessi, University of Dundee, Dr H Isaacs, University of York

Progression to invasiveness and metastasis in epithelial tumours is characterised by cell growth and survival independent of growth factors (GFs) and cell-cell and cell-matrix contact. These functions are modulated by Protein Kinase B (PKB/Akt), activated by phosphatidylinositol 3-kinase (PI3K), and by MAP kinases erk1/2 activated through ras, raf and MEK1/2.

Using a series of prostate-derived cell lines, we previously demonstrated that prostate tumour cells show reduced dependence of PKB/Akt activation on GFs or substrate adhesion, concomitant with enhanced sensitivity to the PI3K inhibitor LY294002. We have further investigated the molecular basis for these changes using specific inhibitors of the EGF and IGF1 receptors, src, PI3K, and MEK1/2, monitoring their effects upon PKB and erk1/2 activation.

In non-tumour cells EGFR activates PI3K/PKB via src and MEK/erk independently of src, while IGF1R activates PI3K/PKB (but not MEK/erk) via src-dependent and -independent pathways. Activation of MEK/erk downregulates PI3K/PKB; this effect includes direct interactions of MEK with PI3K pathway components independent of its erk-kinase activity.

In P4E6 well differentiated tumour cells, alterations to effectors between the GF receptors and src leads to GF-independent PKB activation.

In LNCaP metastatic cells, progressive alterations permit PI3K/PKB activation by both src and MEK, while PC3 cells have acquired src-dependent MEK/erk activation.

Prostate cancer progression thus involves complex alterations to signalling pathways regulating proliferation and survival. Elucidation of the molecular lesions and altered patterns of cross-talk between mechanisms activating PKB and erk1/2 in tumours will allow formulation of combinations of specific protein kinase inhibitors to target prostate cancers and their metastases.

Three dimensional in vitro modelling of the prostate
Dr S. Lang, Mrs K. Hyde, Dr S Bryce, Prof. N.J. Maitland
Collaborators: Dr A. Collins, Prostate Research Group, Dept. Surgery, The Medical School, University of Newcastle
Mr N. Clarke, Department of Surgery, Christie Hospital NHS Trust and Department of Urology, Withington Hospital, Manchester
Mr M. Stower, Dept. of Urology, York District Hospital, Wigginton Road, York
Prof. D. Edwards and Mr A. Riddick, Dept of Biological Sciences, University of East Anglia
Prof. A. Turner, Dr B. Usmani and Ms L. Dawson, School of Biochemistry and Molecular Biology, University of Leeds
Prof. Colin Cooper, Institute of Cancer Research, Sutton, Surrey
Prof, M. Stratton and Dr R Wooster, Sanger Centre, Cambridge

We have reported generation of an in vitro three dimensional model of the prostate, by co-culturing epithelial and stromal cells derived from normal and malignant human tissue. However these primary cultures have a limited life span and are not suitable for large scale analysis. Several of the novel immortal cells, of both normal and tumour origin have also demonstrated an ability to differentiate partially within this system, such as PC3 and several lines of E6 extended life span prostatic epithelium, such as Shmac5. When co-cultured in Matrigel, acinus-like spheroids are formed, closely resembling the prostate in vivo, and producing many 'prostate-specific' protein products. This is most remarkable in the undifferentiated PC3 tumour cell line, which has basal epithelial properties in monoculture and forms hollow spheroids with a single epithelial layer. Complete differentiation cannot be achieved however. By addition of specific growth factors to the growth media, we have also been able to induce branching morphogenesis, particularly in the primary cultures. These interactions lie at the core of metastatic development. The response remains heterogeneous, and not all epithelial cells can be induced synchronously.

The specific influence of prostatic stroma is central to the differentiation and morphogenesis, and distinct differences have been noted between normal and tumour derived (or reactive) stroma, in both biological (see Hall et al, Cancer Research, 2002) and genetic (see Macintosh et al, Cancer Research 1998) analysis in our Unit. The biological analyses have indicated specific modulation, not only of differentiation, but also of invasion and motility. The specific genes modulating this activity are currently being sought.

Our ongoing collaborations with the Universities of Leeds and East Anglia are providing further evidence of the role played by different families of secreted and cell surface proteases in modulating the stromal:epithelial interactions in both normal and malignant stroma.

We have continued our collaboration with Dr A. Collins (University of Newcastle) as part of the NCRI Centre of excellence to examine whether epithelial cells selected from primary populations with markers of stem cell lineage such rapid integrin adhesion and AC133 expression are required to generate spheroids (see Collins et al, J.Cell Science 2001). In late 2002, Dr Collins joined the group in York, integrating the stem cell studies with current projects defining patterns on gene expression in response to the in vitro differentiating stimuli. Cataloguing of global changes at both the mRNA and protein level are currently under way employing the facilities in the Technology facility in the Biology Department, and in collaboration with Prof Colin Cooper, Institute of Cancer Research, as part of the NCRI Prostate programme.

The unique prostate tumour and normal cells, derived in York and confirmed by microsatellite genotyping, are supplied to the Wellcome Trust Cancer Genome Project for mutation analysis. This International collaborative genetic analysis will complement the functional signalling studies carried out by Dr Sharrard and provides a molecular basis for specific inhibitor/therapeutic intervention.


Yorkshire Cancer Research P53 Research Laboratory
Director: Professor J. Milner

P53 is a chromatin accessibility factor for nucleotide excision repair of DNA damage.
Dr. C. Rubbi, Prof. J. Milner

One of the longest standing problems in DNA repair is how cells relax chromatin in order to make DNA lesions accessible for global nucleotide excision repair. Since chromatin has to be relaxed for efficient lesion detection, the key question is whether chromatin relaxation precedes lesion detection or vice versa. Chromatin accessibility factors have been proposed but not yet identified. Here we show that p53 acts as a chromatin accessibility factor, mediating UV-induced global chromatin relaxation. Using localised sub-nuclear UV-irradiation we demonstrate that chromatin relaxation is extended over the whole nucleus and that this process requires p53. We show that the sequence for initiation of global nucleotide excision repair is as follows: transcription-associated lesion detection; p53-mediated global chromatin relaxation; global lesion detection. The tumour suppresser p53 is crucial for genomic stability, a role partially explained by its pro-apoptotic capacity. We demonstrate here that p53 is also a fundamental component of DNA repair, playing a direct role in rectifying DNA damage.

Several regions of p53 are involved in repression of RNA polymerase III transcription.
Dr. L.J. Warnock, Prof. J. Milner

The tumour suppressor p53 has been shown to regulate RNA polymerase (pol) III transcription both in vitro and in vivo. We have characterised the regions of p53 that contribute to this effect. Repression of pol III transcription in vivo does not require residues 13-19 near the N-terminus of p53 that are highly conserved through evolution. However, amino acids 22 and 23 in the adjacent transactivation domain do contribute to the inhibition of pol III activity. Deletions within the central DNA-binding core domain (residues 102-292) of p53 can entirely abolish the repression function in these assays, despite the fact that pol III templates contain no recognised p53 binding site. Deletion or substitution within the C-terminal domain of p53 can also compromise its ability to repress pol III activity in vitro and in transfected cells. These observations reveal that repression of pol III transcription is a complex function involving multiple regions of p53 extending throughout much of the protein.

Selective silencing of viral gene expression in HPV-positive human cervical carcinoma cells treated with siRNA, a primer of RNA interference.
Dr M. Jiang, Prof. J. Milner

Selective silencing of mammalian gene expression has recently been achieved using short interfering RNA (siRNA). Synthetic siRNA targets homologous mRNA for degradation and the process is self-replicative. Here we demonstrate, for the first time, siRNA silencing of pathogenic viral gene expression in human cells. As a clinically relevant, well-characterised model we chose cervical carcinoma cells positive for human papillomavirus type 16. Over 90% human cervical cancers are positive for papillomavirus and abnormal cell proliferation is driven by co-operative effects of viral E6 and E7 genes. We have now designed siRNAs to target E6 and E7 and show selective viral gene silencing with loss of E6 mRNA and E7 mRNA respectively. Importantly, anti-viral siRNA has no discernible effect on normal cells. In contrast, E6 siRNA reduces growth and E7 siRNA induces apoptosis in viral infected human cervical carcinoma cells. Our results identify siRNA as a powerful tool (i) for studying the molecular and cellular pathology of viral-induced diseases, and (ii) for the development of novel anti-viral therapies.

P53 acetylation
Dr. J. Ford, Prof. J. Milner

Acetylation of p53 forms part of the cellular stress response; p53 is acetylated under most conditions that stabilise and activate p53 as a tumour suppressor, and it has been established that acetylation of p53 is required for efficient p53-dependent apoptosis. Acetylation of p53 is also controlled by deacetylase enzymes; class I deacetylases (histone deacetylases) and class III deacetylases (sirtuins). The human sirtuin hSIRT1 interacts with p53, and inhibition of hSIRT1 function enhances p53-dependent apoptosis. Remarkably, deacetylation of p53 is currently the only known function of hSIRT1.

We are analysing the role of acetylation in p53-dependent apoptosis, using RNA interference (RNAi). RNAi is a sequence-specific post-transcriptional silencing mechanism, which can be initiated in cultured human cells by transfection of a short double-stranded RNA (short interfering RNA; 'siRNA'). The process efficiently and specifically silences target mRNA without altering the expression of other genes and without influencing apoptotic processes. We are using RNAi to stabilise p53 acetylation by specifically silencing hSIRT1. We have achieved greater than 50% degradation of hSIRT1 mRNA, and are analysing the phenotypic consequences of hSIRT1 silencing under normal conditions of cell growth and during the cellular stress response. We aim to determine if silencing of hSIRT1 can produce a global sensitisation of cells to p53-dependent apoptosis. We are analysing the acetylation of naturally occurring mutant p53 and, using RNAi, investigating whether stabilisation of acetylated forms of mutant p53 can stimulate or rescue p53-dependent apoptosis in response to cellular stress in human cancer cells.

Selective inhibition of oncogenic functions of mutant p53
Dr. S.J. Allison, Prof. J. Milner

Some mutants of p53, in addition to losing the tumour suppressor function of wild-type p53, display oncogenic properties. To investigate further tumour promotion by mutant p53, a major focus of our work has been to search for novel molecular or biochemical markers of gain-of-function mutant p53. As a starting point for this, isogenic clones of p53+/+ and p53-/- cells were chosen to screen for differences between cells with wild-type p53 and cells that lack p53. Differences detected here indicate a role for wild-type p53 which in turn may be subverted by mutant p53 to a gain-of-function effect. We are currently investigating whether p53 has any effect on cellular histone modifications since it has recently been demonstrated that p53, essential for the maintenance of the genomic integrity of the cell, can act as a chromatin accessibility factor for DNA repair.

Silencing the expression of mdm2 and its splice variants in cancer cell lines using RNA interference.
Dr. M. Seymour, Dr. M. Jiang, Prof. J. Milner

The oncogenic potential of mdm2 has been demonstrated and hence mdm2 represents an important target in human cancers for therapeutic intervention. In order to investigate the role of mdm2 in cancer cell growth, we are developing a system whereby mdm2 mRNA can be selectively targeted for depletion using small interfering RNA (siRNA). We have now designed anti-mdm2 siRNA molecules and have successfully employed these to silence mdm2 gene expression by RNA interference. We have observed changes in cellular morphology and intercellular interactions following selective knockdown of mdm2 mRNA (and protein consequently) in a human colorectal cancer cell line. These changes occur both in a wild type p53 background and also a p53 null background, suggesting p53 independence. Future studies will determine the effects of RNA interference of mdm2 in normal and in human cancer cells in vitro.

The effects of N- and C- terminus phosphorylation state on C-terminus acetylation of p53.
Miss S.R. Raines, Dr L.J. Warnock, Prof. J. Milner

Post-translational modifications, such as phosphorylation and acetylation have been shown to play a role in regulating the activation of p53 and DNA repair but the relationship between the two processes is unknown. Using p53 constructs, which have undergone in vitro mutagenesis at specific sites, within the Baculoviral Expression Vector System we have been able to produce p53 protein which mimics the phosphorylated or dephosphorylated state at these sites. Employing polyclonal antibodies specific to acetylated epitopes at the C- terminus of p53, we are currently investigating the relationship between specific sites of p53 phosphorylation and acetylation.