research report

University of Leeds

2004/5

FACULTY OF MEDICINE AND HEALTH
Dean of the Faculty: Professor E.W. Hillhouse

The Leeds Institute of Genetics, Health and Therapeutics (LIGHT)
Director: Professor P.J. Grant

Centre for Epidemiology and Biostatistics
Head of Centre : Professor C.P. Wild

Genotoxic effects of binary exposures of carcinogens on skin cells.
Dr M.N. Routledge, Prof. E. Ingham, Prof. C.P. Wild

It is known that UV radiation in sunlight is an important risk factor in skin cancer, a very common type of cancer in the UK. We are examining the potential enhancing effects of the combined exposure of skin cells to a chemical carcinogen, benzo[a]pyrene (BP) that is a common environmental pollutant with UV. In earlier work we had shown that there was a significant enhancement of UV induced mutations by BP in a non-skin system, using two wavelengths of UV, UVB and UVC. Because 95% of sunlight is composed of UVA, the longer UV wavelength, we have included UVA in our studies on primary human skin cells. Interestingly, we have found that UVA in combination with BP has a greater enhancement of DNA damage in skin cells than UVB. However, whereas UVB on its own causes the skin cells to enter G2 arrest, allowing for damage to be repaired prior to DNA replication, in the presence of BP, UVB did not induce this cell cycle arrest. This may allow co-exposed cells to replicate when DNA damage is still present, making the induction of potentially carcinogenic mutations more likely. As UVA is not generally thought to be a major risk factor for skin cancer, our results may be significant for future prevention strategies.

Investigation of the risks of genetic damage associated with chromoendoscopy in Barrett's Oesophagus patients.
Dr R.G. Sturmey, Dr L.J. Hardie & Prof. C.P. Wild

Methylene Blue (MB) has been used in chromoendoscopy for a number of years, and although there are some reported minor side effects it was largely assumed safe for use. In the case of Barrett's Oesophagus (BE), MB-chromoendoscopy has been suggested as a particularly useful method since it is proposed to label only cells associated with BE. This may improve diagnosis of BE by increasing the accuracy and targeting of areas for biopsy. However, previous work by this group has shown that patients undergoing routine MB-chromoendoscopy display elevated levels of oxidative DNA damage in BE biopsy samples following the procedure. Free Oxygen Radicals generated by illumination of MB are thought to be responsible.

The current grant proposes to answer the following questions

1. Does a reduction in the concentration of MB used reduce DNA damage?
2. Can the light source used to illuminate the oesophagus during chromoendoscopy be modified to minimise DNA-damaging effects?
3. Does the inclusion of antioxidants reduce the amount of DNA damage.

Using cell lines we have mimicked chromoendoscopy conditions and measured DNA damage by the comet assay. We have shown that, under in vitro conditions a 10-fold reduction of MB reduces DNA damage by approximately 50%. Additionally, we have demonstrated that it is the red portion of endoscopic light that induces the DNA-damaging properties of MB, and the inclusion of N-acetyl cysteine prevents some DNA damage.

We are now investigating the effects of filtering out the red portion of endoscopic white light in an attempt to reduce DNA damage resulting from the procedure.

A UK patent application has been filed to protect this work, and a preliminary account of the work was presented as a plenary poster at the Annual Meeting of the British Society of Gastroenterology in Birmingham (April 2005).

The Institute of Molecular Medicine, Epidemiology and Cancer Research
Director: Professor P.J. Selby

Academic Unit of Molecular Medicine
Director: Professor D.T. Bonthron

Functional characterisation of human EB1 homologues.
Dr J.M. Askham and Dr E.E. Morrison

Mutations in APC are both involved in inheritable and spontaneous forms of colorectal cancer. Resulting truncated APC proteins cannot interact with EB1, a microtubule-associated protein participating in the sub-cellular targeting of APC to microtubule tips. Microtubule tip-associated APC at discrete sub-cellular locations appears to be crucial in defining cell polarity and in directed cell migration. Hence, loss of the APC/EB1 interaction might be linked to the aberrant migration of mutant colonic epithelial cells as well as to chromosomal instability, common features of this type of cancer.

Two EB1 homologues (EB3 and RP1) have been identified in human cells, but remain relatively uncharacterised. We have demonstrated that all EB proteins interact with normal but not mutant APC. Live-cell video microscopy of GFP-tagged normal and chimeric EB proteins has shown that all exhibit a dynamic localisation to polymerising microtubule tips and that this requires EB protein dimerisation. Furthermore, whilst EB1 and EB3 bind to the dynactin component p150Glued, RP1 does not. This interaction does not significantly contribute to the microtubule tip localisation of EB proteins but does mediate the EB1/3 centrosomal localisation.

Additionally we have defined in APC a second binding site for EB proteins and demonstrated that interaction at both sites is required for a functional association between APC and EB proteins.

Our data demonstrate the complexity of molecular interactions at the C-terminus of APC, all of which are lost in colorectal cancer cells. This work has important implications for the understanding of the migratory defects and chromosomal instability commonly observed in colorectal cancer cells.

Investigation of the role of TOGp in colorectal cancer.
Dr E.E. Morrison, Dr J.M. Askham and Dr M. Adams

TOGp is a microtubule-associated protein overexpressed in colorectal and hepatic carcinoma. Based on studies of orthologues in other organisms, it is thought to play a major role in promoting microtubule dynamics. We have previously identified both a centrosomal localisation and a localisation to microtubule tips using immunofluorescence. Imaging a TOGp fragment fused to GFP in living cells revealed that the centrosomal localisation is mediated by the C-terminal third of the protein. This region is known to interact with TACC proteins and we have now identified a novel interaction with gamma-tubulin. Overexpression of this fusion protein both protected microtubules against nocodazole-induced depolymerisation and inhibited microtubule regrowth after drug washout. Further analysis revealed the existence of 2 non-overlapping sequences in the TOGp C-terminal region capable of mediating centrosomal targeting. When overexpressed, neither protected microtubules against nocodazole challenge but both antagonised microtubule regrowth after drug washout. Overexpression of a TOGp N-terminal fragment disrupted the radial microtubule array in transfected cells. We next cloned the TACC domain from TACC1 and expressed it in transfected cells as a GFP fusion protein. As described previously, overexpression induced its aggregation but we also noted severe disruption of the microtubule cytoskeleton. At low expression levels in living cells the fusion protein localised to both centrosomes and growing microtubule tips, replicating the distribution of TOGp and indicating that it can associate with microtubule ends while bound to a TACC domain. Our data support the view that TOGp both promotes microtubule plus-end dynamics and organises microtubule minus ends.

Characterisation of OSC1, a novel tumour suppressor in oral squamous cell carcinoma and other cancers.
Dr S.M. Bell & Dr A.P. Jackson

Loss of heterozygosity at 8p23 is a frequent event in Oral Squamous Cell Carcinomas (OSCC) and a number of other cancer types including breast cancer. We have identified a large candidate tumour suppressor gene (OSC1) with a novel domain structure, which has homologies to adhesion proteins already implicated in carcinogenesis.

Two peptide antibodies for the N-terminus and C-terminus of OSC1 have been commercially produced, and their specificity for OSC1 confirmed. These antibodies have been used to determine the protein expression pattern of OSC1 in normal breast mucosa and tumour specimens. Reduced expression of OSC1 has been identified in 65% of breast cancers. Interestingly down regulation of OSC1 is significantly associated with reduced differentiation and lymph node metastasis. These results are now being confirmed in a larger series of breast cancer samples using tissue micro arrays (TMAs).

Short interfering RNA (siRNA) has been successfully used to investigate the biological effects of knocking down OSC1 expression in two breast cancer cell lines. In the breast cancer cell line BT20 reduced OSC1 expression resulted in a decrease in adhesion to extracellular matrix proteins. Reduced OSC1 expression had no effect on cell proliferation. The affects of down regulation of OSC1 on invasion is currently been investigated.

This study indicates that the tumour suppressor gene OSC1 is an adhesion molecule and plays a significant role in cellVcell adhesion and tumour differentiation in breast cancer.

Expression of oncogenic versions of c-myb and AML1 in in vitro differentiated mouse embryonic stem cells.
Prof. C. Bonifer, Ms D. Clarke

An important approach for advancing our understanding of how specific transcriptional regulation contributes to haemopoiesis has been to mutate transcription factor genes in mouse embryonic stem (ES) cells and to study subsequent effects on in vitro differentiation and target gene regulation. In this work we concentrated on the role of normal and aberrant versions of c-Myb and AML1. We have differentiated ES cells homozygous null for the c-myb locus and showed that definitive precursors accumulate in c-myb-/- embryoid bodies. We have also shown that c-myb has also a haploinsufficiency phenotype. We could show that the timing of erythropoiesis in c-myb knockdown ES cells is deregulated.

A translocation that fuses the AML1 to the ETO gene generates a chimaeric protein with leukaemogenic properties. Mice and ES cells, which carry the human AML1-ETO fusion protein knocked into the wild-type AML1 locus die at day E12. In order to study the effect of AML-ETO expression of haemopoiesis, we have studied the effect of AML1-ETO on target gene expression. We studied human cells carrying the translocation and could for the first time demonstrate how AML1-ETO deregulates target genes.

Activation of the GM-CSF locus in acute myeloid leukaemia.
Dr P. N. Cockerill

GM-CSF is a myeloid cell growth factor. In T cells and monocytes, the GM-CSF gene is highly inducible, and is regulated by an inducible enhancer which undergoes chromatin remodelling to form an inducible DNaseI hypersensitive site (DHS). In acute myeloid leukaemia (AML), the enhancer typically exists as a constitutive DHS, and GM-CSF is often abnormally expressed as an autocrine growth factor. This suggests that the GM-CSF locus exists in a primed state in AML.

We have studied normal inducible mechanisms controlling the GM-CSF enhancer. We found that the enhancer occupies two positioned nucleosomes that become destabilised upon activation. We demonstrated that the enhancer is activated via two essential composite NFAT/AP-1 binding sites, an Sp1 site and an AML1 site, which function in a highly cooperative manner. We demonstrated that NFAT, AP-1, Sp1 and AML1 bind in a coordinated fashion when the enhancer is activated, and that NFAT functions by increasing chromatin accessibility.

To study abnormal GM-CSF regulation in AML we purified blast cells or CD34 positive cells from 10 AML patients and measured GM-CSF expression. Unexpectedly, we found that 9 of the AML samples did not express GM-CSF, and one sample expressed a low level of GM-CSF. When these cells were assayed by in vivo footprinting, we were unable to detect the binding of transcription factors. We suggest, therefore, that the GM-CSF locus is regulated by paracrine mechanisms in AML, whereby additional factors mediate activation of a primed GM-CSF locus.

Inhibition of Helicobacter pylori induced epithelial proliferative responses: a novel approach for cancer chemoprevention.
Dr J.E. Crabtree, Prof. M.F. Dixon, D. P.A. Robinson

H.pylori has been recently classified as a category I human carcinogen playing a causative role in the development of gastric cancer. Long-term infection of an in vivo model with H.pylori results in the development of gastric cancer. We have demonstrated that H. pylori stimulates EGFR transactivation in gastric epithelial cells through a 'triple membrane passing signal' (TMPS) cascade. H. pylori induced release of EGFR ligand, HB-EGF, is dependent on metalloprotease, EGFR and Mek-1 activities. This signalling pathway is likely to lead to autocrine/paracrine signalling cascades promoting enhanced gastric epithelial cell proliferation and decreased apoptosis and ultimately neoplasia. We hypothesise that blocking this TMPS cascade will prevent H. pylori induced epithelial hyperproliferation associated with malignant transformation of gastric epithelial cells. This hypothesis is being tested in a model system using a specific EGFR-tyrosine kinase inhibitor. The effects of the inhibitor on H. pylori induced changes in gastric pathology, epithelial cell proliferation, apoptosis, activation of downstream signalling pathways and gene expression is being examined. To investigate the importance of H. pylori induced EGF receptor signalling in the model system, ADAMs genes, EGF receptor (EGFR), EGF ligands and COX2 have been characterised in the model. Long term infection with H. pylori is associated with marked increases in gastric expression of ADAM17, COX-2 and EGFR, demonstrating the importance of this signalling pathway in the model.

A novel transgenic model for analysis of Helicobacter pylori induced gastric carcinogenesis.
Dr J.E. Crabtree, Dr P.L. Coletta, Dr P.A. Robinson

Helicobacter pylori has a major aetiological role in gastric carcinogenesis. We demonstrated using a model in vivo mutagenesis system that H. felis infection induces mutations in gastric tissue DNA. Long-term infection with H. pylori in a model system results in gastric cancer and the pathological changes associated with H. pylori infection in the model are very similar to those observed in human infection. We are developing a new model mutagenesis system using the lambda/lacI transgene. This model will be used to assess the mutagenicity of H. pylori in vivo. The developed novel model will permit detailed analysis of gastric mutation frequency induced by specific H.pylori virulence factors, and the investigation of therapeutic strategies to reduce H. pylori induced genotoxicity. This model will allow analysis of the specific molecular mechanisms that regulate H. pylori induced gastric carcinogenesis in vivo.

Towards a better definition of the human haemopoietic stem cell; surface marker expression and functional activity.
Dr E.A. de Wynter and Dr G.P. Cook

The identification of human haemopoietic stem cells (HSC) has important implications for cancer treatment. For example, improved methods of HSC purification will lead to isolation of more homogeneous populations of HSC, leading to more efficient stem cell transplantation protocols, a key therapy for leukaemias. Currently, several populations have been identified with stem cell activity but it is not known to what extent these populations overlap. The aim of this project, which started in December 2004, is to determine the relationships between these populations based on their cell surface phenotype and functional activity.

Our initial experiments are focussed on a population of cells known as the side population or 'SP' cells. This population of cells is known to have potent stem cell activity in the mouse. The SP cells can be identified by flow cytometry based on their ability to efflux the fluorescent dye Hoechst 33342. We have identified this population amongst cord blood mononuclear cells. However, the SP cells are an extremely rare population. Our aim is to develop methods by which this rare population of cells can be enriched. Preliminary experiments suggest that we can achieve a thirty-fold enrichment using a simple procedure we have developed. We are currently evaluating this procedure, assessing stem cell activity in the SP enriched population and comparing these cells with stem cell populations isolated using conventional techniques.

Potassium channels as biomarkers of carcinogenesis in human colon.
Dr J.E. Linley, Dr M. Hunter and Prof. G.I. Sandle

Colorectal cancer is a major cause of death in the UK and USA. Many start as adenomas, growing and becoming dysplastic, with an increasing risk of malignant change. Dysplastic transformation is associated with depolarisation of crypt epithelial cells, indicating a change in regulation or expression of proteins responsible for maintaining the hyperpolarized membrane potential. One such group of proteins is potassium (K+) channels, which allow selective movement of potassium across the cell membrane. Three K+ channel species have been identified in human colonic crypt cells by functional and RT-PCR studies: the KCNQ1/KCNE3 complex, KCNN4 and KCNMA1.

We have measured levels of channel mRNA from human colonic cancers and adjacent macroscopically normal tissue using quantitative, real-time PCR. Preliminary data show a greater than 2-fold increase in mRNA for the basolateral K+ channels KCNQ1, KCNE3 and KCNN4 in colonic cancer tissue compared with normal tissue. In contrast, the apical K+ channel KCNMA1 was down-regulated in the cancers. Future studies will investigate whether there are similar changes in protein expression by Western blotting and immunocytochemistry. If the changes in K+ channel expression observed in colon cancer are mirrored in adenomatous polyps, then these channels may provide a biomarker of malignant potential in patients with recurrent polyps, and other groups at high risk of developing colorectal cancer. In patients with borderline dysplasia, evaluation of K+ channel expression by imunohistochemistry may be a useful new method of deciding the best management strategy.

Further defining the role of estrogen receptorb in breast cancer: using variant isoforms to predict hormone response.
Dr V. Speirs and Dr D.J. Scott

ERb variants are expressed in breast cancer, but little is known about their biological role. The aim of this project was to identify the expression and function of ERb variants in tamoxifen (TAM) resistance. The expression of ERb and EERbcx has been measured immunohistochemically in tissue microarrays encompassing TAM-sensitive and resistant breast tumours and this is currently being evaluated by a pathologist. We also developed an in vitro model of TAM-r through continuous culture of the TAM-sensitive cell line MCF-7 with 0.1mM TAM. Real time PCR analysis was used to determine the expression levels of ERa, ERa3D, ERa7D, ERb1, ERb2/cx, ERb5, ERb5D in TAM-r cells compared with TAM-sensitive cells. Although no significant difference in ERa andb and their variant isoforms was observed, subsequent microarray analysis, in collaboration with Dr Indra Poola, Howard University, Washington DC, and using Affymetix GeneChips has revealed several candidate genes that may provide an insight into the mechanism of TAM-r. We identified 131 genes which were up-regulated and 156 which were down-regulated in TAM-r cells by at least 3-fold. Real time RT-PCR analysis of a subset of these genes positively validated the microarray data. Genes that were consistently up-regulated included AIB1, CAV1, HMGCS2, SGP28, CEACAM6, EP4, COX-1, SOX9, TIMP3 and CLIC3, while IGF1R, NPY1R, EP3 and GREB1 were down-regulated. Additionally, the prostaglandin and IGF-1 signalling pathways were differentially regulated at multiple levels. These genes have significance as predictors to determine the likelihood of relapse in breast cancer patients receiving TAM.

Academic Unit of Oncology & Haematology
Head of Academic Unit: Prof. P.J. Selby

Molecular epidemiology of DNA repair and bladder cancer.
Dr A.E. Kiltie, Prof. D.T. Bishop and Mr S.C. Sak

Failure to repair DNA damage caused by chemical carcinogens in cigarette smoke and from occupational exposures could produce mutations in bladder epithelium which could trigger carcinogenesis. We have conducted a case control study of 547 bladder cancer cases and 579 control subjects to investigate the association between genotypes of two DNA repair genes, XRCC1 and XPC, and susceptibility to bladder cancer. We found no evidence that XRCC1 polymorphisms were associated with bladder cancer. Analysis of the XPC gene showed strong linkage disequilibrium (LD) between the Ala499Val variant and two single nucleotide polymorphisms (SNPs) in the 3'UTR (Ex15-177 and Ex15-184). All three polymorphisms were associated with increased bladder cancer risk (odds ratio [95% confidence intervals], 1.61[1.03-2.53], 1.82[1.12-2.97], 1.83[1.12-2.96] respectively) but the strong LD made the causative mutation difficult to determine. Further, smokers carrying the homozygote variant for Ala499Val had a three-fold increased risk of bladder cancer compared to non-smokers with the homozygote wildtype genotype (OR[95%CI], 3.16[1.78V5.62]). Also, cases carrying one or more Val alleles were likely to have more advanced disease (stage T2 and above) compared to homozygote wildtypes (p=0.03). We also confirmed LD between XPC SNPs Lys939Gln, polyAT and IVS11-6 in 1086 individuals but none of these polymorphisms was associated with bladder cancer risk.

We are currently performing functional assays to assess the effect on DNA repair efficiency of the mutation at XPC codon 499. Plasmids containing wildtype or mutated XPC cDNA will be transfected into XPC-deficient and control lymphoblastoids and repair efficiencies assessed using a host cell reactivation luciferase assay.

A study of the prevalence of Occult B-cell lymphoproliferative disorders in the healthy population and in the family members of patients.
Dr P. Hillmen, Dr A.S. Jack, Dr A.C. Rawstron, Dr M.S. Gilthorpe

The surface protein expression profile of normal and neoplastic B-cells has been studied using a panel of antibodies identified from previous reports that showed differential expression either at the mRNA level from microarray studies, or directly from protein expression analysis. Bone marrow from normal individuals, from patients with follicular lymphoma (FCL), diffuse large B-cell lymphoma, marginal zone lymphoma (MZL), mantle cell lymphoma, chronic lymphocytic leukaemia (CLL), monoclonal gammopathy of uncertain significance, myeloma, and 3 normal lymph nodes have been analysed. Protein expression levels have been analysed with dChip software. Hierarchical clustering was used to group antigens, and different clusters were identified that could provide additive separation when combined. 66 cell surface markers have been assessed on samples from 54 B-cell leukaemia or lymphoma cases. Disease specific profiles are evident for common B-cell disorders including CLL, FCL and MZL (Waldenstrom's macroglobulinaemia). All disorders, with the exception of hairy cell leukaemia, down-regulate markers which are normally expressed on circulating B-cells and which are down-regulated on normal germinal centre B-cells. Over 1000 normal individuals have been screened with antigen combinations which depend upon light chain restriction (i.e. relatively insensitive assays) and clonal non-CLL lymphoma cells are found in at least 1% of healthy adults. The newly developed specific assays are being applied to normal individuals to establish the frequency of non-CLL monoclonal B-lymphocytosis. In conclusion, the protein expression profile of neoplastic B-cells is different from that of normal circulating B-cells suggesting a common originating stage of differentiation for several apparently unrelated B-cell disorders.

FACULTY OF BIOLOGICAL SCIENCES
Dean of the Faculty: Professor A.J. Turner

School of Biochemistry and Microbiology
Acting Head of the School: Professor R.A. Killington

Use of post genomic technologies to discover new targets for colorectal cancer immunotherapy.
Dr S.E. Perry, Dr G.P. Cook, Prof. P. Quirke, Dr A. Melcher and Dr G.E. Blair

Despite significant advances in surgery, radiotherapy and chemotherapy, cure rates for colorectal cancer (CRC) remain low and novel approaches to treatment are urgently needed. Immunotherapy holds great promise for cancer therapy, if suitable targets can be identified. We screened microarray databases with the aim of selecting targets (tumour associated antigens) that were uniquely expressed or over-expressed in colon carcinoma cell lines and tissues. Several targets of interest were selected. Here we describe one such protein termed CDCP1, a Cub Domain Containing Protein 1. This protein has recently been described to be over-expressed in lung and colon cancer, and has also been associated with later stages of carcinogenesis. Its function is not clearly understood although it is thought to be involved in cellular signalling and interactions. Other well known published markers of colon cancer are also being studied, namely, CEA and Ep-Cam.

We have used flow cytometry to quantitate expression levels in a panel of human colorectal and other tumour cell lines as well as primary fibroblasts, endothelial and peripheral blood mononuclear cells. Expression was not detected in normal human cells but extensive variation in surface expression of CDCP1, CEA and Ep-Cam was noted among colorectal and gastric cell lines. In addition, very high levels of expression of CDCP1 were found in cervical carcinoma cell lines. These results were confirmed by real-time RT-PCR, Western blotting and more recently by fluorescence microscopy. We are extending these expression studies to include other primary cell types and staged colorectal tumour samples. However the high expression levels on cells derived from several tumour cell types suggest that these markers may represent a novel target for T cell-mediated immunotherapy. In order to look at the T cell responses against these antigens (CDCP1, CEA and Ep-Cam), we are currently optimising the experimental conditions by which we can transfect dendritic cells without compromising the viability of the cell. We hope to eventually use them as APC's to generate cytotoxic responses in autologous T cells.

The variations noted in CDCP1, CEA and Ep-Cam expression between different colorectal cell lines highlight the importance of targeting particular immunotherapeutic strategies to selected groups of patients.

Studies on the expression and intracellular localisation of viral and cellular oncoproteins.
Dr G.E. Blair, Dr S.C.Wright, Dr A. Whitehouse and Dr E.H. Hewitt

This project aims to provide cell biology equipment for the study of viral and cellular oncoproteins in mammalian cells. We have purchased an Amaxa nucleofection system for the high-efficiency transfer of plasmid DNAs into primary human cells and optimised transfer of plasmids encoding potential tumour antigens (CDCP1, CEA and Ep-CAM) into human dendritic cells. We have now purchased a digital microscopy system and are visualising protein expression in mammalian cells. The current projects that are using this microscopy system include the intracellular localization of viral oncoproteins (HPV16 E7, adenovirus E1A and HVS ORF73) in mammalian cells and myc-related cellular oncoproteins. Both instrument systems are improving our capacity to rapidly characterize proteins and marker gene products (mainly EGFP) in transfected and transformed cells.

Functional genomics of a proteinase gene family (M13; NEP) in human prostate cancer.
Dr. B.A. Usmani and Professor A.J. Turner

In human prostate cancer perturbations of stromal-epithelial interactions in the developing tumour can contribute to cancer invasion and metastasis via small bioactive peptides such as endothelin-1 (ET-1) exerting autocrine (epithelial) or paracrine (stromal) growth influences. ET-1 is generated from big-ET-1, by endothelin-converting enzyme-1 (ECE-1). ECE-1 has 4 recognised isoforms, ECE-1a, ECE-1b, ECE-1c and ECE-1d. All isoforms are derived from a single gene through the use of alternative promoters but each isoform has a specific cellular localisation pattern. We investigated the interaction between prostate cancer epithelial and stromal cells, both of which express ECE-1, using Matrigel invasion chambers. Cells derived from malignant and benign stroma from radical prostatectomy were also exploited. Previously we reported that stromal ECE-1 inhibition significantly reduced PC-3 and Du145 invasion and supplementation with ET-1 significantly increased PC-3 and Du145 invasion.

Recently we used siRNA to reduce ECE-1 protein expression by 70% in stromal cells with a consequent 70% reduction in PC-3 invasion through Matrigel. No significant effect was observed using a nonsense siRNA control. Further isoform observations recognised ECE-1c as the major invasion-promoting isoform whereas, ECE-1a and ECE-1b decreased PC-3 cell invasion by ~50%. Transient expression of ECE-1c in low invasive PNT1-a cells, which lack ECE-1, increased their invasion by 150%. Co-transfection of either ECE-1a or ECE-1b together with ECE-1c completely suppressed ECE-1c influence on invasion. The suppressive ability of particular ECE-1 isoforms in PC invasion is currently being investigated further.

(Primary cell cultures and novel cell lines kindly provided by Prof. N. Maitland, YCR Unit of Cancer Research, University of York.)

Transcriptional targets of mammalian Mad family proteins.
Dr. S.C.Wright

Myc and Mad family proteins are transcription factors that function as central regulators of cell proliferation, differentiation and apoptosis. Myc promotes cell cycle progression and is overexpressed in many human cancers. The activity of Myc is antagonized by the Mad family (Mad1, Mxi1, Mad3 and Mad4) proteins that are thought to function as tumour suppressors. The identification of transcriptional targets of Myc and Mad proteins is essential for the ultimate design of therapeutic strategies aimed at tumours associated with deregulation of Myc or Mad. We are using a combination of chromatin immunoprecipitation and genomic microarrays to identify genes that are bound and transcriptionally regulated by the various Mad family members in normal and tumour cells.

Regulation of the Miz-1 transcription factor.
Dr. S.C.Wright

The c-Myc proto-oncogene product normally functions to regulate cell proliferation and differentiation; the overexpression of Myc in many human tumours is associated with both an increased rate of cell cycle progression and a block in cellular differentiation. Myc may act as either a transcriptional activator or a repressor to regulate the expression of target genes; transcriptional repression by Myc is thought to be important in cell transformation and tumourigenesis. We have shown that Myc represses transcription of the differentiation-associated Mad family member, Mad4, in proliferating cells; repression of Mad4 by Myc may contribute to the perturbation of cellular differentiation in tumours associated with Myc overexpression. Myc mediates the transcriptional repression of Mad4 via its association with the zinc-finger protein, Miz-1. We have characterized the mechanism whereby Miz-1 recognises diverse target genes and have identified other proteins involved in transcriptional regulation by Miz-1; this will be relevant for the design of therapeutic strategies aimed at tumours associated with aberrant Myc expression.

Yorkshire Cancer Research Photodynamic Therapy Group
Centre for Photobiology and Photodynamic Therapy

(A multidisciplinary group of staff from the School of Biochemistry and Microbiology, the School of Biomedical Sciences, The Department of Colour Chemistry, the Research School of Medicine, Leeds General Infirmary)

A multidisciplinary approach to the development of photodynamic therapy for cancer
Prof. S. B. Brown, Prof. J. Griffiths, Dr D. I. Vernon, Dr D. J. H. Roberts, Dr F. Cairnduff, Mr P. Marks, Dr A. Bell, Dr S. Gorman.

As discussed in earlier YCR Annual Reports, some of the photodynamic therapy (PDT) drug candidates discovered in our Centre with YCR funding, are now being developed commercially by Photopharmica Ltd. Work in our Centre supported by YCR is therefore now focusing on discovery of new and improved photosensitisers and the development of new technology which has future potential for cancer treatment, using the principles of PDT.

In one such approach, methods have been developed for the production of photosensitiser V peptide conjugates for the specific targeting of tumours. Tumour neovasculature has a characteristic integrin population and a suitable photosensitiser has been linked to a peptide that specifically binds to one of these. We have been able to demonstrate that the level of conjugate associated with different cell lines parallels their integrin expression. It is hoped that targeting tumour vasculature in this way can improve the efficacy of PDT treatment and minimise damage to surrounding normal tissue.

Another new approach is known as photochemical internalisation (PCI) and the remainder of this report will focus on this emerging field. Photochemical internalisation is a photodynamic method for the efficient delivery of biomolecules to suitable intracellular targets. Many molecules such as cytotoxins, immunotoxins and genes are taken into cells via endosomes and subsequently processed in lysosomes where they are degraded. This leads to inactivation of the biomolecule and prevents delivery to its intracellular target. This new PDT approach relies on the activation of lysosomal bound photosensitisers in such a way that the lysosomes are disrupted and prevented from degrading the biomolecule, without significant effect on cell viability. This disruption releases the cytotoxin or DNA into the cytoplasm where it can relocate to the desired target. In our study we have initially concentrated on the ribosome inactivating protein, saporin, as a model system. We have screened over 20 photosensitising drugs and have selected two that have the most effective properties. In studies in vitro we have demonstrated that a combination of PDT and saporin treatment is extremely toxic to cells. Under the same conditions the PDT effect alone is minimal and there is no cytotoxicity of the saporin alone. This technique could potentially provide a light targeted release of cytotoxins, immunotoxins or genes in tumours.