research report

UNIVERSITY OF LEEDS    

2007/08

Faculty of Medicine and Health
Dean of the Faculty: Professor E.W. Hillhouse  

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



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

Barrett's epithelium (BE), the recognised metaplastic precursor for oesophageal adenocarcinoma, selectively absorbs methylene blue (MB) dye.  In some clinics, the dye is used to aid visualisation of this lesion and to target biopsy sampling to areas of dysplasia.  We have previously shown that patients undergoing MB-chromoendoscopy have increased levels of DNA damage in Barrett's epithelium following this procedure.  The aim of this study was to establish conditions to minimise DNA damage and identify genes showing altered expression during MB chromoendoscopy.  

Utilising MB exposed oesophageal cells we have demonstrated that MB has DNA damaging effects which are dose, light and time dependent.  In particular, the red light portion of endoscopic light is responsible for the majority of DNA-damaging properties of MB.  Utilising filters which exclude only the red light portion of white light, we can minimise DNA damage during MB exposure without compromising visualisation of tissue in vitro.  In addition, utilising Affymetrix microarrays we have identified genes which show modified expression in oesophageal cells following exposure to methylene blue.  These include increased expression of genes involved in DNA damage recognition and repair (BRCA1, c-jun oncogene), NOTCH signalling, cell survival (sirtuin) and inflammatory responses (IL6R).  We are currently assessing whether expression of these genes is increased in clinical specimens from patients following MB chromoendoscopy and are reduced when red light is removed.  Some of the results from these studies have been submitted for publication in Mutagenesis.  

Molecular pathology of lung cancer: environmental causes of tumour suppressor mutations  
Prof. C.P. Wild, Prof. M.C. Hollstein and Dr M.N. Routledge  

Mutations in the p53 tumour suppressor gene are common in human cancers, and mutation profiles offer an opportunity to make inferences on carcinogenic exposures. We utilise a novel mutation assay to select and identify base changes induced by mutagens in human p53 sequences, based on the fact that senescence bypass occurs readily in mouse cells, and requires only one crucial genetic step such as loss of p53 function. The procedure we use scores mutations in immortalized cultures derived from mouse fibroblasts isolated from Hupki (human p53 knock-in) mouse strain embryos (HUFs), treated with potential mutagens.  

We have demonstrated that exposure to B(a)P (a mutagenic constituent of cigarette smoke)  results in a predominance of strand-biased G to T transversions (18/36 mutations). Mutations recurred at the smokers' lung tumour hotspot codons 157 and 273, supporting the hypothesised role of B(a)P as a major mutagen present in cigarette smoke.   The preferential mutation or LOH of the codon72 polymorphic variants of p53 has been hypothesised to play a role in cancer risk and response to chemotherapy; however, a consistent picture has yet to emerge. Since both variants of this polymorphism are available in the Hupki model, this system permits further exploration of this issue. To date, a preference has been observed for mutation of the proline allele, with loss of the arginine allele (21/33 mutations).  

Further work will involve determining the mutation spectra from complex mixes such as diesel exhaust, and further elucidating the role of the codon72 polymorphism in carcinogenesis.    



The Leeds Institute of Molecular Medicine
Director: Professor T. Rabbitts  


MCPH1, a potential predictor for response to cancer chemotherapy
Dr S.M. Bell and Dr V. Speirs  

Previously we have identified MCPH1, a DNA damage response protein involved in the regulation of BRCA1, as the defective protein in one form of microcephaly.  The MCPH1 locus (8p22-p23) is frequently deleted in many tumour types and this is associated with a poor prognosis and a reduced response to chemotherapy in breast cancer.  

Chemotherapeutic agents such as taxanes and topoisomerase II poisons require a functional spindle checkpoint for the induction of apoptosis in cancer cells.  Our data indicates MCPH1 plays a role in resistance to chemotherapeutic agents through its involvement in the spindle checkpoint and apoptosis.  We have demonstrated reduced MCPH1 expression in 17 out of 54 (32%) breast cancers, particularly in higher grade tumours.  Interestingly we also identified high levels of MCPH1 expression in 8 out of 36 (22%) low grade breast cancers, which was associated with a good prognosis.  We are now studying MCPH1 expression in a larger series of breast cancer samples using tissue micro arrays to validate our initial results and relate them to patient outcome.  

We found that reduced expression of MCPH1 causes premature chromosomal condensation (PCC).  Using time-lapse imaging we have identified further mitotic defects including slower mitotic progression displaying aberrant chromosomal congression and micronuclei formation in MCPH1 deficient cells.  This mitotic phenotype suggests that loss of MCPH1 function in tumours could cause mitotic errors resulting in aneuploidy development.  

We therefore hypothesise that, while germ line defects in MCPH1 cause microcephaly, somatic defects may cause aneuploidy development and resistance to chemotherapy.  

Epigenetic priming of the GM-CSF locus in acute myeloid leukaemia.  
Prof. P.N. Cockerill  

GM-CSF is a growth factor that controls myeloid cell growth and function.  In acute myeloid leukaemia (AML) GM-CSF is often abnormally expressed as an autocrine growth factor.  The GM-CSF gene is regulated by an enhancer that undergoes chromatin remodelling in AML cells to form a DNaseI hypersensitive site (DHS) where the local nucleosome organisation has been disrupted.  We have shown that a region of the enhancer spanning 3 nucleosomes is disrupted within the DHS in AML.  This region includes several previously characterised NFAT/AP-1 sites, 2 GATA sites, a novel type of Runx-1 binding site, and previously unidentified NF-kB and Ets sites.  

To study paracrine mechanisms of transcriptional activation of the GM-CSF locus we performed in vivo footprinting of the GM-CSF enhancer in activated myeloid cells.  We found that the Ets, NF-kB and Runx-1 sites all associated with transcription factors within activated cells.  We also performed site-directed mutagenesis of sites within the enhancer and tested function in transfection assays.  We found that the GATA, NF-kB and Runx-1 sites all contributed to activity.  Interestingly, the Runx-1 site represented a novel class of binding site that has not been observed before.  This site actually has two overlapping Runx-1 binding elements on opposite strands of the DNA.  This element appears capable of binding two molecules of Runx-1, and we found that both sites are required for function.  

The role of tumour cell syndecan-1 in conferring susceptibility to natural killer (NK) cells.  
Dr G.P. Cook  

Syndecan-1 (CD138) is a proteoglycan expressed by a number of cell-types including plasma cells and their malignant counterpart, myeloma cells.  We have shown that leukaemic cells which lack syndecan-1 become much more susceptible to lysis by natural killer cells when transfected with the syndecan-1 molecule.  Furthermore, this effect is mediated, in part, by the carbohydrate component of the O-linked glycans.  Using RNA interference, we have now shown that reducing syndecan-1 expression on myeloma cells is associated with decreased susceptibility to NK cells.  However, these results also show that syndecan-1 expression is required for myeloma cell growth, this is probably due to the ability of syndecan-1 to capture essential growth factors such as IL-6.  Modest reductions in syndecan-1 allow cell survival but are not associated with reduced NK cell susceptibility.  In addition, we have shown that in other cancers, such as Ewing's sarcoma and cervical carcinoma, syndecan-1 is not essential for cell survival or susceptibility to NK cells.  Investigation of the interactions between myeloma cells and NK cells indicates that the NK cell receptors DNAM-1, NKG2D and NKp46 play a major role in determining susceptibility to killing and that the contribution of syndecan-1 to this process is dependent on the level of expression of the ligands of these receptors on the myeloma cell.  Experiments are in progress to dissect the role of these individual molecules in NK cell susceptibility of myeloma and other tumour cell types.  

Targeting tumour antigens and adjuvants to exosomes; novel delivery vehicles for tumour vaccines.  
Dr G.P. Cook and Prof. A.A. Melcher  

Exosomes are small vesicles (50-80nM in diameter) that are released by many cell types and which contain a spectrum of the proteins expressed by the cell of origin.  Exosomes have attracted a lot of attention as vehicles for vaccine delivery because they potently stimulate dendritic cells and initiate T cell responses.  Exosomes from tumour cells may carry tumour antigens and help to boost anti-tumour immunity.  However, there is no guarantee that exosomes will contain the relevant tumour antigen.  We have identified a protein that can carry other proteins into exosomes allowing the production of tailored recombinent exosomes for tumour vaccines.  We have now generated exosomes containing various tumour antigens and reporter proteins and are using these vesicles to test anti-tumour T cells responses in vitro. Two model systems are being used: a human based model using a melanoma TAA (gp100) and a mouse based model using the well-defined SIINFEKL epitope derived from ovalbumin.  In addition, we have found that exosomes have potent adjuvant activity and we are using exosomes to enhance anti-tumour responses by natural killer cells.  We are investigating the molecular basis of this effect in the hope of exploiting both the adjuvant and tumour antigen delivery capacity in immunotherapy.  

A novel transgenic model for analysis of Helicobacter pylori induced gastric carcinogenesis  
Prof. J.E. Crabtree, Mr D.A. Brooke, Dr P.L. Coletta and 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.  Significant advancement has been made toward the development of this novel transgenic model.  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.  

Prostaglandin E₂-EP4 receptor signalling and colorectal carcinogenesis  
Dr G. Hawcroft, Dr P.L. Coletta and Prof. M.A. Hull  

Prostaglandin (PG) E₂, a major product of cyclooxygenase (COX) during human colorectal carcinogenesis acts via four G protein-coupled receptors termed EP1-4.  We have previously implicated the EP4 receptor in tumorigenic behaviour of colorectal cancer (CRC) cells (Oncogene 2007; 26:3006-19).  

Following this, we have identified endogenous EP receptor (EP1-4) protein expression in vitro in a panel of human CRC cell lines including SW480, HCT116, HCA7, LoVo and LIM 1863 cells.  Functional EP4 activity has been demonstrated by immunoassay of intracellular cAMP levels showing a response to PGE₂ which was inhibited by the specific EP4 antagonist (ONO-AE3-208).  We are in the process of testing the anti-neoplastic activity of ONO-AE3-208 on a selection of our EP4 receptor-positive cell lines in assays including proliferation, apoptosis, invasion and growth on matrigel.  In addition, we are testing the ability of short hairpin RNA to inhibit expression of EP4.  

EP4 receptor expression has also been investigated in an in vivo model of the early stages of intestinal tumorigenesis by IHC.  

Increased understanding of PGE₂-EP4 receptor signalling is vital in order to target the EP4 receptor for anti-CRC therapy, thereby harnessing the anti-neoplastic activity of COX inhibition but avoiding the cardiovascular toxicity linked to the systemic inhibition of COX.  

Investigations into the role of cyclin D2  
Dr R. Sarkar and Mr D. Jayne  

Previous DNA microarray analysis has suggested that the D-type cyclins are up-regulated in colorectal cancers which display an aggressive invasive phenotype.  The current work aims to expand on the hypothesis that D-type cyclins are involved in determining the invasive behaviour, independent of any effect on cell proliferation. Several colorectal cell lines have been screened for constitutive expression of cyclins D1, D2 and D3.  Cyclins D1 and D3 were found to be ubiquitously expressed.  However, only 4 of 9 cell lines were found to express cyclin D2 by semi-quantitative and real time PCR.  This has been supported by western blot analysis.  

We have observed that cyclin D expression is reduced in tumour cells exposed to 1% hypoxia as compared to normoxia.  To determine whether this is an effect of hypoxia-inducible factor (HIF) up-regulation, we are analysing cyclin-D expression (D1, D2 and D3) in hypoxic cells following HIF1α and HIF2α transfection.  

To determine the influence of the D-type cyclins on tumour cell behaviour we are using siRNAs specific to D1, D2 and D3 to knock-down gene expression and to assess the effects on cell proliferation, invasive capacity, differentiation, angiogenesis and epithelial-mesenchymal transformation.  This has been optimised for Lovo and HRT18 cell lines, with transfection efficiency confirmed by real time PCR.  In addition, a method has been devised for transfecting the LIM1863 colorectal tumour cell line, which can exhibit either a mesenchymal or epithelial phenotype depending on culture conditions.  

Role of the ASPM protein in cancer  
Dr E.E. Morrison, Dr J. Bond and Dr S.M. Bell  

Mutation of the ASPM gene on chromosome 1q31 causes microcephaly, a syndrome characterised by defective neurogenic cell division. ASPM is a spindle pole protein with functional and structural similarities to the NuMA protein. NuMA overexpression is known to drive chromosomal instability in human cancer by promoting mulipolar mitosis. Gains and amplifications at 1q31 are common in cancer and recent evidence suggests that ASPM may also be overexpressed in many tumours. Exploiting a unique combination of reagents, expertise and resources we wish to examine whether ASPM is overexpressed at both the mRNA and protein level in cancers where gains at 1q31 are common (such as bladder, ovarian and breast), determine whether this overexpression is due to ASPM gene amplification, correlate ASPM overexpression with centrosomal abnormalities and multipolar mitoses in tumour tissue and cell lines and finally investigate whether ASPM overexpression contributes to aneuploidy in human cancers. Although this project has just started, we have found evidence that ASPM expression is variable in different cell lines and that high levels of expression correlate with mitotic abnormalities. We have also demonstrated that immunohistochemistry is a feasible method for investigating ASPM expression levels in tumour samples and have optimised the procedures for doing this in breast cancer samples.  

Transcriptional regulation of the tumour suppressor gene ZAC  
Dr E.M.A. Valleley and Prof. D.T. Bonthron  

ZAC, also known as LOT1 (lost on transformation) or PLAGL1 (pleiomorphic adenoma gene-like) is a tumour suppressor gene that has been implicated in the pathogenesis of a number of different cancers.  ZAC is an imprinted gene, and transcripts from its differentially-methylated promoter (P1) are derived solely from the paternal allele.  Down-regulation of ZAC expression occurs in tumours, and epigenetic modification of P1 has been found in ovarian cancer.  However, we recently identified a novel ZAC promoter (P2), which unlike P1, generates biallelic transcripts.  The P1 and P2 promoters are frequently concurrently active, although their relative transcription levels vary between tissues.  Peripheral blood leukocytes are distinctive, in that ZAC expression is derived almost exclusively from P2.  This provides an explanation for the observation that ZAC expression in this tissue is biallelic.  It also means that apparent "relaxation" of imprinting in tumours could occur by promoter switching in addition to epigenetic changes at P1.  Our current studies are aimed at understanding more about how ZAC functions in normal and cancerous cells.  As well as characterising the P2 promoter in more detail, we are investigating the relationship between the promoters by blocking the activity of each one separately, in vitro.  The information obtained should give us a greater insight into mechanisms that govern the behaviour of this gene in tumorigenesis.    

Report from the YCR Centenary Chair of Pathology,
Prof. P. Quirke  

YCR support has enabled us to develop our work on staging colorectal cancer.  Staging is important as it determines the risk of relapse and thus the need for early treatment with chemotherapy.  We have demonstrated the importance of routine data collection from pathologists in the Yorkshire region and the importance of such observations (see British Journal of Cancer, 97: 1393-1398).  We investigated the relative importance of the patient, surgeon and pathologist as factors in identifying patients who would benefit from adjuvant therapy (see Journal of Clinical Oncology, 25: 2573-2579) and investigated which patients should be treated with adjuvant chemotherapy in stage II disease (see Gut, 56: 1419-1425).  These articles led to a critique of current staging methods (see Lancet Oncology, 8: 651-657) which was within the top 20 articles cited in Lancet Oncology in 2007.  We also published guidance on reporting colorectal cancer (see Histopathology, 50: 103-102 and 51: 141-149).  

We have continued to increase our participation in clinical trials with approval of the new phase III studies NCRI FOXTROT (neoadjuvant therapy in colonic cancer) and NCRI ARISTOTLE (neoadjuvant radiochemotherapy in rectal cancer).  We have written the pathology protocols and will be conducting translational research in both these studies looking at response to therapy and its prediction.  ENROL (laparoscopic enhanced recovery trial) and the low rectal cancer study are continuing to recruit.  In collaboration with Cardiff University, a new phase II trial to randomise patients on the basis of their molecular defects has been approved by the MRC and the testing will be done in our laboratory in Leeds and in Cardiff.  

Work investigating the quality of surgery is ongoing but we have shown that a new operation to remove low rectal cancer is better as it enables complete excision to occur more frequently and tearing of the specimen occurs less frequently.  This is in press in Journal of Clinical Oncology.  We have also investigated the factors leading to removal of the anal sphincter and demonstrated clear links to the volume of surgery undertaken by individual surgeons as well as social class and this is in press in Gut.  

Molecular analyses are still underway in the Quasar trial with current results showing that predictive factors should be assessed in the first two years of treatment.  We have added a range of new markers of DNA double strand break repair to those already analysed.  Further block collection has been undertaken extending the number of cases available for study to 2,000.      



Faculty of Biological Sciences
Dean of the Faculty: Professor S.W. Homans  

Institute of Molecular and Cellular Biology
Director: Professor N.M. Hooper  


Assessing the role of human papillomavirus (HPV) oncoproteins in regulating cell-mediated immunity: implications for cervical cancer.  
Dr N. James, Ms J. Jarvis, Dr G.P. Cook and Prof. G.E. Blair  

Previous work from this laboratory has investigated the modulation of cell surface Major Histocompatability Complex class I (MHC class I) expression by the E7 oncoprotein from high risk Human Papillomavirus 16 (HPV16).  We utilized two complementary cell-based systems to induce or knock-down expression of the E7 protein, and determined changes to MHC class I expression.  We have already shown that siRNA-mediated inhibition of E7 in HPV16- and HPV18-transformed cells results in significant up-regulation of cell-surface MHC class I molecules. We have also shown that induction of HPV16 E7 using a tetracycline-repressor system results in a significant down-regulation of cell surface MHC class I molecules.  In addition, analysis of total cell MHC class I heavy chain protein by Western blotting has confirmed the results obtained by flow cytometry on cell-surface MHC class I. Interestingly, total cellular levels of the invariant chain of the MHC class I heterodimer, ß₂-microglobulin, appeared to be unaffected by E7 expression, suggesting a selective effect on molecules encoded in the MHC locus (since ß₂-microglobulin is expressed outside the MHC locus).  We are currently examining levels of other components of the antigen processing and presentation pathway (TAP, LMP2, LMP7 etc).  

We have also examined the functional implications of MHC class I down-regulation by HPV E7 using a co-culture system with peripheral blood Natural Killer (NK) cells.   These co-culture experiments revealed that the cell surfaces changes in MHC class I levels mediated by E7 observed have a significant effect on susceptibility to NK cells, consistent with "the missing self hypothesis".  Thus E7-knockout cells that have increased MHC class I were more resistant to NK cell lysis.  Conversely, E7-induced cells that have reduced MHC class I were less resistant to NK cells. We are now investigating NK ligands on HPV-transformed cells to elucidate the mechanism of NK interaction with these cancer cells.   

Defining the role of CUB domain containing protein 1 (CDCP1) in human cancer  
Mr D. Orchard-Webb, Ms J. Jarvis, Dr G.P. Cook and Prof. G.E. Blair  

CUB domain containing protein 1 (CDCP1), is a transmembrane glycoprotein which is up-regulated in colorectal and lung cancer tissues and cell lines, compared to their normal counterparts. The CUB (Complement protein subcomponents Clr/Cls, Urchin embryonic growth factor and Bone morphogenetic protein 1) domain is a protein module characterised by a novel immunoglobulin-like fold and has been proposed to be involved in protein:protein and protein:carbohydrate interactions. Up-regulation of CDCP1 gene expression in colorectal cancer (CRC) tissue was first established by quantitative real time RT-PCR. It was found to be as significant as that found for the extensively studied tumour-associated antigens carcinoembryonic antigen (CEA, or CEACAM5) and Ep-CAM/GA-733-2, also cell-surface glycoproteins. We were also able to describe for the first time the presence of two mRNA isoforms of CDCP1 expressed in CRC tissue and a panel of human carcinoma cell lines. Another interesting finding was that levels of CDCP1 mRNA varied between the cell lines tested and that some degree of correlation was found with that of CEA and Ep-CAM expression. This was further substantiated by analysing expression of the protein products using flow cytometry. We devised an RT-PCR assay that distinguishes isoforms 1 and 2. This showed that both isoform mRNAs were expressed in a coordinate manner and their level correlated with cell-surface CDCP1.  

We are now examining the possible function of CDCP1 in tumorigenesis. Several protein binding partners have been recently described (e.g. src family tyrosine kinases and protein kinase Cd). We have developed RNA interference methods to reduce expression of CDCP1 in colorectal tumour cells and we are studying the phenotype of such modified cells.  

Perturbation of ß-catenin signalling by the hepatitis C virus NS5A protein: implications for development of hepatocellular carcinoma  
Prof. M. Harris  

Hepatitis C virus is a significant cause of hepatocellular carcinoma.  However, the molecular mechanisms linking chronic viral infection to development of liver tumours are poorly understood.  The NS5A protein performs a significant role in this process by activating the phosphatidylinositol 3-kinase (PI3K) pathway, resulting in stabilisation of the proto-oncogene ß-catenin.  

We have previously shown that NS5A stimulates the transcriptional activity of ß-catenin in a PI3K dependent manner.  More recently we have shown that an NS5A mutant lacking an N-terminal amphipathic helix that is unable to bind cytoplasmic membranes also activates ß-catenin.  We have also shown that blockade of the PI3K-Akt pathway resulted in the degradation of both NS5A and ß-catenin, raising the possibility that NS5A might interact directly with ß-catenin and the complex was then targeted for degradation in a PI3K dependent fashion.  Co-immunoprecipitation analysis revealed that NS5A was indeed able to bind to both the p85 subunit of PI3K and ß-catenin.  Deletion analysis showed that both the N-terminal helix and the C-terminal 178 residues of NS5A were dispensable for both interactions, suggesting that the site(s) of binding lie in the central region of the protein.  

We propose that NS5A may bind to both PI3K and ß-catenin in order to activate ß-catenin signalling.  Current experiments are investigating whether NS5A acts as an adaptor protein, directly linking PI3K and ß-catenin, or whether these are two independent interactions.  

Expression profiling of the NEP family of metalloproteinases in prostate cancer  
Dr B.A. Usmani and Prof. A.J. Turner  

Cross-talk between tumour and stromal cells contributes to cancer invasion and metastasis. ET-1 exerts both an autocrine (epithelial) and a paracrine (stromal) influence on growth. ET-1 is generated from its inactive precursor big-ET-1 by endothelin-converting enzyme (ECE-1), which exists as four distinct isoforms; ECE-1a, ECE-1b, ECE-1c and ECE-1d. They differ only in their N-terminal regions and are derived from a single gene through the use of alternative promoters. This study has investigated the interaction between metastatic PC epithelial cells and stromal cells, both of which express ECE-1, using Matrigel invasion chambers. Using Matrigel invasion assays we have shown that ECE-1 expressing stromal cells from benign and malignant prostate can greatly increase invasive potential and that specific inhibition of endogenous ECE-1 activity in these stromal cells can significantly reduce PC-3 invasion. Investigation into the role of the ECE-1 isoforms on cell invasion shows significant correlations between the presence and absence of ECE-1c and/ or ECE-1a in either stromal and/or epithelial compartments.  ECE-1c increased invasion by 20% whereas ECE-1a decreased invasion by approx 50%. Overall, the data from this project support the hypothesis that the ECE-1 isoforms may be relevant therapeutic targets for prostate and other cancers and warrants further investigation. Of the seven NEP-family members in the human genome, tissue microarray (TMA) results have suggested that ECE and its related metalloproteinase NEP are the most useful indicators of tumour progression. Evidence has now accumulated that the key diagnostic factor in this regard is the NEP:ECE ratio and future studies will be pursuing this discovery and extending its applicability to other tumour types.  

Invasion promoting and suppressing properties of ECE-1 isoforms in prostate cancer  
Dr B.A. Usmani and Prof. A.J. Turner  

The specific aims of the project are: (a) to determine the influence of ECE-1 isoform expression on invasive behaviour using stromal/epithelial co-culture experiments in vitro and ex vivo; (b) to modulate invasion using a specific inhibitor of ECE-1 and siRNA molecules targeted to ECE-1; (c) to evaluate the importance of ECE-1 isoform localization on invasion; and (d) to construct and deliver siRNA vectors.  

We have explored the influence of each of the ECE-1 isoforms on PC-3 cell invasion through Matrigel. Only ECE-1a and ECE-1c significantly altered invasion. Expression of the ECE-1c isoform transformed low-invasive cells to cells with a highly invasive phenotype.  Transient expression of ECE-1a, which is normally absent from PC-3 and stromal cells, suppressed the invasion of PC-3 cells.  We are currently trying to identify the diffusible element influencing invasion. 3D models of invasion incorporating PC cells in semi-solid Matrigel have been established.  ECE-1 isoforms have been transiently expressed in PC-3 cells and the spheroid formation observed and quantified.  We have developed techniques and methods for immunofluorescence studies of spheroids in the Matrigel and are currently trying to determine the localisation of ECE-1 and its isoforms within the spheroids.  Invasion of ECE-1 was significantly reduced using siRNA duplexes targeted to stromal or epithelial ECE-1.  Known targeting signals within the N-terminal regions of ECE-1b and 1c isoforms have been mutated using designed primers to alter their localisation and the subsequent impact on cell invasion has been assessed.  Finally, we have successfully cloned a synthetic 64nt short hairpin RNA (shRNA) targeted to ECE-1 into a herpesvirus saimiri (HVS)-based vector through a three-stage protocol.  The HVS-BAC-siRNA virus reduced endogenous ECE-1 transcription and expression in a range of prostate cancer cells and decreased the invasion of PC-3 cells through Matrigel.   

Cellular genes involved in Kaposi's sarcoma associated reactivation and lytic replication  
Dr A. Whitehouse, Dr J. Boyne, Dr S. Harrison and Miss F. Gould  

The etiological agent of Kaposi's sarcoma, Kaposi's sarcoma associated herpesvirus virus (KSHV), is the most recently identified human tumour virus.  Like other herpesviruses, KSHV has two distinct forms of infection, latent persistence and lytic replication. Eeactivation and lytic replication plays an important part in the pathogenesis and spread of KSHV infection.  

The KSHV Rta protein is the key gene product which regulates viral lytic gene expression. Analysis of the KSHV Rta protein has shown that it is capable of autoregulating its own expression and has the ability to activate transcription of various viral and cellular genes.  It activates transcription directly following interactions with promoters containing a specific sequence motif, or indirectly by interactions with cellular transcription factors.  

We are currently investigating the role of two cellular proteins which interact with Rta.  Hey-1 is a transcriptional repressor protein which interacts with KSHV Rta.  The interaction is a particular intriguing one, as Hey-1 represses KSHV lytic promoters, therefore why would Rta interact with a transcriptional repressor protein?  However, we have demonstrated that KSHV Rta acts as an ubiquitin E3 ligase, targeting Hey-1 for proteasomal degradation to allow the initiation and completion of the lytic replication cycle.  The second cellular protein KSHV Rta interacts with is HMG-A.  We have shown HMG-A can synergistically enhance KSHV Rta transactivation of lytic promoters, by directly binding to these lytic promoters and enhancing KSHV Rta binding to these promoters.  These results demonstrate that the recruitment of cellular genes by KSHV Rta are essential for KSHV reactivation and lytic gene expression.  

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

Myc and Mad family proteins are transcription factors that normally regulate cell proliferation, differentiation and apoptosis.  Myc promotes cell cycle progression; the overexpression of Myc in many human tumours results in uncontrolled cell division and a loss of differentiation.  In contrast, the Mad family protein, Mnt, normally acts as an antagonist to Myc, and functions to repress proliferation and promote differentiation.  Many human tumours are characterised by loss of Mnt, which is thought to act as a tumour suppressor.
The identification of transcriptional targets of Myc and Mnt will enable the design of therapeutics that target tumours associated with de-regulation of these proteins.  We have identified such transcriptional targets using chromatin immunoprecipitation and tiled promoter genomic microarrays.  We have deciphered the mechanisms whereby Myc and Mnt regulate their target genes during normal cellular proliferation and differentiation.  

Knowledge of transcriptional targets of Myc and Mnt will improve the accurate classification of many human cancers, and will lead to the development of specific therapeutics that target tumours associated with deregulation of these proteins.