research report

University of Leeds

2002/3

Research School of Medicine, Dentistry, Psychology and Health
Dean of Research: Professor M.A. Smith

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

EB1 is a microtubule-associated protein which interacts with the APC tumour suppressor protein. The interaction participates in the subcellular targeting of APC to the tips of microtubules which are actively extending. Microtubule tip-associated APC at discrete subcellular locations appears to be crucial in defining cell polarity and in directed cell migration. APC mutations commonly associated with colorectal cancer result in a loss of the APC/EB1 interaction and one of the earliest defects seen in mutant colonic epithelial cells is aberrant migration.

At least two EB1 homologues are also found in human cells, but these proteins remain relatively uncharacterised. We are currently performing a targeted mutational analysis of these homologues to refine our understanding of the structure and function of EB1 proteins. Crucially, this analysis, coupled with protein/protein interaction studies, has identified areas of functional overlap between EB1 and its human homologues. Indeed, we have now identified which of the human EB1 homologues are capable of interacting with APC and have new data concerning the nature of this interaction. Hence, just as mutant APC can no longer interact with EB1, mutant APC is also unable to interact with certain other human EB1 homologues. The consequences of the loss of these interactions must also be considered when investigating the cellular defects observed in colorectal cancer. This work has important implications for the understanding of both the migratory defects and genomic instability commonly observed in colorectal cancer cells.

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. We have identified a large candidate gene (OSC1) with a novel domain structure, which has homologies to proteins already implicated in carcinogenesis. Additionally, multiple intragenic homozygous deletions have been demonstrated in OSCC cell lines.

Mutation analysis of this very large 70 exon gene is on going in collaboration with Dr Nalin Thakker's group in Manchester. A number of OSCC cell lines and tumour samples have been screened by SSCP and sequence changes confirmed by sequencing. Additionally, OSC1 expression is also been investigated at both the RNA and protein level in oral samples and a range of tissue types. At least 5 different transcripts have been identified so far, and their expression pattern in normal and malignant samples is been analysed. Two peptide antibodies for the N-terminus and C-terminus of OSC1 have been commercially produced, their specificity for OSC1 is currently been confirmed by ELIZA, immunofluorescence co-localisation and Western blotting using OSC1 GFP constructs. These OSC1 antibodies will then be used to determine protein expression in normal oral musosa, premalignant lesions and tumour specimens.

Characterisation of this very large and complex tumour suppressor gene OSC1 will increase our understanding of the mechanisms involved in carcinogenesis in oral squamous cell carcinomas and other malignancies.

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

A useful 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. In this study we concentrate on the role of normal and aberrant versions of c-Myb and AML1. We differentiated ES cells homozygous null for the c-myb locus and showed that definitive precursors accumulate in c-myb-/- embryoid bodies. In contrast, unilineage precursor commitment and differentiation characteristic of primitive haemopoiesis was differentially affected in the macrophage and the erythroid lineages.

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 and are difficult to study. In order to study the effect of AML-ETO expression of haemopoiesis, we are examining the differentiation properties of knock-in ES cells. Our experiments indicate that in contrast to the c-myb knock-out, myelopoiesis in AML1-ETO ES cells is greatly diminished. We are also establishing stromal differentiation cultures, which allow to isolate mutant precursor cells. We have established a purification system for precursors (Tagoh et al., (2002)) allowing to examine the features of mutant precursor cells.

Tumour vascularity, cellular proliferation and the role of angiopoietins in colorectal cancer
Dr D. Burke, Mr J. Uddin, Dr P. Jones, Prof. P. Quirke

Tumour angiogenesis is an important process for tumour progression and is regulated by angiogenic proteins secreted by the malignant cells. Angiopoietins are recently identified angiogenic proteins that act upon a distinct endothelial cell receptor -Tie2 receptor. Tumour angiogenesis is thought to play an important role in colorectal cancer development, but the mechanism by which this occurs is uncertain. The relationship between tumour angiogenesis, Angiopoietin and Tie2 expression, and epithelial cell proliferation may be vital in understanding tumour biology but is not well understood.

In conjunction with Professor P Quirke (University of Leeds) and Dr P Jones (University of Leeds) immunohistochemical studies have been performed upon colorectal tumour specimens from a large, national clinical trial. These studies quantify expression of the angiopoietins, Tie2 receptor, tumour vascularity, cellular proliferation and apoptosis in these specimens. The tumour data from these studies will shortly be available. It will then be possible to analyse tumour biological characteristics with respect to data regarding tumour recurrence and patient survival. In parallel, the novel and important technique of tissue microarrays has been developed. The data from these arrays will be compared to that obtained from conventional immunohistochemical techniques. This should enable improved use of tumour specimens for clinical research.

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

GM-CSF is a growth factor that normally promotes the growth of myeloid cells in the bone marrow. GM-CSF is also often abnormally expressed as an autocrine growth factor in acute myeloid leukaemia (AML).

We are investigating mechanisms that contribute to the normal inducible regulation of the GM-CSF gene in white blood cells, and to the abnormal constitutive activation of the GM-CSF gene in AML.

We have shown that GM-CSF expression in myeloid cells is dependent upon an inducible transcriptional enhancer 3 kb upstream of the GM-CSF gene. When normal myeloid cells are activated, the chromatin structure of the enhancer is remodelled and a DNaseI hypersensitive site (DHS) is formed. In several AMLs that we have studied, the enhancer exists as an aberrantly activated constitutive DHS. We have used in vivo footprinting to identify transcription factors that are bound to the enhancer inside living cells. Starting with normal cells we have demonstrated that 2 composite NFAT/AP-1 sites are rapidly occupied by NFAT following activation. We also demonstrated that NFAT alone can remodel chromatin. Furthermore, we detect the binding of the constitutively expressed transcription factors AML1 and Sp1 shortly after the association of NFAT. Hence, NFAT binding appears to be essential for the formation of a multi-protein complex that assembles on the enhancer and mediates its activity. We are currently collecting an assaying a panel of AML samples to study abnormal mechanisms of enhancer activation and GM-CSF gene expression.

Tumour initiation and progression in the ApcMin mouse model of familial adenomatous polyposis
Dr P.L. Coletta, Dr C. Bonifer, Dr M.A. Hull

The Adenomatus polyposis coli tumour suppressor gene is mutated in the hereditary form of colorectal cancer, familial adenomatous polyposis (FAP) and in the majority of sporadic colorectal cancers. It is also mutated in a number of mouse models of FAP including the ApcMin mouse. We have previously shown that the inducible form of cyclooxygenase, Cox-2, is upregulated in macrophages early in this pathway in the ApcMin mouse. We have also shown that the majority of human colorectal adenomas express Cox-2 and that expression is localised to interstitial cells including macrophages. We have been characterising the population of Cox-2 expressing cells in ApcMin mice in order to determine the mechanism of Cox-2 upregulation and the functional consequences of Cox-2 expression.

To characterise the Cox-2 expressing population in ApcMin mice, we crossed the ApcMin mutation onto the M-lyslys-EGFP background in which macrophages express EGFP under the control of the macrophage-lysozyme promoter. Using this novel cross, we were able to isolate and sort macrophage populations from the intestine by FACS. In combination with antibody staining for a panel of macrophage-specific antibodies, we isolated the Cox-2 expressing population of cells and showed that Cox-2 mediated PGE2 production was highest in the M-lys expressing cells which also expressed macrophage surface markers. This approach is now being used to analyse gene expression of a number of genes including cytokines in the isolated Cox-2 expressing cells. This will provide cell-specific information on factors that drive the early stages of intestinal tumorigenesis.

Characterisation of host genes associated with the development of Helicobacter pylori induced gastric cancer
Dr J.E. Crabtree, Prof. M.F. Dixon, Dr 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. This model has been established to permit identification of H.pylori induced host genes during gastric cancer progression. Our previous study using cDNA array technology identified hundreds of known and novel host genes which were differentially expressed in gastric epithelial cells following stimulation with H. pylori. H. pylori upregulates both epidermal growth factor (EGF) ligands, such as amphiregulin and heparin binding-EGF, and genes encoding for ADAM metalloprotease-disintegrin proteins in gastric epithelial cells. In collaboration with Prof. Axel Ullrich's group at the Max Plank, Berlin, we have demonstrated that EGF receptor transactivation induced by H. pylori is dependent on extracellular transmembrane metalloprotease cleavge of proHB-EGF and signalling via mature HB-EGF. 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 have identified that the expression of several members of the ADAMs family are increased in patients with H. pylori infection and also in gastric cancer. To investigate the importance of H. pylori induced EGF receptor signalling in the model system, ADAMs genes, EGF receptor (EGFR), EGF ligands TGF alpha and HB-EGF, and COX2 have been characterised in the model. Long term infection with H. pylori was associated with increased HB-EGF in gastric epithelial cells, and marked increases in gastric expression of ADAM17, COX-2 and EGFR, demonstrating the importance of this signalling pathway in the model. The model system will allow future assessment of interventional strategies.

The role of apoptosis in peritoneal invasion
Mr D.G Jayne, Mr D.S. O'Riordain, Prof. P.J. Guillou

Peritoneal carcinomatosis is a frequent mode of dissemination of primary gastrointestinal cancers yet little is understood about the mechanisms involved. Using co-culture models, we have observed that colonic SW480 tumour cells rapidly adhere to peritoneal mesothelial monolayers and, within hours, induce apoptotic mesothelial cell death. This "apoptotic invasion" exposes gaps in the mesothelial barrier, allowing access to the submesothelial connective tissue. We have shown that tumour cells and mesothelial cells both express the death ligand FasL, and its receptor Fas, at the mRNA and protein level. Functionality of mesothelial Fas receptors has been demonstrated by use of stimulatory anti-Fas antibodies and crosslinked sFasL. Pre-treatment of tumour cells with an anti-FasL recombinant protein effectively inhibited mesothelial apoptotic killing, but not to baseline levels, suggesting that addition death ligands may be involved. Confocal microscopy has localise Fas / FasL expression to areas of tumour-mesothelial contact. Mesothelial surface FasL expression is upregulated in areas of tumour cell contact, not as a result of increased transcriptional activity, but probably by its release from cytoplasmic stores. It is speculated that this phenomenon may be responsible for the propagation of the apoptotic message between adjacent cells.

These findings have recently been submitted for peer-review publication, and future work is focusing on Fas/FasL apoptotic invasion utilising a more complex 3-dimensional model of peritoneal carcinomatosis.

The Role of Free Radical Induced DNA Damage during the Evolution of Oesophageal Adenocarcinoma
Dr L.J. Hardie, Dr G.W.B. Clark, Prof. C.P. Wild

The incidence of oesophageal adenocarcinoma (OA) has doubled over the last thirty years, surpassing the rate of increase for any other cancer type in western societies. OA is believed to develop from Barrett's oesophagus, a pre-malignant condition where the normal squamous epithelium of the oesophagus is replaced with a specialised intestinal type, usually in response to severe heartburn or gastro-oesophageal reflux.

The early molecular mechanisms driving the development of Barrett's oesophagus and OA are currently ill-defined. We are testing the hypothesis that gastro-oesophageal reflux stimulates oxidative stress and DNA damage in Barrett's oesophagus rendering this tissue prone to malignant transformation.

Over the last 18 months we have collected oesophageal biopsies from sixty Barrett's oesophagus and sixty dyspeptic control patients and analysed these for levels of DNA damage, including strand breaks and the pro-mutagenic DNA adduct, 8-hydroxydeoxyguanosine. Data reveal significantly increased levels of strand breaks in Barrett's mucosa compared with the normal squamous mucosa of these patients, and is consistent with a role for oxidative DNA damage during this disease process. To extend these exciting findings, we are currently assessing whether certain molecular changes in Barrett's oesophagus tissue, self reported reflux symptoms or medications are important modifiers of DNA damage levels in oesophageal tissue.
In addition to providing an insight into the natural history of this disease process, levels of DNA damage may help us identify those patients at increased risk of disease progression and serve as a biomarker to assess the efficacy of therapeutic regimes.

Down-regulation of?b-catenin underlies chemoprevention of colorectal cancer by non-steroidal anti-inflammatory drugs
Dr G. Hawcroft, Prof. A.F. Markham and Dr M.A. Hull

Non-steroidal anti-inflammatory drugs (NSAIDs) have a significant, but poorly characterised, protective effect against colorectal cancer (CRC). Previously, we have demonstrated that the NSAID indomethacin decreases expression of oncogenic b-catenin and the b-catenin/TCF target gene cyclin D1 in human CRC cells in vitro. Experiments on a panel of NSAIDs have revealed that, although the effect of indomethacin on b-catenin protein levels in human CRC cells in vitro is shared by the related NSAID sulindac sulphide, a significant reduction in b-catenin protein levels does not invariably occur in association with growth arrest and apoptosis induced by other NSAIDs, such as diclofenac,. Furthermore, decreased cyclin D1 protein levels are apparent in NSAID-treated cells prior to significant attenuation of b-catenin levels suggesting that down-regulation of cyclin D1, associated with the anti-proliferative activity of NSAIDs, cannot be explained fully by abrogation of b-catenin-mediated WNT signalling.

GPI Deficient lymphocyte function and opportunistic infection in patients with chronic lymphocytic leukaemia treated with CAMPATH-1H
Dr P. Hillmen and Dr B. Kennedy

CAMPATH-1H is a humanized, therapeutic, monoclonal antibody specific for the Glycosylphosphatidylinositol (GPI) anchored CD52 molecule. It is effective therapy in refractory B-CLL. We have treated over 50 patients within the Yorkshire region.
We have shown GPI-anchor deficient cells form the majority of reconstituting CD3+ T-lymphocytes after CAMPATH-1H therapy in 75% of cases. Such PNH phenotype cells remain in the majority (median 56%) 6 months post therapy. Using 4 colour flow cytometry we have monitored T-cell recovery after CAMPATH-1H at levels from 0.01x109/l. Median CD3+ count during CAMPATH-1H therapy is < 0.01x109/l. At the end of therapy a median of 97% CD3+ cells are GPI deficient. We have reported an 8.9% risk of reactivation of CMV infection, which is associated with significant mortality. Using a CD2/CD2R invitro activation system we found that GPI-deficient T-cells demonstrate normal upregulation of CD69, the earliest marker of T cell activation. We found no change in expression of IL2 receptor, CD71 or HLA DR in GPI positive or negative T-cells after activation. There is an increased proportion of GPI-deficient CD3+ cells expressing the CD45ROHigh memory phenotype. By using flow cytometric analysis of the T-cell receptor Vb expression we have demonstrated that GPI-deficient T-cells are polyclonal although the range of TCRVb appears to be somewhat restricted. GPI positive and negative T-cell populations in the same patient appear to express different TCRVb repertoires. We have performed T-cell receptor excision circle (TREC) analysis on patients at various times after the completion of Campath in order to investigate the maturation of the repopulating T-cells, both GPI-normal and GPI-deficient. In conclusion, the in vitro analysis of T-cell recovery following Campath therapy demonstrates that the majority of the recovering T-cells are abnormal in that they are GPI-deficient and have a somewhat restricted repertoire as defined by their TCR Vb expression. However these T-cells have apparently normal in vitro function.

Investigation and Characterization of Genetic Abnormalities (14q32) in Multiple Myeloma
Prof. G.J. Morgan, Dr J. Fenton

Translocations involving the IgH locus are the commonest genetic lesion in multiple myeloma (MM). Studying the nature of the breakpoints can help us to understand the molecular mechanism leading to MM. We have characterised the genomic breakpoints of nine MM patients, seven with t(4;14) translocations, and two t(11;14). By convention chromosome 14q32 breakpoints in MM are believed to be located in the IgH S? switch region on der (4) or der (11) and a further downstream switch region on der (14), with deletion of intervening DNA occurring as a result of aberrant class switch recombination (CSR). Our analysis showed such breakpoints did occur, but that there was also evidence of more complex recombination events in four of the nine patients. These rearrangements show rearranged hybrid switch region sequences are joined to the foreign partner DNA associated with the IgH translocation event. Although this might suggest a complex three-way translocation, the most probable explanation is that two separate recombination events have occurred. This suggests that primary IgH translocations can be more complex than was previously described are independent of translocation partner and need not necessarily be mediated by an aberrant CSR mechanism.

The role of fibroblast growth factor receptor 3 (FGFR3) in multiple myeloma
Prof. G.J. Morgan, Dr D. Gonzalez de Castro

We are initiating a project looking at the effects of over expressing of the tyrosine kinase receptor FGFR3 at different stages of haemopoietic differentiation. This gene is commonly deregulated in myeloma and these experiments will allow us to understand its effects in cell growth and differentiation as well as the downstream genes, which mediate these effects.??

Both wild type and a constitutively active mutant form of FGFR3 have been cloned into a tetracycline inducible expression system. The appropriate constructs will then be transfected into an embryonic stem cell line and we will examine the effects of turning FGFR3 on and off within this model.

Interaction between the APC tumour suppressor protein and the mitotic spindle component EB1 and the significance of its disruption in colorectal carcinoma
Dr E.E. Morrison and Dr J.M. Askham

Mutation of the APC gene is a frequent event in colorectal carcinoma. These mutations commonly result in the production of a truncated protein that has lost the ability to interact with a number of partners. One of these is a microtubule tip-associated protein called EB1. We have previously demonstrated that EB1 also interacts with the p150glued subunit of the dynein/dynactin microtubule motor complex, and shown that the region in EB1 responsible for this binding overlapped with that required for the APC interaction. By expressing defined EB1 deletion mutants fused to GFP, we have now probed the function of EB1 in transfected cells. In cells expressing the last 84aa of EB1, a fragment capable of binding either APC or p150glued but not microtubules, profound defects in microtubule organisation and centrosomal anchoring were observed. EB1-C84-GFP expression severely inhibited microtubule regrowth, focusing and anchoring in transfected cells during recovery from nocodazole treatment. The recruitment of ?-tubulin and p150glued to centrosomes was also inhibited. None of these effects were seen in cells expressing the last 50aa of EB1 fused to GFP, a fragment that cannot interact with p150glued. Furthermore, EB1-C84-GFP expression did not induce Golgi apparatus fragmentation, indicating that the activity of the full dynein/dynactin complex was not inhibited by the expression of this fusion protein. We propose that a functional interaction between EB1 and p150glued is required for microtubule minus end anchoring at centrosomes during the assembly and maintenance of a radial microtubule array.

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

TOGp is a large microtubule-associated protein that is overexpressed in colorectal and hepatic carcinoma. Based on studies of its homologs in other organisms, it is thought to play a major role in promoting microtubule growth, particularly during mitosis. We have previously defined the intracellular localisation of TOGp by immunofluorescence, identifying a centrosomal localisation. Using a TOGp fragment fused to GFP we have found that this centrosomal localisation is mediated by the C-terminal third of the protein. Overexpression of this fusion protein produced unexpected effects in transfected cells. A peripheral accumulation of microtubules was seen in association with an apparent defect in the organisation of a radial microtubule array. Furthermore, expression of this fusion protein both antagonised microtubule depolymerisation by nocodazole and inhibited microtubule re-growth after drug wash out. At later times after drug wash out a displacement of microtubules to the cell periphery was again seen. This complex set of data is consistent with a role for TOGp in both the growth and organisation of microtubules. It is currently unclear whether TOGp localises to centrosomes directly or via interactions with other proteins. An interaction between the TOGp C-terminus and the TACC family of proteins has been suggested to perform this function. A series of smaller TOGp C-terminal fragments fused to GFP has therefore been constructed and the ability of these fragments to localise to centrosomes and interact with TACC proteins is currently under investigation.

Selection of patients for adjuvant therapy in colorectal cancer. Getting it right with molecular profiling
Prof. P. Quirke and Dr M. Seymour

We have continued to work on the AXIS and QUASAR1 clinical trial series. We have published the AXIS data in the Lancet (Barratt et al 2002) suggesting that 18q loss may relate to a failure to respond to 5FU therapy. In QUASAR1 we have constructed two prognostic groupings, the first a pragmatic histopathological index that identifies good prognosis Dukes stage B carcinoma patients with a 5 year survival of 80% and a poor prognosis group with a 65% 5 year survival. The prognostic indix relies upon peritoneal involvement and extra mural vascular invasion. We have an improved prognostic index when dUTPase staining is included expanding the gap to 82% vs 60%. We are currently confirming the former in a second population based pathology series of 5,500 Yorkshire patients in collaboration with NYCRIS. This grouping should allow the identification of 17% of patients who should benefit from 5FU therapy. We have constructed tissue microarrays and are continuing to add new parameters (SMAD4, TP, DCT etc.) to the Quasar1 series whilst awaiting the response to therapy data and have embarked upon collecting a second data set of a further 800 patients. We expect the response data to be available at the end of 2003. Quasar2 is expected to commence recruitment this year and we will collect material from this trial. Other significant successes are 4 million pounds of NHS funding for a national colorectal education program to build on the pathology/surgery program we have developed around Total Mesorectal Excision with Professor Heald. We hope that Yorkshire will be amongst the first participants in this training. The value of our approach has been demonstrated in Birbeck et al 2002.

The significance of estrogen receptor in breast cancer
Dr V. Speirs, Dr G. Skliris

Two human estrogen receptor (ER) isoforms exist, the "classic" ERa and the more recently identified ERb. ERb is an important prognostic factor in breast cancer, and although ERb is expressed in the breast, its biological role remains undefined. To gain insights into the possible role of ERb during breast carcinogenesis immunohistochemical analysis of ERb was performed on over 500 breast specimens encompassing the full spectrum of breast pathologies (normal gland, ductal carcinoma in situ, invasive cancers, lymph node metastases and recurrences). A gradual reduction, but not a complete loss of expression of ERb was observed during the transition from normal breast and pre-invasive lesions (constitutive expression) to invasive cancers where the receptor was lost in 20% of cases. This was more pronounced in invasive ductal compared to invasive lobular carcinomas where a significantly higher proportion were ERb positive. Treatment of ERb cell lines with DNA methyltransferase inhibitors restored ERb expression, providing experimental evidence that silencing of ERb in breast carcinomas could be due to promoter hypermethylation. Our results suggest that loss of expression of ERb may be one of the hallmarks associated with breast cancer progression. This appears to be a reversible process involving hypermethylation of the ERb promoter.

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

The role of ERb in breast pathobiology still remains unclear. However, it is becoming apparent that the overall estrogenic/antiestrogenic response of a given cell or tissue appears to be dependent on the expression, not only of wild type ERb but also variant isoforms. Despite the presence of these variants in breast cancer, surprisingly little is known about their biological role. The aim of this project is to identify the function of ERb variants and evaluate their distribution and expression in human breast tumours resected from patients who remained sensitive to tamoxifen cells and those who acquired resistance. Heterogeneity in ERb isoforms within tumours may alter ligand binding and may be important predictors of endocrine sensitivity. This study is being carried out in collaboration with Dr Indra Poola, Howard University, Washington DC.

Identification of Barrett's Oesophagus Patients at High Risk of Progression to Oesophageal Adenocarcinoma using Molecular Markers
Prof. C.P. Wild, Prof. D. Forman, Dr M. Scott

Adenocarcinoma of the oesophagus (ADC) develops from the premalignant condition know as Barrett's oesophagus (BO). This change is accompanied by a series of molecular alterations in genes which regulate cell division. These alterations may be used to identify BO patients at particularly high risk of progression to ADC and refine endoscopic surveillance programmes. Using such an approach, we have already shown that cyclin D1 overexpression is associated with a seven fold increased risk of progression to ADC.

Over the last year, we have accessed BO biopsy specimens from ADC cases which arose between 1993 and 1999 in Northern Ireland using hospital tissue archives. These cases have been closely matched to a series of biopsies from BO patients who did not progress to cancer over an equivalent surveillance period. All biopsies have been sectioned and stained for four molecular markers using carefully controlled immunohistochemical approaches. To ascertain the value of these molecular changes as predictive markers of disease progression in BO tissue, these sections are currently being evaluated by two independent histopathologists utilising strict scoring criteria.

It is anticipated that information obtained from this study will confirm whether molecular markers can be used to refine entry criteria into endoscopic surveillance programmes and potentially offer earlier detection and more effective treatment of ADC.

Faculty of Biological Sciences
(School of Biochemistry and Molecular Biology)
Chairman of Faculty: Professor K.T. Holland
Head of the School: Professor A.J. Turner


Investigation into the role of Bcr-Abl-induced abnormalities in glucose transport regulation in the pathogenesis of chronic myeloid leukaemia
Dr K. Barnes, Ms E. McIntosh, Ms D. Itchayanan, Dr. J. Bentley, Prof. S.A. Baldwin

A primary function of haemopoietic growth factors is to maintain rates of glucose uptake sufficient to suppress apoptosis, thus ensuring cell survival under appropriate circumstances. A key step in the development of chronic myeloid leukaemia (CML) is the inappropriate survival of haemopoietic stem cells associated with the expression of the chimaeric tyrosine kinase, Bcr-Abl. We have demonstrated, using haemopoietic cell lines as models for CML, that growth factor stimulation of glucose transport is dependent on PI 3-kinase activation and involves cytoskeleton-mediated GLUT1 translocation to the cell surface. In the absence of growth factor the Bcr-Abl oncoprotein is able to maintain high rates of glucose uptake leading to inappropriate suppression of apoptosis.

To investigate the mechanism of Bcr-Abl action, we are exploiting a haemopoietic cell line, TonB210, developed by our collaborator Dr G. Daley (Massachusetts Institute of Technology), in which expression of Bcr-Abl is tetracycline-dependent. Using these cells and the specific Bcr-Abl kinase inhibitor STI571 (Glivec), a gift from Novartis Pharma, the kinetic properties of the enhanced glucose uptake induced by Bcr-Abl tyrosine kinase activity have been investigated in detail. We have also demonstrated that in Bcr-Abl-activated TonB210 cells the phosphorylation of protein kinase B/Akt lies downstream of PI 3-kinase. In conjunction with Professor G. Morgan (Leeds General Infirmary) we will now measure glucose uptake in clinical samples. If Bcr-Abl-mediated translocation of GLUT1 is shown to play a role in apoptotic suppression in vivo, this should lead to the identification of novel therapeutic targets for CML.

High risk human papilloma virus oncogenes: Studies on antigen processing and presentation in transformed cells, organotypic cultures and tumours
Dr G. Bottley, Mrs J Jarvis and Dr G.E. Blair

The high risk human papillomaviruses (such as HPV 16 and 18) are closely linked with cervical and ano-genital cancers.? Infection of the epithelium cells with these viruses leads to the expression of the transforming oncogenes E6 and E7 which interfere with the regulation of the cell cycle, and therefore increase the incidence of malignancy.? Many tumours, including HPV-associated tumours, display reduced levels of cell surface major histocompatibility complex (MHC) class I molecules. This results in reduced immune activity against the tumour and thus enhances tumour growth and survival.? We are investigating the action of E6 and E7 from both high risk HPV 18 and low risk HPV 6b on the MHC antigen processing and presentation pathway. The target cell of HPV in vivo is the keratinocyte, therefore both HaCaT, a keratinocyte cell line, and primary human keratinocytes form suitable cell systems for investigatation.?

The initial part of the study examined the changes in MHC class I expression by HaCaT cells following transient transfection of the E7 oncogene.? A pilot study revealed the difficulties inherent in a transient system of transfection, including the detection of E7 on a cell-by-cell basis.? The HPV 18 E7 gene construct prepared as part of a previous YCR project was ligated into a bicistronic plasmid vector.? This vector also directs the expression of a truncated H-2Kk molecule at the cell surface which can be detected by flow cytometry or can be used to purify transfected cells using magnetic beads.? It is therefore possible to use two- and three-colour flow cytometry to correlate changes in cell phenotype with E7 expression on a cell-by-cell basis.? Use of the magnetic beads allows transfected cells to be analysed by Western blotting, RT-PCR and other techniques independently of non-transfected cells.? An additional benefit of using the H-2Kk marker is that the magnetic beads can be used to create a stably transfected cell line in a very short time.? We are also constructing inducible cell lines in which expression of E7 is controlled by a tetracycline-regulated promoter. Finally the effect on antigen presentation of knocking out E7 expression is being investigated using RNA interference technology.

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

In human PC, NEP loss contributes to androgen-independent progression by allowing mitogenic peptides such as endothelin-1 (ET-1) to stimulate cell proliferation. ECE-1 processes inactive "big endothelin-1" into active ET-1. ECE-2 has been identified, but its physiological role has been little explored. SEP, a soluble endopeptidase, is highly expressed in prostate, hence its potential involvement in PC progression deserves exploration. Human PHEX gene expression is high in bone and given that bone is a primary site of metastasis, PHEX could contribute to, or modulate metastatic progression. Two other NEP/ECE-like genes of unknown function (ECE-L1 and XCE) complete the known complement in the human genome.

Co-culture matrigel invasion data show that stromal/epithelial interactions can influence the invasive ability of PC cells as a consequence of their altered NEP and ECE-1 activities. Transient expression of ECE-1c isoform only, in PC cells increased invasion. Primary malignant stroma increased PC cell invasion more than primary benign stroma. Immunohistochemical analysis show that ECE-1 expression is up-regulated in stroma adjacent to tumour epithelia, and expression extends throughout the stroma as the tumour becomes more aggressive. ECE-1c isoform is the only detectable isoform in tumour sections. RT-PCR studies show: NEP is absent in highly metastatic cells; ECE-1 is present only in metastatic cells; PHEX is present only in metastatic cells; SEP is present in all cells tested; XCE is undetectable in all cells tested.
(Archive tissue biopsies kindly provided by Prof. P. Quirke, LGI; Primary cell cultures and novel cell lines kindly provided by Prof. N. Maitland, York Cancer Unit)

Kaposi's sarcoma: investigating the Human Herpesvirus-8 latent-lytic switch
Dr A. Whitehouse and Dr D. Goodwin

Kaposi's sarcoma-associated herpesvirus or human herpesvirus 8 (HHV-8) is a lymphotrophic?g-2 herpesvirus associated with Kaposi's sarcoma (KS), a major neoplasm particularly in AIDs patients. KS is an endothelial neoplasm. However, the primary reservoir of latent virus is established in B lymphocytes. Reactivation to lytic HHV-8 infection from the latently infected lymphoid reservoir is essential for the spread of infectious virus to the endothelium. Therefore, reactivation from the latent state into the lytic cycle, allowing active HHV-8 gene expression and replication is a necessary antecedent step in KS development. However, little is known about the nature of the latent-lytic switch. We have demonstrated that the ORF 50 gene products in?g-2 herpesviruses are essential transcriptional regulators, activating the lytic replication cycle. Moreover, forced expression of this protein under the control of an heterologous promoter, has the ability to trigger reactivation of the lytic replication cycle from an established latent infection. This suggests that regulation of the ORF 50 promoter itself is a key component of the switch from latency to a lytic replication cycle and is an important determinant of the natural history of the viral infection. We are now performing detailed analysis of the ORF 50 promoter using DNase I footprinting. Specifically work is being performed to determine whether a number of tentatively identified cellular factors play central roles in regulating the latent-lytic switch in gamma-2 herpesviruses. Moreover, we have identified a novel feedback mechanism which autoregulates the ORF 50 promoter, via the ORF 50 protein itself.

Development of an adenovirus vector which utilises the Herpesvirus saimiri episomal maintenance system to allow efficient long term expression of therapeutic tumour suppressor genes in carcinoma cells
Dr D.A. Matthews, Dr K.T. Hall, Dr A. Whitehouse

This study aims to develop an improved adenovirus cancer gene therapy vector. This will initially involve determination of the minimal components of the Herpesvirus Saimiri (HVS) episomal maintenance system. We have demonstrated that HVS will maintain episomes in a wide range of human, primate and murine cells. Thus the episomal maintenance elements will have broad utility for the delivery of therapeutic genes to a wide range of malignant cell types. We aim to combine this attractive feature with an already established gene therapy vector, adenovirus. Currently, this vector is not capable of maintaining a delivered transgene over successive rounds of cell division, a critical requirement when attempting to moderate tumour cell growth with tumour suppressor genes. We have now identified the minimal elements required for HVS episomal maintenance. These comprise a trans-acting factor encoded by ORF 73, which co-localises with host mitotic chromosomes and a cis-acting element contained with the terminal repeats of the HVS genome. Future work will involve combining the episomal maintenance elements into a recombinant adenovirus and characterise its utility to deliver and maintain a range of therapeutic transgenes in an episomal state over several rounds of cell division in carcinoma cells.

Identification of genes that are transcriptionally repressed by Mxi1
Dr. S.C. Wright

The c-myc oncogene is overexpressed in diverse human tumours. Myc acts as a transcriptional activator to accelerate cell cycle progression; its biological activity is importantly antagonised by the Mad family of transcriptional repressors (Mad1, Mxi1, Mad3 and Mad4). The aim of the current project is to identify target genes that are transcriptionally repressed by Mxi1. We have generated cell lines that express inducible alleles of either Mxi1 or a dominant-negative Mxi1 gene, and are currently using these in conjunction with microarray analysis to identify Mxi targets. In addition, we are identifying Mxi1 targets using a chromatin immunoprecipitation approach.

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

(A multidisciplinary group of staff from the School of Biochemistry and Molecular Biology, the School of Biomedical Sciences, the Department of Colour Chemistry, the Research School of Medicine, Cookridge Hospital, Leeds General Infirmary, St James's Hospital and Goole and District Hospital.)

A multidisciplinary approach to the development of photodynamic therapy for cancer treatmentS B Brown (Principal Investigator) et al

Yorkshire Cancer Research has provided major support for the development of photodynamic therapy (PDT) by its funding of the Leeds Centre for Photobiology and Photodynamic Therapy for more than 15 years. This Centre which is the largest of its kind in Europe and among the global leaders in the development of PDT, now has around 35 scientists and clinicians in the Yorkshire region and many collaborations both in the UK and internationally.

As outlined in previous years, PDT is a developing approach which uses a combination of light and a photosensitising drug (a sensitiser) to selectively destroy unwanted tissue such as tumours. The approach is able to achieve a high degree of targeting partly because of the ability of photosensitisers to concentrate preferentially in tumours relative to normal tissue, but also because of our ability to direct light very accurately to the tumour area.

Further progress has been made in the development of the Centre's portfolio of photosensitisers towards clinical trials. The Centre's drugs consist of families of phthalocyanines and families of phenothiazines. In each case, the central macrocyclic structure largely determines the absorption and sensitising properties and the substituents around the molecule determine the important biological properties. We are anticipating the first use of these drugs in patients for both oncology and for anti-infection within 15 - 18 months, provided that funding can be obtained to support the toxicology and other studies required to be carried out by Contract Research Organisations. In this report, we focus on the continuing novel chemical synthetic work, studies of the mechanisms of PDT and studies on light dosimetry, which has proved very important for successful PDT.

In our Chemistry section, research has focussed on the synthesis of further novel photosensitiser systems for PDT. In the zinc phthalocyanine series, the development of non-isomeric systems has been undertaken, in an effort to overcome one of the main regulatory approval problems facing sensitisers of this class. Two types of structure have been examined: (a) of mono-substituted zinc phthalocyanines, and (b) tetrasubstituted derivatives containing fused five-membered rings. Examples of both types have now been prepared successfully. The phenothiazinium photosensitisers have no such isomer problems, but their PDT activity is much more critically dependent on subtle structural modifications. A number of derivatives have now been identified with excellent PDT properties (i.e. high tumour selectivity and low skin photosensitisation).? The effect of the counter-anion in these salts has also been investigated.

The efficacy of photosensitising compounds used in PDT depends on several factors. The mechanisms involved in the cellular uptake of these drugs is important as this determines the cellular localisation, accumulation and, to some degree, the mode of cell death. Investigations of sensitisers from the phthalocyanine series have concentrated on the mechanism of cell death including cell cycle analysis and apoptosis/necrosis evaluation using flow cytometry techniques. These studies have been further supported by comet assays and Ames tests. The data so far have demonstrated that PDT using this series of photosensitisers in a human cervical cell line (SiHa) results in apoptosis, inhibition of cell growth and G0/G1 cell cycle arrest in a time-dependent manner. These finding provide evidence for the involvement of cell cycle deregulation.

In collaboration with Dr Clive Kelty, Dr Roger Ackroyd, Dr Nicola Brown and Professor Malcolm Reed (Royal Hallamshire Hospital, Sheffield) we have performed a series of in vivo light dosimetry measurements during ALA-PDT of Barrett's oesophagus. Laser light (635 nm from a diode laser) was delivered via a cylindrical diffusing fibre located within the centre of a proprietary oesophageal balloon. A detector fibre located at the balloon/tissue interface is connected to a light detection system that is designed to record the fluence rate at 3 specific wavebands.? Data was successfully recorded from 14 patients, indicating that the in situ light dose is approximately 70% greater than that delivered, due to light scattering within the oesophagus. Also it is observed that the fluorescence signal from ALA-induced protoporphyrin IX is progressively reduced during light exposure, at a rate that implies complete photobleaching after only 30% of the light exposure. This raises important questions concerning the light dose necessary to achieve effective PDT.

As previously reported, it is anticipated that the Centre's drugs will be most suited to early cancer and pre-cancerous conditions, such as the treatment of Barrett's oesophagus which can lead to adenocarcinoma of the oesophagus and CIN which can lead to cancer of the uterus. Our novel sensitisers are also being actively developed to treat bacterial and other infections, especially of wounds such as leg ulcers and burns.

Photopharmica, the Company which has been formed to progress these developments commercially, is planning to lead the later stages of drug development into patients. With active involvement of YCR, it is hoped that this company will provide the vehicle by which the major investment of YCR and the University in PDT can be harnessed to benefit large numbers of patients.