research reports

University of Sheffield

2006/7

Faculty of Medicine
School of Medicine and Biomedical Sciences

Director's Introduction
Prof B.W. Hancock

Cancer research in Sheffield has well-established programmes in both scientific and clinical research. The main scientific research programme is based in the YCR Institute for Cancer Studies, which occupies 800 sq. metres of refurbished space on the top floor of the Medical School. As well as research space for seven investigators, the Institute provides core facilities for cancer researchers throughout Sheffield. These facilities include high throughput DNA sample processing, deconvolutional microscopy, and a radiation source.

The Academic Unit of Clinical Oncology is located in the Cancer Research Centre (CRC), a purpose built clinical trials facility at the Weston Park Hospital. The CRC provides facilities for the assessment and treatment of patients, office space for staff with fully networked computer systems, and specialist facilities including laboratory space and bone densitometry. The North Trent Cancer Research Network is co-ordinated from the CRC.

While these two research components are physically separate, excellent interactions between them have already been established, with numerous collaborations based in cancer genetics, genetic epidemiology and new therapies. Other important cancer research teams are based in the Medical School including the Academic Units of Pathology; Infectious Diseases; Ophthalmology & Orthoptics; Surgical Oncology; Urology and in the University Department of Biomedical Science. There are also researchers now established in the School of Clinical Dentistry (Department of Oral Pathology).

The effect of single nucleotide polymorphisms on clinical outcome and survival in the FOCUS clinical trial of metastatic colorectal cancer
Dr. A. Cox and Profs. P. Quirke and M. Seymour, (University of Leeds)

The aim of this project, which was funded in December 2005, is to identify genetic variants that are associated with outcome and survival from metastatic colorectal cancer. The mainstay for many years for first line chemotherapy in colorectal cancer has been 5-fluorouracil, but the FOCUS clinical trial has helped determine the best way to use newer agents such as oxaliplatin and irinotecan. However, many patients do not respond well to these therapies and it would be of benefit to be able to identify which patients are likely to respond. We have previously shown that inherited variation in the genes involved in the cellular response to DNA damage can be related to patient survival. DNA samples are available from around 1200 patients enrolled in FOCUS, and we have obtained approval from the Local Research Ethics Committee for use of these samples to determine whether DNA damage response genes are associated with treatment outcome in the FOCUS patients. We have developed a panel of single nucleotide polymorphisms (SNPs) in 8 genes involved in the DNA damage response, including XRCC2, ATM, XRCC3, RAD51, BRCA2, MRE11, RAD50 and NBS1. The SNPs have been selected to capture the common variation present in the population for each of these genes. We will analyse the SNP genotypes in relation to clinical outcome and survival data from the FOCUS study, to determine whether any of these genes can predict treatment response or survival.

Poly(ADP-ribose) polymerase (PARP-1), ATM and DNA-PK are all involved in responding to DNA damage to activate pathways responsible for cellular survival. Here, we demonstrate that PARP-1-/- cells are sensitive to the ATM inhibitor KU55933 and conversely that AT cells are sensitive to the PARP inhibitor 4-amino-1,8-napthalamide. In addition, PARP-1-/- cells are shown to be sensitive to the DNA-PK inhibitor NU7026 and DNA-PKcs or Ku80 defective cells shown to be sensitive to PARP inhibitors. We show that ATM is activated following inhibition of PARP and that PARP inhibitor-induced recombination repair is abolished in ATM, but not DNA-PK, inhibited cells. Altogether, we suggest that ATM is activated by PARP inhibitor-induced collapsed replication forks and may function upstream of HRR in the repair of certain types of DSBs.

We also show that replication protein A (RPA) is involved in genome maintenance by the homologous recombination repair system in humans. Depletion of the RPA protein inhibited the formation of RAD51 nuclear foci after hydroxyurea-induced replication stalling, leading to persistent unrepaired DNA double-strand breaks (DSBs). In addition, we demonstrate a direct role of RPA in homology directed recombination repair of a DSB. We find that RPA is dispensable for Chk1 activation and that RPA directly binds RAD52 upon replication stress. We show that inhibition of Chk1 with UCN-01 impairs dissociation of RPA from the chromatin and prevents association of RAD51 and RAD52 with DNA. Altogether, our data suggests a direct role of RPA in homologous recombination in assembly of the RAD51 and RAD52 proteins. Also, our data suggest that replacement of RPA with the RAD51 and RAD52 proteins is dependent on Chk1 signalling.

Genetic instability and cancer
Prof M. Meuth

Cells respond to DNA damage by triggering cell cycle arrest, DNA repair, or death. DNA damage response pathways are frequently disrupted during tumour development leading to genetic instability, loss of cell cycle checkpoints, defects in the induction of apoptosis, and altered responses to widely used chemotherapeutic agents. The objectives of our work are to elucidate the interactions of pathways controlling the response to DNA replication fork stress, to determine how these pathways are altered in tumour cells, and to determine how they may be exploited to improve therapy. One of our major lines of work is aimed at understanding the role of one of the primary regulators of the DNA damage response (a protein called ATM) in the cellular response to disruption of DNA replication. We have discovered that a subset of human colon cancer cell lines are defective in ATM activation following replication fork disruption and identified a mutant allele of another regulator (called MRE11) that suppresses ATM activation and ATM-mediated rescue of disrupted DNA replication forks. Our results have important implications for the regulatory pathways and suggest a novel and therapy specific for tumour cells defective in this response. A second line of work is aimed at understanding how cells decide whether to repair DNA and restart DNA replication or die following disruption of replication. We have shown that the efficacy of a clinically relevant agent that disrupts DNA replication (camptothecin) can be enhanced by manipulation of S-phase signalling pathways. DNA replication inhibitors that do not normally induce cell death can be converted into potent inducers of apoptosis by depletion of the Chk1 checkpoint kinase. Thus combinations of agents that disrupt S-phase checkpoints with those that inhibit DNA replication may be particularly effective in the treatment of some types of tumors.

A structural investigation of the papillomavirus replication initiation complexes
Dr C.M. Sanders and Dr A. Anston (Structural Biology Laboratory, University of York)

The papillomaviruses are the causative agents of warts but are also associated with certain cancers, principally carcinoma of the cervix. Two independent approaches are being employed towards combating viral disease. The first is immunological or vaccine based, aimed at stimulating the bodys own immune response to the virus. The second is chemotherapeutic and involves an understanding of viral proteins as targets for anti-viral drugs. We are undertaking a comprehensive structural and functional analysis of two viral proteins, known simply as E1 and E2, that coordinate and control viral DNA replication. We have determined the molecular structure of part of the E2 protein that interacts with E1 and controls viral replication, and now also the molecular motor of E1 that drives DNA unwinding during replication. Our complementary biochemical studies of theses proteins are providing significant insights in to how they function and how their activities are regulated. The current focus is on the E1 protein, and we are testing models for the DNA melting and unwinding activities of the enzyme, that have arisen in the light of our structural analysis. We also seek to understand the molecular and atomic details of how the E1 protein complex interacts with a DNA replication fork.

Kinases, phosphatases and non-enzymatic regulatory molecules control passage through the cell-cycle. As well as co-ordinating the normal events from replication initiation through to cytokinesis, specialised regulatory pathways intercede when a genotoxic stress is perceived. These checkpoint pathways function to avoid or repair mutagenic errors and ensure accurate DNA replication. Loss of function of these pathways occurs frequently during carcinogenesis. Critical to our understanding of these events is the ability to study cells in defined phases of the cell-cycle. Cell sorting and syncronisation can be achieved without drugs or physiological stress by centrifugal elutriation. Our groups have begun to exploit the elutriation device provided by YCR in order to understand further the proteins and processes that govern the mammalian cell cycle.

Academic Unit of Urology

The role of osteoprotegerin on prostate cancer survival
Dr C. Eaton, Professor F.C. Hamdy and Dr I. Holen (Academic Unit of Clinical Oncology, Division of Genomic Medicine)

Studies to evaluate the effects of modulated OPG expression on TRAIL induced apoptosis in prostate cancer cell lines have continued. Cloned populations of transfected prostate cancer cells in which full length OPG has been stably overexpressed or in which endogenous OPG has been knocked down have been studied extensively in vitro. This work has shown that sensitivity to TRAIL in cloned tumour populations does not correlate with OPG production: i.e. clones that secrete high levels of OPG may be either TRAIL sensitive or resistant, and equally, clones that secrete low levels of OPG are similarly variable in their responses to TRAIL. This reflects underlying heterogeneity that we have now shown to pre-exist in the parent tumour cell lines. This work does not support our original hypothesis regarding the central role of OPG in TRAIL resistance in prostate cancer, but has provided valuable insight into tumour heterogeneity and TRAIL responses and has been provisionally accepted for publication.

We have continued to examine responses of tumour cells to TRAIL in vivo and have shown high levels of TRAIL induced killing in both intraosseous and subcutaneous tumour xenografts. Resistant populations identified in these experiments, and the role of OPG this phenotype in vivo, are being studied.

We have examined serum OPG levels in ~500 treated and untreated prostate cancer patients with up to 4 years follow-up. This study has, for the first time, demonstrated that raised serum OPG levels at diagnosis in patients with locally advanced disease, but without overt metastases, is a strong predictor of biochemical relapse within the follow-up period. This is an important finding and could be extremely valuable in planning treatment in this large patient group.

Finally, we have found that one of the monoclonal antibodies widely used for the detection of OPG displays strong cross-reactivity with carbonic anhydrase II (CAII). The latter is a commonly expressed in cancer and the discovered cross-reactivity of the OPG antibody has identified several publications where OPG expression by cells in tissue sections has been wrongly assigned. The final part of the project, to correlate OPG staining in primary cancers with serum OPG in the patient collection referred to above, is now underway using an alternative antibody that is specific for OPG.

Influence of Tumour Cell Death Mechanisms and Immunoregulatory T Cells on Tumour Growth and Immunity
Prof N.J. Brown

The mode of cell death may influence the induction of tumour immunity, as cellular necrosis is an inflammatory stimulus, whereas apoptosis can have anti-inflammatory consequences. This study, which commenced in May 2006, is currently evaluating the influence of cell death mechanisms on the induction of murine CD4+CD25+ regulatory T cells. The first aim of the research proposal was to generate specific apoptotic and necrotic EMT6 populations with a greater than 95 % purity as confirmed by Annexin V/PI flow cytometry. Current experiments involve the administration of specific necrotic and apoptotic EMT6 cells into tumour-bearing mice. Studies are focusing on the distinct effects on tumour growth/survival, and the recruitment, expansion and activities of CD4+CD25+ regulatory T cell populations. Mouse CD4+CD25+ regulatory T cells were therefore isolated and confirmed on the basis of CD4-FITC, CD25-PE and foxp3-APC flow cytometry analysis. This research clearly has international importance and a BACR Travel Fellowship has been awarded in order to gain experience over a three week period at the Technical University in Munich. Collaborative work will investigate the capacity of tumour expressed Hsp70, a ligand related to the specific mode of cell death, to activate CD4+CD25+ immunoregulatory T cells.

Many haemostatic proteins modulate tumour angiogenesis resulting in either stimulation or inhibition, thus understanding the relationships between these processes has implications for cancer therapy. Tissue Factor (TF), a key haemostatic molecule, up-regulates vascular endothelial growth factor (VEGF), a potent pro-angiogenic factor. In colorectal cancer a close correlation exists between TF, VEGF and microvessel density (MVD), a measure of angiogenesis, but this has not been evaluated in breast cancer. Thus far, our studies using immunohistochemistry on sections of normal human breast, benign and pre-malignant hyperplastic tissue, pre-invasive and invasive breast cancer specimens have shown a significant increase in MVD between normal and hyperplastic/pre-invasive breast cancer tissue and between pre-invasive and invasive carcinomas. Both increases in MVD were associated with a significant increase in VEGF expression in breast epithelial and tumour cells respectively. In contrast, TF was not expressed in normal, hyperplastic breast epithelial or pre-invasive cancer cells, but was expressed in approximately 45% of invasive cancer specimens. VEGF was expressed in the endothelium of normal breast, whilst TF was not. However, both VEGF and TF were expressed in endothelial cells in hyperplastic breast and in pre-invasive and invasive carcinomas. These data indicate that angiogenesis is initiated at the earliest signs of dysplasia and increases rapidly between pre-invasive and invasive cancer. VEGF and TF expression patterns suggest these factors play a significant role in this process and in vitro studies are underway to identify the relative contribution of each of these factors to tumour growth and angiogenesis.

Uveal melanomas of the choroid and ciliary body are aggressive tumours causing the death of approximately 50% of patients. In contrast, iris melanomas only infrequently metastasize; why these differences exist is not known. The local environment can regulate cancer growth and development, and for uveal melanoma we have investigated whether the aqueous and vitreous can alter their behaviour and have an important role in their regulation. Cultures of uveal melanoma were exposed to aqueous and vitreous and the effects investigated using invasion and proliferation assays. The vitreous universally promoted uveal melanoma invasion, whilst the aqueous mainly had no effect or was inhibitory. Proliferation was not significantly altered as a result of exposure to the aqueous or vitreous. Chemi-arrays were performed to determine which regulatory factors might influence the process. High expression was found for some regulators including TIMP-2, TIMP-3 and TGF-b2, which have been shown to regulate uveal melanoma behaviour. The findings suggest that in situ uveal melanoma development reflects an interaction between the tumour and the environment of the eye. Exposure to the aqueous would therefore contribute to the benign nature of iris melanomas, whilst potential interaction with the vitreous appears to promote the aggressive behaviour of posterior uveal melanomas.

SECTION OF INFECTION, INFLAMMATION AND IMMUNITY
Academic Unit of Pathology

Development of a macrophage-based system to target therapeutic viruses to prostate cancer
Profs C. Lewis, N. Maitland, F. Hamdy and N. Brown

This project is designed to use human macrophages as gene therapy vehicles to target the synthesis and delivery of therapeutic human adenoviruses to hypoxic regions of prostate tumours. The main advantages of this cell-based delivery system are the bypassing of the liver, the principal target for intravenously administered viral vectors, and the delivery of large quantities of virus into primary and/or metastatic tumours after intravenous injection. To achieve this, macrophages are co-transfected with a hypoxically activated E1A/B plasmid and an E1A/B deficient adenovirus containing a therapeutic gene under the control of a prostate-specific promoter (eg. PSA, PSMA). When such co-transfected cells reach an area of hypoxia, the E1A/B proteins are expressed, thereby activating replication of the adenovirus. The virus is then released by the host macrophage and infects neighbouring tumour cells. The PSA promoter-driven therapeutic gene is then expressed. We have shown that this takes place both in co-transfected human macrophages exposed to hypoxia (but not normoxia) in vitro, and when such cells are allowed to migrate into the central, hypoxic areas of human prostate (PC3) tumour spheroids. This novel approach employs three distinct levels of tumour-specific targeting, the homing of the macrophages to tumours, the synthesis and release of therapeutic adenovirus only in hypoxia tumour areas, and the restriction of therapeutic gene expression to prostate tumour cells.

Novel use of hypoxia response promoters in Salmonella to target genes to hypoxic
Profs C. Lewis and J. Green, Drs J. Harmey and R. Ryan.

Hypoxic areas are a hallmark feature of tumours but are difficult to access (and thus treat) by conventional therapies due to the poor vascular supply. This has prompted the development of alternative approaches to target therapeutic genes to destroy these otherwise inaccessible areas. We have developed a novel strategy for targeting gene therapy specifically to these sites using a non-pathogenic strain of the bacterium, Salmonella typhimurium, which accumulates and thrives in hypoxic sites in diseased tissues. In these bacteria we have placed either a reporter (lacZ) or three therapeutic genes (ie. the two anti-angiogenics, alphastatin and endostatin or a powerful cytotoxin) under the control of a highly hypoxia-responsive bacterial promoter, FFP* This restricts expression of these genes to sites of tumour hypoxia. In the first year of this project we have constructed stable expression plasmids with these reporter/therapeutic constructs, and shown that all three are functional following there release by hypoxic bacteria in vitro. We are currently conducting in vivo experiments to test the efficacy of this approach in vivo, using a murine syngeneic mammary tumour model (4T1). In these, tumour-bearing have been given Salmonella containing these 3 therapeutic genes under the regulation of FFP+ (or a control promoter). We are currently assessing bacterial distribution, tumour cell death, tumour growth and tumour blood vessel density in these mice.

Macrophage responses to hypoxia: relative contribution to tumour angiogenesis and progression.
Prof C. Lewis, Drs C. Murdoch and R. Johnson.

Recent studies have shown tumour-associated macrophages (TAM) play a crucial part in driving tumour angiogenesis and metastasis. A key microenvironmental factor that promotes angiogenesis in tumours is hypoxia (the low oxygen tension found at multiple sites in tumours). We showed recently that TAM gather in such hypoxic tumour sites where they are stimulated by hypoxia to express a pro-angiogenic phenotype. This is thought to be mediated by their upregulation of the hypoxia-inducible factors (HIFs) 1 and 2, which in turn trigger the expression of a wide array of pro-angiogenic and pro-metastatic genes. However, the exact role of each form of HIF in this pattern of hypoxic gene expression has yet to be outlined. We have successfully devised a way to block the expression of HIF-1 and HIF-2 mRNA in primary human macrophages (using transient traanfection with siRNA), and are currently using Affymetrix cDNA arrays to examine the effect of this on hypoxic gene expression in these cells. We have also discovered that hypoxic macrophages express elevated levels of Tie-2 the cell surface receptor for angiopoietins. This was hitherto thought to be restricted to endothelial cells. Moreover, we then found that the pro-angiogenic cytokine, angiopoietin-2 (which is expressed by hypoxic tumour cells) exerts profound inhibitory effects on hypoxic macrophages and essentially switches off their ability to perform anti-angiogenic functions (eg to release the potent anti-angiogenic cytokine, IL-12). This work has been accepted for publication in the Journal of Immunology (Murdoch et al. in press). Together these studies will show how tumour hypoxia stimulates macrophages to perform various pro-tumour functions.

Gangliosides are small sugary molecules present on a range of human cells, but present in higher numbers on many tumour cells. Ganglioside based vaccines have potential application for therapy of a variety of tumours, but tend to be very poorly immunogenic, and hence methods are required to induce stronger immune responses to these so-called therapeutic vaccines. We have received funding from YCR for a pilot study to address this issue.

We have shown in other work that conjugates of vaccine antigens with antibodies able to bind an immune cell surface molecule called CD40 (CD40mAb), are very highly immunogenic. The pilot study is aimed at producing conjugates of CD40 mAb and ganglioside antigens, with the end result hopefully being that the resultant conjugates are much more immunogenic. Ganglioside based vaccines have potential applications for therapy of a variety of tumours. Chemical conjugates produced between the gangliosides and the antibody must retain ganglioside antigen recognisable to the immune system, and also CD40 binding activity. We have now produced conjugates of gangliosides with CD40 antibodies, and preliminary studies indicate enhanced immune responses.

Improved idiotype vaccines for cancer
Dr A.W.Heath and Dr J.Carlring

Antibody idiotypes present on the surface of certain tumours of the immune system, such as B cells lymphomas, are highly tumour specific antigens. Vaccines derived from tumour idiotypes have had some success in clinical trials, but their success has been limited by the poor immunogenicity of the vaccine. We have shown that CD40 antibodies are very potent adjuvants, enhancing immune responses against a variety of antigens attached to them. This project involves the production of a vaccine with tumour idiotype and CD40 binding antibody within the same construct. This approach should very strongly enhance the immune response against the tumour idiotype. and ultimately the efficacy of idiotype vaccines.

DEPARTMENT OF BIOMEDICAL SCIENCES

Regulation of Pluripotency in Malignant Stem Cells
Prof P. W. Andrews

Teratocarcinomas are a subset of germ cell tumours, comprising a mixture of differentiated cell types and undifferentiated embryonal carcinoma (EC) cells. Undifferentiated EC cells constitute the malignant stem cell component of the tumour, whilst their differentiated derivatives display a limited proliferative capacity and are non-tumorigenic. Aberrations which promote EC self-renewal in preference to differentiation may play an important role in germ cell tumour progression. A further understanding of the pathways involved in stem cell self-renewal and differentiation is important in order to enhance our current knowledge of tumorigenesis and aid the development of novel therapies. The cyclin dependent kinase inhibitor (CDKI) protein p27KIP1 is a negative regulator of cell cycle progression, whose expression is frequently lost in a number of tumour types. Differentiating cells often contain elevated p27KIP1 protein levels, suggesting an additional role in the control of cellular differentiation. An aim of this present study is to investigate further the mechanisms by which p27KIP1 mediates differentiation in EC cells. Recently, however, we have also found that, on prolonged passage, human embryonic stem (ES) cells may acquire chromosomal abnormalities that mirror those found in EC cells from GCT. It seems likely that these common chromosomal abnormalities arise because of a common selection pressure for variant cells that undergo self renewal at the expense of differentiation. They therefore offer a novel approach to identifying the key genes that may influence the progression of GCT and we are extending our study to identify these.