Wilms tumour is a childhood kidney tumour with an incidence of I in 10,000. I investigated a child with bilateral, young-onset Wilms tumour and a de novo 5:6)(q2l;q2l) translocation. Next generation sequencing localized the breakpoints to within 1.7 kb. Using long-range PCR, I identified that HACEI was transected by the 6q breakpoint and was therefore a credible candidate Wilms tumour predisposition gene. I screened HACEI for mutations in 450 Wilms tumour cases and identified one child with a truncating HACEI mutation. These data suggest that HACEl is a rare Wilms tumour predisposition gene. DICERl mutations cause familial pleuropulmonary blastoma (PPB), a rare childhood lung tumour. I sequenced constitutional DNA from a total of 823 cases with various tumour types. I identified DICERL mutations in 19 cases, from a subset of tumour types; 1/114 with PPB,2l3 with cystic nephroma, 417 with ovarian Sertoli- Leydig type tumours, 11243 with Wilms tumour (this individual also had Sertoli- Leydig tumour), l/l with globe medulloepithelioma (this individual also had PPB), 1/86 l/172 medulloblastoma and 11172 with germ cell tumour. Analysis of tumours mutation-positive individuals demonstrated retention of the wild-type allele. I investigated the inheritance in 17 families and identified mutations in 25 relatives; 17 of which were unaffected. Constitutional DICERI haploinsufficiency thus predisposes to a range of tumours, making a substantial contribution to PPB, cystic nephroma and ovarian Sertoli-Leydig tumours. Most mutation carriers are unaffected indicating that the risk of tumours is modest. institutional mutations in PTCHI and SUFU are associated with childhood medulloblastoma. I investigated the contribution of constitutional PTCHI and SUFU mutations to medulloblastoma by mutational analysis in three familial pedigrees and 83 sporadic cases. I identified no mutations in the familial cases and two truncating SUFL mutations in the sporadic cases. These data indicate that familial medulloblastoma is genetically heterogeneous and that inactivating SUFU mutations cause;:rse approximately 2-3%o of sporadic medulloblastomas

The aims of this study were to undertake the first comprehensive in vitro, in vivo and clinical investigation into the effects of the PARP-1 inhibitor, AG014699, in human cancers defective in homologous recombination (HR) DNA double strand break (DSB) repair. HR deficient cells were 9-fold more sensitive to AG014699 than HR proficient cells (mean LC50 = 3.26 μM vs. 29.68; P < 0.0001), confirming the theory of synthetic lethality. BRCA1 methylated UACC3199 breast cancer cells were also sensitive to AG014699 with mean LC50 significantly lower than the HR proficient cells (7.6 μM vs. 29.68; P = 0.002). AG014699 inhibited PARP activity by > 95% and induced DNA DSBs in all 11 cell lines studied. Evidence of HR (by Rad51 foci) was observed only in cells with functional BRCA1/2. A prolonged schedule of AG014699 (10 mg/kg daily for five days of a seven-day cycle for six cycles) more effectively delayed the growth of BRCA2 mutated xenografts than a ten day AG014699 schedule (tumour growth delay (TGD) = 27.5 vs. 12.5 days; P = 0.02). AG014699 significantly delayed UACC3199 tumour growth compared to untreated controls (mean time to relative tumour volume 5 = 35.8 vs. 25.2 days; P = 0.05); confirming in vitro findings that BRCA1 methylated cancer cells are sensitive to PARP inhibition. Clinical trial data from 38 patients demonstrated that AG014699 is non-toxic and efficacious with a clinical benefit rate of 34%. Higher baseline PARP-1 activity was associated with response to AG014699. The major findings of these studies are: the confirmation of the selective cytotoxicity of PARP inhibitors in BRCA mutated cancers; the results in UACC3199 cells which suggest that cancers with other HR defects could benefit from single agent PARP inhibitors, and finally the concept that length of exposure to (not just degree of) PARP inhibition is important for single agent anti-tumour activity. Furthermore, these data have formed the basis for a major amendment to the clinical trial; the result of which is eagerly awaited.

Nek2 and CDK2 are serine/threonine protein kinases that are implicated in cell cycle control and cancer. The Nek family of enzymes contains 11 known serine/threonine protein kinases (Nek1-11) and are involved in mitotic cell cycle control. There is evidence that of these 11 kinases, Nek2, Nek6, Nek7 and Nek9 play an important role in the regulation of mitotic events through microtubule control. Particular interest has been placed upon Nek2, which has been shown to have a critical role in mitosis, assisting centrosome disjunction. 6-Ethnyl purine 39, was identified as a submicromolar irreversible inhibitor of Nek2 (IC50 = 150 nM), exhibiting good selectivity over other members of the Nek family. It was believed that this compound formed a triplet of H-bonds with the amino acids of the Nek2 ATP-binding site hinge region via the 2-amino N-H, the purine N3 and the imidazole N9H, allowing an initial noncovalent binding interaction, which facilitates subsequent covalent modification. To validate this binding motif and to act as control compounds, N-methylated purines 52-54 were synthesised, along with the 2-phenoxypurine 55 and the 2-benzylpurine 56. These compounds were all essentially inactive in the Nek2 inhibition assay, which demonstrated that the purine 2-amino N-H group and imidazole N9H were essential for non-covalent binding interactions. 5 To investigate the influence of the purine heterocycle upon Nek2 inhibitory activity and ethynyl group reactivity, heterocyclic derivatives of 39 were synthesised. Deazapurines 87 and 88 were weakly active in the Nek2 inhibition assay (IC50 = 30 μM and 24 μM, respectively), whilst pyrazolopyrimidine 89 and triazolopyrimidine 90 had sub-micromolar activity comparable with the parent purine 39. The 5- formylpyrimidine 101 and triazine 91 were modest inhibitors of Nek2 (IC50 = > 10 μM and 17 μM, respectively). By contrast, pyrimidine 95 (10% inhibition at 100 μM) demonstrated weak Nek2 inhibitory activity. Quantitative 1H NMR (q1H-NMR) kinetic studies were performed to model the reaction of ethynyl-functionalised heterocycles with Cys-22 at the Nek2 ATP-binding site. Compounds 54, 56, and 88-89 were reacted with N-acetylcysteine methyl ester in DMSO-d6 at 24 oC under pseudo first order reaction conditions. The most reactive compounds were triazolopyrimidine 90 and pyrazolopyrimidine 89. The 2- benzylpurine 56 and N-methylanilinopurine 54 had moderate reactivity, comparable with the parent purine 39. The least reactive compound was pyrrolopyrimidine 88, which was approximately 150-fold less reactive than purine 39. 6 CDK2 is a member of the cyclin-dependent kinase (CDK) family of enzymes and a mediator of cell cycle progression. Of the 11 known human CDKs, four (CDK1, CDK2 CDK4 and CDK6) have been directly implicated in cell cycle regulation. Mutation of genes coding for CDKs are common in a variety of cancers, making CDKs attractive chemotherapeutic targets. Purines 249 and 37 were identified as moderate (IC50 = 17 μM) and potent (IC50 = 5 nM) ATP-competitive inhibitors of CDK2, respectively. The focus of this research was to determine whether subtle changes to the core heterocycle would allow retention of CDK-inhibitory activity. CDK2 inhibitors based on the pyrazolopyrimidine and triazolopyrimidine heterocycles (256, 252, 257 and 253) were of comparable potency with the corresponding purines (249 and 37). The most potent compound was triazolopyrimidine 253 with an IC50 value of 3 nM against CDK2. Interestingly, the pyrrolopyrimidine 255 proved only weakly active (IC50 > 100 μM), whereas the functionalised pyrrolopyrimidine 251 had potent CDK2 inhibitory activity (IC50 = 26 nM). Finally, imidazopyridine 254 exhibited activity comparable with purine 249; however, imidazopyridine 250 was found to be approximately 30-fold less potent (IC50 = 140 nM) than the parent 37, for reasons that remain unclear. N.B. diagrams not included in this abstract

Oxaliplatin based chemotherapy has demonstrated remarkable efficacy in down staging colorectal liver metastases such that patients initially considered to have inoperable disease are able to undergo a potentially curative resection. In addition the use of neoadjuvant Oxaliplatin based chemotherapy has been shown to improve progression free survival following liver resection. Taken together this means that ever increasing numbers of patients are presenting for liver resection having received multiple cycles of chemotherapy. Whilst this approach has many advantages the use of pre-operative chemotherapy has been associated with the development of sinusoidal obstruction syndrome (SOS) in the liver parenchyma in up to 40% of patients. It is thought that the presence of SOS significantly increase the risks associated with major liver resection. More recent data also suggests that the presence of SOS within the liver may result in poorer disease specific outcomes in the long term and in particular a higher risk of early intra-hepatic recurrence. At present the pathogenesis of SOS in this context is not understood and no treatment exists to either prevent its development or reverse the histological changes in the liver associated with it. The aim of the current study was to develop a reproducible in vivo experimental model of Oxaliplatin induced SOS and to interrogate this to identify the pathophysiological mechanisms which underpin its development. With this knowledge it was hoped that potential therapeutic strategies could be suggested to ameliorate the development of SOS in patients treated with Oxaliplatin based chemotherapy. C57BL/6J mice treated with weekly intraperitoneal injections of Oxaliplatin and 5- FU/Leucovorin for 5 weeks develop histological changes of SOS when maintained on a iv purified, but not chow, diet. This is associated with increased expression of key matrix remodelling genes within the liver parenchyma such as MMP2, MMP9, TIMP1, TGFβ and Procollagen I. The development of these gene expression changes is accelerated in the presence of tumour within the liver perhaps as a consequence of increased production of inflammatory mediators such as CXCL1. The presence of SOS is associated with a dramatic increase in expression of the serine protease family member PAI-1 which is involved in a variety of processes including matrix remodelling, thrombus formation and cellular senescence. Immunohistochemistry revealed endothelial cells in areas of sinusoidal injury stained positive for the cell cycle inhibitor p21CIP1/WAF1 in keeping with senescence in these cells. This process was associated with depletion of hepatic glutathione and decreased expression of the antioxidant transcription factor NRF2 suggesting a role for oxidative stress in the pathogenesis of SOS. To explore this further the experiment was repeated but this time using dietary supplementation with either the thiol donor N-Acetylcysteine (NAC) or the NRF2 activator butyrated hydroxyanisole (BHA) alongside FOLFOX treatment. Whilst supplementation with NAC had no effect on the development of SOS its development was completely prevented by supplementation with BHA suggesting that NRF2 activating antioxidants may be a useful therapeutic strategy in preventing the development of SOS. In conclusion this study has described the first reproducible experimental model of Oxaliplatin induced SOS which accurately reflects the pathogenesis of the disease in humans. Through interrogation of this it has been possible to identify therapeutic strategies which may be of value in preventing the development of SOS in patients with colorectal liver metastases.

The NF-κB family of transcription factors can induce or repress target gene expression by binding DNA as homo- or hetero-dimers. In some cancer cells the p52 (NF-κB2) subunit, which is derived by proteolytic processing of its precursor p100, regulates the expression of genes having a role in cell proliferation, such as Cyclin D1 and is also involved in the regulation of G2/M phase. Depletion of p100/p52 by siRNA leads to an increase of cells in G2/M phase and defects in mitosis. These defects in mitosis can be visualised as aberrant chromosome segregation, disruption of the microtubule network and poor alignment of chromosomes on the metaphasic plate. Moreover, p100/p52 depletion results in an increase in multinucleate cells, as well as aberrant centrosome structures. To elucidate this mechanism, I have investigated p100/p52 regulation of various genes involved in the cell cycle and centrosome duplication. I discovered that p100/p52 siRNA depletion reduces the expression of a number of cell cycle regulators, including Polo-like kinase 4 (PLK4) or Spindle assembly abnormal protein 6 (SAS6), a member of the SAS proteins family. PLK4, in conjunction with the cyclin-dependent kinase CDK2 and the PLK4 effector SAS6, is a key regulator of centriolar duplication. I demonstrated by reporter- gene and chromatin-immunoprecipitation analyses that the PLK4 promoter is a direct target for multiple NF-κB subunits and its activity depends upon NF-κB expression. Moreover PLK4 and p100/p52 mRNA and protein expression are cell cycle regulated and NF-κB subunits bind the PLK4 promoter in a cell cycle dependent manner. These results reveal PLK4 as a new NF-κB target, providing a direct link between NF-κB and centrosome duplication with implications for the role of p52 in tumorigenesis.

Poly(ADP-ribose) polymerase 1 (PARP1) is a DNA-binding enzyme involved in DNA repair by the base-excision pathway. The inhibition of PARP1 is being investigated as a cancer treatment. Rucaparib (CO338) is a potent PARP 31 inhibitor currently in Phase II clinical development. In this thesis P in vivo MR Spectroscopy (MRS) and Dynamic Contrast Enhanced (DCE) MRI were used to study acute effects of rucaparib on energy metabolism and tumour vasculature. 1 31 18 18 Ex vivo H and P-MRS, and in vivo [ F]FLT and [ F]FDG-PET, were used to study effects of treatment with rucaparib on tumour metabolism and proliferation. A2780 and SW620 tumours implanted in mice were scanned in a horizontal Varian 7T MR system. Two i.v. injections of the MRI contrast agent gadoteridol were given 90 minutes apart with dynamic phosphorus MRS acquired following the injection of rucaparib, temozolomide or both drugs in combination. The 18 18 same tumours were evaluated by [ F]FLT- and [ F]FDG-PET after 5 daily treatments with rucaparib, temozolomide or the combination, and the livers of PARP1 knock out (KO) and wild type (WT) mice treated in a similar manner 1 31 were analysed by ex vivo H and P-MRS. Tumour uptake of gadoteridol changed significantly after treatment with hydralazine and higher doses of rucaparib in SW620 tumours, and following 31 hydralazine and 1mg/Kg of rucaparib in A2780 tumours. P-MRS studies revealed an increase in the inorganic phosphate (Pi) to β-NTP ratio, consistent with impairment of tumour energy metabolism following hydralazine treatment. 18 [ F]FLT-PET demonstrated a significant reduction in the SUV values in the 18 rucaparib/temozolomide combination group in SW620 tumours, and [ F]FDG- PET revealed a non-significant reduction in tumour metabolism in A2780 1 tumours. H ex vivo MRS demonstrated an increase in the liver NAD concentrations after treatment with rucaparib, but a decrease following the treatment with temozolomide, regardless of the PARP1 status. Together, these pre-clinical imaging studies have shown that MR can be used 18 to investigate the acute anti-vascular effects of rucaparib, that [ F]FLT-PET predicted subsequent changes in tumour volume following combined rucaparib 1 and temozolomide treatment, and that ex vivo H-MRS can be used in mechanistic studies of PARP inhibition. Both MRI/MRS and PET are potential pharmacodynamic and surrogate response imaging biomarkers for PARP inhibitors.

Oral precancer (OPC) is a collective term for a number of disorders that may precede oral cancer. Treatment is aimed at preventing malignant transformation however, this is complicated by a lack of robust evidence concerning both treatment effectiveness and future cancer risk. Uncertainties surrounding prognosis and treatment options might be expected to impact on a patient’s experience of their disease, as well as creating challenges for their management. The aim of this research was to explore the experience of OPC through the eyes of the patient and clinician to assess the impact of living with oral precancer and enable the identification of opportunities to improve outcomes. The project comprised two qualitative studies, each employing semi-structured interviews. 28 patients with OPC, were recruited for study A, while 11 Oral and Maxillofacial Consultants were involved in study B. Data collection and analysis was iterative, following the principles of the ‘constant comparative’ method (Glaser 1965). Data collection stopped when data saturation was achieved. The data were analysed using thematic analysis. The results indicated that during the diagnosis and management of OPC, clinicians were faced with challenges. These included: communicating a diagnosis, (particularly in terms of terminology), conveying risk meaningfully, meeting patients’ additional information needs, encouraging behaviour change and making treatment decisions. The patient data indicated that for some, OPC represents a devastating diagnosis leading to feelings of fear and uncertainty impacting significantly on the individual’s life. In addition, analysis also allowed a disease journey to be mapped and directly related to the findings from the clinician group thereby indicating opportunities where changes in practice may improve patient care. These points included: the diagnosis, where understanding terminology and comprehending risk were problematic, following a diagnosis, where meeting information needs was a challenge and during the management and review stages when treatment decisions were made and carried out.

Scaffold attachment factor B1 (SAFB1) and SAFB2 are oestrogen receptor (ER) corepressors that bind and modulate ER activity through chromatin remodelling or interaction with the basal transcription machinery. However, little is known about the fundamental characteristics and function of SAFB1 and SAFB2 proteins in breast cancer. In this study, an investigation of the characteristics and function(s) of SAFB1 and SAFB2 was undertaken; their expression profile was first assessed in ER-positive (MCF-7) and ER-negative (MDA-MB-231) breast cancer cell lines. Results show that SAFB1 and SAFB2 are themselves regulated by an active metabolite of oestrogen, 17β-oestradiol, in both ER positive and ER negative breast cancer cells. Using a combined approach of RNA interference and gene expression profile studies, 12 novel targets closely linked to tumour progression were identified for SAFB1 and SAFB2. Expression levels of the following genes, CDKN2A, CLU, ESR1, IGFBP2, IL2RA, ITGB4, KIT, KLK5, MT3, NGFR and SPRR1B increased while IL-6 expression and secretion decreased when cells were depleted of SAFB proteins. This observation supports their primary role as transcriptional repressors with SAFB2 playing a prominent role in transcriptional regulation in MDA-MB-231 cells. This study has also established a novel link between SAFB proteins and ITGB4 and IL-6 expression. Both SAFB proteins have an internal RNA-recognition motif but little is known about the RNA-binding properties of SAFB1 or SAFB2. To investigate this, the concluding part of the project utilised crosslinking and immunoprecipitation (CLIP) coupled with high-throughput sequencing. This experimental approach enabled a transcriptome-wide mapping of SAFB1 protein-RNA interactions in breast cancer cells. SAFB1 crosslink sites are significantly enriched in ncRNAs, particularly within snRNAs. A putative RNA-binding motif for SAFB1 that contains adenine-rich sequences, highly similar to the RNA-binding motifs for SR proteins was also identified. In summary, this study has defined the characteristics of SAFB1 and SAFB2 in both ER positive and ER negative breast cancer cell lines. It has also identified transcriptional targets and the RNA-binding ability of SAFB1 and SAFB2 in human breast cancer cells.

Background: Hepatocellular cancer [1] is increasing worldwide. The majority are detected too late for curative surgery as effective identification of the at risk population and their subsequent surveillance is inadequate. My specific aims were to study and characterise our own patients, evaluating the known and predicted biomarkers for HCC in patients with non-alcoholic fatty liver disease and exploring novel biomarker methodologies for the detection of either cirrhosis or HCC complicating it. Methods: Details of all HCC patients presenting to the Newcastle hepatopancreatobiliary (HPB) multidisciplinary team were recorded in an NHS intranet database, and a subset were consented for tissue collection. Established and candidate biomarkers were assessed in serum by western blotting and/or ELISA assay and the role of microarray analysis of HepG2 cells, and 2D-gel electrophoresis of immunodepleted serum in novel candidate biomarker identification were explored. Results: The number of HCC cases referred has increased dramatically over the last decade, as has the numbers arising in a background of non-alcoholic fatty liver disease (NAFLD). In our cohort, patients with earlier stage cancers detected by either - surveillance or incidentally had a better prognosis, but only 10% were eligible for definitive treatment. While HCC patients treated with liver transplant had an overall 5 year survival was 62.1%, the median survival of the transplant cohort was 137 months. Exploration of serum levels of Glypican 3, Squamous Cell Carcinoma Antigen-1 and follistatin were poor surveillance biomarkers, but the combination of Alpha-fetoprotein and Protrhombin induced by vitamin-K absence was encouraging. Serum proteomic analysis in a small subgroup identified four isoforms of apolipoproteins as well as CD5L as differentially expressed between patients with no cirrhosis, cirrhosis and HCC, although subsequent CD5L ELISA failed to confirm its ability to specifically detect HCC. Discussion: While the incidence of HCC is increasing, the prognosis for those affected remains poor. Biomarkers identifying both the at risk individuals and those with cancer are urgently required.

This study aimed to improve outcomes for colorectal cancer (CRC) patients in Greater Birmingham (1996-2000). A two-phase design was adopted. Phase I involved a detailed data collection exercise, which sought to capture the patient's journey from presentation through to primary surgical treatment (n=578). During phase I, it was noted that one of the hospitals showed considerable improvement in the quality of histopathology reporting by adopting a proforma based report. This finding generated the hypothesis that proforma based reporting improved the quality of histopathology reporting. This hypothesis was tested in phase II. A major finding from phase I was that 113 of CRC patients underwent emergency treatment with subsequent poor outcomes. Patient flow analysis showed that half the emergency cases were initially referred by their GP as elective cases, but delays in their pathway of care allowed acute deterioration resulting in emergency admission and treatment. This finding generated the hypothesis that changing GP referral patterns and reducing times to surgical treatment could reduce the number of emergency CRC patients. This hypothesis was also tested in phase II of the study that involved interventions in primary and secondary care to effect a change in referral practice and reduce delays. With reference to emergency CRC patients, the present study challenges two widely held viewpoints - a) that reducing delays is ineffective at improving survival and that b) the emergency CRC is fundamentally different to its elective counterpart. Given that emergency cases account for half the total mortality burden of CRC, the findings of this work merit consideration.