Promyelocytic leukaemia (PML) nuclear bodies are nuclear structures found in a variety of normal tissues and cell lines. They have been implicated in diverse human diseases. In particular, the major constituent, the PML protein, forms a fusion product with another protein in acute promyelocytic leukaemia (APL). These bodies however, also recruit over thirty different proteins with disparate functions. As such, no definite role of these bodies has been discovered, although proposed functions include gene transcription, cell cycle control and deoxyribonucleic acid (DNA) repair. This thesis describes the association of PML bodies with different genomic loci, using principally confocal microscopy and a novel statistical model. The aim was to use such associations to determine if a functional basis exists for the intranuclear pattern of PML bodies. By analyzing loci-PML body distances for different gene loci, it was found that the distance between a locus and its nearest PML body correlates with the transcriptional activity and gene density around the locus. This was confirmed when regions of specific gene activation were examined. However, using RNA-FISH (ribonucleic acid- fiuorescence in situ hybridisation) methodology and RNA interference (RNAi) knockdown studies, PML bodies were found not to be directly involved in gene transcription. Furthermore, cells in S-phase were examined in more detail, and it was found that PML bodies also associated statistically with actively replicating loci. The experiments performed suggest a non-random and functional basis for the positioning of PML bodies. This thesis proposes that PML bodies are multifunctional structures that lie predominantly in nuclear compartments of high transcriptional activity, although they also associate with regions of DNA replication. Finally, this work strengthens the model of the nucleus as a highly organised structure.

Chronic Myeloid Leukaemia (CML) arises as a consequence of the expression of a chimaeric fusion protein; p210BCR-ABL. Many publications report that p210BRR-ABL forms complexes with multiple cytoplasmic proteins which affect signalling pathways demonstrated in cell lines or transduced cells. This has been necessary because primary haemopoietic cell lysates contain a degradative activity which rapidly and permanently destroys p210BCR-ABL. We have identified that the degradative enzymes located in the cell lysosomes and have demonstrated substantial inhibition of the p210BCR-ABL-degradative activity by high pH lysis conditions. We show to the best of our knowledge, the first set of data demonstrating expression and immunoprecipitation of p210BCR-ABL and co-immunoprecipitation of adaptor proteins CBL, CRKL and GRB2. The degradative activity also affects ABL protein, preventing analysis of protein complexes of normal ABL but using the high pH lysis we have shown that normal ABL complexes with GRB2 potentially mediated via BCR or a direct association with ABL, this is different from some cell line data published which implies that proteins that complex with p210BCR-ABL do not complex with normal ABL. We also analysed complexes in two Ph-negative, acute lymphocytic leukaemia (ALL) cell lines, KG1a and HL60 cells and found ABL forms complexes with CBL, CRKL and GRB2 similar to CML. This data suggests that adaptor proteins which complex with p210BCR-ABL also form complexes with ABL in Ph- leukaemic cells and some of them form complexes in non-leukaemic cells. Thus, there may be CML-specific, leukaemia-associated and normal interactions with ABL proteins. Using confocal microscopy and a junction specific anti-BCR-ABL (b2a2) antibody we analysed the subcellular distribution of p210BCR-ABL and have shown that p210BCR-ABL is arranged in discrete foci in the cell cytoplasm in various CML cell lines and primary CML cells. Many studies have implicated CRKL as an important target of p210BCR-ABL that can be used as an indirect indicator of p210BCR-ABL protein tyrosine kinase activity. When we analysed co-localisation of p210BCR-ABL and CRKL, we found that CRKL also formed foci. However, the CRKL and p210BCR-ABL foci were completely or partially associated or separate in different regions of the same cell. Since CRKL in CML is phosphorylated by association with p210BCR-ABL, these data imply that binding of CRKL and p210BCR-ABL maybe in a state of dynamic equilibrium. Heterogeneity of protein complexes from patients in blast crisis (BC) was observed, where the p210BCR-ABL-CBL complex was absent in one CML BC patient. Effects of imatinib on protein complexes were analysed and we found that p210BCR-ABL-CRK1 and p210BCr-ABL- CBL complex dissociates over time of treatment however, the GRB2 remains in complex within the 24 hour treatment period which maybe a potential target in imatinib resistant CML cells.

Chronic myeloid leukaemia (CML) is a chronic myeloproliferative disease that is characterized by the presence of the BCR-ABL1 protein, with subsequent increased tyrosine kinase activity. The Introduction of imatinib and second generation tyrosine kinase inhibitors revolutionized CML therapy; however, prognostic factors that can predict disease outcome are still measured using scores that have been developed during the chemotherapy and IFN-α era. This raises the need for modern scores that match the current era. The EUTOS score that has been recently introduced by the ELN promised significant accuracy in determining patients’ outcomes. One drawback however is that the EUTOS score hasn’t been independently validated; hence it failed to predict outcomes in our patient sample. Knowing that the achievement of an early response is an important factor in determining outcomes, we investigated the use of molecular markers as a prognostic factor and we managed to prove that the measurement of transcript levels at 3 months is the single most important factor to determine patients’ outcomes. Current therapy has managed to convert CML into merely a chronic illness in most of the cases, therefore patients and clinicians are faced by problems that are characteristic of long term therapy, including issues of non-adherence and fertility. We managed to show that non-adherence to imatinib can lead to significant loss of responses and treatment failure. We also managed to show that an adequate response is required prior to imatinib interruption for those women who wish to conceive. Unfortunately, not all patients manage to achieve responses on imatinib, and therefore need to change to second line therapy. Funding agencies argue about the survival benefit of second generation TKIs, so here we don’t only show that second generation TKIs offer a survival benefit, but also that the responses achieved on 2G – TKIs are also durable and long lasting. Finally, we explore of the use of TKIs as third line therapy in CML. With future research directed towards the introduction of even better TKIs, especially those targeted against the T315I mutations and vaccination therapy, perhaps in the near future a cure for CML might be achieved.

Chronic myeloid leukaemia (CML) is a stem cell (SC) disorder initiated by the reciprocal translocation between chromosome 9 and 22, giving rise to the Philadelphia (Ph) chromosome and the resulting expression of the oncogenic fusion protein BCR-ABL. The current first line of treatment is imatinib mesylate (IM), a tyrosine kinase inhibitor (TKI) that competes with ATP to block ABL kinase activity, which in turn prevents tyrosine phosphorylation of downstream molecules and selectively induces apoptosis of BCR-ABL cells. However, despite excellent cytogenetic responses, only a minority of patients achieve complete molecular response (CMR). We have previously identified a population of quiescent (q) Ph+ SC found in chronic phase (CP) CML that are relatively insensitive to IM and other TKIs and which may be responsible for the molecular persistence of this disease. This population may be insensitive because TKIs do not reach therapeutic concentrations within the cell. Such resistance to classical chemotherapeutic drugs, the phenomenon of multidrug resistance (MDR), is mediated by ABC transporters. In this study we have investigated whether CML SC express the clinically relevant ABC transporters and determine their interaction with TKIs. In addition, we determined whether the inhibition of these transporters increased the efficacy of TKI against CML SC. Using CML CD34+ cells isolated from newly diagnosed patients, normal CD34+ cells and cell lines transduced with specific transporters as controls, the relative expression of drug transporters were determined in CML CD34+ cells and intracellular staining confirmed protein expression. The interaction of drug transporters with TKIs was assessed using a combination of substrate displacement assays and radiolabelled assays. The effect of transporter inhibitors 3 with TKIs on the growth and differentiation of q34+ and more mature CD34+ cells from CML patients in CP were assessed with regard to cell division, apoptosis and BCR-ABL kinase activity. When compared to normal CD34+ cells, CML CD34+ cells over-expressed ABCG2 mRNA. In contrast MDR1 expression was reduced in CML CD34+ cells and MRP1 was detected at similar expression levels in both populations. All three drug transporters were expressed at the protein levels in CML CD34+ cells. It was determined that at therapeutic concentrations (5μM) IM and nilotinib both inhibited ABCG2 and MDR1 and nilotinib also inhibited MRP1. Neither drug was a substrate for any of the transporters. In contrast, dasatinib was shown to be a substrate for ABCG2 and MRP1, but had no effect on MDR1. Therefore activity and concentration of dasatinib but not IM or nilotinib may be altered by the activity of these proteins. In keeping with their inhibitory activity, neither IM nor nilotinib demonstrated significantly increased efficacy when combined with specific ABC transporter inhibitors (FTC or PSC 833). Surprisingly, although dasatinib was a substrate for ABCG2 and MRP1, dasatinib did not further increase apoptosis, or reduce the qSC population. Therefore, although MDR1, MRP1 and ABCG2 were found to be expressed and functional in CML CD34+ cells and to interact with TKI, the co-treatment of TKIs with drug transporter inhibitors did not further increase apoptosis, reduce BCRABL kinase activity or reduce the qSC population. Therefore, modulation of individual transporter activity is unlikely to reverse the resistance of this population of cells to TKI and will not improve the clinical response to these drugs.

Cellular senescence is an end point of a signal transduction programme leading to irreversible cell cycle arresst accompanied by characteristic alterations to cell morphology, biochemical properties and gene expression profile. This phenotype can be triggered by a variety of stimuli including telomere shortening, DNA damage or activated oncogenes. Senescence is now recognised as a tumour suppressor mechanism mediated by p53 and pRB pathways which act to prevent the proliferatio of cells that are at risk of tumourigenic transformation. RUNX1 is a transcription factor essential for definitive hematopoiesis and is frequently targeted in human leukaemias by chromosomal rearrangements. RUNX1 has been also demonstrated to act as a dominant oncogene in mice and the ectopic expression of RUNX1 in murine embryonic fibroblasts has been shown to cause senescence. The central aim of this study was to investigate the mechanism of senescence induction by RUNX1 and its fusion derived leukaemogenic oncoproteins in primary fibroblasts. My work showed that RUNX1 induces a strong senescence-like response in murine and human primary fibroblasts that requires intact DNA binding, CBFB interaction and C-terminal transcriptional activation/repression domains. However, surprising differences were found between the major RUNX1 fusion oncoprotein derivatives. The N-terminal fusion protein TEL-RUNX1 fails to induce senescence despite retention of a virtually full-lenght RUNX1 moiety, while the senescence-inducing potential is exaggerated in the truncated C-terminal fusion protein RUNX1-ETO (AML1-ETO). The potential to drive senescence is retained by the deletion mutant RUNX1-ETO[]469 which lacks critical corepressor binding sites suggesting that the repression of target genes may be a primary mechanism implicated in RUNX1-ETO induced senescence. Interestingly, CBFB-MYH11 fusion oncoprotein that affects RUNX1 indirectly by targeting CBFB cn also induce senescence when ectopically expressed in human primary cells. The RUNX1 and RUNX1-ETO induced senescent phenotypes differ from archetypal H-Ras [superscript v12] as arrest occurs without a preliminary phase of proliferation and the arrested cells lack prominent foci of DNA strand breaks and chromatin condensation. Notably however, RUNX1 and RUNX1-ETO display differences in their potency and the extent of engagement of p53 and Rb effector pathways. RUNX1-ETO is highly dependent on p53 function and unlike RUNX1 drives senescence in cells lacking intact p16Ink4a. RUNX1-ETO appears to exert its unique effects through potent induction of reactive oxygen species and p38MAPK phosphorylation. These findings illustrate the heterogeneous manifestations of senescence-like growth arrest and elucidate the distinctive biology and oncogenic properies of RUNX1 and its fusion derivatives.

HA22 is an immunotoxin composed of an anti-CD22 variable fragment linked to a 38 kDa truncated protein derived from Pseudomonas exotoxin A. The mechanisms of cytotoxicity and resistance of HA22 against Acute Lymphoblastic Leukaemia (ALL) and Burkitt’s lymphoma were studied. Using a bone marrow mesenchymal cell culture assay to support ALL cell viability, I? investigated the in vitro cytotoxicity of HA22 against ALL blasts from newly diagnosed and relapsed patients. There was interpatient variability in sensitivity to HA22. There was no significant difference in HA22 sensitivity between diagnosis and relapse samples but peripheral blood ALL blasts were more sensitive to HA22 than those from bone marrow. The mechanisms of resistance to HA22 were studied, using cell lines as a model. The number of CD22 sites/ cell and the rates of immunotoxin internalisation did not affect HA22 cytotoxicity. HA22 mutants with resistance to lysosomal degradation and enhanced targeting to the endoplasmic reticulum had improved cytotoxicity. The role of apoptosis pathways proteins in HA22-mediated cell death was studied. Their role is complex but raised levels of the anti-apoptotic pathway protein Bcl-2 were found in the most resistant NALM6 cell line. Penetration of HA22 into Burkitt’s lymphoma masses was studied using a flow cytometric based method. HA22 rapidly penetrated into the lymphoma masses, however a barrier to further uptake is present which could not be overcome by the addition of adriamycin or taxol in the murine xenograft model. The ability of Acute Myeloid Leukaemia (AML) blasts to create an immunosuppressive niche was investigated using a cell line model and primary patient samples. AML blasts suppress T cell proliferation through altered arginine metabolism, dependent on the enzymes arginase II and iNOS. Small molecule inhibitors to arginase and iNOS restored T cell proliferation in vitro. AML further enhances its immunosuppressive niche by transforming surrounding monocytes into an M2-immunosuppressive phenotype, in an arginase dependent manner. The immunomodulatory protein Serum Amyloid A (SAA) was secreted by AML blasts, and leads to AML chemotaxis, IL-1production, and release of S100A9 protein. Finally, invariant Natural Killer T cells (iNKT) were shown to be cytotoxic to some AML blasts, in the presence of Galactosylceramide, and thus able to restore T cell proliferation. The results provide a strong rationale for the clinical testing of these novel immunotherapeutic and immunomodulatory strategies in patients with haematological malignancies.

Acute myeloid leukaemia is a clonal disorder characterised by recurrent chromosomal translocations. One of the commonest, is the t(8;21) which results in part of the AML1 gene being juxtaposed to most of the ETO gene with the resultant chimeric protein, AML1-ETO, acting predominantly as a transcriptional repressor. Despite the extensive literature available, the exact mechanism by which the chimeric protein results in AML has not been fully elucidated. By using exon arrays and high throughput sequencing as tools it was hoped to gain further insights into the molecular basis of this disease. Gene expression profiling using the exon arrays highlighted molecular pathways and specific genes that play a key role in the pathogenesis in t(8;21). Exon arrays were also used to profile individual exon expression of the ETO gene. This demonstrated that the genomic breakpoint of ETO in the t(8;21) is variable between different patients. This technique also resulted in the discovery of a new exon in the ETO gene. This novel exon results in formation of alternative transcripts of AML1-ETO and was shown in mouse models to play a key role in leukaemogenesis. Chromatin immunoprecipitation followed by high throughput sequencing revealed novel aspects of AML1-ETO binding. A number of novel genes that bind AML1-ETO were recognized and in conjunction with the expression data, a number of hypothesis on how AML1-ETO binding effects gene expression are made.

Mutations in the ATM gene have previously been identified in CLL tumours. In this project, I have demonstrated that their detection would have prognostic value. With a prevalence of 12%, ATM mutations represent the commonest single gene defect to be detected in CLL tumours and they identified a subgroup of CLL patients that had a significant reduction in both treatment free and overall survival. Furthermore, ATM mutations provided prognostic information that was independent of age, clinical stage, the mutation status of the IGVH genes and TP53 mutations. The temporal acquisition of the ATM mutations and their relationship with loss of an ATM allele via a chromosomal 11q deletion provides clues into their mechanism of action. There was only a partial correlation between CLL tumours with mutations in the ATM gene and those with a chromosome 11q deletion. In certain cases, the ATM mutations represented germ-line changes and in others were acquired very early in the disease course raising the possibility that they might contribute to the initial clonal transformation process. However, in some CLL tumours, the ATM mutations had been acquired after the development of the tumour clone during disease progression indicating that there may be a step-wise acquisition of ATM allelic defects during the ontogeny of CLL. The ATM protein is the key coordinator of the cellular response to DNA double strand breaks. In this study, I showed that bi-allelic defects in the ATM gene lead to deficient ATM dependent responses, including the up regulation of p53, following both ionising irradiation and also treatment with the chemotherapeutic drug, Fludarabine. Thus an important mechanism accounting for the poor outcome in CLL patients with ATM mutations is likely to relate to chemo-resistance. Interestingly, there were differential responses to DNA damage with both irradiation and fludarabine amongst the category of tumours with an 11q deletion according to the status of the remaining ATM allele. Therefore, ATM mutations can stratify tumours with a chromosome 11q deletion into two functional subgroups. The identification of CLL tumours with ATM mutations would therefore predict those patients that will have a poor clinical outcome and be both more likely to require early treatment for their disease. Patients whose tumours had bi-allelic ATM defects will be expected to have deficient responses to DNA damaging chemotherapeutic drugs, while those with mono-allelic ATM defects might identify a group in whom the use of DNA damaging agents could provide selective pressure for the emergence of sub-clones that have subsequently acquired bi-allelic ATM defects.

No description available.

Many haematological malignancies are associated with widespread alterations of the transcriptional and epigenetic programmes. Changes in DNA methylation provide the clearest example of epigenetic changes, but the mechanism(s) underlying such changes is unknown. To investigate this I studied DNA methylation across an ~80kb segment of the genome which is not known to be mutated in haematological malignancies. Methylation was perturbed in 35-100% of samples of DNA from individuals with a wide range of haematological malignancies but not in non-malignant haematological disorders. DNA methylation was comprehensively assessed by Southern blot analysis, classical bisulphite sequencing and using a newly developed capture bisulphite sequencing protocol. The results were also compared with analysis by MeDIP, an immunoprecipitation-based technique. These analyses provide methylation status at various levels including individual CpG resolution. This showed both gain and loss of methylation at CpG dinucleotides. Of interest, hypomethylation was most frequently seen in intergenic regions corresponding to transcription factor binding sites and areas of increased chromosome accessibility. These observations suggested that hypomethylation of the genome in haematological malignancies could arise from aberrantly expressed DNA binding proteins which, recruited to sequences in regions of open chromatin, would protect the underlying CpG dinucleotides from the methylation machinery. This, in turn, could lead to passive demethylation accumulating with increasing cell divisions. This hypothesis was tested with electrophoretic mobility shift assays using oligonucleotides representing the DNA underlying one such region. This showed that, compared to nuclear extracts from the lymphocytes of normal individuals, those from patients with CLL were enriched for a protein which binds to oligonucleotides containing the underlying sequence. Using a mass spectrometry approach, I identified a variety of proteins that may bind such regions and account for their passive demethylation in haematological malignancies.