DNA is a highly unstable molecule. Endogenous souces of DNA damage, such as reactive oxygen species (ROS), can cause DNA damage and it has been estimated that 20000 lesions occur in a cell per day. BER is the major pathway for the repair of these lesions and therefore maintains genome stability, thus preventing the development of human diseases such as neurodegenerative diseases and cancer. Therefore, if BER cannot accomplish the repair, accumulation of DNA damage occurs, triggering different cellular responses, such as cell cycle delay and senescence. The ARF tumour suppressor protein, the gene of which is frequently mutated in many human cancers, plays an important role in the cellular stress response by orchestrating upregulation of p53 protein. Moreover, ARF expression is upregulated in senescent cells, suggesting that ARF induction might be triggerred in response to persistent DNA damage. Although ARF has been reported to be important in the regulation of proteins involved in the DNA damage response, its role is still controversial. Here, it has been shown that ARF gene transcription is induced by DNA strand breaks (SBs) and that ARF protein accumulates in response to persistent DNA damage generated by disabling BER. These data suggest that PARP1-dependent poly(ADP-ribose) synthesis at the sites of SBs initiates DNA damage signal transduction by reducing the cellular concentration of NAD⁺, thus inhibiting SIRT1 activity and consequently activating E2F1-dependent ARF transcription. These findings suggest a vital role for ARF in DNA damage signalling, and furthermore explain the critical requirement for ARF inactivation in cancer cells, which are frequently deficient in DNA repair and accumulate DNA damage.

DNA polymerase η (pol η), a crucial component of the cellular translesion synthesis pathway, allows cells to bypass and thereby temporarily tolerate DNA damage. Inherited deficiency of pol η, as reported in the variant form of xeroderma pigmentosum, predisposes to UV light-induced skin cancers. To date, pol η is the only DNA polymerase shown to exhibit a causal link to the formation of cancers in humans. However, the role of pol η in the cellular response to forms of DNA damage other than UV-induced lesions is largely unknown. In the first part of this thesis, it is shown that cells deficient in pol η are resistant to ionising radiation. Deficiency in the polymerase was associated with accumulation of cells in S phase of the cell cycle. Cells deficient in pol η demonstrated increased homologous recombination-directed repair of DNA double-strand breaks created by ionising radiation, and depletion of the homologous recombination protein X-ray repair cross-complementing protein 3 (XRCC3), abrogated the radioresistance observed in pol η-deficient cells compared to pol η-complemented cells. These findings suggest that homologous recombination mediates S phase-dependent radioresistance associated with pol η-deficiency. In the second part of this thesis, it is shown that pol η-deficient cells have increased sensitivity to the chemotherapeutic compound, oxaliplatin, compared to pol η-deficient expressing cells, but not to the drug 5-fluorouracil that is usually administered in combination with oxaliplatin in the clinical setting. Despite the importance of pol η for cellular survival following exposure to oxaliplatin, the drug did not upregulate the enzyme after either short-term or long-term exposure. Inhibition of pol η activity by siRNA-mediated knockdown of the protein sensitised cells to oxaliplatin treatment, and partially reversed acquired resistance in oxaliplatin-resistant tumour cell lines. These data suggest that pol η is an interesting target whose function can potentially be interfered with to optimise oxaliplatin-based chemotherapy. In the third part of this thesis, clinical samples obtained from oesophageal cancer patients before and after treatment with oxaliplatin-containing chemotherapy were analysed for POLH mRNA levels encoding pol η protein. Malignant tissue specimens obtained before treatment demonstrated a significantly higher level of POLH mRNA than matched normal oesophageal tissue samples. Contrary to the preclinical data, high POLH mRNA expression before therapy was shown to correlate with increased overall and disease-free survival of the patient cohort in the clinical trial. Additionally, patients with high POLH mRNA-expressing cancers had better therapeutic responses (measured by PET-CT) to oxaliplatin-based treatment than those with low levels. These data suggest that POLH mRNA expression should be tested as a biomarker to predict survival and therapeutic responses in oesophageal cancer patients treated with oxaliplatin-containing chemotherapy.

The Saccharomyces cerevisiae Rmi1 protein is a component of the highly conserved Sgs1-Top3-Rmi1 complex, which is required for the maintenance of genome stability. The rmi1Δ deletion mutant has proven difficult to study because it exhibits very poor growth, and rapidly accumulates second site suppressor mutations. Furthermore, deletion of the putative HJ resolvase genes, MUS81-MUS81 or SLX1-SLX4 in rmi1Δ mutants causes synthetic lethality. In order to study phenotypes caused by loss of functional Rmi1, and to explore the genetic interactions between RMI1 and the MUS81, MUS81, SLX1 and SLX4 genes, a temperature sensitive mutant of RMI1 was isolated, named rmi1-1. Similar to rmi1Δ deletion mutants, rmi1-1 cells are highly sensitive to the DNA damaging agent, MMS and the replication inhibitor, HU. In addition, rmi1-1 mutants accumulate replication-associated branched DNA structures, and arrest in G₂/M after a transient exposure to MMS. These cells are proficient in DNA damage checkpoint activation. Deletion of SLX1, SLX4, MUS81 or MUS81 in the rmi1-1 strain causes synthetic lethality, which is associated with cell cycle defects. Following a transient exposure to MMS, rmi1-1 mutants accumulate homologous recombination intermediates. These intermediates are slowly resolved at the restrictive temperature, revealing a redundant resolution activity in the absence of functional Rmi1. This resolution depends upon Mus81-Mms4, but not on Slx1-Slx4 or Yen1. I propose that while the Sgs1-Top3-Rmi1 complex constitutes the main pathway for removal of homologous recombination intermediates following a perturbed S-phase, Mus81-Mms4 can act as a back up for resolution of these intermediates, which most likely represent double Holliday junctions. In this study, I also present screens for high copy suppressors of rmi1-1 phenotypes, and for novel Rmi1 interaction partners.

The type 1 insulin-like growth factor receptor (IGF-1R) is a receptor tyrosine kinase that mediates diverse cellular functions including growth, differentiation, migration and apoptosis protection. IGF-1R signalling has been implicated in tumorigenesis in a variety of cancers, and IGF-1R inhibitory drugs are currently undergoing clinical evaluation. Previous work in our laboratory has shown IGF-1R over-expression in urological cancers at both the mRNA and protein level, thus making it a potential therapeutic target. The first aim of this project was to develop a protocol for IGF-1R immunohistochemistry, investigate the expression and cellular distribution of the IGF-1R receptor in clear cell renal cell carcinomas (ccRCC), and assess correlation with clinical parameters. In tissue microarray analysis, IGF-1R was detected in ~90% of 195 ccRCCs, with signal in the plasma membrane, cytoplasm and also in the nucleus. The presence of nuclear IGF-1R in up to 50% of ccRCCs and its association with adverse prognosis was a novel finding, and suggests that nuclear IGF-1R may influence ccRCC biology. Further investigations will clarify its role in the nucleus and its potential as a prognostic biomarker. The second aim was to investigate effects of IGF-1R inhibition on radiosensitivity and DNA repair, following previous work in our laboratory showing that IGF-1R depletion enhances chemo- and radio-sensitivity, delays double strand break (DSB) resolution, and may play a role in the homologous recombination (HR) pathway of DNA DSB repair. However, the repair defect seen in these early experiments was larger than could be entirely explained by a defect in HR. The current project used a small molecule IGF-1R tyrosine kinase inhibitor AZ12253801 (AstraZeneca), which blocked IGF-1 induced IGF-1R activation and inhibited cell survival. AZ12253801 enhanced the radiosensitivity of prostate cancer cells, which appeared to be independent of effects of IGF-1R inhibition on cell cycle distribution and apoptosis induction. IGF-1R inhibition delayed the resolution of γH2AX foci, supporting a potential role for the IGF-1R in DSB repair. This delay in focus resolution was apparent at early time-points (less than 4 hr), and was epistatic with DNA dependent protein kinase (DNAPK) inhibition in prostate cancer cells and DNAPK deficiency in glioblastoma cells. These results suggest a role for the IGF-1R in the non-homologous end-joining (NHEJ) pathway of DNA DSB repair. A cell-based reporter assay in HEK-293 cells confirmed that IGF-1R inhibition suppressed DSB repair by NHEJ, helping to explain the radiosensitization demonstrated upon IGF-1R inhibition. There was lack of support for a transcriptional effect, with no significant change observed in gene expression on microarray analysis. Although the mechanism of this effect remains unclear, the observed inhibition of NHEJ has implications for the use of IGF-1R inhibitors in combination with DNA damaging agents in cancer treatment.

Linear genomes end in characteristic structures consisting of repetitive DNA and proteins: the telomeres. These play two critical roles: on one hand they avoid the of loss of genetic information due to the incomplete replication of the chromosome ends and on the other, they provide capping structures for chromosome termini, differentiating them from double strand breaks. Telomeres contain specialized proteins (the shelterin complex), as well as proteins present elsewhere on the chromosomes (chromatin remodelling, DNA damage repair and response factors). Interestingly, several DNA damage factors are required for proper telomere maintenance, drawing a thin line between telomere protection and their recognition as broken DNA ends. Loss of telomere integrity has severe consequences for the cell, namely it can induce replicative senescence and cellular aging, or it can contribute to tumorigenesis. How telomeres are capped and how they are perceived by the cell when they become dysfunctional is essential for our understanding of the contribution of loss of telomere integrity to aging and disease. In order to unravel new factors involved in telomere maintenance, siRNA screens were performed. The optimization process has confirmed both telomeric foci and telomere dysfunction-induced foci (TIFs) as suitable readouts and the screens performed generated a list of potential candidate genes involved in telomere biology. Although some of the candidate genes tested in this work failed the validation process, other genes deserve further analysis. In addition this work also studied the role of several DNA damage factors at uncapped telomeres. Furthermore, BRCA1, CtIP and EXO1 were found to be critical for the formation of end-to-end fusions generated after TRF2 inactivation. The requirement of this proteins in this process, suggests that not only that not only the classical non-homologous end joining (C-NHEJ) pathway is active at TRF2-depelted telomeres, but emphasises the multiplicity of mechanisms that act to repair dysfunctional telomeres.

The DNA damage response (DDR) is a complex signalling cascade triggered in response to stress, in an attempt to maintain genomic integrity. Components of this pathway, such as ATM-mediated signalling, have been proposed to act as a barrier in the early stages of tumourigenesis. Regions of low oxygen concentrations (hypoxia) occur in most solid tumours and are associated with a poor prognostic outcome. Here, we investigated the DDR induced following hypoxia-induced replication stress in an attempt to decipher the mechanism of ATM activation in response to physiological stresses that do not induce DNA damage. We hypothesized that hypoxia-mediated chromatin changes could impact on ATM signalling. We have characterised H3 methylation in response to hypoxia and found oxygen dependent changes in H3K9me3, including both global and replication fork associated increases in this histone modification. Importantly, we have found that decreases in H3K9me3 result in loss or attenuation of ATM activation. Notably, in a background of replication stress and increased H3K9me3, ATM inhibition or loss leads to accumulation of DNA damage and a significant decrease in replication rates in hypoxia. We propose that when replication stress occurs in the presence of hypoxia-induced chromatin changes, ATM activation is facilitated by the induction of H3K9me3. In this context, we propose a novel and stress specific role for ATM-mediated signalling in maintaining replication and preventing the generation of DNA breaks that may compromise genomic integrity. Moreover, the biological consequences of the hypoxia-induced chromatin context and in particular hypoxia-induced H3K9me3 include the repression of APAK, a negative regulator of p53. Activation of p53 is a key consequence of the hypoxia-induced DDR. Here we found that SETDB1, one of the H3 methyltransferases induced by hypoxia, mediates APAK repression. We propose that H3K9me3 plays a role in regulating APAK expression to allow optimal induction of p53 dependent apoptosis in hypoxic conditions suggesting a further role for H3K9me3 in facilitating DDR signalling in hypoxia. Together, these data suggest that the hypoxic chromatin context is critical for the role of the DDR as a barrier to tumourigenesis and predict that altering the chromatin landscape in combination with DNA damaging therapies would be efficacious in the treatment of hypoxic tumours.

RASSF1A is a tumour suppressor, the silencing of which occurs through promoter methylation in a variety of human cancers. Loss of RASSF1A is associated with decreased sensitivity to DNA damaging agents and worse prognosis in breast, colon and lung cancers amongst others. RASSF1A functions in a number of cellular processes, promoting apoptosis in response to DNA damage or death receptor signalling, or cell cycle arrest at both G1/S and pro-metaphase checkpoints. As a scaffold protein, RASSF1A imparts these functions through direct interaction with target proteins. We have identified a novel interaction between RASSF1A and the SRC activator, OSSA. Further studies identify a role for RASSF1 in SRC signalling. We find that a second isoform of RASSF1, RASSF1C, the expression of which is maintained in cancers, is able to activate SRC. We also identify a novel tumour suppressor role for RASSF1A inhibiting SRC activation through binding of RASSF1C. SRC activation by RASSF1C expression promotes internalisation of adherens junctions leading to subsequent loss of tight junctions and cell polarity markers from sites of cell-cell contact. -catenin is also found to be re-localised throughout the cells from where it is hypothesised to be able to upregulate pro-proliferative genes. In addition, we find that RASSF1C expression promotes cell motility in both scratch wound and transwell assays. Finally, we show that RASSF1C expression enhances tumour cell aggressiveness using a mammosphere growth assay. We conclude that RASSF1C is an oncogene that can promote EMT through the activation of SRC family kinases. This function is inhibited by the tumour suppressor RASSF1A. This work highlights why RASSF1A is lost through epigenetic mechanisms and not mutation and why loss of RASSF1A is associated with more aggressive, metastatic cancers.

Many anticancer therapies, including radiotherapy, act by damaging the deoxyribosenucleic acid (DNA) that is fundamental to cell function and proliferation. H2AX is a histone protein associated with DNA that is phosphorylated to produce γH2AX in response to DNA double strand breaks (DSBs), the most lethal lesions caused in cancer cells. This thesis examines the translation of γH2AX detection assays to clinical situations in order to provide biomarkers of response that might help to guide the treatment of cancer patients. γH2AX immunohistochemistry was developed in preclinical xenograft models, and validated over a range of radiation doses and over time after irradiation. The method was prepared for translation to archived clinical biopsy and surgical specimens. The DSB Biomarkers Pilot Study was established in order to develop a method for γH2AX quantification in direct tumour cell specimens obtained using the clinical technique of fine needle aspiration (FNA) cytology. Eleven patients undergoing anticancer therapy were recruited to the study, and the method evaluated. The coefficient of variation of the measure was 49%. Non-invasive imaging for γH2AX would allow DNA damage to be quantified in all tumour sites, and on multiple occasions. An antibody-based nuclear medicine imaging agent was re-engineered using Fab fragments of the antibody. The novel agent demonstrated improved pharmacokinetics when compared to the whole antibody agent, but reduced target specificity. The findings further develop the potential to exploit DNA damage biomarker measurements in clinical oncology.

Nucleotide synthesis is a conserved and highly regulated response to DNA damage, required for the efficient repair of DNA double strand breaks (DSB) by homologous recombination (HR). This is essential to prevent loss of heterozygosity (LOH) and maintain genome stability. The aim of this study was to identify new genes important for HR through roles in damage-induced nucleotide synthesis. A screen was performed to identify S. pombe gene deletion strains whose DSB sensitivity was suppressed by deleting the ribonucleotide reductase (RNR) inhibitor spd1⁺ to promote nucleotide synthesis. The screen identified a number of genes including ddb1⁺, cdt2⁺, rad3⁺ and csn1⁺ which have known roles in nucleotide synthesis. Distinct roles were identified for the DNA damage checkpoint in suppressing LOH. rad3⁺, rad26⁺, rad17⁺ and the rad9⁺, rad1⁺ and hus1⁺ genes encoding the 9-1-1 complex were required for DNA damage-induced nucleotide synthesis through Cdt2 induction to promote Spd1 degradation. The HR repair defect of rad3⁺ and rad26⁺ deletion strains was partially suppressed by spd1⁺ deletion. However, the HR repair defect of rad17⁺, rad9⁺, rad1⁺ and hus1⁺ deletion strains was not suppressed. An additional role was confirmed for Rad17 and the 9-1-1 complex in preventing LOH by promoting DSB resection. A role was identified for the Gcn5 histone acetyl transferase (HAT) protein module, consisting of Gcn5, Ngg1, Ada2 and Sgf29, in suppressing DSB sensitivity by promoting nucleotide synthesis. This was independent of Cdt2 or RNR protein levels. The Gcn5 HAT module was also found to regulate DSB repair pathway choice consistent with previous observations. Deletion of gcn5⁺, ngg1⁺ or ada2⁺ decreased HR and increased non-homologous end joining. Surprisingly, deletion of spd1⁺ in a gcn5∆, ngg1∆ or ada2∆ background also promoted HR. This predicts a role for nucleotide pools in regulating DSB repair pathway choice. Eleven other candidates showed repeatable suppression of DSB sensitivity following spd1⁺ deletion. However many of these candidates did not show reduced nucleotide levels. This suggests deleting spd1⁺ may also suppress DSB sensitivity by a different mechanism.

Maintaining replication fork integrity is vital to preserve genomic stability and avoid cancer. Physical DNA damage and altered nucleotide or protein pools represent replication obstacles, generating replicative stress. Numerous cellular responses have evolved to ensure faithful DNA replication despite such challenges. Understanding those responses is essential to understand and prevent or treat replication-associated diseases, such as cancer. Re-priming is a mechanism to allow resumption of DNA synthesis past a fork-stalling lesion. This was recently suggested in yeast and explains the formation of gaps during DNA replication on damaged DNA. Using a combination of assays, we indicate the existence of re-priming also in human cells following UV irradiation. The gap left behind a re-primed fork must be stabilised to avoid replication-associated collapse. Our results show that the checkpoint signalling protein CHK1 is dispensable for stabilisation of replication forks after UV irradiation, despite its role in replication fork progression on UV-damaged DNA. It is not known what proteins are necessary for collapse of an unsealed gap or a stalled fork. We exclude one, previously suggested, endonuclease from this mechanism in UV-irradiated human fibroblasts. We also show that focus formation of repair protein RAD51 is not necessarily associated with cellular sensitivity to agents inducing replicative stress, in rad51d CHO mutant cells. Multiple factors are required for replication fork stability, also under unperturbed conditions. We identify the histone methyltransferase SET8 as an important player in the maintenance of replication fork stability. SET8 is required for replication fork progression, and depletion of SET8 led to the formation of replication-associated DNA damage. In summary, our results increase the knowledge about mechanisms and signalling at replication forks in unperturbed cells and after induction of replicative stress. / At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 1: Submitted. Paper 2: Submitted. Paper 3: Manuscript. Paper 5: Submitted.