Rapid clearance of dying cells is a vital feature of apoptosis throughout development, tissue homeostasis and resolution of inflammation. The phagocytic removal of apoptotic cells is mediated by both professional and amateur phagocytes, armed with a series of pattern recognition receptors that participate in host defence and apoptotic cell clearance. CD14 is one such molecule. It is involved in apoptotic cell clearance (known to be immunosuppressive and anti-inflammatory) and binding of the pathogen-associated molecular pattern, lipopolysaccharides (a pro-inflammatory event). Thus CD14 is involved in the assembly of two distinct ligand-dependent macrophage responses. This project sought to characterise the involvement of the innate immune system, particularly CD14, in the removal of apoptotic cells. The role of non-myeloid CD14 was also considered and the data suggests that the expression of CD14 by phagocytes may define their professional status as phagocytes. To assess if differential CD14 ligation causes the ligand-dependent divergence in macrophage responses, a series of CD14 point mutants were used to map the binding of apoptotic cells and lipopolysaccharides. Monoclonal antibodies, 61D3 and MEM18, known to interfere with ligand-binding and responses, were also mapped. Data suggests that residue 11 of CD14, is key for the binding of 61D3 (but not MEM18), LPS and apoptotic cells, indicating lipopolysaccharides and apoptotic cells bind to similar residues. Furthermore using an NF-kB reporter, results show lipopolysaccharides but not apoptotic cells stimulate NF-kB. Taken together these data suggests ligand-dependent CD14 responses occur via a mechanism that occurs downstream of CD14 ligation but upstream of NF-?B activation. Alternatively apoptotic cell ligation of CD14 may not result in any signalling event, possibly by exclusion of TLR-4, suggesting that engulfment receptors, (e.g. TIM-4, BAI1 and Stablin-2) are required to mediate the uptake of apoptotic cells and the associated anti-inflammatory response.

The apoptosome is a caspase-activating complex consisting of Apaf-1, caspase-9 and cytochrome c, which is essential for the induction of stress-mediated apoptosis. This complex ranges in size from ~ 700 kDa to ~ 1.4 MDa, possibly due to the stable association of modulatory proteins. In the current study I employed two strategies to isolate and characterise the active ~ 700 kDa apoptosome in vitro. Firstly, I used GST-Casp91-130, which binds to the CARD domain of Apaf-1 in a dATP and cytochrome c-dependent manner, to affinity-purify an apoptosome containing only Apaf-1XL and cytochrome c. This result was confirmed by the second approach, which used an antibody to the caspase-9 to immunoprecipitate of the native apoptosome, which contained Apaf-1 and caspase-9 (p34/p35). However, in the absence of SMAC and Omi, the native apoptosome also contained caspase-3 and XIAP, both of which associated via the catalytic domains of caspase-9. When isolating the apoptosome from apoptotic cells using TAP-tagged caspase-9 variants, I discovered that the location of the TAP-tag can affect the ability of caspase-9 to interact with known binding partners and consequently can influence the induction of cell death. I have also studied the role of the apoptosome in the caspase-3 null cell line, MCF-7, and have demonstrated that an active ~ 700 kDa apoptosome complex is formed in both dATP-activated cell lysates and an apoptotic MCF-7 cells. Furthermore, the active apoptosome can directly process and activate caspase-7.

All organisms have developed regulated mechanisms to maintain homeostasis. At the cellular level, this normal functioning of cells is regulated by expression of regulatory genes that are required for normal cell function. Most cells in multicellular organisms are capable of altering gene expression in response to extracellular signals such as elevated temperature, ischaemia/reperfusion, inflammation, infection, cytokines, and amino acid analogues. In this thesis the effects of cellular stresses in the form of elevated temperature or simulated ischaemia have been investigated. Previous studies show that elevated temperature or simulated ischaemia can induce expression of heat shock proteins (Hsps) in order to prevent misfolding of cellular proteins. Moreover, it has been shown that the stress responsive transcription factor heat shock factor-1 (HSF-1) is phosphorylated and translocates to the nucleus to bind to heat shock elements within hsp gene promoters. In addition, HSF-1 can interact with other transcription factors such as the signal transducer and activator of transcription-1 (STAT-1), which is a latent cytoplasmic transcription factor activated in response to regulatory cytokines such as interferon gamma (IFNgamma). Preliminary data shows that elevated temperature can induce expression of Hsp90 in the STAT-1 deficient cell line (U3A) treated with IFNa (activates STAT-1 and STAT-2), but reduces the levels of Hsp90 expression in the U3A cell line treated with IFNa and IFNgammain combination. These findings suggest that there may be competition between STAT-1 homodimers and STAT-1/STAT-2 heterodimers and will require further investigation The STAT-1 transcription factor has previously been demonstrated to play a role in stress-induced apoptosis. In this study, STAT-1 is shown to be required for stress-induced apoptosis using the STAT-1 deficient U3A cell line. Cells lacking STAT-1 show reduced cell death/apoptosis in response to elevated temperature or simulated ischaemia. However, expression of STAT-1 in these cells restores sensitivity to stress-induced death. The C-terminal domain alone of STAT-1 is also able to enhance stress-induced cell death, and may be acting via a novel co-activator-type mechanism. Many protective agents have been identified that are able to reduce cell death due to ischaemic injury. Cardiotrophin-1 (CT-1), a member of the IL-6 family of cytokines, has been shown to protect rat neonatal cardiomyocytes subjected to simulated ischaemia via the p42/p44 MAPkinase and PI-3 Kinase pathways. In addition, the unrelated peptide urocortin (Ucn) also protects cardiomyocytes via the same pathway as CT-1 in response to simulated ischaemia and both CT-1 and Ucn induce Hsp expression. In this study, Ucn has been shown to be able to induce enhanced expression of CT-1 at mRNA and protein levels in response to simulated ischaemia. Moreover, the effect is mediated by activation of the CT-1 promoter and requires the transcription factor C/EBPp /NFIL-6. This finding indicates that a common pathway exists for these two protective agents with Ucn inducing CT-1 synthesis. Overall, the work performed indicates that multiple interacting pathways modulate the cellular stress response with either protective or damaging effects.

The serine protease HtrA2 is released from mitochondria following apoptotic stimuli. Once in the cytosol, HtrA2 has been implicated in promoting cell death by a caspase-dependent and -independent mechanism. However, mice lacking expression of HtrA2 show no evidence of reduced rates of cell death. On the contrary, loss of HtrA2 causes mitochondrial dysfunction leading to a neurodegenerative disorder with parkinsonian features. This suggests that the protease function of HtrA2 in the mitochondria, and not its pro-apoptotic action in the cytosol, is critical. Mammalian HtrA2 is therefore likely to function in vivo in a manner similar to its bacterial homologues, which are involved in protection against cell stress. The bacterial DegS homologue senses unfolded proteins, activating a proteolytic cascade leading to induction of stress response genes. Transcriptional profiling of wild type and HtrA2 knockout (KO) cells identified the stress-inducible transcription factor CHOP being differentially regulated when mitochondrial stress was triggered. CHOP up-regulation was found in HtrA2 KO mouse brains but not in other tissues. Transcriptional profiling of brain tissue revealed a number of putative ATF4 target genes being up-regulated in HtrA2 KO, among these CHOP. Promoter analysis identified a C/EBP-ATF composite site in the majority of the genes within this signature. Therefore, loss of HtrA2 might impact on nuclear gene expression specifically in brain, subverting normal cellular homeostasis leading to disease. In humans, point mutations in HtrA2 are a susceptibility factor for Parkinson's disease (PD) resulting in partial loss of proteolytic activity. Affinity purification shows that the mitochondrial kinase PINK1 interacts with HtrA2. PINK1 mutations are associated with the PARK6 PD susceptibility locus. HtrA2 is phosphorylated in a PINK1-dependent manner at residues adjacent to positions found mutated in PD patients. Phosphorylation of HtrA2 and thereby modulation of its proteolytic activity seems necessary for the function of HtrA2 in the mitochondria contributing to increased resistance of cells to mitochondrial stress.

Nutrient deprivation of eukaryotic cells elicits a rapid survival response, including the induction of autophagy. Autophagy, or "self-eating", involves the formation of autophagosomes from an unknown membrane source and the sequestration of cytosolic components, including organelles such as mitochondria, endoplasmic reticulum and small vesicles. Autophagosomes then fuse with protease-containing endosomes and their contents are degraded. This allows the cell to recycle amino acids and to reuse them for the synthesis of new proteins. To study the formation and fusion of autophagosomes, I have developed in vivo and in vitro assays, based on primary rat hepatocytes cultures. For the in vitro assay, the aim of which is to identify proteins involved in fusion, I have designed specific markers and internalized them into endosomes and autophagosomes. Both vesicle populations have been purified and used to reconstitute fusion in vitro. For the in vivo experiments I expressed the GFP-tagged autophagosomal marker, LC3, in cultured primary rat hepatocytes. By measuring the translocation of GFP-LC3 from a cytosolic pool to newly formed autophagosomes, using a high throughput-imaging system, and by assaying for the lipidation of GFP-LC3, I was able to quantify the rate and magnitude of autophagosome formation and fusion. Starvation led to an increase in the rate of autophagosome formation, and the total number of autophagosomes per cell increased more than two-fold. Autophagosome formation was independent of mTOR, a negative regulator of autophagy in yeast and many cell lines, and could be strongly inhibited by leucine, a regulatory amino acid. I then investigated the role of microtubules in the formation and fusion of autophagosomes, using the microtubule-depolymerising drugs nocodazole and vinblastine. I found that nocodazole treatment reduced the rate of autophagosome formation and completely inhibited their mobility. In addition, both drugs inhibited fusion with endosomes, showing that an intact microtubule network is also required for fusion. Interestingly, vinblastine also strongly stimulated autophagosome formation, even in nutrient-rich medium. This effect was independent of mTOR activity, but required the autophagy proteins Atg5 and Atg6, suggesting that vinblastine affects a novel signalling pathway upstream of autophagy proteins.

TRAIL selectively induces apoptosis in cancerous cells through the formation of the death-inducing signalling complex (DISC). Apoptosis can progress through either a Type I (direct procaspase-3 cleavage) or Type II (Bid cleavage) route. However, factors acting at the DISC, which may determine the selectivity of TRAIL and/or which pathway (Type I/II) is activated, are currently ill-defined. Therefore, the aim of this thesis was to investigate the initial phase of TRAIL signalling through protein purification, mass spectrometry and functional assays. A tagged variant of FADD (DISC component) was stably expressed, allowing interactions with calpain small subunit 1 (calpain S1) and heat shock cognate 71 kDa protein (HSC70) to be identified by mass spectrometry. It was speculated that FADD and calpains may interact via calpain S1, whilst HSC70 may contribute to TRAIL- induced subcellular redistribution of FADD. Mass spectrometry of the TRAIL DISC identified the known components and a potential interaction between transferrin receptor and TRAIL-R1/-R2. It was hypothesised that this interaction may help target TRAIL-R1/-R2 and DISC for endocytosis. Increased TRAIL DISC formation in Type I versus Type II cells was unlikely to be the result of other proteins, since no significant differences in DISC composition were detected. Further studies indicated that low TRAIL DISC formation in the Type II Jurkat cell line was likely related to the absence of TRAIL-R1 expression and inefficient recruitment of FADD to TRAIL-R2. Functional assays suggested that Type I and Type II TRAIL DISC have a similar substrate preference. Therefore, in contrast to previous reports, these data suggest that the lower level of DISC formed in a Type II cell line does not preferentially cleave Bid to activate the intrinsic pathway. Instead lower caspase-8 activation at the Type II DISC results in overall less substrate cleavage, delaying apoptosis and thus enabling the intrinsic pathway to contribute to caspase-3 activation.

Kinases found in signalling pathways leading to growth and proliferation can become mutated leading to aberrant cell cycle progression and tumourigenesis. Inhibitors of these oncogenic kinases have already shown clinical effect, with imatinib, an inhibitor of the BCRlAbl fusion kinase found in chronic myeloid leukaemia producing excellent clinical responses. The molecular basis behind this effect has been shown to be due to induction of intrinsic apoptosis, through upregulation of the pro-apoptotic BH3-only proteins of the Bcl-2 family. Our aim is to investigate the effect of kinase inhibition on cells with other oncogenic kinases, including two novel kinases. Malignancies that express oncogenic kinases include mutant epidermal growth factor receptor (EGFR) in non-small cell lung cancer (NSCLC) and B-Raf in colon cancer and melanoma. Currently it is not well understood how signalling pathways link the induction of apoptosis through the upregulation of BH3-only genes. To address this, the key signalling pathways regulated after kinase inhibition were studied, focussing on the ERK, AKT, p38 an JNK pathways and a comprehensive study of the levels of all the BH3-only proteins was carried out. Inhibition of MEK1 alone kills B-Raf mutant melanoma and colon cancer cells, whereas B- Raf wildtype cells were more resistant, with the induction of Bim and Bmf correlating with mutant status. Inhibition of mutant EGFR in NSCLC with Iressa resulted in cell death, with Bim and Puma becoming induced. The combination of MEK and AKT inhibition of EGFR mutant NSCLC cells did not result in appreciable cell death, although Bim was induced, indicating that a decrease of pro-survival members of the Bcl-2 family by another pathway is also needed to induce cell death. In addition, the signalling pathways evoked by two novel fusion kinases CD? 4-ROS and SLC34A2-ROS were also investigated as additional potential oncogenic kinases. By utilising a variety of techniques, including flow cytometry with various anti-CD? 4 antibodies, immunofluoresence using antibodies against c-ROS and cell surface biotinylation assays, a lack of cell surface expression was revealed, contrary to earlier predictions. Furthermore, neither c-ROS fusion protein appeared to confer robust pro-survival signalling, although they were able to stimulate ERK activation and ROS kinase inhibition did not induce apoptosis. Finally, CD?4-ROS was unable to transform Ba/F3 or NIH-3T3 cells, suggesting these two fusion kinases are not a good candidate to target in therapeutic strategies. Together, these studies further our knowledge of oncogenic kinases and indicate that signals downstream of these kinases are complex, being both cell and context dependent. Not only induction of the pro-apoptotic BH3-only proteins is required to induce cell death but also a decrease in the pro-survival members, plus release from any sequestration. We anticipate that the knowledge gained will help to develop future therapies, suggesting combination therapies will be the most effective.

Activation of the mitotic checkpoint by microtubule-interfering drugs (MIAs) such as r- axol causes mammalian cells to undergo apoptosis. Previous studies have shown that tr atment with such drugs activate the extra-cellular regulated kinase (ERK) pathway, causing the cells to arrest in mitosis and then undergo apoptosis. ERK1I2 activation in response to MIAs has been implicated as having a role in mediating MIA-induced cell survival in certain cancer cell types such as lung cancers. Using Hela cells, the results in this project confirm the activation of ERK1I2 upon introduction of MIAs such as Nocodazole, Taxol, Vinblastine and Vincristine. GFP-Erk 1 expression showed a rapid cytoplasmic to nuclear translocation of ERK1I2 in response to MIAs, further suggesting a direct role for ERK1I2 in cell survival. Quantifying cell death in response to MIAs with or without the MEKlERK inhibitor U0126 showed that ERK1/2 inhibition may increase cell death in comparison to MIA treatment alone. Subsequent data has shown, for the first time, that upon ERK1I2 inhibition by U0126, cell death occurs via the intrinsic apoptotic pathway, as seen by caspase-9 and caspase-3 cleavage after 12 hours of treatment with Taxol plus U0126. Furthermore, inhibiting ERK1I2 activation in Taxol- induced mitotic cells using the UO 126 compound increases cell death suggesting that cells need to reach mitosis before the onset of apoptosis. Preliminary data in this report also supports a role for active ERK 112 in mediating cell survival via phosphorylation of the pro-apoptotic protein Bim and possibly caspase-9. Collectively, the results shown in this report provides potential new targets in the development of anti-cancer therapies.

Apoptosis can often be disrupted in cancer cells. Caspase-8 is an initiator caspase, which is part of the extrinsic apoptotic pathway. It has been shown that variants in CASP8 are involved in susceptibility to breast cancer. Recently, a four single nucleotide polymorphism (SNP) haplotype was found to increase breast cancer risk. The aim of this project was to fine-map the CASP8 gene region, in relation to breast cancer, by genotyping a panel of tagging SNPs in cases and controls, and by sequencing exons in haplotype carriers. Another aim was to test whether the 4-SNP haplotype, and two other CASP8 variants shown to be significantly involved in breast cancer susceptibility, might be responsible for the variations in risk via effects on the apoptotic response. The final aim was to carry out a replication study of CASP8 variants in prostate cancer. Fine-mapping of CASP8 was performed using DNA sequencing and multiplex genotyping. To test the apoptotic response; peripheral blood Iymphocytes from whole blood were induced into apoptosis and fluorescence activated cell sorting was used to measure levels of apoptosis and caspase-8 induction. The sequencing of all exons in CASP8 in 24 risk haplotype carriers and 23 protective haplotype carriers identified 22 variants in CASP8, 2 of which were novel. Fine- mapping of 1,015 breast cancer cases and 1,034 controls identified 6 SNPs significantly associated with breast cancer, two at p<O.OI. We found no association between genotype and caspase-8 activity or apoptosis. A replication study of seven CASP8 SNPs in 2,533 prostate cancer cases and 4,078 controls provided support for one SNP, p = 0.0004, but the evidence was less conclusive for the other six. These results support the hypothesis that variants in CASP8 are involved in breast and prostate cancer susceptibility, however further work is needed to identify the functional variants involved

The human AIFM2 protein (formerly known as AMID) is an apoptosis associated flavoprotein with a 6-hydroxy FAD cofactor. AIFM2 is a NAD(P)H-binding oxidoreductase with some sequence similarities to A1FM1 (formerly known as AIF, Apoptosis Inducing Factor), a mitochondrion-associated enzyme which relocates to the cell nucleus during apoptosis and Is considered to be a key player in the progression of cell death (1). Uniquely, AIFM2 binds 100% 6-hydroxy FAD as a cofactor. The complete 6-hydroxylation of the FAD cofactor in AIFM2 is auto-catalysed only in the presence of NADPH, and AIFM2 is purified from an E Coli expression system as the completely 6-hydroxylated form, which Is green in colour (2).