RKIP was initially discovered as an endogenous inhibitor of the ERK and NF-κB pathways,and was also shown to prolong the activation of GPCRs via inhibition of the GRK2 protein. Now increasing evidence has linked RKIP to a metastases suppressing and chemo-sensitising role in cancer cells.The chemo-sensitising effect of RKIP was investigated in a colon carcinoma cell line using a variety of chemotherapeutic agents from conventional agents to newer targeted therapies. Initial results suggested that role of RKIP in the modulation of chemotherapeutic drug response was at the level of apoptosis; there did not appear to be great observable effects in the cell proliferative response and the cell cycle distribution of the colon carcinoma cells after treatment with selected agents. Apoptosis modulation by RKIP occurred after treatment with doxorubicin, FasL, paclitaxel and TRAIL. TRAIL-treated colon carcinoma cells displayed increased cell death as the levels of RKIP within the cell were increased. In contrast, doxorubicin, FasL and paclitaxel-treated cells displayed a scaffold-like response as the levels of RKIP were increased in the cell; with WT RKIP-expressing cells being more sensitive to doxorubicin, FasL and paclitaxel-induced apoptosis than low or high RKIP-expressing colon carcinoma cells. There was no modulation of 5-FU, cisplatin and etoposide-induced apoptosis by RKIP. Indeed, these three agents did not appear to induce cell death in this colon carcinoma cell line. RKIP modulation of chemo-sensitivity has never been shown before in a colon carcinoma cell line and this is the first time that doxorubicin and FasL-induced apoptosis has been shown to be modulated by RKIP. Further, it is shown here, for the first time, that the modulation of chemotherapy-induced apoptosis by RKIP can change depending upon the cytotoxic drug employed as treatment. TRAIL and FasL, both members of the TNF super-family, were selected for further analysis due to the distinctive cell death responses observed as a consequence of the levels of RKIP within the cell. WT RKIP cells were sensitive to FasL treatment, and high RKIP cells were most sensitive to TRAIL administration. Increased sensitivity of high RKIP-expressing colon cells to TRAIL treatment appeared to involve up-regulation of the DR5 receptor; down-regulation of the anti-apoptotic molecule Bcl-xl; pIKK which activates the NF-κB pathway; and TRAF2 which has been shown to activate the NF-κB pathway. Whether RKIP directly interacts with these molecules is unknown however RKIP has been shown to bind upstream activators of the NF-κB pathway and another TRAF subtype TRAF6. YY1 expression was evident in the TRAIL-treated cells but the expression was unchanged as the levels of RKIP within the cell were altered. The FasL-treated cells also displayed decreased pIKK levels as the levels of RKIP were increased; it is possible that NF-κB was behaving as both pro- and anti-apoptotic within this cell line. Thus RKIP inhibition of the NF-κB pathway may have prevented FasL-induced apoptosis in the high RKIP-expressing colon carcinoma cells. The expression of TRAF6, which has been shown to bind RKIP, displayed a scaffold-like response with WT RKIP-expressing cells having the highest TRAF6 expression. This was also the case for the transcriptional regulator YY1, thus it is possible that both YY1 and TRAF6 were behaving in a pro-apoptotic-like manner in the WT RKIP-expressing cells. TRAF2 was also evident in the FasL-administered cells but the expression did not change regardless of the levels of RKIP within the cell. Overall, it appears that differential expression of TRAF adaptor proteins is responsible for the contrasting responses of TRAIL and FasL-treated cells with low, WT and high RKIP expression. Utilisation of particular TRAF adaptors or TRAF combinations by the TRAIL and Fas receptors may also account for the pro- and anti-apoptotic roles of the NF-κB pathway, and the recruitment or down-regulation of other proteins dependent upon the cell stimulus. How RKIP affects these proteins requires further investigation, however these results are exciting and novel, and strengthen evidence surrounding the role of RKIP in chemosensitivity. On another note, RKIP has been shown to bind the PDE5 inhibitor PF-3717842, therefore investigation of the effects of the PDE5 inhibitors sildenafil citrate and vardenafil citrate on RKIP inhibition of the ERK pathway in a colon carcinoma cell line were examined. The effects of the PDE5 inhibitors were compared to the cell migration inhibitor locostatin that has been shown to bind and inhibit RKIP, and prevent the RKIP-Raf-1 interaction. With TPA and EGF stimulation, locostatin appeared to act in a manner consistent with its known function as an RKIP inhibitor. The PDE5 inhibitors sildenafil citrate and vardenafil citrate displayed a similar trend to that of locostatin, although their effects on the ERK pathway were not as potent. It is possible that after EGF stimulation, the strong activation of B-Raf was over-shadowing the subtle effects of the drug treatments. Under growth conditions, the RKIP inhibitor locostatin did not appear to behave as an inhibitor of RKIP nor did the PDE5 inhibitors sildenafil citrate and vardenafil citrate. It is possible that the strong activation of various growth and proliferative cascades was impinging upon the ERK pathway, were overshadowing the drug effects, or resulting in off-target (RKIP-unrelated) effects of the drugs. In summary, the role of RKIP within the cell is becoming an increasingly exciting avenue of research and is consistently yielding new and interesting roles and interactions within the cell. Understanding and elucidating the roles of this intriguing protein within the cell will not only strengthen our knowledge of signal transduction regulation and modulation, but may also provide a new source of targeted therapy and means of manipulation in the treatment of cancer and chemotherapeutic drug resistance.

The mechanisms by which cancer cells hijack the actin cytoskeleton to invade and disseminate to distant sites of metastasis remains one of the great frontiers in cancer research. Many actin-regulating proteins have been identified to be important in cancer cell invasion and metastasis. However the role of a major actin assembly promoting complex, Scar/WAVE regulatory complex (WRC) in cancer cell invasion is poorly understood. WRC has a well-known motility-promoting role in 2D planar cell migration, but a recent study on human epithelial cancers suggests WRC may be anti-invasive in vivo. To investigate the controversy, human epithelial cancer cells with reduced WRC expression were tested in multiple 3D cell motility assays. Interestingly, WRC demonstrates a robust anti-invasive role in these exciting experiments. To understand how loss of WRC promotes invasion, the molecular mechanism is investigated. N-WASP is the other major actin assembly promoting protein. Unlike WRC, N-WASP is interestingly not required for 2D planar cell migration, but is important for motility in 3D. The interplay of the two major actin assembly promoting proteins has not been explored in 3D cell motility. I report here that loss of WRC promotes hyper-activation of focal adhesion kinase that leads to N-WASP accumulation and activation at the invasive front. This chain of events results in enhanced invasion providing a molecular mechanism of WRC’s anti-invasive function.  In addition to this FAK-N-WASP core mechanism, I also identified a novel pro- invasive role of HSPC300 independently of WRC. Loss of WRC possibly releases free HSPC300 that could subsequently interact with and regulate N-WASP activation during invasion providing a potential direct molecular link between the two proteins. Furthermore, WRC also supresses focal adhesion kinase mediated cell transformation and tumour formation in vivo. In this thesis I therefore demonstrate novel anti-invasion and anti-tumourigenesis functions of WRC. I also show how a novel WRC binding protein, NHS, could negatively regulate WRC function.

Drosophila melanogaster has been an important model organism for over a century, cumulating in a vast array of mutant and transgenic stocks, the publication of the genome, its subsequent annotation and more recently the production of the online gene expression database, FlyAtlas. Much of what we know about developmental biology was pioneered in Drosophila and it is possibly the most well studied and understood model organism, in terms of development, genetics and physiology. The so-called ‘omics’ era of biology has resulted in a relatively data poor discipline quickly becoming a data rich one. Therefore the need for a good model organism, which offers the balance between genetic power and relevance has never been more important, as scientists begin to evaluate and analysis this data. We will argue that Drosophila melanogaster offers the best opportunity to study the relevance of omics data. FlyAtlas is an online resource, which allows scientists to look at tissue specific gene expression in the fruit fly Drosophila melanogaster. Unexpected expression patterns of previously characterised genes may hint at novel functions, thus helping to close the phenotype gap. To test this hypothesis we looked at the neuronal gene Fasciclin 2 (fas2), which has been exhaustively characterised (over 500 papers), with neural functions ranging from axonal growth in development to synapse stabilization in the adult. Surprisingly FlyAtlas showed fas2 is predominately expressed in the Malpighian tubule (a renal, rather than neural, tissue), hinting at a previously unreported function in this tissue. Results suggest fas2 may play an important role in apical microvilli development and stability in the principal cells of the tubules. We have also shown that Fas2 may be involved in actin localisation. Fas2 shows dynamic localisation in response to cAMP and over expression of the protein results in a significant increase in secretion when tubules are stimulated with cAMP. We also present evidence that Fas2 co-localises with F-Actin bundles in response to cAMP, hinting at a role for the actin cytoskeleton in secretion. Proteomics experiments carried out in order to determine Fas2’s, interacting partners proved problematic. For this reason 2D Blue Native PAGE and sucrose gradient techniques were optimised in order to facilitate this problem. 4 Unfortunately we were unable to isolate Fas2, however we have shown that BN-PAGE offers a robust protocol for the isolation of protein/protein complexes. We can also conclude from these experiments that 2D BN-PAGE offers an ideal comparative data source for transcriptomics data such as FlyAtlas. The second gene tested in this study is the sex determination transcription factor Doublesex (dsx). Dsx has been extensively studied in its role in differentiation of both the soma and to some extent the nervous system in males and females. FlyAtlas results indicate that it is also expressed in the Malpighian tubules, again hinting at previously unknown function in this tissue. Further to this the male and female transcripts of dsx are expressed in a sex specific manner. Our results confirm these observations and dsx was localised to the principal cells of the main and lower segments of the tubules. Male tubules however do not express dsx in the transitional segment whereas females do, suggesting that perhaps this segment of the tubule constitutes a previously unknown sex specific function. We have determined that Tra RNAi is effective at knocking down the female transcript in female tubules, allowing for the study of masculinised tubules in an otherwise female fly. Experiments concluded that although males and females show differential survival in response to bacterial infection, this is not controlled by dsx expression in the tubules. Preliminary results also suggest that two genes CG8719 and YP3 are differentially expressed in male and female tubules and offer ideal candidates to study dsx role in sexually dimorphic gene expression in the tubules. In conclusion this study verifies the use of FlyAtlas to determine novel functions for well-known genes in D.melanogaster. In turn this indicates the importance of omics data, as a staring point for further functional analysis of both genes and proteins.

Neuronal function, survival and architecture all critically depend on the precise transport of intracellular proteins to a vast array of synaptic connections. Disrupted intracellular transport leads to deficits in synaptic transmission, irregular cell morphology, misallocated organelles and cell death. In addition, axonal transport deficits have been noted in the early stages of several debilitating neurological conditions, thus, axonal transport deficits may contribute to disease progression. This makes it important that we understand the contribution of axonal transport to both physiological and pathophysiological cellular processes and to the transport of essential organelles. As such the aims of this project were as follows: to investigate the long-term transport properties of visualised synaptic vesicles, to investigate whether vesicle transport could be modulated by changes in neuronal activity, to examine whether vesicle transport deficits exist in certain disease models and to develop novel assays for focusing the study of vesicle transport to specific neuronal cell types. To investigate the transport properties of visualised synaptic vesicles we exploited a lentiviral vector to express a fluorescently tagged version of an abundant synaptic vesicle transmembrane protein, synaptophysin. Using synaptophysin-GFP (syp-GFP) as a synaptic vesicle marker we then tracked the movements of synaptic vesicles in the axons of dissociated hippocampal neurons. Synaptophysin-GFP expression revealed two fluorescent vesicle populations, one population that moved in a rapid and bi-directional manner and one population that accumulated into clusters of stationary vesicles at putative presynaptic sites. Each vesicle population was analysed independently. Moving vesicles were termed motile particles, whilst vesicle accumulations were termed vesicle clusters. To investigate potential activity-dependent changes in vesicle transport and vesicle cluster localisation we used acute or co-culture application of the GABAA receptor antagonists bicuculline (bic) (20µM) or Gabazine (gbz) (20µM), which can generate increased neuronal activity or epileptiform-like activity in vitro. As a result of bic treatment we observed a significant decrease in the size of stationary presynaptic vesicle clusters. Under control conditions the average size of vesicle clusters was 14.7±1.67µm2, reducing to 12.1±1.41µm2 following 10 hours of increased neuronal activity (p=0.0042, Wilcoxon-matched pairs test, n=80, 8 experiments). In addition, increased neuronal activity also led to a significant increase in vesicle cluster turnover, which increased from 28±6.89% under control conditions to 44±8.46% as a result of increased neuronal activity (p=0.0261, unpaired student t-test, n=25, 11 experiments). However, these changes were not accompanied by any alteration in vesicle transport, with the speed, the density and the proportion of motile particles remaining unaffected by increased neuronal activity (table 3.1). This suggests that each vesicle population may therefore be differentially modulated by increased neuronal activity. To probe deeper for potential activity-dependent vesicle transport changes we restricted our study of vesicle transport to a specific axonal subtype, the hippocampal mossy fiber. To visualise mossy fiber vesicle transport, lentivirus expressing syp-GFP was pressure injected directly into the cell body layer of the dentate gyrus (DG) in hippocampal organotypic slice cultures. This revealed syp-GFP positive vesicles occupying both small (2-15µm3) and large (˃15µm3) mossy fiber synaptic terminals, which were found in and along the stratum lucidum. By examining the distribution of vesicle clusters at different time points following gbz or bic treatment (0hrs, 4hrs, 12hrs, 24hrs and 48hrs) we were able to show that epileptiform activity caused a delayed (>12 hours) but significant decrease in the proportion of large vesicle clusters. By 24 and 48 hours there was a significant decrease in the proportion of large vesicle clusters following bic treatment, decreasing from 9.4±1.21% under control conditions (n=11, 5 experiments) to 4.84%±0.72% after 24hrs (n=10, 4 experiments) and to 3.3±0.73% after 48hrs (n=12, 5 experiments), P<0.001, one-way ANOVA. This decrease in the proportion of large vesicle clusters may represent an important pathophysiological change triggered by epileptiform activity. Importantly, we also observed the same decrease in the proportion of large vesicle clusters in a mouse model of Rett syndrome, which models a severe neurodevelopmental disorder caused by a mutation in the gene coding MeCP2. As a consequence of bic treatment we observed a significant decrease in the proportion of large vesicle clusters from 7.2% ±1.78% in control cultures (n=6, 2 experiments), down to 0.9% ±0.6% in 48hr bic treated cultures (n=8, 3 experiments) and recovering to 6.9%±1.5% following bic wash out (n=11, 3 experiments); p<0.0001, one way ANOVA. Interestingly, Mecp2Stop/y hippocampal organotypic slices showed a greater decrease in the proportion of large vesicle clusters following 48hrs of bic treatment. The proportion of large vesicle clusters in 48hr bic treated WT slices was 3.3%±0.73%, whilst in 48rs bic treated Mecp2Stop/y slices it was 0.9%±0.6%, p=0.01, two-way ANOVA. These observations suggest that Mecp2Stop/y hippocampal organotypic slices are more sensitive to epileptiform activity than WT slices and may possess deficits in the vesicle transport system. Primary dissociated hippocampal cell cultures benefit from being both optically and experimentally accessible but lack a defined cellular arrangement. This hampers both the identification and study of specific cell types and specific synaptic connections. To overcome this limitation we developed a modified dissociated cell culture assay for defining the arrangement of dissociated hippocampal neurons. We cultured purified DG and CA3 cell populations in close opposition using a magnetic barrier, but transduced only DG granule cells with lenti-synaptophysin-GFP in order to visualise vesicle transport specifically in mossy fibers. Immunocytochemistry and vital dyes were used to confirm that specific cell populations could be cultured in close proximity, to confirm that lentiviral transduction was highly selective to DG granule cells and to post-hoc identify that vesicle trafficking was occurring specifically in mossy fibers. Using this method it was possible to image vesicle transport specifically in mossy fibers and to investigate vesicle cluster dynamics at putative MF-CA3 synapses. We conclude that this method is a significant improvement to previous techniques because dissociated cells can be arranged to form physiologically relevant synaptic connections, whilst remaining highly accessible to both live imaging and experimental manipulation.

Preconditioning is the phenomenon whereby tolerance to lethal insults is induced by exposing the tissue to a prior sublethal stimulus. This exists in several forms, such as ischaemic preconditioning, adenosine preconditioning and excitotoxic preconditioning. Adenosine preconditioning is known to be mediated by activation of A1 receptors and ATP-sensitive potassium channels whilst excitotoxic preconditioning mainly involves stimulation of NMDA receptors, nitric oxide and most likely ATP-sensitive potassium channel activation. ATP-sensitive potassium channel openers such as pinacidil and diazoxide are also known to exert preconditioning against various types of insults. There have been several models of ischaemia used to study preconditioning in vivo and in vitro leading to some confusion over the effects of preconditioning agents. High concentrations of glutamate or NMDA have been used as models of excitotoxicity in many experimental paradigms. Some molecular changes are associated with preconditioning phenomena, the most prominent being an increased expression of heat shock protein 72 (HSP72). The aims of the current study were to: 1) investigate the effects of exogenous glutamate and other depolarizing agents in the slice preparation and their validity for use as toxic agents 2) examine any potential preconditioning neuroprotection induced by adenosine against various depolarizing agents and elucidate the underlying mechanisms where relevant 3) examine the excitotoxic preconditioning phenomenon and possible underlying mechanisms 4) look at the effectiveness of other known preconditioning agents e.g. ATP-sensitive potassium channel openers against depolarizing agents and identify the underlying mechanisms of protection 5) identify any molecular changes that may occur during acute models of chemical ischemia or acute preconditioning. The rat hippocampal slice preparation was used to investigate the effects of depolarizing agents and preconditioning paradigms upon the extracellularly evoked field epsps, orthodromic and antidromic population spikes. Western blotting was used to detect any changes in the levels of HSP72 in the slices that may have occurred as a result of the depolarizing agents or the preconditioning treatments. It was first established that 5mM and 10mM glutamate induced depressions in the amplitudes of orthodromic population spikes which recovered to a stable plateau. The degree of recovery of the spikes depended partially upon the initial size of the response. As adenosine is known to be released in response to glutamate receptor stimulation, the effects of 5mM glutamate upon the orthodromic spikes were studied in the presence of the A1 receptor antagonist, DPCPX. It was observed that DPCPX did not attenuate the depression of the response during glutamate perfusion but there was a significant elevation in the post-glutamate recovery of the response. This effect was not observed when the protocol was applied to antidromic population spikes and field epsps, both of which showed a depression in response during 5mM glutamate perfusion but recovered fully when glutamate was removed. The field epsps showed a trend whereby smaller epsps recovered to a far greater degree than population spikes. Although this effect was not significant, the NMDA receptor blocker, MK-801, was co-perfused with glutamate during epsp recordings to examine this further. The degree to which MK-801 alone affected the response correlated with the post-glutamate recovery. To study this effect, isolated NMDA-receptor mediated epsps were recorded and the effects of 5mM glutamate upon them were studied. There was a similar tendency for small NMDA-receptor mediated epsps to recover to a higher level following glutamate treatment compared with larger potentials. In the presence of DPCPX, the larger potentials showed a significant elevation in recovery following treatment with glutamate. It was also shown that the post-5mM glutamate recovery of the orthodromic population spikes was elevated by the presence of the A2a receptor antagonist, SCH 58261. Further experiments using the ATP-sensitive potassium channel blocker, glibenclamide, indicated that this effect may be due to increasing the opening of these channels. Adenosine preconditioning was attempted using 10mM glutamate as an insult. It was shown that adenosine could not precondition against this effect in antidromic or orthodromic population spikes. The effects of the sodium-potassium ATPase inhibitor, ouabain, upon the evoked responses were studied as an alternative insult. It was shown that ouabain induced depressions in field epsps, orthodromic and antidromic population spikes. The antidromic population spikes showed significantly smaller depressions than the orthodromic responses. Further experiments using the glutamate receptor antagonist, kynurenic acid, showed that glutamate receptors mediated the effects of ouabain upon the orthodromic population spikes but not the antidromic spikes. Adenosine preconditioning was attempted against ouabain. It was shown that adenosine preconditioned against the effects of ouabain upon orthodromic and antidromic population spikes but not field epsps. Further experiments were conducted using antidromic population spikes. It was shown using various antagonists, that adenosine protection against ouabain was mediated by A1 receptors, ATP-sensitive potassium channels, NMDA receptors and nitric oxide. To extend these results further, preconditioning using the ATP-sensitive potassium channel opener, pinacidil, was attempted against 10mM glutamate and ouabain. It was shown that pinacidil was able to precondition the antidromic population spike against either insult. Using the NMDA receptor antagonist, DL-AP5, showed that the preconditioning effect of pinacidil against ouabain was mediated by NMDA receptors. Another preconditioning paradigm was attempted to see if glutamate could precondition against ouabain. It was shown that pre- treatment with glutamate resulted in enhancing the depressant effect of ouabain upon field epsps and antidromic population spikes. To further examine the effects of ouabain upon antidromic population spikes, ouabain was co-perfused in the presence of the intracellular calcium chelator, BAPTA-AM. This resulted in enhancing the depressant effect of ouabain upon the response. A similar result was observed when the calcium concentration in the perfusion medium was lowered to 0.5mM from 2.5mM whereas increasing the concentration to 5mM attenuated the depressant effect. Ouabain was also co-perfused in the presence of charybdotoxin, a blocker of large-conductance calcium activated potassium channels. It was observed that charybdotoxin enhanced the effect of ouabain upon the antidromic spikes. No changes were detected in HSP72 expression in the slices in response to ouabain treatment, 10mM glutamate treatment, pinacidil preconditioning treatment or glutamate preconditioning. The present results show that glutamate and ouabain can induce depressions in the evoked responses from the rat hippocampal slice and that the effects of 5mM glutamate can be attenuated by adenosine receptor antagonists. In addition, adenosine can precondition against ouabain but not glutamate and this effect involves A1 receptors, NMDA receptors, nitric oxide and ATP-sensitive potassium channels. It has also been observed that pinacidil can precondition against ouabain or glutamate and NMDA receptors may be involved in this effect. The inability of glutamate to precondition against ouabain in evoked responses was also demonstrated. The study highlights the effectiveness of preconditioning agents against different depolarizing agents and the interactions between adenosine and glutamate receptors which play a role in preconditioning.

Background: Exercise may acutely and chronically up-regulate appetite and energy intake in overweight and obese women preventing body mass reduction in the long term. Overweight and obese women may be most prone to compensatory responses to exercise but the possible mechanism for this is unclear. Appetite regulating hormones have been investigated as a possible mechanism but to date the evidence is somewhat mixed. Identifying compensatory energy intake responses in overweight and obese women is complicated by the high prevalence of dietary under-reporting in this group. The laboratory-based buffet meal method has frequently been used in research studies that have assessed food intake in these women, but this method has only undergone preliminary validation. Dietary restraint may also affect individual appetite responses to exercise; it has been theorised that restraint may be a behavioural adaptation to diminished energy requirements, and differences in physical activity levels could also contribute. Evidence thus far has produced mixed results, possibly because two distinct sub-groups of restrained eaters exist, those with flexible and rigid control of restraint. It is not known if there are differences in energy requirements between these two sub-groups. Participants and Methods: Participants in all studies were sedentary, healthy, pre-menopausal, overweight and obese, adult women. Study 1: Fourteen women completed four trials; two exercise and two control, following the same protocol as study 1. Energy intake at three buffet meals and subjective appetite ratings were measured, and the reproducibility of these values under control and exercise conditions was tested using intraclass correlation coefficient (ri). Study 2: Twenty-nine women completed two trials in a randomised, counterbalanced order; exercise and control. Each trial lasted 24 hours spanning 2 days; the afternoon of day 1 and morning of day 2. An exercise session to expend 1.65 MJ was completed on day 1 of exercise trials, and three buffet meals were served during each trial to measure energy intake. Appetite was assessed using a visual analogue scale and blood samples were taken to determine acylated ghrelin (n=15) and peptide YY (n=10) concentrations. A repeated measures ANOVA was used to investigate the effects of trial and time on appetite hormones, EI and appetite. Study 3: Fifteen women participated in a sixteen week exercise intervention to expend 8360 kJ week-1. Participants exercised unsupervised in the University gym, and compliance was measured via heart rate monitoring. Sub-maximal fitness and body composition assessments were carried out at baseline, and after 8 and 16 weeks of exercise. Energy expenditure, energy intake, appetite, and acylated ghrelin (n=14) and peptide YY concentrations (n=11) were measured at baseline and after 8 weeks of exercise. Paired t tests were used to assess differences in time-averaged AUC for appetite, total and relative EI, metabolic rate, and exercise responses between trials. Repeated measures ANOVA was used to assess changes over time in body composition, appetite ratings, EI, acylated ghrelin, peptide YY, and cardiovascular fitness levels. Study 4: Forty-one sedentary women in a one week observational study. Participants were classed as restrained or unrestrained using the three factor eating questionnaire, and the former group were further classified as having flexible or rigid control of restraint. All participants completed a food frequency questionnaire, sub-maximal fitness test, body composition assessment and two fasted metabolic rate measurements. Average daily energy expenditure was calculated from a seven day physical activity diary combined with continuous heart rate data. Differences between restrained and unrestrained eaters, and restrained eaters with flexible and rigid control, were assessed using a paired t-test. Results Study 1: The ri for energy intake in control trials was significant but had large associated confidence intervals (ri 0.50 (95% CI 0.03, 0.80) p=0.0003). The ri was for energy intake in exercise trials was (ri 0.04 (95% CI -0.53, 0.55; p=0.45) and for the difference between control and exercise trials was (ri -0.05 (95% CI -0.54, 0.48; p=0.57) this was not significant. The ri values for satiety, fullness and desire to eat were significant in both control and exercise trials (p<0.05), but the associated confidence intervals were large. Study 2: There was no effect of exercise on subjectively rated appetite, acylated ghrelin, or peptide YY concentrations (all p>0.05). Total energy intakes were not significantly different between trials (exercise: 10.9 ± 0.5 MJ, control: 10.8 ± 0.5 MJ; mean ± SEM). Study 3: Total exercise energy expenditure during the intervention was 80.8 ± 7.7 MJ, which resulted in a significant reduction in total body mass (-1.9 ± 0.9 kg), fat mass (-1.7 ± 0.7 kg) and BMI (-0.7 ± 0.4 kg m-2). However individual changes in body and fat mass ranged from +2.8 to -9.9kg, and +1.78 to -6.55 kg respectively. There were no significant differences in appetite, energy intake, or expenditure after 8 weeks of exercise (p>0.05). Study 4: There were no differences in metabolic rate, daily energy expenditure or physical activity patterns between restrained and unrestrained eaters (p>0.05), or between restrained eaters with flexible and rigid control of restraint (p>0.05). Conclusions: Study 1: The laboratory-based buffet meal method of measuring energy intake does not provide reliable, reproducible values in overweight and obese, pre-menopausal women either under control or exercise conditions. Study 2: A walking-based exercise session which induces a moderate energy deficit of 1.65 MJ does not appear to affect subsequent twenty four hour energy intake, subjectively rated appetite, or plasma acylated ghrelin and peptide YY concentrations during the subsequent twenty four hours. Study 3: This study concluded that 16 weeks aerobic exercise in overweight and obese women produces a small, but significant, reduction in body and fat mass (-1.9 ± 0.9 kg); however the extent of these changes varies greatly between individuals (+2.8 to -9.9kg). No evidence of compensatory changes in energy intake or expenditure, subjective appetite ratings, or circulating levels of acylated ghrelin and peptide YY was apparent after 8 weeks of exercise. Study 4: This study concluded that there is no evidence of a difference in body composition or energy requirements between overweight and obese female restrained and unrestrained eaters, or between sub-groups of restrained eaters. Dietary restraint does not appear to be an adaptation to diminished energy requirements.