The importance of IL-17 is underscored both by its resistance to control by regulatory T cells (Treg) and the propensity of Treg to produce this highly inflammatory cytokine. I addressed whether Th17 cells are inhibited by Treg from anti-TNF treated rheumatoid arthritis (RA) and psoriatic arthritis patients (PsA) and defined the underlying mechanisms. Th17 responses were inhibited by Treg isolated from RA patients responding to the anti-TNF antibody adalimumab, but not by Treg from healthy individuals or patients with active RA. Furthermore, in patients with RA, response to adalimumab therapy was associated with a reduction in RORC+ Th17 cells and an increase in FOXP3+ Treg lacking Helios and CD62L expression. These Treg suppressed Th17 cells through inhibition of monocyte-derived IL-6, but independently of IL-10 and TGF-ß, which mediated suppression of Th1 responses. Surprisingly, therapy with the anti-IL-6 receptor, tocilizumab, did not result in a reduction in RORC+ cells in RA patients. Rather, tocilizumab reduced T cell IL-21 production, which was associated with a diminished memory B cell population. The acquisition of IL-17 suppressor function by Treg was not observed in RA patients responding to etanercept, a modified TNF receptor, or in PsA patients treated with either adalimumab or etanercept. Moreover, response to therapy was not associated with an increase in Treg number in these patients. In RA patients treated with etanercept the inability of Treg to suppress Th17 responses was associated with high levels of IL-17 production and high levels of RORC+ Th17 cells ex vivo. In contrast, there was a reduction in IL-17 production and RORC+ Th17 cells ex vivo in PsA patients treated with both adalimumab and etanercept. Furthermore, depletion of Treg from PBMC showed that Treg from healthy controls, patients with active PsA and PsA patients responding to adalimumab can modulate the production of IL-22, a key cytokine in inflammatory skin disorders. However, PsA patients responding to etanercept have an impaired ability to regulate production of this cytokine. In conclusion, the induction of IL-17 suppressing Treg by anti-TNF is both therapy and disease specific. These data provides a rationale for the therapeutic benefit of switching between different anti-TNF agents. Furthermore, the induction of highly potent Treg may offer an explanation as to why patients treated with adalimumab have an increased risk of developing Mycobacterium tuberculosis (TB).

Studies on mitochondrial dysfunction in HIV-infected adults exposed to anti-retroviral therapy. A significant proportion of HIV-infected patients who require anti-retroviral therapy are or have been exposed to nucleoside analogue reverse transcriptase inhibitors (NRTIs). It has been consistently suggested that most of the NRTI-attributed adverse drug reactions (ADR) are due to mitochondrial dysfunction. In a sub-analysis of a large randomised clinical trial (Delta) the incidence of peripheral neuropathy (PN) was constant over time in all study arms, which does not support the hypothesis of cumulative toxicity previously proposed for NRTI-induced ADR. Patients taking zidovudine (AZT)/zalcitabine (ddC) combination were more likely to develop PN than patients on AZT monotherapy (RH= 2.30 95%CI= 1.62 - 3.28). The incidence of PN among patients exposed to zidovudine/didanosine (AZT/ddl) combination was not different from that observed in patients on AZT. In a multi-centre case-control study including 110 cases of lactic acidosis (LA) or severe hyperlactataemia (HL) patients with < 200 CD4 cell/pl were more likely to develop HL/LA than patients with higher levels of CD4 cells (OR=3.44 95%CI= 1.64 - 7.22). Female patients were found to be at higher risk for HULA than men (OR= 4.75 95%CI= 1.96 - 11.53). Patients exposed to either d4T, ddl or the combination of these two were four to six times more likely to develop HL/LA than patients taking other NRTIs based combinations. Interestingly, cases of HL/LA were exposed to d4T for shorter periods of time than controls. Almost 10 % of the cases included in the study were asymptomatic at the time of diagnosis. All these symptom-free cases had blood lactate ranging between 5 and 7 mmol/l. Therefore, case definitions for HL or LA based on clinical presentation may underestimate the magnitude of the problem.

Localizing the source of epileptic discharges is important in gaining a greater understanding of the disease, classifying epilepsy, and identifying areas suitable for potentially curable surgical resection. Functional imaging measures haemodynamic, metabolic or neurochemical correlates to localise neural activity. Combining EEG with functional MRI (EEG-fMRI) allows the localisation of haemodynamic correlates of neuronal events recorded on surface EEG. The work in this thesis aims to identify the spatial haemodynamic correlates of interictal epileptiform discharges (IED) in patients with epilepsy using EEG-fMRI. Five studies form the main body of this thesis. In the first study, 46 patients with frequent generalised spike wave activity (GSW) were studied with EEG-fMRI on a 1.5 Tesla scanner. The main finding was of a characteristic pattern of fMRI signal decrease in frontal, parietal and posterior cingulate cortex, areas of association cortex, during GSW. In the second study, 4 patients from this first series were re-studied with a 3 Tesla scanner. A high degree of reproducibility was seen in the spatial distribution of fMRI changes. Perfusion MRI with an arterial spin label sequence was used that showed a decrease in blood flow to these areas during GSW. In the third study, a novel method for the analysis of fMRI data in epilepsy, temporal clustering analysis (TCA) was assessed. The technique was confounded by subject motion, and we were unable to reliably detect correlates of IED. The fourth study moves away from correlating visually identified IEDs on the EEG, and correlates power fluctuations in the delta frequency band with simultaneously acquired fMRI. Finally a combination of EEG-fMRI and MR tractography were used to study a patient with temporal lobe epilepsy. The issues surrounding potential use of EEG-fMRI as a clinical tool are discussed.

Temperature measurement in an electromagnetically hostile environment, such as a magnetic resonance imaging (MRI) scanner, is challenging. The monitoring of a patient's temperature within an MRI is necessary to ensure the patient is in a comfortable condition due to the following problem. Unfortunately, the radio frequency (RF) radiation in the imaging cavity of an MRI will induce heat into both the body tissue of a patient as well as into metallic elements of any sensors. This thesis presents a feasibility study on the production of an optical phosphor-based sensor for temperature monitoring within such a hostile environment. The concept of a desired optical sensor is based on the thermal dependence of phosphor luminescence. A pulsed ultraviolet light source is used to excite phosphors coated on a thin membrane which is in contact with the patient's skin. The skin temperature (over 27-37 ·C) can then be determined by calculating the exponential decay of luminescence intensity with time after the excitation pulse has ceased. In this research, two thermographic phosphors have been investigated, namely europium doped lanthanum oxysulphide (La202S: Eu) and terbium doped lanthanum oxysulphide (La20 2S: Tb). A wide range of dopant concentrations, 0.005-15 mol% for La20 2S: Eu and 0.005-50 for mol% La202S: Tb, have been characterised in terms of photoluminescent emission, decay time and crystallinity in determining their temperature dependent characteristics. Both phosphors have shown a dependency to dopant concentration through variance of peak intensity and decay time measurements over a low temperature range of 5 to 60°C. Although maximum brightness of the temperature dependent lines is achieved at dopant concentrations of 1 and 10 mol% for La202S: Eu and La2~S: Tb respectively, results have shown that optimum temperature dependency is at a lower mol% of 0.1 for La202S: Eu with a high quenching rate of 24.03 moe-I. Therefore, 0.1 mol% L1l202S: Eu appears to be an ideal candidate for use within an optical sensor. A key aspect of this research is that in comparison to conventional phosphor temperature dependent characteristic, it has been shown for the first time that La20 2S: Tb has an increase in decay time with respect to temperature for concentrations above 2 mol%. A hypothesis of this discovery based on the differences in energy transfer pathways is described. X-ray diffraction (XRD) analysis has demonstrated a linear relationship between the c-axis lattice parameter of a unit cell of the phosphors with dopant concentration. As dopancy increases, the (100) and (002) reflections merge and there is a reduction in the c-axis parameter as well as the crystallite size. In addition, a positive thermal expansion behaviour is observed for a and c parameters as well as the volume of the unit cell over the temperature range of interest

The release of the cytokine Tumour Necrosis Factor-α (TNF-α) by alveolar macrophages, is central to the inflammatory response within the lung. The Alveolar macrophage is derived from peripheral blood monocytes. This study investigated how pulmonary surfactant (composed of approximately 90% phospholipids and lipids and 10% surfactant proteins) modulates the lipopolysaccharide (LPS) induced production of TNF- α in the human monocytic cell line MonoMac-6 (MM6). The artificial surfactants Curosurf™, Survantar™ and Exosurf™, were all shown to significantly inhibit TNF- α (p<0.05). The major lipid components of surfactant, phosphatidylcholine and cholesterol, were both also seen to modulate TNF-a release in these cells but the effect of cholesterol was not seen at physiological concentrations. Further incubation with 1,2 dipalmitoylphosphatidylcholine (DPPC), the major disaturated species of phosphatidylcholine present in surfactant also significantly reduced TNF- α. In contrast 1- palmitoyl-2-arachidonoyl phosphatidylcholine, (PAPC) an unsaturated species of phosphatidylcholine, did not significantly modulate TNF- α release in MM6 cells. DPPC was also seen to down-regulate TNF- α release in human peripheral blood monocytes. DPPC was also able to significantly reduce membrane fluidity (p<0.05) and also reduce TNF- α mRNA production. Transmission electron microscopy demonstrated that incubation with DPPC altered the monocyte membrane ultrastructure. Chromatography also demonstrated increased concentrations of DPPC in the cells after incubation with DPPC. The LPS induced release of another inflammatory cytokine IL-β and the regulatory cytokine IL-10 were unaffected by incubating MM6 cells with DPPC. The LPS induced expression of the monocyte LPS receptor CDl4, was also reduced by DPPC. LPS induced release of prostaglandin E2 (PGE2), an important regulator of TNF- α, was significantly (P<0.05) increased in DPPC treated cells. Inhibition of PGE2 production by aspirin and the specific COX-2 inhibitor NS-398, reversed the modulation of LPS induced TNF- α release by DPPC. COX-2 the enzyme responsible for PGE2 synthesis, was seen to be directly upregulated by DPPC whereas PPAR-y, an important ligand for polyunsaturated fatty acids and regulator of COX-2 expression, was reduced by DPPC. Taken together these results suggest an important role for the disaturated phosphatidylcholine species and important surfactant phospholipid DPPC, in regulating the TNF- α, in a mechanism that involves upregulation of PGE2 down-regulation of PPAR-y, and increased production of COX-2.

Chemokines are chemoattractant cytokines which act on specific receptors and play an important role in tumour biology. The aim of this project was to determine whether the chemokine receptors XCRl, CXCRl and CXCR2 and their respective ligands lymphotactin, IL-8 (CXCRl&2) and GRO-a regulate the behaviour of normal and malignant oral epithelial cells. XCRl, CXCRl and CXCR2 mRNA and surface protein expression was detected in normal and oral cancer cell lines. Lymphotactin, IL-8 and GRO-a facilitated intracellular activation of ERK1/2 signaling pathway and stimulated migration, invasion and proliferation of all cells. These effects were mediated through XCRl for lymphotactin, CXCRl and CXCR2 for IL-8 and CXCR2 for GRO-a. The cancer cells showed a greater response than normal cells and a direct relationship between receptor expression and migration, invasion and proliferation was observed. XCRl but not lymphotactin was expressed by epithelial cells in normal oral mucosa in vivo and both were expressed and upregulated in inflammation and cancer. Constitutive expression of both XCRl and lymphotactin was found in regional lymph nodes and on metastatic tumours. Lymphotactin mRNA al}d constitutive intracellular protein was detected in normal and cancerous oral cells. Exposure of normal cells to lymphotactin resulted in increased adhesion to fibronectin but not collagen and stimulated MMP-2 and -9 release whereas exposure of cancer cells resulted in increased adhesion to both collagen and fibronectin and stimulated MMP-2, 9 and MMP-7 release. These findings show for the first time that XCRl and its ligand lymphotactin are expressed by epithelial cells in a range of oral conditions and strongly suggest that they play an important role in regulating the behaviour of normal and malignant epithelial cells. Similarly CXCRl and CXCR2 are upregulated on malignant oral cells in vitro and may be important in the biology of oral cancer.

Ischaemia reperfusion injury (IRI) is an important source of morbidity and mortality and contributes to renal transplant failure. Renal IRI is known to be mediated by complement, which causes direct injury to the transplant and stimulates the adaptive immune response against transplant antigens. However uncertainty exists as to the exact trigger mechanisms of complement activation in this situation, and to the relative importance of certain effector molecules generated by complement activation. Based on the results of previous studies, I hypothesized that there would be greater involvement of the lectin pathway than previously recognized, as a trigger mechanism that could both initiate complement activation and lead to amplification via the alternative pathway in renal IRI. I also predicted that, whereas C5a has been shown to play a part in the genesis of IRI, C3a could be shown to have an important effect on innate cells that form a bridge with adaptive immunity. The aim of the work described in this thesis was therefore to investigate: (a) the role of the lectin pathway in IRI through the enzyme MASP-2; and (b) the role of C3a linking up the innate and adaptive immune responses following transplantation. I first investigated renal IRI in mice with deficiencies of the classical (C4), alternative (FB) and lectin (MASP-2) pathways of complement activation in order to explore the notion that the injury was mediated by the lectin pathway. Absence of C4 and FB did not protect the mice in a transplant model of IRI; but absence of MASP-2 was markedly protective suggesting that the lectin pathway - or rather a variant of it that did not involve C4 - has a significant role in the mouse kidney transplant IRI. As a result, I also investigated the therapeutic effect of anti-MASP-2 antibody in the mouse kidney isograft model. Absence of C3a-receptor in a native kidney model of IRI did not convey a substantial protective effect, suggesting no major effect of C3a in this injury. However, exploring a possible effect of C3a on the adaptive immune response, I found that stimulation of monocytes with C3a in conjunction with LPS resulted in significant IL-1β production which in turn induced Th-17 cells. I identified that the increase in monocyte IL-1β production is ATP and caspase-1 dependent and that the change in ATP is modulated by Pannexin-1 channel. My findings identify MASP-2 as an early step in the pathogenesis of complement-mediated mouse kidney IRI and support its identity as a therapeutic target. Furthermore, as both monocytes and Th-17 cells have been previously identified as major contributors to acute kidney rejection after transplantation, my results could explain, at least in part, how complement activation induced by tissue stress during the transplant procedure has an impact on the alloimmune response.

Though mammalian JAKs and STATs have been extensively studied over the past 20 years, many aspects of their in vivo function remain unclear. In particular, their roles in control of infection with pathogens remain murky and confusing. One difficulty in understanding how these pathways regulate inflammation is the presence of complex compensatory mechanisms between the different JAK and STAT proteins. In the Dionne lab, we use the fruit-fly Drosophila melanogaster as a model to study the in vivo functions of the JAK/STAT pathway in mycobacterial infection. Flies contain only one JAK (hop), one STAT (STAT92E), and three identified interleukin-like signals to activate signalling (upd, upd2, upd3). I show that, in Drosophila, the STAT-activating cytokine Upd3 is harmful to the host upon mycobacterial infection. Flies lacking upd3, or in which the JAk/STAT pathway is inhibited in phagocytes, show improved survival, decreased mycobacterial numbers, and delayed immune cell death. Strikingly, I find that JAK/STAT signalling acts in concert with other inflammatory signals to regulate expression of Atg2 in Drosophila phagocytes. In isolation or upon infection, STAT activation inhibits Atg2 expression and the ability of other, unknown signals to promote Atg2 expression. Increased Atg2 expression, as is seen in infected animals lacking either the cytokine Upd3 or STAT92E, promotes killing of intracellular bacteria and accumulation of large lipid droplets of unusual shape. I suggest an autophagy-independent mechanism by which Atg2 could reduce bacterial growth, involving the control of lipid body morphology. In this thesis, I show a mechanism by which JAK/STAT controls bacterial growth through inhibition of autophagy gene expression and demonstrate that this inhibition is detrimental to the survival of the host. In addition, I demonstrate that upd3 signalling is also required for glucose homeostasis suggesting a role for Upd3 in regulating gluconeogenesis and glycolysis.

Magnetic resonance imaging (MRI) has recently been shown to provide valuable information for image-guided ablation therapy used to treat patients suffering from cardiac arrhythmia. This requires isotropic high resolution anatomical information about complex structures such as atria and pulmonary veins. In addition, the visual-isation of fibrotic tissue in the myocardium can be important for treatment planning and assessment. One of the main challenges to obtaining images with a high isotropic resolution is respiratory motion. Although a wide range of different methods to minimise res-piratory motion artefacts has been presented, irregular breathing can still lead to unacceptably long scan times and scan abortions. Respiratory motion in patients must also be taken into consideration during the ablation procedure to ensure accu¬rate image guidance. Furthermore, arrhythmia leads to pathological changes in the electrical excitation of the heart. This can cause irregular heart beat variations and result in very long scan times for the functional assessment of the heart in patients suffering from arrhythmia. This thesis presents new MRI methods which overcome these problems and allow for the characterisation and compensation of physiological motion even in patients with highly irregular respiratory and cardiac cycles. A new high resolution 3D whole-heart acquisition scheme is introduced. It reduces scan times by 36% in both volunteers and patients with irregular breathing motion due to a higher respiratory navigator efficiency compared to a commonly used respiratory gating method. Furthermore, this approach not only yields anatomical information but also provides additional respiratory motion information without an increase in scan time. This information can be used to assess, and compensate for, respiratory motion during ablation procedures. This method was also modified for 3D high resolution assessment of myocardial scar tissue which led to a Furthermore, a novel technique to assess cardiac function without ECG using image-based navigation is presented. This allows for a synchronised multi-slice ac-quisition of the heart without the need of an external ECG and could provide new methods to address arrhythmic heart beats. In conclusion, the approaches detailed in this thesis provide imaging methods which may not only be beneficial for patients suffering from arrhythmia but also improve the accuracy, outcome and procedure time of other percutaneous procedures.

In dual modality PET-CT imaging, respiratory motion can introduce blurring in PET images and create a spatial mismatch between the PET and CT datasets. Attenuation correction errors can result from this mismatch, which can produce severe artefacts that potentially alter the clinical interpretation of the images. Various approaches of reducing these effects have been developed. Many involve respiratory gated acquisitions which generally require a measure of the respiratory cycle throughout imaging. In this work, a retrospective respiratory gating technique was devel¬oped for both PET and CT which extracts the respiratory cycle from the acquired data itself, removing the requirement for hardware that measures respiration. This data-driven gating method was validated with phantom and patient data, and compared with a hardware based approach of gating. Extensions to the method facilitated the gating of multi-bed position, 3D clinical PET scans. Finally, 60 Ammonia cardiac PET/CT images were used to compare several different ap¬proaches of reducing respiratory induced attenuation correction errors and motion blur. The data-driven respiratory gating method accurately substituted a hardware based approach, and no significant difference was found between images gated with either methods. Gating 11 clinical 3D whole body PET images validated the extended data-driven gating methods and demonstrated successful combination of separate PET bed-positions. All evaluated approaches to reduce respiratory motion artefacts in cardiac imaging demonstrated an average improvement in PET-CT alignment. However, cases were found where alignment worsened and artefacts resulted. Fewer and less severe cases were produced when the 4D attenuation correction data was created from a 3D helical CT and PET derived motion fields. Full motion cor¬rection produced a small effect on average, however in this case no detrimental effects were found.