Cardio-embolic stroke accounts for nearly a third of all ischaemic strokes. The most clinically important cardio-embolic sources are non-valvular atrial fibrillation (AF) and chronic heart failure. Strokes due to these conditions are associated with greater disability and more mortality, as compared to stroke of other aetiology. This thesis is aimed at addressing some of the challenges faced by clinicians when dealing with stroke in patients with AF or heart failure, using an extensive range of historical data. Chapter 1 provides an introduction to stroke, AF and heart failure, including current prevalences, aetiology, and their complex intertwine relationship. The current acute stroke management in patients with AF or heart failure is also outlined within the chapter. In chapter 2, the data sources and statistical methods that were common to the studies in the thesis are outlined. The justifications of using historical data in the absence of evidence from robust clinical trials are also detailed. Chapter 3 explores the relevance of antithrombotic treatment on patterns and outcomes of acute stroke patients with AF. A non-randomised cohort analysis was conducted using data from the Virtual International Stroke Trials Archive (VISTA). The associations of antithrombotic treatment with the modified Rankin scale (mRS) outcome, and the occurrence of recurrent stroke and symptomatic intracerebral haemorrhage, at 90 days after stroke were described. Combined sequential antithrombotic therapy (i.e. oral anticoagulant and antiplatelet treatment), was associated with favourable outcome on ordinal mRS and significantly lower risk of recurrent stroke, symptomatic intracerebral haemorrhage and mortality by day 90, compared to the patients who did not receive any antithrombotic treatment. The relative-risk of recurrent stroke and symptomatic intracerebral haemorrhage appeared highest in the first 2 days after stroke before attenuating to become constant over time. Thus, early introduction of oral anticoagulant treatment (2-3 days after stroke), and to a lesser extent antiplatelet agents, was associated with substantially fewer recurrent stroke events over the following weeks but with no excess risk of symptomatic intracerebral haemorrhage. Chapter 4 seeks to describe the current prescribing patterns in stroke survivors with AF, with particular emphasis on socio-demographic associations. A cross-sectional analysis of city-wide Glasgow primary care data for the year 2010, was conducted. This chapter highlights that oral anticoagulant treatment was under-used in this high risk population, especially those of older age and affected by deprivation. Strategies need to be developed to improve prescription of oral anticoagulant treatment. Chapter 5 investigates the incidence of stroke within the available heart failure trials spanning a 30 year period, according to AF status at baseline. Individual patient data were pooled from 11 trials conducted in patients with heart failure and reduced ejection fraction (HF-REF); and, 3 trials performed in patients with heart failure and preserved ejection fraction (HF-PEF). Stroke incidence has not significantly declined over time in patients with HF-REF enrolled to trials, despite greater use of evidence-based heart failure and oral anticoagulant therapies. However, anticoagulation proportions remain under 70% among HF-REF patients with documented AF. Similar trends of stroke incidence were observed for patients enrolled in HF-PEF trials. Some patients with heart failure but without atrial fibrillation may be at high risk of stroke and may potentially benefit from oral anticoagulant treatment. Chapter 6 provides a comprehensive description of the current incidence of and risk factors for stroke in patients with HF-REF but without AF. Data from two large and contemporary heart failure trials, the Controlled Rosuvastatin in Multinational Trial Heart Failure (CORONA) and the Gruppo Italiano per lo Studio della Sopravvivenza nell'Insufficienza cardiac- Heart Failure trial (GISSI-HF), were pooled to enable the analysis. The new simple clinical predictive model for stroke showed that about one-third of patients without AF have a risk of stroke similar to patients with AF. The predictive model was also validated in an independent large data set. The high risk of stroke in patients without AF might be reduced by individualised and safer oral anticoagulant treatment. Correspondingly, Chapter 7 explores the risk-model for stroke in a contemporary cohort of patients with HF-PEF but without AF. Data were pooled from the Candesartan in Heart failure Assessment of Reduction in Mortality and Morbidity- Preserved trial (CHARM-Preserved) and the Irbesartan in Heart Failure with Preserved Systolic Function trial (I-Preserve), for patients with ejection fraction ≥45% only. The analysis showed that the simple clinical model developed in Chapter 6, for patients with HF-REF, is also applicable to patients with HF-PEF. There are concerns that systemic thrombolysis might not achieve clinically-important outcome among chronic heart failure patients with acute ischaemic stroke. Chapter 8 evaluates the relevance of chronic heart failure on the outcome of acute stroke patients who received thrombolysis. A non-randomised cohort analysis was conducted using data obtained from the Virtual International Stroke Trials Archive (VISTA). The associations of outcome among chronic heart failure patients with thrombolysis treatment using the mRS distribution at day 90, stratified by presence of AF, were evaluated. Chronic heart failure was associated with a worse outcome with or without thrombolysis. However, acute stroke patients who received thrombolysis had more favourable outcome regardless of heart failure status, compared to their untreated peers. The findings should reassure clinicians considering systemic thrombolysis treatment in hyper-acute ischaemic stroke patients with chronic heart failure. This thesis has summarised and extended our knowledge of the complex relationship between stroke and the heart, focusing on atrial fibrillation and heart failure. It has answered many questions and generated many more. The reported studies may assist clinicians who are dealing with stroke in patients with atrial fibrillation or heart failure. These conditions are common and each carry poor prognosis. Thus, even small advances in their treatment may have a useful societal impact.

Background: The syndrome of heart failure is common, and is associated with high morbidity and reduced life expectancy. Patients can experience high symptom burden, low mood, and impaired quality of life. Repeated, and often prolonged, hospitalisations due to exacerbations of heart failure and other co-morbidities are common. Over the last 25 years, the evidence base for the treatment of heart failure has increased, with an associated improvement in prognosis. However, many patients with heart failure still have a poor prognosis. International guidelines for the treatment of heart failure now suggest referral to palliative care services, particularly in patients thought to have a poor prognosis and impaired quality of life. Despite these recommendations, few patients with this condition have access to specialist palliative care services in the United Kingdom. However, not every patient with heart failure will have palliative care needs, therefore the extent of the problem of unmet palliative care needs in patients with heart failure is unknown. I systematically reviewed the published literature to identify studies describing the palliative care needs, including prevalence, of patients with heart failure. Although my search identified over 60 publications describing the palliative care needs of patients with heart failure, most of the studies were of highly selected cohorts, did not include descriptions of therapy, or descriptions of severity of heart failure such as ejection fraction, natriuretic peptides, prognostic scores or clinical outcomes. Most studies used a cross-sectional approach to describe the potential palliative care needs, and therefore, were unlikely to appreciate the variable clinical course of patient with heart failure. Although the studies identified were informative, a definitive description of the prevalence of palliative care needs in a well described, contemporary cohort of patients with heart failure is lacking. My systematic review also identified a number of preliminary randomised controlled clinical trials, assessing the effect of early palliative care in patients with heart failure. However, these studies included small numbers of participants, and only had qualitative endpoints such as change in quality of life measures without assessment of clinical outcomes such as death or hospitalisation. Again, although these preliminary trials are informative, a definitive evidence base comparing palliative care to standard care in heart failure is not available. Aims: The primary aim of this study was to inform the design of a randomised controlled clinical trial of palliative care in patients with heart failure. The first step in this process was to define the clinical problem and identify a suitable target population by describing the prevalence of patients with heart failure who have palliative care needs. I then aimed to describe whether these patients could be identified from data collected during an index hospital admission. The final aim of my study was to identify useful outcome measures which could be used in a randomised controlled clinical of palliative care in heart failure. Methods: This was a prospective, longitudinal study of the prevalence of possible palliative care needs, defined using quantifiable patient reported outcome measures. An unselected cohort of patients admitted to hospital with a primary diagnosis of heart failure were recruited and extensively characterised. The World Health Organisation definition of palliative care was used to identify patients with heart failure who had palliative care needs. I made objective assessments of quality of life (using the Kansas City Cardiomyopathy and Short Form 12 questionnaires), mood disturbance (using the Hospital Anxiety and Depression Scale), symptom burden (using the Edmonton Symptom Assessment Scale), and caregiver strain (using the Zarit Burden Interview questionnaire). These assessments were made at baseline and repeated every four months for the duration of the study. Patients were identified as having palliative care needs if they had persistently severe impairment of any patient reported outcome measure without improvement, or severe impairment of any patient reported outcome measure followed by death. End-of-life care was assessed using the Views Of Informal Caregivers Evaluation of Services questionnaire, and by comparing preferred place of end of life care to actual place of death. Multivariate logistic regression analysis was used to determine if baseline prognostic markers, physician completed assessments, or patient reported outcome measures could identify patients with palliative care needs. Results: Between January 9th 2013 and December 1st 2014, 313 near consecutive patients with heart failure were enrolled in the study. Of these, 272 (86.9%) completed patient reported outcome measures at baseline and agreed to attend study visits. Patients were elderly, with a median [interquartile range] age of 76 [70-82] years, and 47% of participants were female. 56% of patients did not have a previous diagnosis of heart failure. Most participants had heart failure with reduced ejection fraction (67.3%) compared to heart failure with preserved ejection fraction (32.7%). Use of disease modifying pharmacotherapy was high, especially in participants with heart failure with reduced ejection fraction. Participants suffered from a number of physical and psychological symptoms, as recorded using patient reported outcome measures. The most common physical symptoms were shortness of breath and fatigue, followed by drowsiness and lack of appetite. Although less frequent, pain and nausea were also common. Participants reported higher scores for depression and anxiety compared to studies using similar mood assessments in the general population. Quality of life was impaired in most participants at baseline, with 77.9% of participants being classified as having moderate or severe impairment as assessed by the Kansas City Cardiomyopathy questionnaire. At baseline, 114 (41.9%) participants scored severe in at least one patient reported outcome measure. Of these, 95 (83%) participants scored severe on the Kansas City Cardiomyopathy Questionnaire. Participants were invited to attend study visits, or have home study visits, every four months for the duration of the study. The minimum number of study visits offered was two for the last participant enrolled. A total of 691 study visits were performed. 37% of these assessments were home visits. Participants were also followed up passively using record linkage to report number and cause of hospitalisations, and cause and location of any deaths. Participants were followed up for a minimum of one year. During follow-up, 217 (79.8%) participants were re-admitted to hospital. The median number of admissions was 3. Most hospitalisations were due to non-cardiovascular causes. During passive follow-up until December 1st 2015, there were 103 (37.8%) deaths. Most (60.2%) deaths were due to cardiovascular causes. 73 (26.8%) participants met my criteria for having palliative care needs. These patients had worse summary scores at baseline for all patient reported outcome measures. Patients who met my definition of palliative care needs spent fewer days alive and out of hospital than the group who did not meet the definition of palliative care needs. The median [IQR] days alive out of hospital in the group meeting the definition of palliative care needs was 394 [172-586], compared to 638 [420-809] in the group not meeting the definition of palliative care needs (p<0.001). After adjusting days alive out of hospital for quality of life, patients in the palliative care needs group had fewer days of good health as a percentage of total follow-up, median 12 [3-22] % of potential follow-up, compared to 47 [25-68] % in those not meeting my definition of palliative care needs (p < 0.001). / Most participants expressed a wish to spend the end of their life at home, but despite this, most died in hospital. 17 caregivers completed the Views Of Informal Caregivers Evaluation of Services questionnaire. Overall care in the last few months of life was assessed as fair or poor by 35.3%. Of the 272 participants who participated in the whole study, 33 (12.1%) had access to specialist palliative care services. Of the 73 participants who met the definition of PC needs, 19 (26.0%) accessed specialist PC services. 6 (2.2%) participants used hospice care during the duration of the study. Using multivariate logistic regression analysis, a low Kansas City Cardiomyopathy Questionnaire summary score and a low Australia Modified Karnofsky Performance Scale (a physician completed assessment) score, were predictive of patients with palliative care needs. Conventional prognostic markers, such as natriuretic peptides or ejection fraction, were not predictive of patients with palliative care needs. Physicians, using their clinical judgement, were only modestly accurate at predicting patients with heart failure who had or would go on to develop palliative care needs. Physicians were better at predicting prognosis than need for palliative care.

Background: The process of decreasing ineffective or harmful healthcare (deimplementation) may require different approaches than those used to promote uptake of new procedures (implementation) but research into different approaches is currently lacking. It has not been determined if there is a theoretical and evidence-based rationale for designing different interventions for implementation and de-implementation. Objectives: The objectives of this thesis were to: 1) Investigate whether there is a theoretical basis for identifying different mechanisms of change by which behaviour might increase versus decrease; 2) Assess whether predictors of health professional behaviour differ depending on whether behaviour was one clinicians should implement or behaviours clinicians should de-implement; and 3) Identify the Behaviour Change Techniques of published implementation and de-implementation interventions to determine if there is a difference between the techniques reported in these interventions. Methods: Study 1 used Critical Interpretive Synthesis to investigate whether a theoretical rationale exists for identifying different mechanisms of change by which interventions may work for implementation and de-implementation. Study 2 investigated whether the theoretical constructs commonly used to predict health professional behaviour differ based on whether the behaviours should be implemented or de-implemented. It was an exploratory study involving secondary analysis on 13 existing questionnaire datasets from a variety of healthcare professional groups in primary care settings in the United Kingdom and Canada. Study 3 involved the classification and frequency of Behaviour Change Techniques in implementation and de-implementation interventions from selected Cochrane Effective Practice and Organisation of Care systematic reviews. Findings from these three studies were interpreted using the concurrent triangulation approach to report on the key findings. Results: Operant Learning Theory (OLT) proposes different approaches to increasing and decreasing behaviour changes and therefore implementation and de-implementation interventions (Studies 1 & 3), despite a number of commonly used theories being poor predictors of behaviours for implementation and de-implementation (Study 2). Additionally, whilst the range of techniques was limited, the technique Behaviour substitution was frequently used to decrease health professionals’ behaviours (Study 3) and also identified as a strategy commonly used to decrease behaviour in general (Study 1). Conclusion: Whilst the findings suggest that OLT may be promising in developing different interventions for implementation and de-implementation, how best to use these OLT principles is unclear. In instances whereby Behaviour substitution is part of a de-implementation intervention, it is not clear how best to identify the substitute behaviour. Additional investigation is required to better inform the design of implementation and de-implementation interventions.

The increasing global prevalence of diabetes poses a huge challenge to health services. The diagnosis is accompanied by a reduction in life expectancy, primarily due to cardiovascular disease which is inextricably linked to microvascular complications such as diabetic nephropathy (DN). Microalbuminuria (MA) is generally accepted as the primary clinical hallmark of DN, but despite widespread prescribing of agents blocking the renin angiotensin aldosterone system (RAAS) in these patients many continue to progress towards end-stage renal disease (ESRD). Clinical trials evaluating early initiation of RAAS blocking agents in untargeted, nonalbuminuric diabetic patients have shown potential for delaying disease progression but these effects are generally counterbalanced by side effects and adverse events associated with these therapies. Discovery of novel biomarkers to identify individuals at highest risk of DN who would stand to benefit most from targeted preclinical intervention would be a significant step towards implementation of personalised medicine in this population. One technique which shows promise is proteomics, based on the concept of separation and quantification of peptides in a biological sample to produce a disease-specific pattern. A panel of 273 urinary peptides (CKD273) has been shown to have potential for identification of nonalbuminuric diabetic patients who are at risk of progression to overt DN. However, many such novel biomarkers are described in the literature and to date none have successfully made the transition from research studies to routine clinical practice. In order to be considered for clinical implementation novel biomarkers are required to be subject to a rigorous evaluation process. In brief there are several key steps beginning with proof-of-concept studies; progressing through validation in independent populations to demonstration of incremental value beyond the current guideline-endorsed tests; thereafter proof of clinical applicability in determining treatment strategies and cost-effectiveness are required. The work contained within this thesis is designed to address each of these aspects with regard to use of the CKD273 proteomic panel as a biomarker for early detection of DN.

No description available.

In Graves’ orbitopathy (GO) tissue remodelling by increased proliferation, excess adipogenesis and hyaluronan overproduction, cause exophthalmos. My initial work followed reports by others of enophthalmos in some glaucoma patients treated with Bimatoprost (prostaglandin F2α, PGF2α) eye drops. I hypothesized that this could be due to reduced proliferation or adipogenesis and/or increased lipolysis; any of which could improve GO. In vitro models used to investigate possible mechanisms demonstrated that PGF2α reduced proliferation by prolongation of G2/M (flow cytometry) and adipogenesis (evaluated morphologically, by oil red O staining and QPCR measurement of adipogenesis markers) of 3T3-L1 and human orbital fibroblasts (OF). These effects were reversible upon drug withdrawal. GO OFs proliferated significantly more rapidly and also displayed higher adipogenic potential than non-GO. There was no effect of PGF2α, on basal or norepinephrine-induced lipolysis in 3T3-L1 or human OFs, either GO or non-GO. The data helped us secure NISCHR funding for a clinical trial ‘Prostaglandin F2-alpha eye drops (Bimatoprost) in thyroid eye disease: a randomised controlled double blind crossover trial’. Following informed consent, 31 clinically inactive (late phase) GO patients were randomised to receive Bimatoprost or placebo eye drops daily for three months followed by a two-month drug washout period before switching to the opposite treatment for three months. I concluded that 3 months Bimatoprost treatment is not effective in reducing proptosis in late phase GO patients. Mysubsequent work investigated truncated TSHR variants (TSHR_v2) and thyrostimulin. The former could act as TSH/TRAB binding proteins whilst the latter could bind to and activate the TSHR and contribute to GO pathogenesis. I found no evidence of a role for thyrostimulin in GO. TSHR_v2 transcripts and protein are more abundant than full length TSHR in OF and during adipogenesis are significantly higher in GO than non-GO. TSHR_v2 may be secreted and provide a binding protein for TSHR ligands and thus alter intracellular TSHR signalling. GD patients lose weight during active disease but gain more following treatment. Our group reported that TSHR activation leads to a ‘browning’ of fat from various depots. I performed ex vivo analysis on neck fat for brown, beige and white adipose tissue markers. The samples were GD patients (previously hyperthyroid/positive TSAB), toxic goitre (previously hyperthyroid/no TSAB) and euthyroid. I found no difference in expression of UCP1 and PGC1α. There were general reductions in ZIC1, CITED1, HOXC9 and LEPTIN markers in GD which may explain the altered body composition in these patients.

This medical doctorate thesis contains clinical studies to broaden the application and to improve the safety and efficacy of ultraviolet (UV) A phototherapies, the main focus being to enhance the current clinical practice of topical psoralen photochemotherapy (PUVA). The thesis includes three studies: 1. The validation of a semi-automated Minimal Phototoxic Dose (MPD) Tester for topical photochemotherapy Thirty seven psoriasis patients referred to the phototherapy unit at St. Woolos, Newport were recruited. Patients had two sets of minimal phototoxic dose (MPD) tests performed on symmetrical, contralateral sites on the lower back. MPD test results from a panel of PUVA-lamps with a UV-opaque template and windows were compared to those from the modified hand-held MPD tester. The hand-held MPD results were linearly related to the PUVA-panel MPD results and this was therefore shown to be a convenient and reliable method of assessing MPD. However, the difference in MPD between the PUVA lamp and the modified handheld MPD tester (CFL TL-10 lamp) was much less than predicted from the PUVA action spectrum of previously published studies suggesting that formal re-evaluation of the erythema action spectrum for PUVA was now appropriate.

Alzheimer’s disease (AD) is the most common form of dementia among the elderly. The disease is characterized by amyloid-β (Aβ) plaques, neurofibrillary tangles, loss of synapses and neurons, and chronic neuroinflammation. The significance of neuroinflammatory processes in the on-set and progression of AD have been debated, as activated and reactive glial cells demonstrated protective and damaging properties. However, patients with AD treated with anti-inflammatory drugs demonstrated a lesser than average extent of disease development, indicating an important role of neuroinflammation in AD. Current strategies for the treatment of Alzheimer’s disease are minimally effective, as we lack the methods of diagnosis at an early stage. Numerous cases present significant neuronal loss prior to diagnosis and, as restoration of function is unlikely, treatment options focus on limiting further neuronal loss. It is my hypothesis that deposition of Aβ peptide can activate the innate immune system via pattern recognition receptors (PRRs), including complement, and evoke Alzheimer's pathology. In the current study, we focused on the role of the innate immunity system of the brain in the initiation and the propagation of inflammatory process in AD and the interplay between TLRs/NLRs and the complement system. Silencing the expression of various receptors demonstrated the involvement of NLRP3 in Aβ recognition, and use of confocal microscopy confirmed the association of Aβ interactions with this inflammasome and complement receptors. These results were confirmed in human astrocytes stimulated with Aβ and in brain tissue slides of patients with AD. Using human astrocytes and fluorescence resonance energy transfer (FRET), the associations of cell surface TLRs (such as TLR2, TLR4)with CRs (C5aR1, C3aR) were investigated; TLR2 and TLR4 were 6 | P a g e found to interact with C5aR1. In addition, Aβ-induced responses were augmented in the presence of complement and Aβ appeared to induce the interactions of TLR4 and C5aR1 on the cell surface. The trafficking of PRRs and Aβ in human astrocytes was further investigated to determine whether internalization and trafficking of Aβ is crucial for triggering a proinflammatory response. The results demonstrated that, along with Aβ and TLR4, complement internalized in the Rab5+ endosomes, and recruited NF-kB-inducing kinase (NIK) to these endosomes in a complement-dependent manner. MAC complexes were previously demonstrated to activate an Akt+NIK+ signalosome in human endothelial cells. To the best of our knowledge, this is the first to identify such signalosomes in Rab5+ endosomes in response to complement and Aβ. These results suggest a novel possibility of a therapeutic target in the case of an Aβ-induced neuroinflammation. C5a is the most potent anaphylatoxin, thus upon the identification of the complement-inflammasome interactions, the role of C5a in the Aβ recognition was investigated in human astrocytes. The results demonstrated that C5a augmented the Aβ-induced IL-1β production, activated pro-IL-1β and was caspase-1-dependent. In addition, a C5a-induced IL-1β release was observed following Signal 2 stimulation by Ca2+ and lysosomal damage. Since the signal activation resulting in inflammation in AD appears to involve multiple receptors, the therapy should be combinational and not targeting single molecules. Thus, C5aR1 blocking antibodies and PMX53 were utilized to identify a combinatory therapeutic approach in human astrocytes. The results demonstrated that the secretion of IL-1β was further reduced in the combinatory treatment of PMX53+MCC950 7 | P a g e compared with their single treatments, suggesting a combinatory treatment to tackle the effect of Aβ in human cells. The results of the present study add on to the current knowledge of the molecular mechanisms involved in AD, regarding Aβ association, and may lead to the design of targeted therapeutic interventions for AD.

The aim of this thesis was to assess the relationships between glycaemic exposure and β-cell function with prevalence, incidence and progression of diabetic retinopathy (DR) over 5 years in newly-diagnosed treatment-naïve subjects with type2 diabetes mellitus (T2DM). At diagnosis, we studied 544 subjects and demonstrated DR was independently associated with fasting and postprandial hyperglycaemia and reduced fasting β-cell function during standardized meal and intravenous glucose challenge. The insulin-independent component of glucose tolerance (SG) was also impaired and independently associated with presence of DR at diagnosis. We followed up 233 subjects over 5 years and established independent association between chronic glycaemic exposure (HbA1c,/fasting/ postprandial hyperglycaemia) at diagnosis and incident DR during this period. We also demonstrated that fasting and postprandial β-cell responsiveness to nutrient challenge along with SG at baseline was independently associated with development of DR over 5 years. There was no difference in glycaemic status between those with or without DR at 5 years highlighting the relevance of early history of glycaemic exposure in our subjects to future incidence of DR. Finally, in 45 T2DM subjects with DR at diagnosis, fasting, postprandial glucose and HbA1c along with fasting, postprandial β-cell responsiveness at diagnosis were all independently associated with DR progression. Thus, this study has indicated that hyperglycaemia resulting from pancreatic β-cell deficiency contributes to the risk of development and progression of DR. The data emphasises the need for earlier diagnosis of T2DM and cautious normalization of glycaemia to eliminate glucotoxicity on the already impaired β-cell function. The evidence indicates the potential value of supporting β-cell function aiming to achieve near-normal glycaemia and thus preventing the onset and progression of DR in subjects with T2DM.

Regulatory B cells, producing IL-10, have been studied in many autoimmune diseases. However, less is known about these cells in Type 1 diabetes (T1D), a disease characterized by the destruction of beta-cells by the immune system, leading to deficient production of insulin. Although B cells play a role in development of T1D, most previous investigations have focused on their pathogenic involvement. B cell depletion has been shown to be protective against diabetes development. To examine regulatory B cells in T1D, we used Non-Obese Diabetic (NOD) mice and non-diabetes-prone B6g7 mice as controls. We compared both strains for the production of cytokines and expression of putative regulatory phenotypes in spleen B cells cultured with various stimulants, at different ages. We observed that NOD mice that were > 35 weeks old and naturally protected against T1D had more IL-10-producing B cells than B6g7 and diabetic NOD mice, and this number increased even more on stimulation with lipopolysaccharide (LPS) stimulation. When LPS-stimulated B cells from protected mice were cultured in vitro with CD8 T cells and DCs, their potential for suppression of T cell cytotoxic activity was higher than unstimulated B cells and B cells from diabetic mice. This inhibitory effect was associated with higher levels of IL-10. Lastly, we carried out an investigation where B cells were transiently depleted in transgenic NOD mice expressing human CD20, to enable depletion using a human anti-CD20 monoclonal antibody. We evaluated the effect of depletion and repopulation on regulatory B cells, testing whether the protection afforded by B cell depletion was due to a change in regulatory B cell number or function. B cells with putative regulatory phenotypes were susceptible to depletion and, although the treatment with anti-CD20 reduced the incidence and delayed the onset of diabetes, there was no difference in the IL-10 producing B cell population by the time of full repopulation of B cells. Thus, this protective effect of B cell depletion was unlikely to be due to IL-10-producing B cells. In conclusion, for the first time, regulatory B cells were extensively studied in NOD mice and we demonstrated that protected NOD mice had higher frequencies of spleen B cells producing IL-10 than diabetic NOD mice. Further investigation is warranted to understand how these IL-10-producing B cells contribute to protection against type 1 diabetes.