Biofluid analysis to differentiate brain disease

It has been demonstrated that by using 1H NMR spectroscopy in combination with multivariate statistical modelling (PLS) it is possible, using urine samples obtained from rats, to distinguish between different types of CNS lesions. Against this background this thesis will explore whether the combination of 1H NMR and PLS modelling on biofluids can be used to identify biomarkers in different neurological diseases and in clinically relevant animal models of neurologic disease. The results in this thesis demonstrate that it is possible to separate sets of animals at different stages of disease in models of multiple sclerosis and to identify the presence of early brain metastasis. The same methodology was also applied to human biofluids. In MS patient cohorts (RR-MS, PP-MS and SP-MS) it was also possible to differentiate between RR-MS and SP-MS as well between MS and healthy controls. Therapy for these two stages of MS are very different and therefore a rapid test to determine a patient's stage of MS would be hugely beneficial in the clinic. Further investigation revealed that it is possible to separate MS patients from individuals with Alzheimer’s disease. Metabolomics was then combined with other co-variants in a study of cerebrospinal fluid obtained from patients with HIV associated dementia (HAD) to discover whether disease progression could be followed in this manner. The results show that it is possible to detect neurocognitive changes in patients with HAD. Indeed, the results demonstrate that metabolomics is a far more sensitive tool for the following progression than other non-PLS biomarker techniques and should provide a useful method for early diagnosis of CNS disease and the evaluation of therapy in prospective studies.