Metabolomics studies of ALS a multivariate search for clues about a devastating disease

Abstract: Amyotrophic lateral sclerosis (ALS), also known as Charcot’s disease, motor neuron disease (MND) and Lou Gehrig’s disease, is a deadly, adult-onset neurodegenerative disorder characterized by progressive loss of upper and lower motor neurons, resulting in evolving paresis of the linked muscles. ALS is defined by classical features of the disease, but may present as a wide spectrum of phenotypes. About 10% of all ALS cases have been reported as familial, of which about 20% have been associated with mutations in the gene encoding for CuZn superoxide dismutase (SOD1). The remaining cases are regarded as sporadic. Research has advanced our understanding of the disease, but the cause is still unknown, no reliable diagnostic test exists, no cure has been found and the current therapies are unsatisfactory. Riluzole (Rilutek®) is the only registered drug for the treatment of ALS. The drug has shown only a modest effect in prolonging life and the mechanism of action of riluzole is not yet fully understood. ALS is diagnosed by excluding diseases with similar symptoms. At an early stage, there are numerous possible diseases that may present with similar symptoms, thereby making the diagnostic procedure cumbersome, extensive and time consuming with a significant risk of misdiagnosis. Biomarkers that can be developed into diagnostic test of ALS are therefore needed. The high number of unsuccessful attempts at finding a single diseasespecific marker, in combination with the complexity of the disease, indicates that a pattern of several markers is perhaps more likely to provide a diagnostic signature for ALS. Metabolomics, in combination with chemometrics, can be a useful tool with which to study human disease. Metabolomics can screen for small molecules in biofluids such as cerebrospinal fluid (CSF) and chemometrics can provide structure and tools in order to handle the types of data generated from metabolomics. In this thesis, ALS has been studied using a combination of metabolomics and chemometrics. Collection and storage of CSF in relation to metabolite stability have been extensively evaluated. Protocols for metabolomics on CSF samples have been proposed, used and evaluated. In addition, a new feature of data processing allowing new samples to be predicted into existing models has been tested, evaluated and used for metabolomics on blood and CSF. A panel of potential biomarkers has been generated for ALS and subtypes of ALS. An overall decrease in metabolite concentration was found for subjects with ALS compared to their matched controls. Glutamic acid was one of the metabolites found to be decreased in patients with ALS. A larger metabolic heterogeneity was detected among SALS cases compared to FALS. This was also reflected in models of SALS and FALS against their respective matched controls, where no significant difference from control was found for SALS while the FALS samples significantly differed from their matched controls. Significant deviating metabolic patterns were also found between ALS subjects carrying different mutations in the gene encoding SOD1.