Proteomic investigation of blood-based biomarkers for the diagnosis of HIV-associated neurocognitive disorders

Includes bibliographical referencesBackground: South Africa has the fourth highest prevalence of Human Immunodeficiency Virus (HIV) in the world, with over 340 000 new infections in 2013 alone. Implementation of Highly Active Antiretroviral Treatment (HAART) regimens, has improved the lives of many HIV-infected individuals, and has led to a decrease in the incidence of extreme forms of HIV-associated dementia. However, the extended lifespan afforded by HAART has also resulted in an increase in the prevalence of milder forms of HIV-associated Neurocognitive Disorders (HAND). Clinical diagnosis of such cognitive decline currently relies on neurocognitive assessments, as no definitive biochemical test exists. A point-of-care test designed for use with easily accessible samples (e.g. blood) would be of great utility in both HAND research and clinical diagnosis, prognostication, and provision of timeous therapy. A blood-based biomarker test would also be less subjective than neurocognitive testing, which can be adversely affected by the patient's level of education as well as operator competency. Recently, Professor Simon Lovestone (University of Oxford) developed a panel of blood-based biomarkers for the diagnosis of Alzheimer's disease. Due to pathological similarities between HAND and Alzheimer's disease, it is hypothesised that there may be applicability of this panel to diagnosis of HAND. The differential protein expression profile of patients with and without HAND is thus compared. Aims and Objectives: The aim of this study was to assess the applicability of a set of Alzheimer's disease biomarkers to the diagnosis of HAND. In order to achieve this, targeted proteomic assays were developed to measure the markers in human serum. This assay was then applied to defined patient cohorts. Methods: Targeted proteomic assays were developed in order to quantify candidate markers in patient blood serum. Parallel Reaction Monitoring assays were developed on a Thermo Q Exactive Quadrupole-Orbitrap mass spectrometer, for application to defined patient samples collected from Groote Schuur Hospital (Cape Town, South Africa). Patient samples were divided into groups according to HAND severity (based on the HIV Dementia Scale), namely: Normal (HIV+), Minor Neurocognitive Disorders and HIV-associated Dementia. Discovery proteomic analysis was performed on pooled patient samples in order to determine the optimal peptides for identification and quantitation of candidate proteins. Targeted methods were then developed and refined using the same pooled samples. Finally, patient samples (N=81) from the three classes were analysed in a statistically rigorous manner. Data were normalised to correct for possible loading inconsistencies by two independent methods. Quantitative differences were investigated using t-tests and non-parametric statistical tests where appropriate. Results: Statistical assessment of final data indicated no significant differences in proteins between patient classes. Discussion and Conclusions: It is likely that the inflammatory nature of HIV itself influences the levels of these markers in HIV-positive patient samples to a greater degree than neurocognitive effects. Alternatively, possible co-infection by TB may have confounded the results. Regardless, it was concluded that the candidate Alzheimer's biomarkers were not applicable to diagnosis of HAND. Further analysis on a larger sample cohort is recommended, utilising the assays optimised herein. Prospective studies to obtain viable biomarkers for definitive diagnosis may lie in proteomic analysis of model infection systems and cerebrospinal fluid