Quantitative proteomics of delirium cerebrospinal fluid

OPEN Citation: Transl Psychiatry (2014) 4, e477; doi:10.1038/tp.2014.114 © 2014 Macmillan Publishers Limited All rights reserved 2158-3188/14 www.nature.com/tp ORIGINAL ARTICLE Quantitative proteomics of delirium cerebrospinal fluid A Poljak1,2,3, M Hill2, RJ Hall4, AM MacLullich4, MJ Raftery1, J Tai2, S Yan2 and GA Caplan5,6 Delirium is a common cause and complication of hospitalization in older people, being associated with higher risk of future dementia and progression of existing dementia. However relatively little data are available on which biochemical pathways are dysregulated in the brain during delirium episodes, whether there are protein expression changes common among delirium subjects and whether there are any changes which correlate with the severity of delirium. We now present the first proteomic analysis of delirium cerebrospinal fluid (CSF), and one of few studies exploring protein expression changes in delirium. More than 270 proteins were identified in two delirium cohorts, 16 of which were dysregulated in at least 8 of 17 delirium subjects compared with a mild Alzheimer’s disease neurological control group, and 31 proteins were significantly correlated with cognitive scores (mini-mental state exam and acute physiology and chronic health evaluation III). Bioinformatics analyses revealed expression changes in several protein family groups, including apolipoproteins, secretogranins/chromogranins, clotting/fibrinolysis factors, serine protease inhibitors and acute-phase response elements. These data not only provide confirmatory evidence that the inflammatory response is a component of delirium, but also reveal dysregulation of protein expression in a number of novel and unexpected clusters of proteins, in particular the granins. Another surprising outcome of this work is the level of similarity of CSF protein profiles in delirium patients, given the diversity of causes of this syndrome. These data provide additional elements for consideration in the pathophysiology of delirium as well as potential biomarker candidates for delirium diagnosis. Translational Psychiatry (2014) 4, e477; doi:10.1038/tp.2014.114; published online 4 November 2014 INTRODUCTION Delirium is a common cause and complication of hospitalization, particularly prevalent in the elderly and within intensive care units. Although it is an acute condition from which most patients recover, it is also associated with multiple adverse outcomes including higher risk of future dementia as well as progression of existing dementia and higher mortality.1–3 The aetiology of delirium is thought to involve neurotransmitter changes arising from diverse sources including hypoxaemia, metabolic derangements, disturbance of the sleep–wake cycle, drug effects or systemic inflammation.4,5 However there is a paucity of data relating to biochemical changes in delirium cerebrospinal fluid (CSF).6 There is much speculation; however, little evidence is available on which biochemical pathways are dysregulated in the brain during delirium episodes, whether there are protein expression changes common among delirium subjects and whether there are protein changes which correlate with delirium severity. Further, there is no available biochemical test which could facilitate delirium detection. Quantitative proteomics using iTRAQ tags is a well-established discovery-based tool, which allows unbiased evaluation of dysregulated protein expression in complex biological samples such as cellular and tissue extracts and body fluids. It has wide application in the biological and biomedical sciences and is increasingly applied to the study of neurodegenerative diseases such as Alzheimer’s disease (AD).7–9 However, to date, few studies of delirium have utilized proteomics approaches.10,11 None have utilized CSF which is the body fluid most likely to reflect biochemical changes in conditions which disrupt central nervous system function, and only one proteomics study has been reported on delirium subjects.10 This study utilized urine from post-cardiac surgery intensive care unit subjects, and applied MALDI-TOF mass spectrometric profiling to the intact proteins, which provided some quantitative data but precluded sequence identification. Furthermore few proteins were quantified, none of which were delirium specific. Another proteomics study of delirium plasma is in the planning phase,11 and will utilize SELDI-TOF, which is typically used for proteomics profiling, but will preclude protein sequence identification. In the current study, we used CSF samples from delirium subjects and neurological controls (Sydney study group),12 to identify proteins with altered expression levels relative to a mild dementia control group. We then validated the observed changes using a second study group based in Edinburgh (Edinburgh study group). Dysregulated proteins were analysed using bioinformatics tools DAVID and STRING to determine if functional relationships would emerge. This approach provides both protein sequence identification and quantification of relative protein expression changes, facilitating quantification of potential biomarkers and a better understanding of the pathology at a molecular level. MATERIALS AND METHODS Subjects Sydney study. Patients admitted to the Geriatric Medicine Unit at the Prince of Wales Hospital were screened for delirium in the Emergency Department,12,13 where they were admitted for a variety of medical diagnoses triggering the delirium, including infections, metabolic problems and adverse drug reactions, but on average were suffering from at least two identifiable causes of delirium, and on the Geriatric Medicine 1 Bioanalytical Mass Spectrometry Facility, University of New South Wales, Sydney, NSW, Australia; 2School of Medical Sciences, University of New South Wales, Sydney, NSW, Australia; 3Center for Healthy Brain Ageing, University of New South Wales, Sydney, NSW, Australia; 4Edinburgh Delirium Research Group, University of Edinburgh, Edinburgh, Scotland, UK; 5Prince of Wales Clinical School, University of New South Wales, Sydney, NSW, Australia and 6Department of Geriatric Medicine, Prince of Wales Hospital, Sydney, NSW, Australia. Correspondence: Dr A Poljak, Bioanalytical Mass Spectrometry Facility, University of New South Wales, Anzac Pde, Kensington, Sydney, NSW 2052, Australia. E-mail: Received 23 May 2014; revised 5 September 2014; accepted 11 September 2014 Quantitative proteomics of delirium A Poljak et al 2 Table 1. Baseline characteristics of Sydney and Edinburgh Delirium Study groups Sydney Delirium Study Age Sex (F:M) Mini-mental state exam (/30) Informant questionnaire on cognitive decline (/5) Confusion assessment method Delirium index APACHE III (acute physiology and chronic health evaluation III) index Geriatric Depression Scale (/15) Barthel score (/20) Instrumental activities of daily living (/12) Charlson comorbidity index Edinburgh Delirium Study Age Sex (F:M) Mini-mental state exam (/30)b Infor (...truncated)