Proteomic Approaches to Biomarkers of Alzheimer s Disease

Size: px
Start display at page:

Download "Proteomic Approaches to Biomarkers of Alzheimer s Disease"

Transcription

1 Proteomic Approaches to Biomarkers of Alzheimer s Disease

2 Disclosure: patent pending in conjunction with Rules Based Medicine Funding: Texas Alzheimer s Research & Care Consortium, U.S. Environmental Protection Agency, CH Foundation

3 Review Recent Findings Discuss Limitations to Current Approaches Provide an Overview to the STAR-B Working Group

4 A wealth of literature links a range of proteins with AD status More recent work has taken advantage of advancements in bioinformatics to combine multiple markers These recent advancements have yielded impressive progress towards blood-based screeners

5 Ray and colleagues (2007) filter-based arrayed sandwich ELISAs 18 plasma markers 89% overall classification accuracy 81% correct prediction of MCI AD conversion O Bryant and colleagues (2010, 2011, 2011) RBM Luminex xmap 108 serum proteins 91% overall classification accuracy for biomarkers alone 95% correct classification full algorithm 30 serum proteins 91% accuracy for biomarkers alone 94% full algorithm 11 serum plasma proteins (TARCC & ADNI) 0.70 accuracy biomarkers alone 0.88 accuracy

6 Laske et al 2011 Individual ELISAs 3 serum markers 87% overall correct classification Reddy et al 2010 Identified IgG antibody biomarkers in serum (AD peptoids 1-3) Overall accuracy was 99% Others Buerger 2009 expression of microcirculation peptides Booij et al 2011 gene expression analysis Rye et al 2011 gene expression analysis AIBL Study with cross-validation in ADNI in press

7 Dementia Risk CRP, fibrinogen, and Plasma Aβ have been linked to AD/dementia risk Combination of multiple markers, clinical variables, and demographics will likely yield optimal predictive models (i.e. Framingham-like) Use in Clinical Trials Blood-based biomarker profiles may point towards endophenotypes for tailored treatment Inflammatory endophenotypes Neurotropic endophenotypes Do they vary by APOE4 genotype? Biomarker of disease severity Biomarker risk score that accounts for 40% of the variance in CDR scores Biomarkers of cognitive functioning CRP linked to memory and executive functioning prospectively We have generated specific biomarker algorithms for cognitive domains 50% variance in immediate and delayed memory 27-37% variance in language scores 30-32% of variance in executive functioning scores Bettcher 2012; Komulainen 2007; Wershing 2010; Wilson 2009; Graff-Radford 2007; Schmidt 2002; O'Bryant 2011; Nilsson 2011; Van Oijen 2005; Yaffe 2011

8 Pre-Analytic Methods Serum versus plasma Our data explicitly demonstrates the comparability and drastic discrepancies between markers across blood fractions We have created a blood-screener that works across both Fasting versus non-fasting Sample handling procedures Assay Platforms No set standards for multi-plex platform performance Batch effects No prior work validating multi-plex findings across platforms Ray et al findings did not cross-validate Impact of APOE4 genotype on these biomarkers Data in press Our unpublished data CRP APOE μg/ml versus APOE4-2.92μg/mL O Bryant, 2011; Soares 2009; O Bryant et al in prep; Barber in prep

9 Lack of Consensus on Statistical Approach Capitalization on cutting-edge bioinformatics is a plus, but how to implement them on an individual basis? Should have comparable findings across methods Lack of normative data on most markers CRP levels vary by ethnicity Normalized versus raw data? Failure to take into account stage of disease Lack of consideration of application intended for marker/approach Clinic based versus community based issues (i.e. base rates) Screen follow-up with other modalities? Lack of Consideration of Race/Ethnicity APOE4 genotype varies by region, race/ethnicity CRP related to CVD events among non-hispanic whites only Fibrinogen related to CVD health outcomes in non-hispanic whites, African Americans and Hispanics CRP related to poorer cognition and increased disease severity among AD cases CRP appears not to be linked to these measures among Mexican American AD cases Coutinho 2011; O Bryant 2010; Ward 2012; Haan 2003; Veeranna 2012

10 International working group created to provide working document that Where appropriate, provides explicit standards Where standards not yet ready, provides guidelines Will be updated as needed based on scientific evidence Assist pushing markers beyond individual labs to practical implementation

11 Topics to be addressed Uses of markers Pre-analytic methods Assay methods/platforms Statistical methods Implementation considerations Race/ethnic considerations Economic considerations Longitudinal considerations Sid O Bryant Ralph Martins Colin Masters Michael Weiner William Potter John Trojanowski Monique Breteler Kristine Yaffe Holly Soares Dwight German Pankaj Mehta Contributors Nicole Schupf Pankaj Oberoi Robert Umeck Chrostoph Laske Thomas Kodadek Moola Reddy Neill Graff-Radford George Perry Donald Royal Guanghua Xiao Fan Zhang

12