Poster Presentations P3S526

P3-254 A NEW APPROACH TO TRACKING PROGRESSION

OF MOLECULAR PATHOGENESIS IN

ALZHEIMER’S DISEASE IDENTIFIES

HIPPOCAMPAL IRON DYSHOMEOSTASIS AS

A NOVEL POTENTIAL BIOMARKER

Elizabeth A. Milward, Martin Gomez Ravetti, Osvaldo A. Rosso,

Regina Berretta, Daniel Johnstone, Pablo Moscato, The University ofNewcastle, Callaghan, Australia. Contact e-mail: liz.milward@newcastle.

edu.au

Background: Several lines of evidence suggest brain iron accumulation oc-

curs in Alzheimer’s disease. Whether this is accompanied by alterations in ex-

pression of iron-related genes is unknown. Methods: Combinatorial

optimization methods were used to analyze a published microarray gene ex-

pression dataset of hippocampal tissue from 22 AD patients and 9 controls.

MaxCover (a,b)-k Feature Set analysis determined gene signatures discrim-

inating controls from ‘‘severe AD’’. Samples classed as ‘‘incipient AD’’ or

‘‘moderate AD’’ using MMSE and NFT scores were assessed based on Jen-

sen-Shannon divergences for divergence from the control signature and con-

vergence to the severe AD signature, providing an alternative estimate of

progression of molecular pathogenesis. Results: The 1,372-probe gene ex-

pression signature generated by the Feature Set approach included genes re-

lating to iron metabolism. Transcripts for the iron transport proteins

lactoferrin and SLC11A1 were higher in many AD cases than controls. Ex-

pression of genes encoding the subunits of the iron storage protein ferritin

(FTL, FTH1) correlated positively with the degree of divergence of AD brains

from controls and with molecular changes consistent with neurodegeneration.

Ferric iron taken up into cells must be reduced to ferrous iron before storage in

ferritin. The duodenal ferrireductase CYBRD1, which reduces ferric iron in

some tissues, likewise correlated positively, although ferriductase activity

of CYBRD1 remains to be confirmed in brain. Expression of genes encoding

mitochondrial proteins utilizing iron (e.g. ICSA1, ACO2) inversely correlated

with disease progression as gauged by convergence to the severe AD signa-

ture. There was also altered expression of several genes, notably TPP1 and

CLN8, which are causatively linked to the neuronal ceroid lipofuscinoses.

These severe and sometimes fatal neurodegenerative genetic diseases are

characterized by excessive lipofuscin accumulation in lysosomes, a process

thought to be accelerated by iron derived from ferritin degradation. Conclu-

sions: The general increases in ferritin gene expression from incipient through

moderate into severe AD suggest disease progression may be accompanied by

increased hippocampal iron in ferritin and possibly also in lipofuscin. Since

hippocampal iron can now be assessed non-invasively by MRI R2 relaxivity

technology, iron dyshomeostasis may provide a new marker for progression

of molecular pathogenesis in the AD brain.

P3-255 ALZHEIMER’S CSF MARKERS IN OLDER

SCHIZOPHRENIA PATIENTS

Annapaola Prestia1, Giovanni B. Frisoni1, Cristina Geroldi1,

Andrea Adorni1, Roberta Ghidoni1, Giovanni Amicucci2, Matteo Bonetti3,

Andrea Soricelli4,5, Paul E. Rasser6,7, Paul M. Thompson8,

Panteleimon Giannakopoulos9,10, 1IRCCS - Centro S. Giovanni di Dio

Fatebenefratelli, Brescia, Italy; 2Sant’ Orsola Fatebenefratelli, Brescia,

Italy; 3Istituto Clinico Citta di Brescia, Brescia, Italy; 4SDN Foundation,

Naples, Italy; 5Institute of Diagnostic and Nuclear Development, Naples,Italy; 6Schizophrenia Research Institute, Sydney, Australia; 7University of

Newcastle, Newcastle, Australia; 8UCLA School of Medicine, Los Angeles,

CA, USA; 9University Hospitals of Geneva, Geneva, Switzerland; 10Uni-versity of Lausanne, Lausanne, Switzerland. Contact e-mail: aprestia@

fatebenefratelli.it

Background: Cognitive impairment is prevalent in older schizophrenia pa-

tients but its biological basis is unknown. Neuropathological studies have

not revealed Alzheimer disease (AD) lesion burden but in vivo data are lack-

ing. We investigated the concentrations of CSF biomarkers of brain amyloid-

osis (Abeta42) and neurodegeneration (total and p-tau) in a group of older

schizophrenia patients and related them to cognitive and structural magnetic

resonance imaging (MRI) measures. Methods: Older schizophrenia patients

(n¼ 11), AD patients (n¼ 20) and healthy elderly controls (n¼ 6) underwent

cognitive testing, lumbar puncture, and MRI scanning. Abeta42 and total and

p-tau concentrations were assayed in the CSF. MRI volumes were assessed

using both voxel-based (cortical pattern matching) and region-of-interest

analyses. Results: CSF tau concentration in older schizophrenia patients

was within normal limits (total tau 171 6 51 pg/ml, p-tau 32 6 8 pg/ml), while

CSF Abeta42 (465 6 112 pg/ml) levels were significantly lower compared to

healthy older persons (638 6 130 pg/ml) but higher than in AD patients (352

6 76 pg/ml). There was a strong positive relationship between CSF total or p-

tau levels and MMSE scores in schizophrenia patients but not in AD, where

higher concentrations of total tau were correlated with higher volumes in the

occipital cortex (r ¼ .63, p ¼ .036). In AD, lower Abeta42 concentrations

were correlated with lower gray matter volume in the cingulate and lateral or-

bital cortices (r>.46, p < .05). Conclusions: Older schizophrenia patients

show a peculiar pattern of CSF Abeta42 and tau concentrations that relates

to cognitive and structural markers but is not consistent with neurodegenera-

tion and could be secondary to neurodevelopmental or drug treatment effects.

P3-256 TAU SECRETION TO THE EXTRACELLULAR

SPACE IS MODULATED BY THE PRESENCE

OF THE EXON 2 ALTERNATIVELY

SPLICED INSERT IN CELL CULTURE AND

IN SITU TAUOPATHY MODELS

Garth F. Hall, WonHee Kim, Sangmook Lee, U. Mass. Lowell, Lowell, MA,USA. Contact e-mail: Garth_Hall@uml.edu

Background: While the secretion of some neurodegeneration-associated

proteins (prion protein, beta amyloid and alpha synuclein) to the extracel-

lular space has been established, it is only very recently that tau secretion

(Kim et al., 2010 J. Alz. Dis.) uptake (Frost et al., 2009 J. Biol Chem.) and

interneuronal transfer (Clavaguera et. al. 2009 Nat. Cell Biol, Kim, loc cit)

have been demonstrated. Secretion from viable neurons of N terminal tau

fragments resembling those in AD CSF (Kim, loc cit) raises the possibility

that early stage AD patients with elevated CSF-tau may remain candidates

for an AD ‘‘cure’’. Here we show that the presence of the exon 2 (E2)

amino terminal insert in full length tau and amino terminal tau fragments

is disproportionately retained in cell culture and in situ tauopathy models.

Methods: We used Western Blotting of immunoprecipitated and concen-

trated cellular and media fractions from NB2A and M1C neuroblastoma

lines and confocal microscopic analysis of transverse sections through

the somata and dendrites of tau-expressing lamprey neurons in this study.

Results: We found that transient expression of the amino terminal (1-255)

resulted in much greater secretion (w10-30x) when exons 2/3 were absent

than when they were included. Secretion of full length E2- and E2+ tau iso-

forms of nearly identical sizes (3RN1 and 4RN0) from an inducibly ex-

pressing cell line (M1C) under selective induction conditions confirmed

that full length 3RN1 tau was secreted significantly less efficiently

(w10-20x) than 4RN0 tau. Retrospective analysis of existing samples

from lamprey ABCs expressing E2/3+ and E2/3- tau constructs of all sizes

and types also showed a highly significant negative correlation between

the presence of the E2/3 sequence and the frequency of secretion.

Conclusions: We propose that the absence of E2 (and possibly E3) selec-

tively identifies secreted tau, and suggest that the presence of E2/3- tau in

the CSF of human patients might be a more sensitive prospective predictor

of AD and other tauopathies than currently available CSF-based diagnos-

tics, thus improving the sensitivity of current clinical trials and widening

the effective therapeutic window for these conditions once drugs capable

of arresting AD/tauopathy pathogenesis become available.

P3-257 MEASUREMENT OF AMYLOID-b AND NEURONAL

SPECIFIC ENOLASE AS A PREDICTOR OF

OUTCOME IN PATIENTS SUFFERING FROM

TRAUMATIC BRAIN INJURY

Joshua Gatson1, James Simpkins2, Jane Wigginton1, 1UT Southwestern,

Dallas, TX, USA; 2University of North Texas Health Science Center, FortWorth, TX, USA. Contact e-mail: joshua.gatson@utsw.edu

Poster Presentations P3 S527

Background: Traumatic brain injury (TBI) is a multi-phasic disease and

following the initial impact, secondary injury is thought to contribute to

death and disability. Increased oxidative stress and inflammation in the

brain leads to the demise of neuronal populations and ultimately decreased

brain function. Identification of neural markers such as beta-Amyloid (Ab)

and neuronal specific enolase (NSE) at early time-points after injury can

help us better understand the multitude of such secondary injuries follow-

ing TBI and may help predict outcome in these patients. Methods: Serial

CSF and blood samples were collected from patients with severe TBI (GCS

3-8) who required placement of a therapeutic ventriculostomy. Following

ventriculostomy placement, CSF and blood samples were obtained every

4 hours for the 1st 24 hours post-injury, then every 8 hours for post-injury

days 2-5. Following collection of these samples, the levels of Ab and NSE

were measured in these samples using the ELISA method. Results: Fol-

lowing sample analysis, we found that in TBI patients that had a good neu-

rological outcome (as indicated by the Glasgow Coma Score), there was

a great increase in the levels of Ab (w390 pg/ml) compared to patients

with a bad outcome (w10 pg/ml). With respect to NSE, patients with

a good outcome, exhibited high levels of NSE (between 25 and 60 ng/

ml) in the CSF at 12 hours post-injury which subsided to w5 ng/ml (nor-

mal range) at around 20 hours and sustained until day 5. In contrast, pa-

tients with a bad outcome or death, had high levels of NSE (w225 ng/

ml) in the CSF, at the later time-points, which correlated with a decrease

in neurological status. Conclusions: In the TBI patients, an increase in

Ab levels correlated with improved outcome, whereas a decrease in the

levels of NSE correlated with a good outcome. Detection of such bio-

markers may be a useful tool for determining injury severity and outcome

in TBI patients. As a future direction, more patients will be studied in order

to validate these findings.

P3-258 BLOOD BIOMARKERS OF ALZHEIMER’S DISEASE

AND MILD COGNITIVE IMPAIRMENT

Milan Fiala, UCLA, Los Angeles, CA, USA. Contact e-mail: fiala@ucla.edu

Background: The best biomarker test for AD should make possible early

detection of the process leading to neurodegeneration, not the outcome of

neurodegeneration. We have shown that the innate immune system cells

(macrophages and monocytes) of AD patients are defective in clearance

of amyloid-beta (Abeta), and these immune defects may underlie brain ac-

cumulation of Abeta and neurodegeneration. We have developed a novel

test of Abeta phagocytosis by peripheral blood mononuclear cells

(PBMC’s), which measures the phagocytosis of fluorescent Abeta and

shows the difference between normal and abnormal clearance. The macro-

phages and monocytes of AD patients are further separated into Type I,

which have the most profound defect of Abeta phagocytosis and transcrip-

tion of the gene MGAT-III but are improved by bisdemethoxycurcumin,

and Type II, which have lesser defects but are not improved. Methods:

The flow cytometric test measures phagocytosis of fluorescent Abeta by

PBMC’s in vitro. The PCR test measures MGAT-III transcription in

PBMC’s. Results: In a pilot study (70 subjects), we tested AD patients,

caregivers, age-matched active professors, Mild cognitive impairment

(MCI) and Amyotrophic lateral sclerosis (ALS) patients. The score of

<400 mean fluorescence intensity units (MFI U) was determined as

a break-point between positive and negative AD biomarker. Using this

break-point, the test was positive in 94% patients diagnosed as probable

AD by neurological and neuroimaging criteria (mean score 6SEM 198.6

6 25.5 MFI U), in 60% MCI patients (mean score 301 6 106 MFI U),

and in 75% caregivers (511 6 178 U). Control subjects, active University

professors, were 100% negative (1348 6 174 MFI U). Conclusions: Im-

mune-based tests may evaluate the process, not just the state of neurode-

generation. A normal result of the flow cytometric test might suggest

low risk of AD, whereas an abnormal result might suggest either estab-

lished AD or the risk of AD. The results in caregivers are unclear but ge-

netic relationship and stress levels may explain abnormal results. Ongoing

studies in larger samples will evaluate the sensitivity, specificity and pre-

dictive value of the test and its use for detection of the risk of progression

to AD in MCI patients.

P3-259 DEVELOPMENT AND IMPLEMENTATION OF

A TMT-SRM ASSAY FOR THE QUALIFICATION OF

CANDIDATE BIOMARKERS OF ALZHEIMER’S

DISEASE

Darragh P. O’Brien1,2, Andreas Guentert1, James Campbell2,

Karsten Kuhn3, Malcolm A. Ward2, Helen L. Byers2, Simon Lovestone1,1Institute of Psychiatry, King’s College London, London, United Kingdom;2Proteome Sciences Plc, London, United Kingdom; 3Proteome Sciences

R&D GmbH&CoKG, Frankfurt am Main, Germany.

Contact e-mail: darragh.o’brien@kcl.ac.uk

Background: Biomarker discovery studies have resulted in large numbers of

candidate proteins which are implicated in Alzheimer’s disease (AD). In order

to quickly select the most relevant candidates, selective assays are required

that will ideally measure multiple proteins in a single analysis. TMT-SRM us-

ing triple quadrupole mass spectrometry allows for the multiplexed quantita-

tion of many proteins in complex protein digests such as plasma, with

exceptional performance criteria. Here, we apply such an assay for the quan-

titation of eight candidate biomarkers of AD across a clinical cohort.

Methods: Eight candidate biomarkers of AD were selected from a long-list

of >30 proteins arising from initial proteomic analysis of plasma. Data

from these discovery studies were mined to select the best target peptides

for the quantitation of each candidate protein. Peptides which were proteo-

typic and had suitable characteristics for targeted SRM quantitation were in-

cluded (at least three per protein). Experimental samples were digested and

labeled with light TMT. Known amounts of synthetic versions of the target

peptides were labeled with heavy TMT and spiked into the experimental sam-

ples, providing internal standardisation and absolute quantitation. In an initial

study, 10 AD and 10 non-demented controls were analysed. A total of 3 an-

alytical and 3 technical repeats were performed for each sample. Results: The

multiplexed assay was shown to be reproducible (CV’s<15%) and highly se-

lective for each analyte. Quantitation was linear over two orders of magnitude

and within the endogenous concentration range of each protein. For one of the

candidate biomarkers, gelsolin, the results verified discovery exercises and

were in good agreement with western blot measurements on the same sample

set. The TMT-SRM assay verified the seven other AD candidate biomarkers,

which were also in-line with discovery and ELISA results. Conclusions:

Here we demonstrate a highly selective and precise assay for the quantitation

of eight candidate biomarkers of AD. We have now moved to the analysis of

a more extensive clinical cohort. Two time points have been selected for each

patient sample to monitor potential markers that track with disease progres-

sion (in terms of MCI converters/non-converters and AD fast/slow decliners).

Apolipoprotein E and clusterin genotype will also be considered.

P3-260 BRAIN VENTRICULAR SUB-REGION

SEGMENTATION: A LONGITUDINAL STUDY

USING DATA FROM THE ALZHEIMER’S DISEASE

NEUROIMAGING INITIATIVE

Amanda F. Khan1, Matthew Smith2, Michael Borrrie2,3, Jennie Wells2,

Robert Bartha1, 1Robarts Research Institute, The University of Western

Ontario, London, ON, Canada; 2Lawson Research Institute, London, ON,

Canada; 3Schulich School of Medicine and Dentistry, London, ON, Canada.Contact e-mail: akhan@imaging.robarts.ca

Background: Brain ventricular enlargement has been shown to be an objec-

tive and sensitive measure of ventricular enlargement associated with mild

cognitive impairment (MCI) and Alzheimer’s disease (AD). This study com-

pared ventricular volume in normal elderly controls (NEC), MCI and AD pa-

tients by sub-region over a two-year time period. The 10 defined sub-regions

(Figure 1) were: right-lateral anterior (RLA), right-lateral middle (RLM),

right-lateral posterior (RLP), left-lateral anterior (LLA), left-lateral middle

(LLM) and left-lateral posterior (LLP). The temporal horns were divided

into left-anterior horn (LAH), right-anterior horn (RAH), left-posterior

horn (LPH) and the right-posterior horn (RPH) regions. Methods: Baseline,

12-month and 24-month three-dimensional T1-weighted magnetic resonance

images (MRI) were obtained from 48 subjects (NEC n¼ 13, MCI n¼ 19 and

AD n ¼ 16) participating in the Alzheimer’s Disease Neuroimaging