Memory loss in normal aging is related to hippocampalmediated amyloid deposition

1
Memory loss in normal aging is related to
hippocampal-mediated ?-amyloid deposition

• Beth Mormino
• Jagust Lab

2
Background

• Alzheimers disease (AD) is found in 10 over
  60, 40 over 80
• Most common cognitive deficit is episodic memory
  loss
• Neuron loss and dysfunction are typically found
  in the medial temporal lobe and posterior
  association cortices
• ?-amyloid (A?) pathology may be the primary
  etiological event in AD (Amyloid hypothesis of
  AD)
• APP ? A? protein ? A? oligomers ? A? plaques
• A? plaques found diffusely in association
  cortices

3
Background Measuring A? plaques
A? plaques
PET Imaging - 11C6-OH-BTA-1 (PIB)
Histology - Thioflavin T
4
Background
A? plaques are common in normal older people
(10-40)
AD
NC
PIB
NC
Mintun et al., 2006
Aizenstein et al., 2008
A? accumulation may reflect early AD development
(preclinical AD)
5
Background

• Hippocampus atrophy and episodic memory loss are
  common in aging, and are associated with AD
  conversion (preclinical AD markers)

HV
EM
Age
Age
Hypothesis If A? burden reflects early AD
development, elevated PIB PET will coincide with
smaller hippocampus volume and reduced episodic
memory within older people
6
Study Overview

• Berkeley Aging Cohort (BAC) normal controls
  completed
• PIB-PET Scanning
• 90 min dynamic sequence
• Structural MRI Scanning
• 3 MPRAGE sequences averaged
• Cognitive testing
• Episodic memory (CVLT LDFR, WMS VRII)
• UCSF AD subjects meeting NINDS criteria for
  probable AD used to compare PIB distribution

7
Methods

• Automated region of interest (ROIs) labeled
• on structural image using Freesurfer Software
• Hippocampus volume averaged across
• hemispheres and adjusted by total intracranial
  
• volume
• Mean PIB value from cortical ROIs (prefrontal
  cortex, parietal, lateral temporal, cingulate)
  averaged to yield PIB Index for each subject
• Regression analyses conducted (controlling for
  age, education, and gender)
• PIB Index v. HV
• PIB Index v. EM

8
Demographics
PIB Index Distribution
plt0.001
PIB
NC
AD
9
BAC NC PIB Index v. Hippocampus volume/Episodic
memory
n18
n20
HV
HV
p 0.0015 R2 0.46
p 0.027 R2 0.28
PIB Index
PIB
PIB
EM
EM
p 0.0085 R2 0.36
p 0.47 R2 0.03
PIB
PIB
10
Replication Group
Alzheimers Disease Neuroimaging Initiative (ADNI)

• www.adni-info.org
• Multi-site longitudinal study investigating AD
  progression
• (200 NC, 400 MCI, 200 AD)
• A subset have received PIB-PET/MRI imaging
• Replicated analysis
• NC (n17)
• PIB MCI (n39)

11
Replication Group ADNI
PIB Index Distribution
plt0.05, plt0.01
PIB
MCI
AD
NC
12
ADNI NC
p0.045 R2 0.22
HV
EM
p0.52 R2 0.03
PIB
PIB
ADNI PIB MCI
plt0.001 R20.27
p0.046 R20.11
EM
HV
PIB
PIB
13
Regression summary
14
Mediation Analysis
1c
1a
1b
Amyloid
HV atrophy
EM decline
2
(Baron Kenny, 1986)
15
Mediation Analysis Results
1c
1a
1b
EM
PIB
HV
2
PIB v. EM
HV v. EM
EM
PIB
HV
16
Summary and Conclusions

• Consistent relationship between A? and HV.
• PIB v. HV relationship suggests that A? is
  associated with hippocampus damage
• A direct effect of A??
• Correlation with another pathological process
  (neurofibrillary tangle formation)?
• Outlier and mediation analysis suggests that the
  relationship between A? and EM is indirect and
  may be mediated by HV atrophy.
• Longitudinal studies are needed to determine
  whether elevated A? is a marker of preclinical
  AD, and whether A? precedes hippocampus atrophy
  and EM decline in aging.

17
Thank you!

• Jagust Lab
• Bill Jagust
• Jenny Kluth
• Cindee Madison
• Suzanne Baker
• Ansgar Furst
• Amynta Hayenga
• Irene Yen

UCSF Bruce Miller Mike Weiner Gil Rabinovici
ADNI Chet Mathis Bob Koeppe