Download as a PDF

Effect of Phenserine Treatment on Brain
Functional Activity and Amyloid in
Alzheimer’s Disease
Ahmadul Kadir, MD,1 Niels Andreasen, MD, PhD,1,2 Ove Almkvist, PhD,1–3 Anders Wall, PhD,4
Anton Forsberg, MSc,1 Henry Engler, MD,5 Göran Hagman, MSc,2 Marie Lärksäter, BSc,2
Bengt Winblad, MD, PhD,1 Henrik Zetterberg, MD, PhD,6 Kaj Blennow, MD, PhD,6
Bengt Långström, PhD,4,7 and Agneta Nordberg, MD, PhD1,2
Objective: The effects of (⫺)-phenserine (phenserine) and placebo/donepezil treatment on regional cerebral metabolic rate for
glucose (rCMRglc) and brain amyloid load were investigated by positron emission tomography in 20 patients with mild Alzheimer’s disease in relation to cerebrospinal fluid (CSF) and plasma biomarkers, and cognitive function.
Methods: The first 3 months of the study was a randomized, double-blind, placebo-controlled phase, during which 10 patients
received phenserine (30mg/day) and 10 patients the placebo. Three to 6 months was an open-label extension phase, during
which the placebo group received donepezil (5mg/day) and the phenserine group remained on phenserine. After 6 months, all
patients received phenserine treatment up to 12 months. The patients underwent positron emission tomography examinations to
measure rCMRglc (18F-FDG) and amyloid load (11C-PIB) at baseline and after 3 and 6 months of the treatment. Neuropsychological and biomarker data were collected at the three times of positron emission tomography imaging.
Results: Statistically significant effects on a composite neuropsychological test score were observed in the phenserine-treated
group compared with the placebo and donepezil group at 3 and 6 months, respectively. Values of rCMRglc were significantly
increased in several cortical regions after 3 months of phenserine treatment, compared with baseline, and correlated positively
with cognitive function and CSF ␤-amyloid 40 (A␤40). Cortical Pittsburgh Compound B retention correlated negatively with
CSF A␤40 levels and the ratio A␤/␤-secretase–cleaved amyloid precursor protein. In CSF, A␤40 correlated positively with the
attention domain of cognition.
Interpretation: Phenserine treatment was associated with an improvement in cognition and an increase in rCMRglc.
Ann Neurol 2008;63:621– 631
Currently, cholinesterase inhibitors (ChEIs) are clinically used for symptomatic treatment in Alzheimer’s
disease (AD). Donepezil, rivastigmine, and galantamine
have shown benefits for functional, behavioral, and
cognitive measurements.1 The drug studied in this
clinical trial, (⫺)-phenserine (phenserine), has a dual
mode of action: an acetylcholinesterase inhibitor
(AChEI), and an inhibitor of the formation of
␤-amyloid precursor protein (␤-APP), the source of
neurotoxic ␤-amyloid (A␤) plaque.2 Phenserine interferes with the 5⬘ untranslated region of human APP
messenger RNA, decreasing the amount of translation
into the actual protein, hence reducing newly synthesized APP. In an experimental study, both (⫺)- and
cholinergically inert (⫹)-phenserine have been shown
to decrease A␤ levels in cell cultures and mouse brain.3
Other substances, such as monoclonal antibodies and
vaccines directed against A␤, have led to a reduction of
plaques in transgenic mice4,5 and nonhuman primates,6 and therefore have been used in the study of
patients with AD.7 ChEIs have also been shown to
have an effect on APP processing in cell culture, reducing levels of A␤.8 –10
Although several experimental studies have shown
From the 1Department of Neurobiology, Care Sciences and Society,
Karolinska Institutet, Karolinska University Hospital Huddinge;
2
Department of Geriatric Medicine, Karolinska University Hospital
Huddinge; 3Department of Psychology, Stockholm University,
Stockholm; 4Uppsala Imanet AB, Imanet, GE Healthcare; 5Department of Nuclear Medicine, Uppsala University Hospital, Uppsala;
6
Institute of Neuroscience and Physiology, Section of Psychiatry
and Neurochemistry, Göteborg University, Göteborg; and 7Department of Biochemistry and Organic Chemistry, Uppsala University,
Uppsala, Sweden.
This article includes supplementary materials available via the Internet at http://www.interscience.wiley.com/jpages/0364-5134/suppmat
Received Sep 12, 2007, and in revised form Dec 28. Accepted for
publication Jan 2, 2008.
Published online Feb 25, 2008, in Wiley InterScience
(www.interscience.wiley.com). DOI: 10.1002/ana.21345
Address correspondence to Dr Nordberg, Karolinska Institutet,
Department of Neurobiology, Care Sciences and Society, Division of Alzheimer Neurobiology, Karolinska University Hospital Huddinge, Novum Floor-5, S-14186 Stockholm, Sweden.
E-mail: [email protected]
© 2008 American Neurological Association
Published by Wiley-Liss, Inc., through Wiley Subscription Services
621
the efficacy of (⫺)- and (⫹)-phenserine treatment on
A␤ metabolism and neurogenesis,3,11,12 there have
been no reports regarding effect of (⫺)-phenserine on
cortical amyloid plaque in AD patients. Pittsburgh
Compound B (11C-PIB) has been used as a promising
positron emission tomography (PET) ligand to measure in vivo amyloid plaque.13,14 The primary aim of
this explorative study was to use PET to measure the
effects of (⫺)-phenserine treatment (30mg daily) and
placebo/donepezil (5mg daily) on regional cerebral metabolic rate for glucose (rCMRglc), using
2-(18F)fluoro-2deoxy-D-glucose (18F-FDG), and on
amyloid load, using 11C-PIB, in the brains of mild AD
patients. The secondary objectives were to measure cerebrospinal fluid (CSF) and plasma biomarkers, to
evaluate cognitive function, and to find possible interrelations between PET parameters, CSF and plasma biomarkers, and cognitive function.
Subjects and Methods
Study Design
This was a 12-month study involving 20 patients with mild
AD (Supplementary Fig 1 illustrates the study design). The
first 3-month period was a double-blind, placebo-controlled,
randomized period. The patients were grouped according to
the treatment they received: placebo group (n ⫽ 10) and
phenserine group (n ⫽ 10). The second 3-month period
(3– 6 months) was an open-label extension phase, during
which patients in the placebo group received donepezil and
patients in the phenserine group remained on phenserine.
After 6 months, all patients received phenserine treatment
from 6 to 12 months and were clinically followed up with
neuropsychological tests.
During the dose-escalation periods, in the double-blind
phase, the phenserine-treated patients received 5mg twice
daily from weeks 1 to 4, 10mg twice daily from weeks 5 to
8, and 15mg twice daily from weeks 9 to 12. In the openlabel phase from week 13, the phenserine-treated patients remained on 15mg twice daily. The placebo-treated patients
switched to donepezil at 5mg/day for 3 months; then they
had a 1-week washout before being titrated up to a 15mg
twice-daily dose of phenserine.
Patients
Twenty patients with a diagnosis of mild AD (Mini-Mental
State Examination [MMSE] score ⱖ 21) were recruited from
the Department of Geriatric Medicine, Karolinska University
Hospital Huddinge, Stockholm, Sweden. All patients were
referred for assessment of dementia because of a memory
problem and underwent a thorough clinical investigation including medical history, cognitive screening (MMSE), physical and neurological examination, laboratory blood tests,
apolipoprotein E genotyping, neuropsychological assessment,
lumbar puncture, and magnetic resonance imaging/computed tomography scans. The diagnosis of AD was made by
exclusion of other dementia diseases, in accordance with criteria from the National Institute of Neurological and Communication Disorders and Stroke-Alzheimer’s disease and
Related Disorders Association (NINCDS-ADRDA).15
In the study, we used PIB binding data from six agematched (mean ⫾ standard deviation, 67.3 ⫾ 8.8; range,
57–77 years) healthy control subjects to compare with that
of patients with mild AD at baseline.
Fig 1. Regional cerebral glucose metabolism (rCMRglc, normalized to the pons) in the different cortical brain regions. Data are
expressed as mean ⫾ standard error. Posttreatment rCMRglc values were expressed as percentage changes from the baseline (pretreatment) value by using the following formula: %rCMRglc ⫺ 100 ⫽ (rCMRglc (f)/rCMRglc(b) ⫻ 100) ⫺ 100, with f and b indicating the rCMRglc values at follow-up and baseline, respectively. Positive rCMRglc values indicate increased rCMRglc. Solid bars
indicate the phenserine group at 3 and 6 months of treatment, respectively. Hatched bars indicate the placebo (A) and donepezil
(B) groups. *p ⬍ 0.05 indicates significantly increased in rCMRglc within the group compared with baseline; #p ⬍ 0.05 indicates
a significant difference between the groups at 3 months. ctx ⫽ cortex; lt ⫽ left; rt ⫽ right.
622
Annals of Neurology
Vol 63
No 5
May 2008
All patients and their responsible caregivers provided written informed consent to participate in the study; it was conducted according to the Declaration of Helsinki and subsequent revisions, and was approved by the Ethics Committee
of Karolinska University Hospital Huddinge and the Faculty
of Medicine and Radiation Hazard Ethics Committee of
Uppsala University Hospital, Uppsala, Sweden.
Assessment of Safety
Safety was assessed throughout the study. The safety variables were adverse events (AEs), vital signs, physical examination, laboratory safety screen, and electrocardiogram.
Positron Emission Tomography Methods
PET studies were performed at the PET Center, Uppsala
Imanet AB, Imanet GE Healthcare, Uppsala, Sweden, at
baseline, and after 3 and 6 months of the study.
Radiotracers
Production of FDG and PIB was conducted according to
standard good manufacturing processes at Uppsala Imanet.
Synthesis of PIB was performed by means of a method described previously.13,16
Positron Emission Tomography Scanning
The PET scans were performed using Siemens ECAT EXACT HR⫹ scanners (CTI PET-Systems), with an axial field
of view of 155mm, providing 63 contiguous 2.46mm slices
with 5.6mm transaxial and 5.4mm axial resolution. The patients were scanned after fasting for 4 hours under resting
conditions in a dimmed room. The orbitomeatal line was
used to center the heads of the patients. The data were acquired in three-dimensional mode. The scanner protocol for
transmissions and reconstructions has been described previously.13 The emission scans consisted of 7 or 24 time frames
with a progressive increase in frame duration (1 ⫻ 60, 1 ⫻
140, and 5 ⫻ 300 seconds) and a total duration of 45 minutes for FDG; and 4 ⫻ 30, 9 ⫻ 60, 3 ⫻ 180, and 8 ⫻ 300
seconds with a total duration of 60 minutes for PIB. The
average tracer doses administered were 328 ⫾ 34.6 and
210 ⫾ 64.8Mbq for PIB and FDG, respectively (mean ⫾
standard deviation).
Regions of Interest
A set of standardized regions of interest (ROIs) were defined
using the Scanditronix program (IDA, Images Display and
Analyses GE 1994). All ROIs were paired for the right and
left hemispheres, except for the pons and whole brain. The
ROIs placement procedure has been described in detail previously.13,17
In this article, we focused the data analysis on those ROIs
that are considered to be most critical in AD, and for presentation of PET data, these regions were grouped into four
major areas: (1) frontal cortex (including the ROIs of frontal,
anterior cingulate, and frontal association cortices), (2) parietal cortex, (3) parietotemporal cortex, and (iv) temporal cortex (including the ROIs of posterior/anterior lateral, inferior,
and medial temporal cortices). The mean cortical was calculated as a composite of the above four regions.
Data Management
For FDG, venous arterialized blood samples were obtained
and parametric maps were generated by means of the Patlak
method, using the time course of the tracer in the arterialized
venous plasma as an input function.18 The rCMRGlc values
were normalized to the pons value (ROI/ref) to allow interindividual and intraindividual comparisons.19 PIB retention
was measured by the Patlak reference method,20 assuming
irreversible binding of the tracer to the brain tissue. The
binding of PIB to the target tissue can be considered as irreversible during the measurement period of 60 minutes.
Specific uptake of PIB in the cerebellum is, however, not
negligible. The net accumulation rate in the reference region
was accounted for by means of a correction based on previous measurements of the accumulation rate constant of PIB
in this region.21 The cerebellar cortex was chosen as reference
because of its previously reported lack of Congo red– and
thioflavin-S–positive plaques.22,23 Among several different
PIB evaluation methods, the late scan reference method has
shown a large effect size.24
In this study, percentage changes in the CMRglc and PIB
retention values were used, defined by the following formula:
%PET data ⫺ 100 ⫽ (PET data(f)/PET data(b) ⫻ 100) ⫺
100, with f and b indicating CMRglc and PIB retention values at the times of follow-up and at baseline, respectively.
Cerebrospinal Fluid and Plasma Biomarkers
CSF and plasma samples were collected by lumbar puncture
and venipuncture, respectively, at baseline and after 3 and 6
months of treatment. All samplings were performed in the
morning or 2 to 4 hours after intake of morning medication.
A volume of 12ml CSF was collected in polypropylene
tubes and gently mixed to avoid gradient effects. CSF samples with more than 500 erythrocytes/␮l were excluded. The
blood samples were collected in tubes containing EDTA as
anticoagulant. All samples were centrifuged, aliquoted, and
stored in polypropylene tubes at ⫺80°C pending biochemical analyses, without being thawed and refrozen.
Levels of ␤-amyloid 1-42 (A␤42), total-tau (T-tau), and
phosphorylated tau (P-tau181) in CSF were analyzed using
the INNO-BIA AlzBio3 assay (Innogenetics, Ghent, Belgium), as described previously in detail.25
Levels of CSF ␣- and ␤-secretase–cleaved amyloid precursor protein (␣-sAPP and ␤-sAPP, respectively) were measured using Meso Scale Discovery electrochemiluminescence
detection technology (sAPP␣/sAPP␤ multiplex assay; Meso
Scale Discovery, Gaithersburg, MD) and a SECTOR Imager
2400 instrument, following the instructions of the manufacturer. The ␤-sAPP assay is based on a ␤-sAPP–specific capture antibody, recombinant ␤-sAPP as standard, and an
N-terminal APP-specific SULFO-TAG–labeled antibody as
detector. The ␣-sAPP assay is based on 6E10 monoclonal
antibody (MAb) as capture antibody, recombinant ␣-sAPP as
standard, and an N-terminal, APP-specific, SULFO-TAG–
labeled antibody as detector.
Levels of CSF A␤40 were measured by using Meso Scale
Discovery human A␤40 ultrasensitive assays (Meso Scale
Discovery) and a SECTOR Imager 2400 instrument, following the instructions of the manufacturer. This assay is based
on 6E10 MAb as capture antibody and an A␤40-specific
SULFO-TAG–labeled antibody as detector.
Kadir et al: PET Study of Phenserine in AD
623
Plasma A␤40 levels were determined by using highsensitivity enzyme-linked immunosorbent assay kits
(TK40HS; The Genetics Company, Schlieren, Switzerland).
In this enzyme-linked immunosorbent assay, MAb W02 (A␤
epitopes 5– 8) is the capture antibody and MAb G2-10 is the
detection antibody. The latter has negligible cross-reactivity
with A␤ species other than A␤40.26,27 Plasma A␤42 concentrations were determined by using the high-sensitivity INNOTEST ␤-AMYLOID(1-42) assay (Innogenetics, Ghent,
Belgium). The assay and its characteristics have been described previously in detail.28
Cognitive Function
Global cognition was estimated by means of MMSE29 at
baseline and at 3, 6, 9, and 12 months of the study. Neuropsychological tests of episodic memory (Word Recall [2and 5-second presentation rate], Word Recognition-d [2and 5-second presentation rate]), attention (Digit Symbol,
Trail Making Test-A [time]), and visuospatial ability (Clock
Recognition) were performed at baseline and during treatment at months 2, 3, 6, 9, and 12. Further details of the
neuropsychological procedure have been described previously.30
To make comparisons between various neuropsychological
test results, we Z-transformed all cognitive raw scores (including MMSE score) by using reference data from the Geriatric Clinic, Karolinska University Hospital Huddinge,
Stockholm, Sweden.31
Statistical Analysis
Data are expressed as mean values and standard error of the
mean. The equivalence of groups in demographic variables
was checked by means of one-way analysis of variance.
Between-group comparisons at specific time points were conducted by means of analysis of covariance. In this model,
percentage changes from baseline (for the CMRglc and PIB
retention) or absolute value changes from baseline (for cognitive function, as well as CSF and plasma biomarkers) were
used as dependent variables, with baseline value as a covariate. For within-group comparisons, Wilcoxon signed rank
test was used to compare baseline values versus those at other
time points. This nonparametric method was chosen for
within-group analysis because of the small sample size (n ⫽
10 per group) and because we could not assume that all the
parameters were normally distributed. Two-tailed Pearson’s
correlation coefficients and Spearman’s rank correlations
where appropriate were used in correlation analysis, which
was then visualized graphically using simple regression plots.
All statistical tests were two-tailed at the 0.05 level of significance, and no corrections for multiple comparisons were
made because this was an explorative type of study.
Results
Demographic Data
The demographic characteristics of the patients are presented in the Table. The demographic data did not
show any significant differences between the groups (all
p ⬎ 0.10 except for MMSE score, p ⫽ 0.08, nominally favoring the group initially receiving phenserine).
624
Annals of Neurology
Vol 63
No 5
May 2008
Adverse Events
Two episodes of nausea, one episode of vomiting, and
gastritis commenced during treatment with 15mg
phenserine twice daily in three patients, but for two
patients, these symptoms gradually resolved when the
dose was reduced to 20mg/day. One episode of depression, nasopharyngitis, and headache occurred during
treatment with placebo in one patient; one episode of
pruritus commenced during treatment with 5mg/day
donepezil in one patient. None of these adverse events
was severe, and all patients completed the 12-month
study.
No serious adverse events occurred because of the
treatment, and no clinically relevant changes in laboratory values, vital signs, or electrocardiographic readings
were observed during the study.
Changes in Regional Cerebral Metabolic Rate
for Glucose
At 3 months, a generalized increase (7 ⫾ 3%; range,
4 –10%) in cortical rCMRglc was observed in the
phenserine group compared with baseline. The increase
in rCMRglc was significant in the left frontal, right
parietal, and the right/left parietotemporal cortices (all
p ⬍ 0.05) (Fig 1A), whereas no significant effects were
observed in brain regions that are less affected in AD,
that is, primary visual, sensory motor, thalamus, putamen, and cerebellum (data not shown). The left parietotemporal cortex showed a significant difference between the phenserine- and placebo-treated groups (F ⫽
6.28; degrees of freedom [df] ⫽ 1; p ⫽ 0.02). No significant effects were observed in the placebo group
compared with baseline (see Fig 1A). The rCMRglc in
the cortical regions did not differ from baseline after 6
months of treatment with phenserine or 3 months of
treatment with donepezil (see Fig 1B).
Changes in Cortical 11C-PIB Retention
Because we did not observe any difference in PIB retention between different cortical regions, we aggregated the four studied regions (frontal, parietal, parietotemporal, and temporal) and presented the data as
mean cortical PIB retention. At baseline, both groups
showed significantly high PIB retention compared with
the healthy control group (see Supplementary Fig 2A).
After 3 and 6 months of treatment, no significant
changes in PIB retention were observed in either group
compared with baseline or between the groups at these
time points.
At 3 months, all placebo-treated patients except one
showed results within an estimated test-retest value of
⫾5%.32 In the phenserine group, four patients showed
reduced PIB retention compared with baseline (see
Supplementary Fig 2B).
Supplementary Figure 3 illustrates the parametric
image of cortical PIB retention in an AD patient after
3 months of phenserine treatment and showed a 12%
reduction in cortical PIB binding compared with baseline.
Three patients after 6 months of phenserine treatment and two patients after 3 months of donepezil
treatment showed reduced PIB retention compared
with the baseline below the test-retest value (see Supplementary Fig 2C).
Changes in Global Cognition
The phenserine-treated group showed improvement in
MMSE Z-score, which reached statistical significance
at 9 months ( p ⬍ 0.02; Fig 2A) compared with baseline. After 12 months (Z-score ⫽ 0.4 ⫾ 0.6), no difference was observed in the phenserine-treated group
compared with baseline. No significant change in
MMSE score was observed in the placebo/donepezil/
phenserine-treated group compared with baseline, the
MMSE score at 12 months being comparable with that
in the phenserine group (see Fig 2A).
Fig 2. Mean (⫾ standard error [SE]) changes in Z-scores
from baseline in Mini-Mental State Examination (MMSE)
score (A) and in composite of seven neuropsychological tests (B)
over time. *p ⬍ 0.02 indicates a significant improvement in
the MMSE Z-score in the phenserine-treated group at 9
months compared with baseline (A). #p ⬍ 0.05 indicates a
significant difference in the composite neuropsychological
Z-score between the groups at the specified treatment intervals
(B). Solid squares represent phenserine group; open circles
represent 0 to 3 months ⫽ placebo group, 3 to 6 months ⫽
donepezil treatment, and 6 to 12 months ⫽ phenserine treatment.
Changes in Neuropsychological Performance
Because neuropsychological performance was included
as a secondary objective, and because of the limitation
of sample size, the outcome of the seven neuropsychological subtests was calculated as mean composite
Z-score and presented as changes from baseline (see Fig
2B).
At 3months, analysis of covariance of the composite
Z-scores yielded significant effect (F ⫽ 5.72; df ⫽ 1;
p ⫽ 0.03) between the placebo- and the phenserinetreated group. The significant group effect was found
because of improvement in the phenserine-treated
group (Z-score ⫽ 0.25 ⫾ 0.13) and decline (Z-score ⫽
⫺0.16 ⫾ 0.13) in the placebo-treated group after 3
months of treatment compared with baseline.
Furthermore, a significant group effect (F ⫽ 7.05;
df ⫽ 1; p ⫽ 0.02) was observed between 6 months of
treatment with phenserine (improvement, Z-score ⫽
0.33 ⫾ 0.16) and 3 months of treatment with donepezil (decline, Z-score ⫽ ⫺0.24 ⫾ 0.11) compared
with baseline.
At 12 months, phenserine-treated patients showed
stabilization (Z-score ⫽ ⫺0.09 ⫾ 0.4) compared with
baseline (see Fig 2B). After 6 months, when the
donepezil-treated group switched to phenserine treatment for another 6 months, they showed a nonsignificant decline (Z-score ⫽ ⫺0.33 ⫾ 0.17) compared
with baseline (see Fig 2B).
Correlation between Positron Emission Tomography
Measures (Regional Cerebral Metabolic Rate for
Glucose and Pittsburgh Compound B) and
Cognitive Function
At baseline, rCMRglc values positively correlated with
the composite Z-score of the seven neuropsychological
Kadir et al: PET Study of Phenserine in AD
625
right parietal cortex (r ⫽ 0.51; p ⬍ 0.02; see Fig 3B).
In addition, at this follow-up point, the changes in rCMRglc correlated positively with changes in plasma
A␤40 levels in the following brain regions: right frontal
tests in the following brain regions: the right parietal
(r ⫽ 0.50; p ⬍ 0.03), the left parietal (r ⫽ 0.51; p ⬍
0.03), the right parietotemporal (r ⫽ 0.44; p ⬍ 0.05),
and the left parietotemporal cortices (r ⫽ 0.48; p ⬍
0.04).
At 3 months, the changes in rCMRglc in all four
cortical regions correlated positively with the composite
Z-score. Significant correlations were observed in the
right parietotemporal (r ⫽ 0.44; p ⬍ 0.05) and left
parietotemporal cortices (r ⫽ 0.45; p ⬍ 0.05).
At 6 months, similar positive correlations were observed between the changes in rCMRglc and composite
Z-score. Furthermore, significant positive correlations
were observed in the left frontal (r ⫽ 0.45; p ⬍ 0.05;
see Supplementary Fig 4) and right temporal cortices
(r ⫽ 0.44; p ⬍ 0.05). These findings demonstrate that
AD patients with a high rCMRglc also showed high
cognitive function at the explicit time point. No significant correlations between cortical PIB binding and
rCMRglc or cognitive function were observed throughout the study.
Changes in Cerebrospinal Fluid and
Plasma Biomarkers
No statistically significant differences were observed in
any of the parameters between the placebo/donepeziland phenserine-treated groups or in the within-group
changes from baseline during the study period (see
Supplementary Table 1). However, there was a tendency toward increased values of CSF A␤40 and
plasma A␤40 after 3 months of treatment with
phenserine compared with baseline.
Correlation between Regional Cerebral Metabolic
Rate for Glucose and Cerebrospinal Fluid and
Plasma Biomarkers
At baseline, rCMRglc values in the right parietal cortex
correlated positively with CSF A␤40 levels (r ⫽ 0.46;
p ⬍ 0.04; Fig 3A). At 3 months, changes in rCMRglc
in all four cortical regions correlated positively with the
changes in CSF A␤40 concentrations, and a significant
positive correlation was observed with changes in the
Fig 3. Positive correlations between regional cerebral glucose
metabolism (rCMRglc) in the right parietal cortex and cerebrospinal fluid (CSF) ␤-amyloid 40 (A␤40) concentrations
(pg/ml) at baseline (A); changes in rCMRglc in the right parietal cortex and changes in CSF A␤40 levels at 3 months (B);
changes in rCMRglc in the right temporal cortex and changes
in CSF A␤40 levels at 6 months (C). At 3 and 6 months, all
values on both axes are changes from individual baseline values. Solid squares at 3 and 6 months represent phenserine
group; open circles at 3 months represent placebo group; open
circles at 6 months represent donepezil group. (A) r ⫽ 0.46;
p ⬍ 0.04. (B) r ⫽ 0.51; p ⬍ 0.02. (C) r ⫽ 0.44; p ⬍
0.05.
626
Annals of Neurology
Vol 63
No 5
May 2008
‹
(r ⫽ 0.65; p ⬍ 0.003), left frontal (r ⫽ 0.64; p ⬍
0.003), right parietal (r ⫽ 0.64; p ⬍ 0.003), left parietal (r ⫽ 0.64; p ⬍ 0.003), right parietotemporal (r ⫽
0.65; p ⬍ 0.002), and left temporal cortices (r ⫽ 0.66;
p ⬍ 0.002).
At 6 months, positive correlations were observed between changes in cortical rCMRglc and changes in
CSF A␤40 levels, and a significant positive correlation
was observed with changes in the right temporal cortex
(r ⫽ 0.44; p ⬍ 0.05; see Fig 3C). No significant correlations between changes in cortical rCMRglc versus
CSF or plasma A␤42 concentrations were observed
throughout the study.
Correlation between Cerebrospinal Fluid and Plasma
Biomarkers, and Cognitive Function
At baseline, CSF A␤40 levels correlated positively with
the patients’ performance in the domain of attention
(composite of Digit Symbol and Trail Making Test-A)
Z-score (r ⫽ 0.59; p ⬍ 0.01; Fig 4A). At 3 months,
changes in plasma A␤40 levels correlated positively
with changes in the attention Z-score (r ⫽ 0.46; p ⬍
0.05; see Fig 4B).
At 6 months, changes in CSF A␤40 levels correlated
positively with changes in the attention Z-score (r ⫽
0.58; p ⬍ 0.01; see Fig 4C). In addition, at this time
point, changes in plasma A␤40 levels, only in the
phenserine group, correlated positively with attention
Z-score (r ⫽ 0.83; p ⬍ 0.004). These findings suggest
that an AD patient with a high A␤40 level in CSF or
plasma also had a high attention function. No significant correlation was observed among the changes in
CSF A␤42, plasma A␤42, and the attention Z-score.
Correlation between Pittsburgh Compound B
Retention and Cerebrospinal Fluid and
Plasma Biomarkers
A pattern of negative correlation, although not significant, was observed between mean cortical PIB retention and CSF A␤40 level at baseline (r ⫽ ⫺0.21; p ⬎
0.10), and between changes at 3 months (r ⫽ ⫺0.11;
p ⬎ 0.10; Fig 5A). However, a significant correlation
was found at 6 months (r ⫽ ⫺0.48; p ⬍ 0.03; see Fig
5B). This analysis indicates that patients with lower
PIB retention had increased CSF A␤40 levels. No significant correlations were observed between cortical
Š
Fig 4. Positive correlations between composite attention
Z-score (Digit Symbol and Trail Making Test-A) and cerebrospinal fluid (CSF) ␤-amyloid 40 (A␤40) concentrations (pg/
ml) at baseline (A); changes in composite attention Z-score
and changes in plasma A␤40 levels at 3 months (B); changes
in composite attention Z-score and changes in CSF A␤40
levels at 6 months (C). At 3 and 6 months, all values are
absolute value changes from individual baseline values. Solid
squares at 3 and 6 months ⫽ phenserine group; open circles
at 3 months ⫽ placebo group; open circles at 6 months ⫽
donepezil group. (A) r ⫽ 0.59; p ⬍ 0.01. (B) r ⫽ 0.46; p
⬍ 0.05. (C) r ⫽ 0.58; p ⬍ 0.01.
Kadir et al: PET Study of Phenserine in AD
627
Table. Demography Data
Characteristics
Total Subjects
Placebo/Donepezil
Phenserine
20
10
10
5/15
3/7
2/8
Age (yr)
68 ⫾ 2
70 ⫾ 3
66 ⫾ 3
Education (yr)
12 ⫾ 1
12 ⫾ 1
12 ⫾ 1
Duration of disease (yr)
4⫾1
4⫾1
4⫾1
First-degree relative with AD (Y/N)
6/14
2/8
4/6
Total subjects (n)
Sex (M/F)
Apolipoprotein E ε4 carriers (⫹⫹/⫺⫹/⫺⫺)
6/12/1
3/7/0
3/5/1
MMSE score at baseline
24 ⫾ 1
23 ⫾ 1
25 ⫾ 1
Data are presented as mean ⫾ standard error.
AD ⫽ Alzheimer’s disease; MMSE ⫽ Mini-Mental State Examination.
Fig 5. Negative correlations between changes in mean cortical Pittsburgh Compound B (PIB) retention and changes in cerebrospinal
fluid (CSF) ␤-amyloid 40 (A␤40) concentrations at 3 (A) and 6 months (B). Negative correlations between changes in mean cortical PIB retention and changes in the CSF total ␤-amyloid (A␤40 ⫹ A␤42) and ␤-secretase-cleaved amyloid precursor protein (␤sAPP) ratio at 3 (C) and 6 months (D). All values are changes from individual baseline values. Solid squares ⫽ phenserine
group; open circles at 3 months ⫽ placebo group; open circles at 6 months ⫽ donepezil group. (A) r ⫽ ⫺0.11; p ⬎ 0.10. (B)
r ⫽ ⫺0.48; p ⬍ 0.03. (C) r ⫽ ⫺0.45; p ⬍ 0.05. (D) r ⫽ ⫺0.42; p ⬍ 0.06.
628
Annals of Neurology
Vol 63
No 5
May 2008
PIB retention and levels of CSF A␤42 or plasma A␤40
and A␤42 throughout the study.
Finally, the possible relation between mean cortical
PIB retention and the CSF A␤ (A␤40 ⫹ A␤42)/␤sAPP ratio was investigated. At baseline, a pattern of
negative correlation was observed between mean cortical PIB retention and the CSF A␤/␤-sAPP ratio (r ⫽
⫺0.24; p ⬎ 0.10). Later, changes in mean cortical PIB
retention correlated negatively with changes in the CSF
A␤/␤-sAPP ratio at 3 months (r ⫽ ⫺0.45; p ⬍ 0.05;
see Fig 5C) and at 6 months (r ⫽ ⫺0.42; p ⬍ 0.06;
see Fig 5D). The analysis indicates that patients with
lower PIB retention had an increased CSF A␤/␤-sAPP
ratio.
Discussion
To date, this is the first study in which the new PET
amyloid tracer PIB has been used for evaluation of antiamyloid drug treatment in mild AD patients. Moreover, multitracer PET assessment was applied to investigate the relations between biological factors such as
rCMRglc and amyloid load, with CSF and plasma biomarkers, and cognitive function of AD patients at
regular follow-up intervals.
We observed an increase in rCMRglc in cortical regions after 3 months and maintenance after 6 months
of treatment with phenserine compared with baseline.
This finding is due to the ChEI effect of phenserine
treatment. Because the results of 3 months of placebo
treatment showed a decrease in rCMRglc in cortical
regions, this indicates a possible decrease in synaptic
activity during the course of AD, although 3 months is
a relatively short time to detect disease deterioration in
all patients. No significant effect was observed on rCMRglc when the placebo group switched to donepezil
treatment. The lack of effect on rCMRglc might be a
result of the low dose of donepezil (5mg) and/or variation in this small number of patients per group.
Several ChEIs, such as tacrine,33,34 donepezil,35 rivastigmine,36,37 and galantamine,38 have shown shortterm enhancement or long-term stabilization effects on
rCMRglc; they have also shown increases in cholinergic
neurotransmitter activity30,33,39 in AD patients. Thus,
the increase in rCMRglc observed after short-term
treatment with phenserine probably reflected a similar
effect as observed with other ChEIs, that is, an increase
in cholinergic neuronal signaling and activities, which
is highly energy dependent.
The neurotoxic peptide A␤ is a major component of
the extraneuronal plaques that characterize AD pathologically. Reduction of A␤ levels in the brain is a major
focus in the design and development of effective therapies for AD.40 In this study, no significant changes in
mean cortical PIB retention were observed after
phenserine or donepezil treatment, but at the individual level, changes in PIB retention correlated negatively
with the CSF A␤/␤-sAPP ratio. However, both in
vitro2,11 and in vivo animal41 studies have shown the
positive effect of phenserine on reduction of ␤-APP
and A␤. A plausible explanation for the lack of a significant decrease in PIB retention in vivo in AD patients is likely to be related to the dual mode of action
of phenserine (ie, AChE inhibition and APP reduction); AChE inhibitors are by nature dose limiting,
which prevent achievement of the necessary concentration for optimally inhibiting A␤ production in humans. The results of experimental studies show that
phenserine reduces the production and secretion of
newly synthesized, rather than already produced,
␤-APP and A␤ in a concentration- and timedependent manner with an inhibitory concentration of
50% (IC50) that is greater than its action on AChE.11
In this study, we found that composite neuropsychological test results correlated positively with rCMRglc,
but not with PIB retention. 18F-FDG PET reflects the
functional integrity of synapses, rather than the presence or quantity of plaques, which best correlates with
cognitive function in patients with AD.42 It has also
been reported that, in neuropathological studies, there
is a stronger correlation between degree of dementia
and neurofibrillary tangle density than amyloid
load.43,44
We did not find any significant treatment effect on
CSF and plasma biomarkers. However, we found a
tendency toward increased CSF and plasma A␤40 after
3 months of treatment with phenserine compared with
baseline. We also observed positive correlations between CSF and plasma A␤40, and rCMRglc and cognitive subtest attention, demonstrating that the patients
with the greatest A␤40 levels in CSF and plasma also
showed the greatest increase in rCMRglc and cognition. To our knowledge, no data are available concerning the direct interrelations among CSF and plasma
A␤, rCMRglc, and cognitive function. Our results suggest that, in AD patients, a high level of A␤40 in the
CSF or plasma is a better diagnostic marker. We also
showed that a high level of CSF A␤40 correlated with
low PIB retention in the brain. It should be noted that
A␤42, tau and p-tau did not correlate with imaging
and clinical measures. Whether high level of A␤40 in
the CSF or plasma is a feature of improvement in patients with AD will require further testing with larger
patient groups. However, the role of A␤40 in AD
pathogenesis is not as well established, but recent experimental45,46 and clinical47 studies have provided evidence for a protective role of A␤40 in AD.
In this study, the phenserine-treated AD patients
showed improvement in the composite neuropsychological performance compared with the placebo-treated
patients at 3 months. Within the group, the
phenserine-treated AD patients showed an improvement for up to 6 months compared with baseline. Af-
Kadir et al: PET Study of Phenserine in AD
629
ter 12 months, the performance was comparable with
the baseline level. These findings are in agreement with
our earlier studies on AD patients treated with the
ChEIs galantamine30 and rivastigmine48; however, untreated AD patients showed significant deterioration in
performance in similar neuropsychological tests after
12 months.48 On the other hand, once the placebo
group switched to active treatment after the end of 3
months, they showed nonsignificant changes (because
of the lack of power) compared with baseline data. The
global cognition was assessed by MMSE score, and the
power of this study was inadequate to draw clinical
conclusions on data addressing changes in global cognition.
An important limitation of this study is that relatively large numbers of statistical analyses were performed on PET parameters in relation to the CSF and
plasma biomarkers, and cognitive function, which were
not corrected for multiple comparisons, and this may
render some of the results liable to be obtained by
chance. Therefore, result patterns were taken into consideration rather than isolated findings. Conversely,
this study was designed primarily as a PET investigation, and thus included a small group of mild AD patients. These features make it necessary to be cautious
when interpreting the data because small sample size
may result in a lack of statistical power, although this
might make the significant results even more interesting.
In summary, we showed that phenserine (30mg/day)
results in a generalized increase in rCMRglc in cortical
regions after 3 months of treatment. The changes in
cortical rCMRglc were associated with cognitive function. Mean cortical PIB retention correlated negatively
with the total CSF A␤/␤-sAPP ratio. CSF and plasma
A␤40 concentrations were positively associated with
rCMRglc and the attention domain of cognition; conversely, CSF A␤40 levels were inversely correlated with
PIB retention in AD patients. An important outcome
of this study is information concerning the direct consequence of antiamyloid therapy on brain amyloid load
using 11C-PIB-PET technique.
This research was sponsored by the Swedish Research Council
(project no. 05817) (A.N.), the old Servants (A.N.), the KI foundations (A.N.), the Gun and Bertil Stohne’s Foundation (A.N.),
Swedish Brain Power (A.N.), and Axonyx Inc./Torreypines Therapeutics (A.N.).
We thank Drs N. Greig, M. Murphy, A. Marutle, and C. Hernandez for critically reviewing the manuscript and for their constructive
comments.
References
1. Ringman JM, Cummings JL. Current and emerging pharmacological treatment options for dementia. Behav Neurol 2006;17:
5–16.
630
Annals of Neurology
Vol 63
No 5
May 2008
2. Shaw KT, Utsuki T, Rogers J, et al. Phenserine regulates translation of beta-amyloid precursor protein mRNA by a putative
interleukin-1 responsive element, a target for drug development. Proc Natl Acad Sci U S A 2001;98:7605–7610.
3. Lahiri DK, Chen DM, Maloney B, et al. The experimental Alzheimer’s disease drug posiphen [(⫹)-phenserine] lowers
amyloid-beta peptide levels in cell culture and mice. J Pharmacol Exp Ther 2007;320:386 –396.
4. Chauhan NB, Siegel GJ. Intracerebroventricular passive immunization with anti-Abeta antibody in Tg2576. J Neurosci Res
2003;74:142–147.
5. Lombardo JA, Stern EA, McLellan ME, et al. Amyloid-beta
antibody treatment leads to rapid normalization of plaqueinduced neuritic alterations. J Neurosci 2003;23:10879 –10883.
6. Lemere CA, Beierschmitt A, Iglesias M, et al. Alzheimer’s disease abeta vaccine reduces central nervous system abeta levels in
a non-human primate, the Caribbean vervet. Am J Pathol
2004;165:283–297.
7. Gilman S, Koller M, Black RS, et al. Clinical effects of Abeta
immunization (AN1792) in patients with AD in an interrupted
trial. Neurology 2005;64:1553–1562.
8. Pakaski M, Kasa P. Role of acetylcholinesterase inhibitors in the
metabolism of amyloid precursor protein. Curr Drug Targets
CNS Neurol Disord 2003;2:163–171.
9. Racchi M, Mazzucchelli M, Porrello E, et al. Acetylcholinesterase inhibitors: novel activities of old molecules. Pharmacol Res
2004;50:441– 451.
10. Zimmermann M, Gardoni F, Marcello E, et al. Acetylcholinesterase inhibitors increase ADAM10 activity by promoting its
trafficking in neuroblastoma cell lines. J Neurochem 2004;90:
1489 –1499.
11. Greig NH, Sambamurti K, Yu QS, et al. An overview of
phenserine tartrate, a novel acetylcholinesterase inhibitor for the
treatment of Alzheimer’s disease. Curr Alzheimer Res 2005;2:
281–290.
12. Marutle A, Ohmitsu M, Nilbratt M, et al. Modulation of human neural stem cell differentiation in Alzheimer (APP23)
transgenic mice by phenserine. Proc Natl Acad Sci U S A 2007;
104:12506 –12511.
13. Klunk WE, Engler H, Nordberg A, et al. Imaging brain amyloid in Alzheimer’s disease with Pittsburgh Compound-B. Ann
Neurol 2004;55:306 –319.
14. Nordberg A. PET imaging of amyloid in Alzheimer’s disease.
Lancet Neurol 2004;3:519 –527.
15. McKhann G, Drachman D, Folstein M, et al. Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA
Work Group under the auspices of Department of Health and
Human Services Task Force on Alzheimer’s Disease. Neurology
1984;34:939 –944.
16. Mathis CA, Wang Y, Holt DP, et al. Synthesis and evaluation
of 11C-labeled 6-substituted 2-arylbenzothiazoles as amyloid
imaging agents. J Med Chem 2003;46:2740 –2754.
17. Engler H, Lundberg PO, Ekbom K, et al. Multitracer study
with positron emission tomography in Creutzfeldt-Jakob disease. Eur J Nucl Med Mol Imaging 2003;30:85–95.
18. Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-brain transfer constants from multiple-time
uptake data. J Cereb Blood Flow Metab 1983;3:1–7.
19. Minoshima S, Frey KA, Foster NL, Kuhl DE. Preserved pontine glucose metabolism in Alzheimer disease: a reference region
for functional brain image (PET) analysis. J Comput Assist Tomogr 1995;19:541–547.
20. Patlak CS, Blasberg RG. Graphical evaluation of blood-to-brain
transfer constants from multiple-time uptake data. Generalizations. J Cereb Blood Flow Metab 1985;5:584 –590.
21. Blomquist G, Engler H, Wall A, Langstrom B. Graphical analysis of time-activity data using a reference region with known
irreversible binding. Annales universitatis Turkuensis Medicaodontologica Abstracts of the IX Turku PET symposium, May
25–28, 2002, Turku, Finland.
22. Mirra SS, Gearing M, McKeel DW Jr, et al. Interlaboratory
comparison of neuropathology assessments in Alzheimer’s
disease: a study of the Consortium to Establish a Registry for
Alzheimer’s Disease (CERAD). J Neuropathol Exp Neurol
1994;53:303–315.
23. Yamaguchi H, Hirai S, Morimatsu M, et al. Diffuse type of
senile plaques in the cerebellum of Alzheimer-type dementia
demonstrated by beta protein immunostain. Acta Neuropathol
(Berl) 1989;77:314 –319.
24. Lopresti BJ, Klunk WE, Mathis CA, et al. Simplified quantification of Pittsburgh Compound B amyloid imaging PET
studies: a comparative analysis. J Nucl Med 2005;46:
1959 –1972.
25. Olsson A, Vanderstichele H, Andreasen N, et al. Simultaneous
measurement of beta-amyloid(1-42), total tau, and phosphorylated tau (Thr181) in cerebrospinal fluid by the xMAP technology. Clin Chem 2005;51:336 –345.
26. Ida N, Hartmann T, Pantel J, et al. Analysis of heterogeneous
A4 peptides in human cerebrospinal fluid and blood by a newly
developed sensitive Western blot assay. J Biol Chem 1996;271:
22908 –22914.
27. Jensen M, Hartmann T, Engvall B, et al. Quantification of Alzheimer amyloid beta peptides ending at residues 40 and 42 by
novel ELISA systems. Mol Med 2000;6:291–302.
28. Vanderstichele H, Van Kerschaver E, Hesse C, et al. Standardization of measurement of beta-amyloid(1-42) in cerebrospinal
fluid and plasma. Amyloid 2000;7:245–258.
29. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state.” A
practical method for grading the cognitive state of patients for
the clinician. J Psychiatr Res 1975;12:189 –198.
30. Kadir A, Darreh-Shori T, Almkvist O, et al. PET imaging of
the in vivo brain acetylcholinesterase activity and nicotine binding in galantamine-treated patients with AD. Neurobiol Aging
2007 Mar 20, doi: 10.1016/j.neurobiolaging.2007.02.020
31. Bergman I, Blomberg M, Almkvist O. The importance of impaired physical health and age in normal cognitive aging. Scand
J Psychol 2007;48:115–125.
32. Engler H, Forsberg A, Almkvist O, et al. Two-year follow-up of
amyloid deposition in patients with Alzheimer’s disease. Brain
2006;129:2856 –2866.
33. Nordberg A, Amberla K, Shigeta M, et al. Long-term tacrine
treatment in three mild Alzheimer patients: effects on nicotinic
receptors, cerebral blood flow, glucose metabolism, EEG, and
cognitive abilities. Alzheimer Dis Assoc Disord 1998;12:
228 –237.
34. Nordberg A, Lilja A, Lundqvist H, et al. Tacrine restores cholinergic nicotinic receptors and glucose metabolism in Alzheimer patients as visualized by positron emission tomography.
Neurobiol Aging 1992;13:747–758.
35. Tune L, Tiseo PJ, Ieni J, et al. Donepezil HCl (E2020) maintains functional brain activity in patients with Alzheimer
disease: results of a 24-week, double-blind, placebo-controlled
study. Am J Geriatr Psychiatry 2003;11:169 –177.
36. Potkin SG, Anand R, Fleming K, et al. Brain metabolic and
clinical effects of rivastigmine in Alzheimer’s disease. Int J Neuropsychopharmacol 2001;4:223–230.
37. Stefanova E, Wall A, Almkvist O, et al. Longitudinal PET evaluation of cerebral glucose metabolism in rivastigmine treated
patients with mild Alzheimer’s disease. J Neural Transm 2006;
113:205–218.
38. Mega MS, Dinov ID, Porter V, et al. Metabolic patterns associated with the clinical response to galantamine therapy: a
fludeoxyglucose f 18 positron emission tomographic study.
Arch Neurol 2005;62:721–728.
39. Kadir A, Darreh-Shori T, Almkvist O, et al. Changes in brain
(11)C-nicotine binding sites in patients with mild Alzheimer’s
disease following rivastigmine treatment as assessed by PET.
Psychopharmacology (Berl) 2007;191:1005–1014.
40. Sambamurti K, Greig NH, Lahiri DK. Advances in the cellular
and molecular biology of the beta-amyloid protein in Alzheimer’s disease. Neuromolecular Med 2002;1:1–31.
41. Haroutunian V, Greig N, Pei XF, et al. Pharmacological modulation of Alzheimer’s beta-amyloid precursor protein levels in
the CSF of rats with forebrain cholinergic system lesions. Brain
Res Mol Brain Res 1997;46:161–168.
42. Terry RD, Masliah E, Salmon DP, et al. Physical basis of cognitive alterations in Alzheimer’s disease: synapse loss is the major correlate of cognitive impairment. Ann Neurol 1991;30:
572–580.
43. Berg L, McKeel DW Jr, Miller JP, et al. Clinicopathologic
studies in cognitively healthy aging and Alzheimer’s disease: relation of histologic markers to dementia severity, age, sex, and
apolipoprotein E genotype. Arch Neurol 1998;55:326 –335.
44. Cummings BJ, Pike CJ, Shankle R, Cotman CW. Beta-amyloid
deposition and other measures of neuropathology predict cognitive status in Alzheimer’s disease. Neurobiol Aging 1996;17:
921–933.
45. Zou K, Gong JS, Yanagisawa K, Michikawa M. A novel function of monomeric amyloid beta-protein serving as an antioxidant molecule against metal-induced oxidative damage. J Neurosci 2002;22:4833– 4841.
46. Zou K, Kim D, Kakio A, et al. Amyloid beta-protein
(Abeta)1-40 protects neurons from damage induced by
Abeta1– 42 in culture and in rat brain. J Neurochem 2003;87:
609 – 619.
47. Kumar-Singh S, Theuns J, Van Broeck B, et al. Mean age-ofonset of familial Alzheimer disease caused by presenilin mutations correlates with both increased Abeta42 and decreased
Abeta40. Hum Mutat 2006;27:686 – 695.
48. Almkvist O, Darreh-Shori T, Stefanova E, et al. Preserved cognitive function after 12 months of treatment with rivastigmine
in mild Alzheimer’s disease in comparison with untreated AD
and MCI patients. Eur J Neurol 2004;11:253–261.
Kadir et al: PET Study of Phenserine in AD
631