Báo cáo khoa học: Anti-amyloid precursor protein immunoglobulins inhibit amyloid-b production by steric hindrance ppt

Anti-amyloid precursor protein immunoglobulins inhibit
amyloid-b production by steric hindrance
Rhian S. Thomas
1
, J. Eryl Liddell
2
and Emma J. Kidd
1
1 Welsh School of Pharmacy, Cardiff University, UK
2 Monoclonal Antibody Unit, School of Biosciences, Cardiff University, UK
Introduction
Alzheimer’s disease (AD) is characterized pathologi-
cally by an over-accumulation in the brain of intracel-
lular neurofibrillary tangles, amyloid-b (Ab)-containing
extracellular senile plaques and neuronal loss [1]. The
‘amyloid hypothesis’ suggests that Ab accumulation in
the brain is an initiating event in AD [2], although it
does not explain all aspects of AD pathology [3].
Despite this, it is still the dominant theory used to
explain the disease and many therapeutic strategies
have therefore concentrated on attempting to modify
Ab accumulation in the brain [4].
Ab is a 38–43-mer peptide that is cleaved from
amyloid precursor protein (APP) [5]. APP can be
processed by one of two proteolytic pathways. The
Keywords
Alzheimer’s disease; amyloid precursor
protein; amyloid-b; monoclonal antibodies;
b-secretase cleavage site
Correspondence
E. J. Kidd, Welsh School of Pharmacy,

Cardiff University, Redwood Building, King
Edward VII Avenue, Cardiff CF10 3NB, UK
Fax: +44 29 20874149
Tel: +44 29 20875803
E-mail:
Website: />contactsandpeople/fulltimeacademicstaff/
kidd-emmanew-overview_new.html
(Received 15 July 2010, revised 30
September 2010, accepted 27 October
2010)
doi:10.1111/j.1742-4658.2010.07942.x
The cleavage of amyloid precursor protein (APP) by b- and c-secretases
results in the production of amyloid-b (Ab) in Alzheimer’s disease. We
raised two monoclonal antibodies, 2B3 and 2B12, that recognize the
b-secretase cleavage site on APP but not Ab. We hypothesized that these
antibodies would reduce Ab levels via steric hindrance of b-secretase. Both
antibodies decreased extracellular Ab levels from astrocytoma cells, but
2B3 was more potent than 2B12. Levels of soluble sAPPa from the non-
amyloidogenic a-secretase pathway and intracellular APP were not affected
by either antibody nor were there any effects on cell viability. 2B3 exhib-
ited a higher affinity for APP than 2B12 and its epitope appeared to span
the cleavage site, whereas 2B12 bound slightly upstream. Both of these
factors probably contribute to its greater effect on Ab levels. After 60 min
incubation at pH 4.0, most 2B3 and 2B12 remained bound to their antigen,
suggesting that the antibodies will remain bound to APP in the acidic
endosomes where b-secretase cleavage probably occurs. Only 2B3 and
2B12, but not control antibodies, inhibited the cleavage of sAPPa by b-sec-
retase in a cell-free assay where the effects of antibody internalization
and intracellular degradation were excluded. 2B3 virtually abolished this
cleavage. In addition, levels of C-terminal APP fragments, generated

following b-secretase cleavage (bCTF), were significantly reduced in cells
after incubation with 2B3. These results strongly suggest that anti-cleavage
site IgGs can generically reduce Ab levels via inhibition of b-secretase by
steric hindrance and may provide a novel alternative therapy for Alzhei-
mer’s disease.
Abbreviations
Ab, amyloid-b; AD, Alzheimer’s disease; APP, amyloid precursor protein; BACE1, beta-site APP cleaving enzyme; bCTF, b-cleaved C-terminal
APP fragments; CI, confidence interval; MTS, [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt;
PBS, phosphate-buffered saline; PBST, phosphate-buffered saline with 0.05% Tween 20.
FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS 167
non-amyloidogenic route involves cleavage of APP by
a-secretase [6] within the Ab region to release sAPPa [7].
In the amyloidogenic pathway, b-secretase, identified as
beta-site APP cleaving enzyme (BACE1) [8–10], first
cleaves APP to liberate sAPPb and C99 [11]. The c-sec-
retase complex [12–15] then cleaves C99 to produce Ab
and a C-terminal intracellular fragment [11]. Although
there is still some debate regarding the co-localization of
the enzymes and substrates involved, it is generally con-
sidered that, after synthesis, a proportion of APP is
transported to the cell membrane and is then internal-
ized for processing in the endosomal–lysosomal system,
where it may be further processed to Ab [16,17].
Current therapy in the UK is limited to symptom-
atic treatment with acetyl-cholinesterase inhibitors for
moderate AD only [18]. Recently, there has been much
attention given to the development of novel immuno-
therapeutic approaches in AD. Ab vaccination, both
passive and active, has been used successfully in trans-
genic mice to reduce Ab plaque deposition [19] and

improve cognition [20–24]. This led to a phase IIA
clinical trial involving vaccination with aggregated
Ab42, AN1792, but this was halted when several
patients developed meningoencephalitis [25]. Alterna-
tive immunotherapeutic approaches are therefore
required.
Here we present data relating to two monoclonal
antibodies, 2B12 and 2B3. Unlike the previous
approaches, our antibodies do not bind to Ab, but
bind to APP in the vicinity of the b-secretase cleavage
site. We previously demonstrated that 2B12 reduced
levels of extracellular Ab40 from cell lines endoge-
nously expressing native APP in a time- and concen-
tration-dependent manner. The mode of action was,
however, unclear [26]. Here we present data character-
izing 2B3 and comparing the two antibodies. We dem-
onstrate that 2B3 also reduces Ab levels from a native
cell line and is more potent than 2B12. This effect on
Ab levels was specific to the cleavage site antibodies.
We suggest that these antibodies bind to the extracellu-
lar region of APP when it is transported to the cell
membrane and become internalized with the protein
into the endosomal ⁄ lysosomal system where they inhi-
bit BACE1 cleavage via steric hindrance. This will
drastically reduce levels of Ab produced, representing
an alternative therapeutic strategy to treat AD.
Results
2B3 and 2B12 have different epitopes on APP
Both 2B3, an IgG1 isotype, and 2B12, an IgG2b iso-
type, detected full-length APP from MOG-G-UVW

cell lysates in a western blot (Fig. 1). Both antibod-
ies detected bands of 103 kDa (± 2.5 kDa) and
56 kDa (± 0.6 kDa). However, 2B3 recognized the
56 kDa fragment, possibly a thrombin cleavage frag-
ment [27], more strongly than it did the 103 kDa
fragment of APP, whereas 2B12 recognized the
103 kDa fragment more strongly than it did the
56 kDa fragment. Neither antibody detected Ab40 in
an ELISA or western blot [26]; data not shown. The
antibodies differentially recognized sAPPa and
sAPPb in an ELISA (Fig. 1). There was no signifi-
cant difference in the recognition by 2B12 of sAPPa
or sAPP b. Similarly, there were no significant differ-
ences in the amount of 2B3 bound to sAPPa when
compared with the amount of 2B12 bound to either
sAPPa or sAPPb. However, significantly less 2B3
bound to sAPPb than sAPPa (P < 0.01). The
amount of 2B3 bound to sAPPb was also signifi-
cantly less than the binding of 2B12 to either sAPPa
(P < 0.001) or sAPPb (P < 0.001).
2B3 2B12
103 kDa
A
B
56 kDa
0
2
4
6
8

10
12
14
16
Absorbance (% of standard antibody)
2B3
2B12
sAPPα sAPPβ
**
***
***
Fig. 1. (A) Representative western blot of MOG-G-UVW lysate
(50 lg) detected with either 2B3 or 2B12 at 2 lgÆmL
)1
on a 10%
gel. Both antibodies detected full-length APP at 103 kDa (± 2.5 kDa)
and a smaller protein at 56 kDa (± 0.6 kDa), probably a thrombin
cleavage fragment of APP. 2B3 preferentially recognized the 56 kDa
fragment, n = 4–5. (B) Differential 2B3 and 2B12 recognition of
sAPPa and sAPPb as determined by indirect ELISA. Data are
expressed as a mean (± standard error of the mean) percentage of
the standard antibody (6E10) at 0.1 lgÆmL
)1
. 2B12 recognized
sAPPa and sAPPb equally well. However, significantly more 2B3
bound to sAPPa than to sAPPb. The amount of 2B3 bound to
sAPPb was also significantly less than the amount of 2B12 bound
to either sAPPa or sAPPb.**P < 0.01, ***P < 0.001 after one-way
ANOVA and Bonferroni post-hoc tests, n =4.
Antibodies inhibit amyloid-b production R. S. Thomas et al.

168 FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS
2B3 recognizes full-length APP and a peptide
spanning the cleavage site on APP more strongly
than 2B12
To determine the relative affinities of 2B3 and 2B12
for APP, we first compared their binding to a peptide,
Kb, which contains the b-secretase cleavage site on
APP (Fig. 2). The antibodies differentially bound to
the peptide, as evidenced by their differential Hill
slopes, 0.90 for 2B3 and 0.63 for 2B12 (P < 0.001).
The antibody concentration at which half-maximal
binding was reached was significantly lower for 2B3,
1.279 lgÆmL
)1
[95% confidence interval (CI) 1.153–
1.418], than for 2B12, 2.963 lgÆmL
)1
(95% CI 1.696–
5.177) (P < 0.001) and the maxima reached for 2B3,
103.7% (95% CI 100.7–106.7) was significantly greater
than that for 2B12, 73.44% (95% CI 62.34–84.54)
(P < 0.001). Furthermore, significantly more 2B3 than
2B12 was detected bound to Kb at all antibody con-
centrations greater than 0.1 lgÆmL
)1
(P < 0.05).
We next compared their binding efficiency to full-
length APP in a competition ELISA using MOG-G-
UVW cell lysate as a source of APP. Both 2B3 and
2B12 interfered with the binding of a commercial

detection antibody for APP in a concentration-depen-
dent manner (Fig. 3). At 5, 10 and 20 lgÆmL
)1
, 2B12
significantly reduced the binding of the commercial
antibody to 90.5% (P < 0.01), 89.13% (P < 0.001)
and 82.9% (P < 0.01) of control levels, respectively.
This was similar to previously reported levels [26]. At
5, 10 and 20 lgÆmL
)1
, 2B3 also significantly reduced
the binding of the second commercial antibody to
87.4% (P < 0.01), 82.33% (P < 0.001) and 72.94%
(P < 0.001) of control levels. Even at the highest con-
centration, 20 lgÆmL
)1
, a control IgG did not signifi-
cantly reduce the binding of the second detection
antibody when compared with control levels. The three
antibodies differentially inhibited the binding of the
APP detection antibody at 1 lgÆmL
)1
(P < 0.05),
5 lgÆmL
)1
(P < 0.001), 10 lgÆmL
)1
(P < 0.001) and
20 lgÆmL
)1

(P < 0.001). 2B3 and 2B12 produced a
significantly greater inhibition in the binding of the
APP detection antibody than did the control IgG at
all concentrations tested (P < 0.05). At 10 lgÆmL
)1
(P < 0.01) and 20 lgÆmL
)1
(P < 0.001), 2B3 also sig-
nificantly inhibited binding of the detection antibody
more than 2B12.
The majority of both 2B3 and 2B12 remained
bound to their antigen, Kb, after incubation at
different pH values
To determine the effects of pH on the antibodies, 2B3
and 2B12 were bound to the Kb peptide at pH 7.4 and
then incubated for 0 or 60 min at pH 4.0 or 7.4. There
0
20
40
60
80
100
120
0.00001 0.0001 0.001 0.01 0.1 1 10 100
Antibody concentration (µg·mL
–1
)
Absorbance (% of standard antibody)
2B3
2B12

*
*
*
*
*
*
*
*
*
Fig. 2. Half-maximal binding of 2B3 and 2B12 to a peptide, Kb,
spanning the b-secretase cleavage site, as determined by ELISA.
Data are expressed as a mean (± standard error of the mean) per-
centage of the standard antibody (6E10) at 0.05 lgÆmL
)1
. The con-
centration at which half-maximal binding was reached was
significantly lower for 2B3 than for 2B12 (P < 0.001). *Significant
differences between 2B3 and 2B12 with two-tailed Student’s
t-tests, P < 0.05, n = 3–4.
70
75
80
85
90
95
100
105
0 5 10 15 20 25
Antibody concentration (µ
g

·mL
–1
)
APP (% of control)
2B3
2B12
Control IgG
a
aa
a
b
b
***
***
**
b
**
b
b
c
**
b
***
c
Fig. 3. Binding of 2B3, 2B12 or a control IgG to APP as determined
by a competition sandwich ELISA for APP. Data are expressed as
mean (± standard error of the mean) percentage of the media con-
trol. One-way ANOVA indicated that the antibodies differentially
inhibited the binding of a commercial anti-APP IgG at each antibody
concentration. Antibody data points followed by different letters

(either a, b or c) differ significantly from each other in their ability to
inhibit the commercial antibody at that particular antibody concen-
tration after one-way ANOVA and Bonferroni (P < 0.05). 2B3 and
2B12 also significantly reduced the binding of the commercial anti-
body in comparison with media controls at 100%. **P < 0.01,
***P < 0.001, significantly different from media controls using two-
tailed Student’s t-tests. The control IgG did not significantly inter-
fere with the binding of the detection antibody, n =4.
R. S. Thomas et al. Antibodies inhibit amyloid-b production
FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS 169
were significant differences between the persistence in
binding of the antibodies to Kb under the various con-
ditions tested (P < 0.001) (Fig. 4). There was also a
significant interaction between the antibody type (2B3
or 2B12) and pH (P < 0.05). The complex formed
between 2B12 and Kb was not significantly affected by
pH or incubation time. The complex formed between
2B3 and Kb was also not significantly affected by the
incubation period, but there was a significant reduction
in 2B3 binding at pH 4.0 after 60 min (P < 0.05).
Incubation of Kb with phosphate-buffered saline with
0.05% Tween 20 (PBST) alone at pH 4.0 for 1 h, prior
to incubation with antibody at pH 7.4, did not affect
the binding of either 2B3 or 2B12 (data not shown),
indicating that the above results are not due to degra-
dation of the antigen. Importantly, even at pH 4.0,
there was still significantly more 2B3 bound to Kb
than 2B12.
2B3 is more effective at reducing extracellular
Ab40 and Ab42 levels in cell culture media than

2B12
The antibodies tested differentially inhibited levels of
extracellular Ab40 (P < 0.001) (Fig. 5A). 2B12 signifi-
cantly reduced levels of Ab40 in MOG-G-UVW cell
media to 65.3% of media control levels (P < 0.05).
This was similar to levels previously reported [26]. 2B3
significantly reduced levels of extracellular Ab40 to
36.8% of media control levels (P < 0.001). Neither
the control IgG nor the anti-N-terminal APP IgG had
any significant effect on Ab40 levels. Both 2B3 and
2B12 significantly reduced Ab40 more than the anti-
N-terminal APP IgG (P < 0.001, P < 0.05, respec-
tively) and 2B3 significantly reduced Ab40 levels more
than 2B12 (P < 0.05).
The antibodies also differentially inhibited levels of
Ab42 (P < 0.05). 2B12 significantly reduced Ab42 lev-
els to 54.8% of the media control (P < 0.01). 2B3 was
again more effective and reduced these levels to 21.9%
of media control levels (P < 0.01) (Fig. 5B). Ab42 levels
0
10
20
30
40
50
60
70
80
90
100

0 min 60 min 0 min 60 min
Absorbance (% of standard antibody)
2B3
2B12
***
***
*
*
pH 7.4 pH 4.0
Fig. 4. Effect of pH and incubation time on persistence of 2B3 and
2B12 binding to Kb as determined by ELISA. Data are expressed
as mean (± standard error of the mean) percentage absorbance of
the standard antibody, 6E10 (0.05 lgÆmL
)1
). Antibodies were first
allowed to form a complex with Kb at pH 7.4 and the complex was
then incubated with buffers of pH 4.0 or 7.4 for 0 and 60 min. The
complex formed between Kb and 2B3 was not affected by incuba-
tion time, but was significantly affected by the pH of the buffer.
The persistence in binding of 2B12 to Kb was not significantly
affected by either incubation time or pH. At all values tested, signif-
icantly more 2B3 remained bound to the Kb peptide than 2B12.
*P < 0.05, ***P < 0.001, 2B3 significantly different to 2B12. Data
were analysed for statistical significance with Generalized Linear
Model univariate analysis and Bonferroni post-hoc tests, n = 3–4.
0
20
40
60
80

100
120
140
A
B
2B3 2B12 IgG N
Aβ40 (% of control)
***
**
a
b
0
20
40
60
80
100
120
140
2B3 2B12 N
Aβ42 (% of control)
**
**
c
Fig. 5. Levels of extracellular Ab40 (A) and Ab42 (B) from MOG-
G-UVW culture media after incubation with 2B3, 2B12, an irrelevant
mouse IgG (Ab40 only) or an anti-N-terminal APP IgG (N), all at
10 lgÆmL
)1
for 48 h. Data are expressed as mean (± standard error

of the mean) percentage of media control Ab levels as detected in
a sandwich ELISA and corrected for total cell protein concentration.
Both 2B3 and 2B12 significantly reduced both forms of Ab from
media controls. Neither of the control antibodies had any significant
effect on Ab levels. *P < 0.05, **P < 0.01, ***P < 0.001, signifi-
cantly different from media controls (100%) with two-tailed
Student’s t-tests.
a
P < 0.05, significantly different from all other
groups;
b
P < 0.05 significantly different from 2B3 and N;
c
P < 0.05
significantly different from N after ANOVA and Tukey’s Honestly
Significant Difference, n = 3–6 (A) and n = 4 (B).
Antibodies inhibit amyloid-b production R. S. Thomas et al.
170 FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS
remained at 100.1% of control levels after incubation
with the anti-N-terminal APP IgG. Again, 2B3 signifi-
cantly reduced Ab42 levels more than the anti-N-ter-
minal IgG (P < 0.05).
Anti-b-secretase cleavage site IgGs do not alter
levels of APP, sAPPa, or affect cell viability, but
reduce b-cleaved C-terminal APP fragment (bCTF)
levels
Neither 2B12, 2B3 nor the irrelevant IgG had any effect
on levels of intracellular APP as measured in an ELISA
(Fig. 6A). Furthermore, they did not have any signifi-
cant effect on the number of viable cells, as measured

using an MTS assay with [3-(4,5-dimethylthiazol-2-yl)-
5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetra-
zolium, inner salt (Fig. 6B). The anti-N-terminal APP
IgG appeared to either increase the number of cells or
alter the levels of metabolic activity above control levels
(P < 0.01). Neither 2B12 nor the control IgG had
any significant effect on the levels of sAPPa (Fig. 7A).
However, incubation with 2B3 significantly reduced
bCTF levels to 64.6% (P < 0.05) of control values in
MOG-G-UVW cells (Fig. 7B).
Anti-b-secretase cleavage site IgGs reduce BACE1
cleavage via steric hindrance in a cell-free assay
Neither 2B3, 2B12, the control IgG nor the anti-N-ter-
minal APP IgG had any detectable effects on total lev-
els of BACE1 in MOG-G-UVW cells, as investigated
by western blotting (data not shown). However, differ-
ences were observed in BACE1 activity after in vitro
incubation of sAPPa with BACE1 and the antibodies.
The antibodies tested differentially inhibited the cleav-
age of sAPPa by BACE1, as detected by western blot-
ting (P < 0.001) (Fig. 8A,B). There was a decrease in
0
20
40
60
80
100
120
A
B

2B3 2B12 IgG
APP (% of control)
0
20
40
60
80
100
120
140
160
180
200
2B3 2B12 IgG N
Absorbance (% of control)
**
Fig. 6. Levels of intracellular APP as determined by ELISA (A) and
cell viability as determined by MTS assay (B) in MOG-G-UVW cells
after incubation with 2B3, 2B12, an anti-N-terminal APP IgG (N,
MTS only) or a control irrelevant mouse IgG, all at 10 lgÆ mL
)1
for
48 h. Data are expressed as mean (± standard error of the mean)
percentage of control (media only). APP levels were corrected for
total cell protein concentration. None of the antibodies tested had
any significant effect on levels of APP or cell viability, apart from the
anti-N-terminal APP IgG, which appeared to stimulate growth.
**P < 0.01, significantly different from media control using two-
tailed Student’s t-tests, n =3.
0

20
40
60
80
100
120
140
160
180
200
A
B
2B12 IgG
sAPPα (% of control)
0
20
40
60
80
100
120
2B3 2B12
βCTF (% of control)
*
Fig. 7. Levels of extracellular sAPPa (A) and intracellular bCTF (B)
as determined by ELISA from MOG-G-UVW cells after incubation
with 2B12, 2B3 or a control irrelevant mouse IgG, all at 10 lgÆmL
)1
for 48 h. Data are expressed as mean (± standard error of the
mean) percentage of control (media only). All levels were corrected

for total cell protein concentration. Neither 2B12 nor the IgG anti-
body significantly affected levels of sAPPa. However, 2B3 signifi-
cantly reduced levels of bCTF from control levels. *P < 0.05,
significantly different from media controls using two-tailed
Student’s t-tests, n =3.
R. S. Thomas et al. Antibodies inhibit amyloid-b production
FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS 171
sAPPa levels from control levels (sAPPa alone) after
incubation with BACE1 (media control) of )41.91 A
units. Neither the control IgG nor the anti-N-terminal
APP IgG had any significant effect on the activity of
BACE1 and decreases in sAPPa on the addition of
BACE1 were similar to those observed in the presence
of control media alone (–)6.9 and )44.4 A units,
respectively). The addition of both 2B3 and 2B12 sig-
nificantly inhibited the action of BACE1 from control
conditions. 2B12 significantly reduced this decrease in
sAPPa levels to only )17.29 A units (P < 0.05). 2B3
virtually abolished the reduction in sAPPa caused by
BACE1 and the decrease in sAPPa levels was only
)0.05 A units (P < 0.01). Levels of sAPPa were very
similar to the levels observed when sAPPa was incu-
bated with 2B3 alone.
Discussion
Both 2B3 and 2B12 were raised to the same immuno-
gen, yet they clearly have different epitopes, as evi-
denced by their binding profile to APP in western
blotting and by their differential binding to sAPPa
and sAPPb. 2B3 binds significantly less to sAPPb than
it does to sAPPa (25% less), whereas 2B12 recog-

nizes sAPPb and sAPPa equally well. The evidence
suggests that 2B12 binds upstream of the b-secretase
cleavage site (towards the N-terminal of APP), as it
recognizes both sAPPa and sAPPb. Its epitope, there-
fore, is in a region common to both peptides.
Although 2B3 also recognizes both of these APP frag-
ments, more 2B3 bound to sAPPa than to sAPPb.
However, 2B3 does not recognize the Ab peptide. This
suggests that 2B3 binds across the b-secretase cleavage
site, and only slightly into the Ab region (Fig. 9).
Results from a BLAST (.
gov) search on the Ka peptide used to raise both anti-
bodies indicated that they should be specific to APP as
this sequence of amino acids is highly conserved in
APP and not found in other mammalian sequences
(data not shown). In addition, we have previously
demonstrated that there was no cross-reaction between
2B12 and a range of peptides tested by ELISA or wes-
tern blotting [26].
2B3 binds more effectively to APP than 2B12, as
demonstrated by a competition assay, in which 2B3
interfered with the binding of a second detection anti-
body for APP more efficiently than 2B12. Although
these results could be due to the differential epitopes
or isotypes of 2B3 and 2B12, meaning that 2B3 might
interfere more effectively with the binding of the APP
detection antibody, they do suggest that 2B3 has a
higher relative affinity for APP than 2B12. This is sup-
ported by the results of the affinity rankings of the
two antibodies for the Kb peptide, a 15-mer frag-

ment spanning the b-secretase cleavage site on APP.
Significantly more 2B3 than 2B12 bound to Kb at all
sAPPα and BACE1 incubated with :-
–60
–50
–40
–30
–20
–10
0
10
2B3 2B12 IgG N
Media
control
Change in sAPPα from relevant
control (OD units)
**
*
12
A
B
34 5678 910
Fig. 8. (A) Representative western blot of sAPPa incubated with or
without BACE1 and 2B12, 2B3, control IgG, anti-N terminal APP
IgG (N) (all at 1 lgÆmL
)1
) or media control. Lane 1, sAPPa alone;
lane 2, sAPPa and BACE1 (media control); lane 3, sAPPa and
2B12; lane 4, sAPPa, BACE1 and 2B12; lane 5, sAPPa and 2B3;
lane 6, sAPPa, BACE1 and 2B3; lane 7, sAPPa and IgG; lane 8,

sAPPa, BACE1 and IgG; lane 9, sAPPa and N; lane 10, sAPPa,
BACE1 and N. Only 2B12 and 2B3 inhibited the action of BACE1 in
this system. (B) Quantification of western blots showing the mean
(± standard error of the mean) change in sAPPa from relevant con-
trol levels (i.e. sAPPa and relevant antibody alone) after sAPPa
incubation with BACE1 and either 2B3, 2B12, control IgG, anti-N
terminal APP IgG (N) or media alone. *P < 0.05, significantly differ-
ent from media control and N; **P < 0.01 significantly different
from media control, IgG and N, after one-way ANOVA and Bonfer-
roni post-hoc tests, n = 3–6.
α
α
-secretase
cleavage site
2B3
A
β
COOHNH
2
2B12
β
-secretase
cleavage site
sAPP
β
sAPP
α
APP
Fig. 9. Hypothesized epitopes of 2B3 and 2B12 on APP. 2B3 binds
across the b-secretase cleavage site, whereas 2B12 binds

upstream of this. Neither antibody recognizes Ab when cleaved
from APP.
Antibodies inhibit amyloid-b production R. S. Thomas et al.
172 FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS
concentrations greater than 0.1 lgÆmL
)1
and the con-
centration for half-maximal binding was significantly
lower for 2B3. All these data suggest that 2B3 has a
higher relative affinity for APP than 2B12.
We hypothesize that 2B3 and 2B12 will bind to the
APP ectodomain after APP has trafficked to the cell
membrane and will be endocytosed into the cell with the
protein. Indeed, Tampellini et al. [28] demonstrated that
antibodies to the N- and mid-region of Ab bound first
to the ectodomain of APP and were then internalized.
This immunocomplex would be formed at ‘normal phys-
iological pH’, presumably around pH 7.4. Once inter-
nalized, however, the complex may enter organelles,
where it would be subjected to much lower pH values,
potentially as low as pH 4.5 [29,30]. Low pH is well
known to affect antibody binding [31]. Therefore, we
tested the persistence of the immunocomplex, after it
had formed at pH 7.4, at two different pH values. The
persistence of the immunocomplex formed when 2B12
bound to Kb was not significantly affected by pH or
incubation time. Similarly, the Kb ⁄ 2B3 immunocom-
plex was not affected by incubation time, but a decrease
in pH did significantly reduce its persistence. Neverthe-
less, significantly more 2B3 than 2B12 remained bound

to the Kb peptide at all pH values and time points
tested, and 2B3 also retained nearly 70% of its original
binding capacity. We therefore suggest that both anti-
bodies would retain a large proportion of their biologi-
cal activity, even under the low pH conditions found in
the endosomal ⁄ lysosomal system [29,30]. Furthermore,
the lack of an effect of time suggests that the immuno-
complexes formed will persist for a biologically relevant
period of time.
Having demonstrated that 2B3 bound more effi-
ciently to APP than 2B12, we then investigated
whether it reduced the production of Ab40 in a similar
manner to 2B12. Both 2B3 and 2B12 significantly
reduced extracellular levels of Ab40. However, neither
the anti-N-terminal APP IgG nor the control mouse
IgG had any effect on Ab40. This suggests that it is
not sufficient to have an antibody that binds to APP
in order to reduce Ab, but that the antibodies must
bind in the vicinity of the b-secretase cleavage site to
accomplish this effect. We also investigated whether
either antibody affected the more aggregatory species
of Ab,Ab42, as there have been suggestions that the
majority of this peptide is cleaved from APP within
the trans-golgi network or the endoplasmic reticulum,
prior to the trafficking of APP to the cell membrane
[32]. If this were indeed the case, then our anti-cleav-
age site IgGs might be ineffective against this species
of Ab. Again, both 2B3 and 2B12 significantly reduced
Ab42 from control levels, but the N-terminal antibody
had no effect. This would suggest that at least a por-

tion of this peptide is produced elsewhere in the cell
rather than in the secretory pathway, and after APP
translocation to the cell membrane. It is interesting to
note that, in both cases, 2B3 reduced levels of Ab40
and Ab42 more than 2B12, although this was only sig-
nificant in the case of Ab40. This could be because of
the higher relative affinity of 2B3 for APP over 2B12.
Alternatively, it could be a function of its epitope, as
by binding closer to the cleavage site than 2B12, it
may block the access of BACE1 to APP via steric hin-
drance more effectively than 2B12. Clearly, informa-
tion regarding the mode of action of the antibodies is
important.
There are three predominant theories that have been
used to explain how Ab-specific antibodies may bring
about the clearance of Ab [33]: disruption of Ab aggre-
gates or neutralization of Ab oligomers, Fc-receptor-
mediated phagocytosis of Ab by microglia and the
peripheral sink hypothesis, in which the sequestration
of circulating Ab causes an efflux of Ab from the brain
to the plasma [34,35]. 2B3 and 2B12 are unlikely to
exert their effects by any of these mechanisms, as nei-
ther binds to Ab. We previously demonstrated that
2B12 was not toxic to cells in culture [26] and neither
the cleavage site antibodies nor the control IgG anti-
body affected cell viability levels here. The cleavage
site antibodies are therefore not reducing levels of Ab
by initiating cell death. In contrast, the anti-N-terminal
APP IgG led to increased absorbance levels in the
MTS assay. However, we did not explore this finding

any further because we had already demonstrated that
this commercially available antibody did not affect Ab
levels and therefore was likely to behave in a very dif-
ferent manner to the cleavage site antibodies. Further-
more, 2B12 did not alter detectable cleavage by the
non-amyloidogenic pathway in MOG-G-UVW cells, as
sAPPa levels remained unchanged.
It has been demonstrated that anti-N-terminal Ab
Igs, which can reduce Ab pathology
in vivo, can
reduce intracellular levels of Ab in vitro only when
internalized into cells [28], yet none of the theories
described above fully explain how intracellular Ab lev-
els may be reduced. We have demonstrated that 2B3
and 2B12 can reduce levels of extracellular A b.Itis
therefore probable that our antibodies are also reduc-
ing levels of intracellular Ab. This is particularly likely
as neither of the antibodies bind to the Ab peptide
itself and cannot therefore be increasing its extracellu-
lar degradation, or interfering with the assay. We
were, however, unable to measure intracellular Ab40
or Ab42 because of the low levels detectable in our
native cell lines.
R. S. Thomas et al. Antibodies inhibit amyloid-b production
FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS 173
Internalized anti-Ab Igs or anti-cleavage site IgGs
may inhibit the action of b-secretase. However, Tam-
pellini et al. [28] saw no evidence of this with anti-Ab
Igs. Tampellini et al. [28] observed that Ab and APP
ectodomain antibodies induced increased APP internal-

ization from the cell surface, which actually led to
enhanced cleavage by b-secretase and subsequently to
enhanced clearance of the antibody-bound bCTF frag-
ments in the lysosomal system. 2B3 and 2B12 did not
alter intracellular APP levels. Therefore, it seems unli-
kely that their mode of action is via increased degrada-
tion of APP. Unlike Tampellini et al. [28], we observed
a significant decrease in bCTF levels after incubation
with 2B3. This would imply that the cleavage site anti-
bodies do not induce increased internalization of APP
leading to enhanced cleavage by b-secretase, but that
they are reducing Ab by a different mechanism. It sug-
gests that they are inhibiting the cleavage of APP by
b-secretase. The lack of a significant effect after incu-
bation with 2B12 may be a result of the smaller effect
that this antibody has on Ab levels.
We hypothesized that 2B3 and 2B12 were blocking
the action of b-secretase by steric hindrance. We there-
fore devised a simple cell-free system to investigate this
hypothesis that avoided complications from other
cellular components and overcame the low levels of
APP fragments in the native cell lines. Both 2B3 and
2B12 drastically reduced or nearly abolished BACE1
cleavage of sAPPa. This clearly demonstrates that
anti-cleavage site IgGs are capable of inhibiting
BACE1 in vitro. The presence of a large protein (IgG)
did not nonspecifically block the action of BACE1.
Crucially, our results demonstrate that antibody epi-
topes are vitally important to this inhibition, as the
anti-N-terminal APP IgG, which binds some distance

from the cleavage site, had no such effect on BACE1.
In this simple in vitro system, any effects of APP inter-
nalization or enzyme ⁄ substrate co-localization are
eliminated. In conjunction with the observed decrease
in bCTF levels, results from the cell-free assay system
suggest that the mode of action of our cleavage site
antibodies is probably via steric hindrance.
Similar effects on Ab levels were obtained by Arbel
et al. [36], who produced monoclonal antibodies using
a peptide containing part of the Swedish mutation at
the b-secretase cleavage site. They also demonstrated a
reduction in both extracellular and intracellular Ab
levels, but in cell lines over-expressing APP. These
antibodies have also been shown to improve cognition
in the Tg2576 Swedish mutation mouse model of AD
pathology [37] and to reduce Ab levels in the V717I
London mutation mouse model [38]. We believe that
our use of model cell lines that do not over-express
APP is very important, as the majority of cases of AD
occur in people with much lower levels of APP than
those associated with transfected cells. As far as we are
aware, we have demonstrated for the first time that the
most likely mode of action for such antibodies is via
steric hindrance.
Immunotherapy for AD remains an exciting pros-
pect, despite the failure of the AN1792 clinical trial
[39]. Passive immunization with b-secretase cleavage
site antibodies might alleviate some of the problems
associated with this trial, such as the T-lymphocyte
meningoencephalitis and cerebral micro-haemorrhages

[40,41]. These antibodies would not bind to existing
Ab and would not therefore stimulate the T-lympho-
cyte response or lead to the excessive complement
activation that some believe would be a problem with
Ab antibodies [42]. In conjunction with other immuno-
therapeutic strategies to reduce plaque load, such anti-
bodies may have a considerable impact on the
development of disease-modifying treatments for AD.
Materials and Methods
Materials and cell culture
All chemicals and reagents were purchased from Sigma-
Aldrich (Poole, UK) or Fisher Scientific (Leicester, UK)
and all reactions were performed at room temperature
unless otherwise specified.
Astrocytoma cells, MOG-G-UVW (ECACC, Porton
Down, UK), were cultured in a 1 : 1 mix of Ham’s F10
and Dulbecco’s modified Eagle’s medium supplemented
with 10% fetal bovine serum (Perbio Science UK Ltd,
Cramlington, UK) and 2 mml-glutamine.
Antibody production and isotyping
Full details of the immunization protocol and hybridoma
development are detailed elsewhere [26]. Antibodies (2B12
and 2B3) were raised to a 15-mer peptide spanning the
b-secretase cleavage site on APP, EEISEVKMDAEFRHD,
termed Ka. Both antibodies were concentrated from
culture medium using Amicon Centriplus YM-100 filters
(Millipore, Watford, UK) with a nominal molecular mass
cut-off of 100 kDa and the isotype determined using the
Isostrip mouse monoclonal antibody isotyping kit (Serotec,
Oxford, UK).

Western blotting was performed using standard meth-
ods. Briefly, samples were resolved on 10% polyacrylamide
gels, transferred on to 0.2 lm nitrocellulose membranes
(Amersham Biosciences, Little Chalfont, UK), incubated
with the relevant antibody and detected as previously
described [26].
Antibodies inhibit amyloid-b production R. S. Thomas et al.
174 FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS
Determination of antibody epitopes
The epitopes of 2B3 and 2B12 on APP were investigated
by western blotting, as above, and by comparing the
relative binding profiles of the antibodies with cleavage
products of APP in an indirect ELISA. Recombinant
sAPPa and sAPPb (Sigma-Aldrich) were adsorbed to a
96-well microtitre plate (Greiner Bio-One, Stonehouse,
UK) at 5 lgÆmL
)1
in carbonate ⁄ bicarbonate buffer (15 mm
Na
2
CO
3
,35mm NaHCO
3
, pH 9.8) overnight at 4 °C.
Plates were blocked with 1% nonfat milk powder for 1 h,
then 2B3 or 2B12 was subsequently incubated at 1 lgÆmL
)1
for 2 h. Antibodies were detected with a secondary anti-
mouse IgG conjugated to horseradish peroxidase, 1 : 2500

(Pierce, Rockford, IL, USA) for 1 h and visualized with
the enzyme substrate, o-phenylenediamine in a 0.1 m citrate
phosphate buffer (24 mm citric acid, 51 mm Na
2
HPO
4
,pH
5.0), incubated for 20 min. The reaction was stopped with
2.5 m H
2
SO
4
and the absorbance determined at 492 nm.
The diluent used on day 2 was phosphate-buffered saline
(PBS; 137 mm NaCl, 1.5 mm KH
2
PO
4
,8mm Na
2
HPO
4
,
2.5 mm KCl, pH 7.4) with 0.05% Tween 20 (PBST). All
results were expressed as a proportion of the standard anti-
body (6E10, Cambridge BioScience Ltd, Cambridge, UK)
at 0.1 lgÆmL
)1
bound to sAPPa, to correct for any inter-
plate variation.

Quantification of APP
APP was quantified using the APP DuoSet (R&D Systems,
Abingdon, UK) following the manufacturer’s guidelines
[26]. Briefly, the capture antibody was used at 4 lgÆmL
)1
in
PBS overnight. Plates were blocked with 1% bovine serum
albumin and 5% sucrose in PBS and samples were quanti-
fied using a six-point standard curve. The biotinylated
detection antibody was used at 300 ngÆmL
)1
and detected
using streptavidin–horseradish peroxidase and o-phenylene-
diamine.
Affinity ranking of 2B3 and 2B12 for an APP
fragment
Affinity ranking of the two antibodies was accomplished by
comparing their binding properties to a peptide, Kb, which
spans the b-secretase cleavage site on APP, in an indirect
ELISA. This peptide represents a 15-mer sequence (SEV-
KMDAEFRHDSGY), slightly further into the Ab region
of APP than Ka, the immunizing peptide. ELISA methods
followed those detailed above with the following exceptions.
Kb was adsorbed to a 96-well microtitre plate at a concen-
tration of 10 l g ÆmL
)1
in carbonate ⁄ bicarbonate buffer
overnight at 4 °C. 2B3 or 2B12 was incubated for 2 h at
concentrations ranging from 0.00001 to 30 lgÆmL
)1

and
detected as above. All results were expressed as a propor-
tion of the standard antibody (6E10) at 0.05 lgÆmL
)1
.
Binding of 2B3 and 2B12 to full-length APP
A competition assay, in conjunction with the sandwich
ELISA for APP (R&D Systems) described above, was used
to determine relative binding of 2B3 and 2B12 to APP from
cell lysates. MOG-G-UVW cells were lysed and concen-
trated through a filter with a nominal cut-off of 100 kDa
(Millipore) to provide predominantly full-length APP at a
concentration of 30 ngÆmL
)1
, as described previously [26].
After formation of the APP ⁄ capture antibody complex on
the 96-well plate and prior to incubation with the detection
antibody, the test antibodies, 2B3, 2B12 or control IgG
(Pierce), were incubated at concentrations ranging from 1
to 20 lgÆ mL
)1
for 1 h. Binding of these antibodies was then
inferred by a decrease in binding of the detection antibody
compared with the PBST control alone.
Persistence of 2B3 and 2B12 binding at different
pH values
The binding persistence of 2B3 and 2B12 to Kb was investi-
gated at two different pH values using an indirect ELISA.
The methods followed those detailed above with the follow-
ing modifications. Kb was adsorbed to a 96-well plate and

blocked as above. 2B3 and 2B12 (5 lgÆmL
)1
) were incubated
with Kb for 1 h in PBST (pH 7.4) and the antibody solution
was aspirated. The immunocomplex was then incubated for
a further 1 h in PBST at either pH 7.4 or 4.0 for 0 or
60 min. To ensure that the antigen was not degraded or dis-
sociated from the plate by the pH treatment, Kb was also
incubated with PBST alone at pH 4.0 for 1 h prior to incu-
bation with 2B3 or 2B12 at pH 7.4. Binding of both antibod-
ies was detected as above and all results were expressed as a
proportion of a standard antibody (6E10) at 0.05 lgÆmL
)1
.
PBST was adjusted to the correct pH with H
3
PO
4
.
Effects of 2B3 and 2B12 on levels of Ab40, Ab42,
sAPPa and bCTF
All experiments were performed in 24-well cluster plates, in
triplicate, with a starting density of 25 000 MOG-G-UVW
cells per well. Cells were allowed to attach overnight and
were then incubated with control media, 2B3, 2B12, an
anti-N-terminal APP IgG (22C11, Millipore) or an irrele-
vant control mouse IgG (Sigma-Aldrich) (all at
10 lgÆmL
)1
) for 48 h at 37 °C. This was repeated on a min-

imum of three different passage numbers, where each
n = 1 passage. For analysis of Ab40, media was subjected
to immunoprecipitation and ELISA as described previously
[26]. Briefly, the ELISA employed the N-terminal Ab anti-
body 6E10 (5 lgÆmL
)1
) as the capture antibody and affin-
ity-purified BAM401AP (0.45 lgÆmL
)1
Autogen Bioclear,
Calne, UK), specific to the C-terminus of human Ab40, as
the detection antibody. For analysis of Ab42, media was
collected after antibody treatment and tested in a sandwich
R. S. Thomas et al. Antibodies inhibit amyloid-b production
FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS 175
ELISA (Biosource, Invitrogen, Paisley, UK). To determine
the effect of the antibodies on sAPPa, MOG-G-UVW cells
were incubated as before and media was tested in a sand-
wich ELISA (IBL, Hamburg, Germany). All cells were
lysed and intracellular APP and bCTF (IBL) levels were
detected by ELISA, as described above. All data were nor-
malized to total cell protein concentration as determined by
bicinchoninic acid protein assay (Pierce).
Effect of 2B3 and 2B12 on MOG-G-UVW cell
viability
Viability studies were performed on MOG-G-UVW cells
after incubation with 2B3, 2B12, control IgG (Sigma-
Aldrich), anti-N-terminal APP IgG (22C11, Millipore) or
media control, all at 10 lgÆmL
)1

, using the CellTiter 96
Ò
MTS Aqueous One Solution Cell Proliferation Assay (Pro-
mega, Southampton, UK) in 96-well cluster plates. MOG-
G-UVW cells were first allowed to adhere overnight after
plating at a concentration of 2000 per well, and were then
incubated with treatments for 48 h. Viability was assessed
following the manufacturer’s guidelines.
Effect of 2B3 and 2B12 on BACE1 activity in a
cell-free assay
The effect of 2B3 and 2B12 on BACE1 levels was investi-
gated, after antibody treatment as detailed above, and analy-
sed by western blotting. 12.5 g total protein was run on the
polyacrylamide gel and detected with anti-BACE1 IgG
(0.27 lgÆmL
)1
, Santa Cruz Biotechnology, Santa Cruz,
USA). The effect of the antibodies on BACE1 activity was
also investigated in a cell-free assay. Recombinant human
sAPPa (4 lgÆmL
)1
, R&D Systems), containing the b-secre-
tase cleavage site, was incubated in the presence or absence
of BACE1 (82.5 lgÆmL
)1
, R&D Systems) for 1 h at 37 °C,
with the addition of one of the following treatments, 2B3,
2B12, control IgG (Pierce), anti-N-terminal APP IgG or
media control. Prior to the addition of BACE1, the antibod-
ies and sAPPa were allowed to form a complex for 3 min.

The total reaction volume was 20 lL; all reactions were per-
formed in 50 mm C
2
H
3
O
2
Na (pH 4.5) and all antibodies
were used at 1 lgÆmL
)1
. The reaction was stopped by the
addition of 3· Laemmli sample buffer [43]. A volume equiv-
alent to 26.67 ng starting sAPPa was analysed in a western
blot and detected with the anti-N-terminal APP IgG
(33.3 ngÆmL
)1
).
Statistical analyses
Data generated in ELISA assays were quantified by com-
paring data with standard curves included on each plate,
using graphpad prism
Ò
4. The results were first normalized
to total protein concentration, where relevant, and
expressed as a percentage of media control values. MTS
and ELISA results were then analysed using a Student’s t-
test at the two-tailed significance level to determine if con-
centrations were significantly different to media controls
(100%). Where relevant, ELISA data were subsequently
analysed with one-way ANOVA.

To compare the relative affinities of 2B3 and 2B12 for
the Kb peptide, log antibody concentration was plotted
against the percentage absorbance (of standard antibody)
and a sigmoidal dose–response curve fitted to the data
using graphpad prism
Ò
4. Curve parameters were com-
pared using the F-test and differences between 2B3 and
2B12 at each concentration were compared using a Stu-
dent’s t-test. The persistence in binding of 2B3 and 2B12 to
Kb at different pH values was investigated using General-
ized Linear Model univariate analysis, with absorbance as
the dependent variable and pH, antibody and time as fac-
tors. Both antibodies were subsequently investigated inde-
pendently with ANOVA and Bonferroni.
Western blots for the inhibitory effects of 2B3 and 2B12
on BACE1 were quantified using nih imager. All bands were
first normalized to, and expressed as a percentage of, sAPPa
alone, to allow comparisons between blots. The normalized
control density of sAPPa and antibody was subtracted from
the relevant experimental condition of sAPPa, antibody and
BACE1 to determine the change in sAPPa after incubation
with BACE1. The resulting change in sAPPa was analysed
using one-way ANOVA and Bonferroni.
Where necessary, data were transformed to fulfil the
assumptions of normality and homoskedasticity and, there-
fore, to allow the use of parametric testing.
Acknowledgements
This work was funded by grant number 79 from the
Alzheimer’s Society, UK. We would like to thank

Katrin Hack, Pavlina Doubkova, Lynne Murphy and
Shahista Jaffer for their technical assistance with this
project.
References
1 Selkoe DJ (2001) Alzheimer’s disease: genes, proteins,
and therapy. Physiol Rev 81, 741–766.
2 Hardy J & Allsop D (1991) Amyloid deposition as the
central event in the aetiology of Alzheimer’s disease.
Trends Pharmacol Sci 12, 383–388.
3 Bennett DA, Schneider JA, Wilson RS, Bienias JL &
Arnold SE (2004) Neurofibrillary tangles mediate the
association of amyloid load with clinical Alzheimer
disease and level of cognitive function. Arch Neurol 61,
378–384.
4 Hamaguchi T, Ono K & Yamada M (2006) Anti-
amyloidogenic therapies: strategies for prevention and
Antibodies inhibit amyloid-b production R. S. Thomas et al.
176 FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS
treatment of Alzheimer’s disease. Cell Mol Life Sci 63,
1538–1552.
5 Mori H, Takio K, Ogawara M & Selkoe DJ (1992)
Mass spectrometry of purified amyloid beta protein in
Alzheimer’s disease. J Biol Chem 267, 17082–17086.
6 Asai M, Hattori C, Szabo B, Sasagawa N, Maruyama
K, Tanuma S & Ishiura S (2003) Putative function of
ADAM9, ADAM10, and ADAM17 as APP alpha-
secretase. Biochem Biophys Res Commun 301, 231–235.
7 Allinson TM, Parkin ET, Turner AJ & Hooper NM
(2003) ADAMs family members as amyloid precursor
protein alpha-secretases. J Neurosci Res 74, 342–352.

8 Sinha S, Anderson JP, Barbour R, Basi GS, Caccavello
R, Davis D, Doan M, Dovey HF, Frigon N, Hong J
et al. (1999) Purification and cloning of amyloid precur-
sor protein beta-secretase from human brain. Nature
402, 537–540.
9 Vassar R, Bennett BD, Babu-Khan S, Kahn S, Mendiaz
EA, Denis P, Teplow DB, Ross S, Amarante P, Loeloff
R et al. (1999) Beta-secretase cleavage of Alzheimer’s
amyloid precursor protein by the transmembrane aspar-
tic protease BACE. Science 286, 735–741.
10 Yan R, Bienkowski MJ, Shuck ME, Miao H, Tory
MC, Pauley AM, Brashier JR, Stratman NC, Mathews
WR, Buhl AE et al. (1999) Membrane-anchored aspar-
tyl protease with Alzheimer’s disease beta-secretase
activity. Nature 402, 533–537.
11 Citron M (2004) Strategies for disease modification in
Alzheimer’s disease. Nat Rev Neurosci 5, 677–685.
12 Edbauer D, Winkler E, Regula JT, Pesold B, Steiner H
& Haass C (2003) Reconstitution of gamma-secretase
activity. Nat Cell Biol 5, 486–488.
13 Kimberly WT, LaVoie MJ, Ostaszewski BL, Ye W,
Wolfe MS & Selkoe DJ (2003) Gamma-secretase is a
membrane protein complex comprised of presenilin,
nicastrin, Aph-1, and Pen-2. Proc Natl Acad Sci USA
100, 6382–6387.
14 Takasugi N, Tomita T, Hayashi I, Tsuruoka M, Niim-
ura M, Takahashi Y, Thinakaran G & Iwatsubo T
(2003) The role of presenilin cofactors in the gamma-
secretase complex. Nature 422, 438–441.
15 Chen F, Hasegawa H, Schmitt-Ulms G, Kawarai T,

Bohm C, Katayama T, Gu Y, Sanjo N, Glista M, Rog-
aeva E et al. (2006) TMP21 is a presenilin complex
component that modulates gamma-secretase but not
epsilon-secretase activity. Nature 440, 1208–1212.
16 Yamazaki T, Koo EH & Selkoe DJ (1996) Trafficking
of cell-surface amyloid beta-protein precursor. II. Endo-
cytosis, recycling and lysosomal targeting detected by
immunolocalization. J Cell Sci 109, 999–1008.
17 Koo EH, Squazzo SL, Selkoe DJ & Koo CH (1996)
Trafficking of cell-surface amyloid beta-protein
precursor. I. Secretion, endocytosis and recycling as
detected by labeled monoclonal antibody. J Cell Sci
109, 991–998.
18 Accessed 1 July
2010.
19 Schenk D, Barbour R, Dunn W, Gordon G, Grajeda
H, Guido T, Hu K, Huang JP, Johnson-Wood K,
Khan K et al. (1999) Immunization with amyloid-beta
attenuates Alzheimer disease-like pathology in the
PDAPP mouse. Nature 400, 173–177.
20 Buttini M, Masliah E, Barbour R, Grajeda H, Motter
R, Johnson-Wood K, Khan K, Seubert P, Freedman S,
Schenk D et al. (2005) Beta-amyloid immunotherapy
prevents synaptic degeneration in a mouse model of
Alzheimer’s disease. J Neurosci 25
, 9096–9101.
21 Bard F, Cannon C, Barbour R, Burke RL, Games D,
Grajeda H, Guido T, Hu K, Huang J, Johnson-Wood
K et al. (2000) Peripherally administered antibodies
against amyloid beta-peptide enter the central nervous

system and reduce pathology in a mouse model of Alz-
heimer disease. Nat Med 6, 916–919.
22 Janus C, Pearson J, McLaurin J, Mathews PM, Jiang
Y, Schmidt SD, Chishti MA, Horne P, Heslin D,
French J et al. (2000) A beta peptide immunization
reduces behavioural impairment and plaques in a model
of Alzheimer’s disease. Nature 408, 979–982.
23 Morgan D, Diamond DM, Gottschall PE, Ugen KE,
Dickey C, Hardy J, Duff K, Jantzen P, DiCarlo G, Wil-
cock D et al. (2000) A beta peptide vaccination prevents
memory loss in an animal model of Alzheimer’s disease.
Nature 408, 982–985.
24 Lambert MP, Viola KL, Chromy BA, Chang L, Mor-
gan TE, Yu J, Venton DL, Krafft GA, Finch CE &
Klein WL (2001) Vaccination with soluble Abeta oligo-
mers generates toxicity-neutralizing antibodies. J Neuro-
chem 79, 595–605.
25 Schenk D (2002) Amyloid-beta immunotherapy for
Alzheimer’s disease: the end of the beginning. Nat Rev
Neurosci 3, 824–828.
26 Thomas RS, Liddell JE, Murphy LS, Pache DM &
Kidd EJ (2006) An antibody to the beta-secretase cleav-
age site on amyloid-beta-protein precursor inhibits amy-
loid-beta production. J Alzheimers Dis 10, 379–390.
27 Chong YH, Jung JM, Choi W, Park CW, Choi KS &
Suh YH (1994) Bacterial expression, purification of full
length and carboxyl terminal fragment of Alzheimer
amyloid precursor protein and their proteolytic process-
ing by thrombin. Life Sci 54, 1259–1268.
28 Tampellini D, Magrane J, Takahashi RH, Li F, Lin

MT, Almeida CG & Gouras GK (2007) Internalized
antibodies to the Abeta domain of APP reduce neuro-
nal Abeta and protect against synaptic alterations.
J Biol Chem 282, 18895–18906.
29 Geisow MJ & Evans WH (1984) pH in the endosome.
Measurements during pinocytosis and receptor-medi-
ated endocytosis. Exp Cell Res 150, 36–46.
30 Murphy RF, Powers S & Cantor CR (1984) Endosome
pH measured in single cells by dual fluorescence flow
R. S. Thomas et al. Antibodies inhibit amyloid-b production
FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS 177
cytometry: rapid acidification of insulin to pH 6. J Cell
Biol 98, 1757–1762.
31 Dejaegere A, Choulier L, Lafont V, De Genst E & Alt-
schuh D (2005) Variations in antigen-antibody associa-
tion kinetics as a function of pH and salt concentration:
a QSAR and molecular modeling study. Biochemistry
44, 14409–14418.
32 Haass C, Lemere CA, Capell A, Citron M, Seubert P,
Schenk D, Lannfelt L & Selkoe DJ (1995) The Swedish
mutation causes early-onset Alzheimer’s disease by
beta-secretase cleavage within the secretory pathway.
Nat Med 1 , 1291–1296.
33 Weiner HL & Frenkel D (2006) Immunology and
immunotherapy of Alzheimer’s disease. Nat Rev
Immunol 6, 404–416.
34 Wilcock DM, DiCarlo G, Henderson D, Jackson J,
Clarke K, Ugen KE, Gordon MN & Morgan D (2003)
Intracranially administered anti-A beta antibodies
reduce beta-amyloid deposition by mechanisms both

independent of and associated with microglial activa-
tion. J Neurosci 23, 3745–3751.
35 Wilcock DM, Rojiani A, Rosenthal A, Levkowitz G,
Subbarao S, Alamed J, Wilson D, Wilson N, Freeman
MJ, Gordon MN et al. (2004) Passive amyloid immuno-
therapy clears amyloid and transiently activates micro-
glia in a transgenic mouse model of amyloid deposition.
J Neurosci 24, 6144–6151.
36 Arbel M, Yacoby I & Solomon B (2005) Inhibition of
amyloid precursor protein processing by beta-secretase
through site-directed antibodies. Proc Natl Acad Sci
USA 102, 7718–7723.
37 Rakover I, Arbel M & Solomon B (2007) Immunother-
apy against APP b-secretase cleavage site improves
cognitive function and reduces neuroinflammation in
Tg2576 mice without a significant effect on brain Ab
levels. Neurodegener Dis 4, 392–402.
38 Solomon B & Frenkel D (2010) Immunotherapy for
Alzheimer’s disease. Neuropharmacology 59 , 303–309.
39 Holmes C, Boche D, Wilkinson D, Yadegarfar G,
Hopkins V, Bayer A, Jones RW, Bullock R, Love S,
Neal JW et al. (2008) Long-term effects of Abeta42
immunisation in Alzheimer’s disease: follow-up of a
randomised, placebo-controlled phase I trial. Lancet 372,
216–223.
40 Nicoll JAR, Wilkinson D, Holmes C, Steart P, Mark-
ham H & Weller RO (2003) Neuropathology of human
Alzheimer disease after immunization with amyloid-beta
peptide: a case report. Nat Med 9, 448–452.
41 Ferrer I, Boada Rovira M, Sanchez Guerra ML, Rey

MJ & Costa-Jussa F (2004) Neuropathology and
pathogenesis of encephalitis following amyloid-beta
immunization in Alzheimer’s disease. Brain Pathol 14,
11–20.
42 McGeer PL (2008) Amyloid-beta vaccination for
Alzheimer’s dementia. Lancet 372, 1381–1382.
43 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
Antibodies inhibit amyloid-b production R. S. Thomas et al.
178 FEBS Journal 278 (2011) 167–178 ª 2010 The Authors Journal compilation ª 2010 FEBS