Cotinine reduces amyloid-β aggregation and improves memory in Alzheimer's disease mice.

Abstract

Alzheimer's disease (AD) affects millions of people world-wide and new effective and safe therapies are needed. Cotinine, the main metabolite of nicotine, has a long half-life and does not have cardiovascular or addictive side effects in humans. We studied the effect of cotinine on amyloid-β (Aβ) aggregation as well as addressed its impact on working and reference memories. Cotinine reduced Aβ deposition, improved working and reference memories, and inhibited Aβ oligomerization in the brains of transgenic (Tg) 6799 AD mice. In vitro studies confirmed the inhibitory effect of cotinine on Aβ1-42 aggregation. Cotinine stimulated Akt signaling, including the inhibition of glycogen synthase kinase 3β (GSK3β), which promotes neuronal survival and the synaptic plasticity processes underlying learning and memory in the hippocampus and cortex of wild type and Tg6799 AD mice. Simulation of the cotinine-Aβ1-42 complex using molecular dynamics showed that cotinine may interact with key histidine residues of Aβ1-42, altering its structure and inhibiting its aggregation. The good safety profile in humans and its beneficial effects suggest that cotinine may be an excellent therapeutic candidate for the treatment of AD.

Full text

Page 1

Uncorrected Author Proof
characterized by memory loss, synaptic dysfunction,
and neuropathological changes, such as the presence
of plaques of aggregated amyloid-? (A?) peptide,
amyloid angiopathy, and neurofibrillary tangles of
phosphorylated tau protein in the brain [2, 3]. At
least some of the loss of cognitive abilities has been
attributed to deterioration of the cholinergic system
induced by the toxic forms of A? [1, 4, 5]. The cur-
rently available treatments for AD include the use of
Journal of Alzheimer’s Disease 23 (2011) 1–19
DOI 10.3233/JAD-2011-102136
IOS Press
1
Cotinine Reduces Amyloid-? Aggregation
and Improves Memory in Alzheimer’s
Disease Mice
1
2
3
Valentina Echeverriaa,b,∗, Ross Zeitlina, Sarah Burgessa, Sagar Patela, Arghya Barmanc,
Garima Thakurc, Magorzota Mamcarzd, Li Wangd, David B. Sattellee, Daniel A. Kirschnerf,
Takashi Morig, Roger M. Leblancc, Rajeev Prabhakarcand Gary W. Arendashd
aBay Pines VA Healthcare System, Bay Pines, FL, USA
bDepartment of Molecular Medicine, University of South Florida, Tampa, FL, USA
cDepartment of Chemistry, University of Miami, FL, USA
dDepartment of Cell Biology, Microbiology & Molecular Biology, University of South Florida, Tampa, FL, USA
eFaculty of Life Sciences, University of Manchester, Manchester, UK
fDepartment of Biology, Boston College, Massachusetts, USA
gDepartments of Biomedical Sciences and Pathology, Saitama Medical Center and Saitama Medical
University, Kawagoe, Saitama, Japan
4
5
6
7
8
9
10
11
12
13
14
Accepted 23 January 2011
Abstract. Alzheimer’s disease (AD) affects millions of people world-wide and new effective and safe therapies are needed.
Cotinine,themainmetaboliteofnicotine,hasalonghalf-lifeanddoesnothavecardiovascularoraddictivesideeffectsinhumans.
We studied the effect of cotinine on amyloid-? (A?) aggregation as well as addressed its impact on working and reference
memories. Cotinine reduced A? deposition, improved working and reference memories, and inhibited A? oligomerization in
the brains of transgenic (Tg) 6799 AD mice. In vitro studies confirmed the inhibitory effect of cotinine on A?1-42aggregation.
Cotinine stimulated Akt signaling, including the inhibition of glycogen synthase kinase 3? (GSK3?), which promotes neuronal
survival and the synaptic plasticity processes underlying learning and memory in the hippocampus and cortex of wild type
and Tg mice. Simulation of the cotinine–A?1-42complex using molecular dynamics showed that cotinine may interact with
key histidine residues of A?1-42, altering its structure and inhibiting its aggregation. The good safety profile in humans and its
beneficial effects suggest that cotinine may be an excellent therapeutic candidate for the treatment of AD.
15
16
17
18
19
20
21
22
23
24
Keywords: Alzheimer’s disease, amyloid-?, cotinine, neurodegeneration, oligomerization
25
INTRODUCTION
26
Alzheimer’s disease (AD) is the main cause of
dementia in the elderly [1]. This devastating disease is
27
28
∗Correspondence to: Valentina Echeverria, 10,000 Bay Pines
Blvd. Bldg. 22 Rm. 123, Bay Pines, FL 33744, USA. Tel.:
+1 727 398 6661, ext. 4425; Fax: +1 727 319 1161; E-mail:
valentina.echeverria@va.gov.
29
30
31
32
33
34
35
36
37
ISSN 1387-2877/11/$27.50 © 2011 – IOS Press and the authors. All rights reserved

Page 2

Uncorrected Author Proof
of mice treated with cotinine. We investigated Akt
becauseitiscriticallyinvolvedinneuronalsurvivaland
mediates synaptic plasticity changes during learning
and memory processes [28]. For example, Akt targets
pro-apoptotic proteins, such as Bcl-2 and caspase-9,
protecting neurons from apoptosis [29] and inhibits
thepro-apoptotictaukinase,glycogensynthasekinase
3? (GSK3?) by phosphorylation [30, 31]. Finally, to
understand the effect of cotinine on A? plaque depo-
2
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
acetylcholinesterase inhibitors (i.e., donepezil, galan-
tamine,ortacrine)[6–9]andtheN-methyl-D-aspartate
(NMDA) antagonist, memantine. Unfortunately, these
drugsonlymarginallyamelioratethecognitivedeficits
andhavemostlyshort-termpositiveeffects[6,10–12].
Epidemiological studies have shown a negative
correlation between tobacco consumption and the
development of AD [13]. It has also been found post-
mortem that the levels of soluble and insoluble A?
peptides were significantly decreased in the brains
of smoking AD patients compared to non-smokers
with the disease [14]. The putative beneficial effect of
tobacco has been mainly attributed to nicotine, which
has been reported to improve cognitive abilities and
reduce plaques in a mouse model of AD [15]. Since
nicotinic receptors play an important role in attention,
learning, and memory, the positive effects of nico-
tine on memory were mostly credited to the activation
of these receptors. Unfortunately, due to its toxicity,
addictive properties, and side effects, nicotine has not
been considered an attractive therapeutic agent against
AD [16–18].
In mammals, more than 80% of nicotine is metabo-
lized into cotinine, a metabolite with a longer half-life
(nicotine, 2–3h; cotinine, 19–24h) and much lower
toxicity [19]. For these reasons, we speculate that the
effects attributed to nicotine in vivo may be at least
in part due to cotinine, which persists for longer in
the brain. However, cotinine is a very weak agonist of
nicotinic acetylcholine receptors (nAChRs) [20–22],
and thus the mechanism of its possible beneficial
effects on memory has been elusive and controversial
[23, 24].
Here, we examined the actions of cotinine on
plaque deposition and cognitive impairment using
the amyloid-? protein precursor (A?PP)/presenilin 1
(PS1) transgenic (Tg) 6799 mice, which express 5
familialAD(FAD)mutations[25].WeusetheTg6799
model because these mice show measurable whole
brain levels of A? and brain A? deposition by 2–4
months, and fast cognitive impairment by 4–8 months
of age [26, 27]. We have also explored Akt activa-
tionbyphosphorylationinthehippocampusandcortex
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
sition, we also investigated the effects of cotinine on
A? aggregation in vitro and model the interaction of
cotinine with A?1-42.
We found that cotinine reduced cerebral A? deposi-
tionandamelioratedcognitiveimpairmentinTgmice.
These beneficial actions may result from its capacity
toreduceA?aggregation.Computationalmodelingof
the cotinine–A?1-42complex suggests that cotinine’s
inhibitionofA?aggregationmayberelatedtoitsabil-
ity to bind to amino acid residues that participate in
the aggregation of the peptide. The implications of
these findings for developing a new therapy for AD
are discussed.
90
91
92
93
94
95
96
97
98
99
100
101
102
MATERIALS AND METHODS
103
Drugs
104
Cotinine was purchased from Sigma-Aldrich Cor-
poration (St. Louis, MO).
105
106
Mice
107
WeusedtheTg6799mice,whichexpressthehuman
A?PP and PS1 genes containing five FAD muta-
tions [25], including three FAD mutations in A?PP
(Swedishmutation:K670N,M671L;Floridamutation:
I716V; London mutation: V717I); and PS1 (M146L,
L286V) (The Jackson Laboratories, Bar Harbor, ME)
[26]. The Tg6799 lines were maintained as hemizy-
gotesonaB6/SJLhybridbackground.Malemicewere
usedasheterozygoteswithrespecttothetransgeneand
non-Tg (NT) wild type littermate mice served as con-
trols.Allbehavioralanalyseswereperformedbetween
5.5 and 6.5 months of age by investigators blind to the
genotypeandtreatmentofmice.Miceweremaintained
on a 12-h dark and 12-h light cycle with ad libitum
access to food and water. All protocols were previ-
ously approved by the Institutional Animal Care and
UseCommitteesoftheUniversityofSouthFloridaand
Bay Pines Veterans Affairs Healthcare System.
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
Cotinine treatment
126
At 2 months of age, Tg mice were started on daily
treatment with cotinine (2.5mg/kg) dissolved in PBS
or vehicle alone orally via gavage. Treatment was
administered for 3.5 months, as well as for the ensuing
one-month period of behavioral testing and subse-
quently for two weeks until euthanasia (total of 5
months of treatment).
127
128
129
130
131
132

Page 3

Uncorrected Author Proof
Radial arm water maze (working memory)
For the RAWM task of spatial working memory, an
aluminuminsertwasplacedintoa100cmcircularpool
to create 6 radially distributed swim arms emanating
from a central circular swim area [32]. An assortment
of 2-D and 3-D visual cues surrounded the pool. The
latency and number of errors prior to locating which
one of the 6 swim arms contained a submerged escape
platform (9cm diameter) was determined for 5 tri-
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
3
Behavioral testing
133
A one-month battery of sensorimotor, anxiety, and
cognitive-based tasks was performed beginning at 3.5
months into treatment (5.5 months of age) [32, 33]. In
the order of their performance, the tasks administered
were: sensorimotor (open field, balance beam, string
agility,andY-maze),anxiety(elevatedplusmaze),and
cognition-based [Morris water maze (MWM), circu-
lar platform, platform recognition, radial arm water
maze (RAWM)] [32], and cognitive interference [33].
Since there were no transgene or treatment effects on
any of the sensorimotor or anxiety tasks, and there
were no cognitive effects of treatment observed in
NT or Tg mice in MWM or platform recognition,
we only describe the three cognitive tasks that did
exhibit significant effects of treatment (circular plat-
form, RAWM, and cognitive interference) as detailed
below.
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
Circular platform (spatial reference
learning/memory)
A walled 69cm circular platform, with 16 equidis-
tantly spaced holes along its periphery, was encircled
by a black curtain [34]. Visual cues, located on the
blackcurtainandplatformwallscanbeusedbytheani-
maltofindtheoneholethroughwhichitcanescapethe
platform surface to avoid the aversive stimuli of bright
lights and fan wind. During a single 5-min maximum
daily trial, the total number of errors (head pokes into
non-escape holes) and latency to find the escape hole
were recorded. Although the escape hole remained
constantforanygivenanimaloverthe8daysoftesting,
it was relocated after each animal’s trial to control for
olfactory cues. To minimize interference from olfac-
tory cues, the maze was cleaned with a dilute vinegar
solution following each animal’s trial. Performance
for four 2-day blocks of testing was analyzed statis-
tically with Analysis of Variance (ANOVA), followed
by post hoc comparisons done with the Fisher’s least
significant difference test.
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
als/day over 8 days of testing, with statistical analysis
involving performance over all 8 days of RAWM per-
formance. There was a 30-min time delay between the
4th trial (T4; final acquisition trial) and 5th trial (T5;
memory retention trial). The platform location was
changed daily to a different arm, with different start
arms for each of the 5 trials semi-randomly selected
from the remaining 5 swim arms. During each trial
(60-smaximum),themousewasreturnedtothattrial’s
startarmuponswimmingintoanincorrectarmandthe
number of seconds required to locate the submerged
platform was recorded. If the mouse did not find the
platform within a 60-s trial, it was guided to the plat-
formforthe30-sstay.Thelatencyandnumberoferrors
during T4 and T5 are both considered indices of work-
ing memory and are temporally similar to the standard
registration/recall testing of specific items used clini-
cally in evaluating AD patients.
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
Cognitive interference task (proactive/retroactive
interference)
Wedesignedthistaskbasedonacognitivetaskused
to discriminate normal aged, mild cognitive impair-
ment, and AD patients from one another [35]. Our
interference task for mice involves two RAWM set-
ups in two different rooms, each with different sets of
visual cues [33]. The task requires animals to remem-
ber a set of visual cues, so that following interference
with a different set of cues, the initial set of cues
can be recalled to successfully solve the RAWM task.
A set of four behavioral measures was examined.
Behavioral measures were: A3 (the last of three recall
trials performed in RAWM “A”), “B” (proactive inter-
ference measure attained from a single trial in RAWM
“B”), A4 (retroactive interference measure attained
duringasingletrialinRAWM“A”),and“A5”(delayed-
recall measure attained from a single trial in RAWM
“A” following a 20-min delay between A4 and A5).
As a distraction between trials, animals are placed in a
Y-maze and allowed to explore for 60s between suc-
cessive trials of the three-trial recall task, as well as
duringtheproactiveinterferencetask.Aswiththestan-
dard RAWM task, this interference task involves the
platform location being changed daily to a different
armforbothoftheRAWMset-upsutilized,anddiffer-
ent start arms for each day of testing for both RAWM
set-ups. For A1 and B trials, the animal was initially
allowed one min to find the platform on their own
before they were guided to the platform. The actual
trial was then performed in each case. As with the
standard RAWM task, animals were given 60s to find
the escape platform for each trial, with the number
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231

Page 4

Uncorrected Author Proof
The membranes were blocked in TBS with 0.05%
Tween20(TBST;Bio-RadLaboratories,Inc.)contain-
ing 10% dry skim milk and incubated in TBST with
primary antibodies overnight at 4◦C and secondary
antibodies for 1h at RT. Rabbit monoclonal anti-
bodies directed against Akt phosphorylated at serine
473 (pAkt [Ser473]) (1:500) and total Akt (1:1,000)
were obtained from Cell Signaling Technology. A
rabbit polyclonal antibody directed against phospho-
4
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
of errors and escape latency recorded for each trial.
Given the very close correspondence between error
and latency scores in individual animals for both the
RAWM and cognitive interference tasks, only latency
scoresarepresentedinthisreport.Animalsweretested
for cognitive interference performance on two succes-
sive days, with statistical analysis performed on the
resultant 2-day block.
232
233
234
235
236
237
238
239
Brain tissue preparation
240
Following behavioral testing of seven-month-old
mice, animals were euthanized and perfused with a
cold physiological saline. The left frontal half of each
brain was placed in 4% paraformaldehyde in 0.1M
phosphate buffer (pH 7.4) overnight, wherein tissues
remained until the paraffin embedding process for A?
immunohistochemical analyses. The remaining left
portion of the brain, as well as the entire right side
of the brain, was dissected out into regions (e.g., hip-
pocampus and cortex), quickly frozen, and stored at
–80◦C for later neurochemical analyses.
241
242
243
244
245
246
247
248
249
250
251
Western blot analysis of tissue extracts
252
TissueswereanalyzedbyWesternblotaspreviously
described [33]. Briefly, brains were rapidly removed,
and tissues dissected and disrupted by sonication in
radioimmunoprecipitation assay (RIPA) buffer (Tris
50mM pH 7.4; NaCl 150mM; SDS 0.1%; NaDeoxy-
cholate 0.5%; Triton X-100 1%; Cell Signaling
Technology, Danvers, MA) with a complete protease
inhibitor cocktail (Roche Molecular Biochemicals,
Indianapolis, IN). Brain extracts were incubated on
ice for 30min and centrifuged at 20,000×g for
30min at 4◦C. Equal amounts of protein from
the supernatant were separated by sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-
PAGE)usingeither10–20%Tris-Tricinegel(Bio-Rad
Laboratories, Inc., Hercules, CA) for A?PP/A? anal-
yses or 4–20% Tris-Glycine gel (Thermo Fisher
Scientific Inc., Rockford, IL) for Akt and ?-tubulin
analyses and transferred to nitrocellulose membranes.
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
GSK3? (Serine 9) and total GSK3? were obtained
from Cell Signaling Technology. A mouse mono-
clonal antibody directed against A? (6E10) (1:5,000;
Covance, Emeriville, CA) that also recognizes A?PP
and several forms of the peptide was used to detect
A? oligomers. A monoclonal mouse antibody against
?-tubulin (1:10,000; Promega, Madison, WI) was
used to control protein sample loading and trans-
fer efficiency. The monoclonal antibody against total
A?PP (clone 22C11) (1:2,000; Millipore, Temecula,
CA) was used to confirm the immunoreactivity of
total A?PP. Membranes were washed with TBST and
incubated with LI-COR’s goat anti-mouse IRDye sec-
ondaryantibodies(LI-CORBiosciences,Lincoln,NE)
for 1h, washed with TBST and TBS, and images were
acquired using an Odyssey Infrared Imaging System
(LI-COR Biosciences) or the Kodak Image Station
440CF (Eastman Kodak Company, Rochester, NY)
usingaMolecularImagingSoftwareversion4.0(East-
man Kodak Company), and analyzed using an NIH
Image J software.
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
Analysis of cotinine levels in the brain
301
The levels of cotinine in brain tissues of mice were
estimated using ELISA. Briefly, after treatments and
behavioral testing, mice were euthanized, and brains
were removed, dissected, and disrupted by sonication
inRIPAbuffer(CellSignalingTechnology)containing
a complete protease inhibitor cocktail (Roche Molec-
ular Biochemicals). Cortical extracts were incubated
on ice for 30min and centrifuged at 20,000×g for
30min at 4◦C. The resultant supernatants were used
to determine cotinine levels according to the manu-
facturer’s instructions using a commercial ELISA kit
(Calbiochem,SanDiego,CA).Inthissolidphasecom-
petitive ELISA, the cotinine present in the samples
competes with a cotinine–enzyme conjugate for the
binding to a plate coated with anti-cotinine antibody.
Upon the addition of the substrate for the conjugated
enzyme (horseradish peroxidase), the concentration
of the cotinine in the samples is inversely propor-
tional to the intensity of the color developed in the
wells.
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
Analysis of Aβ levels
322
The levels of A?40 and A?42 were quantified in
brain tissues by ELISA. To determine soluble A? lev-
els, brain tissues were homogenized in RIPA buffer,
centrifuged at 20,000×g for 20min at 4◦C and the
supernatants were stored at –80◦C until use. To mea-
323
324
325
326
327

Page 5

Uncorrected Author Proof
and motor cortices) were captured from each animal,
and a threshold optical density was obtained that dis-
criminated staining from background. Each region of
interest was manually edited to eliminate artifacts. For
A? burden analysis, data are reported as percentage of
immunolabeled area captured (positive pixels) relative
to the full area captured (total pixels). Each analysis
was done by a single investigator blinded to sample
identities.
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
5
sure insoluble A? levels, brain tissues were prepared
bysonicationofthesamplesinasolutionof5Mguani-
dineHCl(Sigma-AldrichCorporation)(pH8.0).After
sonication, samples were incubated for 3h at RT and
centrifugedat20,000×gfor20minat4◦C.Thesuper-
natants were stored at –80◦C until use or immediately
diluted using PBS with 5% bovine serum albumin
(Sigma-Aldrich Corporation) and 0.03% Tween 20
supplemented with 1× protease inhibitor cocktail
(Roche). Then samples were analyzed for A? levels
usinganELISAkit(InvitrogenCorporation,Carlsbad,
CA), according to the manufacturer’s recommenda-
tions.
328
329
330
331
332
333
334
335
336
337
338
339
340
Aβ plaque analysis
341
For brain A? immunohistochemical staining and
analysis, we used our well-established protocol as
previously described [32]. Briefly, the frontal cortex
from the left side of the brain was dissected. At the
level of cingulate and motor cortex, three 5-?m sec-
tions (150?m apart) were made from each mouse
brain using a sliding microtome. Immunohistochem-
ical staining was performed using a Vectastain ABC
Elite kit (Vector Laboratories, Burlingame, CA) cou-
pled with the diaminobenzidine reaction, except that
the biotinylated secondary antibody step was omitted
for A? immunohistochemical staining. A biotinylated
human A? monoclonal antibody (clone 4G8; 1:200,
Covance Research Products, Emeryville, CA) was
incubated for 1h at RT. PBS (pH 7.4) or normal rab-
bit serum was used instead of primary antibody or
ABC reagent as a negative control. Brain sections
were treated with 70% formic acid prior to the pre-
blocking step. Quantitative image analysis utilized
previous methods with modifications [36, 37]. Images
were acquired using an Olympus BX60 microscope
with an attached digital camera system (DP-70, Olym-
pus, Tokyo, Japan), and the digital image was routed
into a Windows PC for quantitative analysis using a
SimplePCI software (Hamamatsu Photonics, Hama-
matsu, Shizuoka, Japan). Images of three sections
through both anatomic regions of interest (cingulate
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
Dot blot analysis of Aβ oligomers in brain tissues
of Tg6799 mice
378
379
To investigate the effect of cotinine on the for-
mation of soluble A? oligomers, we performed dot
blot analysis using the A11 antibody (1:1,000; Invit-
rogen Corporation), which is highly specific for the
oligomeric forms of A? and does not recognize the
monomeric or fibrillar forms of the peptide. Briefly,
hippocampal tissues were disaggregated using a pellet
pestle in RIPA buffer and centrifuged at 20,000×g
for 30min. Aliquots of hippocampal extract super-
natants (2?g protein) were applied onto nitrocellulose
membranes and allowed to dry, after which the mem-
branes were blocked for 1h at RT with LI-COR
blocking buffer, washed, and incubated with A11 anti-
bodyovernightat4ºC.Afterwashing,membraneswere
incubated with LI-COR’s IRDye secondary antibod-
ies for 1h and then washed. The immunoreactive dots
were then visualized using an Odyssey Infrared Imag-
ing System (LI-COR Biosciences) and analyzed with
an NIH Image J software.
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
Western blot analysis of Aβ oligomers
399
The Western blot analysis was performed as des-
cribed previously [38]. Briefly, hippocampal tissues
were disaggregated with a pellet pestle in RIPA buffer
and centrifuged at 20,000×g for 30min. Aliquots
of hippocampal extract supernatants (80?g protein)
were mixed with Tris-Tricine 3× sample buffer
(Bio-Rad Laboratories, Inc.), separated in a 10–20%
Tris-Tricine gel by electrophoresis, and blotted using
the 6E10 antibody (1:5,000). The membranes were
incubated with horseradish peroxidase-conjugated
secondaryantibodyfor1h,visualizedusingEnhanced
Chemiluminiscence (SuperSignal Femto Maximum
Sensitivity Substrate, Pharmacia Biotech, Piscataway,
NJ),scannedontheKodakImageStation440CF(East-
man Kodak Company) using a Molecular Imaging
Software version 4.0 (Eastman Kodak Company), and
analyzed with an NIH Image J software.
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
AFM analysis of the effect of cotinine on Aβ
aggregation
417
418
Preparation of Aβ solutions for atomic force
microscopy (AFM)
To investigate the effect of cotinine on A? fib-
rillation, we prepared the A?1-42 solutions using a
method that permits a fast development of fibrils.
419
420
421
422
423

Page 6

Uncorrected Author Proof
site of cotinine on A?1-42is not known, the binding
site of nicotine on A?1-42was used to develop starting
models for the molecular dynamics (MD) simula-
tions. Because S-cotinine was used in the experiments
reported in this study, we utilized this form of cotinine
in the simulations on the cotinine–A?1-42 complex.
The molecular modeling of the interaction between
cotinine and the A?1-42monomer was performed in
the following two steps.
6
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
The first step of this protocol consists in dissolving
A?1-42 in 1,1,1,3,3,3-hexafluoroisopropanol (HFIP;
Sigma-Aldrich Corporation) to obtain a starting solu-
tion containing only A?1-42monomers [39]. Briefly,
lyophilized A?1-42(American Peptide) was dissolved
in HFIP, evaporated, re-dissolved in dimethyl sulfox-
ide (DMSO) (Sigma-Aldrich Corporation), and then
diluted in PBS (pH 7.4) alone or plus cotinine. Peptide
solutions containing A?1-42(1mM) with and without
cotinine (2mM) were incubated for 10 days at RT and
the formation of oligomers and fibers was examined
by AFM.
424
425
426
427
428
429
430
431
432
433
434
435
AFM analysis
20-?l aliquots of A? solutions were deposited on
freshly cleaned and dried silicon wafers (approxi-
mately 1-mm thick). After waiting for 10min, non-
adsorbed portions of the samples were washed with
de-ionized water (2ml). The wet surface of the sili-
con wafer was then dried using a gentle flow of air.
The AFM analysis was performed using an AFM
apparatus (AFM, ?A multimode SPM, Model no.
920-006-101, Veeco, Plainview, NY) that permits the
acquisition of images using a tapping mode approach.
This approach allows intermittent contact of the tip
withthesampleandminimizesthechancesofdeforma-
tion of the peptide samples. The cantilever and the tip
were made of silicon and the cantilever force constant
was approximately 20–100N/m with the resonance
frequency between 200 and 400kHz. The scan rate
was 1.0Hz. The analysis of fibrils and oligomers was
performedusingtheNanoscopeControlsoftware(ver-
sion 5.30) (Veeco). The analysis of the height of the
aggregates was performed using Pico Image software
(Pico View version 1.6.4) from Agilent Technologies
(Mississauga, ON, Canada).
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
Molecular modeling
459
The chemical structure of cotinine is similar to that
of nicotine, and also possesses two different enan-
tiomers (i.e., S and R forms). As the exact binding
460
461
462
463
464
465
466
467
468
469
470
Molecular docking
Cotinine was docked near the His13 and His14
residues of the full-length A?1-42peptide using the
AutoDockprogram(version4.0.TheScrippsResearch
Institute, La Jolla) [40]. The most representative struc-
ture obtained from the previous 50-ns MD simulations
on full-length A?1-42 in aqueous solution was used
in the docking procedure [41]. The reported bind-
ing sites of nicotine on A?1-42were utilized in this
process [42–44]. The AutoDock program performed
the rapid energy evaluation through a pre-calculated
grid and found the suitable binding position of coti-
nine on A?1-42. Polar hydrogens were added using the
hydrogen module in the AutoDock tools for the pep-
tide and the Kollman united atom partial charges were
assigned. The grid was calculated using the Auto Grid
protocol. It was chosen to include all the His residues
(His6, His13, and His14) of A?1-42. The dimension
of the grid was set to 50×50×50˚A with a spac-
ing of 0.375˚A between the two consecutive grids. In
the docking process, A?1-42was kept rigid and coti-
nine was allowed to form all the possible torsional
bonds. The AutoDock Lamarckian genetic algorithm
using the standard protocol with 150 randomly placed
individual initial populations was applied. In total, 50
independent docking runs were performed. The low-
est energy conformer taken from the docked complex
was utilized to perform 50-ns MD simulations on the
cotinine–A?1-42complex in aqueous solution.
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
Molecular dynamics simulations
All MD simulations were performed using a GRO-
MACS software package (General Public License),
utilizing the GROMACS force field [45]. Dundee Pro
Drug Server was used for generating the topology of
the cotinine molecule for the MD simulation and the
partial charges were also calculated using this server
[46]. The cotinine–A?1-42complex was placed in the
center of a box with dimensions 4.9×4.2×4.6nm.
Theboxcontainedover2,852singlepointchargewater
molecules. Some water molecules were replaced by
sodium and chloride ions to neutralize the system and
to simulate an experimentally used ion concentration
of 150mM. The starting structure was subsequently
energy minimized with a steepest descent method for
2,000 steps. The results of these minimizations pro-
duced the initial structure for the MD simulations. The
MD simulations were then carried out with a constant
number of particles, pressure, and temperature. The
SETTLE algorithm was used to constrain the bond
length and angle of the water molecules [47], while
the LINCS algorithm was used to constrain the bond
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521

Page 7

Uncorrected Author Proof
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
7
length of the peptide [48]. The long range electro-
staticinteractionswerecalculatedbytheparticle-mesh
Ewald method [49, 50]. A constant pressure of 1 bar
wasappliedwithacouplingconstantof1.0ps;peptide,
water molecules, and ions were coupled separately to
a bath at 300K with a coupling constant of 0.1ps.
Theperiodicboundaryconditionswereappliedandthe
equationofmotionwasintegratedattime-stepsof2fs.
The secondary structure analyses were performed by
employingthedefinedsecondarystructuresofproteins
protocol [51]. The contact maps and similarity factor
of the most representative structures obtained from a
cluster analysis have also been employed as structural
descriptors. A contact for a pair of amino acid side
chains is considered to form when a minimal distance
between any pair of their atoms is less than 0.5nm.
In the cluster analysis, the trajectories were analyzed
by grouping structurally similar frames [root-mean-
square-deviation (RMSD) cutoff=0.30nm [52]], and
the frame with the largest number of neighbors is
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
denoted as a “middle” structure, which represents that
particular cluster.
542
543
RESULTS
544
Cotinine improves cognitive performance
in Tg6799 mice
545
546
In a one-month battery of cognitive tasks adminis-
tered between 5.5 and 6.5 months of age, there were
no transgene or cotinine treatment effects on sen-
sorimotor or anxiety function. There were also no
effects of cotinine on several basic cognitive tasks
(MWM and Platform recognition). However, signifi-
cant effects of cotinine were detected in three other
tasks (Fig. 1A–C). In the circular platform task of
spatial reference memory, Tg controls were severely
impaired in performance, as indicated by their much
higher number of errors and latencies during the final
547
548
549
550
551
552
553
554
555
556
557
Fig. 1. Cotinine prevented cognitive impairment in Tg6799 mice. Long-term cotinine treatment (2.5mg/kg) protected transgenic (Tg) mice
(n=9–11) against cognitive impairment in multiple tasks of cognitive function. A) In the circular platform test of reference learning/memory,
cotinine-treated Tg mice performed significantly better than Tg controls and no different from non-transgenic (NT) controls (n=9–11).
**p<0.025 or higher level of significance vs. cotinine-treated Tg mice and NT controls. B) Radial arm water maze testing for working memory
revealed that cotinine-treated Tg mice again were superior to Tg controls, although they did not perform to the level of NT controls (†p<0.02 vs.
Tg; *p<0.25 vs. NT). C) In the cognitive interference task, Tg controls were impaired on all four measures, whereas cotinine-treated Tg mice
were no different from NT controls on A3 (Recall) and B (proactive interference) and nearly better than Tg controls in these two measures. For
A4 (retroactive interference) and A5 (delayed recall), both groups of Tg mice were equally impaired. ANOVA was used for statistical analysis
of performance of four 2-day testing blocks, followed by post hoc comparisons done with the Fisher’s least significant difference test.

Page 8

Uncorrected Author Proof
We investigated the effect of cotinine on the acti-
vation of Akt by phosphorylation in the hippocampus
and cortex of Tg and NT mice. The results show that
cotinine stimulates Akt in both tissues. For vehicle-
treated Tg mice, the levels of the active form of Akt
(pAkt [Ser473]) in the hippocampus and cortex were
similar to the values found in the vehicle-treated NT
mice, considered as 100% immunoreactivity (Fig. 2A
and B). By contrast, the treatment of Tg mice with
8
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
block of testing (Fig. 1A). In sharp contrast, Tg mice
that had been chronically treated with cotinine exhib-
ited error and latency scores that were significantly
lower and no different from NT controls. Cotinine
treatment did not affect performance of NT mice in
this any task of the entire test battery.
In the RAWM task of working memory, a modest
beneficial effect of cotinine treatment was present in
TgmiceforworkingmemorytrialT4(Fig.1B).Across
all 8 days of testing, Tg6799 controls were signifi-
cantly impaired (versus NT groups) during trial T4,
while cotinine-treated Tg mice were significantly bet-
terthanTg6799controls.Nonetheless,cotinine-treated
Tg mice were not at the performance level of NT con-
trolsonT4.FordelayedworkingmemorytrialT5,both
Tg groups were impaired in performance.
As a more challenging variant of the RAWM task,
thecognitiveinterferencetaskevaluatesnotonlywork-
ing memory, but proactive and retroactive interference
as well. Over both days of cognitive interference test-
ing, Tg mice were impaired in all four measures
evaluated in comparison to the excellent performance
of NT controls. By contrast, Tg mice chronically-
treated with cotinine were no different from NT
controls for both the final recall trial (A3) and the
proactive interference trial (B) (Fig. 1C). As well, the
performanceofcotinine-treatedTgmiceonthesetrials
trended to be significantly different from Tg controls
at p=0.08 and 0.13, respectively. For the remaining
twomeasuresofcognitiveinterference(i.e.,retroactive
interferenceanddelayedrecall),cotininetreatmentdid
not provide any cognitive benefit, with performance of
treated Tg6799 mice being impaired and no different
from control Tg mice.
Thus, long-term cotinine treatment to Tg mice
provided complete protection from spatial reference
memory impairment in the circular platform task,
while also improving the performance of Tg mice to
the level of NT controls in two working memory and
cognitive interference tasks.
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
Effect of cotinine on Akt phosphorylation
598
599
600
601
602
603
604
605
606
607
cotinine (2.5mg/kg) significantly increased the levels
of pAkt [Ser473] in the hippocampus and cortex by
+51%(p=0.023)and+189%(p=0.019),respectively,
as normalized to total Akt levels (Fig. 2A and B). For
NTmice,cotininealsoinducedasignificantincreasein
thelevelsofpAkt[Ser473]inthehippocampus(+74%,
p=0.023) and cortex (+205%, p=0.033) (Fig. 2C and
D). When normalized against tubulin, no significant
changes in the levels of total Akt were observed in the
cotinine-treated Tg mice when compared to untreated
Tg mice (Fig. 2A and B). However, we observed a
highly significant increase in total levels of Akt as nor-
malizedagainsttubulinintheTgmicewhencompared
to NT mice in the hippocampus (+49%, p=0.0043)
(Fig.2A).Thisincreasewasnotobservedinthefrontal
cortex of the same mice (Fig. 2B). We did not find
changes in total Akt in cotinine-treated NT mice when
compared to untreated NT mice in hippocampus and
cortex (Fig. 2C and D).
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
Effect of cotinine on the inhibition of GSK3β
by phosphorylation
627
628
Over-activation of GSK3? has been associated with
tau hyperphosphorylation and neuronal cell death in
AD brains [53]. It is well-known that Akt inactivates
GSK3? by phosphorylation at Ser 9. Thus, based on
our previous finding that cotinine activated Akt, we
investigated the effect of cotinine on GSK3? phos-
phorylation in the brains of Tg6799 mice. First, we
found no significant changes in the levels of pGSK3?
[Ser9] normalized to total GSK3? in both the hip-
pocampus and cortex of control Tg mice compared to
control NT mice (Fig. 3A and B). However, cotinine
treatment resulted in a significant increase in the lev-
els of pGSK3? [Ser9] in the hippocampus of Tg mice
(+42%, p=0.035) with respect to vehicle-treated Tg
mice(Fig.3A).Also,asignificantincreaseinthelevels
of pGSK3? [Ser9] was observed in the hippocampus
of cotinine-treated NT mice compared to NT controls
(+56%,p=0.048)(Fig.3C).Similarly,inthecotinine-
treated Tg mice we observed a trend of increase in the
levels of pGSK3? [Ser9] in the frontal cortex when
compared to vehicle-treated Tg mice, but the differ-
ences did not reach significance (+54%, p=0.059)
(Fig. 3B). However, cotinine treatment significantly
increased the levels of pGSK3? [Ser9] in the cortex
of NT mice compared to untreated otherwise identi-
calmice(+295%,p=0.0056)(Fig.3D).Nosignificant
differences in the levels of total GSK3? were found in
the hippocampus or cortex among groups of NT mice
(data not shown).
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656

Page 9

Uncorrected Author Proof
To investigate the effect of cotinine on A? levels,
we analyzed the levels of A? in the RIPA-soluble
and insoluble fractions of cortex and hippocampus
tissues of Tg mice treated with cotinine or vehicle
for 5 months. It has been shown that in AD Tg
mice, the Swedish mutation increases the synthesis of
total A?, whereas A?PP London, Florida, and PS1
mutations promote the synthesis of A?42 [25]. In
coherence with this previous evidence, we found that
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
9
Fig. 2. Cotinine increased the active form of Akt in the brains of Tg6799 and non-transgenic (NT) control littermate mice. Tg and NT control
littermate mice were treated with saline or cotinine (2.5mg/kg) for 5 months and RIPA-soluble hippocampal and cortical protein extracts were
analyzed by Western blot using antibodies against ?-tubulin as well as total and phosphorylated Akt. The plots represent the immunoreactivity
values expressed as percentage of control of saline-treated NT mice. Levels of normalized phospho-Akt [Ser473] immunoreactivity against total
Aktlevelsinhippocampal(A,C)andcorticalextracts(B,D)areshown.TherewasaclearactivationoftheAktinthehippocampusandcortexof
both Tg (hippocampus, n=7–8; cortex, n=7–9) and NT (hippocampus, n=8–9; cortex, n=4–6) mice. Dividing lines separate immunoreactive
bands from different parts of the same membrane. The data are expressed as the mean±SEM. Student’s t-test was used to compare the mean
of the values between groups. ns, non-significant change; *p<0.05; **p<0.01. pAkt, phospho-Akt [Ser473].
Effect of cotinine on Aβ levels in the brains
of Tg6799 mice
657
658
659
660
661
662
663
664
665
666
667
the levels of A?42were substantially higher than the
levels of A?40in both the hippocampus and cortex of
7-month-old Tg6799 mice (Table 1). The insoluble
A?42levels were significantly decreased in the cortex
(but not the hippocampus) of cotinine-treated Tg mice
relative to control Tg mice (–26%, p=0.031). How-
ever,forbothbrainareas,nodifferenceswereobserved
for soluble levels of A?42and A?40in the cotinine-
treated Tg mice when compared with untreated Tg
mice. Although no differences in the levels of insol-
uble A?40were observed in the cortex, they showed
a significant increase (+31%, p=0.019) in the hip-
668
669
670
671
672
673
674
675
676
677
678
679

Page 10

Uncorrected Author Proof
Soluble (pg/mg)
A?42
A?40
A?42/A?40
10
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
Fig. 3. Cotinine inhibits GSK3? by phosphorylation at serine 9 in the brains of Tg6799 and non-transgenic (NT) control littermate mice. Mice
were treated with saline or cotinine (2.5mg/kg) for 5 months and RIPA-soluble hippocampal and cortical extracts were analyzed by Western
blot using antibodies against ?-tubulin as well as total and phosphorylated GSK3?. The plots represent the levels of immunoreactivity for
phospho-GSK3? [Ser9] expressed as percentage of the saline-treated NT mice and normalized to total GSK3? levels. Levels of phospho-
GSK3? [Ser9], total GSK3?, and ?-tubulin in the hippocampus (A, C) and cortex (B, D) of Tg and NT mice are shown. Cotinine increased the
levels of phospho-GSK3? [Ser9] in the hippocampus of both Tg (n=8–10) and NT mice (n=8–10) (A, C), as well as in the cortex of NT mice
(n=4–6) (D). After the cotinine treatment, the levels of phospho-GSK3? [Ser9] also showed a trend of increase (p=0.059) in the cortex of Tg
mice (n=8–9) (B). Dividing lines separate immunoreactive bands from different parts of the same membrane. The data are expressed as the
mean±SEM. Student’s t-test was used to compare the mean of the values between groups. *p<0.05; pGSK3?, phospho-GSK3? [Ser9].
Table 1
ELISA analysis of the levels of RIPA-soluble and insoluble A? in the cortex and hippocampus of Tg6799 mice
Insoluble (ng/mg)
A?40
A?42
A?42/A?40
Cortex
Vehicle control
Cotinine
Hippocampus
Vehicle control
Cotinine
The data are expressed as the mean±SEM from 8–10 mice for the insoluble levels and 5–10 for soluble levels of
theA?peptides.Student’st-testwasusedtocomparethemeanofthevaluesbetweengroups.*p<0.05;**p<0.01.
249±54
222±46
6±1
7±1
41±5
31±5
511±49
364±40*
84±13
82±14
6.5±0.4
4.9±0.3**
5829±988
6942±718
35±8
37±5
204±57
194±18
178±22
202±29
34±3
54±6*
5.3±0.4
3.8±0.4*

Page 11

Uncorrected Author Proof
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
11
pocampus of cotinine-treated Tg mice relative to Tg
controls.TheinsolubleratioofA?42/40wasdecreased
by cotinine treatment in the cortex (–25%, p=0.006)
and hippocampus (–28%, p=0.030). No significant
changes were observed in the soluble ratio of A?42/40
in these same brain regions.
680
681
682
683
684
685
Effect of cotinine on amyloid burden in the brains
of Tg6799 mice
686
687
As expected from previous reports [54, 55], coti-
ninecrossedtheblood-brainbarrierasevidencedusing
ELISA analysis of cotinine levels in the brain tissues.
On average, the levels of cotinine in the RIPA-soluble
fraction of the cortex of the NT and Tg mice fluctuated
between 500 and 700ng/mg protein.
To investigate whether the cotinine treatment
reducesamyloidplaquedeposition,Tgmicewereeval-
uated for forebrain A? burden at 7 months of age
(following completion of behavioral testing). Tg6799
mice present A? deposition detectable as early as 2
months of age, which steadily increases through 8
months of age [25]. As shown in Fig. 4, control Tg
mice had robust amyloid burdens of around 3% in
both cingulate and motor cortices. Compared to these
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
Tg controls, Tg mice that had been given 5 months
of cotinine treatment exhibited a significant decrease
inamyloidburdenwithinbothcingulatecortex(–26%,
p=0.034)andmotorcortex(–17%,p=0.048)(Fig.4A
and B, respectively).
703
704
705
706
707
Effect of cotinine on the levels of Aβ oligomers
in the hippocampus and cortex of Tg6799 mice
708
709
Growing evidence supports the view that A?
oligomers are the main neurotoxic form of the pep-
tide [56, 57]. Thus, we investigated whether cotinine
affected the levels of A? oligomers in the brains of the
Tg mice, using the highly specific anti-oligomeric A?
antibodyA11,whichdoesnotrecognizeA?monomers
or A?PP. The results indicate that cotinine induced
an 18–20% reduction in the immunoreactivity for
A11 in the hippocampus (Fig. 5A and B) and cortex
(data not shown) of Tg mice. No significant differ-
ences in 6E10 immunoreactivity were found between
cotinine-treated Tg mice and Tg controls in both the
hippocampus and cortex (data not shown).
Furthermore, Western blot analyses of the RIPA-
soluble fractions from Tg mice untreated or treated
with cotinine (2.5mg/kg) showed a reduction in the
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
A
B
Fig. 4. Cotinine reduces amyloid burden in the cingulate and motor cortex of Tg6799 mice. Tg mice treated with cotinine showed less amyloid
burden in the cortex compared to those without cotinine treatment. The photomicrographs are representative views of the amyloid plaques in
the cingulate (A) and motor cortex (B) of brain sections of untreated and cotinine-treated mice (n=9–11) stained with the A?-specific antibody
4G8. Scale bar=50?m. The graphs at the bottom represent the percentage of amyloid burden in these brain regions. The data are expressed as
the mean± SEM. The Mann-Whitney rank sum test or the Student’s t-test was used to compare the effects of cotinine treatment on amyloid
burden quantified in both brain areas. *p<0.05.

Page 12

Uncorrected Author Proof
evaporation, in PBS alone or containing cotinine, then
incubated at 37ºC for 10 days.
Figure 6 shows representative AFM images (900×
900nm field) of aggregated forms of A? formed in
eithertheabsence(Fig.6AandB)orpresence(Fig.6C)
of cotinine. The analysis of A? fibril formation at day
0 showed no presence of aggregated A? in the ini-
tial solutions (Fig. 6A). After 10 days of incubation
at 37◦C, the height of the A?42aggregates used for
12
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
Fig. 5. Effect of Cotinine on A? oligomers in the hippocampus of Tg6799 mice. The levels of A? oligomers were analyzed in RIPA-soluble
fractions of the hippocampus of vehicle-treated (n=7) and cotinine-treated (n=6) Tg mice by dot blot and Western blot analysis. Representative
images (A) and the graph (B) of the dot blot analysis are shown depicting the relative intensities of anti-oligomeric A? (A11) immunoreactivity.
Western blot image (C) of the A? oligomers and total A?PP detected in hippocampal extracts of Tg and NT mice are shown, as analyzed using
the antibodies 6E10 and 22C11, respectively. The densitometric analysis (D) showed that cotinine treatment significantly decreased the levels of
high molecular weight 20-mer A? oligomers in the hippocampus of the Tg mice (n=4–5). No significant changes were observed in ?-tubulin or
A?PP levels. The data are expressed as the mean±SEM. Student’s t-test was used to compare the mean of the values between groups. *p<0.05.
levels of soluble A? oligomers (Fig. 5C and D). We
found a significant reduction in the levels of high
molecular weight oligomers (i.e., ∼20-mers) (–53%,
p=0.018) (Fig. 5C and D). No changes in the levels
oflowmolecularweightoligomers(<50kDa)(Fig.5C
and D) or the levels of A?PP induced by cotinine were
observedinthehippocampusofTgmiceasdetermined
with the antibodies directed against A? (6E10) and
total A?PP (22C11), respectively (Fig. 5C and D).
726
727
728
729
730
731
732
733
734
Effect of cotinine on Aβ aggregation
735
To investigate whether cotinine affects fibril for-
mation, we used AFM to analyze A? aggregation in
the presence and absence of cotinine. The solutions
were prepared by dissolving A? in HFIP and after
736
737
738
739
740
741
742
743
744
745
746
747
748
the AFM studies fluctuated approximately between 1
and 15nm (data not shown), indicating that the solu-
tions contained a mixture of oligomers, protofibrils
and fibrils. The pre-incubation of the peptide in the
presence of cotinine significantly reduced the length
of A? fibrils. The average length of A? fibrils incu-
bated in the presence of cotinine was significantly
lower (239±34nm) than the average fibril length in
theabsenceofcotinine(486±157nm,Student’sttest,
p=0.026) (Fig. 6D).
749
750
751
752
753
754
755
756
757
758
Molecular modeling of cotinine–Aβ1-42
interaction
759
760
A previous analysis of the interaction of A? with
nicotine by NMR preceded and facilitated the molecu-
lar modeling of cotinine [42, 43]. These NMR studies
suggested that nicotine binds to the segment of A?
betweenaminoacids1–28whenfoldedinanα-helical
conformation. According to the proposed mecha-
nism,nicotineinhibitstheconformationalchangefrom
α-helix to the amyloidogenic ?-sheet conformation
[42]. Here, we used molecular modeling to analyze
the interaction of cotinine with A?1-42and define a
putative mechanism that could explain the effect of
761
762
763
764
765
766
767
768
769
770
771

Page 13

Uncorrected Author Proof
forthefirst25nsthisregionexistsinhelicalconforma-
tionbutitislaterconvertedintothestableturnstructure
(shown by the yellow color in Fig. 7D, bottom). In
the free peptide, the loop region (24–28, VGSNK) is
quite unstable and undergoes a large dynamical trans-
formation between bend and turn. In the presence of
cotinine, in a marked difference, initially (for the first
22ns) this segment exists in the helical form but later
it adopts stable bend and coil conformations. The sec-
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
13
Fig. 6. Atomic force microscopy (AFM) analysis of the effect of cotinine on A?1-42fibrillation. A 900nm field of AFM analysis of A?1-42
peptide solutions (1mM) incubated for 10 days at 37◦C in the absence (A and B) or presence of cotinine 2mM (C). A. represents the analysis
of fibrils at time zero. At 10 days of incubation (B) some long A?1-42fibrils, few short linear protofibrils, and a heterogeneous mix of large
globularaggregateswereobservedwhenthesolutionswereincubatedinabsenceofcotinine.Inthepresenceofcotinine(C)therewasanevident
decrease in the size and length of the peptide aggregates. The difference in length of the A?1-42fibrils under these conditions were significant
with p=0.026 (Student’s t-test) (D). The differences in length of A? fibrils expressed as mean values±SEM were considered significant:
*p<0.05.
cotinine on A? aggregation. The RMSD of the MD
simulation confirmed that the complex is thermody-
namically equilibrated only after 30ns (Fig. 7A). The
most representative structure derived from the simula-
tion indicates that cotinine interacts with His6, Tyr10,
and His14 residues of the A?1-42peptide (Fig. 7B).
As shown in the figure, the pyridine ring of cotinine is
positioned between the imidazole ring of His6 and the
phenyl ring of Tyr10. It interacts with these residues
through strong π–π interactions that are indicated by
thedistancesof4.3˚Aand4.1˚Abetweencotinine–His6
andcotinine–Tyr10aromaticrings,respectively.Inthe
equilibrated region, the distance between the center of
thearomaticringofTyr10andthepyridineringofcoti-
nineremainsaround4.0˚A(Fig.7C).Ontheotherhand,
cotinine interacts with His14 via C–H–π interaction
[(cotinine–C)–H–His14=3.2˚A]. As discussed below,
theinteractionsofcotininewithHis6,Tyr10,andHis14
residues of A?1-42introduce significant changes in the
secondary structure of the peptide (Fig. 7D).
In the free A?1-42 peptide, in the first 38ns,
the Phe20–Val24 region is dominated by bend-and-
turn conformations with sporadic helical structures
(Fig.7D,top).Itisthentransformedintobendandcoil
structures. However, in the cotinine-bound structure,
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
ond hydrophobic domain (29–35, GAIIGLM) in free
A?1-42is dominated by the bend structure for the first
28ns but after that it is converted into a turn with a
partial ?-sheet character. On the other hand, in the
cotinine-boundstructure,theGly29–Ile32fragmentof
this region stays in the stable helical form throughout
thesimulation.TheremainingLeu33–Met35segment,
after 28 ns, is transformed into the stable bend confor-
mation.
806
807
808
809
810
811
812
813
814
DISCUSSION
815
The progressive deterioration of working memory
is one of the main characteristics of AD, and effective
therapies targeting memory loss in AD have been elu-
sive. Here, we investigated the actions of cotinine as
an anti-A? aggregation and memory-enhancing agent.
More specifically, we studied the effect of cotinine on
the deterioration of cognitive abilities, plaque forma-
tion,A?aggregation,andtheactivationofAkt/GSK3?
pathway in the brains of the Tg6799 mice. We found
that chronic treatment with cotinine improved work-
ing and reference memories and reduced both A?1-42
oligomerization and plaque burden in Tg6799 mice.
Furthermore, we found that cotinine stimulated the
activationofAktandtheinhibitionofGSK3?byphos-
phorylation in the hippocampus and cortex of both
Tg and NT littermate mice. We also discovered that
cotinine inhibited A?1-42aggregation into oligomers
and fibrils in vitro. Furthermore, using MD we mod-
eled the cotinine–A?1-42interaction and elucidated a
mechanism by which cotinine may interfere with A?
aggregation.
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836

Page 14

Uncorrected Author Proof
found that long-term treatment of Tg6799 mice with
cotinine beginning in young adulthood (2 months of
age)protectedtheircognitiveabilitiesinmultipletasks
and cognitive domains when tested between 5.5 and
6.5 months of age. It has been reported that at the
age we started cotinine treatment, senile plaques in
Tg6799 mice are beginning to appear with a large
increase occurring thereafter. Thus, it is likely that
our mice have no plaques or were at early stages of
14
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
Fig. 7. Cotinine interacts with key residues of A?1-42involved in its aggregation. Root-mean-square-deviations (RMSD) plotted against time
for the molecular dynamics (MD) trajectory of cotinine–A?1-42complex (A). The most representative structure derived from the 50ns MD
simulation on cotinine–A?1-42complex with specific interactions of cotinine (shown in pink) at the A?1-42binding site (B1 and B2). Distance
between the center of Tyr10 and pyridine nucleus of cotinine with the numbers representing the distances in ˚A (C). Secondary structural
assignment per residue as a function of time: For the free A?1-42monomer (D, top) and the cotinine–A?1-42complex (D, bottom).
In the search for new therapeutic agents against
AD,cotinine,themainmetaboliteofnicotine,attracted
our attention due to its unique pharmacological char-
acteristics and safety profile. Specifically, cotinine
is a poor agonist of the nAChRs [20–22], crosses
the blood-brain barrier [54], is much less prone to
induce respiratory arrest than nicotine, and has min-
imal toxic side-effects in humans [23, 58, 59]. We
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
plaque formation when the treatment started. Regard-
ing behavioral characterization of the Tg6799 mice,
this group reported initial cognitive impairment (in
Y-maze and Morris maze acquisition/retention) to
occur at 4–6 months of age [25, 26], with impairment
intracefearconditioningpresentby5–6monthsofage
[26, 27].
In the present study, we found several cognitive
domains that were impaired in Tg6799 mice between
5.5and6.5monthsofage.Inthetaskstestingreference
(circular platform) and working learning and mem-
ory (RAWM, cognitive interference), treatment with
cotinine administered from early adulthood provided
significant protection against otherwise certain cogni-
tive impairment. We did not observe changes in the
Morris maze likely due to the lower sensitivity of this
test to detect cognitive changes in AD mice. The tasks
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870

Page 15

Uncorrected Author Proof
result of a reduction in A? aggregation, although an
enhancement of A? clearance by cotinine as an addi-
tional mechanism cannot be ruled out at the present
time.
More importantly, we observed that cotinine treat-
ment induced a clear decrease in the ratio A?42/A?40
inthehippocampusandcortexoftransgenicmice.The
decrease in the A?42/A?40ratio can explain at least in
part its beneficial actions, as this ratio has been shown
V. Echeverria et al. / Cotinine Reduces Aβ-Related Memory Impairment
15
wherein cotinine did enhance performance are more
sensitive tasks (i.e., radial arm water maze, cognitive
interference task) and/or analyze different cognitive
abilities (circular platform). Indeed, we have previ-
ously reported that on AD therapeutics some drugs
have effects in these tasks without being effective in
the Morris water maze [60].
Overall, we hypothesize that cotinine may be useful
in preventing cognitive deterioration when adminis-
tered to individuals not yet exhibiting AD cognitive
impairment or those with mild cognitive impairment
at early stages of the disease.
Consistentwithapositiveeffectofcotinineonbrain
homeostasis and function, we have found that cotinine
treatment stimulated Akt in the hippocampus and cor-
texofbothTgandNTmice.Themultipleeffectsofthe
activation of Akt in the brain, including the enhance-
ment of brain plasticity and neuronal survival as well
as the inhibition of GSK3?, may be critical to mediate
the positive effect of cotinine on memory [61].
GSK3?isaproline-directedserine/threoninekinase
considered to be a key protein in both sporadic and
genetic forms of AD. According to this view, the
over-activation of GSK3? leads to an increase in A?
production, tau hyperphosphorylation, Neuroinflam-
mation, and consequently memory impairment [62].
Hyperphosphorylated tau is the main component
of paired-helical filaments forming the neurofibrillary
tangles (NFT), one of the main pathological hallmarks
in AD brains. GSK3? phosphorylates tau at several
sites [63] and its levels positively correlate with the
presence of NFT in AD brains [64]. Furthermore, the
over-expressionofGSK3?inforebrainregionsisasso-
ciated with neurodegeneration [53]. Interestingly, the
activation of the Akt/GSK3 pathway is also induced
by other drugs currently used for the treatment of AD
such as physostigmine and memantine [30].
Wealsoinvestigatedtheeffectoflong-termcotinine
treatmentonplaquedepositioninTg6799mice.Treat-
ment with cotinine significantly reduced the extent of
A? deposition into plaques in the cingulate and motor
cortices and the levels of insoluble A?42in the cortex.
This decrease in A? deposition was most likely the
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
to be important for the development of AD, by deter-
mining the fibrillogenesis and toxicity of A? [65, 66].
For example, it has been reported that A?40affects the
A?42fibrillation, decreasing its conversion to mature
fibrils [65]. In another study, it was found that the
age-of-onset of the pathology in individuals carrying
PS1-linked FAD mutations inversely correlated with
A?42/A?40and absolute levels of A?42, but directly
with A?40levels [67]. Furthermore, clinical studies
havedemonstratedthatthisratiocanalsodeterminethe
distribution of A? (i.e., parenchymal or vascular A?
deposition) in AD brains [60, 68]. Coherent with these
previousfindings,inconjunctionwithadecreaseinthe
A?42/A?40ratio we observed an increase in insoluble
A?40in the hippocampus of cotinine-treated Tg mice,
which can have a further beneficial effect in inhibiting
A?42fibrillation.
Cotinine did not change the total levels of soluble
A? peptides. However, it decreased the levels of A?
oligomers in the brains of the transgenic mice. Two
otherpotentialtherapeutics,whichsuppressA?aggre-
gation and also provide clear cognitive benefit to AD
Tgmice,suchasmelatoninandelectromagneticfields,
reduced deposited/insoluble A? without affecting the
soluble A? [69]. These studies, along with our present
findings, suggest that a small reduction in oligomers
without changes in total A? levels may be enough to
attain cognitive benefits in affected individuals.
The soluble aggregated forms of A?, including
oligomers and protofibrils, have been proposed as the
main pathological species in AD brains, as their accu-
mulation is sufficient to induce synaptic and cognitive
deficits in vivo [2, 57, 70–74]. The aggregation of A?
isrequiredforitstoxicity;infact,aphysiologicalfunc-
tionforthemonomericformofthepeptidestimulating
synaptic plasticity has been suggested [39, 75]. Thus,
compared to other therapeutic approaches, a distinct
advantage of anti-A? aggregation agents like cotinine
isthattheycantargetthetoxicformsofA?withoutdis-
rupting a possible normal function of the monomeric
form of the peptides.
To investigate whether the cotinine-induced reduc-
tion of A? aggregates in the brain was at least in part
due to an inhibition of the peptide aggregation, we
examined the effect of cotinine on A? aggregation
in vitro. It was previously shown using NMR and cir-
cular dichroism techniques in vitro that cotinine binds
to A? with high affinity, inhibiting its aggregation into
fibrils [42, 75, 76]. Using X-ray fiber diffraction, we
have previously found evidence suggesting that coti-
nine may reduce the extent of hydrogen-bonding and
fiber growth for the fibrillogenic peptide A?12-28[77].
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974