Báo cáo khoa học: Modulation of amyloid-b aggregation and toxicity by inosose stereoisomers pot

Modulation of amyloid-b aggregation and toxicity
by inosose stereoisomers
Mark Nitz
1
, Daniela Fenili
2,3
, Audrey A. Darabie
2
, Ling Wu
2
, Julian E. Cousins
2
and JoAnne McLaurin
2,3
1 Department of Chemistry, University of Toronto, Canada
2 Centre for Research in Neurodegenerative Diseases, University of Toronto, Canada
3 Department of Laboratory Medicine and Pathobiology, University of Toronto, Canada
Inositol is a simple polyol with eight naturally occur-
ring stereoisomers, the most common of which are
myo-inositol, chiro-inositol, epi-inositol, and scyllo-ino-
sitol [1]. myo-Inositol is the most abundant isomer and
is a ubiquitous component of all eukaryotic cells. We
have shown that scyllo-inositol may represent a poten-
tial therapeutic agent for amyloid disorders such as
Alzheimer’s disease (AD) [2,3]. We showed that myo-
inositol complexes with amyloid-b (Ab)42 in vitro to
form a small, stable micelle [4]. The ability of inositol
stereoisomers to interact with and stabilize small Ab
complexes was subsequently addressed [5]. CD spec-
troscopy demonstrated that epi-inositol and scyllo-
inositol, but not chiro-inositol, were able to induce a

structural transition from a random to a b-structure in
Ab42. Furthermore, electron microscopy demonstrated
that scyllo-inositol stabilizes small aggregates of Ab42
that are nontoxic to nerve growth factor (NGF)-differ-
entiated PC-12 cells and primary human neuronal
cultures.
We then examined the in vivo effects of the inositol
stereoisomers. myo-Inositol, epi-inositol and scyllo-ino-
sitol were administered to the TgCRND8 mouse model
of AD, which demonstrates age-associated cognitive
deficits and AD-like pathology [2,6]. myo-Inositol was
effective in vitro but was ineffective in vivo [2,5]. epi-
Inositol had some initial positive results as a prophy-
lactic treatment, but these positive effects were not
sustained with disease advancement [2]. scyllo-Inositol
treatment, however, resulted in a significant improve-
ment in cognitive function, synaptic function, and life-
span. Significant decreases in Ab40 and Ab42 levels,
vascular amyloid levels, plaque size and area were also
observed [2]. Therefore, scyllo-inositol is effective at
Keywords
aggregation; Alzheimer’s disease; amyloid;
fibrillogenesis; inosose
Correspondence
J. McLaurin, Centre for Research in
Neurodegenerative Diseases, University of
Toronto, 6 Queen’s Park Crescent West,
Toronto, ON, Canada M5S 3H2
Fax: +1 416 978 1878
Tel: +1 416 978 1035

E-mail:
(Received 21 November 2007, revised 28
January 2008, accepted 5 February 2008)
doi:10.1111/j.1742-4658.2008.06321.x
Amyloid-b (Ab) aggregation and amyloid formation are key pathological
features of Alzheimer’s disease, and are considered to be two of the major
contributing factors to neurodegeneration and dementia. Identification of
small molecule inhibitors that are orally available, have low toxicity and
high central nervous system bioavailability is one approach to the potential
development of a disease-modifying treatment for Alzheimer’s disease. We
have previously identified inositol stereoisomers as exhibiting stereospecific
inhibition of A b aggregation and toxicity in vitro and in vivo. We report
here the effects of inosose versus inositol stereoisomers on Ab fibrillogene-
sis as determined using CD and fluorescence spectroscopy and negative-
stain electron microscopy. The inososes differ from inositols by the
oxidation of one of the hydroxyl groups to a ketone. These molecules help
in the further elucidation of the structure–activity relationships of inositol–
Ab interactions and identify both allo-inositol and epi-2-inosose as in vitro
inhibitors of Ab aggregation.
Abbreviations
AD, Alzheimer’s disease; Ab, amyloid-b; HMIT, hydrogen myo-inositol transporter; NGF, nerve growth factor; SMIT, sodium myo-inositol
transporter; TFE, trifluoroethanol.
FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS 1663
decreasing the pathological and behavioral correlates
of AD in the TgCRND8 model [2]. Our in vivo and
in vitro results were recently confirmed using stable,
soluble Ab oligomers [7].
In vitro and in vivo data demonstrate that inositols
can inhibit the toxic effects of amyloid peptides
through the formation of nontoxic species. The effec-

tiveness of a specific inositol is highly dependent on
the stereochemistry of the hydroxyl groups. scyllo-Ino-
sitol is the most potent inositol, and a single epimeriza-
tion to form myo-inositol renders the compound less
active in vivo [1]. This dependence on subtle changes in
the orientation of hydroxyl groups suggests that a spe-
cific interaction with the amyloid peptides is necessary.
These results further suggest that the hydroxyl groups
may represent key recognition elements in the inter-
action with Ab. Furthermore, the dependence on the
hydroxyl group positioning also suggests that hydro-
gen bonding is important for stabilizing this interac-
tion. Therefore, we hypothesized that investigating the
interactions of structurally related, naturally occurring
inosose compounds with Ab may provide insights
into the importance of hydrogen bonding and the
position of the hydroxyl groups for the formation of
an Ab–inositol complex. Furthermore, small molecule
inhibitors of aggregation may act on Ab oligomer for-
mation and ⁄ or fibril formation; as these two assembly
pathways are independent and distinct [8], alterations
in the structure of the small molecule may then affect
which pathway is targeted.
Results
Structural characteristics of inosose–Ab
complexes
To investigate the interaction of inosose compounds
with Ab, we examined the effects of these compounds
on the random coil to b-sheet transition that is neces-
sary for Ab-nucleation, oligomer formation and fiber

formation. Examination of transitions in b-structure
over time gives an indication of the effect of the com-
pounds on the initiation step of either oligomerization
or fibrillogenesis. Previously, we have shown that incu-
bation with myo-inositol, epi-inositol and scyllo-inosi-
tol induces an immediate structural transition in Ab42
but not in Ab40 [5]. CD spectroscopy was used to
evaluate structural transitions from randomly struc-
tured Ab40 and Ab42 diluted from 40% trifluoroetha-
nol (TFE) into NaCl ⁄ P
i
. The following compounds
were examined: chiro-inositol, chiro-1-inosose, scyllo-
inositol, scyllo-inosose, epi-inositol, epi-1-inosose, epi-
2-inosose, allo-inositol and allo-3-inosose (Fig. 1). We
had previously shown the efficacy of scyllo-inositol
and epi-inositol in inducing structural transitions, but
had not examined allo-inositol; all inositol stereoiso-
mers were included for comparison with the related
inosose compounds, and chiro-inositol was used as a
negative control. We chose to use the same conditions
as in our original studies in order to have a direct
comparison with previously published data [5].
As previously reported by numerous laboratories, in
the absence of any compounds, Ab40 remains mostly
randomly structured [5,9]. As was seen for the inositol
stereoisomers, none of the inosose compounds induced
an immediate structural transition in Ab40 (data not
shown). However after 2 days, allo-3-inosose induced a
b-structural transition. By 3 days, scyllo-inosose, epi-1-

inosose, epi-2-inosose and allo-inositol induced a
strong b-structure in Ab40; as expected, the results
obtained with chiro-inositol and chiro-1-inositol were
similar to those obtained with Ab40 alone, consisting
of a mixture of random coil and b-structured peptide
(data not shown). In order to rule out Ab batch or
Fig. 1. Structures of inositol and inosose compounds.
Modulating amyloid-b aggregation M. Nitz et al.
1664 FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS
day effects, experiments were repeated three separate
times on two different Ab preparations, and gave con-
sistent, reproducible data. These results suggest that
none of the compounds are strong inducers of Ab40
structural transitions, but some may affect the kinetics
of this transition.
In contrast to Ab40, the inosose compounds
induced immediate transitions to b-structure when
incubated with Ab42 (Fig. 2). Ab42 adopted a
b-structure in the presence of scyllo-inositol (Fig. 2C),
epi-inositol (Fig. 2E), epi-1-inosose (data not shown),
epi-2-inosose (Fig. 2F), allo-inositol (Fig. 2G), and
allo-3-inosose (Fig. 2H), but not in the presence of
chiro-inositol (Fig. 2A), chiro-1-inosose (Fig. 2B), and
scyllo-inosose (Fig. 2D). These results suggest that for
epi-inositol and allo-inositol, replacement of a single
hydroxyl with a ketone does not abolish their ability
to induce Ab42 structural transitions, and that the
position of this replacement is not critical, as both
epi-1-inosose and epi-2-inosose were equally effective.
On the other hand, replacement of a single hydroxyl

group on scyllo-inositol by a ketone abolished activity
towards Ab42. Similarly, our previous results sug-
gested that the epimerization of a single hydroxyl
group from scyllo-inositol to myo-inositol substan-
tially decreased efficacy [2,4]. These results also sug-
gest that the all-equatorial position of hydroxyl
groups, as is the case for scyllo-inositol, is optimal
for interaction with Ab and that subtle changes
decrease potency.
In order to determine the affinity of this interaction,
we examined the concentration dependence of Ab42–
inosose structural transitions. We chose to compare
epi-inositol and epi-2-inosose, as both induced a strong
b-structure and their activity is retained upon modifi-
cation to a single ketone group. We have previously
shown that epi-inositol induces a transition to b-struc-
ture at a 1 : 1 (w ⁄ w) ratio [4]. epi-2-Inosose was also
able to induce a structural transition in Ab42 from a
random to a b-structure at a 1 : 1 ratio (w ⁄ w; data not
shown). This corresponds to a peptide ⁄ inosose molar
ratio of 1 : 27 (peptide concentration 10 lm) and sug-
gests a dissociation constant in the high millimolar
range for the inosose–peptide complex. These results
also suggest that the hydroxyl at position 2 of the epi-
inositol ring is not crucial for interaction with Ab.
To determine whether the inosose ⁄ inositol-induced
transition to b-structure corresponds to an increase in
the formation of insoluble Ab species, we utilized the
inherent tyrosine fluorescence of Ab in the presence
and absence of inosose compounds (Fig. 3). Ab40 was

incubated at a 1 : 20 (w ⁄ w) ratio with inosose com-
pounds for 48 h, and the tyrosine fluorescence of each
solution was measured prior to and directly after cen-
trifugation to remove insoluble Ab species (Fig. 3A).
The percentage fluorescence after centrifugation in
comparison to total fluorescence is an indication of the
amount of soluble Ab species. We measured samples
at a 48 h time point, as preliminary aggregation assays
demonstrated that approximately 60% of Ab40 and
Ab42 was pelleted at this time, which allowed us to
distinguish between compounds that increase and
decrease insoluble Ab species. Ab40 remained in solu-
tion only when incubated with epi-2-inosose and allo-
inositol, suggesting that epi-2-inosose and allo-inositol
promote a b-structural transition, leaving the complex
in a soluble form (Fig. 3A). In contrast, solutions trea-
ted with epi-1-inosose had less soluble Ab species than
Ab40 alone, suggesting that the strong b-structural
transition detected with epi-1-inosose promotes further
aggregation and increased insoluble Ab (Fig. 3A). As
previously reported, both epi-inositol and scyllo-inosi-
tol maintain Ab42 in a soluble state (Fig. 3B). Of the
stereoisomers, only allo-inositol maintained Ab42 as a
soluble species; in contrast, with epi-2-inosose, there
was a reduced level of Ab42 in solution (Fig. 3B).
These results, in combination with the CD data,
suggest that allo-inositol may exhibit antiaggregant
properties similar to those seen for other inositol
stereoisomers, and that epi-2-inosose may represent an
inducer of Ab42 but not Ab40 aggregation.

Effect of inosose compounds on Ab fibril
structure
The CD and tyrosine fluorescence studies suggest that
the inositol ⁄ inosose compounds may affect Ab nucle-
ation and assembly. However, they may have variable
effects on oligomerization versus fibril assembly and
growth [8]. Using a large panel of compounds, Glabe
et al. reported that Ab oligomerization and fibrillogen-
esis are not mutually dependent and that compounds
can affect one pathway without inhibiting the other
[8]. The characteristics of Ab40 and Ab42 assembly
products in the presence and absence of inosose com-
pounds were examined by negative-stain electron
microscopy, using two stocks of Ab40 and Ab42 to
ensure that the results obtained were consistent and
representative [5,9]. Unseeded samples of both Ab40
and Ab42 were incubated in the presence of all com-
pounds and alone for up to 96 h. When Ab40 was
incubated alone, it formed fibers of varying length, as
shown by representative electron microscopy images
(Fig. 4A). When Ab40 was incubated with chiro-inosi-
tol, chiro-1-inositol, scyllo-inositol, epi-1-inositol and
allo-3-inosose, no difference could be detected from
M. Nitz et al. Modulating amyloid-b aggregation
FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS 1665
Ab40 alone (data not shown). In contrast, when epi-2-
inosose (Fig. 4C) and allo-inositol (Fig. 4E) were incu-
bated with Ab40, very few fibers were detected after
72 h of incubation. In order to differentiate between
compounds that inhibit fiber formation and those that

decrease the kinetics of fiber formation, samples were
Wavelen
g
th (nm)
Theta
A
500
500
0
–500
–1000
–1500
500
0
–500
–1000
–1500
–2000
4000
2000
0
–2000
–4000
–6000
–8000
1000
500
0
–500
–1000

–1500
–2000
–2500
–3000
1000
500
0
–500
–1000
–1500
–2000
0
–500
–1000
–1500
–2000
0
–100
–200
–300
–400
–500
0
–500
–1000
–1500
–2000
–2500
190 200 210 220 230 240 250 260
190 200 210 220 230 240 250 260

180 190 200 210 220 230 240 250 260
190 200 210 220 230 240 250 260 190 200 210 220 230 240 250 260
190 200 210 220 230 240 250 260
190 200 210 220 230 240 250 260
190 200 210 220 230 240 250 260
B
DC
E
F
GH
chiro-inositol chiro-inosose
scyllo-inositol
scyllo-inosose
epi-inositol
epi-2-inosose
allo-inositol allo-3-inosose
Fig. 2. Determination of the secondary
structure of Ab42 in the presence and
absence of inosose compounds using CD
spectroscopy. Ab42 was randomly struc-
tured in distilled H
2
O after dilution from
40% TFE stock solutions to a final peptide
concentration of 10 m
M (solid line). Repre-
sentative CD spectra of Ab42 immediately
upon addition of chiro-inositol (A), chiro-1-
inosose (B), scyllo-inositol (C), scyllo-inosose
(D), epi-inositol (E), epi-2-inosose (F), allo-

inositol (G) and allo-3-inosose (H) (dashed
line) are presented from three separate
experiments.
Modulating amyloid-b aggregation M. Nitz et al.
1666 FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS
incubated for 14 days (Fig. 4). After 14 days, the num-
ber and structure of Ab40 fibers in the presence and
absence of epi-2-inosose (Fig. 4D) were not different
from those of Ab40 alone (Fig. 4B), suggesting that
this compound alters the kinetics of Ab40 fibrillogene-
sis but does not inhibit fiber formation. In contrast,
only a few fibers could be detected in samples incu-
bated with allo-inositol after 14 days, suggesting that
this compound inhibits fiber formation (Fig. 4F).
Although Ab40 fibrillogenesis was altered by inosose
stereoisomers, Ab42 aggregates at a much higher rate;
depending on the strength of inosose–Ab interactions,
the efficacy of inosose stereoisomers may be altered.
When Ab42 is incubated in buffer alone, it forms vari-
ous long fibers that begin to self-assemble over time
(supplementary Fig. S1). Similar to what was found
for Ab40, incubation with chiro-inositol, chiro-1-inosi-
tol, scyllo-inositol, epi-1-inositol and allo-3-inosose did
not alter the number and structure of fibers formed in
comparison with Ab42 alone. Very few fibers were
detected after a 3 day incubation with epi-2-inosose
and allo-inositol. These results suggest that epi-2-inos-
ose and allo-inositol are either inhibitors of oligomer
formation or inhibitors of both oligomer and fiber
formation. To differentiate between these two pos-

sibilities, we incubated Ab with both epi-2-
inosose and allo-inositol for 14 days. In contrast to the
3 day incubation, when Ab42 was incubated for
14 days with epi-2-inosose, small protofibril species
could be detected but were not abundant (supplemen-
tary Fig. S1). These results suggest that epi-2-inosose is
more effective at decreasing the kinetics of Ab42 fiber
formation than of Ab40. In vitro examination of
increasing molar ratios of epi-2-inosose to peptide
from 1 : 1 to 1 : 100 were investigated by negative-
stain electron microscopy. In the presence of all molar
ratios, epi-2-inosose decreased the kinetics of Ab40
and Ab42 fiber formation by a greater extent than that
of Ab42 alone (supplementary Fig. S1). Similarly, after
14 days of incubation with allo-inositol, A
b42 fibers
were less abundant but were of similar structure and
length to those seen with Ab42 alone. These results
suggest that allo-inositol is a kinetic inhibitor of Ab42
fiber formation but is unable to fully inhibit fiber for-
mation. These results further suggest that allo-inositol
binds to and inhibits fibrillogenesis for Ab40 more
effectively than for Ab42.
Compounds that inhibit Ab aggregation and fibril-
logenesis can also disaggregate preformed fibers. The
latter is important not only for decreasing plaque load
but also for increasing the pool of soluble, potentially
toxic Ab species. To investigate the disassembly of
fibers, three separate preparations of preformed Ab40
and Ab42 fibers were incubated in the presence of

varying ratios of inosose compounds, and disaggrega-
tion was monitored over a 2 week period using nega-
tive-stain electron microscopy (Fig. 5). As was seen
for the inhibition of aggregation, chiro-inositol, chiro-
1-inositol, epi-1-inositol and allo-3-inosose had no
effect on preformed Ab40 and Ab42 fibers (supple-
mentary Fig. S2). Although scyllo-inosose did not
disaggregate preformed Ab40 fibers, it appeared to
laterally aggregate fibers into thicker, rope-like fibers
(Fig. 5B). The fiber bundles had a cumulative dia-
meter of 200 A
˚
, and exhibited an apparent helical
Percentage soluble Aβ
(Soluble T
y
r fluorescence/Total fluorescence x 100)
Aβ40
Chiro-inositol
Scyllo-inosose
Epi-2-inosose
Chiro-1-inosose
Allo-inositol
Epi-1-inosose
Allo-3-inosose
Scyllo-inositol
Epi-inositol
Aβ42
Chiro-inositol
Scyllo-inosose

Epi-2-inosose
Chiro-1-inosose
Allo-inositol
Epi-1-inosose
Allo-3-inosose
Scyllo-inositol
Epi-inositol
*
*
*
*
*
A

B
0 20 40 60 80 100 120
0 20 40 60 80 100 120
Fig. 3. A tyrosine fluorescence assay was used to determine the
extent of Ab40 (A) and Ab42 (B) aggregation in the presence and
absence of inosose compounds. Ab40 and Ab42 were incubated
with various inosose compounds at a ratio of 1 : 20 (w ⁄ w) or a
1 : 500 molar ratio for 48 h. The extent of Ab aggregation was
determined using the ratio of tyrosine fluorescence before and after
centrifugation, with n = 3 per experiment and three separate exper-
iments. *P < 0.05, Fisher’s PLSD.
M. Nitz et al. Modulating amyloid-b aggregation
FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS 1667
twisting with a 150 nm period. The fibers also assem-
bled into large masses in the presence of scyllo-inos-
ose, rather than the disorganized distribution seen

with Ab40 alone (Fig. 5B). epi-2-Inosose and allo-ino-
sitol disaggregate preformed Ab40 fibers, as evidenced
by a decrease in Ab fibers and an increase in small
oligomers (Fig. 5C,D). Similar to the effects on inhibi-
tion of fibrillogenesis, allo-inositol was more effective
than epi-2-inosose at disaggregating Ab40 fibers
(Fig. 5). Similar effects were seen for inosose- and ino-
sitol-induced Ab42 fiber disaggregation (data not
shown), with the exception that allo-inositol and epi-2-
inosose had similar potency. These results suggest that
allo-inositol and epi-2-inosose may alter the equilib-
Day 3
Day 14
Aβ40
Αβ40 +
epi-2-inosose
Αβ40 +
allo-inositol
A
B
C
D
E
F
Fig. 4. Negative-stain electron microscopy of Ab40 in the presence of inosose compounds. Ab40 incubated in buffer alone demonstrates
long fibers with some small aggregates at day 3 (A), whereas only long fibers are detected after 14 days (B). After incubation in the pres-
ence of epi-2-inosose, no fibers could be detected at day 3 (C), whereas no difference was seen at day 14 (D) in comparison to Ab40 incu-
bated alone. In contrast, in the presence of allo-inositol, very few small aggregates and protofibrils could be detected at either day 3 (E) or
day 14 (F). Images are representative of three separate experiments. Scale bars: 50 nm.
Modulating amyloid-b aggregation M. Nitz et al.

1668 FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS
rium between the formation of Ab aggregates and dis-
aggregation of preformed fibers.
Effects on Ab-induced cytotoxicity
The effect of inosose compounds on Ab -induced cyto-
toxicity was investigated because disaggregation of
plaques and inhibition of the kinetics of Ab fibrillo-
genesis could increase the pool of toxic Ab species.
Ab-induced toxicity against cell lines and primary neu-
ronal cultures is well established, with Ab oligomers
and protofibrils representing the toxic species [10].
NGF-differentiated PC-12 cells were incubated in the
presence and absence of Ab40 and Ab42 with increas-
ing ratios of compounds for 24 h. Toxicity was
measured using two well-established assays for Ab-
induced toxicity, Alamar blue and sulfhydryl rhoda-
mine B assays [5]. When Ab40 and Ab42 were
preincubated with chiro-inositol, chiro-1-inositol,
scyllo-inositol, epi-1-inositol and allo-3-inosose, their
toxicity was either unaffected or increased at high
inosose concentrations, whereas the compounds alone
were not toxic (Fig. 6; data not shown). These results
suggest that at high concentrations, the compounds
may subtly increase the level of toxic species of Ab.
epi-2-Inosose rescued PC-12 cells from Ab40 ⁄ Ab42-
induced toxicity at all molar ratios (1 : 38, 1 : 380,
1 : 760), suggesting that when epi-2-inosose binds to
Ab it forms a noncytotoxic complex at molar ratios
similar to that seen previously for epi -inositol, 1 : 25
Ab ⁄ epi-inositol [5] (Fig. 6). allo-Inositol also rescued

cells from Ab40-induced cytotoxicity to the same
extent as epi-2-inosose; however, it was less effective
at rescuing them from Ab42-induced cytotoxicity
(Fig. 6). These results are consistent with the struc-
tural data, in that allo-inositol has a stronger effect
on Ab40 than on Ab42 in all regards.
Transport of compounds by the sodium
myo-inositol transporter (SMIT)-1 and SMIT-2
Inositol enters the brain by passive diffusion and by
constitutively active stereospecific transport systems
[11–15]. SMIT-1 and SMIT-2 are expressed at the
blood–brain barrier and choroid plexus, and thus rep-
resent the main transport systems for inositol and
related compounds [11–15]. Furthermore, these trans-
port systems are expressed by neurons and astrocytes
[12–16], and the latter can be used as a model system
to examine inositol uptake. The astrocytoma cell line
A
C
D
B
Fig. 5. Negative-stain electron microscopy
of preformed Ab40 fibers in the presence
and absence of inosose compounds. Ab40
formed large masses of long fibers that
were often interwoven (A). scyllo-Inosose
caused aggregation of Ab40 fibers into lar-
ger, ordered ropes without evidence of dis-
aggregation (B). epi-2-Inosose (C) and, to a
greater extent, allo-inositol (D) disaggregat-

ed fibers into short protofibrils and small
aggregates. Images are representative of
three separate experiments. Scale bars:
50 nm.
M. Nitz et al. Modulating amyloid-b aggregation
FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS 1669
1321 N1 was shown by RT-PCR to express both
SMIT-1 and SMIT-2 (data not shown), and therefore
we modified a competition uptake assay previously
used in Xenopus oocytes [17]. We examined the trans-
port of the two most effective compounds, allo-inositol
and epi-2-inosose, using the 1321 N1 astrocytoma
cell line (Fig. 7A). In order to silence the hydrogen
myo-inositol transporter (HMIT), the assays were done
in the presence of the protonophore 2,4-dinitrophenol.
In our assay, cold myo-inositol effectively competes
75% of [2-
3
H]myo-inositol uptake over a 3 h time
course, whereas epi-inositol only blocked 25%. Both
allo-inositol and epi-2-inosose block greater than 50%
of radioactive myo-inositol uptake, suggesting effective
transport of these compounds by SMIT-1 and SMIT-2.
When SMIT-1, SMIT-2 and HMIT transport was
inhibited using 2,4-dinitrophenol and phloridizin, no
radioactivity was detected with the astrocytoma cell
line, and this was unaffected by the presence or
absence of competitors (Fig. 7B). Although these
studies are indicative of inositol and inosose uptake
into astrocytoma cytoplasm, we cannot completely

differentiate between competitor-induced transporter
blockage and uptake. These results support the poten-
tial for high central nervous system bioavailability of
allo-inositol and epi-2-inosose, as SMIT-1 and SMIT-2
maintain a high concentration of myo-inositol within
the central nervous system against a concentration
gradient of 100-fold lower concentrations in the
serum [18].
Discussion
Several therapeutic approaches have been developed
for the treatment of amyloidogenic diseases, such as
AD. The small molecule inhibitor approach was ini-
tially based on the finding that small aromatic mole-
cules such as Congo red and thioflavin interact
specifically with amyloid fibrils and inhibit their for-
mation [10,19]. Evaluation of the more potent com-
pounds demonstrated that at least three hydroxyl
groups were present on the aromatic rings, and
that the charge distribution across small molecules
represents the limiting factor for Ab binding [20].
scyllo-Inositol, which was shown to possess potent
antiamyloidogenic activity in vitro and in vivo, has an
amphipathic surface. The all-equatorial orientation of
the hydroxyl groups renders the faces of the molecule
hydrophobic. Hydroxyl groups located at the edge
of the ring are capable of hydrogen bonding with
Ab. allo-Inositol and epi-2-inosose also have single
hydrophobic surfaces and therefore may bind to simi-
lar sites as scyllo-inositol, thus accounting for the
similar effects seen in vitro between these molecules

(Table 1).
Conversion of a single hydroxyl group contained
within a reactive group in an all-equatorial orientation
(scyllo-inositol) to a ketone (scyllo-inosose) renders the
molecule inactive with respect to inhibition of Ab olig-
omerization and fiber formation. This, however, does
not preclude any interaction with Ab. scyllo-Inosose is
able to bind to Ab, as illustrated by the presence of
A
B
Epi-1-inosose
Epi-2-inosose
Epi-2-inosose
Epi-1-inosose
Allo-inositol
Allo-inositol
0 20406080100
Percent cell survival
120
0 20 40 60 80 100 120
*
*
*
*
*
*
*
*
*
*

*
*
*
*
*
*
Fig. 6. Concentration dependence of inosose rescue of Ab40-
induced (A) and Ab42-induced (B) cytotoxicity. Percentage cell sur-
vival was determined using the sulfhydrylrhodamine B assay and
NGF-differentiated PC-12 cells alone as a control. Ab was incubated
in the presence of increasing concentrations of inosose for 3 days
prior to determining cell survival: Ab alone (black bars); Ab ⁄ inosose
molar ratios of 1 : 25–38 (gray bars), 1 : 250–380 (dotted bars),
1 : 500–760 (white bars); and inosose alone (hatched bars). Bars
represent mean ± SEM of three separate experiments. *P < 0.05,
Fisher’s PLSD.
Modulating amyloid-b aggregation M. Nitz et al.
1670 FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS
large, rope-like bundles of Ab when it is incubated
with preformed Ab fibers. This suggests that binding
of scyllo-inosose to Ab fibers is sufficient to stabilize
the macromolecular structure of pre-existing fibers by
interfiber stabilization. This is analogous to extended
glycosaminoglycan polymer aggregation of pre-existing
Ab40 and Ab42 fibers, where the minimum unit com-
prises chondroitin sulfate-derived disaccharides [21].
Together, these studies highlight the importance of the
extent and distribution of charge across the backbone
for the interaction with and stabilization of Ab by sul-
fates for glycosaminoglycans and hydroxyl groups for

inositols [5,20]. However, scyllo-inosose is unique, as
this small six-carbon ring is sufficient to laterally
aggregate Ab, whereas monosaccharides derived from
glycosaminoglycans cannot.
In agreement with previous reports [4,5], we show
that allo-inositol induces a structural transition in Ab
and that replacement of a single hydroxyl for a
ketone (allo-3-inosose) does not inhibit this structural
transition. However, only allo-inositol inhibits Ab
fiber formation, suggesting that, although the inosose
compounds are able to bind Ab, this interaction is
not sufficient to inhibit further assembly to mature
fibers. These results are in agreement with our previ-
ous studies showing that myo-inositol and scyllo-inosi-
tol stabilize small b-sheet-containing Ab conformers
and inhibit progression to Ab fibers [4,5]. Further-
more, a ketone at position 3 of the allo-conformation
changes the activity of this molecule from an inhibi-
tor of fibrillogenesis to a kinetic inhibitor of assem-
bly. epi-2-Inosose accelerated the Ab structural
transition and inhibited fibrillogenesis. The magnitude
of the effect of epi-2-inosose was similar to that of
epi-inositol [4]. l-epi-1-Inosose, as reported above, did
not inhibit fiber formation, suggesting that a hydroxyl
at position 1 but not at position 2 is important for
this function.
Inositol and inosose compounds have unique in vivo
stability and toxicology as compared to compounds
reported in the literature for treatment of AD [22–
24]. For example, the polyphenols are considered to

be antioxidants and are rapidly degraded in the pres-
ence of reactive oxygen species [25] and by cyto-
chrome P450 in the liver [26]. As inositols are
saturated, they are very stable to oxidative conditions,
whereas the inosose compounds may be susceptible to
ketose reactions in vivo. Thus, these compounds,
which probably bind to similar sites on Ab, have very
different pharmacological profiles. In addition, the
inositol and inosose compounds have other advanta-
ges over other small molecule therapies, in that they
are actively transported across the blood–brain
barrier [14,27], have low toxicity, and are readily
available orally. Only a small subpopulation of drugs
has the low molecular weight and high lipophilicity
that are necessary to allow diffusion across the
blood–brain barrier. For these reasons, it has been
estimated that over 98% of all small molecule drugs
developed do not cross the blood–brain barrier.
SMITs, present at the blood–brain barrier, transport
‡ ‡ ‡ ‡ ‡
A

B
Control
++
++
++
*
Myo-inositol
No inhibitors

Control
Myo-inositol
Epi-inositol
Epi-2-inosose
Allo-inositol
Epi-inositol
Epi-2-inosose
Allo-inositol
0 20 40 60 80 100 120
Active transport (% control)
0 20 40 60 80 100 120
Active trans
p
ort
(
% control
)

Fig. 7. Substrate specificity of the SMIT-1 and SMIT-2 transporters
for inositol and inosose compounds. (A) Transport of 100 l
M
[2-
3
H]myo-inositol was measured in the presence or absence of
10 m
M cold myo-inositol as a positive control or 10 mM other inosi-
tol and inosose compounds. (B) Inhibition of 100 l
M [2-
3
H]myo-ino-

sitol in the presence of SMIT-1, SMIT-2 and HMIT inhibition, in the
presence and absence of competitors. Scheffe test, a = 0.05,
*P < 0.05,
à
P < 0.001.
M. Nitz et al. Modulating amyloid-b aggregation
FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS 1671
most of the inositol stereoisomers, as well as related
small molecules, suggesting there is some breadth in
tolerance with the substrate flexibility of these trans-
porters [3,14,27]. Our results demonstrate that both
epi-2-inosose and allo-inositol are effectively trans-
ported by SMIT-1 and SMIT-2.
Collectively, our results demonstrate that the stereo-
chemistry of inositol binding to Ab is unaffected by
substitution of one hydroxyl group for a ketone at var-
ious positions around the carbon ring, but, depending
on the position of the ketone, there are dramatic
effects on the inhibition of fiber formation (Table 1).
These results highlight the potential negative impact of
subtle changes in charge distribution on small
molecules such as inositol when using drug-based
design of new compounds.
Experimental procedures
Peptides
Ab40 and Ab42 were synthesized by solid-phase Fmoc-
chemistry by the Hospital for Sick Children’s Biotechnol-
ogy Centre (Toronto, Canada). They were purified by
RP-HPLC on a C18 lbondapak column. Ab peptides were
initially dissolved in 0.5 mL of 100% trifluoroacetic acid

(Aldrich Chemicals, Milwaukee, WI, USA), to ensure that
the peptide remained monomeric and free of fibril seeds,
diluted in distilled H
2
O, and immediately lyophilized. Ab
peptides were then dissolved at 1 mgÆmL
)1
in 40% TFE
(Aldrich Chemicals) in distilled H
2
O and stored at )20 °C
until use [28].
CD
CD spectra were recorded on a Jasco Circular Dichroism
Spectrometer Model J-715 (Easton, MD, USA) at 25 °C, as
previously described [9]. Spectra were obtained from 200 to
250 nm, with a 0.5 nm step, 1 nm bandwidth and 1 s col-
lection time per step. Spectra were averaged from five
repeat scans. Peptide ⁄ inositol ratios were 1 : 1 (w ⁄ w) or
1 : 25–38 by molar ratio, with a final peptide concentration
of 10 lm. CD experiments were repeated three times on
two separate Ab stocks. The effect of the inositol com-
pounds on peptide conformation was determined by adding
an aliquot of stock peptide solutions into cosolvent solu-
tions suspended in 20 mm sodium phosphate buffer
(pH 7.0) or distilled H
2
O (approximately pH 6.8). CD spec-
tra were examined immediately after addition of Ab and
over a 96 h time course. The contribution of inositol com-

pounds to the CD signal was removed by subtracting the
inositol-only spectra. Ab peptide conformations were
determined in 40% TFE ⁄ H
2
O, in 20 mm phosphate buffer
and distilled H
2
O under the same conditions.
Tyrosine fluorescence assay
Steady-state fluorescence was measured at 20 °C using an
AM-1 fluorescence spectrophotometer (Photon Technology
International, London, Canada) equipped with excitation
intensity correction and a magnetic stirrer [4]. Tyrosine
emission spectra from 290 nm to 340 nm were collected
(excitation wavelength 281 nm, 0.5 sÆnm
)1
, bandpass 4 nm).
A cuvette with a 1 cm path length was used. For centrifu-
gation studies, 50 lm Ab40 or Ab42 was incubated in the
presence or absence of inosose compounds at a peptide ⁄
inosose ratio of 1 : 20 (w ⁄ w) for 24 h with continuous
shaking. Tyrosine fluorescence was measured, and this was
followed by a centrifugation step at 15 600 g for 30 min in
order to sediment insoluble species. The relative amount of
tyrosine fluorescence in the supernatant was then deter-
mined. The fluorescence of the noncentrifuged samples was
used as a measure of total fluorescence. Steady-state fluo-
rescence measurements were repeated in three separate
experiments with a sample size of three per condition
within each experiment.

Table 1. Summary of Ab40 ⁄ Ab42 interaction with inositol and inosose compounds.
Compound
Secondary structure
(CD spectroscopy)
Fiber formation
(electron microscopy) PC-12 cell toxicity
Ab42 Ab40 Ab42 Ab40 Ab42
L-chiro-Inositol Random < Control < Control No effect No effect
L-chiro-1-Inosose Random < Control No effect Inhibits No effect
scyllo-Inositol b-Sheet No effect Inhibits Inhibits Inhibits
scyllo-Inosose Random < Control No effect Increase Increase
epi-Inositol b-Sheet No effect Inhibits Inhibits Inhibits
L-epi-1-Inosose Partial b-sheet No effect No effect No effect Increase
L-epi-2-Inosose Strong b-sheet < Control Inhibits Inhibits Inhibits
allo-Inositol b-Sheet Inhibits < Control Inhibits Inhibits
D-allo-3-Inosose Random < Control No effect No effect No effect
Modulating amyloid-b aggregation M. Nitz et al.
1672 FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS
Electron microscopy
Ab40 and Ab42 were incubated in the presence and absence
of inositol compounds at a final peptide concentration of
100 lgÆmL
)1
.Ab was used directly after solubilization in
water at a concentration of 10 mgÆmL
)1
, or after assembly
into mature amyloid fibers. Serial dilutions of inositol were
added to Ab and incubated at room temperature for up to
2 weeks. The Ab ⁄ inositol ratio was varied up to 1 : 20

(w ⁄ w) or 1 : 500 molar ratio. For negative-stain electron
microscopy, carbon-coated pioloform grids (Canemco-
Marivac, Lakefield, Canada) were floated on aqueous solu-
tions of peptides. After the grids were blotted and air-dried,
the samples were stained with 1% (w ⁄ v) phosphotungstic
acid (Aldrich Chemicals) and examined on a Hitachi 7000
electron microscope operated at 75 kV [5,9]. The results
presented are representative images from three separate
experiments on different Ab40 and Ab42 stocks.
Ab-induced toxicity
PC-12 cells were plated at 500 cells per well in a 96-well
plate and differentiated in 30 ngÆmL
)1
NGF (Alamone
Labs, Israel) diluted in N2 ⁄ DMEM (Gibco ⁄ BRL, Rock-
ville, MD, USA). Cells were differentiated over 5–7 days to
a final cell number of 10 000–15 000 per well. Ab in the
presence and absence of inosose compounds, at molar
ratios of 1 : 25 to 1 : 500, was aged for 3 days at room
temperature to induce fibrillogenesis before addition to
cultures at a final Ab concentration of 0.1 mgÆmL
)1
, and
incubated for 24 h at 37 °C. Toxicity was assayed using the
Alamar Blue Assay (Biosource Inc., Camarillo, CA, USA)
and the sulfhydrylrhodamine B assay [5]. Toxicity assays
were repeated three times with n = 4 per condition within
an experiment.
Inositol transport assay
1321 N1 astrocytoma cells were plated onto 24-well plates at

a density of 10 000 cellsÆmL
)1
and grown for 4 days. On the
fifth day, cells were washed in NaCl ⁄ P
i
, and then incubated
in the protonophore 2,4-dinitrophenol (1 mm, 10 min,
37 °C; Aldrich Chemicals), with the goal of silencing the
HMIT and preferentially leaving active the SMIT-1 and
SMIT-2. Following incubation, the protonophore solution
was removed and replaced with a solution containing:
2,4-dinitrophenol (1 mm), [2-
3
H]myo-inositol (3 lCiÆmL
)1
,
100 lm), BSA (0.1%, used to reduce background radioactiv-
ity), and a sugar competitor (10 mm). The sugar competitors
tested were cold myo-inositol, epi-inositol, epi-2-inosose, and
allo-inositol, as well as an NaCl ⁄ P
i
control. Following incu-
bation (3 h, 37 °C), cells were washed twice with NaCl ⁄ P
i
and were dissolved using 2% SDS. Cells were transferred to
vials containing 5 mL of scintillation fluid, and their radio-
activity was measured using a scintillation counter.
Negative control experiments with 1321 N1 astrocytoma
cells in the presence of 2,4-dinitrophenol (1 m m) and plori-
dizin (2 mm) to inhibit SMIT-1, SMIT-2 and HMIT were

incubated with [2-
3
H]myo-inositol (3 lCiÆmL
)1
, 100 lm),
BSA (0.1%, used to reduce background radioactivity) and
a sugar competitor (10 mm).
Acknowledgements
The authors would like to thank Dr Wang at the
Hospital for Sick Children’s Biotechnology Center for
synthesis of peptides, and Kevin Da Silva for helpful
comments on the manuscript. The authors acknowl-
edge support from the Ontario Alzheimer’s Society
(J. McLaurin), Canadian Institutes of Health Research
(J. McLaurin, M. Nitz), Natural Science and Engineer-
ing Research Council of Canada (J. McLaurin,
M. Nitz), and Cryptic Rite Charitable Foundations
(J. McLaurin).
References
1 Fisher SK, Novak JE & Agranoff BW (2002) Inositol
and higher inositol phosphates in neural tissues: homeo-
stasis, metabolism and functional significance. J Neuro-
chem 82, 736–754.
2 McLaurin J, Kierstead ME, Brown ME, Hawkes CA,
Lambermon MHL, Phinney AL, Darabie AA, Cousins
JE, French JE, Lan MF et al. (2006) Cyclohexanehexol
inhibitors of Ab aggregation prevent and reverse
Alzheimer phenotype in a mouse model. Nat Med 12,
801–808.
3 Fenili D, Brown ME, Rappaport RV & McLaurin J

(2007) Properties of scyllo-inositol as a therapeutic
treatment of AD-like pathology. J Mol Med 85, 603–
611.
4 McLaurin J, Franklin T, Chakrabartty A & Fraser PE
(1998) Phosphatidylinositol and inositol involvement in
Alzheimer amyloid-beta fibril growth and arrest. J Mol
Biol 278, 183–194.
5 McLaurin J, Goloumb R, Jurewicz A, Antel JP &
Fraser PE (2000) Inositol stereoisomers stabilize an
oligomeric aggregate of Alzheimer amyloid beta peptide
and inhibit Ab-induced toxicity. J Biol Chem 275,
18495–18502.
6 Chishti MA, Yang D, Janus C, Phinney AL, Horne P,
Pearson J, Strome R, Zuker N, Loukides J, French J
et al. (2001) Early-onset amyloid deposition and cogni-
tive deficits in transgenic mice expressing a double
mutant form of amyloid precursor protein. J Biol Chem
276, 21562–21570.
7 Townsend M, Cleary JP, Mehtra T, Hofmeister J, Les-
ne S, O’Hare E, Walsh DM & Selkoe DJ (2006) Orally
available compound prevents deficits in memory caused
M. Nitz et al. Modulating amyloid-b aggregation
FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS 1673
by the Alzheimer amyloid-oligomers. Ann Neurol 60,
668–676.
8 Necula M, Kayed R, Milton S & Glabe CG (2007)
Small molecule inhibitors of aggregation indicate that
amyloid b oligomerization and fibrillization pathways
are independent and distinct. J Biol Chem 282, 10311–
10324.

9 McLaurin J & Fraser PE (2000) Effect of amino acid
substitutions in Ab1–28 on fibril formation and glycos-
aminoglycan interactions. Eur J Biochem 267, 6353–
6361.
10 Irie K, Murakami K, Masuda Y, Morimoto A, Ohig-
ashi H, Ohashi R, Takegoshi K, Nagao M, Shimizu T
& Shirasawa T (2005) Structure of b-amyloid fibrils and
its relevance to their neurotoxicity: implications for the
pathogenesis of Alzheimer’s disease. J Biosci Bioeng 99,
437–447.
11 Spector R (1988) Myo-inositol transport through the
blood–brain barrier. Neurochem Res 13 , 785–787.
12 Wiesinger H (1991) Myo-inositol transport in mouse
astroglia-rich primary cultures. J Neurochem 56, 1698–
1704.
13 Rubin LJ & Hale CC (1993) Characterization of a Mg-
dependent, Na-inositol co-transport process in cardiac
sarcolemmal vesicles. J Mol Cell Cardiol 25, 721–731.
14 Uldry M, Ibberson M, Horisberger J-D, Chatton J-Y,
Riederer BM & Thorens B (2001) Identification of a
mammalian H-myo-inositol symporter expressed pre-
dominantly in the brain. EMBO J 20, 4467–4477.
15 Coady MJ, Wallendorff B, Gagnon DG & Lapointe J-Y
(2002) Identification of a novel Na+ ⁄ myo-inositol
cotransporter. J Biol Chem 277, 35219–35224.
16 Isaacks RE, Bender AS, Kim CY & Norenberg MD
(1997) Effect of osmolality and myo-inositol deprivation
on the transport properties of myo-inositol in primary
astrocyte cultures. Neurochem Res 22, 1461–1469.
17 Hager K, Hazama A, Kwon HM, Loo DD, Handler JS

& Wright EM (1995) Kinetics and specificity of the
renal Na+ ⁄ myo-inositol cotransporter expressed in
Xenopus oocytes. J Membr Biol 143, 103–113.
18 Palmano KP, Whiting PH & Hawthorne JN (1977) Free
and lipid myo-inositol in tissues from rats with acute
and less severe streptozotocin-induced diabetes. Biochem
J 167, 229–235.
19 Lorenzo A & Yankner B (1994) b-Amyloid neurotoxic-
ity requires fibril formation and is inhibited by Congo
Red. Proc Natl Acad Sci USA 91, 12243–12247.
20 Porat Y, Abramowitz A & Gazit E (2006) Inhibition of
amyloid fibril formation by polyphenols: structural simi-
larity and aromatic interactions as a common inhibition
mechanism. Chem Biol Drug Des 67, 27–37.
21 Fraser PE, Darabie AA & McLaurin J (2001) Amyloid-
b interactions with chondroitin sulfate-derived mono-
saccharides and disaccharides. Implications for drug
development. J Biol Chem 276, 6412–6419.
22 Taniguchi S, Suzuki N, Masuda M, Hisanaga S, Iwat-
subo T, Goedert M & Hasegawa M. (2005) Inhibition
of heparin-induced tau filament formation by phenothi-
azines, polyphenols, and porphyrins. J Biol Chem 280,
7614–7623.
23 Pickhardt M, von Bergen M, Gazova Z, Hascher A,
Biernat J, Mandelkow EM & Mandelkow E. (2005)
Screening for inhibitors of tau polymerization. Curr
Alzheimer Res 2, 219–226.
24 Howlett DR, George AR, Owen DE, Ward RV &
Markwell RE. (1993) Common structural features deter-
mine the effectiveness of carvedilol, daunomycin and

rolitetracycline as inhibitors of Alzheimer. Biochem J
343, 419–423.
25 Lashuel HA, Hartley DM, Balakhaneh D, Aggarwal A,
Teichberg S & Callaway DJE (2002) New class of inhib-
itors of amyloid-b fibril formation. Implications for the
mechanism of pathogenesis in Alzheimer’s disease.
J Biol Chem 277, 42881–42890.
26 Guengerich FP (2001) Common and uncommon cyto-
chrome P450 reactions related to metabolism and chem-
ical toxicity. Chem Res Toxicol 14, 611–650.
27 Spector R (1978) The transport and metabolism of
scyllo-inositol in the central nervous system. J Neuro-
chem 31, 1113–1115.
28 Jao S, Ma K, Talafous J, Orlando R & Zagorski MG
(1997) Trifluoroacetic acid pretreatment reproducibly
disaggregates the amyloid b-peptide. Amyloid Int J Exp
Clin Invest 4, 240–252.
Supplementary material
The following supplementary material is available
online:
Fig. S1. Negative-stain electron microscopy of Ab42 in
the presence of inosose compounds.
Fig. S2. Negative-stain electron microscopy of
preformed Ab42 fibers in the presence and absence of
inosose compounds.
This material is available as part of the online article
from
Please note: Blackwell Publishing are not responsible
for the content or functionality of any supplementary
materials supplied by the authors. Any queries (other

than missing material) should be directed to the corre-
sponding author for the article.
Modulating amyloid-b aggregation M. Nitz et al.
1674 FEBS Journal 275 (2008) 1663–1674 ª 2008 The Authors Journal compilation ª 2008 FEBS