Development of Gleevec Analogues for Reducing Production of β-Amyloid Peptides through Shifting β-Cleavage of Amyloid Precursor Proteins

Article abstract of DOI:10.1021/acs.jmedchem.8b02007

Imatinib mesylate, 1a, inhibits production of β-amyloid (Aβ) peptides both in cells and in animal models. It reduces both the β-secretase and γ-secretase cleavages of the amyloid precursor protein (APP) and mediates a synergistic effect, when combined with a β-secretase inhibitor, BACE IV. Toward developing more potent brain-permeable leads, we have synthesized and evaluated over 75 1a-analogues. Several compounds, including 2a-b and 3a-c, inhibited production of Aβ peptides with improved activity in cells. These compounds affected β-secretase cleavage of APP similarly to 1a. Compound 2a significantly reduced production of the Aβ42 peptide, when administered (100 mg/kg, twice daily by oral gavage) to 5 months old female mice for 5 days. A combination of compound 2a with BACE IV also reduced Aβ levels in cells, more than the additive effect of the two compounds. These results open a new avenue for developing treatments for Alzheimer's disease using 1a-analogues.

Full text of DOI:10.1021/acs.jmedchem.8b02007

Article
pubs.acs.org/jmc
Cite This: J. Med. Chem. XXXX, XXX, XXX−XXX
Development of Gleevec Analogues for Reducing Production of
β‑Amyloid Peptides through Shifting β‑Cleavage of Amyloid
Precursor Proteins
Weilin Sun, William J. Netzer, Anjana Sinha, Katherina Gindinova, Emily Chang,
and Subhash C. Sinha*
Laboratory of Molecular and Cellular Neuroscience, The Rockefeller University, New York, New York 10065, United States
S
* Supporting Information
ABSTRACT: Imatinib mesylate, 1a, inhibits production of β-amyloid (Aβ) peptides both in cells and in animal models. It
reduces both the β-secretase and γ-secretase cleavages of the amyloid precursor protein (APP) and mediates a synergistic effect,
when combined with a β-secretase inhibitor, BACE IV. Toward developing more potent brain-permeable leads, we have
synthesized and evaluated over 75 1a-analogues. Several compounds, including 2a−b and 3a−c, inhibited production of Aβ
peptides with improved activity in cells. These compounds affected β-secretase cleavage of APP similarly to 1a. Compound 2a
significantly reduced production of the Aβ42 peptide, when administered (100 mg/kg, twice daily by oral gavage) to 5 months
old female mice for 5 days. A combination of compound 2a with BACE IV also reduced Aβ levels in cells, more than the
additive effect of the two compounds. These results open a new avenue for developing treatments for Alzheimer’s disease using
1a-analogues.
minimizing inhibitor dosage. However, the p-glycoprotein
(pGP) present at the blood−brain barrier (BBB) rapidly
pumps out 1a from the brain, thus preventing 1a from
achieving therapeutic concentration in the central nervous
system.20,21 This together with the fact that 1a is a relatively
less explored molecular entity in terms of APP metabolism
prompted us to develop more potent and brain-permeable 1a-
analogues. Initially, we have prepared straight 1a-analogues by
performing simple modifications of 1a and determined and
compared their effects on Aβ production in cellular assays. We
obtained dozens of compounds that show some improvement
in their activities. We focused on five potent 1a-analogues, 2a−
b and 3a−c (Figure 1), that also accumulated moderately in
wild-type mouse brain. Here, we describe the results of our
studies with these compounds showing that: (1) compounds
2a−b and 3a−c reduce Aβ production similarly to 1a by
shifting β-cleavage of APP14 and (2) compound 2a reduces
Aβ42 levels in 5 months old 2xTg female AD mice significantly
when administered 100 mg/kg twice daily for 5 days.
1. INTRODUCTION
Neurotoxic forms of β-amyloid (Aβ) peptides that have been
implicated for Alzheimer’s disease (AD)1−3 are produced by
sequential cleavages of the amyloid precursor protein (APP)
by β-secretase (BACE1) followed by γ-secretase (GS).4−6 N-
and C-terminal fragments, soluble APPβ (sAPPβ)7 and β-C
terminal fragment (βCTF),8 are two other potentially
neurotoxic entities produced by cleavage of APP by
BACE1.5,9 We have previously shown that imatinib mesylate
(Gleevec), 1a, which is a potent bcr-Abl kinase inhibitor,10 and
also an FDA-approved drug for the treatment of chronic
myelogenous leukemia11 and gastrointestinal stromal tumor,12
reduces levels of Aβ peptide in cultured cells and in guinea
pigs.13 Compound 1a lowers Aβ levels by selective, indirect
reduction of both BACE and GS cleavages of full-length APP
and its metabolite, APP-βCTF, respectively.13,14 This dual
targeting and secretase selectivity set 1a apart from typical
BACE and GS inhibitors (BACEi and GSi) that have failed in
clinical trials.15−18 Besides, a combination of 1a and the potent
BACEi, BACE IV, in N2a695 cells acts in a synergistic manner
to lower Aβ production.14 In a clinical setting, such a
combination could reduce the side-effects of a BACEi19 by
Received: December 21, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. Structure of imatinib mesylate, 1a, and selected potent analogues for evaluation.
a
Scheme 1. Development of Active 1a-analogues
a
Fragmentation of 1a across “a” and “b” can afford compounds 4 and 5. Compound 2a can be obtained from 1a via 6a and 7a, involving steps c−e
or via compounds 3a−c (steps f−h). Compound 1a can also be modified to 8a, which is a truncated analog of 2a, and to 9a and 9r. Compound 9a
possesses an A ring modified, whereas 9r has both A and E ring modified. Additional analogues of 9a have one or more rings modified from A, B,
and/or C ring(s), and a few of those also have an E ring modified (see: Table 2)
sequential changes in connection sites between the C/D or D/
E rings in 1a, as shown in Scheme 1, via 6a → 7a → 2a → 3c
→3b →3a and via 3a → 3b → 3c → 2a.
2. RESULTS AND DISCUSSION
Our study started with the simple modification of 1a for
identifying the molecular motifs that are required for reducing
the production of secreted Aβ peptides. We prepared truncated
compounds 4 and 5 (Scheme 1) containing three or four
rings, and their analogs (Figure S1). Compounds 5 was
previously shown to lower Aβ level more potently than 1a in
an assay performed in drosophila.22 Because of the low
molecular size and less number of hydrogen donor and
acceptor atoms (N, O), such compounds could also possess
greater brain permeability and retention. However, we found
that all truncated 1a-analogues, including 5, were inactive or
had lost substantial activity (see: Figure S2A). The results led
us to focus on 1a-analogues that possess all five rings of the
parent compound. Thus, we made step-wise modifications in
one or more rings of 1a and prepared compounds of series 1−
3 that possess ABC rings identical to the parent compound, 1a.
Compounds of series 2 and 3 were developed by making
We further modified 1a to compounds of the 8 and 9 series,
such as 8a and 9a or 9r in Scheme 1. Compound 8a closely
resembles 2a, whereas compound 9a possesses a modified A
ring. Other analogues of 9a can have one or more of ABC rings
modified, and some analogues, such as 9r (Scheme 1) also
have an E ring modified (see Table 2). Because the primary
goal of this study was to identify modules that could enhance
activity, majority of compounds, especially those possessing a
piperazine ring in 1−2, 5−6, and in the 9 series can still
possess pGP activity. Conversely, the parent compounds or
some analogues: (1) lacking a piperazine ring in series 1 or a
donor/acceptor “N”/”O” atom(s) in series 2 and 3 or (2)
having less freedom of rotation across D/E rings in series 3 and
7 are likely to have lower pGP activity. Several 8- and 9-series
compounds possessing smaller molecular size and lacking one
B
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 2. General Synthesis of 1a-Analogues
a
Key: MW, microwave; (a) HATU, DIEA, DMF, RT, 16 h; (b) aq Me2NH (40%, pressure bottle), 120 °C, 16 h; (c) MW, Pd(PPh3)4, aq Cs2CO3
(2 M), dioxane, 100 °C, 45 min; (d) (i) TFA, CH2Cl2, RT, 16 h, (ii) Me2NH (2 M, THF), NaCNBH3, 0 °C-RT, 16 h; (e) 4 N HCl, dioxane-
EtOAc, RT, 16 h; (f) (CH2O)n, NaCNBH3, 0 °C-RT, 16 h; (g) DIEA, THF, RT, 1 h; (h) MW, 100 °C, K2CO3, MeCN, 30 min.
or more donor/acceptor “N” atom could also possess lower
pGP activity.23
and 6−9 can be found in Figure 1, Schemes 1 and 2, Tables 1
and 2.
We prepared compounds 2a−b, 3a−f, 6a−b, and 7a−b by
modifying the known syntheses of compound 1a,24 using
amine 4 and acid 10a−h or acid chloride 14a in 1−4 steps
(Scheme 2). Thus, HATU coupling of 4 with 10a−h or the
usual amide coupling with 14a in the presence of
diisopropylethylamine (DIEA) afforded intermediates 11a−b,
12a, 3d-Boc, 3e-Boc, 3f-Boc, and 7b-Boc, or the target
compounds 3a−b. Homobenzyl and benzyl chloride inter-
mediates, 11a and 11b were substituted with dimethylamine,
N-methylpiperazine and piperidine to yield compounds 2a and
6a−b. Intermediate 12a was converted to compound 2b by
performing Suzuki coupling25,26 with trans-2-ethoxyethenyl-1-
boronic acid pinacol ester,27 followed by trifluoroacetic acid
(TFA)-deprotection of the resulting vinyl ether, 13a, to afford
aldehyde and reductive amination of the latter product with
dimethylamine. Boc intermediates, 3d-Boc, 3e-Boc, 3f-Boc,
and 7b-Boc, were acid-deprotected to yield the target
compounds, 3d−f and 7b. Finally, compounds 3c and 7a
were obtained by reductive amination of 3f and 7b using
paraformaldehyde, (CH2O)n. All 1a-analogues, including
compounds 1c−l, 2c−j, 3g−i, and 6c−g, were prepared
similarly or by modifying one or more steps. We also designed
and prepared compounds, 8a−q and 9a−u. Compounds of
series 8 are nor homologs of 2a and analogues, whereas
compounds 9a−u possess one or more rings of ABC that are
different from the parent compound, 1a, and 9r-u also have an
E ring modified. Structures of all compounds from series 1−3
2.1. Screening of 1a-Analogues Identified Dozens of
Compounds Inhibiting Production of the Aβ Peptide in
N2a695 Cells Similarly or Greater than the Parent
Compound 1a. We examined all compounds using mouse
neuroblastoma N2a695 cells overexpressing human APP695.14
N2a695 cells secrete sufficient quantities of Aβ40 and Aβ42
peptides into cell culture media in 5 h to determine their
concentration with high accuracy using commercially available
enzyme-linked immunosorbent assay (ELISA) kits. Under
these conditions, Aβ40 levels were generally reduced by 35−
45% when cells were treated with 10 μM 1a compared to a
negative dimethylsulphoxide (DMSO) control. Occasionally,
Aβ40 levels were found to be greater than 45% or lower than
35% under similar conditions using 10 μM 1a, and readings for
the test compounds also varied similarly. These differences
may have resulted from testing cell cultures in different growth
phases (i.e., logarithmic growth vs Contact-inhibited cells).
The screening results for all compounds showing their effects
on Aβ40 levels are shown in Figure S2. We have found as
many as 44 additional compounds, in addition to compounds,
2a−b and 3a−c, which reduced Aβ levels by 40−70% or more.
Thus, by comparing the structures of all 75 compounds
(Figure 1, Schemes 1 and 2, Tables 1 and 2, and Figure S1)
and the results of Aβ production, we have made the following
conclusions:
First, all four ABCD rings of 1a and an alkyl or cyclic amine
are required to maintain the Aβ-lowering activity of the parent
C
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Structure of 1a-Analogues (Series 1−3 and 6) Possessing Modified DE Rings
compound. This is evident from the fact that none of the
compounds among 4a-h and 5 (Figure S1) reduced Aβ levels
by more than 10−20% (Figure S2A). Second, the pyridine “N”
in ring “A” of 1a-analogues, which is required in 1a for kinase
binding, is not essential for inhibiting Aβ secretion. In fact,
several compounds having phenyl and substituted-phenyl,
instead of the pyridine “A” ring, were equally or more active
than 1a. These include compounds 9a−b, 9e, 9g−h, and 9j−u
(Table 2 and Figure S2B,C). Third, a guanidine-type
connection between the “B” and “C” rings may not be
required, as compound 9e that is lacking such function was
also found to be quite active (Table 2 and Figure S2C).
Fourth, modification of the “C”-phenyl ring as a pyridine ring,
as in compound 9l or 9n−p, had little effect on the activity.
Fifth, substitution of “D” ring with F, CF3, or OMe, as in
compounds 1c−d and 2j, reduced the activity. However, there
was little effect when the “D−E” ring connection was moved
from the “1,4” to “1,3” position, as shown by comparing
compounds 1a versus 6a (Schemes 1 and 2 and Table 2).
Finally, compounds possessing an alkyl amine or cyclic amine
instead of the piperazine (E) ring were equally as active as the
parent compound. At least one amine “N” was required in the
“E” ring, where the position of “N” was also important. For
example, compound 1k is active, whereas the positional isomer
1g and compound 1j that are lacking “N” are inactive.
Accumulated in Mouse Brain. We focused on five hit
compounds, 2a−b and 3a−c, and examined their effects on
both Aβ40 and Aβ42 levels at 2.5−10 μM concentrations. All
compounds reduced both Aβ40 and Aβ42 significantly in
N2a695 cells at 5 and 10 μM concentrations, with an apparent
IC50 less than 10 μM and greater than 5 μM. All five
compounds have shown cell viability comparable to the DMSO
control confirming that the effect on Aβ levels was not due to
toxicity (Figure 2B).
We determined the brain permeability and retention of
compounds 2a−b and 3a−c, as compared to 1a, in 8 weeks old
C57BL/6J male mice by administering compounds i.p. (50−
150 mg/kg), collecting both mouse brain and plasma 4 h post
injection, and extracting with ethanol or acetonitrile as
described (see: General Methods). 1a is known to not
accumulate efficiently in brain because of rapid efflux by the
BBB,28 and this is reflected in the low brain concentrations of
1a detected 4 h post administration (Figure 2C). In contrast,
the new tested analogues accumulated in the brain at relatively
high concentrations (between 21 and 127 pg/g, when
administered at a dose of 50 mg/kg) and were found in
plasma at higher concentrations (1.0−11.6 μM).
2.3. 1a and Its Analogues Reduce Aβ Production in a
Similar Manner. Next, we determined whether 1a-analogues
lower Aβ levels by affecting β-cleavage of APP similarly to 1a.
For this, we performed western blotting (WB) experiments
with the cell lysates obtained from the compound treatments
2.2. 1a-Analogues, 2a−b and 3a−c Inhibited Pro-
duction of Aβ Peptide in N2a695 Cells Potently and
D
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 2. Structure of 1a-Analogues: Series 8 Possessing Modified DE Rings, and Series 9 Possessing One or More of ABC
Rings Different from 1a with/without Modification in the E Ring
and probed the blots using antibody RU369 that recognizes
APP-CTFs.29 As shown in Figure 3A,B, cell lysates from 1a
and all its analog-treated samples reveal intense bands for APP
CTFs compared to those from the control DMSO-treated
samples. In addition, there are multiple APP CTFs, including
16 kDa CTF-141, comprising 141 amino acid residues and the
10 kDa CTF (weak band), in cell lysates from 1a and all its
analogues tested. It should be noted that the 10 kDa-CTF
migrating near βCTF is not a product of BACE cleavage, and
that the CTF-141 and other long APP CTFs in 1a-treated
samples fractionate with the lysosome.14 Compounds 2a and
3a produced significantly higher levels of the 16 kDa CTF-141
than compounds 1a, 2b, 3b, or 3c, as observed in WB
experiments (Figure 3A,B). Interestingly, a 16 kDa CTF is also
produced from cells transfected with the APP A673T
mutant.14 The latter is a protective mutation found in
Icelandic populations.30 However, the significance of 16 kDa
CTF-141 from APP-WT or with AD mutations remains to be
understood.
Next, we examined the effects of 1a-analogues on β- and γ-
cleavages of APP by transfecting N2a cells with plasmids
expressing FL APP695 (APP-FL) and β-CTF (APP-C99) and
incubating with analogues 2a, 2b, 3a, and 3c. For a
comparison, we also used a BACEi (BACE IV) or a GSi
(DAPT) and determined both Aβ40 and Aβ42 peptides as well
as sAPPα and sAPPβ by measuring them using MSD ELISA.
The results are shown in Figures 3C,D and S3A. As expected,
BACE IV and DAPT reduced the Aβ levels in cells expressing
APP-FL, and DAPT also reduced Aβ in cells expressing APP-
C99 (Figure 3C). Compound 1a and its analogues also
reduced the Aβ levels in cells transfected with APP-FL, but not
as much in cells transfected with APP-C99. However, there is a
E
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 2. In vitro and in vivo evaluations of 1a-analogues. (A) Compounds 2a−b and 3a−c reduced production of Aβ40 and Aβ42 peptides in
N2a695 cells with an apparent IC50 between 10 and 5μM. DMSO and 1a were used as the negative and positive controls, respectively. Results
presented in (A) are the average of two or three experiments. (B) Compounds 1a and 3a−3c are nontoxic to N2a cells at 10 μM concentration,
whereas 2a−2b slightly increased cell density in the 5 h assay. (C) Brain and serum concentration of 1a-analogues as determined in brain and
serum extracts (n = 3 or 4) using liquid chromatography (LC)−mass spectrometry (MS)/MS. Statistical significance (P): *, <0.05%; **, <0.01%;
and ***, <0.001%.
clear trend of Aβ reduction in the latter case also, where
compounds 2b had shown a ∼25% reduction of Aβ40 or of
both Aβ40 and Aβ42 in APP-C99transfected cells. Because
neither 1a nor its five analogues affected BACE1 activity, in
vitro, at 10 μM concentration (Figure S3B), the reduced β-
cleavage of APP could be due to an increase in alternative
cleavages of APP that produce larger CTFs, including CTF-
141 or could be the result of increased trafficking of APP to
lysosomes, thereby reducing APP exposure to secretases.14
Production of the large CTFs might also account for the
observed reduction of both sAPPβ (by 40−75%) and sAPPα
(by 20−45%) levels (Figure 3D). An increase in APP αCTF
levels in 1a and analogue-treated samples suggests that the
larger CTFs preferentially undergo α-cleavage, possibly
because of a reduction in the competing BACE cleavage. An
increase in the APP level (with compound 3c) could be due to
a decrease in APP metabolism.
serum concentration than the latter. Third, when we fed WT
mice with an increasing amount, up to 150 mg of compound
2a orally (gavage) for five days, we did not notice any toxicity.
In fact, all mice appeared more active. Additionally, we were
able to design and achieve a multigram synthesis of compound
2a using relatively cheap starting materials without any
chromatographic separation.
2.4. Compound 2a Possessed Metabolic Stability
Comparable to the Parent Compound, 1a. We
determined and compared the metabolic stabilities of 1a and
2a by adding these to mouse or human liver microsomes and
measuring the remaining unmodified 1a and 2a in reaction
mixtures.31 In brief, 1a and 2a were incubated at two different
concentrations, 1 and 10 μM, with either mouse or human
liver microsomes, and the remaining amounts of 1a and 2a
were determined after 30 and 60 min by high-pressure liquid
chromatography (HPLC)−MS/MS analysis.32 The results
shown in Figure 4A corroborated the prior report that
compound 1a metabolizes more readily in mice than in
humans.33 We have now found that compound 2a also
metabolizes more readily in mice than in humans. Moreover,
both compounds 1a and 2a possess comparable metabolic
We selected compound 2a to carry out a pilot in vivo study
for several reasons. First, compounds 2a and 3a had shown
that they produced the 16 kDa CTF-141 significantly higher
than 1a and the other three compounds, 2b and 3b−c. Second,
both compounds 2a and 3a accumulated in mouse brain in
similar quantity, but the former had a significantly higher
F
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. Effects of 1a-analogues on APP metabolism modeled in N2a695 cells and in N2a cells transiently expressing human APP full length
(APP-FL) or APP-C99. Shown are: WB analysis of N2a695 cell lysates post compound treatment shows formation of the 16 kDa (CTF-141) and
10 kDa CTF and αCTF in [(A) the labels are not in chronological order] and quantification of CTF-141 and the 10 kDa CTF bands produced in
(B). Note that the 10 kDa CTF band runs just below the βCTF (12 kDa). Shown in (C and D) are: Aβ40, sAPPα, and sAPPβ levels in media from
cells, transfected with plasmids for APP-FL or APP-C99 and treated with 1a or its analogues (10 μM), BACEi (BACE IV, 2 μM), or GSi (DAPT, 1
μM). The Aβ40, Aβ42 (data shown in Figure S3A), sAPPα, and sAPPβ levels were measured using the MSD ELISA. Results presented here are the
representatives of two or three independent experiments. Statistical significance (P): *, <0.05%; **, <0.01%.
stability, and compound 2a is suitable for studies in humans, if
found to effectively reduce Aβ levels in an AD mouse model.
2.5. Compound 2a Reduced Aβ Levels in 2xTg AD
Mouse Brain. To determine the effects of 2a on Aβ
production/secretion, we performed a pilot study using the
2xTg AD mouse model. The latter harbors the APP Swedish
and PS1ΔE9 mutations, and begins to form amyloid plaques in
the brain by 6 months of age with abundant plaques present at
9 months.34 We first determined the dosing requirements by
administering 2a to wild-type mice and measuring its
accumulation in the brain and plasma over a 24 h period.
Compound 2a (50 mg/kg as a mesylate salt solution in water
containing 0.1% methylcellulose) was administered by gavage,
and its concentration measured in the brain and plasma
extracts, as described previously using LC−MS/MS, 2, 4, 8, 12,
and 24 h post gavage. As shown in Figure S4, 2a was detected
in nearly steady amounts during the first 6 h in the brain and
12 h in plasma, but it was completely washed out at the 24-h
G
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 4. In vitro and in vivo evaluation of compound 2a. (A) Microsomal stability of 1a and its analog 2a, measured by incubating these
compounds with mouse and human liver microsomes, and determining the remaining compounds periodically using LC−MS/MS. (B) Analog 2a
reduced Aβ40 and 42 levels in 2xTg AD mice. Effects of 2a on insoluble Aβ40/42 were determined in 5-month old 2xTg AD mice (n = 6 or 7).
(C) Metabolic demethylation of 2a affords 2c via 2d. (D) Metabolite 2d is detected in the brain and plasma extracts of AD mice by performing
LC−MS/MS analysis at the end of the 5-day in vivo study, and (E) compounds 2c and 2d reduce Aβ40 and/or Aβ42 weakly in N2a695 cells.
Statistical significance (P): *, <0.05%; **, <0.01%.
time point. Because compound 2a had shown no toxicity when
it was administered orally (gavage), 150 mg/kg once a day for
5 days, we chose to administer 2a-mesylate as aqueous solution
to AD female mice (n = 7) at an elevated dose, 100 mg/kg
twice daily with at least 8 h between the two administrations
for 4 days and once on day 5. We obtained TBS-soluble and
formic acid (FA)-soluble Aβ by performing extraction of brain
tissue using 1% Triton-x100 TBS buffer (pH 7.4) and 70% FA,
respectively, as described,35 and determined the Aβ levels using
ELISA. We found that TBS-soluble Aβ levels were too low to
measure without large errors. In contrast, there were sufficient
amounts of FA-soluble Aβ40 and Aβ42, which could be
measured with great accuracy. As shown in Figure 4B, there
were large reductions in both Aβ40 and Aβ42 levels in the
drug-treated mice compared to the control mice, but only
Aβ42 was significantly reduced (p < 0.01).
We anticipated that compound 2a could undergo metabolic
oxidative demethylation and didemethylation to produce
demethylated and didemethylated products, 2d and 2c (Figure
4C), similarly to 1a that affords demethylated-1a besides other
metabolites.36 Therefore, we performed LC−MS/MS analysis
of the extracts of brain homogenates and plasma samples from
2a-treated AD mice to determine the levels of both 2a and the
potential metabolites, 2d and 2c. As shown in Figure 4D, we
found that approximately 15−20% of the drug in both brain
and plasma consisted of metabolite 2d. We did not examine
the amounts of 2c in these samples. To examine whether these
metabolites retain Aβ-lowering activity, we prepared these
compounds and evaluated them at 10 μM concentration in
N2a695 cells. Indeed, both metabolites, 2d and 2c, reduced the
levels of secreted Aβ42 peptide significantly (p < 0.05),
whereas compound 2c also reduced the levels of Aβ40 peptide
(Figure 4E). Thus, the observed activity of compound 2a in
AD mouse model, in vivo, could still be the combined effects
of all three compounds, that is, the parent drug, 2a, and its
metabolites, 2c−d.
2.6. Perspective. Accumulation of two key neurotoxic
agents, Aβ peptides and hyperphosphorylated tau (pTau)
protein, in an AD brain starts decades ahead of when the actual
sign of the disease becomes evident and is followed by
neuronal loss and dementia.37 Although aging is a major
causative factor of AD,38 the Aβ peptides and pTau protein
levels increase with age both in AD patients and in non-AD
people.39−41 It can be argued that reducing the production of
Aβ peptides and pTau protein starting early on in life could
improve cognition and dementia of both AD patients and of
non-AD individuals. Earlier, compound 1a has been shown to
prevent Abl-dependent tau phosphorylation42 and Aβ fibrils-
induced apoptosis in rat models.43 We anticipate that the 1a-
analogues, 2a−b and 3a−c, which inhibit Abl kinase potently
H
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 5. Effects of (A) 1a-analogues on Abl kinase inhibition (performed by an outsourced company: Luceome Biotechnologies, Tucson, AZ),
and (B) 2a on Aβ40 and Aβ42 production in combination with BACE IV and Abl and ic-Kit kinase inhibition. N2a695 cells were pretreated with
BACE IV for 1 h before treatment with 1a or 2a (and BACE IV) for 5 h. Aβ40 and Aβ42 levels were measured using commercially available ELISA
kits for human Aβ40 and Aβ42. Statistical significance (P): *, <0.05%; **, <0.01%, ***, <0.001%.
(Figure 5A), could also reduce Abl-dependent tau phosphor-
ylation. Separately, Netzer, et al., have shown that 1a reduces
production of Aβ peptides independent of Abl kinase
inhibition.13 By analogy, we argue that compounds 2a−b
and 3a−c could also be inhibiting production of Aβ peptides
independent of Abl kinase inhibition. Nevertheless, a direct
measurement of these compounds and any future 1a-analogues
in Abl knockout cells may be desirable to confirm that the
effects of 1a-analogues on Aβ production are indeed Abl
independent.
Another key feature of 1a is that it can be combined with a
BACEi to reduce production of Aβ peptides by lessening the
BACEi-related toxicity. BACE1 inhibition is considered a
highly desirable strategy for treating or preventing AD, which is
evident from the fact that multiple clinical trials with small
molecule BACEi’s are in progress.44 However, no BACEi has
yet been successful in clinical trials, as BACE1 has many
cellular substrates and by inhibiting processing of all or most of
these substrates, nonselective BACE1 inhibitors cause serious
side effects.45 We have previously shown that a combination of
1a and a BACEi, BACE IV, reduces Aβ levels in N2a695 cells
in a synergistic manner.14 We have now found that a
combination of compound 2a with BACE IV caused greater
reduction in the production of the Aβ peptide, than using
compound 2a alone, similarly to the 1a-BACE IV combina-
tion.14 We will continue screening 1a-analog/BACEi combi-
nations in our future studies to identify those that mediate
synergistic effects at low BACEi concentration.
female AD mice, compound 2a was found to have significantly
reduced brain Aβ42 levels, when given at 100 mg/kg twice
daily for 4.5 days. The observed effect of compound 2a on Aβ
production could indeed be the effect of both the parent drug,
2a, and its metabolites, 2c and 2d. This in vivo study provides
a basis to further examine compound 2a or other analogues in
larger studies for treating AD and for studies for prevention.
The second, a combination therapy using compound 2a and a
BACEi could provide maximum therapeutic benefit, while
reducing the effective dose of both BACEi and 2a (or its
analogues). This approach would require a potent brain-
permeable BACEi. Future development of 1a-analogues will
avoid functionalities known to possess pGP activity, and
include assays to determine pGP activity of such compounds
and Aβ levels in Abl knock-out cells.
4. EXPERIMENTAL SECTION
4.1. General Methods. All commercial chemicals and solvents
were reagent grade and used without further purification. All air-
sensitive reactions were performed under argon protection. Micro-
wave reactions were performed on a Biotage Initiator. Column
chromatography was performed using CombiFlash or Biotage
instruments and silica gel SNAP columns. Analytical thin layer
chromatography was performed on Merck 250 μM silica gel F254
plates, and preparative thin layer chromatography (PTLC) on Merck
1000 μM silica gel F254 plates obtained from EMD Millipore
corporation. Purity of final compounds was checked using Agilent
1260 Infinity Series II preparative HPLC equipped with a 1260
Infinity II Multiple wavelength detector and Gemini 5 μm C18 110 Å
LC column (100 × 4.6 mm), mobile phase: acetonitrile and water
(0.1% TFA), 80:20 isocratic. All final compounds were found ≥95%
pure. The identity of each product was determined using MS (MS:
Thermo Scientific LTQ XL; LC: Thermo Scientific Dionex UltiMate
3000) and NMR (Bruker 400 or 600 MHz instrument) using CDCl3
as the solvent unless otherwise mentioned. Chemical shifts are
reported in δ values in ppm downfield from TMS as the internal
3. CONCLUSIONS
In this study, we have developed chemical analogues of 1a that
possess comparable activities. The new analogues also function
similarly to the parent compound and reduce Aβ levels in
N2a695 cells primarily through reducing the β-cleavage of APP
and shifting APP metabolism toward a nonamyloidogenic
pathway. By performing an in vivo study in 5 months old 2xTg
1
standard. H data are reported as follows: chemical shift, multiplicity
I
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m =
multiplet), coupling constant (Hz), integration.
32.86, and 18.13. HRMS (ESI), m/z: 521.2264 [M + H]+. Purity
(HPLC): >98%.
4.2.3. N-(4-Methyl-3-((4-(pyridin-3-yl)pyrimidin-2-yl)amino)-phe-
nyl)-4-(4-methylpiperazin-1-yl)benzamide, 3a. Amine 4 (1.39, 5.0
mmol) was coupled with acid 10c (1.10 g, 5.0 mmol) similarly as
described above for 11a, using HATU (2.25 g, 6.0 mmol) in
anhydrous DMF (mL) and iPr2EtN (1.75 mL, 10 mmol) at 60 °C (2
h) than at RT overnight to afford 3a (1.20 g, 50%) after the usual
work-up and chromatographic separation using CombiFlash
4.2. Synthesis and Characterization of Compounds. 4.2.1. 4-
(2-(Dimethylamino)ethyl)-N-(4-methyl-3-((4-(pyridin-3-yl)-
pyrimidin-2-yl)amino)phenyl)-benzamide, 2a. DIEA (3.5 mL, 20
mmol) was added to a mixture of amine 4 (2.77 g, 10 mmol), acid
10a (1.85 g, 10 mmol), and HATU (4.94 g, 13 mmol) in anhydrous
dimethylformamide (DMF, 25 mL), and the mixture was stirred
overnight at room temperature (RT). The reaction mixture was
diluted with water, and the resulting solid 11a was collected by
filtration (3.15 g, 70%). A mixture of crude 11a and Me2NH (30 mL,
40% in water) in a pressure bottle was heated at 120 °C overnight.
The reaction mixture was cooled and filtered, and the resulting white
solid was washed with water. The solid was dissolved in MeOH,
coevaporated with toluene, and triturated with MeOH to afford 2a
(2.60 g 59% yield) upon filtration. 1H NMR (600 MHz) of 2a: δ 9.27
(s, 1H), 8.73 (d, J = 3.6 Hz, 1H), 8.61 (s, 1H), 8.55 (d, J = 4.2 Hz,
2H), 7.84 (d, J = 6.0 Hz, 2H), 7.45 (t, J = 6.60 Hz, 1H), 7.35 (d, J =
7.80 Hz, 2H), 7.32 (d, J = 7.80 Hz, 1H), 7.24−7.21 (m, 2H), 7.05 (s,
1H), 2.90 (t, J = 7.80 Hz, 2H), 2.62 (t, J = 7.80 Hz, 2H), 2.38 (s, 3H),
and 2.36 (s, 6H). 13C NMR (150 MHz, DMSO-d6): δ 165.78, 162.07,
161.66, 159.94, 151.85, 148.67, 144.92, 138.26, 137.72, 134.89,
133.17, 132.69, 130.48, 129.04, 128.04, 124.25, 117.66, 117.18,
107.97, 60.88, 45.22, 33.56, and 18.11. HRMS (ESI), m/z: 453.2398
[M + H]+. Purity (HPLC): >98%.
1
(CH2Cl2−MeOH, 15%). H NMR (400 MHz) of 3a: δ 9.19 (br s,
1H), 8.64 (d, J = 4.0 Hz, 1H), 8.50−8.46 (m, 2H), 8.42 (br s, 1H),
7.79 (t, J = 8.0 Hz, 3H), 7.62 (d, J = 8.0 Hz, 1H), 7.42 (t, J = 8.0 Hz,
1H), 7.31 (m, 1H), 7.17 (dd, J = 8.0, 4.0 Hz, 2H), 6.90 (d, J = 8.0 Hz,
1H), 3.39 (br s, 4H), 2.80 (br s, 4H), and 2.49 and 2.32 (s, 3H each).
13C NMR (150 MHz, DMSO-d6): δ 165.30, 162.06, 161.70, 159.93,
153.43, 151.84, 148.66, 138.18, 138.02, 134.89, 132.70, 130.39,
129.47, 129.47, 127.64, 124.24, 117.67, 117.17, 113.96, 107.91, 54.82,
47.40, 46.18, and 18.09. HRMS (ESI), m/z: 480.2496 [M + H]+.
Purity (HPLC): 96%.
4.2.4. N-(4-Methyl-3-((4-(pyridin-3-yl)pyrimidin-2-yl)amino)-
phenyl)-4-(1-methylpiperidin-4-yl)benzamide, 3b. Amine 4 (277
mg, 1.0 mmol) was coupled with acid 10e (220 mg, 1.0 mmol) using
HATU (494 mg, 1.3 mmol) and DIEA (0.350 mL, 1.3 mmol) in
anhydrous DMF (3.0 mL) as described above for 11a to afford 3b
1
(200 mg, 42%). H NMR (600 MHz, DMSO-d6) of 3b: δ 10.10 (s,
1H), 9.25 (s, 1H), 8.94 (s, 1H), 8.66 (d, J = 6.0 Hz, 1H), 8.48 (d, J =
6.0 Hz, 1H), 8.45 (d, J = 6.0 Hz, 1H), 8.05 (s, 1H), 7.86 (d, J = 6.0
Hz, 1H), 7.50 (td, J = 6.0 Hz, 1H), 7.46 (d, J = 6.0 Hz, 1H), 7.40 (d, J
= 6.0 Hz, 1H), 7.37 (d, J = 6.0 Hz, 1H), 7.18 (d, J = 12.0 Hz, 1H),
2.91 (br, 2H), 2.55 (m, 1H), 2.23 and 2.20 (s, 3H each), 2.04 (br,
2H), and 1.76−1.68 (m, 4H). 13C NMR (150 MHz, DMSO-d6): δ
165.75, 162.07, 161.66, 159.94, 151.85, 150.23, 148.67, 138.26,
137.72, 134.89, 133.48, 132.69, 130.48, 124.24, 128.02, 127.17,
124.25, 117.64, 167.15, 107.98, 55.97, 46.33, 41.25, 32.97, and 18.11.
HRMS (ESI), m/z: 479.2550 [M + H]+. Purity (HPLC): >98%.
4.2.5. N-(4-Methyl-3-((4-(pyridin-3-yl)pyrimidin-2-yl)amino)-
phenyl)-4-(1-methylpiperidin-3-yl)benzamide, 3c. Amine 4 (106.0
mg, 0.382 mmol) was coupled with acid 10g (128.0 mg, 0.42 mmol)
using HATU (172.0 mg, 0.42 mmol) and DIEA (0.097 mL, 0.76
mmol) in anhydrous DMF (3.0 mL) as described above for 11a to
afford the Boc-protected precursor (80 mg, 37%) of 3f. The latter was
deprotected using 4 M HCl in dioxane (0.500 mL) as above to afford
3f (50 mg, 60%) as a HCl salt. 1H NMR (400 MHz, CD3OD) of 3f: δ
9.26 (s, 1H), 8.61 (d, J = 3.7 Hz, 1H), 8.57 (d, J = 7.8 Hz, 1H), 8.45
(d, J = 4.9 Hz, 1H), 8.22 (s, 1H), 7.90 (d, J = 7.8 Hz, 2H), 7.52 (m,
1H), 7.39 (m, 3H), 7.33 (d, J = 5 Hz, 1H), 7.24 (d, J = 8.1 Hz, 1H),
3.18 (d, J = 10.8 Hz, 2H), 2.73−2.88 (m, 3H), 2.31 (s, 3H), 2.00 (m,
1H), 1.89 (m, 1H), and 1.73 (m, 2H). MS: m/z 465.3 [M + H]+.
Purity (HPLC): >98%.
Amine 3f (40 mg, 0.107 mmol) was methylated using
paraformaldehyde (10 mg, 0.300 mmol) and NaCNBH3 (20 mg,
0.30 mmol) in CH2Cl2 (3 mL) in a tightly-capped bottle to afford 3c
(10 mg, 20%) after the usual work-up and silica gel PTLC using
CH2Cl2−MeOH−aq NH3 as the mobile phase. 1H NMR (600 MHz,
DMSO-d6) of 3c: δ 10.18 (s, 1H), 9.28 (s, 1H), 8.98 (s, 1H), 8.68
(dd, J = 4.8, 1.2 Hz, 1H), 8.52 (d, J = 5.4 Hz, 1H), 8.48 (d, J = 8.4 Hz,
1H), 8.08 (s, 1H), 7.96 (d, J = 8.4 Hz, 2H), 7.52 (m, 1H), 7.48 (dd, J
= 7.8, 1.2 Hz, 1H), 7.44 (m, 1H), 7.22 (d, J = 8.4 Hz, 1H), 3.50 (d, J
= 10.8 Hz, 1H), 3.45 (d, J = 11.4 Hz, 1H), 3.14 (t, J = 11.4 Hz, 1H),
3.06 (m, 1H), 2.95 (m, 1H), 2.81 (s, 1H), 2.23 (s, 3H), 1.97 (d, J =
12.0 Hz, 1H), 1.92 (d, J = 12.0 Hz, 1H), 1.80 (m, 1H), and 1.66 (m,
1H). 13C NMR (150 MHz, DMSO-d6): δ 165.42, 162.07, 161.65,
159.96, 151.84, 148.67, 145.14, 138.29, 137.56, 134.89, 134.45,
132.68, 130.53, 128.55, 128.17, 127.57, 124.25, 117.70, 117.22,
108.11, 57.92, 53.70, 43.52, 29.07, 23.01, and 18.12. HRMS (ESI), m/
z: 479.2547 [M + H+]. Purity (HPLC): >98%.
4.2.2. 4-(2-(Dimethylamino)ethyl)-N-(4-methyl-3-((4-(pyridin-3-
yl)pyrimidin-2-yl)amino)phenyl)-2-(trifluoromethyl)benzamide,
2b. Acid 10b (1.10, 4.0 mmol) was coupled to 4 (1.10, mmol)
similarly as described above for 11a, using HATU (1.82 g, 4.8 mmol)
and iPr2EtN (1.5 mL, 8 mmol) in DMF (12 mL) to afford
1
intermediate 12a (1.6 g, 75%). H NMR (600 MHz) of 12a: δ 9.19,
(s, 1H), 8.65 (d, J = 6.0 Hz, 1H), 8.50 (d, J = 6.0 Hz, 1H), 8.49 (d, J =
6.0 Hz, 1H), 8.44 (s, 1H), 7.88 (s, 1H), 7.78 (d, J = 6.0 Hz, 1H), 7.54
(d, J = 12.0 Hz, 1H), 7.44 (dd, J = 6.0, 6.0 Hz, 1H), 7.30 (d, J = 6.0
Hz, 1H), 7.22 (d, J = 12.0 Hz, 1H), 7.19 (d, J = 6.0 Hz, 1H), and 2.35
(s, 3H). MS, m/z: 528.0 [M + H]+.
To a degassed mixture of intermediate 12a (528 mg, 1.0 mmol),
(E)-1-ethoxyethane-2-boronic acid pinacol ester (257 mg, 1.30
mmol), and Cs2CO3 2 M solution (1 mL) in 1,4-dioxane (3 mL/
mmol) in a microwave vial was added Pd(PPh3)4 (55 mg, 5 mol %).
After the reaction mixture was heated at 100 °C for 45 min using a
microwave, it was diluted with water and extracted with CH2Cl2. The
organic layer was adsorbed over silica gel and chromatographed
affording the ethoxyvinyl-coupled product, 13a (450 mg, 85%). MS:
m/z 520.4 [M + H]+.
TFA (1.0 mL) in DCM (20 mL) was added to the above-described
intermediate, and the mixture was stirred overnight at RT. Volatile
materials were removed under reduced pressure, and the resulting
aldehyde was taken to next step without further purification. MS: m/z
492.3 [M + H]+.
Me2NH (10 mL, 4 M in THF) was added to a solution of the
above-described aldehyde intermediate in CH2Cl2 (10 mL) in a
pressure bottle. After the mixture was stirred for 1 h at RT, it was
cooled and NaCNBH3 (130 mg, 2 mmol) was added, and stirring was
continued overnight. Solvents were removed, and the residues were
suspended in CH2Cl2 and washed with brine. Combined organic
layers were dried over anhydrous MgSO4, inorganic materials were
separated by filtration, organic layers were concentrated under
reduced pressure, and the residue was chromatographed over SiO2
gel using CH2Cl2−MeOH−aq NH3 to afford compound 2b (235 mg,
45% based on 13a). 1H NMR (600 MHz, CD3OD) of 2b: δ 9.13 (s,
1H), 8.60 (d, J = 3.3 Hz, 1H), 8.53 (d, J = 7.9 Hz, 1H), 8.45 (d, J =
2.0 Hz, 2H), 8.37 (s, 1H), 7.59 (s, 1H), 7.56 (d, J = 7.8 Hz, 1H), 7.53
(d, J = 7.6 Hz, 1H), 7.46 (dd, J = 7.8, 4.9 Hz, 1H), 7.35 (s, 1H), 7.26
(d, J = 7.8 Hz, 1H), 7.19 (d, J = 8.2 Hz, 1H), 7.18 (d, J = 5.2 Hz, 1H),
3.00 (br s, 4H), 2.62 (s, 6H), and 2.31 (s, 3H). 13C NMR (150 MHz,
DMSO-d6): δ 165.99, 162.03, 161.58, 159.97, 151.85, 148.65, 143.43,
138.36, 137.47, 134.88, 134.55, 133.07, 132.71, 130.62, 128.25,
126.79, 126.31, 125.24, 124.20, 123.42, 116.89, 116.45, 60.41, 45.40,
4.3. Evaluation of Compounds in Cells. 4.3.1. Screening and
Evaluation of 1a-Analogues. N2a695 cells were used to screen all
1a-analogues and in the follow-up studies with compounds found
active in the preliminary screen. Cells were cultured in 1:1 Opti-MEM
J
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Reduced Serum Media (Life Technologies): Dulbecco’s modified
Eagle medium ([+] 4.5 g/L D-glucose; [+] L-glutamine; [-] sodium
pyruvate (Life Technologies) supplemented with 5% fetal bovine
serum, 0.4% penstrep, and 0.4% geneticin and incubated at 37 °C in
5% CO2.
or through oral gavage to 8 weeks old C57BL/6J WT mice. The mice
were euthanized 4 h post drug administration, and the brain
hemispheres and plasma were harvested and collected in preweighted
tubes and snap-frozen in liquid nitrogen. To measure brain and
plasma concentrations of the specific compounds, mouse brain tissue
was homogenized and extracted using ethanol, and plasma samples
were extracted using acetonitrile. Concentration of the drug
compounds and metabolites both in brain and in plasma was
determined by LC-MS/MS analysis.
4.4.1.1. Drug Extraction from Brain. After the tubes were weighed
to calculate brain weight and thawed to RT, 1 mL of EtOH (200
Proof) was added to the microcentrifuge tubes containing the
harvested right brain hemispheres. The internal standard (ABG190, a
synthetic analog of 1a, 10 μL, 1 μM) was added to each tube, and
samples were sonicated to homogeneity (∼2 min). Tubes were
shaken at 40 min at RT (1 K rpm) and centrifuged for 8 min at 13 K
RPM. The supernatant (0.9 mL) was transferred to a new collection
tube, and 0.5 mL EtOH was added to the pellet for a second round of
extraction as described above. The supernatant (600 μL) was
combined with the first collection before samples were submitted
for LC−MS/MS analysis.
4.4.1.2. Drug Extraction from Blood. Acetonitrile (300 μL) was
added to the collected blood samples. The internal standard
(ABG190, 10 μL, 1 μM) was added to each tube, and the samples
were sonicated to homogeneity (∼2 min). The tubes were centrifuged
at 13 K RPM for 9 min. The supernatant (300 μL) was collected and
combined with 500 μL of 5 mM ammonium formate before the
samples were submitted for LCMS-MS analysis.
4.4.2. In Vivo Evaluation of Compound 2a in 2xTg AD Mice.
Female mice expressing the human FAD mutations APP KM670/
671NL (Swedish) and PSEN1 M146V were used in this study. Mice
were obtained by performing in vitro fertilization, in house, and those
positive for APP Swedish and PS1ΔE9 genes were identified by PCR-
based genotyping. The WT C57BL/6J mice (The Jackson
Laboratory; stock no. 000664) were used for generating 2xTg
mouse pups from in vitro fertilization.
4.4.2.1. Evaluation. 2xTg AD female mice (n = 7 for each group)
received 100−125 μL solution (based on mouse weight) via oral
administration (gavage) of 2a-mesylate (20 mg/mL in water
containing 0.1% methylcellulose, 100 mg/kg/mouse) or with vehicle
(water containing 0.1% methylcellulose, 100−125 μL) twice a day, 8
h apart in the morning and afternoon, for 4 days and in the morning
on day 5. All mice were euthanized 4 h after the last injection. Brain
hemispheres and plasma were harvested, collected in preweighted
tubes, and the tubes were snap-frozen in liquid nitrogen. On the day
of analysis, tubes containing the left-brain hemispheres were weighed
after they were defrosted. Brain mass was submerged in 1% Triton-
x100 TBS Buffer (pH 7.4) (2 mL), homogenized using a Dounce
homogenizer and transferred to ultraclear centrifuge tubes (Beckman
Coulter) in which they were ultracentrifuged at 40 000 RPM for 35
min (TLA. 100.3 100 K RPM rotor). Supernatants were collected to
measure soluble Aβ peptides. Insoluble Aβ was extracted from the
pellets with 70% FA (1 mL) and 20 μL aliquots were neutralized with
200 μL Bis−Tris buffer. Aβ 1−40 and Aβ 1−42 were measured using
ELISA (Invitrogen) in accordance with the manufacturer’s
instructions.
In a typical experiment, 6-well tissue culture plates (Corning) were
seeded at 4.0 × 105 to 4.5 × 105 N2a695 cells/mL, 2 mL/well for
overnight incubation. Media were carefully removed and fresh media
containing 10 μM solution of compounds (prepared from 10 mM
solution in DMSO) were gently layered onto adherent cells (>95%
confluent). Appropriate dilution of compounds was performed in
DMSO before adding to media. After cells were incubated with
compounds for 5 h at 37 °C in 5% CO2, culture media were collected.
To measure soluble Aβ concentrations in the culture media, these
were transferred to strips of a 96-well ELISA plate for human Aβ40
peptide or to a 96-well V-PLEX Plus MSD (Mesoscale Discovery)
plate for Aβ Peptide Panel 1 (6E10) Kit (Catalog number K15200G)
and processed as per manufacturer instructions. Signals for Aβ were
measured using a PerkinElmer EnVision and SQ120 MSD ELISA
reader. Follow-up studies with N2a695 cells were performed similarly.
4.3.2. In Vitro Metabolic Stability of 1a-Analogues. Compound
1a and analogues (1 and 10 μM) were incubated with human or CD-
1 mouse liver microsomes (0.5 mg protein/mL) and appropriate
cofactors (2.5 mM NADPH and 3.3 mM MgCl2) in 100 mM
phosphate buffer, pH 7.4, at 37 °C. The incubation contained a final
organic solvent concentration of 0.1% DMSO. Reactions were started
with the addition of an NADPH/MgCl2 mix and stopped by
removing 100 μL aliquots at selected time points (0, 30, and 60 min)
and mixing with 200 μL of acetonitrile. Following brief vortexing and
centrifugation, an aliquot of the supernatant was transferred to a tube
and further diluted 2-fold (for 1 μM, using 10% acetonitrile
containing internal standards) or 20-fold (for 10 μM, using 37%
acetonitrile containing internal standards). Subsequently, amounts of
unmodified 1a or analogues in the quenched reactions were
determined using LC-MS/MS analysis.
4.3.3. Effects of 1a-Analogues on APP Metabolism. N2a695 cells
were treated with compounds for 5 h as described above, and media
were aspirated out (or collected for determination of Aβ levels). Cells
were scraped in cold Dulbecco’s phosphate-buffered saline (PBS)
buffer (1 mL) containing mini ethylenediaminetetraacetic acid-free
protease inhibitor (Roche) and centrifuged for 1 min at 13 000 rpm at
4 °C to form a cell pellet. The buffer was aspirated, and the cell pellets
were lysed in 3% SDS by sonication for two rounds of 20 s on a low
setting. Protein concentrations were measured using the Pierce BCA
Protein Assay (Thermo Fisher) Kit in accordance with the
manufacturer’s instructions.
To perform WBs, N2a695 cell lysates from 1a and analogue-treated
samples were run on a 16.5% Tris-Tricine gel (Criterion) and
electrotransferred to PVDF membranes (EMD Millipore) overnight
at 30 V. The PVDF membranes were incubated in PBS containing
0.25% glutaraldehyde (Sigma) for 30 min after electrotransference,
blocked for 30 min in milk PBST, incubated with primary antibody
RU369 for 1 h at RT followed by washing and incubation with an
HRP-linked secondary antibody, and detected with enhanced
chemiluminescence ECL reagents. WB images were analyzed using
ImageJ to quantify the prominent bands.
4.4.3. Statistical Analysis. The data are presented as means
SEM. The data were analyzed by Student’s t-test for single
comparison and one-way ANOVA for multiple comparisons, and
those showing P value < 0.05 were considered significant.
4.4.4. In Vitro Kinase Activity Assay. The assay was performed by
Luceome Biotechnolgies, LLC, using the general methods, as
described.46 Typically, 10 mM stock solutions of the compounds
were diluted in DMSO to a concentration of 250 μM. Before initiating
the assay, all test compounds were evaluated for a false positive
against split-luciferase. For kinase assays, each Cfluc-Kinase was
translated along with Fos-Nfluc using a cell-free system (rabbit
reticulocyte lysate) at 30 °C for 90 min. An aliquot (24 μL) of this
lysate containing either 1 μL of DMSO (for no-inhibitor control) or
compound solution in DMSO (10 μM final concentration) was
To determine effects of the compounds on BACE1 versus GS
inhibition, we used N2a cells transfected with FL APP (APP-FL) or
with APP99 (APP-βCTF). After 48 h, media were removed and fresh
media containing compound 1a and analogues were added. Following
5 h of incubation, cell supernatants were collected, and analyzed using
MSD-ELISA for Aβ and for sAPPα and β.
4.4. Evaluation of Compounds Using Mouse Models, in
Vivo. All procedures involving animals were approved by The
Rockefeller University Institutional Animal Care and Use Committee
and were in accordance with the National Institutes of Health
guidelines.
4.4.1. In Vivo Brain Permeability and Retention of 1a and
Analogues. Mesylate salts of 1a analogues (1 or 3 mg/mL in water,
125 μL, 50 or 150 mg/kg) were administered intraperitoneally (i.p.)
K
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
incubated for 30 min at RT followed by 1 h in the presence of a
kinase-specific probe. Luciferin assay reagent (80 μL) was added to
each solution, and luminescence was immediately measured on a
luminometer. The percent inhibition was calculated using the
following equation: % inhibition = (ALUcontrol − ALUsample × 100)/
De Strooper, B. Neurotoxicity of Alzheimer’s disease Aβ peptides is
induced by small changes in the Aβ42 to Aβ40 ratio. EMBO J. 2010,
29, 3408−3420.
(3) Hardy, J.; Selkoe, D. J. The amyloid hypothesis of Alzheimer’s
disease: progress and problems on the road to therapeutics. Science
2002, 297, 353−356.
ALUcontrol
.
(4) Golde, T. E.; Eckman, C. B.; Younkin, S. G. Biochemical
detection of Aβ isoforms: implications for pathogenesis, diagnosis,
and treatment of Alzheimer’s disease. Biochim. Biophys. Acta, Mol.
Basis Dis. 2000, 1502, 172−187.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jmed-
chem.8b02007.
■
S
(5) Haass, C. Take five-BACE and the γ-secretase quartet conduct
Alzheimer’s amyloid β-peptide generation. EMBO J. 2004, 23, 483−
488.
General synthetic procedures and characterization of
synthesized compounds including H NMR spectra and
(6) Annaert, W.; De Strooper, B. A cell biological perspective on
Alzheimer’s disease. Annu. Rev. Cell Dev. Biol. 2002, 18, 25−51.
(7) Nikolaev, A.; McLaughlin, T.; O’Leary, D. D. M.; Tessier-
Lavigne, M. APP binds DR6 to trigger axon pruning and neuron death
via distinct caspases. Nature 2009, 457, 981−989.
(8) Lauritzen, I.; Pardossi-Piquard, R.; Bourgeois, A.; Pagnotta, S.;
Biferi, M.-G.; Barkats, M.; Lacor, P.; Klein, W.; Bauer, C.; Checler, F.
Intraneuronal aggregation of the β-CTF fragment of APP (C99)
induces Aβ-independent lysosomal-autophagic pathology. Acta Neuro-
pathol. 2016, 132, 257−276.
1
MS data, truncated Gleevec analogues containing 3 or 4
rings, Gleevec analogues containing 4 or 5 rings and
ABC rings identical to 1a and different from 1a, and
figures showing screening data of all synthetic analogues
and evaluation of selected compounds (PDF)
SMILES data for compounds 1a−l, 2a−j, 3a−i, 4a−h, 5,
6a−g, 7a−b, 8a−q, and 9a−u (CSV)
(9) Cai, H.; Wang, Y.; McCarthy, D.; Wen, H.; Borchelt, D. R.;
Price, D. L.; Wong, P. C. BACE1 is the major β-secretase for
generation of Aβ peptides by neurons. Nat. Neurosci. 2001, 4, 233−
234.
(10) Capdeville, R.; Buchdunger, E.; Zimmermann, J.; Matter, A.
Glivec (STI571, imatinib), a rationally developed, targeted anticancer
drug. Nat. Rev. Drug Discovery 2002, 1, 493−502.
(11) Cohen, M. H.; Williams, G.; Johnson, J. R.; Duan, J.; Gobburu,
J.; Rahman, A.; Benson, K.; Leighton, J.; Kim, S. K.; Wood, R.;
Rothmann, M.; Chen, G.; U, K. M.; Staten, A. M.; Pazdur, R.
Approval summary for imatinib mesylate capsules in the treatment of
chronic myelogenous leukemia. Clin. Cancer Res. 2002, 8, 935−942.
(12) Dagher, R.; Cohen, M.; Williams, G.; Rothmann, M.; Gobburu,
J.; Robbie, G.; Rahman, A.; Chen, G.; Staten, A.; Griebel, D.; Pazdur,
R. Approval summary: Imatinib mesylate in the treatment of
metastatic and/or unresectable malignant gastrointestinal stromal
tumors. Clin Cancer Res 2002, 8, 3034−3038.
AUTHOR INFORMATION
Corresponding Author
*E-mail: ssinha@rockefeller.edu. Phone: (212)327-8121.
ORCID
Subhash C. Sinha: 0000-0001-8916-5677
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
■
We are thankful to Dr. Victor Bustos of the Rockefeller
University for helpful discussions, and to Mondana H. Ghias
for determining Aβ of some compounds in cellular assays early
on. Funding support from JPB (#322 and #839 to SCS) is duly
acknowledged.
(13) Netzer, W. J.; Dou, F.; Cai, D.; Veach, D.; Jean, S.; Li, Y.;
Bornmann, W. G.; Clarkson, B.; Xu, H.; Greengard, P. Gleevec
inhibits -amyloid production but not Notch cleavage. Proc. Natl. Acad.
Sci. U.S.A. 2003, 100, 12444−12449.
(14) Netzer, W. J.; Bettayeb, K.; Sinha, S. C.; Flajolet, M.;
Greengard, P.; Bustos, V. Gleevec shifts APP processing from a β-
cleavage to a nonamyloidogenic cleavage. Proc. Natl. Acad. Sci. U.S.A.
2017, 114, 1389−1394.
(15) Henley, D. B.; Sundell, K. L.; Sethuraman, G.; Dowsett, S. A.;
May, P. C. Safety profile of semagacestat, a gamma-secretase
inhibitor: IDENTITY trial findings. Curr. Med. Res. Opin. 2014, 30,
2021−2032.
(16) Cummings, J. L.; Morstorf, T.; Zhong, K. Alzheimer’s disease
drug-development pipeline: few candidates, frequent failures.
Alzheimer’s Res. Ther. 2014, 6, 37.
ABBREVIATIONS
■
Aβ, β amyloid or amyloid β; AD, Alzheimer’s disease; APP,
amyloid precursor protein; Aq Me2NH, aqueous dimethyl-
amine; BACE1, β-secretase; BACEi, β-secretase inhibitor;
BBB, blood−brain-barrier; Boc, butoxycarbonyl; βCTF, β-
carboxy terminal fragment; CH2Cl2, dichloromethane;
(CH2O)n, paraformaldehyde; Cs2CO3, cesium carbonate;
DIEA, diisopropyethylamine; DMF, dimethylformamide;
DMSO, dimethylsufoxide; ELISA, enzyme-linked immuno-
sorbent assay; EtOAc, ethylacetate; GS, γ-secretase or gamma
secretase; GSi, γ-secretase inhibitor; HATU, 1-[bis-
(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]-
pyridinium 3-oxide hexafluorophosphate; HCl, hydrochloric
acid; K2CO3, potassium carbonate; MW, microwave;
NaCNBH3, sodium cyanoborohydride; Pd(PPh3)4, tetrakis-
(triphenylphosphine)palladium(0); pGP, p-glycoprotein; TFA,
trifluoroacetic acid; THF, tetrahydrofuran
(17) Harrison, R. K. Phase II and phase III failures: 2013-2015. Nat.
Rev. Drug Discovery 2016, 15, 817−818.
́
́
(18) Godyn, J.; Jonczyk, J.; Panek, D.; Malawska, B. Therapeutic
strategies for Alzheimer’s disease in clinical trials. Pharmacol. Rep.
2016, 68, 127−138.
REFERENCES
(19) Zuhl, A. M.; Nolan, C. E.; Brodney, M. A.; Niessen, S.;
Atchison, K.; Houle, C.; Karanian, D. A.; Ambroise, C.; Brulet, J. W.;
Beck, E. M.; Doran, S. D.; O’Neill, B. T.; Am Ende, C. W.; Chang, C.;
Geoghegan, K. F.; West, G. M.; Judkins, J. C.; Hou, X.; Riddell, D. R.;
Johnson, D. S. Chemoproteomic profiling reveals that cathepsin D off-
target activity drives ocular toxicity of β−secretase inhibitors. Nat.
Commun. 2016, 7, 13042.
■
(1) Younkin, S. G. The role of Aβ42 in Alzheimer’s disease. J.
Physiol. 1998, 92, 289−292.
(2) Kuperstein, I.; Broersen, K.; Benilova, I.; Rozenski, J.;
Jonckheere, W.; Debulpaep, M.; Vandersteen, A.; Segers-Nolten, I.;
Van Der Werf, K.; Subramaniam, V.; Braeken, D.; Callewaert, G.;
Bartic, C.; D’Hooge, R.; Martins, I. C.; Rousseau, F.; Schymkowitz, J.;
L
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(20) Dai, H.; Marbach, P.; Lemaire, M.; Hayes, M.; Elmquist, W. F.
Distribution of STI-571 to the brain is limited by P-glycoprotein-
mediated efflux. J Pharmacol Exp Ther 2003, 304, 1085−92.
(21) Breedveld, P.; Pluim, D.; Cipriani, G.; Wielinga, P.; van
Tellingen, O.; Schinkel, A. H.; Schellens, J. H. M. The effect of Bcrp1
(Abcg2) on the in vivo pharmacokinetics and brain penetration of
Imatinib mesylate (Gleevec): Implications for the use of breast cancer
resistance protein and P-glycoprotein inhibitors to enable the brain
penetration of imatinib in patients. Cancer Res. 2005, 65, 2577−2582.
(22) Arslanova, D.; Yang, T.; Xu, X.; Wong, S. T.; Augelli-Szafran, C.
E.; Xia, W. Phenotypic analysis of images of zebrafish treated with
Alzheimer’s gamma-secretase inhibitors. BMC Biotechnol. 2010, 10,
24.
(36) Nebot, N.; Crettol, S.; d’Esposito, F.; Tattam, B.; Hibbs, D. E.;
Murray, M. Participation of CYP2C8 and CYP3A4 in the N-
demethylation of imatinib in human hepatic microsomes. Br. J.
Pharmacol. 2010, 161, 1059−1069.
(37) Jack, C. R.; Knopman, D. S.; Jagust, W. J.; Petersen, R. C.;
Weiner, M. W.; Aisen, P. S.; Shaw, L. M.; Vemuri, P.; Wiste, H. J.;
Weigand, S. D.; Lesnick, T. G.; Pankratz, V. S.; Donohue, M. C.;
Trojanowski, J. Q. Tracking pathophysiological processes in
Alzheimer’s disease: an updated hypothetical model of dynamic
biomarkers. Lancet Neurol. 2013, 12, 207−216.
(38) Guerreiro, R.; Bras, J. The age factor in Alzheimer’s disease.
Genome Med. 2015, 7, 106.
(39) Rodrigue, K. M.; Kennedy, K. M.; Devous, M. D.; Rieck, J. R.;
Hebrank, A. C.; Diaz-Arrastia, R.; Mathews, D.; Park, D. C. β-
Amyloid burden in healthy aging: Regional distribution and cognitive
consequences. Neurology 2012, 78, 387−395.
(23) Raub, T. J. P-Glycoprotein recognition of substrates and
circumvention through rational drug design. Mol. Pharm. 2006, 3, 3−
25.
(40) Davis, D. G.; Schmitt, F. A.; Wekstein, D. R.; Markesbery, W.
R. Alzheimer neuropathologic alterations in aged cognitively normal
subjects. J. Neuropathol. Exp. Neurol. 1999, 58, 376−388.
(41) Bennett, D. A.; Schneider, J. A.; Arvanitakis, Z.; Kelly, J. F.;
Aggarwal, N. T.; Shah, R. C.; Wilson, R. S. Neuropathology of older
persons without cognitive impairment from two community-based
studies. Neurology 2006, 66, 1837−1844.
(24) Deadman, B. J.; Hopkin, M. D.; Baxendale, I. R.; Ley, S. V. The
synthesis of Bcr-Abl inhibiting anticancer pharmaceutical agents
imatinib, nilotinib and dasatinib. Org. Biomol. Chem. 2013, 11, 1766−
1800.
(25) Miyaura, N.; Suzuki, A. Palladium-catalyzed cross-coupling
reactions of organoboron compounds. Chem. Rev. 1995, 95, 2457−
2483.
(42) Gardner, L. E.; White, J. D.; Eimerbrink, M. J.; Boehm, G. W.;
Chumley, M. J. Imatinib methanesulfonate reduces hyperphosphor-
ylation of tau following repeated peripheral exposure to lip-
opolysaccharide. Neuroscience 2016, 331, 72−77.
(26) Martin, R.; Buchwald, S. L. Palladium-catalyzed Suzuki−
Miyaura cross-coupling reactions employing dialkylbiaryl phosphine
ligands. Acc. Chem. Res. 2008, 41, 1461−1473.
(27) Whelligan, D. K.; Thomson, D. W.; Taylor, D.; Hoelder, S.
Two-step synthesis of aza- and diazaindoles from chloroamino-N-
heterocycles using ethoxyvinylborolane. J. Org. Chem. 2010, 75, 11−
15.
(28) Bihorel, S.; Camenisch, G.; Lemaire, M.; Scherrmann, J.-M.
Influence of breast cancer resistance protein (Abcg2) and P-
glycoprotein (Abcb1a) on the transport of imatinib mesylate (Gleevec
) across the mouse blood-brain barrier. J. Neurochem. 2007, 102,
1749−1757.
(29) Netzer, W. J.; Powell, C.; Nong, Y.; Blundell, J.; Wong, L.; Duff,
K.; Flajolet, M.; Greengard, P. Lowering β-amyloid levels rescues
learning and memory in a down syndrome mouse model. PLoS One
2010, 5, No. e10943.
(43) Cancino, G. I.; Toledo, E. M.; Leal, N. R.; Hernandez, D. E.;
́
Yevenes, L. F.; Inestrosa, N. C.; Alvarez, A. R. STI571 prevents
apoptosis, tau phosphorylation and behavioural impairments induced
by Alzheimer’s β-amyloid deposits. Brain 2008, 131, 2425−2442.
(44) Vassar, R. BACE1 inhibitor drugs in clinical trials for
Alzheimer’s disease. Alzheimer’s Res. Ther. 2014, 6, 89.
(45) Yan, R. Physiological functions of the beta-site amyloid
precursor protein cleaving enzyme 1 and 2. Front. Mol. Neurosci. 2017,
10, 97.
(46) Jester, B. W.; Cox, K. J.; Gaj, A.; Shomin, C. D.; Porter, J. R.;
Ghosh, I. A coiled-coil enabled split-luciferase three-hybrid system:
applied toward profiling inhibitors of protein kinases. J. Am. Chem.
Soc. 2010, 132, 11727−11735.
(30) Jonsson, T.; Atwal, J. K.; Steinberg, S.; Snaedal, J.; Jonsson, P.
V.; Bjornsson, S.; Stefansson, H.; Sulem, P.; Gudbjartsson, D.;
Maloney, J.; Hoyte, K.; Gustafson, A.; Liu, Y.; Lu, Y.; Bhangale, T.;
Graham, R. R.; Huttenlocher, J.; Bjornsdottir, G.; Andreassen, O. A.;
̈
Jonsson, E. G.; Palotie, A.; Behrens, T. W.; Magnusson, O. T.; Kong,
A.; Thorsteinsdottir, U.; Watts, R. J.; Stefansson, K. A mutation in
APP protects against Alzheimer’s disease and age-related cognitive
decline. Nature 2012, 488, 96−99.
(31) Li, D.; Edward, H. K.; Xuewen JoAnn, M.; Youping, H.; Guy,
T. C. Applications of high throughput microsomal stability assay in
drug discovery. Comb. Chem. High Throughput Screening 2008, 11,
469−476.
(32) Dahal, U. P.; Jones, J. P.; Davis, J. A.; Rock, D. A. Small
molecule quantification by liquid chromatography-mass spectrometry
for metabolites of drugs and drug candidates. Drug Metab. Dispos.
2011, 39, 2355−2360.
(33) Paniagua, R. T.; Sharpe, O.; Ho, P. P.; Chan, S. M.; Chang, A.;
Higgins, J. P.; Tomooka, B. H.; Thomas, F. M.; Song, J. J.; Goodman,
S. B.; Lee, D. M.; Genovese, M. C.; Utz, P. J.; Steinman, L.; Robinson,
W. H. Selective tyrosine kinase inhibition by imatinib mesylate for the
treatment of autoimmune arthritis. J. Clin. Invest. 2006, 116, 2633−
2642.
(34) Garcia-Alloza, M.; Robbins, E. M.; Zhang-Nunes, S. X.; Purcell,
S. M.; Betensky, R. A.; Raju, S.; Prada, C.; Greenberg, S. M.; Bacskai,
B. J.; Frosch, M. P. Characterization of amyloid deposition in the
APPswe/PS1dE9 mouse model of Alzheimer disease. Neurobiol. Dis.
2006, 24, 516−524.
́
(35) Izco, M.; Pesini, P.; Perez-Grijalba, V.; Fandos, N.; Sarasa, M.
Optimized protocol for amyloid-β extraction from the brain. J.
Alzheimer’s Dis. 2013, 34, 835−839.
M
DOI: 10.1021/acs.jmedchem.8b02007
J. Med. Chem. XXXX, XXX, XXX−XXX