L. J. Simons et al. / Bioorg. Med. Chem. Lett. 19 (2009) 654–657
R1
655
The in vitro data were obtained from two primary assays, BASSR
and BASST, which were designed to determine the inhibition of
amyloid fibril formation by the test compound. BASSR (Beta-Amy-
loid-Self-Seeding Radioassay) is a radiolabeled 125I-Ab(1-40) filtra-
tion assay. The BASST (Beta-Amyloid-Self-Seeding, Thioflavin T) is
an amyloid-specific fluorescence assay. The test compound was
incubated under the same conditions in both assays with Ab(1–
R1
CO2H
a
+
O2N
CHO
3
2
1
NO2
b
R1
R1
c
R2
40)/buffer (30 lM unlabeled Ab(1–40)/50 mM NaPi-150 mM NaCl,
F
1 M urea, 0.02% w/v NaN3, pH 7.5). In the BASSR assay, a tracer
amount of 125I-labeled Ab(1–40) was added to the assay. In the
N
NH2
H
4
6-42
CO2H
CO2H
R2
BASST assay, 5 lM ThioflavinT (ThT) was added after fibril forma-
5
tion.4 BASST reports on the formation of the b-sheet structure of
amyloid fibrils.5 This measure can be complicated by false positives
that result from the competition of test compounds with the
extrinsic fluorescent ThT probe for binding to amyloid fibrils.
Therefore, an additional assay, BASSR, was used to detect the inhi-
bition of peptide aggregation. BASSR measures the formation of the
Scheme 1. Reagents: (a) Pyr; (b) RaNi, H2, THF; (c) LHMDS, THF.
O
O
CHO
a
radiolabeled fibrils that do not pass through a 0.2 lm filter. These
+
two assays provided complementary information about the aggre-
gation state and fibrillar amyloid conformation with compound
present. Eight concentrations of each compound were used for
IC50 determinations that were calculated by logit transformation
of the titration curves. Variation of individual points was 10%.
The average IC50s of two separate experiments are reported. At
R3
NH2
NH2
R3
45
43
44
b
R4
c
30 lM of Ab(1–40) peptide, an inhibitor with an IC50 of 1 lM tar-
N
H
NH2
R3
R3
gets a species that is less than 3% of the total peptide concentra-
tion. These assays were designed to detect inhibitors at an early
stage in fibril formation with an amount of compound that would
be less than the Ab(1–40) peptide monomer concentration. An IC50
F
R4
CO2H
48-72
46
CO2H
47
of 20 lM is approximately equivalent to a ratio of 0.7 moles com-
Scheme 2. Reagents: (a) NaOH, EtOH or H2SO4, AcOH; (b) Et3SiH, TFA; (c) LiNH2,
pound/peptide monomer and thus, is borderline substoichiomet-
ric.4,5 (The observed IC50 is a product of the true affinity of the
compound, the stoichiometry of binding, and the concentration
of the target aggregation intermediate.)
In vivo efficacy studies were performed in Tg2576 mice that
over-express human bAPP695Swe. These mice show an age-related
increase in cerebral Ab deposition, and therefore, are suitable mod-
els for testing b-amyloid aggregation inhibitors.6 Bioavailable and
brain-penetrant compounds were delivered orally by mixture in
chow.
THF.
Table 1
Modifications of ethyl-linked derivatives
R1
6-19
N
H
CO2H
Our initial efforts in the ethyl-linked series produced a series of
compounds (Table 1) that showed acceptable activity in the BASST
assay, but were inactive in the BASSR assay.
By adding a nitro group to the anthranilic acid ring (Table 2),
potency increased significantly in both assays. However, the nitro
functionality is considered to be a risk for CNS compounds since
it may decrease brain-penetration of the compound or have poten-
tial toxicity. Therefore, other substituents on the anthranilic acid
ring were explored (Table 3).
When the carboxylic acid group was moved to the 4- and 5-
positions on the aromatic ring (entries 32 and 34, Table 3 vs entry
14, Table 1), a 5-fold decrease in potency in BASST was observed.
The methyl ester analog (entry 33) of entry 32 exhibited no activity
in either BASST or BASSR. Some of the most potent derivatives of
14 had a carboxylic acid group in the 6-position and an electron-
deficient anthranilic acid ring (see Table 2 and entries 37, 39 and
41 in Table 3).
Compound
R1
BASSR IC50
(
lM)
BASST IC50 (lM)
6
7
8
4-OMe
N-DHIQa
H
3,4-Me
2-Cl
3-Cl
4-Cl
3-F
3,4-Cl
2,4-Cl
3,5-Cl
3,4-F
4-F, 3-CF3
3-Cl, 4-Me
>100
>100
60
>100
1
10
10
5
19
3
70
2
51
15
2
2
1
9
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
88
10
11
12
13
14
15
16
17
18
19
a
N-decahydroisoquinoline.
By increasing the length of the alkyl linker by one methylene
unit (i.e., propyl-linked, Table 4), a significant increase in potency
was observed in the BASSR assay relative to the similarly substi-
tuted inactive ethyl-linked compounds (entries 50, 52, 54–56 in
Table 4 vs 8–12 in Table 1). In addition to this observed increase
in potency, another advantage of the propyl-linked analogs was
the consistent activity in both BASSR and BASST. However, substi-
tution on the anthranilic ring of the propyl-linked analogs with
electron withdrawing groups (Table 5, entries 65, 66) did not give
2-fluoro benzoic acid 5 in the presence of a strong base to give
ethyl-linked derivatives 6–42.
Propyl-linked derivatives (Scheme 2) were prepared via an acid-
or base-catalyzed Aldol reaction. Reaction of substituted aldehydes
43 and 4-aminoacetophenone 44 generated chalcone 45. Reduc-
tion of chalcone 45 gave propyl-linked anilines 46 that were subse-
quently coupled with 2-fluoro benzoic acids 47 to yield the
corresponding propyl-linked anthranilic acids 48–72.