3
86
C. Han et al. / Bioorg. Med. Chem. Lett. 25 (2015) 384–388
synthesized based on literature procedures (Routes A–C,
Table 1
a
4
0
Structures and activities of 2-benzyl analogs 9
Scheme 1). N-substituted thioquinolinones 7 (Route A) were syn-
thesized via alkylation of commercially available quinolinone 1
and subsequent treatment with phosphorous pentasulfide. The
desired DBPQ 4 and analogs 8a–c, 9a–n, 12m–n were obtained
via condensation with hydrazine followed by N-alkylation with
appropriate alkylating agent. Alternatively, (Route B) quinolinone
Ar1
O
N
N
N
1
was treated with phosphorous oxychloride to yield chloroquino-
line 10. The final DBPQs 11a–d and 12a–12l were obtained by the
reaction between chloroquinoline 10 and the corresponding alkyl-
ated hydrazine followed by N-alkylation. Lastly, isatin analogs
Ar1
pEC50 (±SEM)b
EC50 (nM) AChmax (±SEM)b
Compound
1
4a–c were synthesized by N-alkylation (Route C).
Initial benzyl deletion studies (Fig. 3) of DBPQ (4) demonstrated
4
9
9b
9c
(DBPQ)
a
2,5-di Me
H
Ph
2-FPh
3-FPh
4-FPh
2
Ph 6.32 (0.14)
<5.0
473
>10,000
40 (1.2)
48 (5.7)
weak PAM activity for the 5-benzyl quinilonone (8a), while the 2-
Inactive
6.00 (0.17)
Inactive
Inactive
6.11 (0.12)
Inactive
Inactive
<5.0
(
(
2,5-dimethylbenzyl)-pyrazolone analog 11a was inactive
>30 M), suggesting that the presence of both N-benzyl moieties
1010
783
36 (1.2)
47 (1.1)
9
9
9
9
d
e
f
l
is required for activity. With these initial starting points in hand,
we conducted a survey of 2-benzyl modifications (9) and their
impact on functional PAM activity at hM
tion assay as previously described (Table 1). N-Methyl analog 9a
exhibited similar activity as 8a with some diminution of the AChmax
response (68% vs 48%). Parent benzyl analog 9b, lacking the 2,6-
dimethyl found in the HTS hit, lead to an inactive compound. A
2-CH
3-CH
3
Ph
Ph
g
3
1
using a calcium mobiliza-
9h
9i
4-CH Ph
3
17
2-pyridyl
3-pyridyl
4-pyridyl
>10,000
6310
>10,000
59 (2.2)
65 (2.7)
54 (0.83)
9
9
j
k
5.20 (0.12)
<5.0
a
CHO-K1 cells stably expressing the human M
1
receptor (Ref. 17).
pEC50 and AChmax are average of three independent determinations performed
EC50 values reported as >10,000 nM the corresponding max-
imal percent ACh response is from the highest tested concentration of compound
30 M). All compounds ranked as ‘inactive’ did not induce a >10% increase in the
response at 30 M.
b
‘
mono-fluoro walk’ within the N-benzyl phenyl ring (9c–9e) affor-
in triplicate. For hM
1
ded a weak PAM for 2-fluoro isomer 9c (EC50 = 1010 nM); however,
the 3 and 4-substituted congeners were ineffective. Similarly, a
methyl scan of the N-benzyl phenyl ring demonstrated that the 2-
position was preferred over the 3 and 4-position, with the 2-tolyl
derivative PAM 9f leading to sub-micromolar functional potency
(
l
l
and ML137 series,25,37 and analog 12n bearing the N-linked benzyl
(
EC50 = 783 nM) and comparable AChmax relative to DBPQ (4). Exam-
26
ination of pyridyl ring isomers (9i–9k) demonstrated a significant
loss in potency; however, a notable increase in AChmax within the
series (AChmax 54–65%) was observed relative to DBPQ (4), with 3-
pyridyl 9j demonstrating weak potency (EC50 = 6310 nM).
In parallel with the above studies the 2-(2,5-dimethylbenzyl)
moiety was held constant and the 5-benzyl substituent surveyed
pyrazole regioisomer reported within the BQCA series. Within
the context of the 2-(2,5-dimethylbenzyl)-pyrazolone the 4-meth-
oxy substituent resulted in a slight loss (ꢀ1.7-fold) in potency
(12k). Interestingly, the hybrid analog 12l with an additional
2-flourine substituent was found to be functionally inactive.
Utilizing the 4-(1-methyl-1H-pyrazol-4-yl)benzyl substructure
from ML137 gave 12m. Although less potent than BQCA, ML137,
and DBPQ (ꢀ2–4-fold), 12m proved to have a ꢀ2.0-fold boost in
AChmax response relative to DBPQ (AChmax = 74%). Similarly,
N-linked benzyl pyrazole 12n exhibited slightly diminished potency
and enhanced apparent potentiation (AChmax = 80%). Thus, although
hybrids utilizing the 2-(2,5-dimethylbenzyl)-pyrazolone in the
presence of historical N-(5) benzyl substituents from prior studies
suggest a weak enhancement in cooperativity, these modifications
were not inherently beneficial in leading to additive SAR in regard
to potency.
In an effort to understand if the pyrazolo[4,3-c]quinolin-3(5H)-
one core might accommodate multiple binding modes at the
allosteric binding site through a ‘benzylic translocation’, we tested
a small set of individual analogs incorporating preferred pyrazole
biaryl fragment at the N-(2) or N-(5)-position (Fig. 4). As shown
in Figure 4, translocation of the benzyl to provide either the
4-(1-methyl-1H-pyrazol-4-yl)benzyl analogs 11c or 12a or the 4-
(1-methyl-1H-pyrazol-4-yl)benzyl derivative 11d proved incom-
(Table 2, series 12). Similar to the 2-benzyl series 9, potency was
not improved for series 12 derivatives relative to the original HTS
hit DBPQ (4). However, with the exception of the tolyl derivatives
(12e–12g) which were uniformly inactive, fluoro derivatives 12b–
1
2d and pyridyl derivatives 12h–12j exhibited potency below
1
0 lM, with pyridyl analogs displaying enhanced efficacy (AChmax
P75%). Fluoro derivatives 12b–12d generally had a lower AChmax
from 31% to 47%. Both fluoro-substitution
dyl derivatives within the N-benzyl motif have been utilized with
varying success in a number of other M PAM chemotypes (e.g., 2
and 3) and similarly these modifications proved tolerable, and in
,
15,38
and N-methyl pyri-
1
7
1
15,17,25–28,35
some cases resulting in enhanced AChmax
.
We next wished to impart additional bias between the N-(2) and
N-(5) benzyl pharmacophores within DBPQ through incorporation
of established potency enhancing substituents at the N-(5) benzyl
ring. These included the 4-methoxy found in BQCA to afford 12k, a
hybrid of ML137 and BQCA to produce the 4-methoxy-2-fluoro
derivative 12l, the ML137 comparator 12m bearing a 4-(1-methyl-
1
H-pyrazol-4-yl)benzyl moiety reported as beneficial in the BQCA
1
patible with M activity. In addition, within the context of a
recently developed and structurally distinct (1R,2R)-2-hydrox-
ycyclohexyl moiety at N-(2) with potency enhancing properties
3
3
N
NH
(11b), lack of a benzyl substitution at N-(5) continued to be detri-
O
mental for activity. In comparison to N-(2) analogs 11c and 12a,
progression of a series of related N-(5) analogs from hydrogen to
N
N
N
O
1
methyl to benzyl (8c, 9l–9m) afforded M PAMs with diminishing
potency from 1.2 to 10 M, thus establishing the N-(5) position as
l
N
8
a
H
the preferred site for the benzylic pharmacophore and furthermore
indicating limited flexibility at the allosteric site between these
benzylic N-(2) and N-(5) positional isomers. The preference for
the N-(5) position is also reinforced by the ML137 overlapping
hM1 EC50 > 10 µM
11a
ACh = 68%
hM1 Inactive
Figure 3. DBPQ benzyl deletion analogs 8a and 11a.