2
58
M. K. Ameriks et al. / Bioorg. Med. Chem. Lett. 26 (2016) 257–261
O
R
Cl
variety of substituted benzyl bromides 17 (Scheme 4). Intermedi-
ate 16 derived from enantiopure 15, which in turn was assembled
CF3
N
O
Cl
N
N
in six steps from commercially available N-Boc-L-alaninol 14 as
N
CF3
N
O
N
H
16
previously described. Human and rat P2X7 FLIPR data for com-
pounds 11d–k are shown in Table 2. Although the 2,3-dichlorophe-
nyl analog 11e compares favorably with 11d, removing
substituents from the phenyl ring reduces potency against both
human and rat P2X7. However, comparison of 11d and 11f with
N
N
1
2
, R = H; h/r P2X7 IC50s = 2.1/92 nM
3, R = (S)-CH ; h/r P2X7 IC
h/r P2X7 IC50s = 3.2/320 nM
3
50s = 12/10 nM
1
1g reveals that monosubstitution in the meta-position of the
O
Cl
O
Cl
arene is sufficient to achieve single-digit nanomolar potency at
the human P2X7 receptor, while additional substitution in the
ortho-position can provide a synergistic improvement in rat func-
tional activity.
CF3
N
CF
3
N
N
N
N
N
N
N
3
N
T
N
Having established that the (2-chloro-3-trifluoromethyl)benzyl
group is a preferred lactam substituent, we held this moiety fixed
and optimized the 3-position of the triazole. To this end, com-
pounds 11l–t were prepared by initial activation of amide 13 as
an imidate or thioamide, followed by condensation with a variety
of aryl and heteroaryl hydrazides (Scheme 3). Due to the highly
lipophilic nature of the N-benzyl group in the 7-position, the
majority of the hydrazide inputs were selected to minimize LogP.
Gratifyingly, both five- and six-membered heteroaromatic
substituents are well tolerated by the human P2X7 receptor
(IC50s < 10 nM), but these compounds are also generally 100-fold
less potent against rat P2X7 (Table 3). Similar to the SAR trends
observed during the optimization of the 7-position, small struc-
tural changes in the 3-position dramatically affect rat P2X7 func-
tional activity (compare 11o with 11p, 11q with 11r, and 11s
with 11t). However, despite the consistent disparity in potency
between human and rat orthologs, four compounds (11d, 11m,
11q, and 11s) displayed sufficient P2X7 antagonism in the rat FLIPR
assay (IC50 < 100 nM) to warrant further evaluation.
N
N
4
5
h/r P2X7 IC50s = 2.3/1900 nM
[ H] JNJ-54232334
Figure 1. Aryl and N-benzyl amide based P2X7 antagonists.
silicon-based carbamate protecting group due to the poor solubil-
ity of 9 in the benzylic oxidation step.
Lactams 8, 11a, and 11b were evaluated for their ability to block
Bz-ATP induced Ca flux in 1321N1 glial cells expressing human
or rat P2X7 (Table 1). Although 1,2,3-triazolopyridone 8 is inac-
tive in both cell constructs (IC50s > 10 M), the analogous 1,2,4-
triazole isomer 11a is a sub-micromolar antagonist of the human
P2X7 receptor (IC50 = 440 nM). Furthermore, the introduction of a
methyl group in the 6-position of this scaffold provided >100-fold
improvement in potency for both human and rat P2X7 (compare
2
+
2
0
l
2
1
1
1a and 11b). Racemate 11b was separated into individual anti-
podes 11c and 11d using chiral HPLC, and subsequent profiling
revealed that only the (À) enantiomer 11d is a potent P2X7R
antagonist. The absolute configuration of 11d was established by
asymmetric synthesis, starting from commercially available (S)-
All four lead compounds were predicted to be promising CNS
candidates based on their high MPO (multiparameter optimiza-
tion) scores (Table 4). This algorithm utilizes six calculated physic-
1
,2-diaminopropane 12 (Scheme 3). It is worth noting that this is
ochemical properties (cLogP, cLogD, MW, TPSA, HBD, and pK ) to
a
2
3
the same absolute stereochemistry previously determined for
related 6-substituted 1,2,4-triazolopyrazines containing an exo-
cyclic amide (see compound 3).16 However, there appear to be
important speciation differences in the SAR between the two ser-
rate compounds on a scale of zero to six. In general, experimental
ADME data support these favorable in silico evaluations, as all four
leads exhibit minimal turnover in human and rat liver microsomes,
along with low plasma protein binding and encouraging aqueous
solubility across a physiologically relevant pH range. However, in
the Caco-2 cell permeability assay, only pyrazine 11d shows com-
parable rates of transport in both directions (B–A/A–B = 1.1),
whereas the remaining three compounds are actively effluxed
(B–A/A–B > 10).
ies:
a methyl substituent in the (6S)-position dramatically
improves potency at the rat P2X7R for both chemotypes, but a sim-
ilar beneficial effect for the human ortholog is only observed
within the lactam series (compare compounds 2 vs 3 and 11a vs
2
2
1
1d). As a consequence, even though compound 11d is only
moderately potent against rat P2X7, it is one of the most potent
human P2X7 receptor antagonists reported to date.
In order to explore the SAR around the 7-position of 11d,
analogs 11e–k were prepared by alkylating lactam 16 with a
O
N
N
NBoc
R
NTeoc
R
c
N
a,b
NTeoc
R
N
N
N
N
N
N
O
2
N
N
N
N
N
9, R = H
10, R = (+)-CH3
N
a,b
NH *HCl
N
N
N
3 steps
ref 17
N
N
N
N
N
N
Cl
6
d
N
N
N
N
Cl
e
CF3
Br
O
Cl
O
Cl
O
Cl
CF
3
N
N
d
N
N
NH
CF3
CF3
N
N
N
N
c
N
7
N
N
N
N
N
R
R
N
N
N
11a,b
N
N
N
N
8
N
N
Scheme 2. Synthesis of compounds 11a and 11b. Reagents and conditions: (a)
1.25 M HCl/EtOH, DCM, rt (61%); (b) TeocSuc, iPr NEt, DMF, rt (57–89%); (c) cat
RuO , NaIO , MeCN/CHCl /H O, rt (29–44%); (d) TFA, DCM, rt (71–95%); (e) Cs CO
DMF, rt (63–79%).
Scheme 1. Synthesis of compound 8. Reagents and conditions: (a) HCO
60 °C (36%); (b) 6 N HCl, rt (94%); (c) 2-chloro-3-(trifluoromethyl)benzyl bromide,
Et N, DMF, rt (82%); (d) cat RuO , NaIO , MeCN/CHCl /H O, rt (43%).
2
H, Et
3
N,
2
1
2
4
3
2
2
3
,
3
2
4
3
2