S. D. Kuduk et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2514–2518
2515
Complete Freud’s Adjuvant (CFA)-induced inflammatory thermal
hyperalgesia model and skin incision model of post-operative
pain.
first carried out on 8, followed by treatment with guanidine to pro-
duce 6-des-chloro amiloride derivatives 9a–k.
A variety of amines (>80) were employed in a library approach
to evaluate the SAR at the 5-position of the pyrazine, and results
for selected amiloride analogs are shown in Table 1. As noted pre-
viously, due to the lack of a high-throughput biochemical assay,
The lack of an effective high-throughput assay to screen for ASIC
inhibition has impaired identification of new small molecule ASIC3
inhibitors. To minimize the use of labor intensive electrophysiolog-
ical assays, we developed an alternative approach by preparing
more potent amiloride derivatives. As a first step, the MRL sample
repository was examined for analogs of amiloride. These deriva-
tives were subsequently evaluated on ASIC3 channels at a single
dose using an automated patch clamp assay.18 Compounds of
interest were followed up at additional doses. An overview of the
key findings is shown in Figure 2.
The 2-acylguandine and 3-amino groups on the pyrazine ring
were required for inhibition of ASIC3. Aryl or alkyl groups extend-
ing off of the guanidine led to decreases in efficacy. The 6-chloro
group was only slightly favored over just having hydrogen present
at this position. The 5-position seemed quite tolerant of change
and as a result became the primary region to explore SAR.19
The chemistry utilized to prepared amiloride derivatives is
shown in Scheme 1. Methyl ester 320 was treated with guanidine
in isopropanol to afford acylguanidine 4. Displacement of the 5-
chloro group in 4 with the appropriate amines, thiols, or phenols
affords amiloride analogs 5a–i0, 6a–l.
Palladium catalyzed cross couplings were investigated on 4 to
prepare 5-carbon-linked analogs, but proved to be unselective for
the chlorines at the 5- and 6-positions leading to a mixture of
products. Accordingly, the SAR for 5-carbon-linked analogs was
investigated using the 6-des-chloro starting material since only a
very modest (less than twofold) drop-off in activity was noted
without this halogen present. When chloropyrazine 8, which was
prepared in two steps starting from commercially available 7,
was treated with guanidine, the major product resulted from a dis-
placement of the 5-chlorine and not from reaction with the methyl
ester to form the acyl guanidine. This was surprising since this
transformation of the dichloro variant 3 to 4 favored attack at
the ester position. To circumvent this issue, Suzuki coupling was
analogs were evaluated for percent inhibition at 1 or 20 lM using
patch clamp electrophysiology.21 In general, aliphatic groups had
enhanced inhibition relative to amiloride 1, shown for compari-
son purposes. Primary (5a–g) or secondary amines (5h–o) were
similar in efficacy, while too much steric bulk had a deleterious
effect (5c).
Aromatic amines had moderately higher inhibition than their
aliphatic counterparts as exemplified by aniline 5r. The SAR for
substituent groups placed on the aniline ring was generally flat,
although reduced lipophilicity was an important factor with pyri-
dines 5u–v showing less activity. The most potent compound
was naphthalene 5w with 62% inhibition at 1 lM, with related loss
of potency with the more polar quinolines (5x–y). A number of
benzylamines (5z–5i0) were also of interest, but again, flat SAR
was observed for more functionalized benzylamines across the
series.
Guided by the results obtained with the amine derivatives in
Table 1, selected amiloride analogs with ether functionality linked
at the 5-position to the pyrazine were evaluated as shown in Table
2. Generally, these derivatives were not as effective inhibitors as
their amine congeners. Aromatic ethers (6f–l) were modestly fa-
vored over their alkyl counterparts, while thioethers were signifi-
cantly less active. Consistent with the amine series, naphthyloxy
analog 6h was the most effective amongst the series, while certain
quinoline or isoquinoline derivatives could be tolerated depending
upon the position of the nitrogen.
Selected amiloride analogs that are carbon linked at the 5-po-
sition to the pyrazine are shown in Table 3. While placement of a
phenyl group (9a) provided 35% inhibition @ 1 lM, consistent
with trends observed with anilines and phenol from Tables 1
and 2, the corresponding naphthalenes 9c–d and benzyl 9e deriv-
atives were very weak ASIC3 channel blockers. SAR around the
ortho, meta, or para positions on the phenyl ring was generally
flat with only xylene 9f showing similar inhibition. One notable
exception was from the biaryl series in which meta-biphenyl 9g
5-NH2 tolerant
of substitution
3-NH2
required
H2N
Cl
N
N
NH2
N
gave the strongest block (72% @ 1 lM) of all amiloride analogs
examined, while ortho- and para-biaryls (9h–k) were much less
active.
A very limited number of the highly potent amiloride analogs
were selected for 5-point titration in electrophysiology. Results
NH2
NH2
2-Acylguanidine
required
6-H or halogen
tolerated
O
N required
are shown in Table 4. Amiloride provided an IC50 of 4.4 lM, while
Figure 2. Amiloride ASIC3 SAR.
X
N
N
NH2
N
Cl
Cl
N
N
NH2
N
Cl
Cl
N
N
NH2
a
b or c
NH2
NH2
NH2
NH2
Cl
CO2Me
O
O
5a-i: X = NR1R2
3
4
6a-l: X = OR1, SR1
R1
N
N
NH2
N
Cl
N
N
NH2
N
NH2
CO2Me
d, e
f, a
NH2
NH2
CO2Me
N
O
7
8
9a-k
Scheme 1. (a) Na(S), i-PrOH, guanidine-HCl. (b) R1R2NH, DMSO, TEA, 80–120 °C. (c) R1SH or R1OH, NaH, DMF. (d) metachloroperoxybenzoic acid. (e) POCl3, DMF, rt. (f)
Pd(dppf)Cl2, K2CO3, DMF, R1B(OH)2.
