Synthesis of novel trans-4-(substituted-benzamido)-3,4-dihydro-2H- benzo[b]-pyran-3-ol derivatives as potential anticonvulsant agents with a distinctive binding profile

Full text of DOI:10.1021/jm960535w

J . Med. Chem. 1996, 39, 4537-4539
Sch em e 1.a Synthetic Route to
4537
Syn th esis of Novel tr a n s-4-(Su bstitu ted -
4
-Amidobenzo[b]pyran-3-ols
ben za m id o)-3,4-d ih yd r o-2H-ben zo[b]-
p yr a n -3-ol Der iva tives a s P oten tia l
An ticon vu lsa n t Agen ts w ith a Distin ctive
†
Bin d in g P r ofile
Wai N. Chan, J ohn M. Evans, Michael S. Hadley,
Hugh J . Herdon, J effrey C. J erman,
Helen K. A. Morgan, Tania O. Stean,
Mervyn Thompson,* Neil Upton, and
Antonio K. Vong
SmithKline Beecham Pharmaceuticals,
New Frontiers Science Park, Third Avenue,
Harlow, Essex CM19 5AW, U.K.
a
Received J uly 25, 1996
Reagents: (i) 3-chloro-3-methylbutyne and procedures as in
In our earlier communication,1 we reported that
ref 4; (ii) for trans racemate, m-CPBA, and procedures as in ref 4;
(
iii) for 3R,4S enantiomers, epoxidation using procedures in ref 5;
fluorobenzamides such as 5f, possessing 3R,4S stereo-
chemistry, showed good anticonvulsant activity in the
mouse maximal electroshock seizure threshold (MEST)
model, while being devoid of antihypertensive proper-
ties. In contrast, the corresponding 3S,4R enantiomers,
like the structurally and stereochemically related acti-
vator of ATP-sensitive potassium channels, levcro-
t
(iv) NH3, EtOH; (v) Et3N, CH2Cl2, R2COCl (method A); (vi) KOBu ,
t
Bu OH, R CONH (method B).
2
2
4
procedures previously described. The exceptions are the
5
c
chiral amino alcohols 4b, 4c, 4f, and 4i (precursors to
5b, 5c, 5f, and 5i), which were prepared by enantiose-
lective epoxidation of the corresponding chromenes 2
2
5
makalim (6), showed potent antihypertensive activity.
using Mn(III) salen catalysts, followed by treatment
The 3S,4R enantiomers additionally showed little an-
ticonvulsant activity. A subsequent communication
with ethanolic ammonia. The amino alcohols 4 were
then converted into the amides shown in the Table 1
by reaction with the appropriate acid chloride (method
A). Alternatively, the benzopyran epoxides 3 were
converted directly into the benzamides 5a and 5e on
treatment with the anion of 4-fluorobenzamide in tert-
butyl alcohol at 40 °C (method B). Optical purity was
determined by HPLC on a Chiracel OD, P45SC (Daicel)
column using phosphate buffer (pH 3)/methanol/aceto-
3
confirmed that a 4S configuration of the fluorobenza-
mide is crucial in conferring anticonvulsant activity
without antihypertensive activity. Although the 3-hy-
droxyl group appears essential for anticonvulsant activ-
ity, its stereochemistry is not important. These early
data clearly indicated that the mechanism of the anti-
convulsant action of compounds such as 5f was unlikely
to involve the modulation of ATP-sensitive potassium
channels. It was clearly important to gain a fuller
understanding of structure-activity relationships (SAR)
within the series.
3
nitrile as eluent. (X-ray analysis of 5b also confirmed
the absolute stereochemistry; data not presented.)
3
6
Ch a r a cter iza tion of [ H]SB 204269 Bin d in g Site.
3
[ H]SB 204269 binding to rat forebrain membranes is
This communication describes the discovery of ana-
logues of 5f such as the 4-fluorobenzamide 5b (SB
of moderate affinity (K 40 nM) and fairly high density
(Bmax 220 fmol/mg of protein). The binding is highly
D
2
04269) with markedly enhanced anticonvulsant po-
enantioselective, since in competition studies the enan-
tency in vivo. In order to facilitate an understanding
tiomers (+)-5b (3R,4S) and (-)-5c (SB 204268, 3S,4R)
3
of the mechanism of their anticonvulsant action, [ H]-
had pK values of 7.3 and <4.5, respectively (Figure 1).
i
3
3
SB 204269 was prepared by [ H] hydrogenolysis of the
In addition to rat forebrain, [ H]SB 204269 binding
3
corresponding 2,5-dibromobenzamide of 5b. [ H]SB
has been observed in mouse, cat, dog, marmoset, and
human brains. In all species, affinities were similar.
Further details will be published elsewhere.6
2
04269 was then used to identify a novel highly enan-
tioselective binding site in the brain of several species
including humans. Extensive pharmacological charac-
None of the standard anticonvulsant agents (diaz-
epam, phenytoin, valproate, and phenobarbitone) or the
newer compounds in clinical development (gabapentin,
lamotrigine, and levetiracetam) have shown any activity
3
terization illustrated the unique nature of the [ H]SB
2
04269 binding site as a mechanistic target. Having
identified this unique binding site, displacement of the
radioligand was used to study the SAR directly for this
intriguing class of compounds. For the series of ben-
zopyrans described here, affinity for this site was found
to correlate with their anticonvulsant activity in rodent
models.
3
6
(at up to 100 µM) in the [ H]SB 204269 binding assay.
In addition, although the benzamides 5 are structurally,
but not stereochemically, related to the benzopyran class
+
of ATP-sensitive K channel activators, none of the
activators of this or other classes (levcromakalim,
aprikalim, or pinacidil) showed significant affinity for
Ch em istr y. Key intermediates for the synthesis
6
(Scheme 1) of the compounds described in the Table 1
the binding site. These studies demonstrate the unique
are the racemic 4-amino-3,4-dihydro-2,2-dimethyl-2H-
benzo[b]pyran-3-ols 4, which were synthesized using
mechanistic nature of this class of anticonvulsant 4(S)-
amidobenzopyran-3(R)-ols.
Resu lts a n d Discu ssion . In vitro potency (Table 1)
†
was measured using the radioligand binding procedure.6
This work was conducted in compliance with the Home Office
(
U.K.) Guidance on the operation of the Animals (Specific Procedures)
Act 1986, and was reviewed and approved by the SB Procedures
Review Panel.
Compounds were evaluated for oral anticonvulsant
activity, in groups of 10-20 naive mice (male CD1-
S0022-2623(96)00535-3 CCC: $12.00 © 1996 American Chemical Society

4
538 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 23
Communications to the Editor
Ta ble 1. Chemical and Biological Data for trans 4-(Substituted-amido)-2H-benzo[b]pyran-3-ols
mouse MEST po % increase in seizure
threshold at dosec
[3
H]SB 204269
compda,b
mp, °C
method
R1
R2
stereochem
binding pKi
10 mg/kg
3 mg/kg
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
a
b
c
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
175
163-4
161
B
A
A
A
B
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
5-COMe
6-COMe
6-COMe
7-COMe
8-COMe
6-CN
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
4-FPh
Ph
racemic
3R,4S
3S,4R
racemic
racemic
3R,4S
racemic
racemic
3R,4S
racemic
racemic
racemic
racemic
racemic
racemic
3R,4S
racemic
3R,4S
3R,4S
5.09 ( 0.07
7.32 ( 0.03
NT
102*
8
NT
5
41*
45*
-7
-8
10
NT
89*
NT
NT
NT
15*
2
NT
NT
NT
NT
NT
<4.5
190-3
262-4
224-5
197-8
239-40
172
164-5
185-6
205-7
217-20
198
5.51 ( 0.03
5.00 ( 0.02
6.00 ( 0.11
6.20 ( 0.07
5.84 ( 0.13
6.35 ( 0.05
5.38 ( 0.15
5.30 ( 0.02
6.80 ( 0.01
6.12 ( 0.06
6.69 ( 0.04
6.06 ( 0.02
7.30 ( 0.05
7.16 ( 0.01
7.65 ( 0.06
8.04 ( 0.05
8.28 ( 0.03
5.90 ( 0.04
6-NO2
6-SO2Ph
6-C2F5
6-F
6-Me
5
6-COEt
6-COPh
6-CO2Me
6-CONH2
6-COMe
6-COMe
6-COMe
6-COMe
6-COMe
6-COMe
35*
NT
59*
6
110*
55*
140*
90*
51*d
NT
NT
NT
65*
NT
94*
78*
NT
NT
>296
188-91
215
4-ClPh
2-ClPh
3-ClPh
3-IPh
193-5
170-2
171-3
209-10
s
t
u
3R, 4S
racemic
cyclohexyl
a
1H NMR spectra were consistent with assigned structures. b All compounds gave satisfactory C, H, N analyses ((0.4%). c *p < 0.05
compared to vehicle control animals see ref 12. In all experiments, the CC50’s for vehicle-treated controls fell within the range of 12-14
mA. Percent changes for drug-treated groups are derived from studies where standard errors were less than 10% of the CC50 values.
Data for the racemate; NT ) not tested.
d
compounds of particular interest were tested as 3R,4S
enantiomers. Thus, in interpreting the data in Table
1
, it should be borne in mind that the racemate typically
has half the potency of the 3R,4S enantiomer; this is
further exemplified with data (not in Table 1) on the
racemate of 5b, both in vitro (pKi 6.91 ( 0.05) and in
vivo (57% increase in seizure threshold at 10 mg/kg po).
The markedly enhanced potency of 5b compared to
that of the earlier lead compound 5f reflects the change
in the 6-substituent from cyano to acetyl. The 6-cyano
group is a key feature of potassium channel activators
with 3S,4R stereochemistry like levcromakalim (6).
3
F igu r e 1. Competition for [ H]SB 204269 binding to rat
forebrain.
Charles River, 25-30g), in the mouse MEST threshold
7
test using an “up and down” method of shock titration.
3
The enantioselectivity of the [ H]SB 204269 binding
discussed above is entirely consistent with previous
reports1 on 6-cyano-4-(fluorobenzamido)benzopyrans,
where good oral anticonvulsant activity was only ob-
served for compounds such as 5f with 4S stereochem-
istry. In contrast to the high affinity and in vivo potency
of the 3R,4S enantiomer 5b, the 3S,4R enantiomer 5c
showed only very weak binding affinity (pKi <4.5) and
was also inactive in the mouse MEST test at a dose of
,3
Other electron-withdrawing 6-substituents known8 to
be beneficial in potassium channel activators such as
nitro (5g), phenylsulfonyl (5h ), and pentafluoroethyl (5i)
gave compounds with no improvement over the cyano
compound. The 6-fluoro (5j) and 6-methyl (5k ) com-
pounds also showed only weak activity. Thus, not only
is there stereochemical differentiation from potassium
channel activators SAR but also there are clear differ-
ences in the optimal nature of the 6-substituent, 6-acetyl
being particularly beneficial. The data suggest that the
carbonyl of the 6-acetyl group is likely to be involved in
a key interaction with the binding site. Homologation
to 6-propionyl (5l) resulted in little loss in affinity,
1
0 mg/kg po. Indeed, for a range of compounds, a
qualitative correlation was found between in vitro
binding affinity and in vivo activity, supporting the
relevance of this novel site to anticonvulsant activity
(Table 1). As it became established that potent anti-
convulsant activity of the trans compounds was associ-
ated only with the 4S enantiomer, many compounds
were tested as trans racemates and generally only

Communications to the Editor
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 23 4539
whereas replacement by 6-benzoyl (5m ) was clearly
detrimental. Replacement by 6-methoxycarbonyl (5n )
gave a compound, isosteric with 5l, with similar affinity,
but replacement by 6-aminocarbonyl (5o) resulted in
markedly reduced affinity.
As the 6-acetyl moiety appeared to be crucial in
conferring high affinity, its position in the benzopyran
ring was investigated. Affinity was reduced by around
supplies of amino alcohol 4b and the Synthetic Isotope
Chemistry Unit, SB, for the provision of radiolabeled
5b.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures, including analytical and spectroscopic data, for the
preparation of all compounds 5 and biological protocol details
(6 pages). Ordering information is given on any current
masthead page.
1
00-fold when it was moved to any of the alternative
Refer en ces
aromatic ring positions (5a , 5d , 5e).
(
1) Blackburn, T. P.; Buckingham, R. E.; Chan, W. N.; Evans, J .
M.; Hadley, M. S.; Thompson, M.; Upton, N.; Stean, T. O.; Stemp,
G.; Vong, A. K. Stereochemical Differentiation of Anticonvulsant
and Antihypertensive Effects in 4-(Fluorobenzoylamino)ben-
zopyrans. Bioorg. Med. Chem. Lett. 1995, 5, 1163-1166.
Looking at the benzamide substitution pattern, re-
moval of the 4-fluoro atom from 5b resulted in no loss
of affinity (5p ), nor did replacement by 4-chloro (5q).
However, the 2-chloro (5r ) and 3-chloro (5s) compounds
showed higher affinity, although the latter compound
did not show the anticipated higher oral potency. This
discovery of higher affinity for the 3-chloro compound
led to the synthesis of the 3-iodo compound (5t) which
showed one of the highest affinities in this series,
although this was not reflected in its in vivo activity
when examined as the corresponding racemate. Re-
placement of the benzamide by cyclohexylamide (5u )
caused a reduction in potency, suggesting a role for the
phenyl ring other than that of a lipophilic binding
interaction.
(2) Ashwood, V. A.; Buckingham R. E.; Cassidy, F.; Evans, J . M.;
Faruk, E. A.; Hamilton, T. C.; Nash, D. J .; Stemp, G.; Willcocks,
K. Synthesis and Antihypertensive Activity of 4-(cyclic amido)-
2
H-benzopyrans. J . Med. Chem. 1986, 29, 2194-2201.
(3) Brown, T. H.; Campbell, C. A.; Chan, W. C.; Evans, J . M.; Martin,
R. T.; Stean, T. O.; Stemp, G.; Stevens, N. C.; Thompson, M.;
Upton, N.; Vong, A. K. Stereochemical Preferences and Require-
ment for the 3-Hydroxyl Group in Novel Anticonvulsant 4-Fluo-
robenzoylamino Benzopyrans. Bioorg. Med. Chem. Lett. 1995,
5
, 2563-2566.
(
4) (a) Ashwood, V. A.; Cassidy, F.; Coldwell, M. C.; Evans, J . M.;
Hamilton, T. C.; Howlett, D. R.; Smith, D. M.; Stemp, G.
Synthesis and Antihypertensive Activity of 4-(Substituted-
carbonylamino)-2H-benzopyrans. J . Med. Chem. 1990, 33, 2667-
2
672. (b) Roxburgh, C. J . The Synthesis of Potassium Modulating
Structures. Synthesis 1996, 307-320.
(
5) (a) Sasaki, H.; Irie, R.; Hamada, T.; Suzuki, K.; Katsuki, T.
Rational Design of Mn-Salen Catalyst (2): Highly Enantio-
selective Epoxidation of Conjugated cis-Olefins. Tetrahedron
Compound 5b (SB 204269) was selected for further
evaluation and studies detailing the preclinical profile,
which suggest potential utility in the treatment of
1
994, 50, 11827-11838. (b) J acobsen, E. N.; Deng, L.; Furukawa,
Y.; Martinez, L. E. Enantioselective Catalytic Epoxidation of
Cinnamate Esters. Tetrahedron 1994, 50, 4323-4334. (c)
European Patent. PCT WO/94/03271, 1994.
9
,10
epilepsy disorders, will be published elsewhere.
The
compound is very effective at impairing pathologically
high levels of neuronal excitability while having mini-
mal effects on normal physiological processes. The
(
6) (a) Herdon, H.; J erman, J . C.; Upton, N.; Stean, T. O.; Chan, W.
N. Identification of a stereospecific CNS binding site for the novel
anticonvulsant SB 204269. Eur. J . Pharmacol. Unpublished
results. (b)European Patent PCT WO/96/18650.
absence of sedative potential has also been confirmed
in a range of studies.11
(7) Upton, N. Mechanisms of action of new antiepileptic drugs:
rational design and serendipitous findings. Trends Pharmacol.
Sci. 1994, 15, 456 - 463
8) Evans, J . M.; Hamilton, T. C.; Longman, S. D.; Stemp, G.
Potassium Channels and their Modulators: From Synthesis to
Clinical Experience, 1st ed.; Taylor and Francis, Ltd.: London,
1996.
9) Stean, T. O.; Upton, N. Unpublished results. For example, in
the more stringent rat supramaximal electroshock test 5b shows
excellent potency (ED50 6.3 mg/kg po) relative to standards (ED50
values of 12.0, g 40, and 60.3 mg/kg po for carbamazepine,
gabapentin, and diazepam, respectively) and also produces a
high level of efficacy. For methods and rationale, see refs 7 and
10.
10) L o¨ scher, W.; Schmidt, D. Which Animals should be used in the
search for new antiepileptic drugs? A proposal based on experi-
mental and clinical considerations. Epilepsy Res. 1988, 2, 145-
181
In conclusion, a novel enantioselective binding site for
trans 3R,4S benzamidobenzopyrans which appears to
mediate their anticonvulsant activity has been identified
in mammalian brain. This has been utilized to study
SAR in this mechanistically unique series of anticon-
vulsants. No other known anticonvusants have been
found to have significant affinity for this binding site,
and compound 5b (SB 204269) has been selected for
clinical evaluation as a pioneer treatment for epilepsy
disorders. Further studies to elucidate the mechanism
of action of this novel series of 4-amido benzopyrans are
underway.
(
(
(
(
11) Stean, T. O.; Upton, N. Unpublished results. For example, the
MED for 5b was >200 mg/kg for any impairment in performance
in the accelerating rotorod test.
12) Litchfield, J . T.; Wilcoxon, F. A Simplified Method of Evaluating
Dose-Effect Experiments. J . Pharmacol. Exp. Ther. 1949, 96,
Ack n ow led gm en t. We are pleased to thank Ana-
lytical Sciences, SB, Harlow, for analytical and spec-
troscopic data and Miss Kellie Shrimpton for assistance
in the preparation of the manuscript. Thanks are also
due to Synthetic Chemistry, SB, Harlow, for generous
(
9
9-113.
J M960535W