1216
B. E. Kornberg et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1213–1216
hydrogen bond donor, was necessary. In-house model-
ing studies suggested an optimal distance of 9-11 ang-
stroms between the two rings and the distance of the
basic nitrogen from either aryl ring was not well defined.
Of the various classes of compounds examined, the
focus of this paper will be specifically on NR1A/2B
receptor antagonists containing an acetylenic linker
between the A- and B-rings.
donor heterocycles. While the oxindole analogue (12)
reduced binding affinity, the benzisoxazolone analogues
(13 and 14) retained better binding affinity and were
active in FT at 30 mg/kg PO.
Substitutions ortho to the B-linker in the phenyl B-ring
(18 and 19) were well tolerated and resulted in highly
active compounds whereas reduced activity was observed
in compounds with substitutions meta to the linker (16
and 17). The 7-azabenzimidazolone analogue, 22, had
similar NR2B affinity as 3 but was inactive in all the in
vivo models similar to 18 and 19. We believe the poor in
vivo activity of 18, 19, and 22 is due to their relative
insolubility, resulting in poor absorption. Further
B-ring modifications such as the thiourea analogue (21)
and attachment of the linker at the 4-position of the
benzimidazolone (23)10 showed no advantage indicating
the preference for a linear motif.
The general procedure used to synthesize the com-
pounds shown in the accompanying tables is depicted in
Scheme 1. Based on the overall metabolic profile of the
compounds described in this communication it was
important to have a fluoro group in the para position of
the A-ring. Shifting the fluorine to the meta position
(i.e., 11) resulted in weaker NR2B activity. As indicated
in Scheme 1, a variety of 4-(4-fluorobenzyl)-piperidine
analogues5 were alkylated with propargyl bromides to
give the corresponding terminal alkynes. Palladium
cross-coupling reactions with a variety of aryl bromides
or iodides gave the desired products in moderate to
good yields.6
Compound 3 was identified as a highly selective NR1A/
2B receptor antagonist with good efficacy in models of
Parkinson’s disease and neuropathic pain. SAR studies
around the structural motif of 3 indicated that con-
formationally constrained A- and B- linkers offered no
significant advantages over 3 in vitro and in vivo.
Alterations within the B-ring resulted in compounds
with excellent potency in vitro, however, in vivo activity
remains elusive. Further work is in progress to optimize
the in vivo activity.
The affinity of target compounds for the NR1A/2B
3
receptor was determined by a H-ifenprodil radioligand
binding assay using native rat brains. Those compounds
with desirable affinity were further tested for in vivo
activity (PO or IP) in the unilaterally 6-hydroxy-dopa-
mine (6-OHDA) lesioned rat model (a measure of anti-
Parkinson’s activity), and the formalin test (FT, a model
of neuropathic pain).
References and notes
It was interesting to note that compound 3 was the only
compound that displayed efficacy in both the 6-OHDA
and FT assays with nanomolar affinity in the ifenprodil
binding assay. However, the in vivo activity in this
compound could be limited due to its poor solubility.
To optimize the in vivo activity and improve overall pK
properties, extensive SAR studies were undertaken. The
rationale here was to improve aqueous solubility by
reducing the planarity as well as incorporating hetero-
atoms in the structural motif. The results are listed in
Tables 1 and 2.
1. Seeburg, P. H. Trends Neurosci. 1993, 16, 359.
2. Monyer, H.; Burnashev, N.; Laurie, D. J.; Sakmann, B.;
Seeburg, P. H. Neuron 1994, 12, 529.
3. For a review on NR2B Selective NMDA Receptor
Antagonists see: Nikam, S. S.; Meltzer, L. Curr. Pharm.
Des. 2002, 8, 125.
4. Chenard, B. L.; Shalaby, I. A.; Koe, B. K.; Ronau, R. T.;
Butler, T. W.; Prochniak, M. A.; Schmidt, A. W.; Fox,
C . B.J. Med. Chem. 1991, 34, 3085.
5. Shanklin, J. R., Jr.; Johnson, C. P., III; Proakis, A. G.;
Barrett, R. J. J. Med. Chem. 1991, 34, 1630.
6. Wright, J. L.; Gregory, T. F.; Kesten, S. R.; Boxer, P. A.;
Serpa, K. A.; Meltzer, L. T.; Wise, L. D.; Espitia, S. A.;
Konkoy, C. S.; Whittemore, E. R.; Woodward, R. M. J.
Med. Chem. 2000, 43, 3408.
7. Wright, J. L.; Gregory, T. F.; Bigge, C. F.; Boxer, P. A.;
Serpa, K.; Meltzer, L. T.; Wise, L. D.; Cai, S.-X.; Haw-
kinson, J. E.; Konkoy, C. S.; Whittemore, E. R.; Wood-
ward, R. M.; Zhou, Z.-L. J. Med. Chem. 1999, 42, 2469.
8. A-ring-linker benzylidene analogue prepared via dehy-
dration of the tertiary carbinol analogue with SOCl2/pyr-
idine.
9. Prepared from 4-[2-(4-fluorophenyl)vinyl]-pyridine fol-
lowed by reduction. J. Med. Chem. 1990, 33, 3133.
10. Prepared from coupling with 4-bromo-1,3-dihydrobenz-
imidazol-2-one, which was made via nitration of 2-
bromoacetanilide and separation of the desired 2-bromo-
6-nitroacetanilide according to the procedure described in
J. Chem. Soc. 1928, 3092.
The non-planar compounds with conformationally
constrained amines (4–7, Table 1) had significantly
weaker NR2B binding than compound 3 indicating a
preference for unconstrained amine. The branching
away from the amine was better tolerated as demon-
strated by the tertiary carbinol7 (8) which had improved
binding affinity in comparison to analogues 4–7. How-
ever compound 8 was inactive in vivo. The benzylpiper-
idine motif for the A-ring and A-linker in compound 3
appears to be optimum since modifications with regard
to unsaturation (9),8 chain length (10),9 and positioning
of the fluorine atom (11) all resulted in reduced NR2B
binding affinity.
SAR studies in the B-ring primarily focused on repla-
cing the benzimidazolone moiety with other H-bond