3216
J. A. Stafford et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3215–3218
assay.14 Attachment of the ester functionality at C3
provided entities with potency/selectivity and the
opportunity to map SAR as a function of changes in the
tether (Table 1). Based on tether length alone, potent
compounds were identified from each group of ana-
logues, and heteroatom-containing linkers delivered
compounds equipotent to carbon analogues. Interest-
ingly, the configuration at the C3-stereogenic center
proved important regardless of the tether length; nat-
ural l-amino acids conferred greater potency than their
unnatural d-counterparts, and this trend continued with
heteroatom linkers as well. In the end, the potency of
the glutamate-based analogues (11a, 11d, 13, and 15)
versus the single atom-spaced versions (8b–d), the cost-
liness of the aminoadipate analogue (16), and the rela-
tively cumbersome synthesis of the amine-linked
analogues (18b and 19b) afforded advantages to this
glutamate-based series.
Figure 1. Structures of midazolam (1) and diazepam (2).
The strategy called for esters that bind with high affinity
to the BZD receptor whereas their corresponding acids
would display reduced affinity. Thus, subsequent com-
parisons of ester-acid pairs were performed and are
represented by Ki (acid)/Ki (ester) ratio (Table 1). A
ratio of ꢁ100-fold in relative activities was targeted.
This filter confirmed the utility of the glutamate-based
analogues and prompted an investigation of selectivity
as a function of the identity of the ester. Because of the
lipophilic character of the BZD receptor,15 aliphatic and
simple aralkyl esters were evaluated. The methyl ester
11a emerged as a particularly interesting compound
owing to its excellent receptor binding affinity and the
remarkable separation in binding affinity relative to its
corresponding acid 11f, which binds the human BZD
receptor with a Ki value of 6.2 mM.
Scheme 1. Reagents and conditions: (a) CHCl3, reflux; (b) Et3N,
CH2Cl2, reflux; (c) 5% HOAc in (CH2Cl)2, 60 ꢂC; (d) [(n-Bu2Sn
(Cl)]2O, ROH, PhCH3, reflux; (e) LiOH, THF, MeOH, H2O.
FMOC protecting group was achieved by treatment
with triethylamine in dichloromethane. The solvent was
removed and replaced with 5% acetic acid in dichloro-
ethane; the resulting solution was warmed to effect
cyclodehydration and provide 11a in 55% yield.10
Transesterifications were executed using the protocol
reported by Otera11 to deliver 11b–e. Saponification
provided acid 11f.
Neither high affinity binding of a ligand to the benzo-
diazepine receptor nor selectivity within an ester–acid
pair characterizes the intrinsic efficacy (full agonist,
inverse agonist, antagonist) of a BZD receptor ligand.
Therefore, determination of intrinsic efficacy was asses-
sed by the ability of a compound to cause loss of the
righting reflex (LRR) in rats, an effect associated with
benzodiazepine full agonism.16 Compounds producing
LRR were evaluated by three measures: (1) time to
onset of LRR, (2) initial recovery from LRR (animal
rights itself three consecutive times after losing its
righting reflex), and (3) total recovery time (the animal
walks without ataxia and pulls itself up three con-
secutive times when suspended from a horizontal wire).
Several compounds were evaluated in the LRR assay
and the recovery behaviors were compared to the clin-
ical standards 1 and 2. Results are tabulated in Table 1.
Compound 11a presented the desired pharmacody-
namic profile offering rapid, uneventful sedation and
significantly shorter recovery times (initial=9 min,
total=24 min) relative to 1 and 2. The finding that the
acid 11f is the only identifiable metabolite from with-
drawn plasma samples bolsters the hypothesis that the
observed pharmacodynamic profile is attributed to the
plasma and/or tissue degradation of the BZD ester. In
addition, in vitro incubation of 11a in rat plasma gives
rise to 11f with a half-life <1 min.
Analogues 18a–c and 19a–c were prepared according to
Scheme 2. In the event, 3-amino-BZD (17)12 was reacted
with methyl bromoacetate or methyl acrylate after
which chiral HPLC was utilized to supply the adducts in
enantiomerically pure form. Subsequently (3S)-amino-
BZD13 was used to synthesize analogues of interest. As
before, saponification of the esters afforded the corre-
sponding acids.
Our project required esters with levels of potency com-
parable to clinically relevant BZDs, targeted as Ki ꢀ 50
nM as measured in our BZD radioligand binding
Scheme 2. Reagents and conditions: (a) methyl bromoacetate, Et3N,
THF; (b) methyl acrylate, EtOH; (c) LiOH, THF, MeOH, H2O.