A series of 1-, 2-, 3-, 4-, 5-, 6-, 7-, 10-, and 12-substituted pyridodiindoles were synthesized and screened in vitro against [3H]diazepam for activity at the benzodiazepine receptor (BzR). In vitro, the 2-substituted pyridodiindoles were found to be the most potent (IC50 less than 10 nM) of this new class of BzR ligands. In vivo, 2-methoxypyridodiindole 19a (IC50 = 8 nM) was found to be the most potent partial inverse agonist (proconvulsant) of the series. The parent compound 2 (IC50 = 4 nM) was only slightly less potent. In addition, 2-hydroxypyridodiindole 21a (IC50 = 6 nM) was found to exhibit potent proconvulsant activity when administered as a prodrug derivative, pivaloyl ester 22. 2-Chloropyridodiindole 16a (IC50 = 10 nM) was devoid of preconvulsant activity; however, 16a was found to be the most potent antagonist of the anticonvulsant effects of diazepam in this class of BzR ligands. From the in vivo data available, substitution on ring E of 2 with electron-withdrawing groups results in antagonists at BzR, while replacement of hydrogen at C-2 with electron-releasing groups provides enhanced inverse agonist activity. The pyridodiindoles were used as "templates" for the formulation of a model of the inverse agonist/antagonist active site of the BzR. The proposed model consists of a hydrogen bond acceptor site (A1) and a hydrogen bond donor site (D2) disposed 6.0-8.5 A from each other on the receptor protein. The hydrogen-bonding sites are believed to be located at the base of a narrow cleft. A large lipophilic pocket at the mouth of the narrow cleft serves to direct molecules into the binding site, while the presence of a small lipophilic pocket permits substitution only at position 2 of the pyridodiindole nucleus for maximum binding potency.

A series of rigid planar azadiindoles (8a, 8b, and 8d), benzannelated pyridodiindoles (11a, 11b, and 11d), and indolopyridoimidazoles (11c, 20, and 24) were synthesized from 4-oxo-1,2,3,4-tetrahydro-beta-carboline 5 via the Fischer indole cyclization with the appropriate arylhydrazines. These analogues were employed as probes ("molecular yardsticks") to define the spatial dimensions of the lipophilic regions of the benzodiazepine receptor (BzR) binding cleft. Benzannelated indoles 11a-d and indolopyridoimidazoles 20 and 24 were important in establishing an area of negative interaction (S1, see Figure 6, part b) in the binding cleft common to the interactions of both inverse agonists and agonists. Data from this chemical and computer-assisted analysis of the pharmacophore (see Figure 6) indicates that inverse agonists and agonists bind to the same binding region, but the pharmacophoric descriptors required for the two activities are different, in keeping with previous studies with these planar ligands. However, the hydrogen bond donating site H1 and the lipophilic region L1 in the receptor binding site are common interactions experienced by both series of ligands. The low affinities of both indolo[3,2-c]carbazole (3a) and indolo[3,2-b]isoquinoline (3b) for the BzR are consonant with the requirements of a hydrogen bond acceptor interaction at donor site H1 and a hydrogen bond donor interaction at acceptor site A2 for potent inverse agonist activity in the beta-carboline series. The hydrochloride salts of 1-aza- 8a (IC50 10.6 nM), 2-aza- 8b (IC50 51.5 nM), and 4-azadiindole 8d (IC50 11.2 nM) were found to be much more soluble in water than the corresponding salt of the parent diindole 2. Moreover, aza analogues 8a and 8b were shown to be partial inverse agonists with proconvulsant potencies comparable to that of the parent diindole 2.

No abstract

The heterocyclic base, 7,12-dihydropyrido[3,2-5:5,4-5 diindole 1, was first reported in 1985.1 It was later discovered that 1 possesses high affinity for the benzodiazepine receptor in vitro and exhibits a broad range of