D. Kuzmich et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6640–6644
6641
1) small structural changes can have a range of effects on potency
and efficacy in cell based assays for agonist activity while GR bind-
ing potency is maintained.7 These changes are in contrast to the
classical GC antagonist, RU-38486 (Fig. 1), which upon binding to
the GR ligand-binding-domain (LBD) causes helix 12 to adopt an
open conformation resulting in functional antagonism or ‘active
antagonism’.8 Similarly, the sulfonamide series of GC modulators
described in this report (e.g., compound 2a) display a range of
effects on potency and efficacy in cell based assays for agonist
activity while GR binding potency is often maintained.
revealed several interesting features. The central NH group of
(S)-2b forms a key H-bond with Asn564 (helix 12) mimicking the
interaction seen with the 11b-hydroxyl group of dexamethasone
(Fig. 2).
The indole moiety of (S)-2b overlays with the steroid A-ring of
dexamethasone forming p-stacking interactions with Phe623. The
tri-substituted phenyl occupies the space of the steroid D-ring.
Based on this analysis and our experience optimizing the trifluoro-
methyl carbinol series (compound 1) we targeted three strategies
for facilitating a functional switch of 2b. These included: (i) intro-
duction of a polar residue on the phenyl sulfonamide moiety which
Our initial report disclosed the identification of a-methyltrypta-
mine sulfonamide derivatives (e.g., 2a and 2b) as potent GR
ligands, that were optimized for binding potency and NR selectiv-
ity.9 These compounds failed to display the desired agonist activity
in functional assays and were assumed to be GR antagonists. Since
our initial publication, a number of groups have reported modifica-
tions to the sulfonamide template that have afforded GR agonist
activity.10,11 In this Letter, we describe our SAR findings for this
class of GR ligands. Specifically, we report that the placement of
function-regulating pharmacophoric points in three distinct
regions of this scaffold afford a switch in functional activity provid-
ing potent and efficacious GR agonists relative to dexamethasone
in a cellular assay.12,13
The IC50 values reported herein for binding to GR, MR and PR
were determined by using a fluorescence polarization competitive
binding assay.14 The GR and MR assays measured the ability of test
compounds to compete with tetramethylrhodamine (TAMRA)-la-
beled dexamethasone. The PR binding assay was run using TAM-
RA-labeled mifepristone (RU-38486). Anti-inflammatory activity
(transrepression) is reported as the inhibition of IL-6 production
in human foreskin fibroblasts (HFF) in response to stimulation by
the pro-inflammatory cytokine IL-1.15 It is known that circulating
IL-6 levels increase during an inflammatory response and IL-6 pro-
duction can be inhibited by GCs such as dexamethasone. In the IL-6
functional assay, potency and efficacy are reported, and efficacy is
expressed as a percentage of the maximum response observed for
dexamethasone.
could engage the Asn564, (ii) replacement of the a-methyl group
with a trifluoromethyl moiety to improve the hydrogen bond
potential of the sulfonamide N–H, and (iii) placement of a polar
residue, useful as a steroid A-ring mimetic, on the indole ring to en-
gage the Arg611 and Gln570 pair. As a proof of concept, compound
3, containing an aniline moiety designed to engage Asn564, dis-
played potent GR binding, good potency (IC50 = 120 nM) and par-
tial efficacy (59%) in the IL-6 assay. In comparison, compounds
2a and 2b are only 12 and eightfold less potent in the GR binding
assay, respectively, but were not active in the IL-6 assay (max. con-
cn = 2000 nM). Replacement of the
a-methyl group of 3 with a
trifluoromethyl group (compound 4) further improved potency
and efficacy in the IL-6 assay relative to compound 3. Interestingly,
the CF3 modification alone was not sufficient to afford potency or
efficacy in the IL-6 assay for compound 5, the 2,4,6-trichlorophenyl
substituted sulfonamide. However, weak efficacy in the IL-6 assay
was observed for compound 6, the 2,4,6-trimethylphenyl substi-
tuted sulfonamide. The gem-dimethyl analog 7 also failed to dem-
onstrate efficacy in the IL-6 assay. Clearly, these results highlight
the value in a multi-component SAR approach to achieving func-
tional activity. It should be noted that compounds 3, 4 and 7 dem-
onstrated excellent selectivity relative to dexamethasone in the PR
and MR assays. For example, compound 3 was 102-fold more selec-
tive for GR over MR in the binding assays (Table 1).
Finally, we focused our attention on the indole moiety which
modeling suggests aligns with the same pocket occupied by the
A-ring of dexamethasone. Our choice of functional groups was dri-
ven by our experience in the trifluoromethyl carbinol series (com-
pound 1).7b In the carbinol series, the cyano group of 1 can serve as
a function-regulating pharmacophore due to critical polar interac-
tions with the Arg611/Gln570 pair and a lipophilic component
(methyl groups) in this same region can influence the level of ago-
nist activity and efficacy observed. The use of a cyano group as a
replacement for the A-ring carbonyl is known from other nuclear
hormone receptors, like progesterone19 and androgen20 receptor.
To study polar and lipophilic modifications to the indole moiety
we chose the 2,4,6-trimethylphenyl sulfonamide analog 6, and pre-
pared methylindoles 8–11 and cyanoindoles 12–15 (Table 2).
Interestingly, the methylated indole analogs 8–11 are all of similar
potency in the GR binding assay (GR IC50’s = 10–37 nM). However,
a range of GR binding potencies is observed for the cyano analogs
12–15. Thus, the 4-cyanoindole 12 was not potent at the highest
To guide our SAR and placement of pharmacophoric points we
docked compound (S)-2b into the published GR-LBD binding pock-
et (Fig. 2).16 A proposed binding pose of compound (S)-2b (Fig. 2b)
concentration tested (GR IC50 >10 lM). However, the 7-cyanoin-
dole 15 (GR IC50 = 10 nM) is 28 and 32-fold more potent than 6-
cyanoindole 14 and 5-cyanoindole 13, respectively. Consistent
with a lipophilic component in this region useful for improving
functional activity, the 5-methylindole (compound 9) afforded par-
tial agonist activity (60% efficacy) in the functional assay. At the
same time, consistent with our binding hypothesis (Fig. 2b) that
a polar component in the steroid A-ring region is useful for func-
tional activity, the 7-cyanoindole analog 15, which can engage
the Arg611/Gln570 pair, displayed potent binding activity and par-
tial agonism (70% efficacy) in the functional assay. Compound 16,
containing both of these function-regulating pharmacophoric
points (the 5-methyl-7-cyanoindole), gave further improvements
Figure 2. Overlay of docking pose of compound (S)-2b (orange) with the co-
complex x-ray structure of GR-LBD and dexamethasone (rose). Docking was
performed using Glide17 and the picture was generated with MOE.18 Dotted lines
represent hydrogen bonds.
Kuzmich, Daniel
