5780
A. S. Felts et al. / Bioorg. Med. Chem. Lett. 23 (2013) 5779–5785
compound known as GRN-529 point to a potential role for mGlu5
NAMs in the treatment of other autism spectrum disorders.13
GRN-529 has also recently proven efficacious in rodent models of
treatment resistant depression (TRD).14 Finally, both MPEP and
MTEP have produced encouraging results in animal models of
addiction with various drugs of abuse, including cocaine,15 nico-
tine,15g,16 methamphetamine,17 morphine,18 and ethanol.19
Multiple mGlu5 NAMs have advanced to clinical trials, with the
most encouraging results thus far observed in GERD,20 FXS,21 and
PD-LID.22 The majority of clinical compounds are from within the
disubstituted alkyne structure class, including each of the three
molecules with confirmed ongoing clinical activity: dipraglurant
(ADX48621), mavoglurant (AFQ056), and RG7090 (RO4917523)
(Fig. 1).3a We have been interested in the identification and optimi-
zation of mGlu5 NAMs within chemotypes that do not contain a
disubstituted alkyne motif. One approach that we have success-
fully employed in this endeavor was based on the development
of hits identified using a functional cell-based high-throughput
screen (HTS) of a collection of 160,000 compounds.23 We have also
used both rational design approaches24 as well as virtual screening
methods to identify new mGlu5 NAM tool compounds.25
Among the confirmed hits from our functional HTS was aryl
ether benzamide 1 (Fig. 2), which demonstrated good potency in
our functional assay. This assay also serves as our primary assay
for lead optimization by measuring the ability of the compound
to block the mobilization of calcium induced by an EC80 concentra-
tion of glutamate in HEK293A cells expressing rat mGlu5.26 As part
of our initial hit evaluation process, the primary amine functional
group was removed to afford analog 2, which was approximately
four-fold more potent than hit 1. Another early structural modifi-
cation involved preparation of the compound with the alternative
orientation of the amide bond of 2 to produce analog 3. Though
compound 3 was more than 20-fold less potent than 2, our antic-
ipation was that optimization within this series might restore lost
potency and yield interesting analogs. Indeed, such an effort was
fruitful and is described in detail herein.
Preparation of aryl ether analogs of 3 was straightforward and
followed the general methods outlined here (Scheme 1).27 Though
certain reactions proceeded in poor to moderate yield, these reac-
tions were not optimized and were sufficient for analog generation.
For selected compounds of interest, scalable routes with improved
yields were developed.28 A nucleophilic aromatic substitution
reaction between pyridine or pyrimidine alcohols 4 and suitable
3-halobenzonitrile compounds 5 afforded ether intermediates 6.
Basic hydrolysis of the nitrile functional group provided the car-
boxylic acid intermediates 7. In certain cases acids 7 were coupled
with primary amines under standard conditions to give the desired
Scheme 1. Reagents and conditions: (a) For X = Br; CuO, K2CO3, pyridine, 80 °C (50–
73%) or CuI, KOtBu, DMG, DMF,
w, 190 °C (35–50%); (b) for X = F; K2CO3, DMF, w,
l
l
150–180 °C (48–72%); (c) aq NaOH, EtOH or dioxane, sealed tube, 100 °C (77–99%);
(d) R3NH2, DIEA, HATU, DMF, CH2Cl2 (15–58%) or R3NH2, POCl3, pyridine, À15 °C
(36–65%); (e) H2SO4, MeOH, reflux (84–92%); (f) Pd(OAc)2, PS–PPh3 or Pd(PPh3)4,
Zn(CN)2, DMF, l
w, 140 °C (32–81%); (g) R3NH2, KN(SiMe3)2, THF (10–65%).
The first area of the chemotype that was targeted for SAR explo-
ration was the eastern secondary amide group (Table 1). Potency
data in the primary assay is presented here as both pIC50 and
IC50 values for convenient evaluation of SAR. Not surprisingly,
the 3-chlorophenyl group (3) can be replaced with a 3-methyl-
phenyl group (9) with little effect on potency. More interesting
was the enhanced potency observed with 2-pyridyl derivative
10. While compound 3 was considered relatively lipophilic
(cLogP = 3.86), the more polar analog 10 was considerably less
lipophilic (cLogP = 2.70).29 Lipophilicity can be an important
parameter to monitor during the course of a CNS drug discovery
program.30 For this reason, the flexibility to install a heteroaryl ring
at this position of the chemotype was considered attractive, and
SAR development continued along those lines. Further modifica-
tion of this ring in the form of pyrimidine 11 was not well
tolerated. Efforts to evaluate substituted analogs of 10 produced
6-methylpyridine 12, which was equipotent to the original hit 1.
Other substituted analogs of 10 (14–16) were two to five-fold less
potent than 10. 4-Pyridyl analog 13 was only weakly active and
more than 25-fold less active than 12, highlighting the importance
of the location of the nitrogen atom in the pyridine ring.
amide compounds directly. Alternatively, acids
7 were first
converted to the corresponding methyl esters 8. Subsequent treat-
ment of 8 with primary amines in the presence of potassium
bis(trimethylsilyl)amide yielded the desired amide compounds.
This alternative route was especially useful for the incorporation
of nitrile groups into the scaffold at a late stage.
In addition to these six-membered ring heteroaryl analogs, sev-
eral five-membered ring heteroaryl analogs were also prepared
(Table 2). Although unsubstituted thiazole 17 was only weakly
active, simple modification of the thiazole ring by installation of
a 4-methly group (18) resulted in substantially enhanced potency.
As can often be the case in allosteric modulator chemotypes, large
changes in potency were observed with quite minor structural
modifications. For example, both thiadiazole 19 and 4-trifluorom-
ethylthiazole 20 were inactive up to the highest concentration
tested (30 lM) in spite of the fact that each compound was quite
closely related to 18 from a structural standpoint. Triazole deriva-
tives 21 and 22 were also both inactive up to the highest concen-
tration tested.
Figure 2. mGlu5 NAM HTS hit and early analogs.
Felts, Andrew S.
