Discovery of 4-arylquinoline-2-carboxamides, highly potent and selective class of mGluR2 negative allosteric modulators: From HTS to activity in animal models

Article abstract of DOI:10.1016/j.bmcl.2020.127066

Antagonism of the mGluR2 receptor has the potential to provide therapeutic benefit to cognitive disorders by elevating synaptic glutamate, the primary excitatory neurotransmitter in the brain. Selective antagonism of the mGluR2 receptor, however, has so far been elusive, given the very high homology of this receptor with mGluR3, particularly at the orthosteric binding site. Given that inhibition of mGluR3 has been implicated in undesired effects, we sought to identify selective mGluR2 negative allosteric modulators. Herein we describe the discovery of the highly potent and selective class of mGluR2 negative allosteric modulators, 4-arylquinoline-2-carboxamides, following a successful HTS campaign and medicinal chemistry optimization, showing potent in vivo efficacy in rodent.

Full text of DOI:10.1016/j.bmcl.2020.127066

Bioorganic&MedicinalChemistryLetters30(2020)127066
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of 4-arylquinoline-2-carboxamides, highly potent and selective
class of mGluR2 negative allosteric modulators: From HTS to activity in
animal models
Youheng Shua, Tracy L. Diamondc, James C. Hersheyb, Shaei Huanga, Brian C. Magliaroc,
Julie A. O'Brienc, Kelly-Ann S. Schlegela, Vanita Purib, Victor N. Uebeleb, Jason M. Uslanerb,
⁎
Cheng Wanga, Antonella Conversoa,
a Departments of Medicinal Chemistry, Merck & Co., Inc., West Point, PA 19486, USA
b Neuroscience Biology Discovery, Merck & Co., Inc., West Point, PA 19486, USA
c Pharmacology, Merck & Co., Inc., West Point, PA 19486, USA
A R T I C L E I N F O
A B S T R A C T
Keywords:
Antagonism of the mGluR2 receptor has the potential to provide therapeutic benefit to cognitive disorders by
elevating synaptic glutamate, the primary excitatory neurotransmitter in the brain. Selective antagonism of the
mGluR2 receptor, however, has so far been elusive, given the very high homology of this receptor with mGluR3,
particularly at the orthosteric binding site. Given that inhibition of mGluR3 has been implicated in undesired
effects, we sought to identify selective mGluR2 negative allosteric modulators. Herein we describe the discovery
of the highly potent and selective class of mGluR2 negative allosteric modulators, 4-arylquinoline-2-carbox-
amides, following a successful HTS campaign and medicinal chemistry optimization, showing potent in vivo
efficacy in rodent.
Metabotropic glutamate receptor 2
mGluR2 antagonists
High throughput screening
Cognition
Alzheimer’s disease
The role of glutamate in cognitive processes including memory ac-
quisition, consolidation and retention has been established both in
animal studies and in the clinic. Inhibiting ionotropic glutamate re-
ceptors, such as NMDA and AMPA receptors, disrupts cognition,1
whereas enhancing ionotropic receptor activity is pro-cognitive.2,3
However, there are clear challenges associated with over-activating
NMDA and AMPA receptors including the potential for inducing sei-
zures. Alternative approaches to elevating glutamatergic tone while
keeping desensitization processes intact have focused on antagonizing
the presynaptic metabotropic glutamate receptors. Indeed, antagonists
of mGluR2/3 show AMPA-dependent efficacy in preclinical cognition
assays.4,5 However, it has not been clear whether antagonism at one or
both receptors is necessary or sufficient for efficacy. Dual mGluR2/3
inhibitors existed in the literature.6 however, modulation of mGluR3
receptors has been implicated in undesired pharmacodynamic effects
including hyperlocomotion and increased wake.7,8 We therefore set out
to identify mGluR2-selective compounds with the potential to differ-
entiate from other glutamatergic mechanisms based on improved safety
margins. Due to the high homology between the mGluR2 and mGluR3
receptors, the team’s strategy to achieve high selectivity was to focus
efforts on the identification of mGluR2 negative allosteric modulators
rather than target the orthosteric site. We describe herein the discovery
of the most potent and selective class of mGluR2 negative allosteric
modulators known to date, the quinoline carboxamides. Since its dis-
covery,9 there has been significant activity in the field and several other
promising selective mGluR2 negative allosteric modulators from the
same class have been reported.10–13
We undertook a FLIPR-based high throughput screen which pro-
vided a number of high-quality classes (see Supporting information for
assay details). One interesting chemotype is represented by compound
1 (Fig. 1), belonging to the 5-lipooxygenase inhibitors class14 pre-
viously discovered in our laboratories. While the in vitro profile of 1 is
very promising, we observed low brain penetration in rat PK studies,
likely due to high plasma protein binding, very low solubility and ac-
tivity in the PGP assay in both human and rat.15
To alleviate these liabilities, we proceeded to optimize 1 by first
addressing the primary amide, whose polar nature we hypothesized
might be at the root of poor brain permeability. However, we quickly
realized that the amide is key to potency as shown in Table 1. Removal
of the amide (8) as well as its mono- (2) and bis- (3) methylation
⁎ Corresponding author.
E-mail address: antonella_converso@merck.com (A. Converso).
https://doi.org/10.1016/j.bmcl.2020.127066
Received 6 December 2019; Received in revised form 20 February 2020; Accepted 25 February 2020
Availableonline28February2020
0960-894X/©2020ElsevierLtd.Allrightsreserved.

Y. Shu, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127066
Table 2
SAR for the replacement of C4 fluorophenyl in 1.
Fig. 1. In vitro profile of compound 1.
Table 1
SAR for the replacement of primary amide in 1.
R
Cmpd
PGP h, r
2,2, 10.9
mGluR2 antag FLIPRI IC50
13 nM
1
9
68 nM
10
144 nM
11
12
36 nM
36 nM
R
Cmpd
mGluR2 antag FLIPRI IC50
13 nM
1
9.1, 4.1
2.1, 1.0
2
3
> 10,000 nM
> 10,000 nM
13
14
15
25 nM
2700 nM
19 nM
2.1, 1.0
4
5
238 nM
16
> 10,000 nM
> 10,000 nM
both enantiomers. A significant breakthrough was the realization that
removing the C7 substituent altogether provides a clear improvement in
brain penetration in both human and rat (21) with only 10-fold loss in
potency. We therefore elected to rebuild the C7 vector and explore non-
heteroaryl substituents to improve the overall profile of the class, in the
hopes of regaining activity while retaining the excellent properties in
21. We indeed found that secondary amines such as 22 and 23 possess
good potency and physicochemical properties while retaining good PGP
values. More strikingly, we also observed exquisite and unique se-
lectivity against mGluR3. We then proceeded to modify this vector and
to lower the basicity of the amine, we synthesized corresponding amide
analogue 24, which maintained activity. Ultimately, the introduction of
a second carbonyl group as in succinimide-substituted compound 25
resulted in a 20-fold potency improvement. Similar effects were ob-
served with the corresponding 6-membered analogues containing pi-
peridine-2,6-dione 27 and piperidin-2-one 26.
6
> 10,000 nM
7
8
409 nM
> 10,000 nM
resulted in complete loss of activity. Interestingly, ketone 4 and nitrile 7
retain moderate levels of potency, while sulfone 5 and pyridine 6 are
not tolerated.
We then proceeded to the exploration of the C-4 fluorophenyl group
on the quinoline (Table 2). Small substituents are tolerated at all po-
sitions on the aryl (9–11), with o- and p-substitution being preferred. At
the p-position, F and Cl are essentially equipotent (13), while in-
troduction of a larger substituent (iPr, 14) causes a drastic potency loss.
Bis-substitution is tolerated as in compound 15. While conversion of the
aryl to a 2-pyridine retains potency, it also decreases brain penetration
potential as measured by PGP (11.5 and 27.8 in human and rat, re-
spectively). This is a general trend with this position, where we saw
comparable PGP values among the various aryl substituted quinolines
but significant erosion in PGP with most heterocyclic substituents we
tested.
The good activity seen with C7-unsubstituted quinoline 21 gave us
an opportunity to assess the core and test the hypothesis of whether an
internal hydrogen bond between the primary amide and the quinoline
nitrogen favors the bioactive conformation (Table 4). This hypothesis
was suggested by the observation that the addition of a second ortho-
nitrogen to the amide as in quinazoline 28 results in complete loss of
potency. To understand whether the loss was due to an unfavorable
interaction of the 3-nitrogen with the binding pocket or a competing
internal hydrogen bonding of the 3-nitrogen to the carboxamide, we
prepared compounds 29 and 30. In 29, the carboxamide group is in-
ternally locked by cyclization. While we did lose activity with this
modification, 29 still presents sub micromolar activity. However,
The C-7 vector was the next area to examine (Table 3) and we
realized that the benzylic triazole has the highest impact on properties
as activity, PGP, and PPB can all be modulated by appropriate mod-
ification of that moiety. Different 5-membered heterocycles are toler-
ated at C7 (e.g. 17–19). Substitution at the α-methyl is also tolerated as
in 20-(+) and 20-(−), but comes with drastic increase in rat PGP for
2

Y. Shu, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127066
Table 3
Table 4
SAR for the replacement of C7 triazole in 1 and selectivity against mGluR3.
Quinoline carboxamide core SAR.
R
Cmpd
28
mGluR2 antag FLIPRI IC50
> 10,000 nM
29
30
31
32
404 nM
R
Cmpd
PGP h, r
mGluR2 antag
FLIPRI IC50
mGluR3 antag
FLIPRI IC50
1
2.2, 10.9 13 nm
3900 nM
> 10,000 nM
17
18
19
1.7, 21
1, 5.2
16 nm
10 nm
16 nm
3500 nM
> 10,000 nM
> 10,000 nM
367 nM
1378 nM
20(+) 6.3, 28.3 12 nm
> 10,000 nM
> 10,000 nM
20(−) 37.23.5
19 nm
21
22
0.4, 0.4
0.4, 0.5
149 nm
61 nm
> 10,000 nM
> 10,000 nM
33
> 10,000 nM
23
24
0.5, 0.9
68 nm
91 nm
> 10,000 nM
> 10,000 nM
25
1.5, 7.0
9 nm
> 10,000 nM
showed complete loss of activity. Ultimately, none of the constrained
analogues designed showed comparable levels of potency to the un-
substituted primary amide, which was again revealed to be impervious
to substitution.
26
27
58 nm
21 nm
> 10,000 nM
> 10,000 nM
The potency and selectivity of 25, coupled with its excellent phar-
macokinetic properties (see Fig. 2), fit the profile for a desirable in vivo
tool compound to leverage towards achieving preclinical in vivo vali-
dation for this mechanism in relevant animal models.
1.1, 3.5
Our first goal was to demonstrate in vivo target modulation with 25.
To this end, we tested the ability of 25 to reverse the effect of
LY37926816, a known mGluR2/3 orthosteric agonist. LY379268 in-
hibits amphetamine-induced hyperactivity and compound 25 is able to
reverse the effect of LY379268 in mouse at a dose of 10 mpk (Fig. 3).
The mouse delayed non-match to position assay was selected as
primary efficacy assay for the program, as it has been shown to be re-
sponsive to non-selective mGluR2/3 antagonists.6 This assay assesses
working memory. Scopolamine, a muscarinic acetylcholine receptor
antagonist, was used to disrupt performance on this task given that
reintroduction of the ortho-nitrogen in 30 caused a complete loss of
potency, suggesting that this substitution is not tolerated and is the
likely cause of the loss in activity in going from 21 to 28. To interrogate
the effect of the amide HBD-HBA pair position, we then designed
matched pair 31 and 32, where we observed a slight preference for a
more extended HBA as in 31. Six-membered cyclic analogue 33 also
3

Y. Shu, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127066
Fig. 2. Overall profile of compound 25. For details of GTPγS assay, see Supplementary materials.
cholinergic deficits accompany Alzheimer’s disease and that the sco-
polamine impairment is sensitive to non-selective antagonists. Com-
pound 25 shows clear efficacy in this model, by affording significant
dose-dependent reversal of scopolamine impairment. Plasma exposure
were collected in separate animals dosed in the same manner and
correlate with compound exposures (see Fig. 4).
In conclusion, via high throughput screening and subsequent SAR
optimization, we discovered quinoline carboxamides as a highly potent
and selective class of mGluR2 antagonists.11 Compound 25 is an orally
bioavailable mGluR2 negative allosteric modulator with excellent brain
permeability. In vivo, 25 reverses the effect of mGluR2 agonist
LY379268 in amphetamine-induced hyperlocomotion and shows good
efficacy in the mouse delayed non-match to position assay at 10mpk.
Since the discovery of this class,8 several related carboxamides with
similar mGluR2 selectivity profiles have been reported by others9–12
and are in the clinical and preclinical space as the field continues to
advance.
Declaration of Competing Interest
Fig. 4. Dose-responsive effects of 25 in delayed non-match to position in sco-
polamine-impaired mice. Values are % change from baseline. * indicates sig-
nificant difference (p < 0.05) from 0.17
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
Fig. 3. Compound 25 displays in vivo mGluR2/3 target engagement by reversing LY379268 inhibition of Amphetamine-induced hyperactivity.
4

Y. Shu, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127066
Appendix A. Supplementary data
WO 2013066736 A1 20130510.
10. Bollinger KA, Felts AS, Brassard CJ, et al. ACS Med Chem Lett. 2017;8(9):919–924.
11. Childress ES, Wieting JM, Felts AS, et al. J. Med. Chem. 2019;62(1):378–384.
12. For PET work with selective mGluR2 NAM compounds see a. Zhang, X.; Kumata, K.;
Yamasaki, T.; Cheng, R.; Hatori, A.; Ma, L.; Zhang, Y.; Xie, L.; Wang, L.; Kang, H. J.;
Sheffler, D. J.; Cosford, N. D. P., Zhang, M.-R., Liang, S. H., ACS Chemical
Neuroscience, 2017, 8(9), 1937-1948; b. Kumata K.; Hatori A.; Yamasaki T.; Zhang Y.
; Mori W.; Fujinaga M.; Xie L.; Nengaki N.; Zhang M. R.; Bioorg Med Chem., 2019,
1;27(3):483-491.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmcl.2020.127066.
References
13. Felts AS, Rodriguez AL, Smith KA, et al. J. Med. Chem. 2015;58(22):9027–9040.
14. Preparation of triazole derivatives as leukotriene biosynthesis inhibitors , Grimm, E.
L.; Ducharme, Y.; Frenette, R.; Friesen, R.; Gagnon, M.; Juteau, H.; Laliberte, S.;
Mackay, B.; Gareau, Y.; U.S. Pat. Appl. Publ. (2008), US 20080188521 A1 20080807.
15. The PGP assay assesses LLC-PK1 B-A/A-B efflux ratio as a measure of endogenous
transport. For a review, see K. Bleasby; X. Y. Chu; R. Evers, (2010) in: Enzyme- and
transporter based drug-drug interactions (Eds. Pang, K. S.; Rodrigues, A. D.; Peter, R.
M., Springer), 237-255.
1. Robbins TW, Murphy ER. Trends Pharmacol. Sci. 2006;27:141–148.
2. Francis PT. Neurodegener. Dis. 2008;5:241–243.
3. Morrow JA, Maclean JK, Jamieson C. Curr. Opin. Drug Discov. Devel. 2006;9:571–579.
4. Higgins GA, Ballard TM, Kew JN, et al. Neuropharmacology. 2004;46:907–917.
5. Shimazaki T, Kaku A, Chaki S. Eur. J. Pharmacol. 2007;575:94–97.
6. Woltering TJ, Wichmann J, Goetschi E, et al. Bioorg. Med. Chem. Lett.
2010;20(23):6969–6974.
7. Wood CM, Wafford KA, McCarthy AP, et al. Neuropharmacology. 2018;140:246–259.
8. Jin LE, Wang M, Galvin VC, et al. Cereb Cortex. 2018;28(3):974–987.
9. Bungard, C. J.; Converso, A.; De Leon, P.; Hanney, B.; Hartingh, T. J.; Manikowski, J.
J.; Manley, P. J.; Meissner, R.; Meng, Z.; Perkins, J. J.; et al, PCT Int. Appl. (2013),
16. Monn JA, Valli MJ, Massey SM, et al. J. Med. Chem. 1999;42(6):1027–1040.
17. Puri V, Wang X, Vardigan JD, Kuduk SD, Uslaner JM. Behav. Brain Res.
2015;287:96–99.
5