Fragment and Structure-Based Drug Discovery for a Class C GPCR: Discovery of the mGlu5 Negative Allosteric Modulator HTL14242 (3-Chloro-5-[6-(5-fluoropyridin-2-yl)pyrimidin-4-yl]benzonitrile)

Article abstract of DOI:10.1021/acs.jmedchem.5b00892

Fragment screening of a thermostabilized mGlu₅ receptor using a high-concentration radioligand binding assay enabled the identification of moderate affinity, high ligand efficiency (LE) pyrimidine hit 5. Subsequent optimization using structure-based drug discovery methods led to the selection of 25, HTL14242, as an advanced lead compound for further development. Structures of the stabilized mGlu₅ receptor complexed with 25 and another molecule in the series, 14, were determined at resolutions of 2.6 and 3.1 ?, respectively. (Figure Presented).

Full text of DOI:10.1021/acs.jmedchem.5b00892

Subscriber access provided by KUNGL TEKNISKA HOGSKOLAN
Article
Fragment and Structure-Based Drug Discovery for a Class C GPCR:
Discovery of the mGlu5 negative allosteric modulator HTL14242
(3-chloro-5-[6-(5-fluoropyridin-2-yl)pyrimidin-4-yl]benzonitrile)
John Christopher, Sarah J Aves, Kirstie A Bennett, Andrew S Doré, James C Errey, Ali Jazayeri, Fiona
H. Marshall, Krzysztof Okrasa, Maria J Serrano-Vega, Ben Tehan, Giselle R Wiggin, and Miles Congreve
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b00892 • Publication Date (Web): 30 Jul 2015
Downloaded from http://pubs.acs.org on August 12, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Fragment and StructureꢀBased Drug Discovery for a Class C GPCR:
Discovery of the mGlu5 negative allosteric modulator HTL14242 (3ꢀ
chloroꢀ5ꢀ[6ꢀ(5ꢀfluoropyridinꢀ2ꢀyl)pyrimidinꢀ4ꢀyl]benzonitrile)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
John A. Christopher,* Sarah J. Aves, Kirstie A. Bennett, Andrew S. Doré, James C. Errey, Ali
Jazayeri, Fiona H. Marshall, Krzysztof Okrasa, Maria J. SerranoꢀVega, Benjamin G. Tehan,
Giselle R. Wiggin, Miles Congreve.
Heptares Therapeutics Ltd., BioPark, Welwyn Garden City, Hertfordshire, AL7 3AX, U.K.
Corresponding Author
*Phone: +44 (0)1707 358731. Fax: +44 (0)1707 358640. Email: john.christopher@heptares.com.
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 47
1
2
3
4
ABSTRACT
5
6
7
8
Fragment screening of a thermostabilized mGlu5 receptor using a highꢀconcentration radioligand
binding assay enabled the identification of moderate affinity, high ligand efficiency (LE)
pyrimidine hit 5. Subsequent optimization using structureꢀbased drug discovery methods led to
the selection of 25, HTL14242, as an advanced lead compound for further development.
Structures of the stabilized mGlu5 receptor complexed with 25 and another molecule in the
series, 14, were determined at resolutions of 2.6 Å and 3.1 Å respectively.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION
Glutamate is the major excitatory neurotransmitter in the central nervous system. When released,
glutamate exerts its effects through ionotropic glutamate receptors (the ligandꢀgated ion channels
2ꢀaminoꢀ3ꢀ(3ꢀhydroxyꢀ5ꢀmethylisoxazolꢀ4ꢀyl)propanoic acid (AMPA) receptor, kainate receptor
and NꢀmethylꢀD
ꢀaspartic acid (NMDA) receptor) and metabotropic glutamate (mGlu) receptors.1
Activation of glutamatergic ion channels leads to fast glutamate transmission whilst activation of
the mGlu receptors is involved in modulating the strength of synaptic transmission.
The metabotropic glutamate receptors are a family of eight G proteinꢀcoupled receptors (GPCRs)
which are subdivided into three groups based on sequence homology, preferred signaling
pathways and pharmacology.2 The mGlu5 receptor, predominately located postꢀsynaptically,
belongs to group I and is linked to the Gq/11 family of G proteins. The mGlu5 receptor is
abundantly expressed in the striatum, hippocampus, amygdala and frontal cortex.3
ACS Paragon Plus Environment
2

Page 3 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structurally, the mGlu receptors consist of a large extracellular binding domain, the Venus
flytrap (VFT) domain, linked via a cysteineꢀrich domain to the seven transmembrane domain
(TMD) which is a feature common to all GPCRs. The VFT domain is a biꢀlobed structure which
contains the orthosteric (glutamate) binding site. Historical drug discovery efforts focusing on
the orthosteric site typically resulted in the identification of acidic glutamateꢀbased compounds.
Generally these compounds have failed to achieve appropriate pharmacokinetic profiles, brainꢀ
penetrant characteristics and subtype selectivity.4 Subsequently, greater success and a wider
diversity of chemotypes have arisen from efforts targeting alternative binding sites (allosteric
sites) within the TMD. 5ꢀ8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compounds binding at an allosteric site which do not activate the receptor in their own right but
potentiate receptor response are termed positive allosteric modulators (PAMs). Compounds
which bind to an allosteric site and act as nonꢀcompetitive antagonists are termed negative
allosteric modulators (NAMs). At least two allosteric sites have been identified in the
transmembrane domain of the mGlu5 receptor, commonly referred to as the ‘MPEP site’ and the
‘CPPHA site’ (due to initially identified compounds which bind to these regions). NAMs bind to
the MPEP site whilst PAMs have been identified which bind to either the MPEP site or the
CPPHA site.9,10 We have recently published the crystal structure of the TMD of mGlu5 with the
NAM mavoglurant bound to the MPEP site.11
Dysfunctions in glutamatergic signaling have been linked to numerous disorders which could at
least be partially corrected by a NAM of mGlu5. NAMs of the mGlu5 receptor have been, or are
being tested clinically in: fragile X syndrome (FXS),12 Parkinson’s Disease levodopaꢀinduced
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 47
1
2
dyskinesias (PDꢀLID),13 anxiety,14 gastroesophageal reflux disease (GERD),15 neuropathic
pain,16 obsessiveꢀcompulsive disorder (OCD),17 migraine,18 chorea in Huntington’s disease19 and
depression.20 The most advanced trials have been in FXS where mavoglurant (1, Chart 1) from
Novartis and basimglurant (RO4917523, 2)21 from Roche have completed phase III and II trials
respectively, however these studies failed to show efficacy.22 Phase II data reported by Roche on
the use of basimglurant as an adjunctive in major depressive disorder shows encouraging results
and strengthens the approach of targeting glutamatergic systems to develop novel antidepressants
with alternative mechanisms of action for use in treatmentꢀresistant depression.20 Clinical
development of Addex Therapeutic’s NAM dipraglurant 3 is ongoing in rare dystonias and
Novartis have evaluated mavoglurant in a phase II OCD trial.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chart 1. Literature mGlu5 NAMs mavoglurant 1, basimgluant 2, dipraglurant 3 and GRNꢀ529 4.
N
F
N
O
H
N
H
HO
O
N
Cl
basimglurant
mavoglurant
2
1
O
N
N
N
N
N
N
O
F
GRN-529
4
dipraglurant
3
Mavoglurant, basimglurant, dipraglurant, and GRNꢀ529 (Chart 1) are all examples of mGlu5
NAM compounds containing a central acetylene flanked by two aromatic or nonꢀaromatic
ACS Paragon Plus Environment
4

Page 5 of 47
Journal of Medicinal Chemistry
1
2
groups. The acetylene moiety has the potential to be susceptible to metabolic activation.23 A
report from Pfizer demonstrates extensive glutathione conjugation occurring at the acetylene of
GRNꢀ529 4 (Chart 1),24,25 implicating this as a structural link to hepatotoxicity observed in 8ꢀ
week rat and nonꢀhuman primate regulatory toxicology studies.26 Workers at BristolꢀMyers
Squibb have reported the formation of glutathione adducts in a series of biheteroaryl acetylene
mGlu5 PAMs.27 A subsequent investigation of the relationship of acetylene reactivity to the
structure of the adjacent aryl groups concluded that GSH conjugation could be deꢀrisked by
flanking the triple bond with only one electronꢀwithdrawing aryl ring. Clinical development of
raseglurant, an earlier acetylene molecule from Addex, was halted in 2009 after raised alanine
transaminase (ALT) levels, which are considered to be predictive of the potential for drug
induced liver injury, were observed in a migraine prophylaxis trial.28 Taken together, these
reports indicate that whilst the inherent bioactivation risk residing in the acetylene can be
managed, an opportunity to identify and progress nonꢀacetylene compounds as mGlu5 NAMs is
evident. For the above reasons, recent efforts in the mGlu5 NAM arena have focused on a second
generation of nonꢀacetylene compounds which lack the bioactivation risk.6,7 Herein we report
our own efforts to identify and characterize a novel, nonꢀacetylene, negative allosteric modulator
of the mGlu5 receptor, identified through fragment screening.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Fragment Screen
Within the past decade there has been a dramatic upsurge in the number of published GPCR Xꢀ
ray structures,29 which has ushered in a new era of structureꢀbased drug discovery (SBDD) for
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 47
1
2
3
4
5
6
7
8
9
this highly important protein family. In tandem with SBDD, fragmentꢀbased drug discovery
(FBDD) represents a powerful strategy for the identification of small, efficient molecules, and
has become well established in soluble protein classes such as proteases and kinases.30ꢀ32 We
have previously reported the discovery of high affinity leads for the β1ꢀadrenergic receptor (β1ꢀ
AR) using structureꢀbased and fragmentꢀbased techniques.33 As there is ample precedent for
mGlu5 to be modulated by lowꢀmolecular weight ligands6,7 we felt that the target was ideally
positioned for a fragment screening approach to identify hit material. In contrast to β1ꢀAR, where
our approach used surfaceꢀplasmon resonance (SPR) as the primary fragment screen, we opted to
use a highꢀconcentration radioligand binding assay with a stabilized receptor (StaR® protein) for
mGlu5 hit discovery. During the course of stabilization of the receptor to enable Xꢀray
crystallographic studies, an intermediate mGlu5 StaR protein was generated with three residues
mutated from the wildꢀtype sequence. The mutations thermostabilized the receptor, significantly
increased expression over wildꢀtype protein, and in common with other StaR proteins we have
used for fragmentꢀscreening, conferred an enhanced tolerance of high DMSO concentrations.
Additionally, as the thermostabilization was undertaken in the presence of the NAM radioligand
[3H]ꢀMꢀMPEP (2ꢀ[(3ꢀmethoxyphenyl)ethynyl]ꢀ6ꢀmethylpyridine),34 the receptor was biased into
the physiologically relevant conformational state for our hit discovery campaign.11 A final assay
concentration of 10% DMSO was viable with the mGlu5 StaR, facilitating screening of ligands at
high concentration whilst ensuring they remained in solution. A set of approximately 3,600
fragments (including a proportion which had been selected specifically for our Class C fragment
hit discovery efforts) was screened at 30 µM and 300 µM concentrations in radioligand binding
format using [3H]ꢀMꢀMPEP in HEK293 cell membranes expressing the mGlu5 StaR protein. No
acetylene containing molecules were present in the screening set. Subsequently, 178 active
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
6

Page 7 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
fragments were screened in 8ꢀpoint concentrationꢀresponse curves to define equilibrium
dissociation constant (pKi) values. A number of promising fragment hits emerged from the hit
discovery campaign, including the ligandꢀefficient,35 moderate affinity, pyrimidine 5 (Chart 2,
pKi 5.6, LE 0.36), the progression of which to an advanced lead molecule is the subject of this
manuscript.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chart 2. mGlu5 fragment screening hit 5.
O
O
5
N
pKi 5.6
N
N
N
LE 0.36
Design and Synthesis
The binding mode of 5 and analogues was initially derived from docking into an experimentally
enhanced homology model of mGlu5, refined using literature site directed mutagenesis
information and our own extensive data set derived from receptor stabilization experiments.
During the course of the project, a number of literature compounds, including mavoglurant 1,11
and compounds within our own series (see below), were successfully coꢀcrystallized in the
receptor which further assisted our efforts. Our initial synthetic strategy for fragment validation
centered on a detailed evaluation of literature mGlu5 chemotypes to identify common motifs
within high affinity molecules. Hybridization of these motifs with 5 yielded compounds which
were examined inꢀsilico in the context of the mGlu5 homology model, with molecules for
synthesis selected by medicinal and computational chemists after consideration of their proposed
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 47
1
2
3
4
5
6
7
8
9
binding modes (Tables 1 and 2). One such motif is 3ꢀfluoroꢀ5ꢀcyanophenyl, which features in
several high affinity mGlu5 NAM acetylene and nonꢀacetylene chemotypes from a number of
different companies, and as such was a good candidate for hybridization with 5. Analogues of 5
were prepared using the routes summarized in Schemes 1 and 2.36 Intermediates 18 (Scheme 1)
were either commercially available or were readily synthesized from 4,6ꢀdichloropyrimidine
under mild SNAr conditions, subsequent SuzukiꢀMiyaura coupling led to 6-14 and 17.
Compounds 15 and 16 were analogously accessed from the commercially available
halopyridines 19 and 20. Scheme 2 describes the routes employed to variants which replace the
pyrazole by pyridines or diazines (Table 2). Pyridines 21ꢀ28 and pyridazine 29 employed
successive palladiumꢀmediated Stille and SuzukiꢀMiyaura couplings (or the reverse sequence) to
yield final compounds, with the aryl stannane being synthesized from the bromoꢀprecursor if
unavailable from commercial suppliers. For 25 replacement of the Stille protocol by a nickelꢀ
mediated Negisihi coupling was additionally used, which removed the use of stoichiometric tin
in the reaction sequence. Bisꢀpyrimidine 30 was synthesized via formation of a βꢀketo ester,
followed by pyrimidine ring synthesis with formamidine and palladiumꢀcatalysed arylation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
8

Page 9 of 47
Journal of Medicinal Chemistry
1
2
Scheme 1. Synthesis of 6ꢀ17.a
3
4
5
6
7
8
9
R2
R1
CN
Cl
Cl
N
b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
N
N
-
,
6 14 17
N
N
N
N
N
N
Cl
N
18
R3
R3
R2
R1
CN
Hal
b
X
X
19 Hal = Cl, X = N, Y = CH
20 Hal = Br, X = CH, Y = N
,
15 16
N
N
N
Y
N
Y
R3
a
Reagents and conditions: (a) Pyrazole or 4ꢀfluoropyrazole, K2CO3, DMF, rt; (b) Substituted
aryl boronic acid or ester, PdCl2(dppf), Cs2CO3 or 1M(aq) Na2CO3, 1,4ꢀdioxane / H2O, 80ꢀ90ºC.
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 47
1
2
Scheme 2. Synthesis of 21ꢀ30.a
3
4
5
6
7
8
9
R2
R1
CN
Cl
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
N
Br
N
SnR3
N
N
b
a
c
X
Y
N
X
Y
X
Y
N
X
Y
21, 22, 26, 29
R2
R1
CN
R2
R1
CN
N
Cl
N
N
N
b or d
c
N
N
Y
Cl
Cl
N
23-25, 27, 28
R2
R1
CN
O
Cl
OEt
N
N
OH
N
N
N
N
f
c
e
O
N
N
O
N
N
N
N
30
a
Reagents and conditions: (a) R3SnSnR3 (R = Me or nꢀBu), Pd(PPh3)4, DME, 80ꢀ110ºC; (b)
Heteroaryl halide or stannane, Pd(PPh3)4, CuI, PhMe, 110ºC; (c) Substituted aryl boronic acid or
ester, PdCl2(dppf)•CH2Cl2, Cs2CO3 or 1M(aq) Na2CO3, 1,4ꢀdioxane, H2O, 80ꢀ90ºC; (d) (i) 2ꢀ
Bromoꢀ5ꢀfluoropyridine, iꢀPrMgCl, ZnCl2, THF, rt, (ii) Substituted chloropyrimidine, Pd(PPh3)4,
THF, 50ꢀ60ºC; (e) (i) Oxalyl chloride, DMF, 0ºC, (ii) EtOAc, LDA, THF, ꢀ78ºC; (f) (i)
H2NCH=NH•HCl, NaOMe, MeOH, rt, (ii) POCl3, rt.
Structure Determination of mGlu5 in complex with 14 and 25
To determine the structure of the mGlu5 TMD in complex with the allosteric modulators 14 and
25 a thermostabilized receptor (StaR) was generated. The receptor was thermostabilized in the
presence of the allosteric radioligand [3H]ꢀMꢀMPEP and contained six mutations, none of which
are located in close proximity to the allosteric site.11 To further aid crystallization in lipidic cubic
phase (LCP), the extracellular domains were removed from the Nꢀterminus (residues 2ꢀ568), the
ACS Paragon Plus Environment
10

Page 11 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
unstructured Cꢀterminus was truncated (residues 833ꢀ1153) and finally T4ꢀlysozyme (T4L)
inserted into intracellular loop (ICL) 2 between Lys678 and Lys679. The structure of mGlu5
TMD bound to 14 and 25 was determined to 3.1 and 2.6 Å respectively, through merging
diffraction data from multiple crystals grown in lipidic cubic phase (LCP). The structures were
solved by molecular replacement using PDB ID: 4OO9.11 Details of data collection and
refinement statistics for both structures are given in Supporting Information Table 3, in both
cases the Rfree set was imported from structure factor data deposited with PDB ID: 4OO9.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Strong and unambiguous electron density was observed for both 14 and 25 in sigmaꢀA weighted
2m|Fo|ꢀ|DFc|, m|Fo|ꢀD|Fc| maps calculated using Phenix37 following molecular replacement using
the program Phaser38 from the CCP4 suite of programs39 (Figure 1). In general the protein
backbone of the two novel structures are in close agreement with the mavoglurant coꢀstructure
with Cα r.m.s.d values of 0.23 Å and 0.26 Å between 4OO9 and the 14 and 25 structures
respectively. In both the mGlu5 14 and 25 coꢀstructures the ligands are found in the allosteric
pocket (comprised of an upper and lower chamber connected by a narrow linker region) ~8 Å
from the receptor surface and in a similar position to that observed previously for mavoglurant.11
Both 14 and 25 are oriented with an approximate 30 ° tilt (relative to the central axis of the TM
bundle) in a pocket defined by residues from TM2, TM3, TM5, TM6 and TM7. The pyrazole
and pyridyl ring systems of 14 and 25 respectively sit in a pocket between Ala8107.40
(superscripts indicate Ballesteros−Weinstein nomenclature40) and Pro6553.40, bordered on one
side by Ile6252.46, Gly6282.49, Ser6543.39, Ser6583.43 and on the other by Tyr6593.44 and Ser8097.39
which makes a hydrogen bond to the nitrogen within the respective pyrazole and pyridyl ring
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 47
1
2
3
4
5
6
systems (Figure 1). A number of these molecular interactions have recently been predicted in
modelling studies.41
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Both pyrazole and pyridyl ring systems are connected to a pyrimidine linker which traverses the
narrow channel in the allosteric pocket formed between Tyr6593.44, Ser8097.39, Val8067.36 and
Pro6553.40. In comparison to the mGlu5 mavoglurant coꢀstructure where the acetylene group of
the small molecule is found crossing the linker region, here the sidechain of Val8067.36 is shifted
outwards from the binding site by ~1 Å, sterically permitting the pyrimidines to traverse the
narrow channel.
The upper chamber of the allosteric pocket for the two novel structures presented here differs
most significantly from that previously determined for mGlu5 in complex with mavoglurant.
TM6 of mGlu5 contains the highly conserved Trp7856.50 equivalent of the central toggle switch
within the FxxCWxP6.50A motif in rhodopsin and other class A receptors. In complex with
mavoglurant Trp7856.50 is found rotated out of the helical bundle towards TM5 with a potential
bifurcated hydrogen bond formed from the mainꢀchain carbonyl of Leu7445.44 to Gly7485.48 to the
NH of the indole ring. In both the 14 and 25 coꢀstructures however the sideꢀchain of Trp7856.50 is
found in a χ1 trans orientation placing the indole ring system towards the centre of the helical
bundle where the NH of the indole moiety forms a hydrogen bond with Ser8097.39. This
rotameric change drastically alters the shape of the allosteric site resulting in a much smaller
pocket defined primarily by Trp7856.50, to the bottom, and Phe7886.53 & Met8027.32, to the top,
with Ser8057.35, Val8067.36 & Leu7445.44 forming the sides. The observed rearrangement of
residues within the allosteric pocket would occlude binding of mavoglurant where the sideꢀchain
ACS Paragon Plus Environment
12

Page 13 of 47
Journal of Medicinal Chemistry
1
2
of Trp7856.50 would severely clash with the previously observed position of the saturated bicyclic
ring system. The overall properties of the upper chamber of the mGlu5 allosteric pocket, whilst
remaining mostly hydrophobic in nature, now exhibit a greater proportion of aromatic character
fitting with the SAR of the series (detailed below) and highlight the significance of utilizing a top
ring system with the correct electronic character. In the case of 14 the 2ꢀfluoroꢀ3ꢀchloroꢀ5ꢀ
cyanophenyl ring system is able to make complementary electronic interactions and additionally
the 5ꢀcyano substituent makes a water mediated hydrogen bond with the backbone carbonyl of
Val7405.40. In a similar fashion the 3ꢀchloroꢀ5ꢀcyanophenyl group of 25 is also able to make
complementary electronic interactions within the pocket and the water mediated hydrogen bond
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
to Val7405.40
.
Whilst the allosteric pockets of the novel structures described here differ significantly in shape to
mGlu5 with mavoglurant bound, the crystallographic water present in the lower chamber of the
allosteric pocket bridging between Thr7816.46, Tyr6593.44 and the backbone carbonyl oxygen of
Ser8097.39 is still observed. The water molecule is adjacent to a region where it has been
proposed that observed changes in ligand pharmacology result from perturbation of the network
around the water, pointing to a potential activation switch holding TM7, TM3 and TM6 together
at the base of the allosteric pocket.11 Such mode switching effects have been observed in a range
of chemotypes,42ꢀ47 but as detailed below in our series this phenomenon has not been observed in
the compounds evaluated for their functional pharmacology.
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 47
1
2
3
4
5
6
7
8
9
Figure 1. The mGlu5-StaR crystal structures with 14 and 25 bound in the allosteric site. A,
Diagram of ligand interactions within the allosteric site of mGlu5 with 14 bound. B, Diagram of
ligand interactions within the allosteric site of mGlu5 with 25 bound. mGlu5 in ribbon
representation coloured green. 14 and 25 in stick representation with carbon, nitrogen, oxygen,
chlorine and fluorine atoms oxygen atoms coloured magenta, blue, red, green and white
respectively. Selected residues from mGlu5 in stick representation with carbon, nitrogen and
oxygen atoms coloured grey, blue and red, respectively. TM helices are labeled and selected
hydrogen bonds depicted as dashed red lines with distances labelled in Å. In both figures the
extracellular portion of TM7 has been removed for clarity. FoꢀFc OMIT density for both ligands
contoured at 2.0σ calculated prior to inclusion of the small molecule in the model. C, D, 2D
diagrams of 14 and 25 ligand interactions in the mGlu5ꢀStaR allosteric pocket respectively.
Green lines denote πꢀstack interactions, dashed purple arrows denotes hydrogen bond (side
chain), solid purple arrows denote hydrogen bond (main chain), residues in cyan spheres denote
polar interactions, residues in green spheres denote hydrophobic interactions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
14

Page 15 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 47
1
2
3
4
Structure-Activity Relationships
5
6
7
8
When tested in the mGlu5 StaR fragment screening format described above, compound 6 (Table
1), which incorporated a 3ꢀfluoroꢀ5ꢀcyanophenyl motif, yielded higher affinity than the initial hit
(pKi 5.6 and 6.9 for 5 and 6 respectively) and was also significantly more ligand efficient (LE
0.36 and 0.47 for 5 and 6 respectively). In a radioligand binding assay under standard conditions
(1% DMSO final assay concentration, wildꢀtype human mGlu5) 6 had comparable affinity (pKi
7.2, Table 1) as did a small number of other molecules profiled in both assay formats, allowing
use of the wildꢀtype screening assay for the remainder of the drug discovery campaign.
Compounds with affinities over pKi 7 were additionally screened in an IPone accumulation
functional assay, which was configured to assess the potency (pIC50) of compounds in
antagonizing an EC80 concentration of the group 1 selective agonist Lꢀquisqualate. In general,
SAR trends in the binding assay were highly comparable to those in the functional assay. As
detailed above, whilst others have reported that subtle changes to compound structure can cause
modeꢀswitching no evidence of this behavior was observed in compounds in our series which
were progressed to the functional assay. The presence and position of the fluorine substituent in
these analogues had a significant impact on binding affinity, with alternative regiochemistries
(including 7), or removal (8) of the substituent, leading to a drop in affinity. Presence of the
nitrile moiety was also crucial, as removal, replacement, or hydrolysis to the primary amide of
this group led to significant drops in affinity (9ꢀ11). These findings indicate that the water
mediated hydrogen bonding contact with Val7405.40 noted above is an important contributor to
activity, along with the electronics of the aromatic ring structure. Keeping the nitrile in place
whilst replacing the 3ꢀfluoro substituent was encouraging, as switching fluoro to methyl (12) or
chloro (13) gave approximately tenꢀfold increases in binding affinity, and in the latter case a
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
16

Page 17 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
further tenꢀfold increase could also be gained by installing an additional 2ꢀfluoro substituent
(14). Investigation of the pyrimidine core of the molecule indicated that replacement by selected
pyridine regiochemistries was viable with a moderate loss of binding affinity (compare 15 and 16
with 12), whilst other pyridine or pyrimidine regioisomers and replacement by pyrazine or
pyridazine resulted in significant drops in affinity, as did substitution of the ring at either of the
available positions (data not shown). This SAR is consistent with the constrained nature of the
allosteric pocket in the region of the narrow linker channel occupied by the pyrimidine, and also
indicate that appropriately matching the electronic characteristics of the binding site in this
region is important. Tight SAR was also apparent in the region of the pyrazole, efforts to
substitute the ring led at best to compounds such as 17 with a moderate drop in affinity compared
to the parent molecule, no substituents examined were beneficial for affinity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R2
R1
Z
X
N
N
Y
R3
Table 1. In vitro profile of compounds 6ꢀ17.
RLM
t1/2
(min)
mGlu5
pKi
mGlu5
pIC50
X
Y
Z
R1
R2
R3
N
N
N
N
N
N
N
N
N
N
CN
CN
CN
H
H
F
F
H
H
F
H
H
H
H
H
7.2
6.6
6.4
nd a
nd
25
6
6
7
H
H
H
6.1
nd
nd
nd
8
5.1
nd
9
OMe
F
< 4.2
nd
10
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 47
1
2
3
4
5
6
7
8
9
N
N
N
N
CONH2
CN
H
H
H
F
F
H
H
H
H
H
H
F
< 4.2
8.4
8.4
9.3
7.6
7.7
7.7
nd
7.9
8.3
8.6
7.4
7.7
7.1
nd
10
12
20
51
43
52
11
12
13
14
15
16
17
Me
Cl
N
N
CN
N
N
CN
Cl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
CH
N
CN
H
H
H
Me
Me
Cl
CH
N
CN
N
CN
a nd = not determined.
Despite these steric constraints, insights from the structure of mGlu5 bound to 14 suggested that
replacement of the pyrazole ring by alternative heterocycles was viable. A focused investigation
was undertaken, which identified that 2ꢀpyridines and certain diazines were viable replacements,
whereas a number of other heterocycles were not beneficial. Compounds were synthesized using
the routes depicted in Scheme 2, and SAR is summarized in Table 2. Installation of a 2ꢀpyridyl
group yielded compounds with comparable affinities and potencies to the analogous pyrazoles
(compare 21 and 22 with 13 and 14 respectively). Substitution at the 4ꢀposition of the pyridine
ring was tolerated with fluoro (25) and cyano (26) substituents yielding an increase in affinity
and potency over the parent compound 21; methyl (23) and chloro (27) substituents were
essentially equipotent. Installation of the extra fluorine in the top aromatic ring did not
significantly impact upon affinity or potency (compare 28 to 25) but did appear to enhance in
vitro microsomal stability (discussed further below), and the 3,5ꢀdicyano compound 24 was also
tolerated. Pyridazine 29 and pyrimidine 30 retained good mGlu5 affinity, but in comparison to
pyridine 21 their functional potencies dropped significantly. Other pyridine regiochemistries
ACS Paragon Plus Environment
18

Page 19 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
were poorly tolerated, and taken together these findings highlight the importance of the hydrogen
bond between Ser8097.39 and the lower ring system.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R2
R1
CN
N
N
X
Y
N
Table 2. In vitro profile of compounds 21ꢀ30.
mGlu5
pKi
mGlu5
pIC50
RLM t1/2
(min)
X
Y
R1
R2
CH
CH
CH
CH
CH
CH
CH
CH
N
CH
CH
CMe
CF
H
F
Cl
Cl
Cl
CN
Cl
Cl
Cl
Cl
Cl
Cl
8.5
8.9
8.6
8.8
9.3
9.2
8.8
9.1
8.0
8.3
8.6
8.8
8.3
8.6
9.2
9.2
8.5
9.4
6.7
7.5
43
31
21
22
23
24
25
26
27
28
29
30
H
H
H
H
H
F
19
> 100
44
CF
CCN
CCl
CF
> 100
35
87
CH
N
H
H
nd a
nd
CH
a nd = not determined.
Pharmacokinetics, Receptor Occupancy and In Vivo Efficacy studies
As part of our routine in vitro profiling, we monitored stability in rat liver microsomes (RLM,
Tables 1 and 2). Across a number of examples it was observed that pyridyl variants of the series
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 47
1
2
3
4
5
6
7
8
9
typically had better in vitro stability in this system, for example compare 21, 25ꢀ28 with 13. To
examine the relationship of rat microsomal stability to in vivo PK parameters, we examined a
number of compounds in rat iv experiments, summarized in Table 3. In general, compounds with
moderate to high in vitro stability had moderate to high in vivo stability, and a small number of
compounds examined with low microsomal stability also exhibited low in vivo stability (for
example 12, RLM t1/2 10 min, Cl 70 mL/min/kg), allowing us to use the RLM experiment as a
filter to PK progression with a reasonable degree of confidence. Metabolite identification studies
on 12 revealed a rapidly appearing single major metabolite, which was identified as the product
of oxidation of the aromatic methyl group. The information allowed us to bias subsequent
compound design towards molecules with less metabolically vulnerable aromatic substituents.
Compounds in the series typically exhibited high brain:plasma ratios, with low, but appreciable,
unbound fractions in the CNS (approximated as the ratio of drug concentration in the
cerebrospinal fluid (CSF) to that in brain at the same time point).48 Selected compounds with
good rat iv PK parameters and appropriate overall in vitro profiles were further evaluated in dog
iv/po PK studies (Table 4). Compound 25 presented the best dog profile, with moderate
clearance, volume and halfꢀlife and excellent oral bioavailability, matching a similarly
encouraging oral profile in rat (25: rat Fpo approximately 100%, po t1/2 4.0 h). Rat ascending dose
oral studies additionally indicated that the compound had approximately linear PK at 3, 10 and
30 mg/kg. In rat, the PK parameters of 25 compare favourably to those of key clinical standards
such as dipraglurant (Fpo 48%, po t1/2 1.1 h)49 and mavoglurant (Fpo 32%, po t1/2 2.9 h).50
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
20

Page 21 of 47
Journal of Medicinal Chemistry
1
2
Table 3. Rat intravenous PK profiles of compounds 12, 21, 24ꢀ26.a
3
4
5
6
7
Cl
Vss
t1/2
(h)
Br : Pl (0.5, 2.0 h)
CSF : Pl (0.5, 2 h)
(mL/min/kg) (L/kg)
8
9
79
32
25
27
5
2.5
4.4
3.3
3.7
4.9
0.9
2.4
1.7
1.7
5.5
2.9
4.7
1.9
1.9
8.2
2.7
6.7
0.02
0.01
0.04
0.01
0.01
0.01
BQLb
0.07
12
21
24
25
26
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1.9
2.3
BQL
BQL
15.1
a
Dosed at 0.5 mg/kg in rat, using 10% DMA + 10% Solutol HS15 + 80% Saline (12, 21, 25,
26) or 10% DMA + 10% Solutol HS15 + 80% (30% (2ꢀhydroxypropyl)ꢀβꢀcyclodextrin) in
saline) (24) as vehicle. b BQL = Below quantifiable limit of 1.00 ng/mL.
Table 4. Beagle intravenous & oral PK profiles of compounds 21, 24, 25.a
iv
po
Dose
Cl
Vss
Dose
(mg/kg)
AUCinf
(ng/h/mL)
t1/2 (h)
F %
(mg/kg) (mL/min/kg) (L/kg)
1a
1a
1b
36
29
27
3.9
5.2
4.9
1a
5a
8c
0.2
6.3
6.5
6
1
21
24
25
423
14
80
3946
a
b
Dosed using 10% DMA + 5% Solutol HS15 + 85% (10% aqueous VEꢀTPGS) as vehicle.
Dosed using 10% DMA + 10% Solutol HS15 + 80% saline as vehicle. c Dosed using 10% DMA
+ 10% Solutol HS15 + 80% (10% aqueous (2ꢀhydroxypropyl)ꢀβꢀcyclodextrin).
Based on the studies detailed above and a consideration of the overall in vitro profile, compound
25 was selected for extensive further evaluation as summarized below and in Table 5. In
common with other members of the series, the high mGlu5 affinity of 25 was matched by high
potency in the functional assay. In comparison, mavoglurant (pKi 8.0, pIC50 8.0) and dipraglurant
(pKi 6.9, pIC50 6.9), key clinical molecules, had approximately 10ꢀ100 fold lower affinities and
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 47
1
2
3
4
5
6
7
8
9
potencies. In P450 inhibition assays 25 was a moderate inhibitor of 1A2 but clean against other
isoforms and in P450 stability assays 25 demonstrated a balanced profile with metabolic
contributions from all isoforms examined.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Moderate permeability in a Cacoꢀ2 monolayer system was observed without efflux, and in a rat
brain homogenate equilibrium dialysis assay 25 demonstrated approximately 1% fraction
unbound,48 findings in line with the observed rat PK profile. The compound was stable in rat
plasma, inactive at the hERG ionꢀchannel, and had a clean profile in in vitro assays of
cytotoxicity in HepG2 cells, and in GreenScreen cytotoxicity / genotoxicity assays in the
presence and absence of S9 fractions.51 Plasma protein binding in human and dog was high, but
lower in rat. 25 was crossꢀscreened in an external panel of 30 GPCR, kinase, ionꢀchannel and
transporter targets, where the compound displayed at least 1,000 fold selectivity against the panel
members (see Supporting Information). To understand the selectivity against other metabotropic
glutamate receptors, 25 was profiled in agonist, positive allosteric modulator (PAM), and
antagonist/NAM modes against all eight mGlu receptors in an external screen (see Supporting
Information). 25 displayed an excellent selectivity profile, with no appreciable receptor
activation in agonist or PAM modes against any of the mGlu receptors. Similarly, no appreciable
inhibition was observed for mGlu1,2,4,6ꢀ8 in antagonist/NAM mode, with activity at the mGlu5
receptor (97% inhibition at 10 µM) confirming our inꢀhouse characterization of the molecule. A
small degree of mGlu3 inhibition was noted (47% inhibition at 10 µM), though doseꢀresponse
experiments were not performed to clarify the degree of selectivity against this receptor. The
excellent selectivity profile across the mGlu family was not an unexpected finding, as a number
of other mGlu5 NAM chemotypes have been reported to have similar profiles.52ꢀ54 Selectivity of
ACS Paragon Plus Environment
22

Page 23 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mavoglurant against mGlu1, the other Group I metabotropic glutamate receptor and hence
historically the most challenging selectivity target for orthosteric mGlu5 ligands, has been
reported and rationalized by comparison of the Xꢀray structures of the two receptors.4,50
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 5. Profile of 25, HTL14242.
mGlu5 pKi 9.3, pIC50 9.2
Clean in HepG2 cytotoxicity assay, TC50 > 90
µMd
P450 pIC50:
Rat fu,brain = 1%e
3A4/2C19/2D6/2C8/2C9 < 4.8, 1A2 5.9
hERG pIC50 < 5.5
Negative in cytotoxicity and genotoxicity
screen,f with and without S9 fractions
Permeabilitya 13x10ꢀ6 cm sꢀ1 (efflux ratio 1.0) Exꢀvivo binding: ED50 0.3 mg/kg po
PPB%b >98, >98, 83
In vivo efficacy (mouse marble burying) ED50
8 mg/kg po
Rat plasma stability 96%c
7ꢀday nonꢀGLP repeat rat toxicity – well
tolerated up to 30 mg/kg/day po
a Cacoꢀ2 monolayer, Papp AꢀB. b Binding to human, canine and rat plasma proteins respectively. c
Percent compound remaining after 60 min incubation with rat plasma. d Concentration causing a
50% reduction in cell viability. e Equilibrium dialysis between buffer and rat brain homogenate. f
GreenScreen HC assay, see reference 51.
To evaluate receptor occupancy in rat CNS from an oral dose, 25 was advanced to an ex vivo
autoradiography experiment. Male SpragueꢀDawley rats were dosed orally with vehicle alone
(10% DMA, 10% Solutol HS15, 80% (10% aqueous (2ꢀhydroxypropyl)ꢀβꢀcyclodextrin)) or 25
(1, 3 or 10 mg/kg, po) and sacrificed 1 h postꢀdose. Sections of the hippocampal CA3 region
were prepared and incubated with [3H]ꢀMꢀMPEP,34 with levels of bound radioactivity in the
sections determined using a beta imager. Receptor occupancy was determined from mean
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 47
1
2
3
4
5
6
7
8
9
specific binding with the vehicle treated control taken as 100%. As depicted in Figure 2, 25
demonstrated excellent, doseꢀdependent, occupancy of mGlu5 receptors from an oral dose, with
an estimated ED50 of 0.3 mg/kg. By comparison mavoglurant returned an ED50 of approximately
3.6 mg/kg po in a comparable study.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. [3H]ꢀMꢀMPEP ex vivo autoradiography in rat hippocampal CA3 region 1 hour
following oral administration of 25.a
a Results are expressed as mean specific binding as a percentage of control taken as 100% (n=5).
For statistical analyses data were square root transformed and analysed by oneꢀway analysis of
variance followed by Williams’ test to compare 25 to vehicle. ***p<0.001. The images are
autoradiograms of coronal halfꢀbrain sections in rat hippocampal CA3 sections and are
representative samples of the 10 mg/kg treatment group.
The encouraging overall profile and high receptor occupancy of 25 prompted us to take the
compound forward into a preliminary assessment of rat repeat dose toxicology. Material for the
study was synthesized using the route depicted in Scheme 2 with Negishi coupling replacing
Stille coupling as the final step, negating the use of stoichiometric tin in the synthesis, and the
batch used was confirmed to have acceptable levels of Pd (< 0.5 mg/kg) by inductively coupled
plasmaꢀmass spectrometry (ICPꢀMS) analysis. A nonꢀGLP 7ꢀday repeat dose toxicity study in
SpragueꢀDawley rats was performed, administering 25 by once daily oral gavage at levels of 3,
ACS Paragon Plus Environment
24

Page 25 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10 and 30 mg/kg/day. 25 was well tolerated, no target organ effects were observed, no clinical
observations or differences in body weight, food consumption, haematology, coagulation,
clinical chemistry, urinalysis, organ weights, gross pathology or histopathological findings that
were considered to be related to treatment with test compound were observed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As an initial in vivo evaluation of 25 we undertook a murine study in a marble burying paradigm,
a model which has been used to probe the antiꢀanxiolytic effect of mGlu5 NAMs.55,56 In brief, 24
marbles were evenly spaced across the surface of sawdust bedding, and male CDꢀ1 mice treated
orally with vehicle (10% Solutol HS15, 90% (10% aqueous (2ꢀhydroxypropyl)ꢀβꢀcyclodextrin))
or 25 (1, 3, 10 or 30 mg/kg) thirty minutes before the start of the test. After preꢀtreatment, mice
were placed individually in a cage containing the marbles, and after thirty minutes mice were
removed and the number of marbles buried by at least two thirds into the sawdust were counted.
Compound 25 was able to strongly attenuate marble burying behavior in a dose dependent
manner, with an estimated ED50 of approximately 8 mg/kg, comparable to that of mavoglurant 1
in our hands in the same model.
In conclusion, stabilization of the mGlu5 receptor enabled a fragment screen, leading to the
identification of pyrimidine hit 5 which was then optimized using a SBDD approach to a highly
potent series of negative allosteric modulators. During the course of the project, multiple crystal
structures of mGlu5 NAMs were generated, including those of 14 and 25. As well as their utility
in the lead optimization process these structures also help to explain several of the inherent
challenges in drug discovery in this area, and pave the way for future SBDD campaigns in the
mGlu field. Compound 25, hereafter known as HTL14242, is a high affinity modulator of mGlu5
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 47
1
2
3
4
5
6
7
8
9
with excellent in vitro characteristics including high selectivity within the metabotropic family of
GPCRs. HTL14242 has good PK characteristics in rat and dog, high receptor occupancy as
elucidated through ex vivo autoradiography studies, demonstrates activity in vivo in a relevant
model, and has a clean profile in a repeat dose rat toxicology study. Based on the studies
described in this manuscript, HTL14242 has been selected as an advanced lead compound for
further development.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
26

Page 27 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
EXPERIMENTAL SECTION
Chemistry
7
8
9
Active compounds have been examined for known classes of assay interference compounds
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
and do not fall into this category.57 19F NMR data for 25, HTL14242 is referenced to C6H5CF3 at
1
ꢀ63.72 ppm. H NMR and LCMS QC data for all compounds, and experimental procedures for
synthesized compounds in Tables 1 and 2 are described in the Supporting Information. Synthesis
of 25, HTL14242 is described below.
3ꢀ(5ꢀAminoꢀ2ꢀchloropyridinꢀ4ꢀyl)ꢀ5ꢀchlorobenzonitrile. A mixture of 4,6ꢀdichloropyrimidine
(9.0 g, 60.4 mmol), 3ꢀchloroꢀ5ꢀ(4,4,5,5ꢀtetramethylꢀ1,3,2ꢀdioxaborolanꢀ2ꢀyl)benzonitrile (17.6 g,
66.8 mmol) and cesium carbonate (35.0 g, 107.4 mmol) were dissolved in 1,4ꢀdioxane / water
(9:1, 50 mL). After degassing with N2, Pd(dppf)Cl2 (2.20 g, 3.0 mmol) was added and the
reaction mixture was stirred at 90°C for 2 h. After cooling to rt the reaction mixture was
partitioned between H2O (250 mL) and EtOAc (150 mL), the phases were separated and the
aqueous layer extracted with EtOAc (2 x 150 mL). The combined organic layers were dried
(Na2SO4), concentrated in vacuo and purification by column chromatography (SiO2, 0 to 10%
ethyl acetate in hexane) yielded the title compound (4.0 g, 16.0 mmol) as offꢀwhite solid. m/z
250.1 (M+H)+; δH (400 MHz; d6ꢀDMSO) 8.27 (s, 1H), 8.53 (s, 1H), 8.59 (s, 1H), 8.66 (s, 1H),
9.16 (s, 1H).
3ꢀChloroꢀ5ꢀ[6ꢀ(5ꢀfluoropyridinꢀ2ꢀyl)pyrimidinꢀ4ꢀyl]benzonitrile (25, HTL14242).
A mixture of 2ꢀbromoꢀ5ꢀfluoropyridine (500 mg, 2.89 mmol) and hexamethylditin (946 mg, 2.89
mmol) in DME (10 mL) was degassed by purging with N2 for 5 min before the addition of
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 47
1
2
3
4
5
6
7
8
9
tetrakis(triphenylphosphine)palladium(0) (166 mg, 0.14 mmol). The reaction mixture was stirred
at 110 °C for 16 h before cooling to rt and partitioning between H2O (50 mL) and EtOAc (25
mL). The phases were separated and the aqueous phase was extracted with EtOAc (2 x 25 mL).
The combined organic phases were dried (Na2SO4) and concentrated in vacuo to yield crude 5ꢀ
fluoroꢀ2ꢀ(trimethylstannyl)pyridine (700 mg) which was used in the subsequent step without
characterization or further purification. 3ꢀChloroꢀ5ꢀ(6ꢀchloropyrimidinꢀ4ꢀyl)benzonitrile (100
mg, 0.39 mmol) and crude 5ꢀfluoroꢀ2ꢀ(trimethylstannyl)pyridine (114 mg) were dissolved in
toluene (15 mL) and the reaction mixture was degassed by purging with N2 for 5 min before the
addition of tetrakis(triphenylphosphine)palladium(0) (46.2 mg, 0.04 mmol) and copper(I) iodide
(7.6 mg, 0.03 mmol). The reaction mixture was stirred at 110 °C for 16 h before cooling to rt and
partitioning between H2O (50 mL) and EtOAc (25 mL). The phases were separated and the
aqueous phase was extracted with EtOAc (2 x 25 mL). The combined organic phases were dried
(Na2SO4) and concentrated in vacuo. Purification by gradient flash chromatography, eluting with
0ꢀ10% EtOAc in hexane yielded the title compound (37 mg, 0.12 mmol) as a pale yellow solid.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
m/z 311.0, 313.0 (M+H)+; H NMR (400 MHz, DMSO) δ: 8.02 (td, J = 8.7, 2.7 Hz, 1H), 8.29
(dd, J = 2.0, 1.4 Hz, 1H), 8.57ꢀ8.67 (m, 2H), 8.76 (t, J = 1.5 Hz, 1H), 8.84 (d, J = 2.7 Hz, 1H),
8.95 (d, J = 1.5 Hz, 1H), 9.44 (d, J = 1.2 Hz, 1H); 19F NMR (376 MHz, DMSO) δ: ꢀ125.85 (dd, J
13
3
= 8.6, 4.6 Hz, 1F); C NMR (101 MHz,) δ: 113.1, 114.1, 117.1, 123.6 (d, JCF = 5.4 Hz, 1C),
2
2
124.6 (d, JCF = 19.2 Hz, 1C), 129.7, 131.5, 134.0, 134.9, 138.2 (d, JCF = 24.5 Hz, 1C), 139.0,
149.4 (d, 4JCF = 3.8 Hz, 1C), 158.9, 160.5 (d, 1JCF = 258.4 Hz, 1C), 160.8, 162.5.
ACS Paragon Plus Environment
28

Page 29 of 47
Journal of Medicinal Chemistry
1
2
3
4
Expression of mGlu5 and Membrane Preparation
5
6
7
8
9
cDNA encoding the human mGlu5 receptor was transfected into HEK293 cells using the
transfection reagent Genejuice (Novagen). Fortyꢀeight hours after transfection, cells were
harvested and washed twice with ice cold phosphateꢀbuffered saline. Membranes were prepared
and stored as previously described.58
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
[3H]-M-MPEP binding assay
Membranes (2.5 µg) prepared from HEK293 cells stably expressing the mGlu5 receptor were
incubated with a KD concentration (~0.75 nM) of [3H]ꢀMꢀMPEP34 in binding buffer (50 mM
HEPES; 150 mM NaCl; pH 7.4). Nonspecific binding was determined by inclusion of 0.1 mM of
MPEP. After 90 min incubation at room temperature (25 °C) assays were terminated by rapid
filtration through 96ꢀwell GF/B filter plates preꢀsoaked with 0.1% polyethyleneimine (PEI) using
a 96ꢀwell head harvester (Tomtec, USA) and plates washed 5 × 0.5 mL with H2O. Scintillant (50
µL Ultima Gold F; PerkinElmer, UK) was added and bound radioactivity was determined using
scintillation spectroscopy on a Microbeta counter (PerkinElmer, UK). Inhibition binding curves
were fitted to a fourꢀparameter logistic equation to determine IC50 values. IC50 values were
converted to Ki values with the ChengꢀPrusoff equation59 using a KD value (0.75 nM) derived
from saturation binding studies.
Functional assay
To measure receptor activation the IPone assay kit from Cisbio (France) was used. The assay
was optimised to measure the ability (potency; pIC50) of antagonists / negative allosteric
ACS Paragon Plus Environment
29