Bicyclo((aryl)methyl)benzamides as inhibitors of GlyT1

Article abstract of DOI:10.1016/j.bmcl.2018.02.029

A series of isoquinuclidine benzamides as glycine uptake inhibitors for the treatment of schizophrenia are described. Potency, lipophilicity, and intrinsic human microsomal clearance were parameters for optimization. Potency correlated with the nature of

Full text of DOI:10.1016/j.bmcl.2018.02.029

Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Bicyclo((aryl)methyl)benzamides as inhibitors of GlyT1
⇑
Jeffrey G. Varnes , Hui Xiong, Janet M. Forst, Christopher R. Holmquist, Glen E. Ernst, William Frietze,
Bruce Dembofsky, Don W. Andisik, William E. Palmer, Lindsay Hinkley, Gary B. Steelman, Deidre E. Wilkins,
Gaochao Tian, Gerald Jonak, William M. Potts, Xia Wang, Todd A. Brugel, Cristóbal Alhambra,
⇑
Michael W. Wood, Chris A. Veale, Jeffrey S. Albert
Medicinal Chemistry, Neuroscience, IMED Biotech Unit, AstraZeneca, Wilmington, DE, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of isoquinuclidine benzamides as glycine uptake inhibitors for the treatment of schizophrenia are
described. Potency, lipophilicity, and intrinsic human microsomal clearance were parameters for opti-
mization. Potency correlated with the nature of the ortho substituents of the benzamide ring, and reduc-
tions in lipophilicity could be achieved through heteroatom incorporation in the benzamide and pendant
phenyl moieties. Improvements in human CLint were achieved through changes in ring size and the N-
alkyl group of the isoquinuclidine itself, with des-alkyl derivatives (40–41, 44) demonstrating the most
robust microsomal stability. Dimethylbenzamide 9 was tested in a mouse MK801 LMA assay and had
Received 20 November 2017
Revised 12 February 2018
Accepted 13 February 2018
Available online xxxx
Keywords:
GlyT1
Schizophrenia
NMDAr
a statistically significant attenuation of locomotor activity at 3 and 10 lmol/kg compared to control.
Ó 2018 Elsevier Ltd. All rights reserved.
N-Methyl-D-aspartate receptor
Glycine uptake inhibition
The N-methyl-D-aspartate receptor (NMDAr) hypofunction
hypothesis of schizophrenia correlates disease symptomology with
glutamatergic neurotransmission dysfunction.1–4 Therefore, inhibi-
tion of glycine transporter 1 (GlyT1) to elevate synaptic levels of
glycine and potentiate NMDAr signaling is a promising therapy
for this disease. Current treatments for schizophrenia are associ-
ated with unwanted side effects and do not provide sufficient relief
across all symptom domains.2,5–7 Additionally, FDA-approved
GlyT1 inhibitors for schizophrenia remain elusive, with Hoffman-
La Roche clinical candidate bitopertin (RG1678) being one of the
most advanced to date. Having reached Phase III, the compound
was ultimately withdrawn from development after failing to
achieve statistical significance in the improvement of negative
symptoms compared to control.1 As such, treatments for the posi-
tive, negative, and cognitive symptoms of schizophrenia remain a
significant unmet medical need.
*
1
2a, R = H; SSR504734
2b, R = Me
3, * = S
4, * = R
racemic
We previously reported a series of azepane sulfonamides as
GlyT1 inhibitors.8 Identified through a virtual screening campaign
and exemplified by 1, these compounds were characterized by
notable glycine uptake efficiency9 (GlyT1 IC50 for 1: 37 6 nM),
low aqueous solubility,10 and favorable brain-to-plasma ratios
(1.6–2.1). As previously noted,8 1 is structurally similar to Sanfoi’s
withdrawn clinical candidate SSR5047341,11 (2a) and the more
potent related N-Me analog 2b.12 We speculated that by capitaliz-
ing on our learnings from development of the former we might
increase the potency of 2b through an increase in piperidine
lipophilicity. However, rather than evoke a ring expansion to form
an azepane, we proposed that conversion to an isoquinuclidine,
such as in 3,13 might also provide the desired outcome. While
⇑
Corresponding authors.
E-mail addresses: jeffrey.varnes@astrazeneca.com (J.G. Varnes), jeffrey.albert@
intellisynrd.com (J.S. Albert).
https://doi.org/10.1016/j.bmcl.2018.02.029
0960-894X/Ó 2018 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Varnes J.G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.029

2
J.G. Varnes et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
quinuclidine isomers of 2b similar to 3 have been reported,14 3 also
had the potential advantage of having a single chiral center. We
were optimistic that this might increase synthetic tractability
and facilitate exploration of our hypothesis.
While the potency and solubility (typically high for the compounds
described herein) of 4 were acceptable, both lipophilicity, as mea-
sured by clogP, and stability on incubation with human liver
microsomes were higher than desired. Reasoning that reductions
to the former might also attenuate the latter, we sought to reduce
clogP (<4) in a variety of ways. This included a) heteroaromatic and
substituent changes to the benzamide aromatic ring, b) conversion
of the unsubstituted phenyl moiety into various heterocycles, and
c) ring contraction of the isoquinuclidine itself.
As shown in Table 1, a comparison of 2b and 3 in a glycine
uptake inhibition assay9 revealed slightly more than a ten-fold
decrease in potency. While this did not confirm our hypothesis,
we were gratified that such a change still afforded a compound
with appreciable in vitro activity. Even more interesting, enan-
tiomer 4 did demonstrate activity comparable to 2b. This was sur-
prising, because the chirality of the benzylic position of 4 (R
enantiomer) is opposite that of the same position in 2b (S,S). Sus-
pecting an error in stereochemical assignment, we validated the
chirality of both 3 and 4 by vibrational circular dichroism (VCD)
and X-ray crystallography (discussed below). A molecular overlay
(recreated in MOE)15 reinforced the concept that similar potencies
between 2b and 4 were not unreasonable as there is good overlap
of benzamide, aryl, and heterocyclic regions (Fig. 1).
In the course of elucidating benzamide SAR to attenuate
lipophilicity, simple heterocyclic replacements (Not shown: 1-
methyl-imidazol-4-yl, thiazol-4-yl; GlyT1 uptake IC50 > 20 lM) of
the aryl ring were insufficient to retain potency. For benzamides
themselves, potency was found to directly depend on the nature
and number of ortho substituents. Specifically, linear free energy
relationship analysis (LFER) identified both the ortho substituent
parameter (Es; inverse correlation)17 and hydrophobic nature (
p
Hammett constant; positive correlation) of the benzamide ortho-
substituents to be good predictors of glycine uptake inhibition.
Where such information was not available, a torsion angle analysis
could be used to roughly bin compounds according to potency
(Fig. 2). In other words, hydrophobic ortho-substituents that per-
turbed the planarity of amide-aryl conjugation generally were
more active at inhibiting glycine uptake.
We were interested in modifying 4 to be a probe suitable for
evaluation in a murine MK801 locomotor activity (LMA) assay.16
Table 1
Initial assessment of isoquinuclidine enantiomers.
Armed with this knowledge, we looked at ways to reduce the
clogP of 4 through nitrogen incorporation while still maintaining
potency via installation of ortho substituents such as chloro,
methyl, bromo, and thiomethyl (Table 2). In general, pyridine ana-
logs were less active (6 vs 7; 9 vs 10; 14 vs 15) than their simple
phenyl counterparts, with 3-pyridyl amides evoking the most dra-
matic potency reductions (10 & 13). The most active compounds,
however, were those containing chloro, methyl, or thiomethyl
ortho substituents (6–9), and, in particular, 4-pyridyl analog 7 rein-
forced the concept that clogP could be reduced in select cases with
only a slight decrease in uptake efficiency.
2b (S,S)
3 (S)
4 (R)
Uptake IC50 (nM)a
Sol (mM)b
clogP
3
41
6
124
5.1
2.9
34
>500
5.4
2.7
34
>482
5.4
2.7
34
logDc
PSA (Å2)
hu CLint
(l
L/min/mg)d
223
103
109
Glycine uptake inhibition assay.9
Equilibrium solubility (pH 7.4).10
a
b
c
Empirically determined partitioning ratio: octanol/water (pH 7.4).
Rate of clearance as measured in human liver microsomes.
d
Despite reductions in lipophilicity, we saw no improvement in
predicted human clearance (6 vs 7; 14 vs 15) for pyridine matched
pairs, and all compounds contained a low to moderate efflux liabil-
ity. Improvements in metabolic stability were only observed in a
very small number of benzamide modifications that involved
introduction of polar ortho substituents such as for 16 and 17
(hu CLint <4 lL/min/mg). While very potent, both compounds car-
ried additional liabilities such as concerns around potentially
mutagenic anilines (16) or excessively high efflux (17).
Fig. 2. Benzamide aryl-carbonyl (red arrow) torsion angle analysis color coded
according to GlyT1 uptake IC50 potency. Green: <100 nM; yellow: 100 nM to 1
lM;
red: >1
l
M. Aryl rings with hydrophobic ortho-substituents (R1 & R2) that hindered
free rotation were generally predisposed to be more potent at inhibiting glycine
uptake, with the preferred conformation trending towards one in which the
aromatic ring is perpendicular to the plane of the amide itself.
Fig. 1. Overlay of 2b (blue) with 4 (magenta) using MOE (Chemical Computing
Group) software.15
Please cite this article in press as: Varnes J.G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.029

J.G. Varnes et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
3
Table 2
Reduction of lipophilicity through benzamide modifications.
R
#
R
Uptake IC50 (nM)a
6
Sol (mM)b
clogP
logDc
2.7
hu CLint
109
(l
L/min/mg)d
MDR1 efflux ratioe
>1.9
4
>482
5.4
5
92
>500
4.4
2.4
197
2.2
6
7
1
6
493
498
5.2
4.3
2.2
2.2
57
95
1.5
1.5
8
3
>500
414
4.0
4.8
4.0
3.8
1.3
1.9
0.8
1.0
76
121
–
6.3
6.2
–
9
1
10
11
494
474
373
>500
45
–
12
75
>500
3.7
–
50
2.9
13
14
15
809
23
>500
>500
481
3.6
4.5
3.6
1.1
1.9
1.4
–
–
91
97
3.1
6.3
45
16
17
8
6
462
495
4.2
4.4
1.3
0.7
<4
<4
3.4
>29
a
b
c
Glycine uptake inhibition assay.9
Equilibrium solubility (pH 7.4).10
Empirical logD measured as a ratio of compound partitioned between octanol and water (pH 7.4).
Rate of clearance as measured in human liver microsomes.
d
e
Transporter efflux assay using canine kidney cells expressing P-glycoprotein.
A similar exercise aimed at reducing lipophilicity and/or explor-
was the first example of a phenyl replacement that appeared to
retain a significant degree of potency while lowering clogP; disap-
pointingly this had an adverse impact on metabolic stability. Gen-
erally, efflux ratios were not investigated for the majority of these
samples as these were racemates.
It is worth noting that 4-pyridine analog 27 is the only chiral
compound in Table 3 and so must be compared to chiral 4 rather
than 4b. There was roughly a threefold decrease in potency upon
ing areas for scaffold elaboration was conducted through modifica-
tion of the pendant phenyl ring using racemic material for ease of
accessibility. Unfortunately, glycine uptake inhibition decreased in
nearly all cases compared to racemic 4b (Table 3). Only data for
meta substitution (20 & 21 vs 4b) suggested that there was room
for a potency increase, and this was accompanied by a commensu-
rate increase in lipophilicity. Furan 26 was also promising as this
Please cite this article in press as: Varnes J.G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.029

4
J.G. Varnes et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Table 3
Profiling phenyl modifications for racemic analogs of 4.
#
R
Uptake IC50 (nM)a
clogP
logDb
hu CLint (l
L/min/mg)c
4b
18
19
20
21
22
23
24
25
26
27
28
29
Ph
39
5.4
5.2
5.9
6.3
5.9
5.2
5.9
5.2
5.5
4.6
3.9
3.9
3.9
2.7
–
3.0
3.3
3.0
–
–
2.7
–
2.0
2.4
2.3
1.6
94
81
2-OMe-Ph
2-Me-Ph
3-Br-Ph
3-Me-Ph
3-OMe-Ph
4-Me-Ph
4-OMe-Ph
n-Butyl
Furan-2-yl
Pyridin-4-yld
Pyridin-3-yl
Pyridin-2-yl
340
514
9
247
58
119
102
69
73
–
128
119
56
29
177
1703
1759
3362
78
18
211
904
55
a
b
c
Glycine uptake inhibition assay.9
Empirical logD measured as a ratio of compound partitioned between octanol and water (pH 7.4).
Rate of clearance as measured in human liver microsomes.
d
Chiral, single enantiomer, absolute configuration not determined.
incorporation of a nitrogen at the 4-position of the phenyl ring, no
significant change in metabolic liability despite a 1.5 log unit drop
in clogP, and efflux liability persisted (efflux ratio: 4.7).
few sulfonamides) such as 32, all suffered from high predicted
human clearance and efflux liabilities. Lastly, pyrazole 33, despite
the promise suggested by bromide 20, failed to realize any gain
through expansion off the meta position of the phenyl ring.
What remained elusive within the isoquinuclidines were struc-
tural modifications leading to a reduction in both clearance and
efflux. We suspected that N-dealkylation was at least partly to
blame, and, through a combination of changes in the nitrogen sub-
stituent and isoquinuclidine itself, we were able to demonstrate
this conclusively (Table 5).
We opted to prepare a small number of additional analogs to
finish our exploration of phenyl ring SAR using the more potent
scaffold of 2,6-dimethylbenzamide 9 (Table 4). Encouraged by
the data for 26, 2-methylfuran 30 was prepared in an attempt to
mitigate any potential metabolism associated with the 2-position
of the furan ring itself. While potency was retained and lipholicity
and human CLint decreased, efflux was unchanged. Similar potency
and efflux results were obtained with 4-pyridine 31. The last two
compounds of Table 4 involved scaffold modifications similar to
those described by Sanfoi11b and Taisho Pharmaceuticals18 respec-
tively. While we synthesized a number of potent sulfones (and a
Thus, for des-methyl analog 34, in vitro microsomal clearance
decreased by fourfold (34 vs 121 lL/min/mg) compared to 9.
Removal of a methylene spacer to afford a series of azabicyclo
[2.2.1]heptanes19 (ABHs; 35–41) reduced clogP by roughly half a
Table 4
Select phenyl ring modifications of benzamide 9.
*
B
A
C
D
E
#
R
Uptake IC50 (nM)a
clogP
logDb
hu CLint
(l
L/min/mg)c
MDR1 efflux ratiod
9
A
B
C
D
E
1
1
1
1
4.8
4.4
3.3
4.1
4.8
1.9
1.5
1.4
2.0
2.1
121
76
–
167
–
6.2
6.0
6.3
14
–
30e
31e
32e
33e
946
a
b
c
Glycine uptake inhibition assay.9
Empirical logD measured as a ratio of compound partitioned between octanol and water (pH 7.4).
Rate of clearance as measured in human liver microsomes.
Transporter efflux assay using canine kidney cells expressing P-glycoprotein.
Single enantiomer; absolute chirality not determined.
d
e
Please cite this article in press as: Varnes J.G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.029

J.G. Varnes et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
5
Table 5
Exploration of isoquinuclidine modifications of benzamide 9.
A
B
C
*
Y
Y
#
R
X
Y
Uptake IC50 (nM)a
clogP
logDb
hu CLint
(
l
L/min/mg)c
MDR1 efflux ratiod
9
A
A
A
B
B
B
B
B
B
B
C
C
C
CH
N
Me
Me
H
Me
Et
iPr
iPr
cPr
H
1
1
3
3
4
1
10
2
58
45
7
4.8
3.3
4.2
4.2
4.7
5.0
3.5
4.6
3.6
2.1
3.6
4.2
3.1
1.9
1.4
1.3
1.9
2.1
1.9
–
1.1
1.4
0.9
1.1
1.1
0.8
121
–
6.2
6.3
11.8
2.9
6.2
4.3
–
1.1
4.7
20
31e
34e
35e
36e
37e
38e
39e
40e
41e
42e
43e
44e
CH
CH
CH
CH
N
CH
CH
N
34
96
145
39
43
142
22
15
43
40
<4
H
Me
Et
CH
CH
CH
7.4
8.3
6.6
3
168
H
a
b
c
Glycine uptake inhibition assay.9
Empirical logD measured as a ratio of compound partitioned between octanol and water (pH 7.4).
Rate of clearance as measured in human liver microsomes.
Transporter efflux assay using canine kidney cells expressing P-glycoprotein.
Single enantiomer; absolute chirality not determined.
d
e
log unit and afforded comparable potency (9 vs 35). As before,
removal of the N-alkyl substituent within the ABHs (40) reduced
predicted human clearances by roughly fivefold. Similarly, decreas-
ing ring size further to establish a series of azabicyclo[2.1.1]hex-
anes20 (ABXs; 42–44) was accompanied by another drop in
lipophilicity and potency relative to parent isoquinuclidines, how-
ever, this series also had overall lower liabilities from a microsomal
stability perspective. Consistent with isoquinuclidine and ABH
compounds, des-methyl ABX 44 had the most favorable result in
this regard but also the largest drop-off in glycine uptake
inhibition.
We also explored various nitrogen substituents with an aim
toward modulating sterics and basicity within the ABH series. No
glycine uptake inhibition was detected with non-basic bicycles
(data not shown). Isopropyl 37 was the best compromise between
desired on-target effect and clearance, however, efflux was still
problematic. We did see a favorable shift in the latter with cyclo-
propyl 39, but microsomal stability was dramatically worse. Lastly,
as with pyridine 31, reduction in clogP through nitrogen incorpora-
tion was unsuccessful. This change adversely impacted potency
(38) and efflux (41) respectively.
Intravenous and oral pharmacokinetic profiles of dimethylben-
zamide 9 were determined in Sprague Dawley rats (Table 6). Com-
pared to in vitro results, a lower than expected in vivo clearance
was observed, which may be indicative of an underlying saturation
mechanism. Consequently, bioavailability was also high while vol-
ume of distribution was low. Consistent with MDR1 efflux results,
the brain-to-plasma ratio was low in rat but, for reasons not well
understood, higher in mouse (brain/plasma: 0.4; CD-1 mice dosed
s.c.). This may be at least partially reflective of the lower plasma-
protein binding observed in the latter species (mu PPB: 32% free).
Additionally, when profiling 9 for off-target pharmacology, only
five out of 100 targets demonstrated moderate activity in an MDS
Pharma panel (% inhib @ 10
lM): human:
a2B (54), RXR-
a
(À5 4) &
5HT2a (62); rat: Ca2+ channel L-type (65) & Na+ channel site-2 (70).
Minimal inhibition (<20%) was observed for related targets such as
adenosine, DAT, NET, and SERT transporters and the NMDA recep-
tor. Activity against glycine transporter-2 (GlyT2) was also low
(GlyT2 IC50 3.8 lM).
Compound 9 was tested in a mouse MK801 locomotor activity
assay to model the treatment of positive and negative symptoms
of schizophrenia.16 Compound 9 showed dose-dependent reversal
of MK801-induced locomotor activity and a statistically significant
Table 6
IV and PO DMPK parameters for 9 in Sprague Dawley rats.a
Route
Parameter
9
dose response at
potency at 3 mol/kg for 9 was comparable to 2b at a higher dose
(10 mol/kg) (Fig. 3).
3 and 10 lmol/kg (p < 0.01). Furthermore,
in vitro
CLint
(l
L/min/mg)b
182
14
l
PPB (% free)c
l
Compounds described herein were prepared using the general
route described in Scheme 1, which involved addition of an aryl
or heteroaryl Grignard or lithium alternative to a chiral or racemic
sulfinamide21 followed by deprotection and acylation. Isoquinu-
clidine/done, ABH, or ABX core aldehydes were prepared from their
respective acid or ester precursors, and these syntheses, along with
those of the fully elaborated final products have been previously
reported.22–24
IV
Clearance (mL/min/kg)
VDss (L/kg)
32
1.8
3.1
0.2
t1/2 (h)
Brain/plasmad
PO
Cmax (nM)
Tmax (h)
F (%)
2452
1.0
105
a
b
c
Dose: IV = 3 lmol/kg; PO = 9 lmol/kg.
Rate of clearance as measured in rat liver microsomes.
Determined by equilibrium dialysis.
Brain-to-plasma ratio; brain concentrations not corrected for 2–3% vascular
An example synthesis of isoquinuclidine 9 is shown in Scheme 2.
Amide 4522a was alkylated with iodomethane and then reduced at
low temperature using lithium aluminum hydride25 to afford alde-
d
contamination.
Please cite this article in press as: Varnes J.G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.029

6
J.G. Varnes et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
hyde 46 directly. Lewis acid-mediated sulfinamide formation to
give 47 was followed by trimethylaluminum-assisted26 addition
of phenyllithium (98% d.e.), amide reduction using catalytic RhH
(CO)(PPh3)4,22a and deprotection using methanolic HCl. Acid
coupling then provided 9.22b Final products derived from racemic
analogs of 47 (isoquinuclidines, ABHs, or ABXs) could be resolved
under chiral chromatography conditions using supercritical fluid
chromatography (SFC) or reverse phase chiral chromatography as
a final step.27
For chiral isoquinuclidine final compounds, stereochemistry
was assigned based on a comparison of intermediate calculated
VCD28 spectra (R, red/upper; S, green/lower) with that derived
experimentally from the amine (prior to amide coupling) which
afforded more active 4 (blue).29
In summary, we have described modifications to a series of iso-
quinuclidine benzamides as part of an effort to develop a novel
in vivo probe for inhibiting glycine uptake in an MK801 LMA assay.
In addition to potency, lipophilicity and human microsomal clear-
ance were parameters for optimization. Heteroatom incorporation
into the benzamide portion of 4 led to equipotent derivatives (7, 8)
with reduced clogP. Conversion of the pendant phenyl ring into a
2-methylfuran or pyridine (30–31) was less well tolerated but still
superior from a potency perspective to all but changes to the meta
position. Modest improvements in predicted human CLint could be
had through changes in ring size and the N-alkyl group of the iso-
quinuclidine itself, with ABH and ABX bicycles lacking an alkylated
nitrogen (40–41, 44) demonstrating the largest improvement in
microsomal stability. Simultaneously achieving a low efflux ratio
remained elusive. Dimethylbenzamide 9, after evaluation for rat
pharmacokinetics, was tested in a mouse MK801 LMA assay and
had a statistically significant attenuation of locomotor activity at
Fig. 3. Dose-response of 9 in a mouse MK801 LMA assay. MK801 and 2b were dosed
at 1.4 and 10 lmol/kg, respectively. Male CD-1 mice were dosed subcutaneously
using a formulation of 20% sulfobutylether-b-cyclodextrin (SBECD).
1) Ar/HetMgBr or
ArLi/HetLi
2) HCl/MeOH
22-92%
Ti(OEt)4
THF, r.t.
5-100%
15-94%
R1 =
Y
3 and 10 lmol/kg compared to control, with favorable comparison
Y
to 2b. The results of further optimization, an investigation of clear-
ance mechanisms, and conditioned avoidance response and novel
object recognition data will be reported in due course.
Scheme 1. General synthetic route & yields to bicyclo((aryl)methyl)benzamides. Y
is either alkyl, tert-butylcarbonate (Boc), or allyl. Allyl groups enabled late stage
diversification of nitrogen alkyl substituents following Pd-mediated
deprotection.22b
Acknowledgements
The authors are grateful to: Jennifer Van Anda, David Coomber,
and Xiaomei Ye for purification & chiral resolution of many of the
final compounds; Michelle Lamb for her assistance with recreating
the molecular overlay in MOE; and Steve Wesolowski and Don
Pivonka for VCD analyses.
d-g
c
a-b
45
46
47
9
References
Scheme 2. a) NaH, MeI, DMF, 81%; b) LAH, THF, À78 °C then 1 N aq HCl, 85%; c) (R)-
tBuSONH2, Ti(OEt)4, THF, 68%; d) PhLi in Et2O, AlMe3, THF, À78 °C; e) RhH(CO)
(PPh3)4, Ph2SiH2, THF, 68% over 2 steps; f) 4 M HCl/diox, MeOH, quant. g) TBTU,
HOBt monohydrate, DIPEA, DMF, 90%.
1. For leading references on GlyT1 inhibitors and clinical candidates, see: Singer P,
Dubroqua S, Yee BK. Curr Pharm Des. 2015;21:3771–3787.
2. Maksymetz J, Moran SP, Conn JP. Mol Brain. 2017;10(15):1–19.
Fig. 4. Overlay of the experimental (blue) VCD spectrum of the benzyl amine precursor of 4 with those predicted for R (red, upper) and S (green, lower) amines.
Please cite this article in press as: Varnes J.G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.029

J.G. Varnes et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
7
3. Kehrer C, Maziashvili N, Dugladze T, Gloveli T. Front Mol Neurosci. 2008;1:1–6.
4. Thomson JL, Blackaby WP, Jennings ASR, et al. Bioorg Med Chem Lett.
2009;19:2235–2239.
5. Dohi T, Morita K, Kitayama T, Motoyama N, Morioka N. Pharmacol Ther.
2009;123:54–79.
20. Racemic ABXs have also been reported by Sanofi: See: Dargazanli G, Estenne-
Bouhtou G, Mafroud A-K. Preparation of N-[(2-azabicyclo[2.1.1]hex-1-yl)(aryl)
methyl]benzamide derivatives as inhibitors of glycine transporters glyt1. From
PCT Int. Appl. (2010), WO 2010092286 A1 20100819.
21. Robak MT, Herbage MA, Ellman JA. Chem Rev. 2010;110:3600–3740.
22. (a) The isoquinuclidine methyl ester was prepared using the following
reference with substitution of iodomethane for benzyl bromide: Casabona D,
Cativiela C. Tetrahedron. 2006;62(42):10000–10004b) For full experimental
procedures, see Refs. 13,19.
23. (a) For an overview of the evolution of the synthesis of ABH compounds such as
35, see: Xiong H, Frietze W, Andisik DW, et al. Tetrahedron Lett.
2010;51:6741–6744b) For full experimental procedures, see Ref. 19.
24. (a) For a review of existing syntheses of the ABX acid, see: Varnes JG, Lehr GS,
Moore GL, Hulsizer JM, Albert JS. Tetrahedron Lett. 2010;51:3756–3758b) For
preparation of racemic versions of 42–44, see Ref. 20 and procedures described
in Refs. 22,23 for resolution.
6. Sur C, Kinney GG. Curr Drug Targets. 2007;8:643–649.
7. Lowe III JA, Hou X, Schmidt C, et al. Bioorg Med Chem Lett. 2009;19:2974–2976.
8. Varnes JG, Forst JM, Hoerter TN, et al. Bioorg Med Chem Lett.
2010;20:4878–4881.
9. GlyT1 uptake inhibition assay: The GlyT1 SPA assay measures [3H]glycine
uptake by CHO cells expressing human recombinant Glycine Transporter 1
(GlyT1). Assay: To the wells of an OptiPlate, 2
compound were spotted. This was followed by addition of 98
suspension (ꢀ1 million/mL final). After incubating cells with compound for
ꢀ15 min, 100 L of the SPA (200 g/well final) and isotope mixture (30 nM
isotope with 10 M cold glycine, final) was added to initiate the glycine uptake.
l
L of DMSO containing a test
l
L
of cell
l
l
l
At 2 h, the plate was read on a TopCount to quantify SPA counts. Results: IC50’s
are reported as the geomean, are the average of two, four, or six experiments,
and the limit of quantitation is approximately 1 nM. See Ref. 8 for additional
context.
25. For literature precedent of a similar reduction of a tertiary ester to an aldehyde
using LAH see: Jackson AC, Goldman BE, Snider BB.
1984;49:3988–3994.
J Org Chem.
26. Borg G, Chino M, Ellman JA. Tetrahedron Lett. 2001;42:1433.
10. Alelyunas YW, Liu R, Pelosi-Kilby L, Shen C. Eur J Pharm Sci. 2009;37:172–182.
27. For compounds obtained from chiral resolution (Tables 3–5), data for the more
potent enantiomer has been reported.
11. (a)
Depoortere
R,
Dargazanli
G,
Estenne-Bouhtou
G,
et
al.
Neuropsychopharmacology. 2005;30:1963–1985(b) Blackaby WP, Huscroft IT,
Keown LE. et al., ‘‘Preparation of benzamide derivatives as GlyT1 inhibitors”
PCT Int. Appl. (2006), WO 2006134341 A1 20061221.
28. Nafie LA, Keiderling TA, Stephens PJ. J Am Chem Soc. 1976;98:2715.
29. (a) Monte Carlo molecular mechanics searches of low energy conformers for R
and S benzyl amines depicted in Figure 4 were conducted using MacroModel
within the Maestro graphical interface (Schrödinger Inc.). The lowest energy
conformers identified were used as starting points and minimized using
density functional theory (DFT) within Gaussian03. Optimized structures,
harmonic vibrational frequencies/intensities, VCD rotational strengths, and
free energies at STP (including zero-point energies) were determined for each
conformer. In these calculations, the B3LYP generalized gradient
approximation (GGA) exchange-correlation density functional was used. The
cc-pVTZ basis set was used for the computations. Simulations of infrared and
VCD spectra for each conformer were generated using an in-house written
program to fit Lorentzian line shapes (16 cmÀ1 line width) to the computed
spectra to allow direct comparisons between simulated and experimental
spectra. Stereochemistry identified by VCD was confirmed by X-ray
crystallography (data not shown); b) All analyses were conducted using the
BioTools ChiralIR instrument. Experimental data were determined as follows:
20 mg of compound were dissolved in 0.25 mL of CDCl3. Analyses were
conducted at 4 cmÀ1 resolution using the dual source, VCD scan protocol. The
instrument incorporated a dual photo-elastic modulator set for polarization
modulation at 37.024 kHz with k/4 retardation (optimized for acquisition of
the spectral region centered around 1400 cmÀ1). Lock-in amplification with a
12. For potency comparisons, elaboration of 2b into a candidate SPECT ligand and
leading references, see: Fuchigami T, Haratake M, Magata Y, Haradahira T,
Nakayama M. Bioorg Med Chem. 2011;19:6245–6253.
13. Albert JS, Alhambra C, Brugel TA, Varnes JG. ‘‘Preparation of
azabicyclooctylbenzyl benzamides as modulators of glycine transporter
receptors”. PCT Int. Appl. (2009), WO 2009013535 A1 20090129.
I
14. Dargazanli G, Estenne Bouhtou G, Magat P, Marabout B, Roger P. Derivatives of
N-heterocyclomethylbenzamide, in particular N-(di/azabicycloalkyl)
methylbenzamides, as specific inhibitors of glycine transporters glyt1 and/or
glyt2. Fr. Demande (2005), FR 2861076 A1 20050422.
15. Modelled in MOE (Molecular Operating Environment, Chemical Computing
Group) using a flexible alignment with default settings and AMBER10-EHT
forcefield.
16. Rung JP, Carlssona A, Markinhuhtab KR, Carlsson ML. Prog
Neuropsychopharmacol Biol Psychiatry. 2005;29:827–832.
17. Data for LFER analyses conducted on isoquinuclidines is beyond the scope of
this publication, however, see the following reference for a discussion of Es:
Sotomatsu T, Fujita T. J Org Chem. 1989;54(18):4443–4448.
18. Sekiguchi Y, Shibata T, Abe K, Yamamoto S, Kashiwa S, Okubo T. Kokai Tokkyo
Koho (2009), JP 2009185010 A 20090820.
30
ls time constant and a 20 kHz high pass and a 4 kHz low pass filter was
19. Albert JS, Alhambra C, Brugel TA. et al. 2-Azabicyclo[2.2.1]heptane compounds
used.
as glycine transport
1 modulators and their preparation and use in the
treatment of diseases. PCT Int. Appl. (2010), WO 2010087762 A1 20100805.
Please cite this article in press as: Varnes J.G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.029