Discovery of MK-8719, a Potent O-GlcNAcase Inhibitor as a Potential Treatment for Tauopathies

Article abstract of DOI:10.1021/acs.jmedchem.9b01090

Inhibition of O-GlcNAcase (OGA) has emerged as a promising therapeutic approach to treat tau pathology in neurodegenerative diseases such as Alzheimer's disease and progressive supranuclear palsy. Beginning with carbohydrate-based lead molecules, we pursued an optimization strategy of reducing polar surface area to align the desired drug-like properties of potency, selectivity, high central nervous system (CNS) exposure, metabolic stability, favorable pharmacokinetics, and robust in vivo pharmacodynamic response. Herein, we describe the medicinal chemistry and pharmacological studies that led to the identification of (3aR,5S,6S,7R,7aR)-5-(difluoromethyl)-2-(ethylamino)-3a,6,7,7a-tetrahydro-5H-pyrano[3,2-d]thiazole-6,7-diol 42 (MK-8719), a highly potent and selective OGA inhibitor with excellent CNS penetration that has been advanced to first-in-human phase I clinical trials.

Full text of DOI:10.1021/acs.jmedchem.9b01090

Subscriber access provided by Nottingham Trent University
Article
Discovery of MK-8719, A Potent O-GlcNAcase
Inhibitor as a Potential Treatment for Tauopathies
Harold Selnick, J. Fred Hess, Cuyue Tang, Kun Liu, Joel Schachter, Jeanine E. Ballard,
Jacob Newman Marcus, Daniel J. Klein, Xiaohai Wang, Michelle Pearson, Mary J. Savage,
Ramesh Kaul, TONGSHUANG LI, David J. Vocadlo, Yuanxi Zhou, Yongbao Zhu, Changwei
Mu, Yaode Wang, Zhongyong Wei, Chang Bai, Joseph L Duffy, and Ernest J. McEachern
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01090 • Publication Date (Web): 05 Sep 2019
Downloaded from pubs.acs.org on September 5, 2019
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 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 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.
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 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of MK-8719, A Potent O-GlcNAcase
Inhibitor as a Potential Treatment for Tauopathies
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
†
†
†
†
Harold G. Selnick , J. Fred Hess , Cuyue Tang , Kun Liu , Joel B. Schachter , Jeanine
†
†
†
†
†
E. Ballard , Jacob Marcus , Daniel J. Klein , Xiaohai Wang , Michelle Pearson , Mary J.
†
‡
‡
‡
‡
Savage , Ramesh Kaul , Tong-Shuang Li , David J. Vocadlo , Yuanxi Zhou , Yongbao
‡
∥
∥
∥
∥
†
Zhu , Changwei Mu , Yaode Wang , Zhongyong Wei , Chang Bai , Joseph L. Duffy* ,
and Ernest J. McEachern*‡
†Merck & Co., Inc., 2015 Galloping Hill Road, Kenilworth, NJ 07033 USA
‡Alectos Therapeutics, 8999 Nelson Way, Burnaby, British Columbia, V5A 4B5, Canada
∥
Pharmaron Beijing Co. Ltd., 6 Taihe Road, BDA, Beijing, China 100176, P.R. China
ABSTRACT: Inhibition of O-GlcNAcase (OGA) has emerged as a promising therapeutic
approach to treat tau pathology in neurodegenerative diseases such as Alzheimer’s
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
disease and progressive supranuclear palsy. Beginning with carbohydrate-based lead
molecules, we pursued an optimization strategy of reducing polar surface area to align
the desired drug-like properties of potency, selectivity, high central nervous system (CNS)
exposure, metabolic stability, favorable pharmacokinetics, and robust in vivo
pharmacodynamic response. Herein, we describe the medicinal chemistry and
pharmacological studies that led to the identification of (3aR,5S,6S,7R,7aR)-5-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(difluoromethyl)-2-(ethylamino)-3a,6,7,7a-tetrahydro-5H-pyrano[3,2-d]thiazole-6,7-diol
42 (MK-8719), a highly potent and selective OGA inhibitor with excellent CNS penetration
that has been advanced to first-in-human Phase I clinical trials.
F
O
S
N
F
NH
HO
OH
42
MK-8719
ACS Paragon Plus Environment

Page 3 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
One of the defining pathological hallmarks of Alzheimer’s disease (AD) is the
accumulation of neurofibrillary tangles (NFTs) in the brains of patients. NFTs are
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
composed primarily of aggregates of the microtubule-associated protein tau. During the
course of disease progression, aggregation of tau appears to be driven by its abnormal
hyperphosphorylation leading to self-assembly into oligomers, which then form the paired
helical filaments that make up NFTs. The presence of pathological hyperphosphorylated
tau aggregates is a common feature of several neurodegenerative disorders that are
collectively termed the tauopathies; these include progressive supranuclear palsy (PSP),
corticobasal degeneration, Pick’s disease, and frontotemporal dementias with
2
parkinsonism linked to chromosome 17. A substantial body of evidence indicates that
pathological tau protein plays a central role in driving neuronal cell death in these
diseases. Consistent with this hypothesis, the extent of NFTs correlate with clinical
3
progression of AD. Moreover, over 44 different tau mutations have been identified that
4
markedly increase the risk of these diseases. Emerging data supports the idea that
soluble hyperphosphorylated tau species are involved in prion-like propagation of tau
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
pathology throughout the brain in AD.3,5 In light of these findings, major efforts are focused
on reducing pathogenic tau species as a disease-modifying therapeutic approach for AD
and related tauopathies.6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A post-translational modification (PTM) of tau that has emerged recently is O-
GlcNAcylation, which involves glycosylation of serine and threonine protein residues with
O-linked N-acetylglucosamine (O-GlcNAc). O-GlcNAcylation is a reversible PTM that
occurs in all multicellular eukaryotes and is found on hundreds of nuclear and cytoplasmic
7
proteins, including tau. In mammalian cells, O-GlcNAcylation is regulated by two highly-
conserved enzymes: the glycosyltransferase O-GlcNAc transferase (OGT), which
catalyzes the addition of O-GlcNAc to protein substrates, and the glycoside hydrolase O-
GlcNAcase (OGA), which catalyzes the hydrolytic cleavage of O-GlcNAc from proteins.
8
These modifications are illustrated by process a) in Figure 1. Because elevated O-
GlcNAc levels have been found in some cases to reduce serine and threonine
phosphorylation, O-GlcNAcylation has been proposed to modulate phosphorylation either
directly, by modifying phosphorylation sites, or indirectly, through glycosylation of
7
residues proximal to these sites. Notably, it has been suggested that O-GlcNAcylation of
ACS Paragon Plus Environment

Page 5 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
tau regulates its phosphorylation state, with increased O-GlcNAc modification correlated
9
with lower tau phosphorylation. In addition, in vitro studies have demonstrated that
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
increased O-GlcNAcylation of tau hinders its propensity for aggregation.10,11 Taken
together, these data suggest that pharmacological blockade of OGA would result in an
increase of O-GlcNAc modified tau and reduce the formation of pathogenic tau species
and tau aggregates (Figure 1), thereby providing a therapeutic benefit in diseases
associated with tau hyperphosphorylation such as AD and other tauopathies.12
Consistent with this idea, multiple independent studies using transgenic mice expressing
mutant forms of human tau have shown that administration of the small-molecule OGA
inhibitor Thiamet-G (1)13,14 confers beneficial disease-modifying effects. These benefits
include reduction in phosphorylated tau species and tau aggregates, reduced levels of
tau in cerebrospinal fluid (CSF), decreased neuronal cell loss, and a reduction in disease-
associated behavioral phenotypes.11,15,16,17,18 On the basis of these findings, OGA
inhibition has emerged as a promising therapeutic strategy that could be used to treat tau
pathology in patients. Herein we report pharmacological studies and medicinal chemistry
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
modification of 1 leading to identification of the first-in-class clinical OGA inhibitor, MK-
719.19
8
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Hypothesis for increased O-GlcNAc modification hindering tau aggregation. a)
Cellular modification of tau protein with O-GlcNAc (G) by OGT maintains the protein in a
stable soluble form. Equilibrium levels of O-GlcNAc are maintained by the antagonistic
actions of OGT and OGA. b) Abnormal disease-associated addition of phosphoryl
residues (P) by kinases, countered by phosphatases, coupled with additional
modifications including addition of nitro (N), acetyl (A) and ubiquitin (U) groups leads to
oligomers and subsequently to c) insoluble tau aggregates.
OGA is a two-domain protein, with a C-terminal domain similar to the GCN5-related
family of N-acetyltransferases and an N-terminal domain that is a member of Glycoside
Hydrolase family 84 (GH84).8,20 It is the N-terminal catalytic domain of OGA that
hydrolyzes the glycosidic bond of O-GlcNAc linked to proteins. As illustrated in Figure 2,
ACS Paragon Plus Environment

Page 7 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
OGA uses a two-step catalytic mechanism involving participation of the substrate
acetamido group, featuring an oxazoline intermediate (A) bound to the enzyme active
site.21,22 During the catalytic cycle, the acetamido group acts as a nucleophile to attack
the anomeric center and displace the protein hydroxyl group leading to the formation of
the oxazoline intermediate A, which is then hydrolyzed to liberate N-acetylglucosamine.
Two enzymic residues play central roles in catalysis.21,23,24,25 Asp174 hydrogen bonds
with the acetamido group of the substrate, orienting this residue to place the carbonyl
oxygen adjacent to the anomeric center and serving as a general base to enhance the
nucleophilicity of this group. A second residue, Asp175, acts as a general acid,
concomitantly aiding expulsion of the hydroxyl leaving group and favoring formation of A.
NAG-thiazoline (B, Figure 2), which resembles the oxazoline intermediate A, was shown
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
to potently inhibit human OGA (K = 70 nM) but to lack selectivity over the functionally
i
21
related human lysosomal -hexosaminidase from GH20 (hHEX K = 70 nM).
i
Concomitant inhibition of -hexosaminidase is not desirable, as loss-of-function mutations
in these enzymes result in the lysosomal storage disorders Tay-Sachs and Sandhoff
diseases. A series of NAG-thiazoline analogues with improved selectivity were identified
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
and shown to be mimics of the OGA transition state.26 Subsequently, aminothiazoline
analogues were prepared that, unlike the corresponding NAG-thiazoline derivatives, are
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
protonated at physiological pH and benefit from an ion-pair interaction of the
aminothiazoline substituent with the anionic Asp174 carboxylate group.13,27 These
inhibitors, including Thiamet-G (1, Figure 2), are stable transition state analogues for
human OGA and exhibit high ligand efficiencies.27 X-ray cocrystal structures of
homologues of OGA23 and the GH84 domain of human OGA in complex with 128,29,30
provide support for the importance of the ionic interaction between the protonated
aminothiazoline and Asp174, as well as providing a clear atomic rationale for the
selectivity over lysosomal -hexosaminidase conferred by increasing the steric bulk at the
2
′-position of the thiazoline moiety.13,21,27 Inhibitor 1 sits deeply within a pocket at one face
of the (βα) -barrel structure where it is held within an extensive network of hydrogen
8
bonds with the 2′-ethylamino substituent of the thiazoline tucked into a discrete pocket.
ACS Paragon Plus Environment

Page 9 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
D175
D175
D175
O
O
O
Protein
O
O
O
Protein
H
H
O
H
O
H
O
H
O
O
HO
H
HO
HO
HO
HO
HO
O
HO
O
HO
HO
N
O
N
N
O
H
H
H
O
H
H
O
O
O
D174
O
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
D174
(A)
D174
O
O
OH
OH
OH
O
O
O
HO
HO
HO
HO
HO
HO
N
N
N
O
S
S
HN
Oxazoline Intermediate (A)
NAG-Thiazoline (B)
Thiamet-G (1)
Figure 2. Mechanism-inspired basis for inhibition of OGA and its substrate-assisted
catalytic mechanism. The catalytic mechanism of OGA involves two carboxyl residues
within the enzyme active site (D174 and D175) that catalyze a two-step process
proceeding through an oxazoline intermediate (A), which leads to cleavage of GlcNAc
from modified proteins. The known inhibitors NAG-Thiazoline (B) and Thiamet-G (1)
resemble the oxazoline intermediate (A) or a closely-related transition state.
RESULTS AND DISCUSSION
Several classes of carbohydrate analogues have been reported that potently inhibit
human OGA in cells and tissues,13,21,31,32,33,34,35,36 but 1 (K = 2.1 nM) has emerged as
27
i
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a particularly effective tool compound. Within cells it has a reported EC value of 21 nM
50
13
for increasing O-GlcNAc modified protein to half-maximal level. In rodent models 1 has
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
been shown to dramatically elevate brain protein O-GlcNAc levels11,13,15,16,17,18 while also
being well tolerated over extended treatment periods and demonstrating beneficial effects
in a range of tauopathy models.11,15,16,17,18,37
While 1 shows several desirable properties as a tool compound, it also possesses
features that could limit its clinical utility as a therapeutic agent for the treatment of
tauopathies. Initial profiling of 1 in an assay to measure its apparent permeability (Papp)
revealed a Papp value below the detection limit of the assay, that is, less than 1.0 x 10-6
cm/s using a monolayer of LLC-PK1 cells transfected with MDR1.38 This result indicated
that 1 possesses comparatively low or slow membrane penetration, which is inauspicious
for achieving high brain exposure. Pharmacokinetic profiling of 1 in rat provided
confirmation, showing an initial brain/plasma Cmax ratio of ~0.1 (Figure 3) following a
single oral gavage dose (po). The consistent brain concentration of 1 suggests a slow
diffusion rate into and out of the CNS from systemic circulation. Additionally, 1 exhibited
a high unbound fraction (f ) in human plasma, with f = 76%. We ascribe both features to
u
u
ACS Paragon Plus Environment

Page 11 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the high polar surface area (PSA) of this molecule. Surveys of CNS penetrant drugs have
2
established an optimal PSA range of 60–70 Å for blood-brain barrier permeability, with
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
an upper limit of 90 Å2 We examined the topological polar surface area (TPSA) of 1 and
39
.
2
40
found that it was outside this range (TPSA = 105 Å ). We hypothesized that we could
lower the TPSA by modifying polar substituents while maintaining high intrinsic potency,
which would lead to greater and more rapid biodistribution into the CNS, and thus a
superior clinical candidate.
1
0
1
1
Plasma
Brain
1
0.1
0.01
0.001
0
5
10
15
20
25
Hours
Figure 3. Rat plasma and brain pharmacokinetics for 1, 10 mg/kg po. Brain concentration
(nmol/g); plasma concentration (μM).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Our initial efforts to reduce the TPSA of 1 focused on the hydroxyl and amino groups of
the molecule, and we prepared a series of analogues in which these groups were either
methylated or, in the case of the OH groups, replaced by H or F (Table 1). All new
compounds were tested for their inhibitory potency against recombinant hOGA as well as
ancillary potency against hHEX. Cell-based activity for OGA inhibition was assessed
using an ELISA-based assay with rat PC12 cells to measure EC values for elevation of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
50
all protein O-GlcNAc levels. The MDR1-LLC-PK1 assay described above was used to
measure apparent permeability (P ) and Pgp efflux for selected compounds. None of
app
the tested compounds were found to be Pgp substrates, although this transporter activity
is difficult to reliably establish for compounds with Papp < 7. All compounds were assayed
for activity against the cardiovascular ion channels hERG (by displacement of
radiolabeled MK-0499, a known hERG blocker), Na 1.5, and Ca 1.2. The ion channel
V
V
results are not presented in the tables, as all the compounds are devoid of significant
activity against these targets (IC50 > 30 M). Similarly, all compounds were tested for
inhibitory potency against the metabolic enzymes CYP3A4, CYP2D6, and CYP2C9, and
were found to be inactive (IC > 50 M). Finally, all compounds were assayed for their
50
ACS Paragon Plus Environment

Page 13 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
propensity to induce CYP3A4 in a PXR activation assay and are similarly inactive (EC50
>
30 M).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
S
N
R1
N
R4
R2
R3
Table 1. Modification of hydroxyl or amino substituents from Thiamet-G.
Compound Structurea hOGA
Ki
rOGA
hHEX
TPSA Papp
cell EC50 Ki
(
Å)2 e (10-6
nM)b
(nM)c
(nM)d
cm/s)f
(
1
Thiamet-
G
0.41
13.5
>
105
< 1.0
10,000
2
3
R = OCH 190
--
--
--
94
91
--
--
1
3
R = OCH > 3,000 --
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
R = OCH 270
--
>
1
91
--
3
3
0,000
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
6
7
R = CH
5.5
69
36.7
--
3,600
93
84
86
1.1
--
4
3
R = H
1
--
>
R = H
2
> 3,000 --
7.0
10,000
8
R = H
3
44
364
>
1
84
2.6
0,000
9
R = F
1
20
176
--
>
84
84
6.1
--
g
10
R = F
2
> 3,000 --
10,000
11
R = (R)-F 29
3
443
>
1
84
2.7
0,000
ACS Paragon Plus Environment

Page 15 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Unless otherwise noted, the substituents R , = OH, R = (S)-OH, R = (R)-OH, and R
1
2
3
4
b
=
<
H. Human OGA enzyme inhibition K values are from > 2 assays with standard deviation
i
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
c
35% of the reported mean. Rat cellular OGA EC values are from > 2 assays with
50
d
standard deviation < 35% of the reported mean. hHEX K values were measured as
i
described in the supplementary material from > 2 assays with standard deviation < 22%
e
of the reported mean. Polar surface area (TPSA) was calculated as the sum of fragment-
f
based contributions. MDR1-LLC-PK1 cell monolayer permeability measurements are
g
from > 2 assays with a standard deviation < 55% of the reported mean. Single
diastereomer, absolute configuration at R not assigned.
2
The structural modifications outlined in Table 1 indicate that some alteration of 1 is
tolerated to reduce the TPSA without a complete loss of OGA activity. The (6S)-hydroxyl
substituent R is critical to the OGA inhibitory potency of 1. Any modification or removal
2
of this substituent, as illustrated by 3, 7, and 10, resulted in profound loss of activity of the
corresponding compound. Modification of the (7R)-hydroxyl substituent R was more
3
tolerated. Methylation of this moiety (4) afforded a reduction in calculated TPSA but also
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a significant loss of intrinsic potency. However, removal of the hydroxyl as in 842 or
42
replacement by fluorine as in 11 afforded compounds with further reduction in calculated
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
TPSA and superior potency as compared to the R modified compounds. The decrease
2
in TPSA correlated to measurable membrane penetration, as measured by Papp in LLC-
PK1 cells transfected with MDR1. Modification of the (5R)-hydroxymethyl substituent (R1)
41
on 1 was more productive. Although methylation of this substituent (2 ) resulted in a
poorly active compound, removal of the hydroxyl provided 642 with improved potency
42
relative to 2 and a further reduction in TPSA. Substitution of the hydroxyl by fluorine (9 )
afforded improved OGA inhibitory potency while maintaining the reduction in TPSA
achieved by 6, and this lower TPSA translated to significantly improved membrane
permeability (P ) as compared to 1. Finally, the addition of a methyl substituent to the
app
14
2-ethylamino moiety (R ) provided 5 with a moderate reduction in both TPSA and
4
inhibitory potency as compared to 1.
This first round of modifications suggested that further improvements in TPSA reduction
could be achieved by combining the structural modifications outlined above. The (6S)-
hydroxyl substituent R was retained as critical for potency, but the combined
2
ACS Paragon Plus Environment

Page 17 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
modifications suggested by the data in Table 1 were extended to additional substituents
to further improve permeability while retaining OGA inhibitory potency. The results from
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
these studies are provided in Table 2.
O
S
N
R5
R1
HO
N
R4
R3
Table 2. Modification of hydroxyl or amino substituents from 1.
Structurea
hOGA
rOGA cell hHEX A/B TPS Papp
b
d
K (nM) EC50
K (nM)
i
A
i
-6
f
(
10 cm/s)
(nM)c
(Å)2 e
R1 R3
R4
R5
1
2
2-F ethyl 5.6
509
152
> 10,000
4,500
105
92
2.2
0.8
1
3
Me allyl
2.8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
4
F
F
F
F
F
H
1,200
180
14
--
--
96
105
82
69
83
71
--
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
5
allyl
propyl
--
> 10,000
> 10,000
1,600
--
1
6
119
177
34.4
--
17
6.8
5.3
--
1
7
Me Me
9.0
1
8
Me
0.55
41
790
1
9
(R)-F Me Me
> 10,000
ACS Paragon Plus Environment

Page 19 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
0
(R)-F
(R)-F
(S)-F
H
propyl
Me
14
445
307
10.6
> 10,000
8,100
83
84
84
98
83
72
--
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
1
3.0
0.53
--
2
2
Me
> 10,000
> 10,000
> 10,000
> 10,000
--
2
3
H
> 3,000 --
--
2
4
H
propyl
50
11
224
2.5
--
2
5
H
Me Me
358
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
6
H
Me
Me
Me
9.4
50
33.1
> 10,000
> 10,000
> 10,000
> 10,000
> 10,000
> 10,000
85
61
61
91
92
93
--
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
7
F
F
(R)-F
(S)-F
--
--
2
8
35
328
--
26
--
2
9
1,800
N
N
N
3
0
> 3,000 --
--
F
3
1
13
165
--
ACS Paragon Plus Environment

Page 21 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
F
F
3
2
1,700
960
--
--
94
93
96
95
94
71
--
N
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
--
> 10,000
> 10,000
--
--
N
N
3
4
3.3
28.0
--
3
5
N
N
N
F
> 3000 --
--
3
6
Me
72
--
--
> 10,000
> 10,000
2.4
--
3
7
F
200
a
b
Unless otherwise noted, the substituents R , = OH, R = (R)-OH, and R = H. Human
1
3
4
OGA enzyme inhibition K values are from > 2 assays with standard deviation < 66% of
i
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
c
the reported mean. Rat cellular OGA EC values are from > 2 assays with standard
50
d
deviation < twofold of the reported mean. hHEX K values were measured as described
i
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
in the supplementary material from > 2 assays with standard deviation < 79% of the
e
reported mean. Polar surface area (TPSA) was calculated as the sum of fragment-based
f
contributions. MDR1-LLC-PK1 cell monolayer permeability measurements are from > 2
assays with a standard deviation < 53% of the reported mean.
Structural modification of the aminoethyl R substituent from 1 afforded no appreciable
5
reduction in TPSA, and diminished potency against OGA as illustrated by comparing
43
1
214,27 with 1 and 13 with 5 (from Table 1). Like the findings from Table 1, replacement
of the primary alcohol R with a fluoro substituent as in 9 was more productive in reducing
1
the TPSA, which correlated with higher membrane permeability in the Papp assay. The
combination of this change with modifications to the aminothiazole substituents appear
42
to amplify the potency impact of the amine position. The unsubstituted amine 14 lost
significant potency compared to the ethylamine substituted derivative 9, and even the
42
corresponding allyl derivative 15 was significantly less potent than 9. The propylamine
ACS Paragon Plus Environment

Page 23 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
derivative 1642 and dimethylamine derivative 1742 afforded comparable potency to 9.
However, the methylamine substituent in 1842 provided a significant improvement in
hOGA potency, as well as in the cell-based assay using the rat enzyme, albeit with
concomitant increased potency for the lysosomal hHEX enzyme.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Modification of the R hydroxyl was investigated for combination SAR with other
3
changes in the molecule. The replacement of the (R)-OH in R with (R)-F afforded 19 -
3
2
1.42 While the corresponding R -F propylamine derivative 20 displayed comparable
3
potency to the R -F counterpart 16, the methylamine derivative 21 and dimethylamino
1
derivative 19 were less potent than their counterparts 18 and 17 respectively. The
diastereomeric (S)-F R3 derivative 2242 was significantly more potent than the
corresponding (R)-F diastereomer 21, with higher selectivity against hHEX as well.
42
Elimination of the R hydroxyl group provided 23 - 26, which were generally less potent
3
than the corresponding R fluoro derivatives. Combination of the fluoro substituents in R
3
1
42
42
and R led to 27 and 28 with a significant reduction in TPSA. This corresponded to
3
high membrane permeability in the Papp assay for 28, but also to a significant loss of
intrinsic potency against hOGA.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The disubstituted amine derivatives 5, 13, 17, 19, and 25 retained significant OGA
potency while reducing TPSA, which led us to consider the potential of cyclic amine
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
44
derivatives. The piperidine compound 29 lost significant activity, whereas the pyrrolidine
product 3144 and azetidine 3444 were more active. Substitution of the cyclic amine
44
44
44
moieties with a fluoro substituent as in 30, 32, 33, and 35 resulted in a significant
loss of hOGA potency relative to the corresponding non-fluorinated derivatives. We
ascribe this to the inductively decreased basicity of the amino substituent, which reduces
the interaction of this moiety with Asp174 in the hOGA binding pocket.27
The high membrane permeability achieved with the doubly fluorinated compound 28
led us to speculate that an alternate arrangement of difluoro substituents on the pyran
moiety may be optimal for both permeability and potency. To investigate this hypothesis,
a variety of analogs with difluoro substituents in R were synthesized, and the results are
1
shown in Table 3.
F
O
S
N
R5
F
N
R4
HO
R3
ACS Paragon Plus Environment

Page 25 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. Investigation of 5-difluoromethyl derivatives of 1.
Structurea
hOGA
rOGA cell hHEX A/B TPS Papp
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
b
d
K (nM) EC50
K (nM)
A
i
i
-6
f
(
10 cm/s)
(nM)c
(Å)2 e
R3
H
R4
R5
3
8
Me
Me
> 3,000 --
> 10,000
> 10,000
> 10,000
> 10,000
47
61
82
82
31
24
12
14
3
9
H
Me
28
74
5.3
--
4
0
allyl
propyl
--
4
1
41.4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
2
ethyl
Me
H
7.9
0.36
760
2.4
58
52.7
1.71
2,400
31.2
--
> 10,000
1,300
80
83
96
69
61
61
6.4
5.7
9.2
14
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
3
4
4
> 10,000
3,900
4
5
Me
Me
Me
Me
4
6
(R)-F
(S)-F
> 10,000
> 10,000
22
4
7
23
58.2
20
ACS Paragon Plus Environment

Page 27 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
8
N
26
--
> 10,000
71
18
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
b
Unless otherwise noted, the substituents R = (R)-OH, and R = H. Human OGA
3
4
enzyme inhibition K values are from > 2 assays with standard deviation < 58% of the
i
c
reported mean. Rat cellular OGA EC values are from > 2 assays with standard
5
0
d
deviation twofold of the reported mean. hHEX K values were measured as described in
i
the supplementary material from > 2 assays with standard deviation < twofold of the
e
reported mean. Polar surface area (TPSA) was calculated as the sum of fragment-based
f
contributions. MDR1-LLC-PK1 cell monolayer permeability measurements are from > 2
assays with a standard deviation < 43% of the reported mean.
From the difluoro-substituted compounds illustrated in Table 3, the highest reduction in
TPSA was achieved with the dimethylamino derivative 38.42 This TPSA reduction
translated to the highest Papp in this lead class, albeit with complete loss of hOGA activity.
The activity was restored with the removal of a single methyl from the amine substituent,
42
affording 39 as a high permeability compound with appreciable potency.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 201
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The aminothiazoline substituents were systematically investigated with 40 - 45,42
indicating that the propyl and ethyl derivatives 41 and 42 maintained a good balance of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
hOGA potency, P , and selectivity vs. hHEX. The potent methylamine derivative 43 was
app
further investigated to improve selectivity against hHEX. Substitution at R with a fluorine
3
42
42
as in 46 and 47 did diminish hHEX potency, but at a significant cost to hOGA potency
42
as well. Finally, the azetidine analog 48 was investigated, but this compound afforded
significantly diminished hOGA potency as compared to the dimethylamine analog 45.
The full collection of Papp data for compound in Tables 1-3 was compared to the
calculated TPSA for these compounds to evaluate the design principle and strategy for
optimization, and this data is provided graphically in Figure 4. As expected, there is a
reasonable correlation between compounds with reduced TPSA and improved
permeability as measured by Papp.
ACS Paragon Plus Environment

Page 29 of 201
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
3
2
2
1
1
5
0
5
0
5
0
5
0
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
40
50
60
70
80
90
100
110
PSA (Å)2
Figure 4. Correlation between Papp and calculated TPSA for OGA inhibitors in Tables 1-
3.
Several compounds from each sub-series were further investigated for their
pharmacokinetic properties in rats, and the results of the studies are provided in Table 4.
All compounds exhibiting hOGA K < 35 nM were selected for rat PK profiling unless a
i
more potent and closely homologous compound was already included in the PK profiling
set. Compounds from each sub-series in Tables 1-3 were evaluated for their free fraction
in rat and human plasma, and these results are also included in Table 4. All of the
compounds tested in the plasma binding assays were found to have a plasma free fraction
ACS Paragon Plus Environment