Discovery of Atabecestat (JNJ-54861911): A Thiazine-Based β-Amyloid Precursor Protein Cleaving Enzyme 1 Inhibitor Advanced to the Phase 2b/3 EARLY Clinical Trial

Article abstract of DOI:10.1021/acs.jmedchem.0c01917

Accumulation of amyloid β peptides (Aβ) is thought to be one of the causal factors of Alzheimer's disease (AD). The aspartyl protease β-site amyloid precursor protein cleaving enzyme 1 (BACE1) is the rate-limiting protease for Aβ production, and therefore

Full text of DOI:10.1021/acs.jmedchem.0c01917

pubs.acs.org/jmc
Drug Annotation
Discovery of Atabecestat (JNJ-54861911): A Thiazine-Based
β‑Amyloid Precursor Protein Cleaving Enzyme 1 Inhibitor Advanced
to the Phase 2b/3 EARLY Clinical Trial
Yuji Koriyama,* Akihiro Hori, Hisanori Ito, Shuji Yonezawa, Yoshiyasu Baba, Norihiko Tanimoto,
Tatsuhiko Ueno, Shiho Yamamoto, Takahiko Yamamoto, Naoya Asada, Kenji Morimoto,
Shunsuke Einaru, Katsunori Sakai, Takushi Kanazu, Akihiro Matsuda, Yoshitaka Yamaguchi,
Takuya Oguma, Maarten Timmers, Luc Tritsmans, Ken-ichi Kusakabe,* Akira Kato, and Gaku Sakaguchi
Cite This: J. Med. Chem. 2021, 64, 1873−1888
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Accumulation of amyloid β peptides (Aβ) is thought to be one of the causal factors of
Alzheimer’s disease (AD). The aspartyl protease β-site amyloid precursor protein cleaving enzyme 1
(BACE1) is the rate-limiting protease for Aβ production, and therefore, BACE1 inhibition is a
promising therapeutic approach for the treatment of AD. Starting with a dihydro-1,3-thiazine-based
lead, Compound J, we discovered atabecestat 1 (JNJ-54861911) as a centrally efficacious BACE1
inhibitor that was advanced into the EARLY Phase 2b/3 clinical trial for the treatment of preclinical
AD patients. Compound 1 demonstrated robust and dose-dependent Aβ reduction and showed
sufficient safety margins in preclinical models. The potential of reactive metabolite formation was
evaluated in a covalent binding study to assess its irreversible binding to human hepatocytes. Unfortunately, the EARLY trial was
discontinued due to significant elevation of liver enzymes, and subsequent analysis of the clinical outcomes showed dose-related
cognitive worsening.
secretase) followed by γ-secretase to generate Aβ peptides. As
BACE1 is a rate-limiting enzyme of Aβ production, it is one of
the prime targets for AD therapeutics, and its inhibition offers
the potential to delay or even prevent the disease progression,
given its critical role in Aβ accumulation in the pathogenesis of
AD.5 The Icelandic mutation of A673T in APP reduces affinity
for BACE1, resulting in approximately 40% inhibition of Aβ
production and is protective against AD and cognitive
impairment in elderly individuals, providing further support
for BACE1 being a promising target for AD therapeutics.6
Since the identification of BACE1 in 1999, a number of
academic and industrial institutes have pursued potent and
centrally active BACE1 inhibitors.7 Initial explorations focused
on transition state and substrate-based drug design, leading to
identification of high affinity peptide and peptidomimetic
inhibitors with high topological polar surface area (TPSA) and
molecular weight. Such physical properties, however, prevented
cellular and brain penetration, thereby resulting in suboptimal
Aβ reduction in the brain following oral administration. The
discovery of amidine-based inhibitors overcame the obstacles
INTRODUCTION
■
Globally, there were over 50 million people living with dementia
in 2019, and the number is expected to increase to 152 million
by 2050.1 Alzheimer’s disease (AD) is the most common cause
of dementia, which accounts for approximately 60−80% of its
cases. The cost of medical care and long-term care for
individuals with dementias, including AD, presents a substantial
impact on public and personal finance. Indeed, in the United
States, the total cost of caring for patients with dementia is
estimated to increase from $305 billion in 2020 to more than
$1.1 trillion in 2050.2 Unfortunately, no therapeutic is yet
available to prevent, reverse, or even slow the disease, although
some symptomatic agents, such as donepezil or memantine, are
being prescribed. Interestingly, one estimation suggests that
delaying the onset by just 5 years may suppress the number of
cases by up to 50% over 50 years, pointing to the need to develop
effective disease-modifying treatments.3
Evidence obtained in the Dominantly Inherited Alzheimer
Network (DIAN) study enrolling individuals who carry
mutations in genes such as APP, PSEN1, or PSEN2 that cause
familial AD has indicated that amyloid β peptide 42 (Aβ42) levels
in the cerebrospinal fluid (CSF) begin to decline 25 years before
cognitive impairment becomes evident, and Aβ deposition in the
brain is detected 15 years before the onset of symptoms, which
points to accumulation of Aβ as a causative of AD.4 Amyloid
precursor protein (APP) is first cleaved by β-site amyloid
precursor protein cleaving enzyme 1 (BACE1, also known as β-
Received: November 5, 2020
Published: February 15, 2021
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
© 2021 American Chemical Society
J. Med. Chem. 2021, 64, 1873−1888
1873

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Chart 1. Representative BACE1 Inhibitors That Advanced into Clinical Trials
Figure 1. Medicinal chemistry progression from hit to clinical candidate.
observed with the peptidomimetic inhibitors, as they have more
optimal TPSA and molecular weight for central nervous system
(CNS) drugs.7 Elaboration of amidine-based inhibitors along
with structure-based drug design and fine-tuning of its basicity
culminated in clinical compounds such as atabecestat (JNJ-
54861911, 1),8 LY-2811376 (2),9 LY-2886721 (3),10 LY-
3202626 (4),11 verubecestat (MK-8931, 5),12 lanabecestat
(AZD-3293, 6),13 elenbecestat (E-2609, 7),14 and umibecestat
(CNP-520, 8)15 (Chart 1).
Unfortunately, clinical trials of all of these BACE1 inhibitors
were discontinued due to findings of lack of efficacy, retinal/liver
toxicity, or even cognitive impairment.7f,16 The clinical trials for
compounds 1, 2, and 3 were terminated because compounds 18d
and 310 exhibited liver toxicity likely due to the potential of
reactive metabolite formation caused by thiazine structures,
while retinal toxicity derived from cathepsin D (CatD)
inhibition was evident with 2.9 This raised the importance of
structural alerts resulting in reactive metabolite formation and
minimizing inhibition of cathepsin D. Compounds 5,12d,e 6,13b
and 714a did not meet their clinical end point or showed no
clinical benefit in the interim analysis in patients with mild
cognitive impairment (MCI) or mild to moderate AD.
Considering that Aβ accumulation begins 15 years before
cognitive symptom onset,4 their administration was too late to
start therapeutic intervention even for patients with MCI. Two
Phase 2/3 clinical trials of compounds 1 and 8 evaluated the
ability to prevent AD in cognitively unimpaired people who were
at high risk for dementia due to AD (i.e., preclinical AD).
Disappointingly, both 18c and 815b led to cognitive worsening,
and other BACE1 inhibitors 4 and 5 also displayed cognitive
liability.12e The clinical evidence points to the importance of
further elucidating the biological function of BACE1 and its
homologue BACE2, for which the clinical BACE1 inhibitors
have suboptimal selectivity. Note that BACE2 was found to be
expressed in the brain and to have several neural substrates.17
Further considerations are presented in Future Directions.
1874
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Table 1. Head Group Exploration to Reduce Basicity
a
BACE1 IC50 values were determined using biochemical homogeneous time-resolved fluorescence (HTRF)-based assay and are the averages of at
least two determinations. For 1, mean SD was calculated from triplicate data. β-Secretase Inhibitor IV (CAS#: 797035-11-1) showed an IC50
b
value of 13 nM. Cellular IC50 values were determined by measuring the levels of secreted Aβ40 in human APP-transfected human neuroblastoma
(SH-SY5Y) cells via an HTRF-based assay and are the average of at least two determinations. For 1, mean SD was calculated from triplicate data.
z
β-Secretase Inhibitor IV showed an IC50 value of 25 nM. Metabolic stability represents percent remaining after 30 min incubation in human liver
d
(HLM) and rat (RLM) microsomes. hERG inhibition was measured at 5 μM in CHO cells transfected with hERG channels using an automated
e
f
patch clamp system. P-gp efflux ratio was determined with MDCK cells transfected with human MDR1 at Absorption Systems. Permeability was
g
measured using wild-type MDCK cells. pKa was determined by capillary electrophoresis (CE).
Herein, we disclose the design, synthesis, and characterization of
thiazine-based BACE1 inhibitors and the rational design based
on a pKa lowering approach which culminated in the discovery
of atabecestat 1 that was advanced into the Phase 2b/3 clinical
trial EARLY for the prevention of AD.
group resulted in 12 possessing a cyanopyridine, which
mitigated CYP2D6 inhibition (11: IC50 = <1 μM, 12: IC50
=
3.6 μM). At the time that 11 and 12 were identified, avoiding
QT interval prolongation caused by inhibition of the hERG
channel became a requirement for drug discovery programs;21 in
addition, P-gp mediated transport was found to be the main
gatekeeper for the blood−brain barrier (BBB).22 Unfortunately,
the lead 12 showed high hERG inhibitory activity (91%
inhibition at 5 μM) and high P-gp efflux (efflux ratio = 52 in
MDCK cells), which required further optimization to mitigate
these issues (Figure 1).
BACKGROUND OF LEAD IDENTIFICATION
■
Shortly after BACE1 was discovered in 1999, our research
program started with high throughput screening (HTS) using a
biochemical HTRF BACE1 assay.18 However, this campaign
drew blanks similarly to those of other pharmaceutical
companies. Most companies subsequently turned to fragment-
or substrate-based drug design utilizing the crystal structure of
OM99-2 bound to BACE1.7a,b Regardless of this trend, we
devoted our efforts to a phenotypic HTS based on a cellular
assay, where cells expressing human wild-type human APP, and
inhibition of Aβ40 levels were measured. In 2003, the HTS
campaign successfully afforded a hit compound 9 with a 1,3-
thiazine scaffold having an IC50 value of 2.6 μM without
significant cellular toxicity (Figure 1).18 In parallel with our
efforts, the research group at Roche also identified 9 in their
HTS campaign.19 In 2005, our hit-to-lead efforts utilizing
BACE1 structures led to the identification of 11, known as
Compound J,20 via an interim lead 10, which gave the first
thiazine-based BACE1 inhibitors disclosed in patent literature.
More importantly, the amide linker contained in 11 is a
commonly used pharmacophore in clinical BACE1 inhibitors
and contributes to excellent selectivity over the antitarget
CatD.7a Further efforts using parallel synthesis of the amide
INHIBITOR DESIGN AND OPTIMIZATION
■
At the onset of our lead optimization effort, several reports
demonstrated that controlling basicity and lipophilicity could be
a successful approach to reducing hERG inhibition.23 Analysis of
physicochemical features related to P-gp efflux had indicated
that a compound with a pKa < 8 would be less likely to be a P-gp
substrate.24 The pKa and LogD at pH 7.4 for 12 were
determined to be 9.0 and 1.2, respectively. Thus, the highly
basic and moderately lipophilic profile of 12 prompted us to
adopt the pKa-lowering approach to reduce both hERG
inhibition and P-gp efflux.7 To test this hypothesis, we pursued
a series of analogues containing a double bond or a fluorine atom
on the thiazine headgroup (Table 1). The two analogues with a
double bond (1 and 13) successfully reduced pKa, leading to a
significantly reduced P-gp efflux, whereas the thiazine 13 with an
exocyclic double bond was accompanied by reduced metabolic
stability in human and rat microsomes. It was unclear why the P-
1875
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Table 2. Tail Group Exploration to Reduce hERG Potency
a
BACE1 IC50 values were determined using biochemical homogeneous time-resolved fluorescence (HTRF)-based assay and are the average of at
least two determinations.. For 1, mean SD was calculated from triplicate data. β-Secretase Inhibitor IV (CAS#: 797035-11-1) showed an IC50
b
value of 13 nM. Cellular IC50 values were determined by measuring the levels of secreted Aβ40 in human APP-transfected human neuroblastoma
(SH-SY5Y) cells via an HTRF-based assay and are the average of at least two determinations. For 1, mean SD was calculated from triplicate data.
z
β-Secretase Inhibitor IV showed an IC50 value of 25 nM. Metabolic stability represents percent remaining after 30 min incubation in human
d
(HLM) and rat liver (RLM) microsomes. hERG inhibition was measured at 5 μM in CHO cells transfected with hERG channels using an
e
f
automated patch clamp system. P-gp efflux ratio was determined in MDCK cells transfected with human MDR1 at Absorption Systems. CYP2D6
inhibition was measured with 5 μM dextromethorphan using human microsomes.
Figure 2. Cocrystal structure of 1 bound to human BACE1 determined at 2.3 Å resolution (PDB: 7DCZ). Key residues of Asp228, Asp32, Tyr71,
Gly230, and Ala335 are shown in stick style. Compound 1 is shown by sticks with green carbons. Oxygen, nitrogen, fluorine, and sulfur atoms are
colored red, blue, sky blue, and yellow, respectively. Hydrogen bonds are shown as dashed lines. The pocket surface colored gray is shown in B.
gp efflux ratio in 14 with a 4-fluoromethyl group remained
moderate-to-high despite a basicity similar to 13. Perhaps, the
higher intrinsic permeability observed for 13 relative to that of
14 could be attributed to the reduced P-gp efflux. Compound 1
achieved a balanced profile in terms of potency, P-gp efflux, and
metabolic stability, whereas the hERG inhibitory activity could
be improved.
The amide substituents (the tail group) of compound 1 were
explored to lower hERG inhibitory activity (Table 2). As a
result, incorporation of substituents at the ortho-position on the
carbonyl in 1, such as 15, lowered hERG activity relative to 1,
but this improvement came at the cost of increased P-gp efflux. A
change to a fluorine instead of the cyano group in 1 was
accompanied by loss of BACE1 and cellular potency. Similarly to
compound 11, the chloropyridine analogue 17 had high hERG
1876
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Figure 3. Effects of compound 1 on CSF Aβ40 levels after single oral administration in male Sprague−Dawley (Crl:CD(SD)) rats (n = 64, vehicle: 0.5%
methyl cellulose (MC); n = 63, untreated control for Aβ measurement; n = 32, each dose). Compound 1 was administered orally, and Aβ40 amounts in
the CSF were measured using an ELISA system at the points of 0.5, 1, 2, 4, 6, 8, 10, and 24 h after administration. Aβ levels are represented as the
percentage of untreated rats at each sampling point. Data shown were mean SEM for 7 to 8 rats in each time point of each group. *p < 0.05, **p <
0.01 versus vehicle group (Dunnett’s test).
potency and CYP2D6 inhibition. A significant reduction in
hERG potency was realized when a methoxypyrazine was
incorporated. Unfortunately, this change was also associated
with a CYP2D6 IC50 value below 1 μM and low metabolic
stability in human microsomes. Although a backup program18,25
for atabecestat later showed us how to balance profiles among
hERG, CYP2D6 inhibition, and metabolic stability by keeping
the methoxypyrazine tail group, this issue was not addressed
during this optimization effort. Therefore, compound 1 (JNJ-
54861911, atabecestat) with the most promising profile was
advanced to in vivo cardiovascular safety studies following
pharmacokinetic (PK) and pharmacodynamic (PD) character-
ization to correctly estimate the cardiovascular risk associated
with its hERG potential.
The cocrystal structure of 1 bound to human BACE1 was
solved (Figure 2). The amidine on the thiazine headgroup forms
hydrogen bond interactions with the catalytic residues Asp32
and Asp228. The amide hydrogen also engaged in the hydrogen
bond interaction, confirming the molecular basis of the amide
pharmacophore. The cyano group in 1 projects deeply toward
the S3 pocket and forms a van der Waals contact with Ala335,
explaining why the fluoro 16 impaired potency relative to 1. The
fluorine atom on the phenyl ring also has a van der Waals contact
with the side chain of Tyr71. The corresponding nonfluoro
analogue showed reduced BACE1 potency (IC50 = 34 nM vs 9.8
nM; data not shown), confirming the importance of the
interaction.
0.50 μM using a manual patch cramp assay; 1 inhibited BACE2
(IC50 = 5.6 nM) with comparable potency as for BACE1,
whereas it was extremely selective over CatD and CatE (IC50
=
>100 μM). Compound 1 was not mutagenic in the Ames test.
Metabolic stability in human, rat, mouse, and dog hepatocytes
was also moderate-to-good (77, 53, 37, and 81% remaining after
120 min incubation), whereas 1 was significantly metabolized in
monkey hepatocytes (4% remaining). The unbound fraction in
serum of 1 was measured to be 0.060, 0.11, 0.083, 0.077, 0.056,
and 0.144 in human, rat, mouse, dog, monkey, and guinea pig.
The ability of 1 to inhibit human CYP isoforms 1A2, 2C9, 2C19,
2D6, and 3A4 was found to be moderate-to-low; compound 1
showed IC50 values of 10, 21, >100, 15, and 33 μM, respectively.
However, a shift in IC50 following 30 min preincubation of 1 was
observed (IC50 = <10 μM), indicating that 1 may cause time-
dependent inhibition (TDI) of CYP3A4, raising concern about
clinical drug−drug interactions (DDIs). In addition, the TDI
potential observed in 1 might result in electrophilic reactive
metabolite (RM) formation,26 which can deplete the
endogenous antioxidant glutathione (GSH) and/or bind
covalently to liver proteins, leading to drug-induced liver injury
(DILI). The covalent binding of the liver proteins can also
trigger an immune response via hapten formation, which is a
causal factor of idiosyncratic adverse drug reactions (IADRs).26
Therefore, to quantitatively estimate the potential of RM
formation, we assessed the covalent binding capability of 1
following PK/PD characterization and cardiovascular safety
study.
PRECLINICAL CHARACTERIZATION
PK/PD Study. Single oral administration of 1 to rat caused
dose-dependent plasma, cortical, and CSF Aβ40 reductions at
doses of 0.1, 0.3, 1, 3, and 10 mg/kg. At 1 mg/kg, reduction in
Aβ40 levels reached approximately 60% in CSF, and significant
Aβ reduction was maintained for 10 h (Figure 3), where a CSF-
to-unbound plasma ratio (Kp,uu,CSF) reached 0.88. The PK/PD
profile of 1 was also evaluated in dog. Repeated oral
administration of 1 for 14 days dose-dependently reduced
CSF Aβ40 and Aβ42 and maintained 50% Aβ reduction for 24 h at
■
In Vitro Profiles. Prior to in vivo preclinical studies, in vitro
pharmacological, physicochemical, and ADME profiles were
characterized. Compound 1 showed comparable potency for
mouse and dog BACE1 with IC50 values of 9.7 and 12 nM,
respectively. Compound 1 was a crystalline material that showed
good aqueous solubility of >1000 and 5.6 μg/mL at pH 1.2 and
6.8, respectively. Its LogD at pH 7.4 was 2.4. Regarding the
safety pharmacology, its hERG IC50 value was determined to be
1877
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Figure 4. Effects of compound 1 on CSF Aβ40 levels after repeated oral administration in TOYO Beagle dogs (n = 3, vehicle; n = 3, each dose).
Compound 1 was administered orally, and Aβ40 amounts in the CSF were measured using an enzyme-linked immunosorbent assay (ELISA) system at
the points of 12 and 24 h after administration at Day 1 (A) and Day 14 (B). Data shown are mean standard error of mean (SEM) levels expressed in
percent from the predose level for 3 dogs in each group. **p < 0.01 versus vehicle group (Dunnett’s test).
Figure 5. Aβ plaque quantification in the brain after repeated administration for 168 days in PS/APP mice (B6;SJL-TgN(APPSWE) × PS1I213T KI; n =
35, vehicle (0.5% MC); n = 15, untreated control for Aβ measurement; n = 35, each dose). Compound 1 was administered orally for 168 days, and the
number of plaques and the plaque area in the cortex (A and B) and hippocampus (C and D) were measured by immunostaining using antiAβ antibody
24 to 32 h after the last administration. Untreated: control mice before administration (24 to 30 weeks old of PS/APP mice). Data shown are mean
standard error (SE) for mice in each group. **p < 0.01 versus 0 mg/kg, once daily (vehicle) group (Dunnett’s test).
Table 3. PK Profiles of 1 in Rats, Dogs, and Monkeys
a
b
iv
po
dose (mg/kg)
CL (mL/min/kg)
Vdss (L/kg)
t1/2 (h)
dose (mg/kg)
AUC (ng·h/mL)
Cmax (ng/mL)
F (%)
c
rat/Sprague−Dawley
dog/beagle
0.3
0.3
0.3
22.8
6.02
16.3
4.46
3.13
2.87
5.52
8.48
2.72
1
1
0.3
386
2610
337
73.9
223
30.7
55
87
32
c
d
monkey/cynomolgus
a
b
Dosed as a solution of DMA/20% HP-β-CD in 100 mmol/L citrate buffer solution (1:9, v/v). Dosed as a suspension of 0.5% (w/v) MC aqueous
c
d
solution. n = 4. n = 3.
doses above 0.3 mg/kg (Figure 4). Plasma levels to lower CSF
Aβ40 or Aβ42 by 50% (EC50) were 22 and 24 ng/mL, respectively
(unbound EC50 = 1.9 and 2.0 ng/mL, respectively). No
significant difference in CSF Aβ reductions and compound
levels (plasma and CSF) between day 1 and day 14 was observed
(see Supporting Information, Table S1). Reflecting the good
brain distribution observed in rat, compound 1 showed high
CNS penetration in dog (Kp,uu,CSF = 0.98).
Amyloid Plaque Formation. Having established the PK/
PD relationship for compound 1, the inhibitory effect of 1 on
1878
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
plaque formation using PS/APP mouse was investigated. Male
PS/APP mice (24 to 30 weeks old at the initiation of dosing)
were given 15 mg/kg bid or 30 mg/kg qd for 168 days, where 1
caused significant inhibition of the number of Aβ plaques and Aβ
plaque load in the cortex and hippocampus after repeated dosing
compared with vehicle treatment (Figure 5). Taken together,
these data demonstrated robust Aβ reduction at low plasma
concentrations in rodent and nonrodent species, leading to
inhibition of Aβ plaque formation with chromic administration
of 1.
PK Study in Multiple Species and Human Dose
Projection. Pharmacokinetic parameters of 1 were determined
in rat, dog, and monkey (Table 3). Reflecting the metabolic
stability in hepatocytes, compound 1 showed good to excellent
oral bioavailability and low to moderate total clearance in rat and
dog, whereas in monkey, moderate-to-high clearance and
moderate oral availability were observed. The volume of
distributions at steady state (Vdss) were relatively unchanged
across the species and showed high values of ∼3.2, indicating
that 1 can be highly permeable to tissues. Using the PK
parameters, allometric scaling along with dog PK/PD data was
used for human dose and PK/PD projections. Assuming an oral
bioavailability of 50% and absorption rate constant (ka) of 0.5,
the target trough level in plasma was set to the dog EC50 value,
which led to a once daily projected dose of 30 mg with estimated
amide hydrolysis could be underestimated in microsomes,
which is a limitation of this study.
Cardiovascular Safety Study. As discussed above,
compound 1 inhibited the hERG channel with an IC50 value
of 0.50 μM in a manual patch clamp assay. Therefore, the
potential of cardiovascular safety associated with hERG
inhibition was evaluated in guinea pig and dog. In urethane-
anesthetized guinea pigs (n = 4), compound 1 was administered
intravenously at doses of 1, 3, and 10 mg/kg over a period of 10
at 30 min intervals, with no notable effect on QTc prolongation
(less than 10%) being observed at doses of up to 10 mg/kg.
Mean Cmax values at 1, 3, and 10 mg/kg were measured to be
264, 1180, and 4680 ng/mL, respectively. Based on the no-
observed-adverse-effect level (NOAEL) of the Cmax at 10 mg/kg
(1783 nM free) and the human projected Cmax (168 nM, 9.42
nM free), the safety margin in this study was calculated to be
189-fold. In conscious dogs (n = 3), 1 was administered orally at
doses of 3, 10, and 100 mg/kg. Although no significant QTc
prolongation was seen at 3 and 10 mg/kg, compound 1
prolonged QTc up to 15.1% relative to vehicle at 100 mg/kg,
with the Cmax values being 438, 1287, and 3467 ng/mL at 3, 10,
and 100 mg/kg, respectively. The safely margin over the
NOAEL (595 nM free) at 10 mg/kg was determined to be 63-
fold. With the favorable cardiovascular safety profile observed in
these studies, compound 1 progressed to tolerability studies.
Repeated Dose Toxicology Study in Rat and Dog. Prior
to the repeated dose studies, oral single dose studies were
performed in rats and dogs. In rats, no adverse effects were
observed at 0, 20, 60, and 200 mg/kg, whereas significant
findings, such as salivation and hypothermia, were observed at
2000 mg/kg. In single dose studies in dogs, compound 1 was
given at 0, 30, 100, and 300 mg/kg, with tonic conversion being
observed at 300 mg/kg. Based on the adverse effects observed in
the single dose studies, doses for repeated dose studies in rats
and dogs for a period of 1-month were set to 0. 20, 60, and 200
mg/kg/day in rats (n = 10 animals/sex/group) and 0, 3, 10, 100
mg/kg/day in dogs (n = 4 animals/sex/group). In the 1-month
study in rats, an NOAEL of 20 mg/kg/day was established, with
AUCs in males and females being 10.6 and 20.6 μg.h/mL,
respectively. At 200 mg/kg, there were several clinical pathology
changes including decrease in red blood cell parameters and
increase in cholesterol and liver enzymes of alanine amino-
transferase (ALT) and aspartate aminotransferase (AST). The
mean safety margin of males and females over the human
projected AUC (0.672 μg·h/mL) was considered to be 23-fold.
In the 1-month dog study, short-lasting clonic convulsions and
QTc prolongation were pronounced at 100 mg/kg. In addition,
slight decreases in red blood cell and serum parameters, such as
ALT, alkaline phosphatase (ALP), and cholesterol, were seen
without any histological correlates. The NOAEL was
determined to be 10 mg/kg/day, with the AUC being 18.4
and 11.5 μg·h/mL in males and females, respectively, and the
mean safety margin of both sexes being 22.3-fold. On the basis of
the safety profiles observed in the tolerability study as well as
other nonclinical results, we selected 1 (JNJ-54861911,
atabecestat) as our candidate for clinical development.
C
max and AUC of 61.8 ng/mL and 672 ng·h/mL, respectively.
Covalent Binding. To estimate the potential for bio-
activation of 1, metabolism-dependent covalent binding (CVB)
was determined (Table 4). The CVB study was carried out with
Table 4. Covalent Bond Binding Values of [14C]-1 and [3H]-
Pioglitazone in Human Liver Microsomes
total
covalent
bond
NADPH
independent
covalent bond
NADPH
dependent
covalent bond
binding
concentration (pmol/mg binding (pmol/ binding (pmol/
compound
(μmol/L)
protein)
mg protein)
mg protein)
[14C]-1
10
50
10
50
425
3804
599
5
24
4
419
3780
594
[3H]-
pioglitazone
3458
12
3446
human liver microsomes at concentrations of 10 and 50 μM of 1.
Radiolabeled compound 1 with a 14C atom on the amidine
carbon was synthesized to measure the CVB levels of 1 based on
the total amount of radio activity (synthesis of [14C]-1; see
Supporting Information). The use of CVB evaluation has been
well-discussed in several articles.26 For comparison, pioglitazone
was included as a control with a low concern of DILI risks.
Although pioglitazone was used as a negative control given its
rare frequency of DILI, it was indeed associated with elevations
of liver enzymes in the clinic.26d As shown in Table 4, compound
1 was found to form RM(s) and showed metabolite NADPH-
dependent CVB values of 419 and 3780 pmol/mg protein at 10
and 50 μM, respectively. Importantly, no significant difference in
CVB burden between 1 and pioglitazone was observed,
indicating that the likelihood of DILI risks might be of the
same level as pioglitazone. However, primary pathways of
metabolism observed in human hepatocytes and in vivo were
found to be hydrolysis of the amide bond to form the aniline and
GSH addition. Based on the current understanding of the
metabolic pathways, microsomal incubations may not be
relevant to estimate the human situations; GSH addition and
SUMMARY OF CLINICAL STUDIES
■
A series of Phase 1 clinical studies to assess PK, safety, and PD
effects on Aβ production was conducted with healthy
participants. In single (SAD) and multiple ascending dose
(MAD) studies, 1 was well-tolerated at doses of up to 150 mg/
day; 1 was given once per day for 14 days in the MAD study.
1879
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Dose-dependent reductions of CSF Aβ40 were observed,
reaching approximately 50, 80−85, 90, and 90−95% reductions
at 5, 30, 50, and 90 mg/day, respectively, on day 14. The mean
over CatD, leading to thiazine-based inhibitors 1, 3, and 7.7a,b
Although the thiazines were found to be potent in humans, the
substructure has the potential to form reactive metabolites,
which may have led to the significant elevation of liver enzymes
observed with compounds 1 and 3, consequently resulting in
their termination.8d,10 In addition to the structural issue, initial
BACE1 inhibitors containing an amidine moiety have high
basicity along with a certain level of lipophilicity, which is likely
to be associated with liabilities on P-gp recognition, CYP2D6
inhibition, and phospholipidosis. Significant medicinal chem-
istry efforts were able to balance basicity and lipophilicity to
mitigate the issues and culminate in compounds 5 and 8, except
for suboptimal selectivity over BACE2.7a,b
C
max values for 5, 30, and 50 mg/day on day 14 were 38.7, 202,
and 364 ng/mL, respectively. Also, the mean AUC values were
548, 2897, and 5509 ng·h/mL, respectively.8a Consistent with
the animal data, compound 1 showed high CNS penetration
(Kp,uu,CSF = 0.69, measured at 30 mg/day on day 14 based on the
AUC values).
A clinical DDI study was performed to evaluate the
interaction potential of 1 as a perpetrator when coadministered
with a drug cocktail of CYP3A4, CYP2C9, and CYP1A2
substrates, i.e. midazolam, caffeine, and tolbutamide. No
significant changes in PK parameters for the drug cocktail with
and without 1 were observed following single-dose admin-
istration, indicating no clinically significant effects on the activity
of the CYP enzymes. The DDI effects observed in the
nonclinical study were not reflected in the clinic.8c
Disappointingly, both compounds 5 and 8 as well as 1
progressed into Phase 3 clinical trials but led to a cognitive
deficit in prodromal and preclinical AD patients,8c,12e,15b
although compound 6 did not lead to cognitive decline in
MCI and mild AD patients.13b Neuropsychiatric adverse events
were evident in the treatment groups compared with the placebo
among the BACE1 inhibitors.8c,12e Such evidence hints at the
involvement of on-target mechanisms. Although initial inves-
tigations of BACE1 KO mice did not disclose major phenotypes,
several mild neuronal phenotypes, such as reduced myelination,
synaptic deficits, impaired memory, and axonal targeting errors,
were revealed in more in-depth studies. BACE1 is found to have
numerous substrates and thereby plays an important role in a
number of functions in the brain.16,28 Indeed, the undesirable
phenotypes observed in BACE1 KO mice, such as hypomyeli-
nation, deficient synaptic plasticity, or axon guidance
abnormalities, can be explained by loss of the function of
BACE1 substrates such as neuregulin 1 (NRG1), seizure protein
6 (SEZ6), or a close homologue of the neural cell adhesion
molecule L1 (CHL1), respectively.16,28 Therefore, the key to
restart development of a BACE1 inhibitor is to identify the level
of BACE1 inhibition that can discriminate achieving efficacy
from causing cognitive worsening. We believe that progress in
biomarkers to measure levels of Aβ and p-tau181/p-tau217 in
the plasma could pave the way to correctly estimate the timing of
the intervention through the development of Aβ antibodies such
as aducanumab or BAN2410. However, considering the
cognitive worsening observed with the BACE1 inhibitors,
biomarkers capable of early detection of cognitive symptoms
or any other neuronal functional changes are required to ensure
further development from the ethical standpoint. An alternative
approach would be the development of a BACE1 inhibitor with
substrate selectivity for APP versus the nondesired BACE1
substrate that causes the cognitive worsening.
The BACE1 inhibitors exhibiting cognitive decline lack
selectivity over BACE2.7f,16 Although compound 8 did not
show hypopigmentation in preclinical models or significant hair
color changes in humans, it has only 3-fold selectivity over
BACE2.15b Regarding the selectivity of 1 and 5, they were
slightly BACE2 selective given their IC50 ratios.12a,b The
physiological functions of BACE2 are much less known than
those of BACE1. Unlike BACE1, BACE2 is highly expressed in
peripheral tissues, such as the pancreas, skin, or eye, but is also
present at low levels in the brain.17 Early investigations identified
peripheral BACE2 substrates, such as the pigment cell-specific
melanocyte protein (PMEL),29 the pro-proliferative type I
transmembrane protein 27 (TMEM27),30 and islet amyloid
peptide (IAPP).31 BACE2 processes PMEL, which is expressed
in the pigment cells of the eye and skin to generate functional
amyloids and consequently cause melanogenesis. Indeed,
A thorough QT study was performed with healthy
participants given a therapeutic dose of 50 mg/day and a
supratherapeutic dose of 150 mg/day. Compound 1 showed
clinically relevant QTc prolongation at 150 mg/kg on Day 7,
whereas the dose of 50 mg/day did not carry regulatory
concerns for QTc prolongation according to the guidelines. The
mean Cmax values for 50 and 150 mg/day following 7-day
administrations were 513 and 1380 ng/mL, respectively. The
unbound Cmax at 150 mg/day was calculated to be 0.21 μM,
which was at a similar level to the hERG IC50 value of 0.50 μM.8c
Prior to large phase 2b/3 global clinical development, a Phase
2a clinical study was started to evaluate PD activity of 1 in the
intended target population of patients diagnosed with early AD,
i.e., preclinical AD (Clinical Dementia Rating (CDR) = 0) or
mild cognitive impairment due to AD (CDR = 0.5) with a
positive CSF biomarker signature (reduced CSF amyloid
levels). The patients given a once-daily dose of 10 mg or 50
mg for 4 weeks achieved 60−70% and 90% CSF Aβ40 reductions
compared with the placebo, respectively.8b The Phase 2b/3
EARLY clinical trial enrolled 557 patients with preclinical AD
randomized to 5 mg or 25 mg 1 or placebo. Significant increase
in ALS and AST (>3 times the upper limit of the normal level
(ULN)) was observed in 11% of patients treated with 1,
compared to 0.5% of patients receiving the placebo; one patient
receiving 25 mg of 1 met the criteria of Hy’s Law (ALT more
than 3 times ULN and total bilirubin more than 2 times ULN),
which led to discontinuation of further development of 1.8d The
EARLY trial evaluated the efficacy of 1 using the Preclinical
Alzheimer’s Cognitive Composite (PACC) score as the primary
end point and the Repeatable Battery for the Assessment of
Neuropsychological Battery (RBANS) score as a secondary end
point. At 6 and 12 months, worsening of cognitive performance
was evident at doses of 5 and 25 mg compared to the placebo.
Decline in the RBANS total score at 3 months was also observed
for the treatment group over the placebo. In addition to the
cognitive worsening, neuropsychiatric adverse events, such as
depression, sleep/dreaming, and anxiety, were confirmed in the
treatment group over the placebo group levels.8c
FUTURE DIRECTIONS
■
To date, all BACE1 inhibitor clinical trials have been
discontinued.7f,16 Initial clinical-stage BACE1 inhibitors such
as 2 and AMG-8718 could not achieve selectivity over CatD,
resulting in retinal toxicity.7b,9,27 Extending substituents into the
S3 pocket using amide groups addressed the issue of selectivity
1880
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
a
Scheme 1. Synthesis of 1 and Its Analogues 15−18
a
Reagents and conditions: (a) (i-Pr)2NH, n-BuLi, t-BuOAc, Ti(Oi-Pr)3Cl, THF, −78 °C, 77%; (b) HCl, MeOH, dioxane, rt: (c) BH3, 0 °C to rt,
then HCl, EtOAc, i-Pr2O, rt; (d) NaHCO3, Fmoc N-hydroxysuccinimide ester, THF/H2O; (e) IBX, DMSO, rt; (f) CH(OMe)3, Amberlite,
MeOH, reflux, then piperidine, DMF, rt, 54% in 3 steps; (g) BzNCS, acetone 0 °C then H2SO4, 40 °C, 78%; (h) Fe, AcOH/H2O, 50 °C, 92%; (i)
RCO2H, DMT-MM, MeOH, rt; or RCO2H, HCl, EDC·HCl, MeOH, 0 °C, 16−88%.
a
Scheme 2. Synthesis of Thiazine 13
a
Reagents and conditions: (a) HCl, H2O, rt; (b) BzNCS, DCM, 0 °C to rt, 95% in 2 steps; (c) I2, DCM, rt; (d) pyrrolidine, THF, 60 °C, 80% in 2
steps; (e) hydrazine hydrate, EtOH, rt, 74%; (f) 5-cyanopicolinic acid, DMT-MM, MeOH, rt, 46%.
BACE2 KO mice exhibited fur depigmentation, implying its
important role during melanogenesis.5,16 As for TMEM27 and
IAPP, they are responsible for glucose homeostasis. Inhibition of
processing of TMEM27 in pancreatic β-cells using BACE2 KO
mice led to increased β-cell production. Both IAPP and insulin
are secreted into the blood, where IAPP has a protective role
against glucose metabolism.31 Recent reports highlighted certain
roles of BACE2 in the brain.17,32 The authors pinpointed the
expression of BACE2 in neurons and glia and identified several
BACE2 substrates in the brain. Of these, the vascular cell
adhesion molecule 1 (VCAM1) was found to play a role in
neuroinflammation, which is one of the hallmarks of AD, as
inflammation stimuli were found to increase VCAM expres-
sion.17 Another report points to BACE2 as a protective enzyme
against amyloidosis.32 People with Down’s syndrome have
triplication of chromosome 21, and 70% develop AD during
their lifetime, whereas 100% of non- Down’s syndrome people
inheriting triplication of APP display AD-like symptoms. The
BACE2 gene also sits on chromosome 21 and is known to have
θ-secretase activity, preventing formation of Aβ and serving as an
Aβ-degrading protease. The authors indicated that BACE2
could contribute to the delayed onset in 30% of individuals with
Down’s syndrome and serve as an AD suppressor gene.32 The
accumulated evidence for BACE2 indicates that chronic
inhibition of BACE2 could cause unwanted side effects and
thereby warrant pursuing BACE1 selective inhibitors. However,
although the above findings suggest the potential involvement of
cognitive function derived from BACE2, there is no direct
evidence linking BACE2 inhibition to the cognitive worsening
that has backed BACE1 inhibitors into a corner. Taken together,
elucidating the biological functions of BACE1, along with
development of biomarkers to measure the unwanted side
effects derived from BACE1 inhibition, are required to identify
the levels at which a given BACE1 inhibitor can achieve efficacy
1881
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
a
Scheme 3. Synthesis of Dihydrothiazine 14
a
Reagents and conditions: (a) LDA, t-BuOAc, Ti(Oi-Pr)3Cl, THF, −78 °C; (b) HCl, dioxane, 60 °C; (c) BH3, THF, 45 °C, 75% in 3 steps; (d)
BzNCS, DCM, 0 °C to rt; (e) Ghosez’s reagent, DCM, 0 °C to rt, 81% in 2 steps; (f) aq. H2SO4, 50 °C, 71%; (g) H2SO4, HNO3, 98%; (h) Fe,
HCl, MeOH, 50 °C; (i) 5-cyanopicolinic acid, HCl, WSCD·HCl, MeOH, rt, 53% in 2 steps.
but avoid cognitive worsening. It also remains a critical challenge
to elucidate BACE2 biology in the brain as well as its
involvement in cognitive function. We believe that accumulation
of such evidence can lead to restarting the development of next-
generation BACE1 inhibitors with high selectivity over BACE2.
followed by condensation with 5-cyanopicolinic acid afforded
the desired thiazine 14 possessing a fluoromethyl group.
CONCLUSIONS
■
Starting with the lead compound 12, we designed and
synthesized thiazine head groups that undergo key interactions
with the catalytic aspartic acid residues through leveraging the
pKa lowering approach, and the subsequent medicinal chemistry
efforts culminated in the discovery of compound 1, which
mitigated high P-gp efflux and consequently demonstrated
robust central Aβ reduction in preclinical models. Compound 1
was a potent inhibitor of BACE2 but completely avoided
inhibition of CatD. Although 1 inhibited hERG at submicro-
molar concentrations, its promising potency ensured a safety
margin in nonrodent cardiovascular models. Time-dependent
inhibition (TDI) of CYP3A4 and covalent binding to human
hepatocytes using radiolabeled 1 were confirmed, implying that
1 has the potential to form reactive metabolites. Human dose
projection along with the CVB study in human microsomes
indicated that the likelihood of DILI risks might be at the same
level as pioglitazone. Compound 1 demonstrated robust and
dose-dependent inhibition of Aβ production in humans and
advanced to the EARLY Phase 2b/3 clinical trial. Unfortunately,
significant elevation of liver enzymes halted further development
of compound 1, and subsequent analysis of clinical efficacy
confirmed dose-dependent cognitive worsening.
CHEMISTRY
■
The medicinal chemistry routes of synthesizing 1 and its
analogues possessing different tail groups are outlined in Scheme
1. Well-established diastereoselective alkylation to an Ellman’s
imine of 2118a enabled construction of the tert-amine 22 having
the desired chiral stereocenter. The amine 22 was treated with
HCl to remove both t-butyl sulfinyl and t-butyl groups, followed
by borane reduction to afford the amino alcohol 23. After 9-
fluorenylmethoxycarbonyl (Fmoc) protection of the amino
group, the primary alcohol was oxidized with 2-iodoxybenzoic
acid (IBX), and the resulting aldehyde was converted to
dimethylacetal, which was subsequently treated with piperidine
to remove the Fmoc group to afford 24. The obtained amine was
reacted with benzoyl isothiocyanate to form N-hydroxypropyl
thiourea followed by cyclization under acidic conditions to give
25. After reduction of the nitro group using iron, the aniline 26
was successfully condensed with the corresponding carboxylic
acids using 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmor-
pholinium chloride (DMT-MM)33 or 1-(3-dimehylaminoprop-
yl)-3-ethylcarbodiimide hydrochloride (EDC·HCl) with 1
equiv of HCl to furnish the targeting thiazine 1 and its analogues
15−18.
EXPERIMENTAL SECTION
■
The thiazine 13, which has an exocyclic alkene moiety, was
synthesized by a procedure outlined in Scheme 2. The thiourea
28 was prepared from 2718a by removal of the sulfinyl group,
followed by thiourea formation. The thiazine ring possessing
exo-olefin 29 was constructed by iodine-mediated cyclization
and the telescoped elimination reaction. After global deacylation
by hydrazine, the resulting aniline 30 was condensed with 5-
cyanopicolinic acid to afford 13.
Synthesis of 4-fluoromethyl dihydrothiazine 14 commenced
with the addition of tert-butyl acetate to the known Ellman’s
imine 31 (Scheme 3).34 The resulting tert-butyl ester was
converted to a carboxylic acid and reduced to the primary
alcohol 32 by borane. The thiazine cyclization was accomplished
by Ghosez’s reagent35 after benzoylthiourea formation to give
33. Debenzoylation, regioselective nitration, and reduction
General Chemistry. All commercial solvents and reagents were
used without further purification. Column chromatography was done
on a purification system using prepacked silica gel columns (Fuji Silysia,
1
Biotage, or Yamazen). H and 13C NMR spectra were recorded on a
Bruker Avance system operating at 400 and 100 MHz, respectively.
Chemical shifts are reported as ppm, and tetramethylsilane was used as
an internal reference. For peak multiplicities, the following abbrevia-
tions are used: singlet (s), doublet (d), double doublets (dd), double
double doublet (ddd), double triplet (dt), triplet (t), quartet (q),
multiplet (m), and broad peak (br). The purity of the compounds used
in Tables 1 and 2 was greater than 95% using the following analysis
method: analytical LC/MS was done on a Shimadzu Shim-pack XR-
ODS (C18, 2.2 μm, 3.0 × 50 mm; a linear gradient from 10% to 100% B
over 3 min and then 100% B for 1 min (A = 0.1% formic acid in water, B
= 0.1% formic acid in MeCN); flow rate = 1.6 mL/min) using a
Shimadzu UFLC system equipped with a LCMS-2020 mass
1882
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
spectrometer, SPD-M20A photodiode array detector (detection at 254
nm), LC-20AD binary gradient module, and SIL-20AC sample
manager. High resolution mass spectra (HRMS) were determined
with a Thermo Fisher Scientific LTQ Orbitrap.
(12 mL) was added benzoyl isothiocyanate (1.70 mL, 12.6 mmol) at 0
°C. After stirring for 30 min, conc. H2SO4 (13 mL) was added, and the
solution was heated to 40 °C overnight. To the reaction mixture were
added ice (40 g), THF (10 mL), and conc. aqueous ammonia (20 mL).
The aqueous layer was extracted with EtOAc, and the organic layer was
washed with brine, dried over MgSO4, and concentrated under reduced
pressure. The crude product was purified by silica gel column
chromatography (n-hexane/EtOAc, 20−50% EtOAc) to afford 25
(2.59 g, 9.69 mmol, 78%) as a yellow foam. 1H NMR (CDCl3) δ 1.70
(3H, s), 6.24 (1H, dd, J = 9.5, 4.6 Hz), 6.34 (1H, d, J = 9.5 Hz), 7.16
(1H, dd, J = 10.5, 8.8 Hz), 8.13 (1H, ddd, J = 8.8, 4.1, 3.0 Hz), 8.43 (1H,
dd, J = 6.8, 3.0 Hz). MS-ESI (m/z): 268 [M + H]+.
(S)-4-(5-Amino-2-fluorophenyl)-4-methyl-5,6-dihydro-4H-
1,3-thiazin-2-amine (26). To a stirred solution of 26 (2.59 g, 9.69
mmol) in AcOH/water (13 mL/1.3 mL) was added iron (2.16 g, 38.8
mmol). After stirring for 1 h at 50 °C, to the reaction mixture were
added conc. aq. NH3 and EtOAc, and the insoluble materials were
filtered off. The aqueous layer was extracted with EtOAc, and the
organic layer was washed with brine, dried over MgSO4, and
concentrated under reduced pressure to afford 26 (2.12 g, 8.93
mmol, 92%) as a yellow solid. 1H NMR (CDCl3) δ 1.67 (3H, s), 3.53
(2H, br s), 6.21−6.30 (2H, m), 6.50 (1H, dd, J = 7.5, 4.3 Hz), 6.74−
6.85 (2H, m). MS-ESI (m/z): 238 [M + H]+. Anal. Calcd for
C11H12FN3S; C, 55.68; H, 5.10; F, 8.01; N, 17.71; S, 13.51. Found: C,
55.36; H, 5.03; N, 17.89; F, 8.00; S,13.40.
(S)-N-(3-(2-Amino-4-methyl-4H-1,3-thiazin-4-yl)-4-fluoro-
phenyl)-5-cyanopicolinamide (1). To a stirred solution of 26 (147
mg, 0.619 mmol) and 5-cyanopicolinic acid (94.0 mg, 0.623 mmol) in
MeOH (5.0 mL) was added 2 M HCl (0.310 mmol, 0.619 mmol) and
cooled to 0 °C. To the solution, EDC·HCl (125 mg, 0.650 mmol) was
added. After stirring for 1 h at the same temperature, 2 M aq. NaOH
(0.31 mL) and 5% aq. K2CO3 were added, and the resulting solid was
filtered and dried to afford 1 (200 mg, 0.544 mmol, 88%) as a white
solid. 1H NMR (DMSO-d6) δ 1.56 (3H, s), 6.16 (1H, dd, J = 9.5, 4.3
Hz), 6.32 (2H, s), 6.48 (1H, d, J = 9.5 Hz), 7.14 (1H, t, J = 10.2 Hz),
7.74−7.78 (1H, m), 7.95−7.92 (1H, m), 8.28 (1H, d, J = 8.3 Hz), 8.58
(1H, d, J = 8.3 Hz), 9.19 (1H, s), 10.77 (1H, s). 13C NMR (DMSO-d6)
δ 28.8 (d, J = 2.9 Hz), 58.9 (d, J = 3.7 Hz), 111.5, 116.1 (d, J = 24.9 Hz),
116.6, 116.7, 120.8 (d, J = 8.8 Hz), 121.0 (d, J = 5.1 Hz), 122. 3, 127.5
(d, J = 2.9 Hz), 133.4 (d, J = 2.9 Hz), 134.8 (d, J = 12.5 Hz), 142.1,
148.1, 151.4, 152.5, 156.8 (d, J = 243.5 Hz), 161.1. MS-ESI (m/z): 368
[M + H]+. HRMS-ESI (m/z): [M + H]+ calcd for C18H15FN5OS+,
368.0976; found, 368.0971. Anal. Calcd for C18H14FN5OS; C, 58.85;
H, 3.84; N, 19.06; F, 5.17; S, 8.73. Found: C, 58.66; H, 3.99; N, 18.84;
F, 5.20; S, 8.67.
(S)-N-(3-(2-Amino-4-methyl-4H-1,3-thiazin-4-yl)-4-fluoro-
phenyl)-3-chloro-5-cyanopicolinamide (15). Compound 15 was
prepared in a manner similar to that for 1 from 26 (165 mg, 0.696
mmol) using 2-chloro-5-cyanopicolinic acid (152 mg, 0.835 mmol) in
85% yield (240 mg, 0.594 mmol). 1H NMR (DMSO-d6) δ 1.54 (3H, s),
6.15 (1H, dd, J = 9.5, 4.7 Hz), 6.31 (2H, br s), 6.45 (1H, d, J = 9.5 Hz),
7.14 (1H, dd, J = 11.4, 8.7 Hz), 7.62−7.70 (2H, m), 8.76 (1H, d, J = 1.7
Hz), 9.06 (1H, d, J = 1.7 Hz), 10.84 (1H, br s). 13C NMR (CDCl3) δ
28.6 (d, J = 3.6 Hz), 59.0 (d, J = 4.4 Hz), 112.7, 114.3, 115. Seven (d, J =
1.4 Hz), 116.9 (d, J = 25.6 Hz), 119.7 (d, J = 3.7 Hz), 119.9 (d, J = 8.8
Hz), 128.7 (d, J = 3 7 Hz), 132.7 (d, J = 2.2 Hz), 132.8, 134.8 (d, J = 13.2
Hz), 143.8, 147.8, 148.5, 150.3, 156.9 (d, J = 245.8), 159.1. MS-ESI (m/
z): 402 [M + H]+. HRMS-ESI (m/z): [M + H]+ calcd for
C18H14ClFN5OS+, 402.0586; found, 402.0586.
tert-Butyl (S)-3-(((R)-tert-Butylsulfinyl)amino)-3-(2-fluoro-5-
nitrophenyl)butanoate (22). To a stirred solution of diisopropyl-
amine (13.1 mL, 21.0 mmol) in THF (8.0 mL) was added n-
butyllithium (1.6 M in n-hexane, 3.26 mL, 27.9 mmol) at −78 °C. After
stirring for 30 min at 0 °C, tert-butyl acetate (3.26 mL, 24.5 mmol) in
THF (8.0 mL) was added at −78 °C. After stirring for 30 min at the
same temperature, Ti(Oi-Pr)3Cl (7.03 mL, 27.9 mmol) in THF (8.0
mL) was added and stirred for 30 min at the same temperature. To the
solution, 2118a (2.00 g, 6.99 mmol) in THF (8.0 mL) was added at −78
°C. After stirring for 2 h, the reaction mixture was poured into aq.
NH4Cl, and volatile solvents were evaporated. Water was added to the
residue, and the aqueous layer was extracted with EtOAc. The organic
layer was washed with water and brine, then dried over Na2SO4. The
crude product was purified by silica gel column chromatography (n-
hexane/EtOAc) to afford 22 (2.16 g, 5.37 mmol, 77%) as a yellow oil.
1H NMR (CDCl3) δ 1.30 (9H, s), 1.86 (3H, s), 3.11 (1H, dd, J = 16.2,
2.1 Hz), 3.26 (1H, dd, J = 16.2, 2.1 Hz), 5.55 (1H, s), 7.18 (1H, dd, J =
11.1, 8.9 Hz), 8.18 (1H, ddd, J = 8.9, 4.0, 2.8 Hz), 8.53 (1H, dd, J = 7.0,
2.8 Hz). MS-ESI (m/z): 403 [M + H]+.
(S)-3-Amino-3-(2-fluoro-5-nitrophenyl)butan-1-ol hydro-
chloride (23). To 22 (14.1 g, 34.9 mmol) were added HCl (4 M in
dioxane, 52 mL) and MeOH (1.7 mL, 41.9 mmol), and the solution was
stirred overnight. To the reaction mixture was added ether (150 mL),
and the resulting precipitate was filtered, washed with ether, and then
dried. To a stirred solution of BH3 (1.0 M in THF 92 mL, 92 mmol), the
residue was added portionwise at 0 °C. After stirring at rt overnight, the
reaction mixture was poured into ice water, and aq 1 M NaOH was
added to basify the reaction medium. The aqueous layer was extracted
with EtOAc, and the organic layer was washed with brine, dried over
MgSO4, and then dried under reduced pressure. To the resulting brown
oil were added diisopropyl ether and HCl (4 M in EtOAc, 8 mL) to give
a precipitate, which was filtered and dried to afford 23 (6.12 g, 23.1
1
mmol, 75%) as a brown solid. H NMR (DMSO-d6) δ 1.78 (3H, s),
2.21−2.34 (2H, m), 3.27−3.50 (3H, m), 7.64 (1H, dd, J = 11.6, 9.1
Hz), 8.36−8.48 (2H, m), 9.06 (3H, br s). MS-ESI (m/z): 229 [M +
H]+.
(S)-2-(2-Fluoro-5-nitrophenyl)-4,4-dimethoxybutan-2-
amine (24). To a stirred solution of 23 (6.12 g, 23.1 mmol) in THF/
water (60 mL/30 mL) were added NaHCO3 (3.88 g, 46.2 mmol) and
Fmoc N-hydroxysuccinimide ester (7.80 g, 23.1 mmol) at room
temperature. After stirring for 2 h at the same temperature, the aqueous
layer was extracted with EtOAc. The organic layer was washed with
brine, dried over MgSO4, and concentrated under reduced pressure. To
the residue in DMSO (30 mL) was added 2-iodobenzoic acid (6.47 g,
23.1 mmol) at room temperature. After stirring overnight, water and
EtOAc were added, and then the insoluble materials were filtered off.
The organic layer was washed with brine, dried over MgSO4, and
concentrated under reduced pressure. To the residue in MeOH (100
mL) were added methyl orthoformate (0.128 mL, 1.16 mmol) and
Amberlite (2.0 g), and the solution was heated to reflux for 2.5 h.
Insoluble materials were filtered off, and the volatile compounds were
removed under reduced pressure. To the residue were added DMF (20
mL) and piperidine (2.75 mL, 27.7 mmol). After stirring for 40 min,
piperidine (0.69 mL, 6.95 mmol) was added, and the solution was
stirred for an additional 30 min. To the reaction mixture were added
AcOH and water. The aqueous layer was extracted with EtOAc, and the
organic layer was washed with water and brine, dried over MgSO4, and
concentrated under reduced pressure. The crude product was purified
by silica gel column chromatography (n-hexane/EtOAc, 50−80%
EtOAc) to afford 24 (3.37 g, 12.4 mmol, 54%) as a yellow oil. 1H NMR
(CDCl3) δ 1.57 (3H, s), 2.09 (1H, dd, J = 14.4, 6.8 Hz), 2.35−2.42 (1H,
m), 3.19 (6H, s), 4.13 (1H, dd, J = 6.8, 4.3 Hz), 7.16 (1H, dd, J = 10.8,
9.0 Hz), 8.13−8.18 (1H, m), 8.62 (1H, dd, J = 6.8, 2.8 Hz). MS-ESI
(m/z): 273 [M + H]+.
(S)-N-(3-(2-Amino-4-methyl-4H-1,3-thiazin-4-yl)-4-fluoro-
phenyl)-5-fluoropicolinamide (16). To a stirred solution of 26
(50.0 mg, 0.211 mmol) in MeOH (1.0 mL) were added 5-
fluoropicolinic acid (35.7 mg, 0.253 mmol) and DMT-MM (72.9 mg,
0.232 mmol). After stirring for 6 h at room temperature, additional
DMT-MM (6.6 mg, 0.02 mmol) was added, and the solution was
stirred for 3 h. To the reaction mixture was added aq. Na2CO3, and the
aqueous layer was extracted with EtOAc. The organic layer was washed
with brine, dried over MgSO4, and concentrated under reduced
pressure. To the crude product were added EtOAc and ether, and the
resulting solid was filtered and dried to afford 16 (31.9 mg, 89.0 μmol,
(S)-4-(2-Fluoro-5-nitrophenyl)-4-methyl-4H-1,3-thiazin-2-
amine (25). To a stirred solution of 24 (3.37 g, 12.4 mmol) in acetone
1883
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
42%) as a pale yellow solid. 1H NMR (DMSO-d6) δ 1.56 (3H, s), 6.16
(1H, dd, J = 9.6, 4.2 Hz), 6.30 (2H, s), 6.47 (1H, d, J = 9.6 Hz), 7.13
(1H, dd, J = 11.2, 9.2 Hz), 7.74 (1H, m), 7.90 (1H, m), 7.98 (1H, m),
8.22 (1H, m), 8.65 (1H, s), 10.54 (1H, s). 13C NMR (CDCl3) δ 28.8,
57.9, 116.0 (d, J = 24 Hz), 116.7, 120.5 (d, J = 8.1 Hz), 120.7 (d, J = 5.1
Hz), 124.6 (d, J = 5.9 Hz), 124.8 (d, J = 18.3), 127.5 (d, J = 2.9 Hz),
133.7 (d, J = 2.9 Hz), 135.8 (d, J = 12.8 Hz), 136.8 (d, J = 24.9 Hz),
146.6 (d, J = 3.7 Hz), 148.0, 156.1 (d, J = 242.8 Hz), 160.8 (d, J = 258.8
Hz), 161.4. HRMS-ESI (m/z): [M + H]+ calcd for C17H15F2N4OS+,
361.0929; found, 361.0924.
(S)-N-(3-(2-Amino-4-methyl-4H-1,3-thiazin-4-yl)-4-fluoro-
phenyl)-5-chloropicolinamide (17). Compound 17 was prepared in
a manner similar to that for 16 from 26 (94.8 mg, 0.400 mmol) using 5-
chloropicolinic acid (50.4 mg, 0.320 mmol) in 16% yield (24.1 mg,
0.0640 mmol) as a white solid. 1H NMR (DMSO-d6) δ 1.56 (3H, s),
6.16 (1H, dd, J = 9.5, 4.5 Hz), 6.31 (2H, s), 6.47 (1H, d, J = 9.7 Hz),
7.13 (1H, dd, J = 11.4, 8.9 Hz), 7.73−7.76 (1H, m), 7.91 (1H, dd, J =
7.2, 2.4 Hz), 8.21−8.13 (2H, m), 8.78 (1H, d, J = 1.8 Hz), 10.60 (1H,
s). 13C NMR (CDCl3) δ 28.7 (d, J = 2.2 Hz), 59.0 (d, J = 4.4 Hz), 115.7,
116.7 (d, J = 24.2 Hz), 119.5 (d, J = 4.4 Hz), 119.6 (d, J = 8.8 Hz),
123.4, 128.8 (d, J = 3.7 Hz), 133.1 (d, J = 3.7 Hz), 134.5 (d, J = 12.5
Hz), 135.2, 137.4, 147.0, 147.9, 150.2, 156.6 (d, J = 245.7 Hz), 161.0.
MS-ESI (m/z): 377 [M + H]+. HRMS-ESI (m/z): [M + H]+ calcd for
C17H15ClFN4OS+, 377.0634; found, 377.0631.
(S)-N-(3-(2-Amino-4-methyl-4H-1,3-thiazin-4-yl)-4-fluoro-
phenyl)-5-methoxypyrazine-2-carboxamide (18). Compound
18 was prepared in a manner similar to that for 16 from 26 (47.4 mg,
0.200 mmol) using 5-methoxypyrazine-2-carboxylic acid (30.8 mg,
0.200 mmol) in 65% yield (48.4 mg, 0.130 mmol) as a white solid. 1H
NMR (CDCl3) δ 1.71 (3H, s), 4.06 (3H, s), 6.28 (2H, d, J = 2.1 Hz),
7.04 (1H, dd, J = 11.2, 8.8 Hz), 7.60 (2H, dd, J = 6.8, 2.8 Hz), 7.83 (1H,
ddd, J = 8.8, 4.1, 2.8 Hz), 8.14 (1H, d, J = 1.2 Hz), 9.0 (1H, d, J = 1.2
Hz), 9.45 (1H, s). 13C NMR (CDCl3) δ 28.5 (d, J = 3.7 Hz), 54.4, 59.1
(d, J = 3.7 Hz), 115.6, 116.8 (d, J = 24.9 Hz), 119.4 (d, J = 4.4 Hz),
119.7 (d, J = 8.8 Hz), 128.8 (d, J = 3.7 Hz), 133.2 (d, J = 2.2 Hz), 133.3,
134.4, 137.3, 141.9, 149.8, 156.7 (d, J = 243.6 Hz), 161.0, 162.3. MS-
ESI (m/z): 374 [M + H]+. HRMS-ESI (m/z): [M + H]+ calcd for
C17H17FN5O2S+, 374.1082; found, 374.1077.
(S)-N-((2-(2-Fluoro-5-(2,2,2-trifluoroacetamido)phenyl)-
pent-4-en-2-yl)carbamothioyl)benzamide (28). Compound
2718a (3.99 g, 10.1 mmol) in 2 M aq. HCl (20 mL, 40 mmol) was
stirred at room temperature for 1 h. After concentration under reduced
pressure, the solution was diluted with EtOAc, and the organic layer was
extracted with water and 2 M aq. HCl. The combined aqueous layers
were basified with K2CO3 and extracted with EtOAc. The organic layer
was washed with brine, dried over Na2SO4, and concentrated. To the
residue in DCM (20 mL) was added at 0 °C benzoyl isothiocyanate
(1.43 mL, 10.6 mmol). After stirring for 5 min, the reaction mixture was
warmed to room temperature and stirred for 45 min. The reaction
mixture was concentrated under reduced pressure, and the crude
product was purified by silica gel column chromatography (n-hexane/
EtOAc, 10−50% EtOAc) to afford 28 (4.34 g, 9.57 mmol, 95% in 2
steps) as a white amorphous substance. 1H NMR (CDCl3) δ 2.00 (3H,
s), 2.73, (1H, dd, J = 13.1, 6.6), 3.18 (1H, dd, J = 13.1, 7.6), 5.24 (1H, d,
J = 10.6), 5.29 (1H, J = 17.2 Hz), 5.67−5.79 (1H, m), 6.99−7.07 (1H,
dd, J = 10.4, 10.4 Hz), 7.45−7.53 (3H, m), 7.57−7.65 (2H, m), 7.83
(2H, d, J = 7.07), 8.39 (1H, s), 8.86 (1H, s), 11.36 (1H, s).
product was purified by silica gel column chromatography (n-hexane/
EtOAc, 10−50% EtOAc) to afford 29 (3.44 g, 7.62 mmol, 80% in 2
steps) as a yellow amorphous substance. 1H NMR (CDCl3) δ 1.71 (3H,
s), 2.71, (1H, d, J = 14.1 Hz), 3.39 (1H, d, J = 14.1 Hz), 5.01 (1H, s),
5.09 (1H, s), 7.10 (1H, dd, J = 10.6, 10.1 Hz), 7.31−7.40 (4H, m), 7.47
(1H, dd, J = 7.0, 7.0 Hz), 8.05 (2H, d, J = 7.6 Hz), 8.08−8.13 (1H, m).
MS-ESI (m/z): 452 [M + H]+.
(S)-4-(5-Amino-2-fluorophenyl)-4-methyl-6-methylene-5,6-
dihydro-4H-1,3-thiazin-2-amine (30). To a stirred solution of 29
(100 mg, 0.222 mmol) in EtOH (1 mL) was added hydrazine hydrate
(54.0 μL, 1.11 mmol). After stirring overnight at room temperature, the
reaction mixture was concentrated under reduced pressure, and the
crude product was purified by silica gel column chromatography (n-
hexane/EtOAc, 50−100% EtOAc) to afford 30 (40.3 mg, 0.160 mmol,
74%) as a yellow amorphous substance. 1H NMR (CDCl3) δ 1.54 (3H,
s), 2.53, (1H, d, J = 13.6), 2.89 (1H, d, J = 13.6), 4.97 (1H, s), 5.09 (1H,
s), 6.45−6.51 (1H, m) 6.73−6.84 (2H, m).
(S)-N-(3-(2-Amino-4-methyl-6-methylene-5,6-dihydro-4H-
1,3-thiazin-4-yl)-4-fluorophenyl)-5-cyanopicolinamide (13).
Compound 13 was prepared in a manner similar to that for 16 from
30 (96.3 mg, 0.391 mmol) using 5-cyanopicolinic acid (75.0 mg, 0.508
mmol) in 46% yield (69.0 mg, 0.181 mmol) as a yellow solid. 1H NMR
(DMSO-d6) δ 2.48−2.53 (3H, m), 2.73 (1H, d, J = 15.2 Hz), 4.95 (1H,
s), 5.08 (1H, s), 6.23 (2H, br s), 7.14 (1H, dd, J = 9.9, 9.9 Hz), 7.75
(1H, br s), 7.91 (1H, d, J = 5.1 Hz), 8.27 (1H, d, J = 8.1 Hz), 8.57 (1H,
d, J = 8.1 Hz), 9.19 (1H, s), 10.73 (1H, s). MS-ESI (m/z): 382 [M +
H]+. 13C NMR (CDCl3) δ 28.3, 57.7, 57.8, 109.8, 111.5, 116.1 (d, J =
25.6 Hz), 116.6, 121.0 (d, J = 8.8 Hz), 122.0 (d, J = 5.1 Hz), 122.3,
133.7 (d, J = 2.9 Hz), 134.3 (d, J = 13.2 Hz), 136.9, 142.1, 147.5, 151.4,
152.6, 156.7 (d, J = 242 Hz), 161.0. HRMS-ESI (m/z): [M + H]+ calcd
for C19H17FN5OS+, 382.1132; found, 382.1128.
(S)-3-Amino-4-fluoro-3-(2-fluorophenyl)butan-1-ol (32). To a
stirred solution of lithium diisopropylamide (2 M in THF/heptane/
ethylbenzene, 6.65 mL, 13.3 mmol) was added at −78 °C tert-butyl
acetate (1.80 mL, 13.3 mmol) in THF (1 mL) over 5 min under N2.
After stirring for 30 min at the same temperature, Ti(Oi-Pr)3Cl in THF
(8.0 mL) was added over 15 min. After stirring for 35 min, 31 (1.15 g,
4.43 mmol) in THF (6.0 mL) was added over 15 min. After stirring for
15 min at the same temperature, the reaction was quenched with
saturated aq. NH4Cl (4.0 mL) and diluted with EtOAc. The insoluble
materials were filtered off, and the resulting organic layer was washed
with brine, dried over MgSO4, and concentrated under reduced
pressure. The residue in 1,4-dioxane (2.0 mL) and HCl (4 M in 1,4-
dioxane, 6.65 mL, 26.6 mmol) was warmed to 60 °C. After stirring for
20 min, 1,4-dioxane (3.0 mL) was added, and the mixture was stirred for
30 min. The reaction was cooled to 0 °C; ether (10 mL) was added, and
the resulting solid was filtered and dried to give an amino acid
hydrochloride (902 mg) as a white solid. To the solid was added BH3
(1.0 M in THF, 10.8 mL, 10.8 mmol) at room temperature. After
stirring at 45 °C for 1 h, the reaction was quenched with MeOH and
water. The aqueous layer was basified with NaHCO3 and extracted with
EtOAc. The organic layer was washed with brine, dried over MgSO4,
and concentrated under reduced pressure to give 32 (659 mg, 3.28
mmol, 92%) as a pale yellow oil. 1H NMR (CDCl3) δ 2.02−2.19 (2H,
m), 3.41−3.49 (1H, m), 3.62−3.68 (1H, m), 4.49 (1H, dd, J = 47.4, 9.0
Hz), 4.71 (2H, ddd, J = 47.8, 9.0, 1.9 Hz), 7.05 (1H, ddd, J = 12.8, 8.2,
1.6 Hz), 7.20 (1H, ddd, J = 8.0, 8.0, 1.2 Hz), 7.28−7.36 (1H, m), 7.57
(1H, ddd, J = 8.0, 8.0, 1.0 Hz). MS-ESI (m/z): 374 [M + H]+.
(S)-N-(4-(Fluoromethyl)-4-(2-fluorophenyl)-5,6-dihydro-4H-
1,3-thiazin-2-yl)benzamide (33). To a stirred solution of 32 (659
mg, 3.28 mmol) in DCM (6.0 mL) was added at 0 °C benzoyl
isothiocyanate (464 μL, 3.44 mmol). After stirring overnight at room
temperature, Ghosez’s reagent (1-chloro-N,N-2-trimethylpropenyl-
amine) (677 μL, 4.91 mmol) was added. After stirring for 40 min,
Ghosez’s reagent (135 μL, 0.984 mmol) was added. After stirring for 1
h, Ghosez’s reagent (90.0 μL, 0.656 mmol) was added and stirred for 25
min. The reaction mixture was concentrated and treated with i-Pr2O.
The resulting solid was filtered and then suspended in ether, filtered,
and dried in vacuo to afford 33 (918 mg, 2.65 mmol, 81% in 2 steps) as a
pale yellow solid. 1H NMR (CDCl3) δ 2.48−2.59 (1H, m), 2.70−2.86
(S)-N-(4-(2-Fluoro-5-(2,2,2-trifluoroacetamido)phenyl)-4-
methyl-6-methylene-5,6-dihydro-4H-1,3-thiazin-2-yl)-
benzamide (29). To a stirred solution of 28 (4.34 g, 9.57 mmol) in
DCM (220 mL) was added at 0 °C iodine (3.64 g, 14.4 mmol). After
stirring at room temperature for 2 h, the reaction mixture was quenched
with aq. Na2S2O3. The aqueous media was basified (pH > 11) with
K2CO3 and extracted with CHCl3. The organic layer was washed with
brine, dried over Na2SO4, and concentrated under reduced pressure. To
the resulting brown amorphous substance in THF (130 mL) was added
pyrrolidine (3.96 mL, 47.9 mmol). After stirring at 60 °C for 2 h, The
volatile materials were evaporated, and water was added. The aqueous
layer was extracted with EtOAc. The organic layer was washed with
water and brine and dried over Na2SO4 and concentrated. The crude
1884
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
(2H, m), 3.04 (2H, ddd, J = 17.8, 7.3, 4.7 Hz), 4.90 (1H, dd, J = 46.2,
10.0 Hz), 4.97 (1H, dd, J = 46.7, 10.1 Hz), 7.16 (1H, dd, J = 12.5, 8.2
Hz), 7.27−7.33 (1H, m), 7.40−7.69 (5H, m), 8.38−8.41 (2H, m),
13.78 (1H, br s). MS-ESI (m/z): 374 [M + H]+.
Institutional Review Boards. Written informed consent was obtained
from all participants before their participation in any clinical study. All
study procedures were in accordance with current ICH guidelines on
Good Clinical Practice (GCP), applicable regulatory and country-
specific requirements, and the principles of the Declaration of Helsinki.
(S)-4-(Fluoromethyl)-4-(2-fluorophenyl)-5,6-dihydro-4H-
1,3-thiazin-2-amine (34). A mixture of 33 (898 mg, 2.59 mmol) and
aq. H2SO4 (18 M, 2.0 mL) was heated to 50 °C. After stirring for 1 h, the
reaction mixture was cooled to 0 °C, and ice and 28% aq. NH3 were
added. The aqueous layer was extracted with EtOAc. The organic layer
was washed with brine, dried over MgSO4, and concentrated under
reduced pressure. The residue was triturated with iPr2O, and the
resulting solid was filtered and dried to afford 35 (447 mg, 1.85 mmol,
71%) as a beige solid. 1H NMR (CDCl3) δ 1.97 (1H, ddd, J = 13.1, 13.1,
4.0 Hz), 2.57−2.74 (2H, m), 2.93 (1H, ddd, J = 12.2, 4.1, 4.1 Hz),
4.47−4.83 (2H, m), 7.03 (1H, ddd, J = 12.5, 7.8, 1.3 Hz), 7.14 (1H,
ddd, J = 7.8, 7.8, 1.2 Hz), 7.28−7.40 (2H, m). MS-ESI (m/z): 243 [M +
H]+.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.jmedchem.0c01917.
■
sı
Experimental procedures for PK/PD study; measurement
of Aβ plaque formation in PS/APP mice; determination
of CVB values in human liver microsomes; synthesis of
1
[14C]-15, H, and 13C NMR spectra for compound 1;
HPLC traces for compounds 1, 13, 14, 15, 16, 17, and 18;
(S)-4-(2-Fluoro-5-nitrophenyl)-4-(fluoromethyl)-5,6-dihy-
dro-4H-1,3-thiazin-2-amine (35). To a stirred solution of 34 (407
mg, 1.68 mmol) in aq. H2SO4 (18 M, 1.6 mL), conc. HNO3 (150 μL,
3.36 mmol) in conc. H2SO4 (0.4 mL) was added at −15 °C over 20 min.
After stirring for 40 min at 0 °C, to the reaction mixture were added ice
and aq. NH3 (28%, 6 mL) dropwise. The aqueous layer was extracted
with EtOAc. The organic layer was washed with brine, dried over
MgSO4, and concentrated under reduced pressure to give 35 (474 mg,
and crystallographic data (PDF)
SMILES strings, BACE1 IC50, and cellular Aβ40 IC50
(CSV)
Accession Codes
The PDB accession code for the X-ray structure of 1 bound to
human BACE1 is 7DCZ.
1
1.65 mmol, 98%) as a yellow foam. H NMR (CDCl3) δ 2.03−2.13
AUTHOR INFORMATION
Corresponding Authors
(1H, m), 2.51 (1H, ddd, J = 13.8, 6.1, 4.8 Hz), 2.71 (1H, td, J = 11.9, 3.6
Hz), 3.02 (1H, dt, J = 13.2, 4.8 Hz), 4.50−4.73 (4H, m), 7.20 (2H, dd, J
= 10.9, 8.9 Hz), 8.22−8.17 (1H, m), 8.40 (1H, dd, J = 6.7, 2.9 Hz). MS-
ESI (m/z): 288 [M + H]+.
■
Yuji Koriyama − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan; Phone: +81-6-6209-6976; Email: yuuji.kooriyama@
shionogi.co.jp
Ken-ichi Kusakabe − Medicinal Research Laboratories,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan; orcid.org/0000-0001-9055-8267;
Phone: +81-6-6331-6190; Email: ken-ichi.kusakabe@
shionogi.co.jp; Fax: +81-6-6332-6385
(S)-N-(3-(2-Amino-4-(fluoromethyl)-5,6-dihydro-4H-1,3-
thiazin-4-yl)-4-fluorophenyl)-5-cyanopicolinamide (14). To a
stirred solution of 35 (58.5 mg, 0.204 mmol) in MeOH (1.5 mL) were
added conc. HCl (12 M, 50 μL) and iron (56.9 mg, 1.02 mmol). After
stirring at 50 °C for 2 h, insoluble materials were filtered off and washed
with MeOH. The filtrate was concentrated, and the resulting residue
was diluted with EtOAc and 2 M aq. NaOH (0.5 mL), which was then
filtered through Celite. The organic layer was washed with brine, dried
over MgSO4, and concentrated under reduced pressure. To a stirred
solution of the residue in MeOH (1 mL) were added 5-cyanopicolinic
acid hydrate (32.2 mg, 0.194 mmol), 2 M aq. HCl (97 μL, 0.19 mmol),
and EDC·HCl. After stirring for 1 h, 2 M aq. NaOH (0.12 mL) was
added at 0 °C. The resulting solid was filtered, washed with MeOH,
water, and ether, and dried in vacuo to give 14 (40.0 mg, 0.103 mmol,
53%) as a yellow solid. 1H NMR (CDCl3) δ 1.97−2.07 (1H, m), 2.60
(1H, ddd, J = 13.7, 4.5, 4.5 Hz), 2.77 (1H, ddd, J = 12.2, 12.2, 3.5 Hz),
2.98 (1H, ddd, J = 12.2, 4.5, 4.5 Hz), 4.49−4.81 (2H, m), 7.10 (1H, dd,
J = 11.6, 8.9 Hz), 7.45 (1H, dd, J = 7.0, 2.8 Hz), 8.04 (1H, ddd, J = 8.9,
4.1, 2.9 Hz), 8.20 (1H, dd, J = 8.1, 2.0 Hz), 8.42 (1H, d, J = 8.2 Hz),
8.89−8.90 (1H, m), 9.84 (1H, br s). 13C NMR (DMSO-d6) δ 21.9,
25.1, 59.8 (dd, J = 17.6, 5.1 Hz), 88.5 (dd, J = 178, 2.9 Hz), 111.5, 116.2
(d, J = 24.9 Hz), 116.6, 121.8 (d, J = 8.8 Hz), 122.3, 123.9, 129.1 (dd, J =
13.2, 5.1 Hz) 134.0, 142.1, 151.4, 151.5, 152.4, 155.9 (d, J = 244.3 Hz),
161.2. MS-ESI (m/z): 388 [M + H]+. HRMS-ESI (m/z): [M + H]+
calcd for C18H16F2N5OS+, 388.1038; found, 388.1032.
Biochemical BACE1 Assay. The biochemical BACE1 IC50 values
were determined by HTRF assay using an APP derived peptide as
described previously.18a
Cellular Aβ Assay. The cellular Aβ IC50 values were determined by
measuring Aβ40 using HTRF assay in SH-SY5Y cells expressing human
APP as described previously.18a
In Vitro ADMET Assays. Metabolic stability in microsomes, P-gp
efflux ratio, protein binding, CYP IC50, and hERG inhibitory activity
were determined according to procedures described previously.18,27
In Vivo Experiments. All the procedures were in accordance with
regulations and established guidelines and were approved by the
Shionogi Animal Care and Use Committee. In vivo experimental
procedures are given in Supporting Information.
Authors
Akihiro Hori − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
Hisanori Ito − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
Shuji Yonezawa − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
Yoshiyasu Baba − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
Norihiko Tanimoto − Medicinal Research Laboratories,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan
Tatsuhiko Ueno − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
Shiho Yamamoto − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
Takahiko Yamamoto − Medicinal Research Laboratories,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan
Naoya Asada − Research Laboratory for Development, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
Clinical Studies. All clinical study protocols and their amendments
were approved by the appropriate Independent Ethics Committees/
1885
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Kenji Morimoto − Research Laboratory for Development,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan
Shunsuke Einaru − Research Laboratory for Development,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan
Katsunori Sakai − Research Laboratory for Development,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan
Takushi Kanazu − Research Laboratory for Development,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan
Akihiro Matsuda − Research Laboratory for Development,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan
Yoshitaka Yamaguchi − Research Laboratory for Development,
Shionogi Pharmaceutical Research Center, Toyonaka, Osaka
561-0825, Japan
Takuya Oguma − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan; orcid.org/0000-0002-6305-2305
Maarten Timmers − Janssen Research & Development, a
division of Janssen Pharmaceutica N.V., B-2340 Beerse,
Belgium
Network; DILI, drug-induced liver injury; ELISA, enzyme-
linked immunosorbent assay; GSH, glutathione; HTS, high
throughput screening; EDC·HCl, 1-(3-dimehylaminopropyl)-3-
ethylcarbodiimide hydrochloride; Fmoc, 9-fluorenylmethoxy-
carbonyl; IADR, idiosyncratic adverse drug reaction; IAPP, islet
amyloid peptide; IBX, 2-iodoxybenzoic acid; MAD, multiple
ascending dose; MC, methyl cellulose; MCI, mild cognitive
impairment; NOAEL, no observed adverse effect level; NRG1,
neuregulin 1; PACC, Preclinical Alzheimer’s Cognitive
Composite; PMEL, pigment cell-specific melanocyte protein;
TMEM27, pro-proliferative type I transmembrane protein 27;
RBANS, Repeatable Battery for the Assessment of Neuro-
psychological Battery; SAD, single ascending dose; SEZ6,
seizure protein 6; TDI, time-dependent inhibition; TPSA,
topological polar surface area; ULN, upper limit of the normal
level; VCAM1, vascular cell adhesion molecule 1
REFERENCES
■
(1) Alzheimer’s Disease International. World Alzheimer Report 2019.
https://www.alz.co.uk/research/WorldAlzheimerReport2019.pdf (ac-
cessed April 30, 2020).
(2) Alzheimer’s Association. 2020 Alzheimer’s Disease Facts and
Figures. https://www.alz.org/media/Documents/alzheimers-facts-
and-figures.pdf (accessed April 30, 2020).
Luc Tritsmans − Janssen Research & Development, a division of
Janssen Pharmaceutica N.V., B-2340 Beerse, Belgium
Akira Kato − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
Gaku Sakaguchi − Medicinal Research Laboratories, Shionogi
Pharmaceutical Research Center, Toyonaka, Osaka 561-0825,
Japan
(3) Winblad, B.; Amouyel, P.; Andrieu, S.; Ballard, C.; Brayne, C.;
Brodaty, H.; Cedazo-Minguez, A.; Dubois, B.; Edvardsson, D.;
Feldman, H.; Fratiglioni, L.; Frisoni, G. B.; Gauthier, S.; Georges, J.;
̈
Graff, C.; Iqbal, K.; Jessen, F.; Johansson, G.; Jonsson, L.; Kivipelto, M.;
Knapp, M.; Mangialasche, F.; Melis, R.; Nordberg, A.; Rikkert, M. O.;
Qiu, C.; Sakmar, T. P.; Scheltens, P.; Schneider, L. S.; Sperling, R.;
Tjernberg, L. O.; Waldemar, G.; Wimo, A.; Zetterberg, H. Defeating
Alzheimer’s disease and other dementias: a priority for European
science and society. Lancet Neurol. 2016, 15, 455−532.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.jmedchem.0c01917
(4) Bateman, R. J.; Xiong, C.; Benzinger, T. L. S.; Fagan, A. M.; Goate,
A.; Fox, N. C.; Marcus, D. S.; Cairns, N. J.; Xie, X.; Blazey, T. M.;
Holtzman, D. M.; Santacruz, A.; Buckles, V.; Oliver, A.; Moulder, K.;
Aisen, P. S.; Ghetti, B.; Klunk, W. E.; McDade, E.; Martins, R. N.;
Masters, C. L.; Mayeux, R.; Ringman, J. M.; Rossor, M. N.; Schofield, P.
R.; Sperling, R. A.; Salloway, S.; Morris, J. C. Dominantly Inherited
Alzheimer Network. Clinical and biomarker changes in dominantly
inherited Alzheimer’s disease. N. Engl. J. Med. 2012, 367, 795−804.
(5) Yan, R.; Vassar, R. Targeting the β secretase BACE1 for
Alzheimer’s disease therapy. Lancet Neurol. 2014, 13, 319−329.
(6) Jonsson, T.; Atwal, J. K.; Steinberg, S.; Snaedal, J.; Jonsson, P. V.;
Bjornsson, S.; Stefansson, H.; Sulem, P.; Gudbjartsson, D.; Maloney, J.;
Hoyte, K.; Gustafson, A.; Liu, Y.; Lu, Y.; Bhangale, T.; Graham, R. R.;
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Hidekuni Yamakawa, Chie Unemura, and Motoko
Hosono for the biological assays; Yumiko Takagi for the
elemental analysis; Atsushi Morimoto for the HRMS analysis;
and Yukiko Kan for the NMR analysis. We also thank Dieder
Moechars and Harrie Gijsen for critically reading the manuscript
and helpful comments. We gratefully thank the members of the
Janssen-Shionogi research collaboration for the support of this
program.
̈
Huttenlocher, J.; Bjornsdottir, G.; Andreassen, O. A.; Jonsson, E. G.;
Palotie, A.; Behrens, T. W.; Magnusson, O. T.; Kong, A.;
Thorsteinsdottir, U.; Watts, R. J.; Stefansson, K. A mutation in APP
protects against Alzheimer’s disease and age-related cognitive decline.
Nature 2012, 488, 96−99.
DEDICATION
■
(7) (a) Oehlrich, D.; Prokopcova, H.; Gijsen, H. J. M. The evolution
of amidine-based brain penetrant BACE1 inhibitors. Bioorg. Med. Chem.
Lett. 2014, 24, 2033−2045. (b) Hall, A.; Gijsen, H. J. M. Targeting β-
secretase (BACE) for the Treatment of Alzheimer’s Disease. In
Comprehensive Medicinal Chemistry III, 3rd ed.; Chackalamannil, S.,
Rotella, D. P., Ward, S. E., Eds.; Elsevier: Oxford, 2017; Vol. 7, pp 326−
383. (c) Hsiao, C.-C.; Rombouts, F.; Gijsen, H. J. M. New evolutions in
the BACE1 inhibitor field from 2014 to 2018. Bioorg. Med. Chem. Lett.
2019, 29, 761−777. (d) Ghosh, A. K.; Osswald, H. L. BACE1 (β-
secretase) inhibitors for the treatment of Alzheimer’s disease. Chem.
Soc. Rev. 2014, 43, 6765−6813. (e) Yuan, J.; Venkatraman, S.; Zheng,
Y.; McKeever, B. M.; Dillard, L. W.; Singh, S. B. Structure based design
of β-site APP cleaving enzyme 1 (BACE) inhibitors for the treatment of
Alzheimer’s disease. disease. J. Med. Chem. 2013, 56, 4156−4180.
(f) Rombouts, F.; Kusakabe, K.-I.; Hsiao, C.-C.; Gijsen, H. J. M. Small-
This paper is dedicated to the memory of our colleague, Yuko
Ninomiya. She is an outstanding toxicologist, and her excellent
efforts on the toxicological studies described in this paper
significantly contributed to the development of atabecestat.
ABBREVIATIONS
■
AD, Alzheimer’s disease; ALP, alkaline phosphatase; ALT,
alanine aminotransferase; APP, amyloid precursor protein; AST,
aspartate aminotransferase; BACE1, β-site amyloid precursor
protein cleaving enzyme 1; CatD, cathepsin D; CDR, Clinical
Dementia Rating; CHL1, close homologue of neural cell
adhesion molecule L1; CSF, cerebrospinal fluid; DDI, drug−
drug interaction; DIAN, Dominantly Inherited Alzheimer
1886
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
molecule BACE1 inhibitors: a patent literature review (2011 to 2020).
Expert Opin. Ther. Pat. 2021, 31, 25−52.
2017; Vol. 8, pp 204−251. (d) Egan, M. F.; Kost, J.; Tariot, P. N.; Aisen,
P. S.; Cummings, J. L.; Vellas, B.; Sur, C.; Mukai, Y.; Voss, T.; Furtek,
C.; Mahoney, E.; Mozley, L. H.; Vandenberghe, R.; Mo, Y.; Michelson,
D. Randomized trial of verubecestat for mild-to-moderate Alzheimer’s
disease. N. Engl. J. Med. 2018, 378, 1691−1703. (e) Egan, M. F.; Kost,
J.; Voss, T.; Mukai, Y.; Aisen, P. S.; Cummings, J. L.; Tariot, P. N.;
Vellas, B.; van Dyck, C. H.; Boada, M.; Zhang, Y.; Li, W.; Furtek, C.;
Mahoney, E.; Mozley, L. H.; Mo, Y.; Sur, C.; Michelson, D.
Randomized trial of verubecestat for prodromal Alzheimer’s disease.
N. Engl. J. Med. 2019, 380, 1408−1420.
(13) (a) Eketjall, S.; Janson, J.; Kaspersson, K.; Bogstedt, A.; Jeppsson,
F.; Falting, J.; Haeberlein, S. B.; Kugler, A. R.; Alexander, R. C.; Cebers,
G. AZD3293: A novel, orally active BACE1 inhibitor with high potency
and permeability and markedly slow off-rate kinetics. J. Alzheimer's Dis.
2016, 50, 1109−1123. (b) Wessels, A. M.; Tariot, P. N.; Zimmer, J. A.;
Selzler, K. J.; Bragg, S. M.; Andersen, S. W.; Landry, J.; Krull, J. H.;
Downing, A. M.; Willis, B. A.; Shcherbinin, S.; Mullen, J.; Barker, P.;
Schumi, J.; Shering, C.; Matthews, B. R.; Stern, R. A.; Vellas, B.; Cohen,
S.; MacSweeney, E.; Boada, M.; Sims, J. R. Efficacy and safety of
lanabecestat for treatment of early and mild Alzheimer disease: The
AMARANTH and DAYBREAK-ALZ randomized clinical trials. JAMA
Neurol. 2020, 77, 199−209.
(14) (a) Alzforum. https://www.alzforum.org/therapeutics/
elenbecestat (accessed August 10, 2020). (b) Suzuki, Y.; Motoki, T.;
Kaneko, T.; Takaishi, M.; Ishida, T.; Takeda, K.; Kita, Y.; Yamamoto,
N.; Khan, A.; Dimopoulos, P. Preparation of Condensed Amino-
dihydrothiazine Derivatives as Inhibitors of β-Site APP-cleaving
Enzyme 1 (BACE1). WO 2009091016 A1, July 23, 2009.
(15) (a) Neumann, U.; Ufer, M.; Jacobson, L. H.; Rouzade-
Dominguez, M.-L.; Huledal, G.; Kolly, C.; Lueoend, R. M.;
Machauer, R.; Veenstra, S. J.; Hurth, K.; Rueeger, H.; Tintelnot-
Blomley, M.; Staufenbiel, M.; Shimshek, D. R.; Perrot, L.; Frieauff, W.;
Dubost, V.; Schiller, H.; Vogg, B.; Beltz, K.; Avrameas, A.; Kretz, S.;
Pezous, N.; Rondeau, J.; Beckmann, N.; Hartmann, A.; Vormfelde, S.;
David, O. J.; Galli, B.; Ramos, R.; Graf, A. Lopez Lopez, C. The BACE-1
inhibitor CNP520 for prevention trials in Alzheimer’s disease. EMBO
Mol. Med. 2018, 10, e9316. (b) Alzforum. https://www.alzforum.org/
therapeutics/umibecestat (accessed Aug 10, 2020).
(16) Hampel, H.; Vassar, R.; De Strooper, B.; Hardy, J.; Willem, M.;
Singh, N.; Zhou, J.; Yan, R.; Vanmechelen, E.; De Vos, A.; Nistico, R.;
Corbo, M.; Imbimbo, B. P.; Streffer, J.; Voytyuk, I.; Timmers, M.;
Tahami Monfared, A. A., Irizarry, M.; Albala, B.; Koyama, A.;
Watanabe, N.; Kimura, T.; Yarenis, L.; Lista, S.; Kramer, L.; Vergallo,
A. The β-Secretase BACE1 in Alzheimer’s disease. Biol. Psychiatry 2020
DOI: 10.1016/j.biopsych.2020.02.001.
(17) Voytyuk, I.; Mueller, S. A.; Herber, J.; Snellinx, A.; Moechars, D.;
van Loo, G.; Lichtenthaler, S. F.; De Strooper, B. BACE2 distribution in
major brain cell types and identification of novel substrates. Life Sci.
Alliance 2018, 1, e201800026.
(18) (a) Fuchino, K.; Mitsuoka, Y.; Masui, M.; Kurose, N.; Yoshida,
S.; Komano, K.; Yamamoto, T.; Ogawa, M.; Unemura, C.; Hosono, M.;
Ito, H.; Sakaguchi, G.; Ando, S.; Ohnishi, S.; Kido, Y.; Fukushima, T.;
Miyajima, H.; Hiroyama, S.; Koyabu, K.; Dhuyvetter, D.; Borghys, H.;
Gijsen, H. J.M.; Yamano, Y.; Iso, Y.; Kusakabe, K. Rational design of
novel 1,3-oxazine based β-secretase (BACE1) inhibitors: Incorporation
of a double bond to reduce P-gp efflux leading to robust Aβ reduction in
the brain. J. Med. Chem. 2018, 61, 5122−5137. (b) Nakahara, K.;
Fuchino, K.; Komano, K.; Asada, N.; Tadano, G.; Hasegawa, T.;
Yamamoto, T.; Sako, Y.; Ogawa, M.; Unemura, C.; Hosono, M.; Ito, H.;
Sakaguchi, G.; Ando, S.; Ohnishi, S.; Kido, Y.; Fukushima, T.;
Dhuyvetter, D.; Borghys, H.; Gijsen, H. J. M.; Yamano, Y.; Iso, Y.;
Kusakabe, K. Discovery of potent and centrally active 6-substituted 5-
fluoro-1,3-dihydro-oxazine β-secretase (BACE1) inhibitors via active
conformation stabilization. J. Med. Chem. 2018, 61, 5525−5546.
(19) (a) Woltering, T. J.; Wostl, W.; Hilpert, H.; Rogers-Evans, M.;
Pinard, E.; Mayweg, A.; Gobel, M.; Banner, D. W.; Benz, J.; Travagli,
M.; Pollastrini, M.; Marconi, G.; Gabellieri, E.; Guba, W.; Mauser, H.;
Andreini, M.; Jacobsen, H.; Power, E.; Narquizian, R. BACE1
inhibitors: A head group scan on a series of amides. Bioorg. Med.
(8) (a) Timmers, M.; Van Broeck, B.; Ramael, S.; Slemmon, J.; De
Waepenaert, K.; Russu, A.; Bogert, J.; Stieltjes, H.; Shaw, L. M.;
Engelborghs, S.; Moechars, D.; Mercken, M.; Liu, E.; Sinha, V.; Kemp,
J.; Van Nueten, L.; Tritsmans, L.; Streffer, J. R. Profiling the dynamics of
CSF and plasma Aβ reduction after treatment with JNJ-54861911, a
potent oral BACE inhibitor. Alzheimers Dement (N Y) 2016, 2, 201−
212. (b) Timmers, M.; Streffer, J. R.; Russu, A.; Tominaga, Y.; Shimizu,
H.; Shiraishi, A.; Tatikola, K.; Smekens, P.; Borjesson-Hanson, A.;
Andreasen, N.; Matias-Guiu, J.; Baquero, M.; Boada, M.; Tesseur, I.;
Tritsmans, L.; Van Nueten, L.; Engelborghs, S. Pharmacodynamics of
atabecestat (JNJ-54861911), an oral BACE1 inhibitor in patients with
early Alzheimer’s disease: randomized, double-blind, placebo-con-
trolled study. Alzheimer’s Res. Ther 2018, 10, 85/1−85/18. (c) Henley,
D.; Raghavan, N.; Sperling, R.; Aisen, P.; Raman, R.; Romano, G.
Preliminary results of a trial of atabecestat in preclinical Alzheimer’s
disease. N. Engl. J. Med. 2019, 380, 1483−1485. (d) Novak, G.; Streffer,
J. R.; Timmers, M.; Henley, D.; Brashear, H. R.; Bogert, J.; Russu, A.;
Janssens, L.; Tesseur, I.; Tritsmans, L.; Van Nueten, L.; Engelborghs, S.
Long-term safety and tolerability of atabecestat (JNJ-54861911), an
oral BACE1 inhibitor, in early Alzheimer’s disease spectrum patients: a
randomized, double-blind, placebo-controlled study and a two-period
extension study. Alzheimer's Res. Ther. 2020, 12, 58.
(9) May, P. C.; Dean, R. A.; Lowe, S. L.; Martenyi, F.; Sheehan, S. M.;
Boggs, L. N.; Monk, S. A.; Mathes, B. M.; Mergott, D. J.; Watson, B. M.;
Stout, S. L.; Timm, D. E.; LaBell, E. S.; Gonzales, C. R.; Nakano, M.;
Jhee, S. S.; Yen, M.; Ereshefsky, L.; Lindstrom, T. D.; Calligaro, D. O.;
Cocke, P. J.; Hall, D. G.; Friedrich, S.; Citron, M.; Audia, J. E. Robust
central reduction of amyloid-β in humans with an orally available, non-
peptidic β-secretase inhibitor. J. Neurosci. 2011, 31, 16507−16516.
(10) May, P. C.; Willis, B. A.; Lowe, S. L.; Dean, R. A.; Monk, S. A.;
Cocke, P. J.; Audia, J. E.; Boggs, L. N.; Borders, A. R.; Brier, R. A.;
Calligaro, D. O.; Day, T. A.; Ereshefsky, L.; Erickson, J. A.; Gevorkyan,
H.; Gonzales, C. R.; James, D. E.; Jhee, S. S.; Komjathy, S. F.; Li, L.;
Lindstrom, T. D.; Mathes, B. M.; Martenyi, F.; Sheehan, S. M.; Stout, S.
L.; Timm, D. E.; Vaught, G. M.; Watson, B. M.; Winneroski, L. L.; Yang,
Z.; Mergott, D. J. The potent BACE1 inhibitor LY2886721 elicits
robust central Aβ pharmacodynamic responses in mice, dogs, and
humans. J. Neurosci. 2015, 35, 1199−1210.
(11) Mergott, D.; Audia, J.; Barberis, M.; Beck, J.; Boggs, L.; Boyer, R.;
Brier, R.; Borders, A.; Daugherty, L.; Dean, R.; Ereshefsky, L.; Erickson,
J.; Garcia-Losada, P.; Gevorkyan, H.; Green, S.; Hembre, E.; Irizarry,
M.; James, D.; Jhee, S.; Lin, Q.; Lopez, J.; Lo, A.; Lowe, S.; Mathes, B.;
May, P.; McKinzie, D.; Monk, S.; Nakano, M.; Porter, W.; Shi, Y.; Stout,
S.; Timm, D.; Watson, B.; Willis, B.; Winneroski, L.; Yang, Z.; Zimmer,
J. Discovery, Structure Disclosure, and Early Clinical Development of
LY3202626, a Low-Dose, CNS-penetrant BACE Inhibitor. Abstracts of
Papers; 256th ACS National Meeting & Exposition, Boston, MA,
United States, August 19−23, 2018. MEDI-330.
(12) (a) Scott, J. D.; Li, S. W.; Brunskill, A. P. J.; Chen, X.; Cox, K.;
Cumming, J. N.; Forman, M.; Gilbert, E. J.; Hodgson, R. A.; Hyde, L.
A.; Jiang, Q.; Iserloh, U.; Kazakevich, I.; Kuvelkar, R.; Mei, H.;
Meredith, J.; Misiaszek, J.; Orth, P.; Rossiter, L. M.; Slater, M.; Stone, J.;
Strickland, C. O.; Voigt, J. H.; Wang, G.; Wang, H.; Wu, Y.; Greenlee,
W. J.; Parker, E. M.; Kennedy, M. E.; Stamford, A. W. Discovery of the
3-imino-1,2,4-thiadiazinane 1,1-dioxide derivative verubecestat (MK-
8931)−A β-site amyloid precursor protein cleaving enzyme 1 inhibitor
for the treatment of Alzheimer’s disease. J. Med. Chem. 2016, 59,
10435−10450. (b) Kennedy, M. E.; Chen, X.; Hodgson, R. A.; Hyde, L.
A.; Kuvelkar, R.; Parker, E. M.; Stamford, A. W.; Cumming, J. N.; Li,
W.; Scott, J. D.; Cox, K.; Dockendorf, M. F.; Kleijn, H. J.; Mei, H.;
Stone, J. A.; Egan, M.; Ereshefsky, L.; Jhee, S.; Mattson, B. A.; Palcza, J.;
Tanen, M.; Troyer, M. D.; Tseng, J. L.; Forman, M. S. The BACE1
inhibitor verubecestat (MK-8931) reduces CNS β-amyloid in animal
models and in Alzheimer’s disease patients. Sci. Transl. Med. 2016, 8,
363ra150. (c) Cumming, J. N.; Scott, J. D.; Strickland, C. O.
Verubecestat. In Comprehensive Medicinal Chemistry III, 3rd ed.;
Chackalamannil, S., Rotella, D. P., Ward, S. E., Eds.; Elsevier: Oxford,
1887
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Chem. Lett. 2013, 23, 4239−4243. (b) Hilpert, H.; Guba, W.;
Woltering, T. J.; Wostl, W.; Pinard, E.; Mauser, H.; Mayweg, A. V.;
Rogers-Evans, M.; Humm, R.; Krummenacher, D.; Muser, T.; Schnider,
C.; Jacobsen, H.; Ozmen, L.; Bergadano, A.; Banner, D. W.;
Hochstrasser, R.; Kuglstatter, A.; David-Pierson, P.; Fischer, H.;
Polara, A.; Narquizian, R. β-Secretase (BACE1) inhibitors with high in
vivo efficacy suitable for clinical evaluation in Alzheimer’s disease. J.
Med. Chem. 2013, 56, 3980−3995.
(20) Kobayashi, N.; Ueda, K.; Itoh, N.; Suzuki, S.; Sakaguchi, G.; Kato,
A.; Yukimasa, A.; Hori, A.; Koriyama, Y.; Haraguchi, H.; Yasui, K.;
Kanda, Y. Preparation of 2-Aminodihydrothiazine Derivatives as β-
Secretase Inhibitors. WO2007049532A1, March 5, 2007.
(27) Fielden, M. R.; Werner, J.; Jamison, J. A.; Coppi, A.; Hickman,
D.; Dunn, R. T., II; Trueblood, E.; Zhou, L.; Afshari, C. A.; Lightfoot-
Dunn, R. Retinal toxicity induced by a novel β-secretase inhibitor in the
Sprague-Dawley rat. Toxicol. Pathol. 2015, 43, 581−592.
(28) Barao, S.; Moechars, D.; Lichtenthaler, S. F.; De Strooper, B.
BACE1 physiological functions may limit its use as therapeutic target
for Alzheimer’s disease. Trends Neurosci. 2016, 39, 158−169.
(29) (a) Rochin, L.; Hurbain, I.; Serneels, L.; Fort, C.; Watt, B.;
Leblanc, P.; Marks, M. S.; De Strooper, B.; Raposo, G.; van Niel, G.
BACE2 processes PMEL to form the melanosome amyloid matrix in
pigment cells. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 10658−10663.
(b) Shimshek, D. R.; Jacobson, L. H.; Kolly, C.; Zamurovic, N.;
Balavenkatraman, K. K.; Morawiec, L.; Kreutzer, R.; Schelle, J.; Jucker,
M.; Bertschi, B.; Theil, D.; Heier, A.; Bigot, K.; Beltz, K.; Machauer, R.;
Brzak, I.; Perrot, L.; Neumann, U. Pharmacological BACE1 and BACE2
inhibition induces hair depigmentation by inhibiting PMEL17
processing in mice. Sci. Rep. 2016, 6, 21917.
(30) Esterhazy, D.; Stuetzer, I.; Wang, H.; Rechsteiner, M. P.;
Beauchamp, J.; Doebeli, H.; Hilpert, H.; Matile, H.; Prummer, M.;
Schmidt, A.; Lieske, N.; Boehm, B.; Marselli, L.; Bosco, D.; Kerr-Conte,
J.; Aebersold; Spinas, G. A.; Moch, H.; Migliorini, C.; Stoffel, M. Bace2
is a β cell-enriched protease that regulates pancreatic β cell function and
mass. Cell Metab. 2011, 14, 365−377.
(31) Rulifson, I. C.; Cao, P.; Miao, L.; Kopecky, D.; Huang, L.; White,
R. D.; Samayoa, K.; Gardner, J.; Wu, X.; Chen, K.; Tsuruda, T.;
Homann, O.; Baribault, H.; Yamane, H.; Carlson, T.; Wiltzius, J.; Li, Y.
Identification of human islet amyloid polypeptide as a BACE2
substrate. PLoS One 2016, 11, e0147254.
(32) Alic, I.; Goh, P. A.; Murray, A.; Portelius, E.; Gkanatsiou, E.;
Gough, G.; Mok, K. Y.; Koschut, D.; Brunmeir, R.; Yeap, Y. J.; O’Brien,
N. L.; Groet, J.; Shao, X.; Havlicek, S.; Dunn, N. R.; Kvartsberg, H.;
Brinkmalm, G.; Hithersay, R.; Startin, C.; Hamburg, S.; Phillips, M.;
Pervushin, K.; Turmaine, M.; Wallon, D.; Rovelet-Lecrux, A.; Soininen,
H.; Volpi, E.; Martin, J. E.; Foo, J. N.; Becker, D. L.; Rostagno, A.;
Ghiso, J.; Krsnik, Z.; Simic, G.; Kostovic, I.; Mitrecic, D.; Francis, P. T.;
Blennow, K.; Strydom, A.; Hardy, J.; Zetterberg, H.; Nizetic, D. Patient-
specific Alzheimer-like pathology in trisomy 21 cerebral organoids
reveals BACE2 as a gene dose-sensitive AD suppressor in human brain.
Mol. Psychiatry 2020 DOI: 10.1038/s41380-020-0806-5.
(21) Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. Medicinal
chemistry of hERG optimizations: Highlights and hang-ups. J. Med.
Chem. 2006, 49, 5029−5046.
(22) Doran, A.; Obach, R. S.; Smith, B. J.; Hosea, N. A.; Becker, S.;
Callegari, E.; Chen, C.; Chen, X.; Choo, E.; Cianfrogna, J.; Cox, L. M.;
Gibbs, J. P.; Gibbs, M. A.; Hatch, H.; Hop, C. E.C.A.; Kasman, I. N.;
LaPerle, J.; Liu, J.; Liu, X.; Logman, M.; Maclin, D.; Nedza, F. M.;
Nelson, F.; Olson, E.; Rahematpura, S.; Raunig, D.; Rogers, S.; Schmidt,
K.; Spracklin, D. K.; Szewc, M.; Troutman, M.; Tseng, E.; Tu, M.; Van
Deusen, J. W.; Venkatakrishnan, K.; Walens, G.; Wang, E. Q.; Wong,
D.; Yasgar, A. S.; Zhang, C. The impact of P-glycoprotein on the
disposition of drugs targeted for indications of the central nervous
system: Evaluation using the MDR1A/1B knockout mouse model.
Drug Metab. Dispos. 2005, 33, 165−174.
(23) Rankovic, Z. CNS Drug design: balancing physicochemical
properties for optimal brain exposure. J. Med. Chem. 2015, 58, 2584−
2608.
(24) Didziapetris, R.; Japertas, P.; Avdeef, A.; Petrauskas, A.
Classification analysis of P-glycoprotein substrate specificity. J. Drug.
Target. 2003, 11, 391−406.
(25) (a) Oguma, T.; Anan, K.; Suzuki, S.; Hisakawa, S.; Takada, A.;
Ogawa, M.; Kusakabe, K.-I. Synthesis of a 6-CF3-substituted 2-amino-
dihydro-1,3-thiazine β-secretase inhibitor by N,N-diethylsulfur tri-
fluoride-mediated chemoselective cyclization. J. Org. Chem. 2019, 84,
4893−4897. (b) Anan, K.; Iso, Y.; Oguma, T.; Nakahara, K.; Suzuki, S.;
Yamamoto, T.; Matsuoka, E.; Ito, H.; Sakaguchi, G.; Ando, S.;
Morimoto, K.; Kanegawa, N.; Kido, Y.; Kawachi, T.; Fukushima, T.;
Teisman, A.; Urmaliya, V.; Dhuyvetter, D.; Borghys, H.; Austin, N.;
Van Den Bergh, A.; Verboven, P.; Bischoff, F.; Gijsen, H. J. M.; Yamano,
Y.; Kusakabe, K. Trifluoromethyl dihydrothiazine-based β-Secretase
(BACE1) inhibitors with robust central β-amyloid reduction and
minimal covalent binding burden. ChemMedChem 2019, 14, 1894−
1910. (c) Tadano, G.; Komano, K.; Yoshida, S.; Suzuki, S.; Nakahara,
K.; Fuchino, K.; Fujimoto, K.; Matsuoka, E.; Yamamoto, T.; Asada, N.;
Ito, H.; Sakaguchi, G.; Kanegawa, N.; Kido, Y.; Ando, S.; Fukushima,
T.; Teisman, A.; Urmaliya, V.; Dhuyvetter, D.; Borghys, H.; Van Den
Bergh, A.; Austin, N.; Gijsen, H. J. M.; Yamano, Y.; Iso, Y.; Kusakabe,
K.-i. Discovery of an extremely potent thiazine-based β-secretase
(BACE1) inhibitor that minimized cardiovascular and liver toxicity at a
low projected human dose. J. Med. Chem. 2019, 62, 9331−9337.
(26) (a) Stepan, A. F.; Walker, D. P.; Bauman, J.; Price, D. A.; Baillie,
T. A.; Kalgutkar, A. S.; Aleo, M. D. Structural alert/reactive metabolite
concept as applied in medicinal chemistry to mitigate the risk of
idiosyncratic drug toxicity: A perspective based on the critical
examination of trends in the top 200 drugs marketed in the United
States. Chem. Res. Toxicol. 2011, 24, 1345−1410. (b) Thompson, R. A.;
Isin, E. M.; Li, Y.; Weidolf, L.; Page, K.; Wilson, I.; Swallow, S.;
Middleton, B.; Stahl, S.; Foster, A. J.; Dolgos, H.; Weaver, R.; Kenna, J.
G. In vitro approach to assess the potential for risk of idiosyncratic
adverse reactions caused by candidate drugs. Chem. Res. Toxicol. 2012,
25, 1616−1632. (c) Orr, S. T. M.; Ripp, S. L.; Ballard, T. E.; Henderson,
J. L.; Scott, D. O.; Obach, R. S.; Sun, H.; Kalgutkar, A. S. Mechanism-
based inactivation (MBI) of cytochrome P450 enzymes: structure-
activity relationships and discovery strategies to mitigate drug-drug
interaction Risks. J. Med. Chem. 2012, 55, 4896−4933. (d) Scheen, A. J.
Drug Saf. 2001, 24, 873−888.
(33) Kunishima, M.; Kawachi, C.; Iwasaki, F.; Terao, K.; Tani, S.
Synthesis and characterization of 4-(4,6-dimethoxy-1,3,5-triazi-2-yl)-4-
methylmorpholinium chloride. Tetrahedron Lett. 1999, 40, 5327−5330.
(34) Robak, M. T.; Herbage, M. A.; Ellman, J. A. Synthesis and
applications of tert-butanesulfinamide. Chem. Rev. 2010, 110, 3600−
3740.
(35) Haveaux, B.; Dekoker, A.; Rens, M.; Sidani, A. R.; Toye, J.;
Ghosez, L. α-Chloro enamines, reactive intermediates for synthesis: 1-
chloro-N,N,2-trimethylpropenylamine. Org. Synth. 1980, 59, 26−34.
1888
https://dx.doi.org/10.1021/acs.jmedchem.0c01917
J. Med. Chem. 2021, 64, 1873−1888