Identification of a Novel Positron Emission Tomography (PET) Ligand for Imaging β-Site Amyloid Precursor Protein Cleaving Enzyme 1 (BACE-1) in Brain

Article abstract of DOI:10.1021/acs.jmedchem.7b01769

Alzheimer's disease (AD) is characterized by accumulation of β-amyloid (Aβ) plaques and neurofibrillary tau tangles in the brain. β-Site amyloid precursor protein cleaving enzyme 1 (BACE1) plays a key role in the generation of Aβ fragments via extracellular cleavage of the amyloid precursor protein (APP). We became interested in developing a BACE1 PET ligand to facilitate clinical assessment of BACE1 inhibitors and explore its potential in the profiling and selection of patients for AD trials. Using a set of PET ligand design parameters, compound 3 (PF-06684511) was rapidly identified as a lead with favorable in vitro attributes and structural handles for PET radiolabeling. Further evaluation in an LC-MS/MS cold tracer study in rodents revealed high specific binding to BACE1 in brain. Upon radiolabeling, [¹⁸F]3 demonstrated favorable brain uptake and high in vivo specificity in nonhuman primate (NHP), suggesting its potential for imaging BACE1 in humans.

Full text of DOI:10.1021/acs.jmedchem.7b01769

Subscriber access provided by MT ROYAL COLLEGE
Article
Identification of a novel positron emission tomography (PET) ligand for
imaging #-site amyloid precursor protein cleaving enzyme 1 (BACE-1) in brain
Lei Zhang, Laigao Chen, Jason K Dutra, Elizabeth M. Beck, Sangram Nag, Akihiro
Takano, Nahid Amini, Ryosuke Arakawa, Michael A. Brodney, Leanne M Buzon,
Shawn D Doran, Lorraine F Lanyon, Timothy J McCarthy, Kelly R Bales, Charles E
Nolan, Brian T. O'Neill, Klaas Schildknegt, Christer Halldin, and Anabella Villalobos
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01769 • Publication Date (Web): 22 Jan 2018
Downloaded from http://pubs.acs.org on January 22, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Identification of a novel positron emission tomography
(PET) ligand for imaging βꢀsite amyloid precursor
protein cleaving enzyme 1 (BACEꢀ1) in brain
Lei Zhang, 1* Laigao Chen, 2 Jason K. Dutra, 3 Elizabeth M Beck, 1 Sangram Nag, 4 Akihiro Takano, 4
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
4
1
3
5
Nahid Amini, Ryosuke Arakawa, Michael A. Brodney, Leanne M. Buzon, Shawn D. Doran,
Lorraine F. Lanyon, 5 Timothy J. McCarthy, 2 Kelly R. Bales, 6 Charles E. Nolan, 6 Brian T. O’Neill, 3
Klaas Schildknegt, 7 Christer Halldin, 4 Anabella Villalobos8
1Medicine Design, Medicinal Chemistry, Pfizer Inc., Cambridge, Massachusetts 02139, USA; Clinical
2
& Translational Imaging, Early Clinical Development, Pfizer Inc., Cambridge, Massachusetts 02139,
3
United States; Medicine Design, Medicinal Chemistry, Pfizer Inc., Groton, Connecticut 06340, United
4
States; Department of Clinical Neuroscience, Center for Psychiatry Research, Karolinska Institutet and
5
Stockholm County Council, SEꢀ17176 Stockholm, Sweden; Medicine Design, Pharmacokinetics,
6
Dynamics and Metabolism, Pfizer Inc., Groton, Connecticut 06340, United States; Internal Medicine,
7
Pfizer Inc., Cambridge, Massachusetts 02139, United States; Pharmaceutical Sciences, Pfizer Inc.,
8
Groton, Connecticut 06340, United States; Medicinal Synthesis Technologies, Pfizer Inc., Groton,
Connecticut 06340, United States.
RECEIVED DATE (to be automatically inserted after your manuscript is accepted if required
Keywords: Positron emission tomography, PET, CNS PET MPO, BACE1, BACE1 selective, LC-MS/MS
cold tracer method, PET imaging
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 36
ABSTRACT:
1
2
3
4
5
6
7
8
9
Alzheimer’s disease (AD) is characterized by accumulation of βꢀamyloid (Aβ) plaques and
neurofibrillary tau tangles in brain. βꢀSite amyloid precursor protein cleaving enzyme 1 (BACE1) plays
a key role in the generation of Aβ fragments via extracellular cleavage of the amyloid precursor protein
(APP). We became interested in developing a BACE1 PET ligand to facilitate clinical assessment of
BACE1 inhibitors and explore its potential in the profiling and selection of patients for AD trials. Using
a set of PET ligand design parameters, compound 3 (PFꢀ06684511) was rapidly identified as a lead with
favorable in vitro attributes and structural handles for PET radiolabeling. Further evaluation in an LCꢀ
MS/MS “cold tracer” study in rodents revealed high specific binding to BACE1 in brain. Upon
radiolabeling, [18F]3 demonstrated favorable brain uptake and high in vivo specificity in nonꢀhuman
primate (NHP), suggesting its potential for imaging BACE1 in humans.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION
Alzheimer’s disease (AD) is a highly debilitating neurodegenerative disorder with significant
unmet medical need for diseaseꢀmodifying therapy.1 During the past decade, the pharmaceutical industry
has made significant investment on the advancement of potential diseaseꢀmodifying therapies as
evidenced by a number of clinical entries. However, the AD field as a whole has suffered from a high
rate of clinical attrition2 which may have been driven in part by the heterogeneous nature of patients
enrolled in large and expensive Phase II/III studies. In AD trials, patients have been typically recruited
under broad diagnostic categories, a practice that may have led to a dilution in signals of clinical efficacy
and inconclusive or negative trial outcomes. To help address this issue, clinical studies have recently
started to employ nonꢀinvasive positron emission tomography (PET) imaging agents to choose more
homogenous patient groups. As such, the use of PET ligands for Aβ plaques3 and neurofibrillary tau
tangles4 have become more routine in clinical trials, providing “PET imagingꢀbased phenotypic
signatures” and enabling better definition of optimal patient groups for lateꢀstage clinical trials. As part
ACS Paragon Plus Environment

Page 3 of 36
Journal of Medicinal Chemistry
of our BACE1 inhibitor program, we became interested in developing a BACE1ꢀselective PET ligand to
further enable clinical studies. A BACE1 PET ligand would serve as a valuable translational tool to help
clinical characterization of BACE1 inhibitors by measuring in vivo target occupancy and its relationship
to reduction of cerebrospinal fluid (CSF) Aβ levels. Furthermore, based on reports that BACE1 enzyme
level is elevated in AD brains5, specifically in presynaptic dystrophic neurites surrounding amyloid
plaques6, we were keen to understand if the brain BACE1 enzyme levels of AD patients could be
quantified by clinical PET imaging. Such information could in turn be combined with other biomarkers
(e.g., amyloid plaque and tau imaging) to ultimately provide a more comprehensive understanding of
AD etiology.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
SAR and Lead
Triaging guided
by CNS PET
Design
LC-MS/MS “Cold
Tracer” Study
for Specific
Binding
NHP PET Imaging,
Dosimetry and
GLP toxicity
studies for eIND
Filing
Brain Bmax and
Bio-
distribution
Determination
Parameters
Assessment
Figure 1. A rational and costꢀeffective CNS PET ligand discovery process.
Development of BACE1 inhibitors has been a major effort across the pharmaceutical industry for
over a decade, which has led to extensive literature disclosures featuring chemical matter with
significant diversity in structure and properties.7 Our BACE1 PET efforts started with a survey of
existing BACE1 inhibitor chemical matters from the literature and our internal effort. Identifying a
potential PET ligand from existing chemical matter was greatly facilitated by using our central nervous
system (CNS) PET ligand discovery process illustrated in Figure 1.8 As a first step in our PET ligand
discovery effort, we sought to determine the maximum concentration of target binding sites (Bmax) of
BACE1 in brain early in the process to inform PET viability. Favorable BACE1 brain expression would
trigger subsequent lead triaging, guided by a set of CNS PET design and selection criteria previously
disclosed by our group.9 Specifically, compounds with the following attributes would be prioritized:
high BACE1 potency and selectivity, favorable physicochemical properties as measured by CNS PET
multiꢀparameter optimization (MPO), fraction unbound in brain (Fu_b) > 0.05 for low risk of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 36
nonspecific binding (NSB), high passive permeability (RRCK Papp AB > 5 x10ꢀ6cm/s) and low Pꢀ
glycoprotein (Pꢀgp) transporter activity (MDR1 BA/AB ≤ 2.5) for brain permeability. PET ligand leads
emerging from the triaging process would be assessed via a liquid chromatographyꢀmass
spectroscopy/mass spectroscopy (LCꢀMS/MS) “cold tracer” method10 in rodents for in vivo specific
binding. Implementation of these steps would allow effective prioritization and make it possible to
advance only one or two best PET leads into subsequent nonꢀhuman primate (NHP) PET imaging and
dosimetry studies, ultimately enabling clinical imaging of BACE1. Herein, we describe our efforts in the
identification of [18F]PFꢀ06684511 ([18F]3) as a promising PET ligand with favorable brain uptake and
high specific binding to BACE1 in NHP, via effective utilization of PET ligand design parameters and
LCꢀMS/MS cold tracer method.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
BACE1 Brain Expression (Bmax) Determinations
As part of our initial efforts in developing a PET ligand for BACE1, we were keen to identify a
tool compound that could be tritiated and used for ex vivo saturation binding studies to determine
BACE1 Bmax in brain tissue. It was necessary for such a tool compound to be potent and selective for
BACE1 and have low NSB in brain tissue, but not necessarily be brain permeable. Compound 1, derived
from a chemotype previously reported by Eisai11, caught our attention. As shown in Figure 2, compound
1 exhibited good potency at BACE1 with a whole cell IC50 of 1.4 nM and a cellꢀfree fluorescence
polarization (FP) IC50 of 12 nM. While 1 did have limited selectivity (4.5x) over BACE212, a BACE1
homolog protein, it demonstrated remarkable selectivity over cathepsin D (CatD) and other targets. As
BACE2 is mainly expressed in peripheral tissues (e.g., kidney, pancreas, etc.) and has minimal
expressions centrally,13 we anticipated that the BACE2 activity of compound 1 would not significantly
interfere with the determination of BACE1 Bmax in brain. Properties such as low lipophilicity (eLogD =
1.1) and fraction unbound in brain (Fu_b = 0.06) suggested a low risk of NSB. Compound 1 displayed
ACS Paragon Plus Environment

Page 5 of 36
Journal of Medicinal Chemistry
moderate transporter activity (MDR1 BA/AB = 3.94), however it is fitꢀforꢀpurpose for ex vivo
saturation binding studies in brain tissue where brain permeability is not required. Structurally,
1
2
3
4
5
3
compound 1 bears a methoxy group amenable for tritiation. The Hꢀlabeling of 1 was readily achieved
6
7
8
via alkylation of the corresponding hydroxyl pyrazine precursor 5 with [3H]CH3I (Chemistry Section,
9
Scheme 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A)
B)
3000
2500
2000
1500
1000
2500
2000
1500
1000
Wildꢀtype
Heterozygous KO
Homozygous KO
Kd= 1.8 nM
Bmax= 172 fmol/mg protein
500
0
500
0
10
0
2
4
6
8
2
4
6
8
nM [3H]1
nM [3H]1
Figure 2. Profile of 1 and binding studies of [3H]1: (A) saturation binding of [3H]1 in brain frontal
cortex homogenates of wildꢀtype (WT), BACE1 heterozygous knockout (KO) and BACE1 homozygous
KO mice; (B) BACE1 expression (Bmax) and Kd determination in WT mice frontal cortex homogenate.
With [3H]1 in hand, we carried out saturation binding studies in mouse brain tissue to understand
the level of BACE1 brain expression. It is well known that BACE1 is expressed throughout various
brain regions.14 The mouse frontal cortex brain homogenate was used as a representative brain region
for Bmax determination. We first carried out binding studies with [3H]1 (vehicle + 3 dosing groups) in
wildꢀtype (WT) as well as BACE1 heterozygous and homozygous knockout (KO) mouse frontal cortex
tissues to confirm its BACE1 specificity in brain and minimum interference due to its BACE2 activity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 36
We were gratified to see that different levels of specific binding of [3H]1 indeed were observed (Figure
2A). The highest brain binding was seen in WT mice which was reduced to roughly 50% in
heterozygous BACE1 KO mice. Importantly, no specific binding was observed in homozygous BACE1
KO mice, confirming that even with comparable activity at BACE2, [3H]1 only labeled BACE1
specifically in brain, likely due to low brain expression of BACE2. These results boosted our
confidence in the Bmax value determined by this radiotracer as an accurate measure of BACE1
expression in brain. Subsequently, a thorough saturation binding (vehicle + 8 dosing groups) using WT
frontal cortex homogenate (Figure 2B) yielded a Kd value of 1.8 nM, comparable to the in vitro whole
cell (WC) potency, and a Bmax value of 144 fmol/mg protein (~7 nM assuming 50 mg protein/g wet
tissue) residing in a range favorable for PET consideration. This Bmax value suggested that a potent
compound, with WC IC50 most likely in the subꢀnM range, would be desirable to achieve a favorable
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
B
max/Kd ratio > 10 which is typically associated with successful PET ligands.15
PET Ligand Lead Triaging
ACS Paragon Plus Environment

Page 7 of 36
Journal of Medicinal Chemistry
pIC50 (BACE Whole Cell)
pIC50 (BACE Whole Cell)
9.5
1
9.5
9
2
3
Compound 3
9
4
5
6
8.5
8.5
7
8
9
8
7.5
7
8
7.5
7
Compound 2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
6.5
6.5
0
1
2
3
4
0
1
2
3
4
LogD
LogD
Color by RRCK:
Shape by MDR:
Filter Criteria:
> 5 x10ꢀ6 cm/s
≤ 5 x10ꢀ6 cm/s
BA/AB ≤ 2.5
BA/AB > 2.5
RRCK > 5 x10ꢀ6 cm/s
MDR BA/AB ≤ 2.5
BACE whole cell IC50 ≤ 5 nM
cFu_b > 0.05
790
BACEi’s
256
BACEi’s
32
BACEi’s
16
BACEi’s
2
leads
RRCK > 5
MDR ≤ 2.5
BACE
IC50 < 5 nM
cFu_b >
0.05
Labeling
moiety
Figure 3. BACE1 PET ligand triaging based on CNS PET ligand design parameters.
With sufficient BACE1 expression in brain confirmed, we turned our attention to identifying
suitable leads for PET consideration. Ideal candidates would retain the desired pharmacology, selectivity
and low NSB profile of 1 and in addition, have high passive permeability and low transporter activities
for better brain permeability. As shown in Figure 3, starting from roughly 800 BACE1 inhibitors
(literature leads and internal Pfizer compounds), we quickly narrowed down the list to 256 compounds
(represented by solid green circles) upon filtering by favorable MDR1 BA/AB (≤ 2.5) and RRCK (> 5
x10ꢀ6 cm/s) values. Among these, 32 compounds displayed BACE1 WC IC50 values less than 5 nM.
Further prioritization using calculated Fu_b16 (> 0.05) narrowed the list down to 16 compounds. From
this cohort, compounds 217 and 3 (PFꢀ06684511, a compound previously reported by Eisai18) were
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
11
chosen for further follow up as they both had structure moieties amenable to either Cꢀ or 18Fꢀlabeling.
The pharmacology and PETꢀspecific properties of compound 2 and 3 are shown in table 1. In terms of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 36
pharmacology, compound 3 was the more potent of the two with subꢀnM WC BACE1 IC50 and around
15ꢀfold selective over BACE2. Both compounds were highly selective against CatD and other targets in
broadꢀspectrum CEREP selectivity panel evaluations. Compounds 2 and 3 showed low lipophilicity
(LogD < 2) and resided in favorable CNS drug space. Compound 3 had a slightly lower CNS PET MPO
score, mainly due to a high number of Hꢀbond donors. Nevertheless, both compounds displayed aligned
in vitro ADME parameters for CNS PET ligand consideration, with low MDR1 effux, high passive
permeability and favorable calculated Fu_b (> 0.05) suggesting low risk of NSB.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Pharmacology and property profile of BACE1 PET ligand leads, 2 and 3 (PFꢀ06684511).
N
Compound
2
3
H
N
S
Pharmacology
O
F
H2N
N
BACE1 WC IC50 (nM)a
BACE1 FP IC50 (nM)b
BACE2 FP IC50 (nM)c
CatD FP IC50 (nM)d
4 ± 0.5
53 ± 7
0.7 ± 0.2
16 ± 2.3
239 ± 18
[
11C]
265 ± 32
F
[
18F]
13,600 ± 3,600 206,000 ± 13,000
2
Physicochemical Properties
F
CNS PET MPOe
eLogDf
3.9
1.2
3.0
1.4
H
S
O
H2N
N
O
F
In vitro PK Properties
11C]
RRCK Papp AB (10ꢀ6 cm/s)g
MDR BA/ABh
cFu_bi
17.7 ± 3
1.92 ± 0.8
0.13
17.5 ± 2
1.55 ± 0.7
0.08
[
HN
N
N
O
3
aIC50 values obtained from BACE1 whole cell (WC) assay; values represent the geometric mean of at
least three experiments with standard error.
bIC50 values obtained from BACE1 cell free fluorescence polarization (FP) assay; values represent the
geometric mean of at least three experiments with standard error.
cIC50 values obtained from BACE2 cell free FP assay; values represent the geometric mean of at least
three experiments with standard error.
dIC50 values obtained from CatD cell free FP assay; values represent the geometric mean of at least
three experiments with standard error.
eCNS PET MPO represents the summation of scores of six commonly used individual
physicochemical properties using the transformed functions described in ref 9.
fExperimentally measured LogD using a reverse phase HPLC method19.
gRRCK cells with low transporter activity were isolated from MardinꢀDarby canine kidney (MDCK)
cells and were used to estimate passive permeability.
ACS Paragon Plus Environment

Page 9 of 36
1
2
3
Journal of Medicinal Chemistry
hRatio from the MSꢀbased quantification of Papp,A→B and Papp,B→A of the test compound (1 µM) across
contiguous monolayers from MDR1ꢀtransfected MDCK cells;
I in silico calculated fraction unbound in brain.
4
5
6
7
8
9
Prompted by their overall favorable profiles, 2 and 3 were further assessed for brain permeability
in vivo. As shown in Table 2, both compounds demonstrated favorable measured fractions unbound in
brain and plasma, confirming good alignment of experimental and in silico predicted values. Indeed,
rodent neuroꢀpharmacokinetic studies showed favorable brain permeability with a free brain to plasma
ratio (Cb,u/Cp,u) of 0.53 for compound 2 and 0.45 for compound 3.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2: Neuropharmacokinetic (NeuroPK) profile of 2 and 3 in CDꢀ1 mice.
Cb,u/Cp,u
Fu_ba
mFu_pb Cp,u (hr*nM) Cb,u (hr*nM)
Compound
c
d
(AUC0–Tlast
)
(AUC0–Tlast)
0.11
0.06
0.51
0.25
1321
701
0.53
0.45
2
3
718
321
aFraction unbound in brain homogenate.20
bFraction unbound in mouse plasma.
cAUC0–Tlastꢀderived ratio of plasma unbound compound concentration in CDꢀ1 mice (10 mg/kg PO).
dAUC0–Tlastꢀderived ratio of brain unbound compound concentration in CDꢀ1 mice (10 mg/kg, PO).
Chemistry
Scheme 1a
aReagents and conditions: (a) 5ꢀhydroxypyrazineꢀ2ꢀcarboxylic acid, HBTU, DIPEA, DMF, rt, 34%;
(b) [3H]CH3I, Ag2CO3, CHCl3, rt; (c) TFA, DCM, rt, 80 Ci/mmol specific activity, >99% radiochemical
purity.
The tritiation of compound 1, illustrated in Scheme 1, started from a known aniline intermediate
411. Amide coupling with 5ꢀhydroxypyrazine 2ꢀcarboxylic acid yielded the hydroxyl pyrazine precursor
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 36
5. The subsequent methylation step was first explored using nonꢀradioactive methyl iodide with
anticipation of an Oꢀ versus Nꢀalkylation selectivity challenge. Typical methylation conditions using
cesium carbonate as base gave predominantly undesired Nꢀmethylation product. Changing the base to
silver carbonate yielded a 1:1 mixture of the desired compound 1 and the Nꢀmethylation product that
could be easily separated. Using this improved methylation condition and subsequent Boc removal,
[3H]1 was obtained in high radiochemical purity (>99%) and specific radioactivity (80 Ci/mmol).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2a
o
aReagents and conditions: (a) HCO2H (neat), 0 C to rt; (b) LiOH, THF/H2O (1/1 v/v), rt; (c)
o
o
perfluorobutanesulfonyl fluoride (NfF), Et3N.3HF, Et3N, THF, 0 C˗40 C, 52% for 3 steps; (d) NaN3,
o
N, N'ꢀdimethylcyclohexaneꢀ1, 2ꢀdiamine, sodium Lꢀascorbate, CuI, EtOH/H2O (3/1 v/v), 70 C, 72%;
o
(e) 5ꢀmethoxypyrazineꢀ2ꢀcarboxylic acid, HATU, DIPEA, CH3CN, rt, 93%; (f) DBU, MeOH, 60 C,
77%.
Our synthesis of 3 was accomplished from known intermediate 621 using modified procedures
from the literature (Scheme 2).22 The trityl group of 6 was removed through a twoꢀstep process
involving solvolysis in formic acid to afford formate ester 7 and subsequent hydrolysis to provide
alcohol 8. The fluoride group was then installed, at this early stage, by activation of alcohol 8 with
ACS Paragon Plus Environment

Page 11 of 36
Journal of Medicinal Chemistry
perfluorobutanesulfonyl fluoride (NfF) to give intermediate 9. Introduction of the amide functionality
was achieved by first converting the bromide of 9 to an azide via copper catalysis followed by
concomitant reduction mediated by Lꢀascorbate to afford aniline 10.23 The amide was then installed
under standard coupling conditions using HATU to afford 11. Benzoyl deprotection in basic MeOH
provided compound 3 with no amide cleavage observed.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3a
aReagents and conditions: (a) NaN3, N, N'ꢀdimethylcyclohexaneꢀ1, 2ꢀdiamine, sodium Lꢀascorbate,
o
CuI, EtOH/H2O (3.5/1 v/v), 70 C; (b) 5ꢀmethoxypyrazineꢀ2ꢀcarboxylic acid, propane phosphonic acid
o
anhydride (T3P), pyridine, DCM, rt, 41% over two steps ; (c) HCO2H (neat), 0 C to rt; (d) LiOH,
o
o
THF/H2O (1/1 v/v), rt, quantitative yield; (e) DBU, MeOH, 60 C; (f) Boc2O, Et3N, THF, 60 C, 58%
over two steps; (g) TsCl, DIPEA, DMAP, DCM, rt, 56%; (h) [18F]fluoride/kryptofix 222, DMSO, 150
o
oC, 5 min, followed by 1.5 M HCl, 100 C, 3 min, 4 and 7% radiochemical yield (> 99% purity), 187
and 417 MBq of delivered radioactivity and 84 and 175 GBq/µmol of specific radioactivity at the time
of injection.
18
For the preparation of the Fꢀlabeling precursor 18 (Scheme 3), we modified the synthesis used
for 3 to enable a lateꢀstage fluorination. As shown in scheme 3, copperꢀcatalyzed azide formation and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 36
subsequent in-situ reduction by Lꢀascorbate converted intermediate 6 to the corresponding aniline 12.
Introduction of the amide was accomplished in the presence of propylphosphonic anhydride (T3P) to
provide 13. In preparation for the lateꢀstage fluorination, the trityl group of 13 was removed in two steps
involving formic acid solvolysis and base catalyzed hydrolysis to give intermediate 15. To better
1
2
3
4
5
6
7
8
9
18
accommodate the short halfꢀlife of F radionuclide (T1/2 = 110 min), we next carried out a protecting
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
group exchange by first removing the benzoyl protecting group with DBU in MeOH to give 16,
followed by Bocꢀprotection under standard conditions to give compound 17 in 58% yield over two steps.
We expected that the eventual tꢀBoc removal would be mild and rapid under acidic conditions. Finally,
18
tosylation of the hydroxyl group provided the desired labeling precursor 18 in good yields. Fꢀlabeling
proceeded by nucleophilic displacement of the tosyl group using [18F]fluoride/kryptofix 222 in DMSO,
followed by rapid Bocꢀremoval by aqueous HCl. Upon purification by semiꢀpreparative HPLC, [18F]3
was generated in high radiochemical purity (> 99%) and molar activity (84 and 175GBq/µmol at the
time of injection).
LCꢀMS/MS “Cold Tracer” Study in Rodents
As both compounds 2 and 3 demonstrated suitable PET properties and good brain permeability, we
sought an early read for BACE1 specific binding to differentiate these two leads prior to triggering
radiochemistry and PET imaging. Toward this end, we utilized a LCꢀMS/MS based “cold tracer”
method10, in which nonꢀradioactive 2 and 3 were dosed intravenously in mice at tracer level (10 µg/kg).
Concentrations in four brain regions (cerebellum, frontal cortex, hippocampus and striatum) and plasma
were quantified. As there is no reference region due to the widespread expression of BACE1 throughout
the brain, we performed studies with baseline (vehicle) and blocking (pretreatment of a BACE1
selective inhibitor) arms to determine whether compound concentrations in the four main brain regions
could be reduced as an indication of BACE1 specific binding. The results of the LCꢀMS/MS studies are
illustrated in Figure 4. As shown in Figure 4A, no significant reduction in the concentration of 2 in the
brain regions examined other than frontal cortex could be observed after pretreatment with a high dose
ACS Paragon Plus Environment

Page 13 of 36
Journal of Medicinal Chemistry
of BACE1 inhibitor A24 (100 mg/kg, SC), suggesting a lack of in vivo specific binding. In contrast, the
concentration of 3 was dose dependently reduced in all four brain regions upon pretreatment of BACE1
inhibitor B25 (1, 10, 80 mg/kg, SC, Figure 4B). To further validate that the specific binding of 3 was
BACE1ꢀdriven, we further examined the distribution of 3 in WT and BACE1 KO mice following the
same protocol (Figure 5). We were gratified to see significant reduction in the concentration of 3 across
the four brain regions of BACE1 KO mice compared to WT mice. These data indicated that compound 3
displayed the desired BACE1 specific binding in vivo and was a good candidate for further PET
evaluation. While 2 had many desirable properties that met our PET ligand criteria, its potency may
have not been sufficient to achieve specific binding in vivo.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
B
Figure 4. In vivo specific binding assessment of 2 and 3 in male 129/sve mice using LCꢀMS/MS “cold
tracer” method: (A) distribution of 2 in cerebellum, frontal cortex, hippocampus, striatum, and plasma,
pretreated with either vehicle or BACE1 inhibitor A22 (100 mg/kg, SC) 60 min prior to ligand
administration; (B) distribution of 3 in cerebellum, frontal cortex, hippocampus, striatum, and plasma,
pretreated with either vehicle or BACE1 inhibitor B23 (1, 10, 80 mg/kg, SC) 60 min prior to ligand
administration. For both studies, brain tissues and plasma were collected at 45 min post administration
(n = 5 per animal group) at which the highest brain tissue/plasma ratios were achieved for 2 and 3 (see
supporting information, section 3).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 36
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. In vivo specific binding assessment of 3 in BACE1 WT and KO mice, using LCꢀMS/MS
“cold tracer” method: four brain regions (cerebellum, frontal cortex, hippocampus, and striatum) and
plasma were collected at 45 min post administration (n = 5 per animal group).
NHP PET Imaging
The favorable binding profile of 3 in the LCꢀMS/MS “cold tracer” experiment prompted us to
advance this lead to subsequent radiochemistry and PET imaging studies. Radiofluorination of 3 was
achieved in high radiochemical purity and delivered radioactivity as illustrated in Scheme 3. The
specific binding of [18F]3 was evaluated in NHP by PET under both baseline and blocking conditions. A
total of two PET scans (one female NHP, one baseline scan and one blocking scan) were performed with
injected radioactivity of 154 and 117 MBq, specific radioactivities at the time of injection of 84 and 175
GBq/µmol, and injected mass of 0.8 and 0.3 µg. The summed PET images of baseline and blocking
conditions, as well as the coꢀregistered MRI image, of the NHP are shown in Figure 6. Consistent with
LCꢀMS/MS study results, [18F]3 demonstrated good brain uptake and widespread distribution in various
brain regions, with enriched signal in hippocampus and striatum (Figure 6B). The peak of the brain
uptake reached at approximately 20 min and reduced to half at 120 min under baseline condition. In
addition, upon administration of a BACE1 inhibitor (LY288672126, 5 mg/kg, PO, 2 hour prior to PET
ACS Paragon Plus Environment

Page 15 of 36
Journal of Medicinal Chemistry
scanning), the radio signal of [18F]3 decreased significantly (Figure 6C) across different brain regions
1
2
3
(Figure 7), demonstrating BACE1 specificity.
4
5
6
7
8
A
B
C
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
0
SUV
Figure 6. MRI (A) and PET images of [18F]3 in NHP summed from 32 min to 180 min under (B)
baseline and (C) blocking (pretreatment with a BACE1 inhibitor, LY2886721, 5 mg/kg, PO) conditions.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 36
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
SUV
32-180min
CB ACC FC
TC CAU PUT THA OC PC AMY HIP
baseline 1.1 2.0 1.5 1.7 1.6 1.8 1.9 1.4 1.5 2.2 2.1
blocking 0.3 0.4 0.4 0.4 0.4 0.5 0.4 0.4 0.4 0.5 0.5
Figure 7. Standardized uptake values (SUVs) of [18F]3 in representative brain regions under baseline
and blocking conditions. CB: cerebellum, ACC: anterior cingulate cortex, FC: frontal cortex, TC:
temporal cortex, CAU: caudate, PUT: putamen, THA: thalamus, OC: occipital cortex, PC: parietal
cortex, AMY: amygdala, HIP: hippocampus
CONCLUSIONS
BACE1 is an aspartyl protease that plays a key role in the generation of Aβ fragments via
extracellular cleavage of the amyloid precursor protein (APP). BACE1 expression level has been
reported to be elevated in AD patients. However, no BACE1 specific PET ligand has been reported to
date which allows in vivo quantification of BACE1 expression and target occupancy measurement.
Through an efficient and costꢀeffective CNS PET ligand discovery process, we have successfully
identified [18F]3 as a novel PET ligand specific to BACE1 in brain.
As a first step, we determined Bmax in brain tissue using an earlier nonꢀbrain penetrant radiotracer
[3H]1 and established the PET viability of this target. The PET ligand selection criteria previously
reported by our group was applied to narrow an extensive collection of literature and inꢀhouse BACE1
ACS Paragon Plus Environment

Page 17 of 36
Journal of Medicinal Chemistry
inhibitors to just two leads, 2 and 3, based on key pharmacological, physicochemical, and ADME
1
2
3
4
5
6
7
8
9
11
18
properties as well as structural amenability to Cꢀ or Fꢀ labeling. Prior to PET radiolabeling, these
two leads were advanced into rodent LCꢀMS/MS “cold tracer” studies for specific binding evaluation,
from which compound 3 emerged as a promising PET ligand lead with the desired level of BACE1
specific binding. The favorable outcome from the LCꢀMS/MS “cold tracer” studies successfully
translated to NHP PET imaging where [18F]3 demonstrated excellent brain uptake and high BACE1
specific binding under baseline and blocking conditions. These results confirmed [18F]3 as a suitable
PET ligand for BACE1 PET imaging in vivo and reinforced the value of our overall PET discovery
approach. This and previous work from our laboratories demonstrate that we have put in place a
rational and costꢀefficient paradigm for the development of novel PET ligands that can successfully
identify a small number of candidate ligands (n = 1ꢀ3) from a large set of compounds. 9, 27
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We envision that BACE1 PET ligand [18F]3 could facilitate the development of AD therapeutics in
multiple ways. First, successful translation of this PET ligand to the clinic would enable target
occupancy measurement of clinical BACE1 inhibitors and relationship of target occupancy to reduction
of Aβ levels in CSF. Furthermore, [18F]3 could enable the examination of brain BACE1 enzyme levels
in AD patients across the disease continuum. A robust correlation of BACE1 level with AD progression
would render this BACE1 PET ligand as a valuable AD biomarker, which in turn could be used in
conjunction with other established imagingꢀbased AD biomarkers, such as amyloid plaque and tau PET
imaging as well as hippocampus volume, to enable a deeper understanding of AD etiology. Based on
the favorable outcome of the first NHP PET imaging studies, [18F]3 has been investigated for detailed
quantitative analysis and a target occupancy study in NHPs. It has also been advanced to testꢀretest and
dosimetry studies in humans. The results from these studies will be communicated in future
publications.
Experimental Procedures
Biology.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 36
The sAPPβ wholeꢀcell assay, BACE1 enzyme cellꢀfree assay, BACE2 enzyme cellꢀfree assay and
1
2
3
cathepsin D enzyme cellꢀfree assay were performed following the protocols that were previously
4
5
described28.
6
7
8
Pharmacokinetics.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NeuroPK study in male CDꢀ1 mice was carried out following the same protocol described in reference
17. The unbound fraction in mixed gender mouse (CDꢀ1) plasma (Fu_p) was determined using the
Rapid Equilibrium Device (RED, Thermo Scientific). A spiked mice plasma (2 µM) sample was
prepared and added (220 µL) to the donor side, followed by addition of buffer (350 µL) to the receiver
side. The RED device was then incubated at 37oC for 4 hours. Following the incubation, plasma (15
µL) and buffer (45 µL) were sampled and mixed with the appropriate blank matrix to prepare mixed
matrix samples. Samples were protein precipitated with acetonitrile containing internal standard (200
µL), vortexed and centrifuged. Supernatants were analyzed using an LCꢀMS/MS assay. The buffer
dilution factor was 0.33 (15 µL/45 µL). Fu_p was then calculated using: (concentration in buffer
sample *0.33) / (concentration in plasma sample). The unbound fraction in brain tissue homogenate
(Fu_b) was determined using the 96ꢀwell equilibrium dialysis method described by Kalvass et al.29 with
the following exceptions. Brain homogenates were prepared from freshly harvested rat whole brains
(Wistar han) following dilution with a 10ꢀfold volume of phosphate buffer and spiked with 2 ꢁM test
compound. The homogenates were dialyzed against an equal volume (150 ꢁL) of phosphate buffer at 37
°C for 6 h. Following the incubation, equal volumes (50 ꢁL) of brain homogenate and buffer samples
were collected and mixed with 50 ꢁL of buffer or control homogenate, respectively, for preparation of
mixed matrix samples. All samples were then precipitated with internal standard in acetonitrile (200
ꢁL), vortexed, and centrifuged. Supernatants were analyzed using an LC–MS/MS assay. A dilution
factor of 10 was applied to the calculation of brain fraction unbound.
ACS Paragon Plus Environment

Page 19 of 36
Journal of Medicinal Chemistry
Saturation Binding: The saturation binding of [3H]1 in mouse brain tissue was performed using
filtration separation of ligand bound in crude tissue homogenate. The source of tissue was frontal cortex
from wild type, heterozygous and homozygous BACE1 KO mice. The [3H]1 was titrated from 0.1 to 10
nM and incubated for 60 min at rt with 2.4 mg of tissue homogenate (homogenized at 12 mg/mL on ice
for 8 seconds using polytron setting 5) in 250 µL of 50 mM Tris, pH 7.4 containing 2 mM MgCl2.
Following incubation, samples were separated over GF/B filters that were preꢀsoaked in 0.1%
polyethyleneimine followed by 3 washes with cold assay buffer. The amount of specific binding was
calculated as the counts bound in the presence of 0.1% DMSO vehicle minus the counts bound in the
presence of excess BACE1 inhibitor (10 µM).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chemistry
General Methods: All solvents and reagents were of reagent grade and were used as supplied by the
manufacturer. All reactions were run under a N2 atmosphere. Organic extracts were routinely dried over
anhydrous Na2SO4 or MgSO4. Concentration in vacuo refers to rotary evaporation under reduced
pressure. Chromatography refers to flash chromatography using disposable RediSepRf 4 to 120 g silica
columns or Biotage disposable columns on a CombiFlash Companion or Biotage Horizon automatic
purification system. Purification by massꢀtriggered HPLC was carried out using Waters XTerra PrepMS
C18 columns, 5 µm, 30 mm x100 mm. Compounds were preꢀsalted as TFA salts and diluted with 1 mL
of DMSO. Samples were purified by massꢀtriggered collection using a mobile phase of acetonitrile and
0.1% TFA in water, with a gradient of 100% aqueous to 100% acetonitrile over 10 min at a flow rate of
20 mL/min. All target compounds were analyzed using ultra high performance liquid
chromatography/ultraviolet/evaporative light scattering detection coupled to time of flight mass
spectrometry (UHPLC/UV/ELSD/TOFMS). LCꢀMS analyses were performed on a Waters Acquity
UPLCꢀMS system with a Waters Aquity HSS T3, 1.7 µm C18 column (50 mm x 2.1 mm). UPLC
conditions: mobile phase A = 0.1% formic acid in water (v/v), mobile phase B = 0.1% formic acid in
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 36
acetonitrile; flow rate = 1.25 mL/min; compounds were eluted with a gradient of 5% B/A to 95% B/A
for 1.1 min. All tested compounds were found to be > 95% pure by this method.
1
2
3
4
5
6
Nꢀ(3ꢀ((4aS,5S,7aS)ꢀ2ꢀaminoꢀ5ꢀ(fluoromethyl)ꢀ4a,5ꢀdihydroꢀ4Hꢀfuro[3,4ꢀd][1,3]thiazinꢀ7a(7H)ꢀyl)ꢀ
7
8
9
4ꢀfluorophenyl)ꢀ5ꢀmethoxypyrazineꢀ2ꢀcarboxamide (3)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To a stirring mixture of 11 (360 mg, 0.667 mmol) in anhydrous MeOH (6.5 mL) was added DBU (71.3
o
mg, 0.468 mmol, 0.070 mL) at rt. The reaction mixture was stirred at 60 C for 6 h during which time a
o
white solid formed. The reaction was stirred at 60 C for an additional 2.5 h then at rt overnight. The
resulted white solid was collected by filtration, rinsed with cold MeOH (3 mL) then dried in vacuo to
afford 3 (224 mg, 77% yield) as a white solid. HRMS: Calculated for C19H20F2N5O3S [M+H]+:
436.1249, found 436.1256; Anal. Calcd for C19H19F2N5O3S: C, 52.41; H, 4.40; N, 16.08. Found: C,
52.16; H, 4.35; N, 15.93; [α]D23.0 = +121.9° (c 0.36, CHCl3); 1H NMR (400 MHz, DMSOꢀd6) δ 10.56 (s,
1H), 8.89 (d, J = 1.3 Hz, 1H), 8.40 (d, J = 1.3 Hz, 1H), 7.89 (dd, J = 7.3, 2.6 Hz, 1H), 7.84 (ddd, J = 8.6,
4.0, 2.8 Hz, 1H), 7.15 (dd, J = 11.9, 8.8 Hz, 1H), 6.03 (s, 2H), 4.68 – 4.43 (m, 2H), 4.34 (dddd, J = 23.5,
7.7, 5.1, 2.4 Hz, 1H), 4.24 (d, J = 8.2 Hz, 1H), 4.02 (s, 3H), 3.79 (dd, J = 8.1, 3.2 Hz, 1H), 3.13 (dd, J =
13
13.5, 3.6 Hz, 1H), 3.00 (dd, J = 13.5, 3.8 Hz, 1H), 2.75 (dt, J = 7.8, 3.7, 3.7 Hz, 1H). C NMR (101
MHz, DMSOꢀd6) δ 161.63 , 161.45 , 155.99 (d, J = 243.6 Hz), 148.36 , 141.53 , 137.92 , 134.37 (d, J =
2.3 Hz), 133.53 , 130.14 (d, J = 12.4 Hz), 122.12 (d, J = 4.1 Hz), 120.91 (d, J = 8.4 Hz), 116.16 (d, J =
25.0 Hz), 83.56 (d, J = 169.7 Hz), 78.52 (d, J = 18.2 Hz), 77.53 (d, J = 6.1 Hz), 65.80 – 65.36 (m),
54.29 , 35.51 (dd, J = 5.1, 2.4 Hz), 22.97 .
tertꢀButyl ((4aR,8aS)ꢀ8aꢀ(2ꢀfluoroꢀ5ꢀ(5ꢀhydroxypyrazineꢀ2ꢀcarboxamido)phenyl)ꢀ4,4a,5,6,8,8aꢀ
hexahydropyrano[3,4ꢀd][1,3]thiazinꢀ2ꢀyl)carbamate (5)
To a stirring solution of 5ꢀhydroxypyrazineꢀ2ꢀcarboxylic acid (74.7 mg, 0.53 mmol) in DMF (2 mL)
were added DIPEA (116 uL, 0.67 mmol), HBTU (230 mg, 0.58 mmol) and 411 (170 mg, 0.45 mmol) in
ACS Paragon Plus Environment

Page 21 of 36
Journal of Medicinal Chemistry
that order. The reaction was flushed with nitrogen and stirred at rt for 3 h. The reaction was diluted with
EtOAc and saturated aqueous NaHCO3 solution and extracted. The organic extract was washed with
water and brine then dried over Na2SO4 and concentrated to afford the crude product. Purification via
flash column chromatography (silica gel, gradient 50% ꢀ 100% EtOAc in heptanes) afforded 5 (76 mg,
34%) as an offꢀwhite solid. LCꢀMS m/z 504.3 [M + H]+. 1H NMR (400 MHz, CD3OD) δ 8.13 (d, J = 1.3
Hz, 1H), 8.08 (d, J = 1.3 Hz, 1H), 7.84 (ddd, J = 2.7, 4.2, 8.8 Hz, 1H), 7.68 (dd, J = 2.6, 7.1 Hz, 1H),
7.18 (dd, J = 8.9, 12.2 Hz, 1H), 4.07 (td, J = 3.9, 11.3, 11.9 Hz, 2H), 3.78 (d, J = 11.8 Hz, 1H), 3.70 (td,
J = 2.2, 12.2, 12.2 Hz, 1H), 3.14 – 3.07 (m, 1H), 2.98 (dd, J = 3.8, 13.0 Hz, 1H), 2.74 (dd, J = 3.0, 12.9
Hz, 1H), 2.15 – 2.03 (m, 1H), 1.62 (ddt, J = 1.6, 1.6, 3.3, 13.0 Hz, 1H), 1.52 (s, 9H).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
((4aS,5S,7aS)ꢀ2ꢀBenzamidoꢀ7aꢀ(5ꢀbromoꢀ2ꢀfluorophenyl)ꢀ4a,5,7,7aꢀtetrahydroꢀ4Hꢀfuro[3,4ꢀ
d][1,3]thiazinꢀ5ꢀyl)methyl formate (7)
A mixture of 619 (10.0 g, 14.131 mmol) in formic acid (70 mL) was stirred at r.t. overnight. LC/MS
showed 6 was consumed completely and the major peak was the desired formate ester. The solid was
removed by filtration and the filtrate was washed with petroleum ether (75 mL×3). The formic acid layer
was concentrated to give crude 7 (7.00 g) as yellow oil which was carried directly into the next reaction.
Nꢀ((4aS,5S,7aS)ꢀ7aꢀ(5ꢀbromoꢀ2ꢀfluorophenyl)ꢀ5ꢀ(hydroxymethyl)ꢀ4a,5,7,7aꢀtetrahydroꢀ4Hꢀ
furo[3,4ꢀd][1,3]thiazinꢀ2ꢀyl)benzamide (8)
A mixture of the crude formate ester 7 (7.00 g, 14.19 mmol) and LiOH (1.70 g, 70.9 mmol) in THF/H2O
(100 mL, 1/1 v/v) was stirred at rt for 1 h. LC/MS showed the reaction was complete. The organic layer
was removed in vacuo and water (50 mL) was added. The mixture was extracted with DCM (150
mL×3). The organic layer was washed with brine solution (100 mL), dried, filtered and concentrated to
afford crude 8 (5.60 g) as a light yellow solid, which was carried directly into the next reaction without
further purification.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 36
Nꢀ((4aS,5S,7aS)ꢀ7aꢀ(5ꢀbromoꢀ2ꢀfluorophenyl)ꢀ5ꢀ(fluoromethyl)ꢀ4a,5,7,7aꢀtetrahydroꢀ4Hꢀfuro[3,4ꢀ
d][1,3]thiazinꢀ2ꢀyl)benzamide (9)
1
2
3
4
5
6
7
8
To a stirred solution of the crude primary alcohol 8 (5.60 g, 12.03 mmol) in dry THF (150 mL) was
added triethylamine (12.2 g, 120 mmol), triethylamine trihydrofluoride (6.6 g, 40.9 mmol) and
perfluorobutane sulfonylfluoride (12.4 g, 40.9 mmol) in that order at 0℃ under nitrogen. After addition,
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
the mixture was stirred at 30 C – 40 ℃ for 3 h. TLC (petroleum ether: EtOAc = 3:1) showed the
reaction was complete. The mixture was poured into iceꢀwater (200 mL), saturated aqueous NaHCO3
(200 mL) was added and the mixture was extracted with EtOAc (200 mL×2). The organic layer was
washed with brine solution (150 mL×2), dried over Na2SO4, filtered and concentrated to give a residue,
which was purified by flash column chromatography (silica gel; gradient 0% ꢀ 30% EtOAc in
petroleum ether) to afford 9 (3.40 g, 60.5%) as a white solid. LCꢀMS m/z 468.7 [M + H]+. H NMR
1
(400 MHz, CDCl3) δ 8.13 – 7.99 (m, 2H), 7.60 – 7.41 (m, 5H), 7.03 (dd, J = 8.7, 11.7 Hz, 1H), 4.73 –
4.53 (m, 3H), 4.52 – 4.46 (m, 1H), 4.03 – 3.95 (m, 1H), 3.36 – 3.28 (m, 1H), 3.27 – 3.17 (m, 1H), 2.85 –
2.76 (m, 1H).
Nꢀ((4aS,5S,7aS)ꢀ7aꢀ(5ꢀaminoꢀ2ꢀfluorophenyl)ꢀ5ꢀ(fluoromethyl)ꢀ4a,5,7,7aꢀtetrahydroꢀ4Hꢀfuro[3,4ꢀ
d][1,3]thiazinꢀ2ꢀyl)benzamide (10)
To a solution of 9 (865mg, 1.81 mmol) in absolute EtOH (35 mL) was added NꢀN'ꢀ
dimethylcyclohexaneꢀ1,2ꢀdiamine (300 mg, 2.11 mmol), followed by addition of sodium azide (1.20 g,
18.46 mmol) and a solution of sodium Lꢀascorbate (185 mg, 0.93 mmol) in water (10 mL) at rt. The
mixture was stirred under nitrogen for 15 min, then copper (I) iodide (175 mg, 0.92 mmol) was added
o
and the mixture was heated at 70 C for 2 hrs. LC/MS showed the reaction was complete. The mixture
was concentrated in vacuo then partitioned between brine solution (50 mL) and EtOAc (100 mL). The
phases were separated and the aqueous layer was extracted three times with additional EtOAc. The
combined organic layers were dried over Na2SO4 filtered and concentrated to give crude product which
was purified by flash column chromatography (silica gel, gradient 0% ꢀ 100% EtOAc in heptanes) to
ACS Paragon Plus Environment

Page 23 of 36
Journal of Medicinal Chemistry
afford 10 (524mg, 72%) as a white solid. LCꢀMS m/z 404.1 [M + H]+. H NMR (400 MHz, CDCl3) δ
12.03 (s, 1H), 8.17 (d, J = 7.1 Hz, 2H), 7.56 – 7.50 (m, 1H), 7.45 (t, J = 7.5, 7.5 Hz, 2H), 6.92 (dd, J =
8.7, 11.9 Hz, 1H), 6.70 (dd, J = 2.6, 6.5 Hz, 1H), 6.61 (dt, J = 3.3, 3.3, 8.5 Hz, 1H), 4.74 – 4.49 (m, 4H),
4.08 – 3.98 (m, 1H), 3.69 (s, 2H), 3.46 – 3.38 (m, 1H), 3.29 (dd, J = 3.4, 13.8 Hz, 1H), 2.81 (dd, J = 3.6,
13.8 Hz, 1H).
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nꢀ(3ꢀ((4aS,5S,7aS)ꢀ2ꢀbenzamidoꢀ5ꢀ(fluoromethyl)ꢀ4a,5ꢀdihydroꢀ4Hꢀfuro[3,4ꢀd][1,3]thiazinꢀ7a(7H)ꢀ
yl)ꢀ4ꢀfluorophenyl)ꢀ5ꢀmethoxypyrazineꢀ2ꢀcarboxamide (11)
To a solution of 5ꢀmethoxypyrazineꢀ2ꢀcarboxylic acid (120 mg, 0.78 mmol) and 10 (300 mg, 0.74
mmol) in acetonitrile (10 mL) was added DIPEA (300 µL, 1.72 mmol) followed by HATU (300 mg,
0.79 mmol) at rt. The mixture was stirred at rt for 30 min. LCꢀMS showed the reaction was completed.
The mixture was concentrated in vacuo, diluted with EtOAc (40 mL), then washed with water (15 mL)
and brine solution (15 mL). The organic extract was dried over dried over Na2SO4 filtered and
concentrated to give crude product which was purified by flash column chromatography (24 g silica gel,
gradient 0% ꢀ 100% EtOAc in heptanes) to afford Compound 11 (374mg, 93%) as a white solid. LCꢀ
MS m/z 540.1 [M + H]+. 1H NMR (400 MHz, CDCl3) δ 12.66 (s, 1H), 9.56 (s, 1H), 9.00 (s, 1H), 8.25 –
8.05 (m, 3H), 7.91 (d, J = 7.7 Hz, 1H), 7.72 – 7.64 (m, 1H), 7.57 – 7.50 (m, 1H), 7.46 (t, J = 7.3, 7.3 Hz,
2H), 7.18 (dd, J = 8.9, 11.5 Hz, 1H), 4.76 – 4.54 (m, 3H), 4.51 (d, J = 9.3 Hz, 1H), 4.12 (q, J = 7.1, 7.1,
7.1 Hz, 1H), 4.06 (s, 3H), 3.50 – 3.32 (m, 2H), 2.94 – 2.81 (m, 1H).
Nꢀ((4aS,5S,7aS)ꢀ7aꢀ(5ꢀaminoꢀ2ꢀfluorophenyl)ꢀ5ꢀ((trityloxy)methyl)ꢀ4a,5,7,7aꢀtetrahydroꢀ4Hꢀ
furo[3,4ꢀd][1,3]thiazinꢀ2ꢀyl)benzamide (12)
To a solution of compound 619 (7.08 g, 10.00 mmol) in EtOH (210 mL) was added NꢀN’ꢀ
dimethylcyclohexaneꢀ1,2ꢀdiamine (1.71 g, 12.0 mmol), followed by addition of solid sodium azide (6.18
g, 95.06 mmol) and a solution of sodium lꢀascorbate (931 mg, 4.70 mmol) in water (60 mL) at rt under a
nitrogen atmosphere. Then copper (I) iodide (895 mg, 4.70 mmol) was added and the mixture was
o
heated at 70 C for 2 h. The mixture was diluted with water (300 mL) and EtOAc (200 mL) and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 36
extracted with additional EtOAc (100 mL x 2). The combined organic layers were washed with brine
solution (200 mL), dried over Na2SO4, filtered and concentrated to give crude aniline 12 (6.44g, 72%
purity) as yellow oil, which was used directly in the next step without further purification. LCꢀMS m/z
666.1 [M + Na]+.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nꢀ(3ꢀ((4aS,5S,7aS)ꢀ2ꢀbenzamidoꢀ5ꢀ((trityloxy)methyl)ꢀ4a,5ꢀdihydroꢀ4Hꢀfuro[3,4ꢀd][1,3]thiazinꢀ
7a(7H)ꢀyl)ꢀ4ꢀfluorophenyl)ꢀ5ꢀmethoxypyrazineꢀ2ꢀcarboxamide (13)
To a mixture of the aniline 12 (6.44 g, ~72% purity, 7.2 mmol) and commerciallyꢀavailable 5ꢀ
methoxypyrazineꢀ2ꢀcarboxylic acid (1.33 g, 8.64 mmol) in DCM (100 mL) was added pyridine (2.28 g,
28.8 mmol) at rt, followed by addition of a solution of propylphosphonic anhydride solution (a.k.a. T3P,
50% by wt. in EtOAc, 13.8 g, 21.6 mmol) dropwise at rt and stirred at that temperature for 16 h. The
mixture was quenched with water (200 mL) and extracted with DCM (2 x 100 mL). The combined
organic layers were washed with brine solution (300 mL), dried and concentrated to give crude product,
which was purified by flash column chromatography on silica gel (40 g silica gel, 0% ꢀ15% EtOAc in
petroleum ether) to afford 13 (2.3 g, 41% over two steps) as a yellow semiꢀsolid. LCꢀMS m/z 780.2 [M
1
+ H]+. H NMR (400 MHz, CDCl3) δ 9.54 (s, 1H), 9.00 (s, 1H), 8.20 – 8.14 (m, 2H), 8.13 (s, 1H), 8.01
– 7.95 (m, 1H), 7.56 (d, J = 5.2 Hz, 1H), 7.53 – 7.50 (m, 1H), 7.49 – 7.42 (m, 7H), 7.35 – 7.29 (m, 5H),
7.28 – 7.21 (m, 5H), 7.16 (dd, J = 9.2, 11.7 Hz, 1H), 4.62 – 4.52 (m, 2H), 4.12 – 4.09 (m, 1H), 4.06 (s,
3H), 3.48 – 3.42 (m, 1H), 3.41 – 3.31 (m, 2H), 3.28 – 3.21 (m, 1H), 2.71 (dd, J = 3.4, 13.9 Hz, 1H).
((4aS,5S,7aS)ꢀ2ꢀBenzamidoꢀ7aꢀ(2ꢀfluoroꢀ5ꢀ(5ꢀmethoxypyrazineꢀ2ꢀcarboxamido)phenyl)ꢀ4a,5,7,7aꢀ
tetrahydroꢀ4Hꢀfuro[3,4ꢀd][1,3]thiazinꢀ5ꢀyl)methyl formate (14)
o
Formic acid (30 mL) was added to 13 (2.30 g, 2.95 mmol) neat at 0 C. After the addition, the mixture
was warmed to rt and stirred at ambient temperature for 16h. TLC (petroleum ether/ EtOAc =3/1)
showed the reaction was complete. The mixture was concentrated in vacuo to give the corresponding
formate ester 14 (1.67 g, crude, quant.) as a yellow solid, which was used directly in the next step.
ACS Paragon Plus Environment

Page 25 of 36
Journal of Medicinal Chemistry
Nꢀ(3ꢀ((4aS,5S,7aS)ꢀ2ꢀbenzamidoꢀ5ꢀ(hydroxymethyl)ꢀ4a,5ꢀdihydroꢀ4Hꢀfuro[3,4ꢀd][1,3]thiazinꢀ
7a(7H)ꢀyl)ꢀ4ꢀfluorophenyl)ꢀ5ꢀmethoxypyrazineꢀ2ꢀcarboxamide (15)
1
2
3
4
5
6
7
8
To a solution of crude formate ester 14 (1.67 g, 2.95 mmol) in THF/water (40 mL, 1/1 v/v) was added
LiOH (141 mg, 5.91 mmol) at rt and stirred at the same temperature for 20 min. TLC (pet ether / EtOAc
= 1:1, UV) showed the reaction was complete. The reaction was diluted with water (100 mL) and
extracted with EtOAc (30mL x 2). The organic layer was washed with brine solution (50 mL x 2), dried
over Na2SO4, filtered and concentrated afford compound 15 (1.59 g, quant. over two steps) as a yellow
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
semiꢀsolid. LCꢀMS m/z 538.0 [M+H]+. H NMR (400 MHz, CDCl3) δ 9.55 (s, 1H), 8.99 (s, 1H), 8.20 –
8.10 (m, 3H), 7.91 – 7.84 (m, 1H), 7.71 – 7.65 (m, 1H), 7.54 – 7.49 (m, 1H), 7.48 – 7.41 (m, 2H), 7.16
(dd, J = 8.8, 11.7 Hz, 1H), 4.57 – 4.52 (m, 1H), 4.50 (d, J = 9.6 Hz, 1H), 4.18 – 4.11 (m, 1H), 4.06 (s,
3H), 3.96 (dd, J = 2.6, 11.8 Hz, 1H), 3.75 (dd, J = 3.5, 11.5 Hz, 1H), 3.44 – 3.33 (m, 2H), 2.87 (dd, J =
2.1, 14.1 Hz, 1H), 2.04 (s, 1H).
Nꢀ(3ꢀ((4aS,5S,7aS)ꢀ2ꢀaminoꢀ5ꢀ(hydroxymethyl)ꢀ4a,5ꢀdihydroꢀ4Hꢀfuro[3,4ꢀd][1,3]thiazinꢀ7a(7H)ꢀ
yl)ꢀ4ꢀfluorophenyl)ꢀ5ꢀmethoxypyrazineꢀ2ꢀcarboxamide (16)
To a solution of compound 15 (700 mg, 1.30 mmol) in MeOH (25 mL) was added DBU (1.98 g, 13.0
o
mmol) at rt. The reaction mixture was stirred at 60 C for 3 h during which time a white solid formed
and LC/MS showed the reaction was complete. The reaction was filtered, and the white solid was dried
to afford crude compound 16 (500 mg, 80% purity by LCꢀMS) which was advanced asꢀis into the next
step. LCꢀMS m/z 456.0 [M + Na]+.
tertꢀButyl ((4aS,5S,7aS)ꢀ7aꢀ(2ꢀfluoroꢀ5ꢀ(5ꢀmethoxypyrazineꢀ2ꢀcarboxamido)phenyl)ꢀ5ꢀ
(hydroxymethyl)ꢀ4a,5,7,7aꢀtetrahydroꢀ4Hꢀfuro[3,4ꢀd][1,3]thiazinꢀ2ꢀyl)carbamate (17)
To a suspension of compound 16 (500 mg, 1.15 mmol) in tetrahydrofuran (200 mL) was added Boc2O
o
(378 mg, 1.73 mmol) and triethylamine (233 mg, 2.31 mmol) at rt. The suspension was stirred at 60 C
for 16 h. The mixture was concentrated to afford the crude product (0.95 g) which was subjected to flash
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 36
column chromatography (4 g silica gel, 0% ꢀ 5% in DCM) followed by reꢀpurification via preparative
HPLC (Column: Agella Durashell C18 250mmx21.2mm 5µm particle size; Mobile phase: from
36%ꢀ56% acetonitrile in aqueous ammonia (pH 10) to 56% acetonitrile in ammonia). The pure
preparative HPLC fractions were concentrated and lyophilized to afford compound 17 (400 mg, 58%
over 2 steps) as a white solid. LCꢀMS m/z 556.1 [M + Na]+. 1H NMR (400 MHz, CDCl3) δ 9.53 (s, 1H),
9.01 (d, J = 1.3 Hz, 1H), 8.18 (d, J = 1.3 Hz, 1H), 7.97 – 7.89 (m, 1H), 7.56 – 7.41 (m, 1H), 7.11 (dd, J
= 8.9, 11.6 Hz, 1H), 4.50 (d, J = 9.0 Hz, 1H), 4.48 – 4.44 (m, 1H), 4.07 (s, 3H), 3.99 – 3.85 (m, 2H),
3.76 – 3.68 (m, 1H), 3.27 – 3.05 (m, 2H), 2.74 (dd, J = 2.7, 12.6 Hz, 1H), 2.14 (s, 1H), 1.52 (s, 9H).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
((4aS,5S,7aS)ꢀ2ꢀ((tertꢀButoxycarbonyl)amino)ꢀ7aꢀ(2ꢀfluoroꢀ5ꢀ(5ꢀmethoxypyrazineꢀ2ꢀ
carboxamido)phenyl)ꢀ4a,5,7,7aꢀtetrahydroꢀ4Hꢀfuro[3,4ꢀd][1,3]thiazinꢀ5ꢀyl)methyl 4ꢀ
methylbenzenesulfonate (18)
To a stirred solution of 17 (135 mg, 0325 mmol) in DCM (1.1 mL) were added tosyl chloride (160 mg,
0.839 mmol), DIPEA (111 mg, 0.861 mmol, 0.150 mL) and DMAP (11mg, 0.09 mmol) in that order.
The reaction was stirred at rt for 20 h. The reaction was diluted with DCM (10 mL) then washed
sequentially with 0.25 N aqueous HCl (3 mL), water (2 mL), saturated aqueous NaHCO3 solution (1
mL) and brine solution (1 mL). The organic extract was dried over Na2SO4, filtered and concentrated to
afford a crude product. Purification via flash column chromatography (silica gel, gradient 0 ꢀ 100%
EtOAc in heptanes) to afforded a white solid (123 mg), which was further purified by flash column
chromatography (silica gel, gradient 5 ꢀ 20% acetonitrile in DCM) to afford 18 (98mg, 56% yield) as a
white solid. HRMS: Calculated for C31H35FN5O8S2 [M+H]+: 688.1906, found 688.1907; Anal. Calcd for
C31H34FN5O8S2: C, 54.14; H, 4.98; N, 10.18. Found: C, 53.88; H, 4.94; N, 9.85; [α]D23.1 = +67.8° (c
0.345, CHCl3); 1H NMR (400 MHz, Acetonitrileꢀd3) δ 9.67 (s, 1H), 8.91 (s, 1H), 8.35 (s, 1H), 8.22 (d, J
= 1.3 Hz, 1H), 7.88 (ddd, J = 8.8, 4.1, 2.8 Hz, 1H), 7.79 (d, J = 8.3 Hz, 2H), 7.62 (dd, J = 7.1, 2.6 Hz,
1H), 7.42 (d, J = 8.0 Hz, 2H), 7.12 (dd, J = 12.0, 8.9 Hz, 1H), 4.37 (ddd, J = 8.2, 5.2, 3.6 Hz, 1H), 4.31
(d, J = 8.9 Hz, 1H), 4.25 – 4.12 (m, 2H), 4.03 (s, 3H), 3.74 (dd, J = 8.7, 2.5 Hz, 1H), 3.00 – 2.91 (m,
ACS Paragon Plus Environment

Page 27 of 36
Journal of Medicinal Chemistry
1H), 2.91 – 2.84 (m, 1H), 2.72 – 2.62 (m, 1H), 2.42 (s, 3H), 1.46 (s, 9H); 13C NMR (101 MHz,
Acetonitrileꢀd3) δ 163.38 , 162.46 , 157.06 (d, J = 243.5 Hz), 146.54 , 142.54 , 138.46 , 135.52 (d, J =
2.5 Hz), 134.63 , 133.64 , 131.03 , 128.92 (d, J = 11.5 Hz), 128.81 , 122.85 (m, br), 122.15 (d, J = 8.7
Hz), 117.65 (d, J = 25.0 Hz), 81.55 , 78.31 , 77.55 (d, J = 6.5 Hz), 71.40 , 67.29 (d, J = 5.1 Hz), 55.12 ,
40.18 (d, J = 3.1 Hz), 28.35 , 23.64 , 21.67.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Radiosynthesis of [3H]1.
A stirred solution of [3H]methyl iodide (1 Ci, 80 Ci/mmol, 12.5 µmol) in 0.5 mL of CHCl3 was
placed under N2, cooled in an ice bath and treated with a suspension of 5 (6.3 mg, 12.5 µmol, 1.0 eq) in
0.5 mL of CHCl3 and solid Ag2CO3 (3.5 mg, 12.7 µmol, 1.0 eq). The reaction mixture was then
warmed to rt and stirred for 30 h. The reaction mixture was filtered and the solvent was removed in
vacuo. The residue was then treated with 1 mL of trifluoroacetic acid and the resulting solution was
stirred at rt for 2 h. The reaction was then concentrated and the residue was dissolved in 4 mL of EtOAc
and washed with 4 mL of saturated aqueous solution of NaHCO3. The organic layer was concentrated
and the residue was purified by reverse phase HPLC to yield 20 mCi of [3H]1 with specific activity of 80
Ci/mmol and radiochemical purity > 99%. The HPLC condition is as follow: Column: Zorbax SBꢀC18,
5 µm, 4.6 x 250 mm; Column temperature: ambient; Mobile phase A: 0.1% formic acid in H2O; Mobile
phase B: acetonitrile; Gradient:10% ꢀ 100% B in 30 min; Flow rate: 1 mL/min; Run time: 30 min;
Detection: 254 nm.
LCꢀMS/MS Cold Tracer Method.
Brain binding of compounds 2 and 3 in mice were evaluated with a LCꢀMS/MS based “cold tracer”
assay. A total of three studies were performed. All procedures in this study were in compliance with the
U.S. Department of Agriculture’s (USDA) Animal Welfare Act (9 CFR Parts 1, 2, and 3); the Guide for
the Care and Use of Laboratory Animals: Eighth Edition, (Institute for Laboratory Animal Research,
The National Academies Press, Washington, DC, 2010); and the National Institutes of Health, Office of
Laboratory Animal Welfare. Whenever possible, procedures in this study were designed to avoid or
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 36
minimize discomfort, distress, and pain to animals. In study 1, male CDꢀ1 mice (25ꢀ35 g) were dosed
with 100 mg/kg of BACE1 inhibitor A22 (100 mg/kg, SC) or vehicle (n = 5 per treatment group),
followed by intravenous administration of the tracer, compound 2, at 10 µg/kg dose 60 min post
blocking compound or vehicle dosing. Mice were euthanized by live decapitation at 45 min post the
tracer administration and four brain tissues (cerebellum, frontal cortex, hippocampus and striatum) were
dissected and the plasma samples were collected to measure the tracer concentration using LCꢀMS/MS.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In study 2, male 129/sve mice (20ꢀ25 g) were dosed with varying doses of BACE1 inhibitor B23 (1, 10 or
80 mg/kg, SC) or vehicle (n = 5 per treatment group), followed by intravenous administration of the
tracer, compound 3, at 10 µg/kg dose 60 min inhibitor or vehicle dosing. Mice were then euthanized by
live decapitation at 45 min post the tracer administration. The same four brain tissues plus plasma
samples used in study 1 were collected to measure the tracer concentration.
In study 3, BACE1 KO and WT mice (n = 5 per group) were evaluated following the same procedure
used in previous two studies, with the exception that no blocking compound was dosed.
PET radiochemistry
Noꢀcarrierꢀadded [18F]fluoride was produced from a GEMS PET trace Cyclotron using 16.4 MeV
18
18
protons via the O(p,n)18F reaction on O enriched water ([18O]H2O). [18F]fluoride was isolated from
[18O]H2O on a preconditioned SepPak QMA light anion exchange cartridge and subsequently eluted
from the cartridge with a solution of K2CO3 (1.8 mg, 13 µmol), Kryptofix 2.2.2 (4,7,13,16,21,24ꢀ
hexaoxaꢀ1,10ꢀdiazabicycloꢀ [8.8.8]hexacosaneꢀK2.2.2) (9.8 mg, 26 µmol) in water (85 µL, 18 MΩ) and
CH3CN (2 mL) to a reaction vessel (10 mL). The solvents were evaporated at 160°C for 10ꢀ12 min
under continuous nitrogen flow (70 mL/min) to form a dry complex of [18F]fluoride/K2.2.2 and the
residue was cooled to rt. The precursor 18 (~ 0.007 mmol, ~5mg) in DMSO (600 µL) was added to the
dry complex of [18F]fluoride/K2.2.2. The closed reaction vessel was heated at 110 °C for 10 min and
cooled down to rt. Removal of the Boc protecting group was carried out by adding aqueous HCl (2.0 N,
ACS Paragon Plus Environment

Page 29 of 36
Journal of Medicinal Chemistry
200 µL) to the reaction mixture and heated at 100 °C for 5 min. The reaction mixture was cooled to rt
and was diluted with water to a total volume of 2.0 mL before injecting to HPLC for purification. A
radioactive fraction corresponding to [18F]3 (retention time = 15ꢀ17 min) was collected and diluted with
sterile water (50 mL). The resulting mixture was loaded onto a preꢀconditioned SepPak tC18 plus
cartridge. The cartridge was washed with water (10 mL) and the desired product [18F]3 was eluted with
EtOH (1 mL) into a sterile vial containing phosphate buffer saline solution (PBS, 9 mL). [18F]3 was
successfully radiosynthesized in 80ꢀ95 min from end of beam (n = 2) in 4 and 7% radiochemical yield (>
99% purity), 187 and 417 MBq of delivered radioactivity, and 84 and 175 GBq/µmol of specific
radioactivity at the time of injection, respectively.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NHP PET Imaging
[18F]3 was evaluated with PET imaging in one female cynomolgus monkey (body weight
approximately 5500g). The monkey was scanned twice with PET, one baseline scan and one blocking
scan with pretreatment of a BACE1 inhibitor LY288672124. Approximately 2 months separated the
baseline and blocking scans. The study was approved by the Animal Ethics Committee of the Swedish
Animal Welfare Agency and was performed according to the “Guidelines for planning, conducting and
documenting experimental research” (Dnr 4820/06ꢀ600) of Karolinska Institutet (KI).
The monkey was immobilized and transported from its housing facility to the KI PET center after
intramuscular injection of ketamine hydrochloride (approximately 10 mg/kg). The anesthesia was
maintained by the administration of a mixture of isoflurane (1.5ꢀ2.0%), oxygen and medical air after
endotracheal intubation at the KI PET center. For the blocking study, 5 mg/kg of LY2886721in 0.5%
methylcellulose with a dosing volume of 2 mL/kg, followed with 2 mL of tap water, was administrated
orally (via gavage) approximately 2 h before PET scanning. The monkey’s head was immobilized with a
fixation device and its body temperature was maintained by a Bair Hugger model 505 (Arizant
Healthcare, MN) and monitored with an oesophageal thermometer. The ECG, heart rate, respiratory rate
ACS Paragon Plus Environment