Discovery of N-(Pyridin-4-yl)-1,5-naphthyridin-2-amines as Potential Tau Pathology PET Tracers for Alzheimer’s Disease

Article abstract of DOI:10.1021/acs.jmedchem.6b01173

A mini-HTS on 4000 compounds selected using 2D fragment-based similarity and 3D pharmacophoric and shape similarity to known selective tau aggregate binders identified N-(6-methylpyridin-2-yl)quinolin-2-amine 10 as a novel potent binder to human AD aggreg

Full text of DOI:10.1021/acs.jmedchem.6b01173

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
Discovery of N-(pyridin-4-yl)-1,5-naphthyridin-2-amines as
potential tau pathology PET tracers for Alzheimer’s Disease.
Frederik J.R. Rombouts, José-Ignacio Andrés, Manuela Ariza, José Manuel Alonso, Nigel Austin, Astrid
Bottelbergs, Lu Chen, Vladimir Chupakhin, Erna Cleiren, Katleen Fierens, Alberto Fontana, Xavier Langlois,
Joseph E. Leenaerts, Jonas Mariën, Carolina Martinez Lamenca, Rhys Salter, Mark E. Schmidt, Paula Te
Riele, Cindy Wintmolders, Andrés A. Trabanco, Wei Zhang, Gregor J. Macdonald, and Diederik Moechars
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01173 • Publication Date (Web): 20 Jan 2017
Downloaded from http://pubs.acs.org on January 21, 2017
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 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of Nꢀ(pyridinꢀ4ꢀyl)ꢀ1,5ꢀnaphthyridinꢀ2ꢀ
amines as potential tau pathology PET tracers for
Alzheimer’s Disease.
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
a
b
Frederik J. R. Rombouts, * José-Ignacio Andrés, Manuela Ariza, José Manuel Alonso, Nigel
c
b
d
c
c
c
Austin, Astrid Bottelbergs, Lu Chen, Vladimir Chupakhin, Erna Cleiren, Katleen Fierens,
b
e
a
e
Alberto Fontana, Xavier Langlois, Joseph E. Leenaerts, Jonas Mariën, Carolina Martínez
a
d
f
e
e
Lamenca, Rhys Salter, Mark E. Schmidt, Paula Te Riele, Cindy Wintmolders, Andrés A.
g
h
a
e
Trabanco, Wei Zhang, Gregor Macdonald, Dieder Moechars
a
Neuroscience Medicinal Chemistry, Janssen Research & Development, Janssen Pharmaceutica
b
N. V., Turnhoutseweg 30, Bꢀ2340 Beerse, Belgium. Discovery Sciences, Janssen Research &
c
Development, JanssenꢀCilag S. A., C/ Jarama 75A, 45007 Toledo, Spain. Discovery Sciences,
Janssen Research & Development, Janssen Pharmaceutica N. V., Turnhoutseweg 30, Bꢀ2340
d
Beerse, Belgium. Isotope Chemistry and Biotransformation, Janssen Research & Development,
e
Welsh & McKean Roads, Spring House PA 19477, USA. Neuroscience Biology, Janssen
Research & Development, Janssen Pharmaceutica N. V., Turnhoutseweg 30, Bꢀ2340 Beerse,
f
Belgium. Neuroscience Experimental Medicine, Janssen Early Development, Janssen
g
Pharmaceutica N. V., Turnhoutseweg 30, Bꢀ2340 Beerse, Belgium. Neuroscience Medicinal
Chemistry, Janssen Research & Development, JanssenꢀCilag S. A., C/ Jarama 75A, 45007
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
h
Toledo, Spain. Neuroscience Biomarker Research, Janssen Research & Development, 3210
Merryfield Row, San Diego CA 92121, USA.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
KEYWORDS Tau, βꢀamyloid, Aβ, PET, positron emission tomography, Alzheimer’s disease,
quinoline, naphthyridine, microdosing.
ABSTRACT. A miniꢀHTS on 4000 compounds selected using 2D fragmentꢀbased similarity and
3
D pharmacophoric and shape similarity to known selective tau aggregate binders identified Nꢀ
(
6ꢀmethylpyridinꢀ2ꢀyl)quinolinꢀ2ꢀamine 10 as a novel potent binder to human AD aggregated tau
with modest selectivity versus aggregated βꢀamyloid (Aβ). Initial medicinal chemistry efforts
identified key elements for potency and selectivity, as well as suitable positions for
radiofluorination, leading to a first generation of fluoroalkylꢀsubstituted quinoline tau binding
ligands with subꢀoptimal physicochemical properties. Further optimization towards a more
optimal pharmacokinetic profile led to the discovery of 1,5ꢀnaphthyridine 75, a potent and
selective tau aggregate binder with potential as a tau PET tracer.
INTRODUCTION
Alzheimer’s disease (AD), a chronic neurodegenerative disorder and the most common cause
of dementia among the elderly population, is characterized by progressive cognitive impairment
including memory loss, compromised executive function, behavioral symptoms, and loss of
1,2
ability for selfꢀcare. Presentation can vary and signs and symptoms of other neurodegenerative
diseases overlap with those of AD and as a consequence the accuracy of clinical diagnosis, when
3
judged by autopsy, varies considerably. Despite the recent advances in the search for relevant
biomarkers to diagnose AD, such as amyloid positron emission tomography (PET) imaging
2
ACS Paragon Plus Environment

Page 3 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
agents, definitive diagnosis of AD can only be made by postmortem histological analysis of
4
human brain samples. Postmortem studies of AD brains have revealed two pathological
hallmarks of the disease: (1) extracellular senile plaques (SPs), composed of Aβ peptides, and (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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
intracellular neurofibrillary tangles (NFTs), composed of paired helical filaments (PHFs) of
5
hyperphosphorylated tau protein. Evidence for AD pathology can also be obtained from
cerebrospinal fluid (CSF) samples and brain imaging. The latter technique is noninvasive and
may be used to identify patients at risk of developing AD prior to the development of indications
6
ꢀ8
of cognitive impairment. Specifically, PET is a sensitive imaging technique that enables in
vivo and in situ noninvasive visualization, characterization, and quantification of physiological
9
processes at the molecular level, and has become an extremely powerful diagnostic tool.
Several PET probes for in vivo imaging of Aβꢀplaques have been described, of which four
11
18
18
representative examples 1 ([ C]PiB), 2 ([ F]Flutemetamol), 3a ([ F]Florbetapir), and 3b
18
10ꢀ14
(
[ F]Florbetaben) are shown in Chart 1.
The latter three were approved by regulatory
agencies for commercial use. However, the density of amyloid plaques does not correlate well
1
5
with the levels of neurodegeneration or cognitive impairment in AD.
HO
11CH3
NH
S
N
HO
S
N
NH
18F
11C-PiB (1)
18F-Flutemetamol (2)
18F
O
NH
O
O
X
X = N: 18F-Florbetapir (3a)
X = CH: 18F-Florbetaben (3b)
Chart 1. Representative Aβ PET probes used in the clinic.
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In contrast to the amyloid pathology, the density and neocortical spread of NFTs correlate very
1
4,16,17
well with progressive neuronal degeneration and cognitive decline in AD patients,
thus
making PET imaging of NFTs a desirable biomarker for AD. Indeed, recently published PET
imaging studies comparing amyloid PET and tau PET reported that the tau PET signal was
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8ꢀ20
highly correlated with cognitive performance and disease severity compared to Aβ imaging.
Whereas Aβ imaging may provide an early marker for disease risk, brain amyloid burden does
not correlate with disease severity. NFT imaging may allow early detection of
neurodegeneration, possibly even before the onset of cognitive symptoms, as well as predict
21
disease progression. NFT imaging may also help to delineate the relationship between the
accumulation of tau aggregates and their spread as a function of time, cognition, and brain
structure across the continuum from normal aging to AD. Tau aggregate imaging may also help
researchers to investigate the effect of diseaseꢀmodifying therapies on tau aggregation and
disease progression. In addition to its involvement in AD, tau aggregation is a characteristic of
2
2
other neurodegenerative diseases known as “tauopathies”, which include tangleꢀonly dementia
(
TD), argyrophilic grain disease (AGD), progressive supranuclear palsy (PSP), corticobasal
degeneration (CBD), Pick’s disease, and frontotemporal dementia and parkinsonism linked to
chromosome 17 (FTDPꢀ17). Therefore, tau aggregate imaging might be useful for the diagnosis
and study of these pathologies as well, although interpretation of the in vivo signal in early PET
studies of other tauopathies with current generation tau PET ligands is an active topic of
23
debate.
The development of small molecule PET imaging radiotracers for the in vivo quantification of
2
4ꢀ28
NFTs has been pursued by numerous academic and industrial groups.
Recently, we have
reviewed these efforts as well as the multiple challenges inherent to the discovery of suitable tau
4
ACS Paragon Plus Environment

Page 5 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
29
PET ligands. Several of these recently discovered radioligands with high binding to NFTs and
selectivity over Aβꢀplaques have been selected and evaluated in early clinical studies (Chart 2).
1
8
The Tohoku University has described a series of Fꢀlabeled “THK compounds”, among which 4
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
8
30ꢀ32
18
33
(
[ F]THKꢀ5117)
and 5 ([ F]THKꢀ5351) are being extensively evaluated in the clinic. The
1
8
18
Siemens group, now Avid/Lilly, has reported two Fꢀlabeled compounds 6 ([ F]AVꢀ1451,
18
34ꢀ38
18
18
25,39
previously known as [ F]T807),
and 7 ([ F]AVꢀ680, previously known as [ F]T808).
The National Institute of Radiological Sciences (NIRS) in Chiba has described a new series of
1
1
11
40,41
Cꢀlabeled phenylbutadienylbenzothiazole (PBB) compounds from which 8 ([ C]PBB3)
has
also been evaluated in human subjects. Very recently Merck & Co. has reported the discovery of
18
42
a new tau PET tracer 9 ([ F]MKꢀ6240) which is currently under clinical evaluation. Finally,
HoffmannꢀLa Roche in collaboration with the Johns Hopkins University School of Medicine,
1
1
reported in vitro and in vivo results with three potential tau tracers, [ C]RO6931643,
11
18
34
[ C]RO6924963, and [ F]RO6958948, but whose structures were not disclosed at the 2015
43
Annual Human Amyloid Imaging conference. From all these radiotracers, 6 has been the most
extensively studied in AD patients. However, although 6 is very selective for tau over Aβ, recent
reports suggest that it shows widespread, high affinity offꢀtarget binding, hence the signalꢀtoꢀ
23,44,45
background ratio in early AD patients may not be optimal.
In addition 6 is reported to have
a rather slow washꢀout from the human brain requiring long scanning times of 80ꢀ100 min, and
thus cortical (but not putamen) timeꢀSUVR curves for 6 may not subsequently plateau within
4
6
1
00 min for mild cognitive impaired (MCI) or AD subjects. These limitations reinforce our
belief that more sensitive and selective tau PET tracers are needed. Towards this objective we
started a medicinal chemistry program and herein describe our hit identification strategy and
subsequent medicinal chemistry optimization that allowed the discovery of a new family of
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
potent and selective 1,5ꢀnaphthyridines as promising tau PET tracers with potential for human
use, in particular for patient selection and monitoring disease progression in clinical trials for
AD.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Chart 2. Representative tau PET tracers evaluated in clinical studies.
SELECTION AND SCREENING
Our initial screening set was assembled from the Janssen compound collection applying two
4
7
different selection strategies: (a) 2D fragmentꢀbased (ECFP6) similarity via internally
developed software and (b) 3D shape/pharmacophore similarity to reference compounds with
4
8
49
known affinity for NFT using FastROCS. Reference compounds 2, 4, 6, 7 and astemizole,
were selected from literature and scientific databases up to January 2013, the moment our search
6
ACS Paragon Plus Environment

Page 7 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for tau PET tracers started. While being a nonꢀselective binder to tau, we selected astemizole to
increase the chemical diversity of our reference set. The FastROCS approach aligns the center of
masses of the conformer shapes of the reference with conformer shapes of the database
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
50
compounds. Then, a sum of 3D shape similarity and 3D pharmacophore features similarity was
calculated giving us a combo score. 3D pharmacophore features include: HBA, HBD, anion,
cation, hydrophobic and aromatic rings. FastROCS is a good tool for hit diversification but it is
sensitive to the size of the reference compounds compared to the 2D similarity methods. Thus,
we have additionally applied 2D fragmentꢀbased similarity methods to get more chemically
diverse compounds from internal compound database. Extended connectivity fingerprints of a
size 6 (ECFP6) were used to encode the compounds and Tanimoto similarity to compare the
reference compounds versus database compounds resulting in a score from 0, when no fragments
are shared between compounds, to a value of 1, when shared fragments are identical between
4
7
compounds.
The two ranked lists were combined using the median of individually scaled similarities,
compounds with molecular weight > 700 and number of rotatable bonds between aromatic
fragments ≥ 3 were discarded. Additional manual filtering was applied to the combined list to
reach a diverse library of ~4000 compounds. This compound set was then screened for binding
5
1
to aggregated tau extracted from human late stageꢀAD brains (Braak VꢀVI) using a filter
3
52
binding assay with [ H]7 as a displacement probe. Compounds with K ≤ 50 nM were
i
subsequently screened for their selectivity versus purified amyloid beta in a similar competition
3
52
assay using [ H]3a as a displacement probe. Following this strategy we identified Nꢀ(6ꢀmethylꢀ
ꢀpyridyl)quinolinꢀ2ꢀamine (10, Chart 3) as an attractive starting point with good binding affinity
for aggregated tau (K = 14 nM), moderate selectivity (~10ꢀfold) versus aggregated Aβ (K = 134
2
i
i
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
nM), and optimal physicochemical properties for achieving brain penetration (MW = 235, PSA =
2
53
3
8 Å , and S+LogP = 3.4 calculated with ADMET predictor™).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Chart 3. Hit identified from a miniꢀHTS with selected compounds with similarity to known
selective tau binders.
For the medicinal chemistry optimization of 10 to potential candidate PET tracers, we defined
a target product profile (TPP) with the following criteria: (i) high affinity tau binding (K ≤ 15
i
nM), (ii) over 100ꢀfold selectivity versus Aβ binding, (iii) the presence of fluorine atom that can
be easily introduced using standard radiochemistry methods, (iv) fast brain uptake and brainꢀtoꢀ
plasma ratio > 1 and (v) fast clearance without formation of radioactive brain penetrant or
defluorinated metabolites.
CHEMISTRY
The synthesis routes for the preparation of compounds 10 and 30-44 are outlined in Scheme 1.
While Nꢀ(pyridinꢀ2ꢀyl)quinolinꢀ2ꢀamine (11) was commercially available, compounds 10 and
30-43 were prepared via BuchwaldꢀHartwig aminations of (het)arylhalides (12-19) with anilines
or aminopyridines (20-27), which were commercially available or obtained by straightforward
and/or reported procedures (Scheme 1, route 1). The 2ꢀnaphthyl derivative 44 was prepared via
8
ACS Paragon Plus Environment

Page 9 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the reverse BuchwaldꢀHartwig coupling between 2ꢀaminonaphtalene (28) and 2ꢀbromoꢀ6ꢀ
methylpyridine (29) (Scheme 1, route 2).
a
Scheme 1. BuchwaldꢀHartwig couplings leading to compounds 10 and 30-44.
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
Reagents and conditions: a) Pd(OAc) , XantPhos, Cs CO , tertꢀbutanol, 90 °C, 10ꢀ16 h.
2
2
3
Compounds 50-52 were synthesized as shown in Scheme 2. Thus, BuchwaldꢀHartwig coupling
of 2ꢀaminoquinoline (45) with (4ꢀchloroꢀ2ꢀpyridinyl)methanol (46) provided alcohol 49 (Scheme
2
) which was subsequently treated with DAST to provide the fluoromethyl derivative 50 with
low yield. Hence for the preparation of 51ꢀ52, fluorination in the last step was avoided and both
compounds were synthesized in satisfactory yields via BuchwaldꢀHartwig coupling of 45 with
the fluorineꢀcontaining pyridylhalides 47 and 48. These in turn were conveniently prepared from
commercial materials (Scheme 2).
a
Scheme 2. Synthesis routes towards 2ꢀfluoroalkyl(oxy)pyridines 50-52.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
OH
OH
N
F
b
N
N
4
6
,
a
Cl
Br
N
N
H
7%
N
N
H
49
50
9
3 %
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
F
F
47
,
a
N
N
NH2
6
1%
N
N
H
45
F
5
1
48
N
, a
Cl
O
F
O
N
20%
N
N
H
52
a
Reagents and conditions: a) Pd (dba) , XantPhos, K PO , THF, 80 °C, 16 h. b) DAST, DCM,
2
3
3
4
0
°C, 30 min.
Compounds 58-60 were prepared via a twoꢀstep procedure (Scheme 3). First, commercially
available 2ꢀchloroquinolinꢀ6ꢀol (53) or 2ꢀchloroquinolinꢀ7ꢀol (54) was Oꢀalkylated with the
corresponding fluorineꢀcontaining primary haloalkane to give fluoroalkylethers 55-57. Finally,
reactions of the 2ꢀchloroquinolinethers 55-57 with 4ꢀaminoꢀpyridine (23) under Buchwaldꢀ
Hartwig conditions yielded the desired fluoroalkyloxyꢀsubstituted quinolines 58ꢀ60.
a
Scheme 3. Synthesis route towards fluoroalkoxyꢀsubstituted quinolines 58-60.
10
ACS Paragon Plus Environment

Page 11 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Reagents and conditions: a) 1ꢀFluoroꢀ2ꢀiodoethane or 3ꢀbromoꢀ1ꢀfluoropropane, Cs CO ,
2
3
DMF, 85 °C, 1ꢀ16 h. b) 23, Pd (dba) , XantPhos, K PO , toluene, 80 °C, 16 h.
2
3
3
4
Compound 63 was prepared from the previously reported 2ꢀbromoꢀ4ꢀnitroꢀquinoline (61)
5
4
following the twoꢀstep procedure shown in Scheme 4. Firstly, BuchwaldꢀHartwig coupling of
1 with 23 led to the nitro substituted quinoline 62. Subsequent nucleophilic aromatic
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
6
substitution of the nitro group in 62 with KF in DMSO yielded the 4ꢀfluoroquinoline 63 in
modest yield.
a
Scheme 4. Synthesis route towards 4ꢀfluoroquinoline 63.
O
O-
O
O-
N+
N
N+
N
F
a
b
N
N
Br
N
N
N
4
7%
34%
H
H
61
62
63
a
Reagents and conditions: a) 23, Pd (dba) , XantPhos, K PO , THF, 90 °C, 1 h. b) KF, DMSO,
2
3
3
3
150 °C (MW), 5 min.
The 1,7ꢀnaphthyridine derivative 68 was prepared from 4ꢀbromoꢀ6ꢀfluoroꢀpyridinꢀ3ꢀamine
55
(
64), as shown in Scheme 5. Thus, Heckꢀcoupling of 64 with methyl acrylate provided 65 in
moderate yield. Next, sodium methoxide promoted intramolecular cyclization of 65 led to the
bicyclic pyridine 66 with excellent yield. The 1,7ꢀnaphthyridone 66 was subsequently
chlorinated with POCl to give 67 which was then finally coupled with 23 to provide the target
3
molecule 68.
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Scheme 5. Synthesis route towards 1,7ꢀnaphthyridine 68.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Reagents and conditions: a) Methyl acrylate, Pd(Amphos)Cl , Et N, DMF, 145 °C, 40 min. b)
NaOMe, MeOH, 60 °C, 6 h. c) POCl , 120 °C (MW), 20 min. d) 23, Pd (dba) , XantPhos,
2
3
3
2
3
K PO , THF, 100 °C, 6 h.
3
4
5
6
1,5ꢀNaphthyridines 74 and 75 were prepared from 2ꢀbromoꢀ6ꢀfluoroꢀpyridinꢀ3ꢀamine (69)
using a similar procedure as the one outlined for 68 (Scheme 6). Thus, Heckꢀreaction of 69 with
methyl acrylate provided 70 in good yield. Cyclization of 70 under the previously described
NaOMe promoted reaction conditions was unsuccessful and did not afford compound 71 but the
corresponding carboxylic acid of 70. On the other hand, a literature procedure using PBu in
3
5
7
AcOH proved effective in providing 71 in 93% yield. Treatment of 71 with POCl yielded a
3
mixture and 72 and 73 which was not separated but subsequently coupled via BuchwaldꢀHartwig
amination with 23 or 4ꢀaminoꢀ2ꢀmethylpyridine (24) providing 74 and 75, respectively, after
isolation from the crude reaction mixture.
a
Scheme 6. Synthesis route towards 1,5ꢀnaphthyridines 74 and 75.
12
ACS Paragon Plus Environment

Page 13 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: a) Methyl acrylate, Pd(OAc) , PPh , Et N, DMF, 140 °C, 120 min.
2
3
3
b) Bu P, AcOH, 110 °C, 60 min. c) POCl , 1,4ꢀdioxane, 110 °C, 60 min. d) 23 or 24, Pd(OAc) ,
XantPhos, Cs CO , tertꢀbutanol, 90 °C, 18ꢀ24 h.
3
3
2
2
3
In view of a potential future radiofluorination, an alternative route to 75 in which the fluorine
label could be introduced in the last step was also explored (Scheme 7). Thus, 1,5ꢀnaphthyridine
76) was treated with mCPBA to provide the bisꢀNꢀoxide derivative 77. Next, rearrangement of
(
both Nꢀoxide functions with NMe in the presence of TFAA led to the bistrimethylammonium
3
58
trifluoroacetate salt 78 in good yield. Interestingly, one of both trimethylamine groups could be
selectively substituted by treatment with the sodium salt of 24, providing after purification by
RPꢀHPLC with 0.1% TFA buffer compound 79 as a triple TFA salt. Displacement of the
remaining trimethylammonium group 79 with fluorine in DMF at 90 °C using TBAF as fluorine
source provided 75 in a satisfactory 65% yield, demonstrating that intermediate 79 could be a
18
suitable radiofluorination precursor for the preparation of [ F]75.
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Scheme 7. Evaluation of potential radiofluorination route for 75.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Reagents and conditions: a) mCPBA, DCM/MeOH, rt, 75 h. b) TFAA, Me N, DCM, 0 °C to
rt, 30 min. c) 24, NaH, DMF, 0 °C to rt, 20 min. d) TBAF, DMF, 90 °C, 30 min.
3
3
Radioligand [ H]39 was prepared as follows. Monobromination of 39 occurred selectively in
the 3ꢀposition of the pyridine after treatment with NBS providing 80. Next, 80 was subjected to
3
Pd/C assisted tritioꢀdebromination by reaction with carrier free tritium gas at rt to afford [ H]39
in high specific activity (23.6 Ci/mmol) after HPLC purification (Scheme 8).
3
a
Scheme 8. Synthesis of radioligand [ H]39.
a
b
N
N
N
N
N
H
N
N
N
N
26%
~40%
H
H
Br
3H
39
80
[3H]39
a
3
Reagents and conditions: a) NBS, CHCl , MeOH, 60 °C, 2 h. b) H , 10% Pd/C, DIPEA,
3
2
DMF.
RESULTS AND DISCUSSION
14
ACS Paragon Plus Environment

Page 15 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Prior to exploring suitable fluorination positions, a focused exploration of 10 aiming to identify
key structural elements for potency and selectivity was carried out. In first instance variations
around the quinoline core were explored. The quinoline nitrogen proved to be crucial for good
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
tau binding, as the naphthalene derivative 44 showed significantly lower affinity for tau (44, K =
i
5
9
92.5 nM). Variation of the nitrogen position across the quinoline core revealed an interesting
and unexpected SAR: while isoquinolines 30, 33 and 34 showed potent binding to tau with Ki
values lower than 35 nM, the other (iso)quinoline variations 31, 32, and 35 had much weaker
affinity for tau. The quinazoline derivative 36, which can be considered a hybrid of the potent
binders 10 and 30, also showed reduced affinity for tau (Table 1). Interestingly compounds 33
and 34 were the only analogues from this exploration showing increased selectivity for tau
versus Aβ aggregates (205 and 33 fold respectively) when compared to the initial hit 10.
Table 1. Binding results for aggregated tau and Aβ from the exploration of the quinoline moiety
a,b
in HTS hit 10.
4
Compound
0 (Hit)
R
Tau K (nM) Aβ K (nM)
Selectivity (ratio)
9.6
i
i
1
14 ± 1.8
134 ± 73
30
8.8 ± 0.4
33.4 ± 6.5
3.8
ꢀ
31
> 3,900
> 4,400
32
815 ± 9.4
4,299 ± 99
5.3
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
3
4
11.6 ± 1.4
34.6 ± 4.0
498 ± 113
433 ± 19.9
92.5 ± 13.7
2,384 ± 327 205
1,150 ± 118 33
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
35
> 4,400
> 8
0.1
0.5
3
6
4
48.8 ± 1.7
44.0 ± 0.0
4
a
K values are a mean of at least two measurements and are given with standard deviation.
i
b
Selectivity ratio calculated as K (Aβ) / K (tau).
i
i
While the similar activity of quinolones 10 and 30 can be explained by the rotational freedom
around the NH bonds, which allows both molecules to present their acceptor nitrogens in similar
positions, further exploration of isomers 33 and 34 suggested a distinct binding mode. This was
confirmed after an extensive SAR exploration, which will be disclosed in a separate
communication.
Next, a focused exploration of the 6ꢀmethylpyridinꢀ2ꢀyl substituent in hit 10 was carried out
(
Table 2). Firstly, the presence of the pyridine nitrogen in 10 was crucial for activity as this was
lost in the toluene analogue 37. The methyl group next to the pyridyl nitrogen in 10 was not
essential for achieving good binding to aggregated tau as the 2ꢀpyridyl derivative 11 displayed
not only improved binding but also increased selectivity versus Aβ binding (63 fold). While
affinity for aggregated tau was lost in the 3ꢀpyridyl regioisomer (38) of 11, improved potency
and remarkable selectivity versus Aβ (> 680 fold) were observed for the 4ꢀpyridyl derivative 39.
Both potency and selectivity were maintained in the 2ꢀmethylpyridinꢀ4ꢀyl derivative 40.
Interestingly, binding was lost in the pyrimidine isomer 41, which displays each of the nitrogens
16
ACS Paragon Plus Environment

Page 17 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
at tolerated positions for activity in compounds 11 and 39. Finally, methylation of the NH linker
in 10 significantly reduced binding affinity to aggregated tau (42, tau K = 2,233 nM).
i
Table 2. Binding results for aggregated Tau and Aβ from the exploration of the aminopyridine
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
moiety in HTS hit 10.
5
6
Compound
R
R
Tau K (nM) Aβ K (nM) Selectivity (ratio)
i
i
1
1
3
0 (Hit)
H
H
H
H
H
H
H
13.9 ± 1.8
134.2 ± 73 10
1
7
3.9 ± 0.2
101 ± 12.7
560 ± 51
6.3 ± 0.7
6.3 ± 0.9
394 ± 80
246 ± 45
119 ± 27
> 4,400
63
1.1
38
> 7
3
9
0
> 4,400
> 680
> 680
2.5
4
> 4,400
41
997 ± 34
42
CH
3
2,233 ± 256 1,184 ± 27 0.5
a
K values are a mean of at least two measurements and are given with standard deviation.
i
b
Selectivity ratio calculated as K (Aβ) / K (tau).
i
i
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
From these initial explorations compounds 39 and 40 were the only examples that met the first
two criteria of the TPP, i.e. K ≤ 15 nM for aggregated tau and > 100ꢀfold selectivity versus Aβ.
i
Compound 39 is a low MW (221) compound with a PSA (38 Ų), logD (2.93) and logP (3.29)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
0
and pK (7.51) well within the ideal boundaries for a CNS penetrant molecule. Based on these
a
favourable properties 39 was further profiled and the results obtained are shown in Table 3.
Metabolic stability of 39 was assessed in human (h), rat (r) and mouse liver microsomes (mLM)
and was found to be low in all three species. High metabolic turnover is a desired attribute in a
PET tracer, provided that the produced radiometabolites are polar and do not enter the brain, as
this is likely to lead to more rapid clearance in vivo and therefore a short halfꢀlife which will help
minimizing scanning time. Compound 39 was administered subcutaneously (sc) to mice at a dose
of 5 mg/kg in a pharmacokinetic (PK) study to assess total brain and plasma concentrations.
Compound 39 showed an excellent exposure (Cmax 5980 ng/g) and a high total brainꢀtoꢀplasma
6
1
ratio (K ) of 8.1. Based on the AUC values in brain (5024 ng.h/g) and plasma (617 ng.h/mL),
p
and the unbound fractions in brain (fu,b rat = 0.016) and plasma (fu,p mouse = 0.12) determined in
vitro, the Kp,uu was estimated to be 1.09 in mouse. This could be expected from the low efflux
ratio of 0.95 observed in an MDCKꢀMDR1 cell line.
a
Table 3. Physicochemical and ADME properties of 39.
MW (D)
PSA (Ų)
LogD (pH 7.4)
LogP
221
38
2.93
3.29
7.51
pKa
18
ACS Paragon Plus Environment

Page 19 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hLM CL (ꢀL/min/mg)
112
> 346
258
0.016
0.13
0.12
int
rLM CL (ꢀL/min/mg)
int
mLM CL (ꢀL/min/mg)
int
fu,b
f (h)
u,p
f (m)
u,p
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
Papp A>B (nm/s)
B>A/A>B ratio
187
0.95
5980
5024 / 617
8.1
b
Cmax (ng/g)
b
AUC0ꢀlast, b / p
c
Kp
d
p,uu
a
K
1.09
b
See experimental section for assay details. PK study in male Swiss mice (n=2 per time point);
dose: 5 mg/kg sc; formulation: 0.5 mg/mL in 20% HPꢀβꢀCD at pH 3.6; time points: 0.25, 0.5, 1,
c
d
2, 4 h. AUC ratio of total drug levels in brain (ng.h/g) and plasma (ng.h/mL). Estimated from
AUC0ꢀlast values, f and f (m).
u,b
u,p
Finally, autoradiography studies were performed to assess whether the high in vitro binding of
39 translated into binding to AD brain tissue sections. To this end a tritiated analog of 39 was
3
prepared ([ H]39, Scheme 8), and its binding profile compared to that of tritiated analogues of
3
3
the selective amyloid binder [ H]3a and selective tau binder [ H]7 on human AD brain tissue
sections (Braak stage VꢀVI as indicated by IHC with phosphoꢀtau antibody AT8 and amyloid
antibody 4G8 performed on adjacent sections). As can be seen from Figure 1, the binding pattern
3
3
of [ H]39 matched that of [ H]7, thus confirming it to be a selective tau binder. Using saturation
3
binding experiments on adjacent sections, K and Bmax values were determined for [ H]7 (K 5.3
d
d
3
nM, Bmax 5823) and [ H]39 (K 8.5 nM, B 3827), see SI.
d
max
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
3
3
A. [ H]3a
B. [ H]7
C. [ H]39
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
D. AT8
E. 4G8
20
ACS Paragon Plus Environment

Page 21 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
3
3
Figure 1. Binding of (A) 3 nM [ H]3a, (B) 10 nM [ H]7 and (C) 10 nM [ H]39 to postꢀmortem
human AD brain sections and IHC with phosphoꢀtau antibody AT8 (D) and amyloid antibody
4
G8 (E) show 39 is selective for tau. Zoom regions have been boxed.
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
With the potent and selective tau ligands 39 and 40 in hand, an exploration looking for suitable
18
positions to introduce F for PET imaging was initiated. The pyridine ring was first explored by
modification with fluoroalkyl (50, 51) and fluoroalkyloxy chains (52). The binding results
determined for aggregated tau and Aβ are shown in Table 4 and suggest that the basicity of the
pyridyl nitrogen is a key factor to obtain good affinity for tau. For instance, the less basic
fluoromethyl derivative 50 presented a much reduced binding to tau (50, tau K = 37.7 nM)
i
compared to the methyl analogue 40 (tau K = 6.3 nM). The slightly more basic 2ꢀfluoroethyl
i
homologue 51 showed somewhat improved binding (51, tau K = 13.4 nM) when compared to
i
5
0, however, not as good as in the case of 40. Compound 52 bearing a fluoroakyloxy chain, also
showed reduced affinity for tau compared to 40, confirming that the presence of a nitrogen with
certain basicity is required for potent tau binding.
a,b
Table 4. Exploration of fluorine tails on the pyridine 2ꢀposition.
7
Compound
R
Tau K (nM)
Aβ K (nM)
Selectivity (ratio)
> 100
> 320
i
i
5
5
5
0
1
2
CH
CH
OCH
2
F
CH
37.7 ± 1.7
13.4 ± 2.0
17.2 ± 4.5
> 4,400
> 4,400
> 4,400
F
2
2
CH
F
> 250
2
2
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
K values are a mean of at least two measurements and are given with standard deviation.
i
b
Selectivity ratio calculated as K (Aβ) / K (tau).
i
i
Modifications around the quinoline core proved to be better tolerated for potency and offered
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
good opportunities for the introduction of fluorine containing aliphatic chains. Recognizing that
the quinoline nitrogen is a key pharmacophoric element, the exploration focused on the distal 6ꢀ
and 7ꢀpositions of the quinoline core (Table 5). Thus, the 6ꢀ(2ꢀfluoroethoxy) derivative 58
maintained potency and selectivity (> 750 fold) when compared to 39. In contrast, the isomeric
7ꢀ(2ꢀfluoroethoxy) derivative 59 did not show any affinity for tau aggregates. The 6ꢀ(3ꢀ
fluoropropyloxy) chain homologue 60 had similar potency and selectivity to the shorter 58.
While both 58 and 60 met the first three requirements of the TPP, i.e. high affinity for tau with >
18
1
00ꢀfold selectivity over Aβ and a tolerated position to introduce a [ F], they showed high
metabolic stability in hLM which, in the absence of other clearance pathways, could result in a
slow clearance from plasma and subsequently slow clearance from brain. Moreover significant
Oꢀdealkylation was observed for 60 upon incubation with human liver microsomes, which risks
to generate a potentially brain penetrant radiometabolite.
Table 5. Exploration of fluorine substituted alkyloxy chains on the quinoline 6ꢀ and 7ꢀ
a
position.
Compound
58
59
60
8
R
R
OCH CH F
H
OCH CH CH F
2
2
2
2
2
9
H
OCH CH F
> 3,900
H
2
2
b
Tau K (nM)
5.7 ± 2.5
4.7 ± 1.5
i
22
ACS Paragon Plus Environment

Page 23 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
b
Aβ K (nM)
> 4,400
> 750
283
> 4,400
ꢀ
283
47
> 4,400
> 900
297
47
4.31
NT
NT
10.3
147
0.0049
0.10 / 0.042
73
1.1
460
853 / 137
6.2
0.31
i
c
Selectivity (ratio)
MW (D)
PSA (Ų)
LogP
47
2.82
2.38
7.67
8.9
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
LogD
pKa
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
hLM CL (ꢀL/min/mg)
mLM CL (ꢀL/min/mg)
int
193
int
fu,b
fu,p m / h
0.018
0.13 / 0.22
132
1.0
792
789 / 178
4.4
0.61
Papp A>B (nm/s)
B>A/A>B ratio
Cbrain (15 min, ng/g)
d
d
AUC
(ng.h/g), b / p
0ꢀlast
e
Kp
f
K
p,uu
NT
a
b
See experimental section for assay details. K values are a mean of at least two measurements
and are given with standard deviation. Selectivity ratio calculated as K (Aβ) / K (tau). PK
study in Male swiss mice (n=2 per time point); dose: 1 mg/kg, sc; formulation: 0.1 mg/mL in
i
c
d
i
i
e
1
(
0% HPꢀβꢀCD at pH > 3.5. AUC ratio of total drug levels in brain (ng.h/g) and plasma
f
ng.h/mL). Estimated from AUC
values, f and f (m).
ꢀlast u,b u,p
0
In view of the challenges due to the low clearance of fluoroalkylꢀsubstitued quinolines the
introduction of aromatic fluorines on activated positions of either the pyridyl or quinoline rings
1
8
was explored. Although the preparation of the Fꢀlabeled compounds could present some
additional challenges, fluorine substituents in aromatic rings offer the advantage of being more
robust against oxidative defluorination which was one of the requirements in an optimal TPP.
Furthermore, a similar consideration led to the progression of 6 instead of 7, where the later
3
9
suffered from oxidative defluorination issues. The introduction of a fluorine atom in the 2ꢀ
position of the pyridyl ring was detrimental for affinity (43, tau K > 1,000 nM), as could be
i
expected due to its strong electron withdrawing effect making the pyridine nitrogen less basic.
Compound 63 which carries the fluorine substituent on the quinoline 4ꢀposition showed a
reduced binding affinity to tau compared to 39 (Table 6, 63, tau K = 18.7 nM vs Table 2, 39, tau
i
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
K = 6.3 nM). As no other activated positions to introduce a fluorine atom were available in the
i
quinoline core, and knowing that the presence of nitrogen atoms in certain positions of the fused
phenyl ring was tolerated for activity in compounds such as 32ꢀ35 (Table 1), a small exploration
focused on fluoroꢀsubstituted naphthyridine cores was initiated. Although the data for the first
representative prepared (68), bearing a 1,7ꢀnaphthyridine core, were rather disappointing (tau Ki
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
=
263 nM), better results were obtained for the isomeric 1,5ꢀnaphthyridines 74 and 75 which
maintained the good potency and high selectivity of the corresponding quinoline analogues 39
and 40.
a,b
Table 6. Exploration of aromatic fluorines at activated positions.
compound structure
Tau K (nM)
Aβ K (nM)
Selectivity (ratio)
> 4.0
i
i
43
1,024 ± 141
> 4,400
63
18.7 ± 0.6
> 4,400
> 230
6
8
4
263 ± 36
11.3 ± 2.8
7.9 ± 2.5
> 4,400
> 4,400
> 4,400
> 16
7
> 380
> 540
75
a
K values are a mean of at least two measurements and are given with standard deviation.
i
b
Selectivity ratio calculated as K (Aβ) / K (tau).
i
i
24
ACS Paragon Plus Environment

Page 25 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Compound 75 was selected for further evaluation as a potential tau PET radiotracer. The
physicochemical properties and ADME profile of 75 are compared with our internally generated
data for 6 in Table 7. 1,5ꢀNaphthyridine 75 is a low MW compound with a significantly lower
experimental LogP value compared to 6 (75, LogP = 3.38 vs 6, LogP = 4.31), and slightly higher
pK (75, pK = 7.89 vs 6, pK = 7.61). In line with the lower LogP, 75 had higher free fraction in
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
a
a
brain homogenate compared to 6 (75, f = 0.045 vs 6, f = 0.015). In addition, 75 was found to
u,b
u,b
have a high unbound fraction of 0.37, 0.32 and 0.39 in mouse, rat and human plasma,
respectively. Interestingly, 75 showed consistently high microsomal clearance across species
(
hLM = 49.9, rLM = 98.6, and mLM = 73.3 ꢀL/min/mg) in contrast to 6, which is cleared much
faster in mLM and rLM (98.6, 181 ꢀL/min/mg) compared to hLM (16.8 ꢀL/min/mg). This made
5 a good candidate for microdosing studies as it was likely to have fast clearance/short halfꢀlife
7
in rodent preclinical models and in human.
a,b
Table 7. Comparison between tau binders 6 and 75.
Compound
6
75
b
Tau K (nM)
1.1 ± 0.5
258 ± 76.1
235
263
42
NT
4.31
7.9 ± 2.5
> 4,400
> 540
254
51
2.18
i
b
Aβ K (nM)
i
Selectivity (ratio)
MW (D)
PSA (Ų)
LogD (pH 7.4)
LogP
3.38
pKa
7.61
7.89
fu,b
fu,p m / r / h
hLM CL (ꢀL/min/mg)
rLM CL (ꢀL/min/mg)
0.015
0.08 / NT / 0.11
16.8
98.6
181
0.045
0.37 / 0.32 / 0.39
49.9
94.4
73.3
int
int
mLM CL (ꢀL/min/mg)
int
Papp A>B (nm/s)
73.5
376
B>A/A>B ratio
1.6
0.68
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
b
See experimental section for assay details. K values are a mean of at least two measurements
i
and are given with standard deviation.
3
3
The specific binding of 75 to tau was assessed by displacement studies with [ H]3a, [ H]7, and
3
3
3
[ H]39 on human AD sections (Figure 2). The binding of selective tau binders [ H]39 and [ H]7
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
could be entirely displaced by treatment with 75 (Figure 2, AꢀD). In contrast the binding of
3
selective amyloid binders [ H]3a was not affected by treatment with 75 (Figure 2, EꢀF). The
presence and distribution of aggregated amyloid and tau in these sections was confirmed by IHC
(
Figure 2, GꢀH). This suggests 75 is a highly selective binder to aggregated tau.
3
3
3
3
A. [ H]39
B. [ H]39 + 75
C. [ H]7
D. [ H]7 + 75
3
3
E. [ H]3a
F. [ H]3a + 75
G. AT8
H. 4G8
26
ACS Paragon Plus Environment

Page 27 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
3
Figure 2. (A) Total binding of 10 nM [ H]39, (B) displacement of 10 nM [ H]39 with 1 ꢀM 75,
3
3
(C) total binding of 10 nM [ H]7, (D) displacement of 10 nM [ H]7 with 1 ꢀM 75, (E) total
3
3
binding of 3 nM [ H]3a, (F) competition of 3 nM [ H]3a with 1 ꢀM 75, (G) AT8 phosphoꢀtau
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
staining on adjacent section, (H) 4G8 amyloid staining on adjacent section.
The early lead 60 showed negligible clearance from cortex following iv dosing in mice in an
acute microdosing study (0.03 mg/kg). Compounds 6 and 75 on the other hand showed
significant clearance from brain compared to compound 60 which was in line with a previous
1
8
PET tracer study with [ F]6 (Figure 3) and reflects the high in vitro CL in mouse liver
int
3
4
microsomes. Of importance is that 75 has similar high turnover in hLM, in contrast to 6, which
is significantly more stable in hLM compared to mLM and rLM (Table 7). This high metabolic
stability of 6 could explain the unfavorable PK reported in human studies. Further allometric
6
5
modelling confirmed a high predicted clearance for 75 in man.
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 3. Cortex clearance for compounds 6, 60 and 75 in an acute microdosing study in NMRI
mouse (0.03 mg/kg, iv, n = 3/time point).
Based on the promising clearance observed for 75 in the acute microdosing study, the
compound was progressed into a more elaborate microdosing study in rat (iv, 0.03 mg/kg), in
which cortex, cerebellum and plasma levels were assessed in a 60 min time course (Figure 4).
These studies were used to extract relevant PK parameters which are summarized in Table 8.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Microdosing study of 75 in SpragueꢀDawley rat (0.03 mg/kg, iv, n = 3/time point) with
sampling of cortex, cerebellum and plasma.
First of all, 75 showed a fast clearance from plasma of 234 mL/min/kg, resulting in a short
halfꢀlife of 0.31 h. In addition, 75 showed good brain exposure given the low administered dose
(AUC cortex = 24.9 ng.h/mL) in line with its high Vd,ss (5.2 L/kg), and a high total cortex to
28
ACS Paragon Plus Environment

Page 29 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
plasma ratio (K = 12). Finally, levels in cortex and cerebellum correlated well, suggesting
p
comparable distribution of 75 between these regions.
Table 8. PK parameters extracted from microdosing of 75 in rat (0.03 mg/kg, iv)
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
CL (mL/min/kg)
Plasma t1/2 (h)
234
0.31
Plasma AUC0ꢀ1 h (h.ng/mL)
Cortex AUC0ꢀ1 h (h.ng/g)
Vd,ss (L/kg)
1.95
23.9
5.2
K cortex/plasma
12
p
A metabolic pathway study was performed for 75 in mouse, rat, dog, monkey and human
hepatocytes (Scheme 9). It was found that the same metabolites were formed in all five species,
one being the result of the hydroxylation of the methyl group (81), and the other an oxidation of
the 1,5ꢀnaphthyridine core (82). No defluorinated metabolites were observed across species in
this assay.
Scheme 9. Metabolites 81 and 82 observed for 75 after incubation with human, rat, mouse, dog
and monkey hepatocytes.
29
ACS Paragon Plus Environment