Identification of Three Novel Radiotracers for Imaging Aggregated Tau in Alzheimer's Disease with Positron Emission Tomography

Article abstract of DOI:10.1021/acs.jmedchem.7b00632

Aggregates of tau and beta amyloid (Aβ) plaques constitute the histopathological hallmarks of Alzheimer's disease and are prominent targets for novel therapeutics as well as for biomarkers for diagnostic in vivo imaging. In recent years much attention has been devoted to the discovery and development of new PET tracers to image tau aggregates in the living human brain. Access to a selective PET tracer to image and quantify tau aggregates represents a unique tool to support the development of any novel therapeutic agent targeting pathological forms of tau. The objective of the study described herein was to identify such a novel radiotracer. As a result of this work, we discovered three novel PET tracers (2-(4-[¹¹C]methoxyphenyl)imidazo[1,2-a]pyridin-7-amine 7 ([¹¹C]RO6924963), N-[¹¹C]methyl-2-(3-methylphenyl)imidazo[1,2-a]pyrimidin-7-amine 8 ([¹¹C]RO6931643), and [¹⁸F]2-(6-fluoropyridin-3-yl)pyrrolo[2,3-b:4,5-c′]dipyridine 9 ([¹⁸F]RO6958948)) with high affinity for tau neurofibrillary tangles, excellent selectivity against Aβ plaques, and appropriate pharmacokinetic and metabolic properties in mice and non-human primates.

Full text of DOI:10.1021/acs.jmedchem.7b00632

Subscriber access provided by EAST TENNESSEE STATE UNIV
Article
Identification of Three Novel Radiotracers for Imaging Aggregated
Tau in Alzheimer’s Disease with Positron Emission Tomography
Luca C. Gobbi, Henner Knust, Matthias Körner, Michael Honer, Christian Czech, Sara Belli, Dieter
Muri, Martin R. Edelmann, Thomas Hartung, Isabella Erbsmehl, Sandra Grall-Ulsemer, Andreas
Koblet, Marianne Rueher, Sandra Steiner, Hayden T. Ravert, William B. Mathews, Daniel P.
Holt, Hiroto Kuwabara, Heather Valentine, Robert F. Dannals, Dean F Wong, and Edilio Borroni
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00632 • Publication Date (Web): 27 Jun 2017
Downloaded from http://pubs.acs.org on June 28, 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 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Identification of Three Novel Radiotracers for Imaging Aggregated Tau in Alzheimer’s Disease
with Positron Emission Tomography
†
†
†
†
†
Luca C. Gobbi,* Henner Knust, Matthias Körner, Michael Honer, Christian Czech, Sara
†
†
†
†
†
Belli, Dieter Muri, Martin R. Edelmann, Thomas Hartung, Isabella Erbsmehl, Sandra Grallꢀ
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
†
†
†
†
‡
Ulsemer, Andreas Koblet, Marianne Rueher, Sandra Steiner, Hayden T. Ravert, William B.
‡
‡
‡
‡
‡
Mathews, Daniel P. Holt, Hiroto Kuwabara, Heather Valentine, Robert F. Dannals, Dean F.
‡
§#
†
Wong, Edilio Borroni
†
Pharma Research and Early Development, Roche Innovation Center Basel, F. HoffmannꢀLa
Roche Ltd., CHꢀ4070 Basel, Switzerland
‡
Department of Radiology, The Johns Hopkins University School of Medicine, Baltimore,
United States
§
Department of Psychiatry, The Johns Hopkins University School of Medicine, Baltimore,
United States
#
Department of Neuroscience, The Johns Hopkins University School of Medicine, Baltimore,
United States
ABSTRACT. Aggregates of tau and beta amyloid (Aβ) plaques constitute the
histopathological hallmarks of Alzheimer’s disease and are prominent targets for novel
therapeutics as well as for biomarkers for diagnostic in vivo imaging. In recent years much
attention has been devoted to the discovery and development of new PET tracers to image tau
aggregates in the living human brain. Access to a selective PET tracer to image and quantify tau
aggregates represents a unique tool to support the development of any novel therapeutic agent
targeting pathological forms of tau. The objective of the study described herein was to identify
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 74
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
such a novel radiotracer. As a result of this work we discovered three novel PET tracers (2ꢀ(4ꢀ
11
11
11
[
C]methoxyphenyl)imidazo[1,2ꢀa]pyridinꢀ7ꢀamine 7 ([ C]RO6924963), Nꢀ[ C]methylꢀ2ꢀ(3ꢀ
1
1
18
methylphenyl)imidazo[1,2ꢀa]pyrimidinꢀ7ꢀamine
8
([ C]RO6931643)
and
[ F]2ꢀ(6ꢀ
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
18
fluoropyridinꢀ3ꢀyl)pyrrolo[2,3ꢀb:4,5ꢀc’]dipyridine 9 ([ F]RO6958948)) with high affinity for tau
neurofibrillary tangles, excellent selectivity against Aβ plaques and appropriate pharmacokinetic
and metabolic properties in mice and nonꢀhuman primates.
INTRODUCTION
The microtubuleꢀassociated protein tau is highly expressed in neurons, where it has an
important function in the assembly and stabilization of axonal microtubules. Six isoforms of tau
exist in the adult human brain, originating from a common gene by alternative splicing, and
1
varying between 352 to 441 amino acids in length. The native soluble forms of the tau protein
exhibit predominantly a random coil conformation. By contrast, in certain types of neurological
disorders, tau is encountered in protein aggregates possessing a cross betaꢀsheet amyloid core
2
structure. Most prominent among these neurological disorders is Alzheimer’s disease (AD), that
is characterized by the presence of tau aggregates in the brain in form of neurofibrillary tangles
3
(NFTs) and neuropil threads (NTs). The second protein aggregates present in AD consist of
amyloidꢀβ (Aβ) plaques, sharing a similar cross betaꢀsheet fibrillary core structure. The
spreading of tau pathology in AD is well documented and follows a predictable pattern, both in
3
terms of timing and spatial localization.
ACS Paragon Plus Environment

Page 3 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
As of today, no cure exists for AD that represents the most common reason for dementia. The
increasing prevalence of AD related to aging populations, in particular in western countries,
4
makes it a major reason of concern for society and a topic for intensive research. Evidence is
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
emerging in recent years for tau NFTs but not Aβ plaques to be indicators of clinical symptoms
5
progression, despite the fact of Aβ pathology temporally preceding the one of tau. Thus several
novel therapeutic approaches to cure or at least modify the progression of AD are focusing on tau
and on preventing the spreading of aggregated tau species or clearing of existing NFTs and NTs.
The development of any such new therapy would greatly benefit from the existence of a sensitive
biomarker able to monitor tau pathology in the living brain. In particular, this would allow for
patient selection and stratification as well as collection of longitudinal data demonstrating the
efficacy of any novel antiꢀtau therapeutic agent at limiting tau pathology. Moreover, such a
biomarker would represent an invaluable diagnostic tool for monitoring AD.
Recently, a number of positron emission tomography (PET) tracers targeting aggregated forms
6ꢀ8
of tau have been identified. Such radiotracers enable for the first time the quantitative
9
monitoring of tau aggregates in the living human brain. For any novel tau PET tracer, selectivity
over Aβ plaques is mandatory, because of the mixed nature of pathology in AD. The first
1
8
clinically evaluated radiotracer in this field was 1 ([ F]FDDNP), a ligand possessing high
1
0, 11
affinity for Aβ plaques and thus of limited use as tau aggregates biomarker (Figure 1).
More
recently several novel structures of selective tau tracers were reported, the most prominent being
1
8
18
12ꢀ14
18
15, 16
2 (Flortaucipir, formerly known as [ F]AVꢀ1451 and [ F]T807),
3 ([ F]T808),
4
18
17
11
18, 19
18
20, 21
(
[ F]THK5351), 5 ([ C]PBB3)
and 6 ([ F]MKꢀ6240).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 74
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
Previous work:
[18
F]1
[ F]2
18
[ F]3
18
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
[18
11
18
F]4
[ C]5
[ F]6
This work:
1
1
11
18
[ C]7
[ C]8
[ F]9
Figure 1. Structure of PET tracers for imaging of tau aggregates reported to have entered
clinical evaluation studies.
While running an internal program to secure access to a selective tau PET tracer, the initial
1
8
18
12, 13, 15
results on [ F]2 and [ F]3 were disclosed.
The two structures were deemed being
promising starting points for a PET tracer discovery program. We herein wish to report on our
work that has culminated in the identification of the three novel and selective tau radiotracers
11
11
22 11
11
23
18
18
24
[
C]7 ([ C]RO6924963), [ C]8 ([ C]RO6931643) and [ F]9 ([ F]RO6958948).
RESULTS AND DISCUSSION
ACS Paragon Plus Environment

Page 5 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Amyloid aggregates are known to possess different binding sites on the crossꢀbeta sheet fibrils
2
5ꢀ27
surface.
Caution is thus mandatory when interpreting the results of any competitive binding
assay, as the addressed binding site depends on the reporter ligand that is used. Our strategy to
find novel high affinity tau NFT ligands was to build a novel medium to high throughput assay
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
3
based on the displacement of [ H]3 from brain sections originating from AD donors. Ligand 3
has been described to possess high affinity of K = 22 nM for tau aggregates as encountered in
d
1
6
AD. New chemical structures were derived from 3 and 2 following the hypothesis that related
scaffolds would share a common binding site on the amyloid fibril surface. Early in the program
3
we demonstrated 2 to indeed compete for binding with [ H]3 (vide infra). For assessing
selectivity of the most promising candidates against Aβ plaques, we relied on low throughput,
direct binding assays with tritiated material. Of particular importance for rapidly progressing the
project was the simplification of the tricyclic aromatic core of the molecules. In addition,
incorporation of heteroatoms in the aromatic rings was investigated to reduce the lipophilicity,
thus aiming at improving the specific to nonꢀspecific binding ratio. The optimal range for the
properties of a PET tracer for use in the central nervous system has been extensively reviewed in
2
8ꢀ30
the past few years and this information guided our design of novel molecules.
Synthesis of non-labeled compounds. The synthesis of imidazo[1,2ꢀa]pyridines typically
started from 2ꢀaminoꢀ4ꢀchloropyridine (10) by reaction with an appropriate amine (Scheme 1).
While 2,4ꢀdiaminopyridine (11) was commercially available, reaction of 10 with methylamine
produced N4ꢀmethylꢀ2,4ꢀdiaminopyridine (12). Condensation of 11 and 12 with αꢀ
bromoacetophenones 13-15 afforded imidazo[1,2ꢀa]pyridines 7 and 16-18. Direct Nꢀalkylation
of 16 to give the dimethylamino derivative 19 proceeded in poor yield (6%) under formation of a
mixture with the monoꢀalkylated compound 17. No attempt was made to optimize this step.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 74
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
MonoꢀNꢀalkylation of 16 with a fluoroethyl group was best achieved in a 3ꢀstep procedure via
the Boc intermediates 20 and 21, thus leading to the final product 22.
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
Scheme 1: Synthesis of Imidazo[1,2-a]pyridines 7 and 16-22
a
b
10
R1
R1
R2
11: NH2
12: NHMe
16:
19:
NH2
NMe2
H
H
c
R2
3: H
4: OMe
5: OH
1
7:
7:
18:
20:
NHMe
NH2
NH2
H
d
1
1
1
OMe
OH
H
H
H
NHBoc
e
f
2
2
1: N(CH CH F)Boc
2: NH(CH CH F)
2
2
2
2
a
Reagents and conditions: (a) for 12: R = NHMe: 40% aq. MeNH , 160 °C, 4 h, 75%; (b) for
1
2
7
, 18: 14 or 15, NaHCO , MeOH or EtOH, refl., 4ꢀ6 h, 22ꢀ35%, for 16, 17: 13, acetone, 65 °C, 5
3
h, 44ꢀ90%; (c) NaH, DMF, 30 min, then MeI, rt, 18 h, 6%; (d) LiHMDS, THF, 0 °C, 25 min,
then Boc O, rt, 1.5 h, 66%; (e) NaH, DMF, rt, 15 min, then 1ꢀbromoꢀ2ꢀfluoroethane, 50 °C, 1 h,
2
95%; (f) TFA, DCM, rt, 3 h, 93%.
Commercially available 4ꢀchloropyrimidinꢀ2ꢀamine (23) or pyrimidineꢀ2,4ꢀdiamine (24)
(Scheme 2) served as starting points for the synthesis of imidazo[1,2ꢀa]pyrimidines. Where
necessary, aromatic nucleophilic substitution of 4ꢀchloropyrimidinꢀ2ꢀamine (23) was used in a
first step to introduce the amine substituents in position R1 to give intermediates 25-28.
ACS Paragon Plus Environment

Page 7 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Condensation of 24ꢀ28 with appropriately substituted αꢀbromoacetophenones in acetone
afforded the final products 8 and 29-36. Phenol 33 was further elaborated by alkylation to the
fluoroethoxy derivative 37 or it’s precursor for radiofluorination, tosylate 38.
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
Condensation of 4,6ꢀdiaminopyrimidine (39) and αꢀbromoacetophenone in EtOH at 80 °C
afforded in one step the regioisomeric imidazo[1,2ꢀc]pyrimidine 40 (Scheme 2).
Scheme 2: Synthesis of Imidazo[1,2-a]pyrimidines 8 and 29-36 and Imidazo[1,2-
a
c]pyrimidine 40
a
b
23
R1
R1
R2
R3
2
2
2
4:
NH2
5: NHMe
6: NMe2
29: NH2
30: NHMe
31: NMe
H
H
H
H
H
OH
H
H
H
Me
Me
H
2
3
2: NH2
2
7: NHcꢀPr
8
: NHMe
3: NHMe
7: NHMe
2
8:
3
3
c
OCH CH F
H
2
2
=
R4
38
:
NHMe OCH CH OTs
2
2
H
H
H
H
d
3
3
3
4: NHcꢀPr
5: NHcꢀPr
6:
OMe
OH
H
R4
e
39
40
a
.
Reagents and conditions: (a) 25: NH Me HCl, K CO , NMP, 120 °C, 2 h, 40%; 26:
2
2
3
.
NHMe HCl, K CO , NMP, 120 °C, 1 h, 43%; 27: NH cPr, K CO , DMF, 100 °C, 18 h, 55%;
2
2
3
2
2
3
.
28: 4ꢀ(2ꢀfluoroethyl)piperidine HCl, K CO , NMP, 120 °C, 12 h, 78%; (b) 8 and 29-36: αꢀBrꢀ
2
3
acetophenones, acetone, 60ꢀ65 °C, 5ꢀ18 h, 35ꢀ94% or 36: αꢀBrꢀacetophenone, pTsOH, acetone,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 74
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
6
5 °C, 12 h, 25%; (c) 1ꢀbromoꢀ2ꢀfluoroꢀethane, Cs CO , DMF, 70 °C, 1 h, 55%; (d)
2
3
(TsOCH ) , Cs CO , DMF, 50 °C, 2 h, 41%; (e) αꢀBrꢀacetophenone, EtOH, 80 °C, 8 h, 13%.
2
2
2
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
The initial synthesis of the tricyclic pyrrolo[2,3ꢀb:4,5ꢀc’]dipyridine core was achieved in two
steps, beginning with a Suzuki crossꢀcoupling of iodide 41 with boronic acid 42 leading to
intermediate 43 (Scheme 3). Ring closure of 43 by intramolecular nucleophilic substitution was
accompanied by concomitant loss of the Boc protective group, thus resulting in tricyclic
intermediate 44. Reꢀprotection with Boc anhydride to obtain 45 was beneficial for further
transformations of this compound via Suzuki crossꢀcoupling reaction towards both the αꢀfluoro
pyridine 47 and the corresponding nitro derivative 49, that represents the radiofluorination
18
precursor of [ F]9. Complete loss of the Boc protective group was observed during the Suzuki
crossꢀcoupling reaction leading to nitro derivative 49, while the synthesis of 9 under identical
conditions led to a mixture of free and NꢀBoc protected material, thus requiring an additional
cleavage step. Full consumption of starting material 45 was observed in the crossꢀcoupling
reactions leading to 47 and 49. The modest yields for these conversions were mainly originating
from the formation of side products and from the poor solubility of the final compounds, that
complicated their isolation.
a
Scheme 3: Synthesis of pyrrolo[2,3-b:4,5-c’]dipyridines 9 and 49
ACS Paragon Plus Environment

Page 9 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
b
42
41
43
4
4: H
c
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
4
5: Boc
d
4
7: Boc
45
e
9
: H
4
6
d
45
49
48
a
Reagents and conditions: (a) [Pd(OAc) ], PPh , Et N, DMF, 100 °C, 3 h, ca. 90% purity,
2
3
3
6
3%; (b) K CO , 18ꢀcrownꢀ6, DMF, 100 °C, 3 h, 63%; (c) NaH, DMF, 0 °C to rt, then Boc O,
2
3
2
DMF, 0 °C to rt, 18 h, 73%; (d) [PdCl (dppf)].CH Cl , K CO , DMF, 90 °C, 17ꢀ18 h, 47: 10%,
2
2
2
2
3
4
9: 13%; (e) TFA, CH Cl , rt, 18 h, 88%.
2
2
To explore structural variations of the tricyclic core of 9, one of the nitrogen atoms was shifted
in the scaffold (Scheme 4). Starting from 4ꢀchloroꢀ5ꢀiodopyrimidine (50) and 3ꢀbromoaniline
51) intermediate 52 was obtained by nucleophilic aromatic substitution (Scheme 4). The
(
31
intramolecular Pdꢀcatalyzed direct arylation gave the tricyclic core of compound 53, that was
reacted under Suzuki crossꢀcoupling conditions with boronic acid 54, to lead to the final product
55 in moderate yield. Protection of the pyrroleꢀNH was not required for this step.
a
Scheme 4: Synthesis of 9H-pyrimido[4,5-b]indole 55
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 74
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
5
1
50
5
2
5
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
54
c
5
5
a
Reagents and conditions: (a) EtOH, reflux, 3 h, 70%; (b) [Pd(OAc) (PPh ) ], NaOAc, DMA,
2
3 2
.
ꢀW 130 °C, 4 h, 11%; (c) [PdCl (dppf)] CH Cl , K CO , DMF, ꢀW 100 °C, 1 h, 20%.
2
2
2
2
3
Further modification of 2 was investigated by replacing the 6ꢀfluoroꢀ4ꢀpyridinꢀ3ꢀyl moiety
with unsaturated fluorinated, cyclic amines (Scheme 5). Thus, 1ꢀacetylꢀ3ꢀbromopiperidinꢀ4ꢀone
.
(
56) and (3ꢀbromophenyl)hydrazine hydrochloride (57) were reacted in the presence of BF Et O,
3
2
leading to the formation of an inseparable mixture of regioisomers having the bromoꢀatom in
positions 7 (58a) or 9 (58b) of the 5Hꢀpyrido[4,3ꢀb]indole scaffold. NꢀTosylation to form 59
allowed for the chromatographic isolation of the desired regioisomer. Buchwald amination of 59
with 4ꢀfluoro piperidine (60) resulted in the tosyl protected 5Hꢀpyrido[4,3ꢀb]indole 61. The
desired final product 62 was obtained after cleavage of the tosyl group.
a
Scheme 5: Synthesis of 5H-pyrido[4,3-b]indole 62
ACS Paragon Plus Environment

Page 11 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a, b
—
HCl
5
6
0
57
59
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
R
1: Ts
2: H
6
6
6
c
d
a
.
Reagents and conditions: (a) BF Et O, 120 °C, 16 h, 69% (mixture of regioisomers); (b)
3
2
.
NaH, DMF, 0 °C to rt, 30 min then TsCl, rt, 1 h, 21%; (c) [Pd (dba) ] CHCl , Xantphos,
2
3
3
Cs CO , THF, 90 °C, 16 h, 76%; (d) NaOH, THF/H O, 50 °C, 20 h, then NaOH, MeOH, reflux
2
3
2
6
h, 25%.
In vitro optimization of binding affinity and physicochemical properties. In order to test
the affinity of novel compounds towards protein tau aggregates possessing the polymorphic form
3
relevant for AD pathology, we set up a novel assay based on the displacement of [ H]3 from
immunohistologically characterized fresh frozen human brain sections derived from AD cases.
Tritiation of 3 is described in the supporting information and delivered the radioligand in 99%
radiochemical purity and a specific activity of 1.49 TBq/mmol (40.3 Ci/mmol). Novel
compounds were tested at a concentration of 10 nM (results in Table 1). As expected, both 2 and
3
3 demonstrated good affinity by displacing 41% and 40% of radioligand ([ H]3) binding,
16
respectively. The affinity of 3 for tau aggregates was previously reported with a K = 22 nM,
d
thus in a comparable range as found with our new assay. In addition to the primary binding
affinity assay, lipophilicity (logD) and passive cell membrane permeability (P in the PAMPA
eff
32
assay ) were determined for all compounds and used to guide the optimization program. We
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 74
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
complemented the standard octanol/water partition coefficient logD results with the
3
3,
measurement of LIMBA logDbrain in the novel brain Lipid Membrane Binding Assay LIMBA.
3
4
2
was found to be rather lipophilic, with a logDoctanol = 3.28, and possessed a moderate passive
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
ꢀ1
permeability P = 1.8 x 10 cm s . By contrast, 3 had a lower logD
= 2.80 and a higher Peff
octanol
eff
ꢀ6
ꢀ1
= 4.6 x 10 cm s . The initial goals for optimization where an increase in affinity combined with
a decrease in lipophilicity to augment the specific to nonꢀspecific binding ratio in a PET
ꢀ6
ꢀ1
experiment, and obtaining a PAMPA P > 2 x 10 cm s to provide high and fast brain uptake.
eff
Initial attempts to simplify the tricyclic core of 3 by replacement with a bicyclic imidazo[1,2ꢀ
a]pyrimidine, flanked by an aromatic ring in position 2, led to the observation that very simple
structures possessing a small amino group in position 6 of the bicyclic heteroaromatic ring
system retained moderate to good affinity (29-31) (Table 1). Interestingly and to our surprise a
direct analogue of 3, compound 36, had only very modest affinity (14% ligand displacement).
These novel structures had an optimal logD (1.42, 2.13 and 2.33 for 29, 30, 31 respectively),
ꢀ6
ꢀ
possessed excellent passive membrane permeability in the PAMPA assay (P ≥ 4.2 x 10 cm s
eff
1
)
and had low molecular weight in the range of 210ꢀ238 g/mol, indicative of a high ligand
3
5
efficiency LE. Based on these favorable properties, a large optimization program was started
on this novel structural class. Attempts to shift one of the nitrogen atoms in the bicyclic ring to a
different position led to compound 40 that had a similar affinity compared to 29. To our delight,
removal of one of the ring nitrogen atoms of 29 giving imidazo[1,2ꢀa]pyridine 16, further
increased affinity (58% ligand displacement vs. 40% for reference 3). Methylation of the amino
group of 16 resulted in 17 and 19 which displayed similar high affinity (60% and 59% ligand
displacement respectively). All these novel imidazo[1,2ꢀa]pyridines possessed appropriate
ACS Paragon Plus Environment

Page 13 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
lipophilicity (logD ≤ 2.64) in combination with excellent passive membrane permeability (P ≥
eff
ꢀ6
ꢀ1
4
.3 x 10 cm s ).
Further modification of the lead structures of 29 (imidazo[1,2ꢀa]pyrimidine) and 16
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
(
imidazo[1,2ꢀa]pyridine) to introduce positions for carbonꢀ11 or fluorineꢀ18 labelling finally led
to the identification of compounds 7, 8, 22, 34 and 37. When compared to 3, these five ligands
possessed comparable or better affinity for tau aggregates in combination with lower
lipophilicity and thus were predicted to display a more favorable specific to nonꢀspecific binding
ratio (SB/NSB). All these structures exhibited very good passive membrane permeability in the
PAMPA assay.
In parallel to the work around leads 16 and 29 derived from 3, attempts were made to modify
the structure and improve the properties of 2. Replacement of the 6ꢀfluoropyridinꢀ4ꢀyl ring of 2
with several aliphatic amines was found to be detrimental, as exemplified by the 4ꢀ
fluoropiperidine derivative 62. Introduction of an additional nitrogen atom in the tricyclic core of
2
failed to give the desired properties in the case of final compound 55, but was successful for
the regioisomeric derivative 9. 9HꢀDipyrido[2,3ꢀb;3',4'ꢀd]pyrrole 9 maintained the high affinity
(44% ligand displacement) in combination with a slight reduction of the lipophilicity (logD =
ꢀ6
ꢀ
3
.22). Passive membrane permeability of 9 was found to be good (PAMPA P = 3.8 x 10 cm s
eff
1
).
A modest linear correlation was found between the results for logD and LIMBA logDbrain
2
(
LIMBA logDbrain = 0.27 + (0.42 x logD); r = 0.32). Interestingly LIMBA logDbrain predicted all
compounds to possess low nonꢀspecific binding (i.e. LIMBA logDbrain < 2.0). By contrast,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 74
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
prediction based on logD would have erroneously flagged several compounds with logD > 3 to
2
9, 30, 34
display high nonꢀspecific binding.
During the course of the program, approximately 550 novel compounds originating from the
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
structural classes discussed above were prepared and tested. In the light of the favorable in vitro
1
1
18
properties, and a reasonably good predicted chance to incorporate a C or F radionuclide, the
six lead tracer molecules 7-9, 22, 34 and 37 were selected for further profiling by in vitro
3
autoradiography of Hꢀlabelled material, in depth pharmacokinetic (PK) evaluation and PET
imaging in nonꢀhuman primates.
a
Table 1. Results for the determination of affinity , lipophilicity (logD), brain lipid membrane
binding (LIMBA logDbrain) and passive membrane permeability (PAMPA).
%
[
inhib.
H]3
b
LIMBA
logDbrain
PAMPA P
ꢀ6
eff
ꢀ1
Compd
Structure
3
a
logD
[x 10 cm s ]
2
41
3.28
1.92
1.8
3
7
40
67
2.80
1.73
1.41
1.57
4.6
3.6
8
48
2.51
1.30
5.6
ACS Paragon Plus Environment

Page 15 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
9
44
3.22
1.52
3.8
16
58
60
1.58
2.23
1.27
1.37
4.7
4.3
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
1
1
7
9
59
2.64
1.56
5.6
2
2
3
2
9
0
40
23
39
2.40
1.42
2.13
1.33
0.72
0.87
4.4
6.1
4.2
31
35
36
2.33
2.55
1.25
0.38
4.9
3.8
3
3
4
6
14
3.14
1.94
4.4
3
4
7
0
43
25
1.83
2.01
1.13
0.87
3.2
5.9
N
N
N
NH2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 74
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
c
c
5
6
5
2
7
3.35
2.56
ꢀ
4.5
4.1
13
ꢀ
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
3
Displacement of [ H]3 (10 nM) from fresh frozen human brain sections derived from AD
cases; test concentration = 10 nM.
b
Determined at pH = 7.4.
c
No data determined on account of the poor affinity for tau aggregates.
In vitro transport and in vivo pharmacokinetic profile. The in vitro transport and in vivo
drug pharmacokinetic profile for the reference 3 and the six lead tracer candidates 7-9, 22, 34
and 37 is summarized in Table 2. Pharmacokinetics in mice after single intravenous (i.v.) bolus
administration were studied at a relatively low dose (~1 mg/kg) to better reflect the micro dosing
conditions encountered in a human PET experiment, yet without compromising the
quantification of the compound concentration in the biological samples. As reported in Table 2,
the tested compounds showed high systemic plasma clearance in mouse in vivo studies typically
exceeding liver blood flow, and short terminal plasma halfꢀlives ranging between 0.20 > t1/2
0.43 h, thus being appropriate for use with a carbonꢀ11 (isotope decay t1/2 = 20 min) or fluorineꢀ
8 (t1/2 = 110 min) label. Using physiologicallyꢀbased PK modeling, that integrates drug
>
1
physicochemical properties, animal/human in vitro input and in vivo PK data from various
ACS Paragon Plus Environment

Page 17 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
species (data not shown), the compounds 7ꢀ9 were expected to undergo high systemic clearance
in humans. The same model predicted a short terminal halfꢀlife of ca 2 h and a large tissue
distribution, all properties that were considered appropriate for a PET tracer.
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
Both in vitro transport and in vivo brain uptake studies in animals (methods in Supporting
Information) confirmed the compounds not to be recognized by the efflux transporter Pꢀ
glycoprotein (Pꢀgp, mouse and human), to possess high passive cellular and PAMPA
permeability and large distribution volume in vivo, and were therefore predicted to reach rapidly
the brain in human.
Free fractions (f , method in Supporting Information) for the new tracer candidates in mouse
u_p
and human plasma were high, in general > 10%, in contrast to the reference compound 3. High
fu_p was deemed to be of advantage, in order to facilitate rapid in vivo clearance of the ligands
from nonꢀspecifically bound and free compartments, both contributing to confound the specific
3
6
binding component in the PET images. Furthermore, the high f values together with the lack
u_p
of interaction with Pꢀgp and high in vitro permeability were considered to be beneficial in
37, 38
promoting rapid brain uptake and large unbound fraction/low nonꢀspecific binding in brain.
In fact, most compounds possessed total brain to plasma AUC ratios (B/P) in rodents > 1, with
the exception of 37. A B/P = 0.5 for the latter compound was deemed to be still sufficient to
progress 37 into further tests.
Table 2. Mouse singleꢀdose PK parameters in plasma (i.v. bolus administration), in vitro Pꢀgp
efflux ratios (ER) and plasma free fractions (f ) of compounds 3, 7ꢀ9, 22, 34 and 37.
u_p
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 74
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
Brain/plasma
Total
P-gp ER
fu_p [%]
i.v. dose Terminal
Vss
total
Compd
plasma Cl
mL/min/kg]
[
mg/kg]
t1/2 [h]
[L/kg] concentration
[
a
human mouse human mouse
ratio (AUC )
3
7
1.1
0.9
1.1
1.3
0.8
0.30
0.24
0.27
0.39
0.43
95
1.4
4.2
2.0
2.1
1.7
3.6
3.6
ꢀ
ꢀ
1.4
24
12
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
250
114
207
541
1.1
1.1
1.0
1.1
1.1
1.5
0.9
1.3
27
10
12
17
8
1.9
9
5.2
22
13.5
16
3
3
4
7
1.0
1.0
0.39
0.20
84
83
2.7
1.2
1.3
0.5
1.1
1.7
1.2
1.9
14
17
11
14
a
AUC: area under the concentrationꢀtime curve.
Tritiation of lead tracer molecules. In order to allow for a detailed and high resolution in
vitro autoradiographical characterization of the novel tracer candidates, the molecules were
initially labelled with tritium. Five out of the six compounds were labelled by a hydrogenꢀtritium
exchange reaction directed by the adjacent nitrogenꢀcontaining heterocycles. The wellꢀknown
Crabtree’s catalyst ([Ir(COD)(PCy )(Py)]PF ) was used in superstoichiometric amounts to
3
6
3
3
3
promote the labeling and achieve high specific activity (Scheme 6). Thus, [ H]7, [ H]9, [ H]22,
3
3
[
H]34 and [ H]37 were obtained in specific activities ranging from 900 to 2000 GBq/mmol. The
position of the label was assigned based on mechanistic considerations, knowing that CꢀH
activation and isotope exchange require the proximity of an heteroatom for complexation of the
III
39
Irꢀatom and preferably proceed via fiveꢀmembered cyclometalated Ir species. Radioligand
ACS Paragon Plus Environment

Page 19 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
3
[
[
H]8 was prepared in high specific activity (SA = 2600 GBq/mmol) by methylation of 32 with
H]methylnosylate.
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
3
3
Scheme 6: Representative tritiations as exemplified for the preparation of [ H]8, [ H]9
3
a
and [ H]37
a
[3H]8
[3H]9
3
9
3
2
b
b
[3
H]37
7
a
3
3
Reagents and conditions: (a) [ H]MeONs, Cs CO , toluene, rt, 24 h; (b) [ H]H ,
2
3
2
[
Ir(COD)(PCy )(Py)]PF , DCM/DMF, rt, 4 h.
3
6
In vitro autoradiography of tritiated compounds and selectivity testing by antibody co-
localization studies. All tritiated compounds were evaluated by in vitro autoradiography using
native freshꢀfrozen brain sections originating from lateꢀstage AD patients (with high tau load)
and healthy control subjects (devoid of tau pathology) in order to evaluate i) tracer binding to tau
aggregates, ii) nonꢀspecific binding (NSB) and iii) lack of binding to other targets, in particular
to Aβ plaques. Analogous experiments performed on brain sections origination from two
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 74
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
transgenic mouse tauopathy models, Tg4510 and TauPS2APP, revealed a lack of binding of the
3
radioligands to the forms of tau present in these animals. Results for the reference [ H]3 are
included here for comparison, although extensive characterization and validation of this tool
compound was done earlier in the project. The results are summarized in Figure 2. All
radioligands were tested at a concentration of 3 nM. On first visual inspection of the macroscopic
autoradiographies, all novel radioligands displayed a binding pattern in AD tissue sections
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
similar to the reference [ H]3. This binding pattern in cortical tissue sections is characterized by
a layered distribution and correlates with the distribution of tau aggregates (neurofibrillary
tangles and neuropil threads) as visualized by immunohistochemical analysis with a tauꢀspecific
antibody (Figure 3A). Specificity of radioligand binding to tau aggregates and absence of
significant offꢀtarget binding to any other CNS target was suggested by very low binding to
cortical tissue sections from healthy control subjects (Figure 2 bottom row). Among the six
3
candidates some differences in NSB levels in healthy control tissue was observed. [ H]9
displayed the lowest NSB among all tracer candidates. Specific binding (SB) to tau aggregates
was assessed by quantifying the difference in binding to tauꢀrich cortical grey matter (GM) in
AD tissue versus tauꢀpoor cortical tissue of healthy controls. Highest SB values of 9726 fmol/mg
3
3
3
protein were identified for [ H]7 followed by [ H]9 and [ H]8 (SB = 5905 and 5500 fmol/mg
protein, respectively). Overall, SB correlated well with the affinity of the compounds as
3
determined by the competition of [ H]3 binding (Table 1). The binding of all six radioligand
3
candidates in advanced AD (Braak V staging ) tissue sections was also characterized by low
NSB in white matter (WM) yielding excellent signalꢀtoꢀnoise ratios (SNR) and image contrast in
vitro. The SNR was quantitatively assessed by calculating GM over WM ratios of radioligand
binding to AD Braak V cortical tissue sections. All compounds possessed favorable GM/WM
ACS Paragon Plus Environment

Page 21 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
3
ratios above or equal to 5, with the highest values found for [ H]34 (GM/WM = 14) and [ H]7
3
and [ H]9 (GM/WM = 11).
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
3
Figure 2. In vitro autoradiography of the reference tracer [ H]3 and all six test ligands (7ꢀ9,
2
2, 34 and 37) in tritiated form using native cortical tissue sections from an AD Braak V patient
(
top row) and a healthy control subject (bottom row). The radioligand concentration was 3 nM
for all compounds.
High binding selectivity of the tritiated ligands for tau versus Aβ aggregates was demonstrated
by coꢀlocalization experiments, assessed by autoradiography and immunohistochemistry on the
same tissue section on both macroscopic and microscopic scale. For immunohistochemistry
40
staining of NFTs and Aβ plaques the proprietary tauꢀ and Aβꢀ specific antibodies pS422 and
4
1
BAPꢀ2 were used. All compounds possessed excellent selectivity for aggregated tau without
3
any evidence of binding to Aβ plaques. Representative images of one radioligand ([ H]8) are
shown in Figure 3.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 74
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
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
40
41
Figure 3. Coꢀlocalization of tau mAb pS422
(A) and Aβ mAb BAPꢀ2
(C)
3
immunohistochemical staining with autoradiographical tracer binding (B) of [ H]8, as a
representative example. Staining and binding experiments were performed on a single freshꢀ
frozen tissue section from an AD Braak V cortical tissue block. The detail (D) shows the
autoradiography superimposed with tau mAb (red) and Aβ mAb staining (green). The black
radioactive binding signal coꢀlocalizes mainly with the tau mAb staining.
ACS Paragon Plus Environment

Page 23 of 74
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
4
0
41
Figure 4. Coꢀlocalization of tau mAb pS422
(A) and Aβ mAb BAPꢀ2
(C)
3
immunohistochemical staining with microautoradiographical tracer binding (B) of [ H]8, as a
representative example. Staining and binding experiments were performed on a single freshꢀ
frozen tissue section from an AD Braak V cortical tissue block. The detail (D) shows the
microautoradiographical signal superimposed with tau mAb staining (red). The black silver grain
deposits (representing radioligand binding) coꢀlocalize with tau mAb staining while areas of Aβ
mAb staining (dotted circle in B) do not coꢀlocalize with microautoradiographical signals.
1
1
18
Radiochemical synthesis of [ C]- and [ F]labelled compounds. A radiosynthesis was
established for five out of the six lead tracer candidates to introduce a fluorineꢀ18 or a carbonꢀ11
nuclide (Scheme 7). In the case of compound 37, offering in principle the possibility for labelling
with any of the two radioisotopes, fluorineꢀ18 was preferred, on account of the longer decay
halfꢀlife, facilitating the logistics of tracer use, and lower positron emission energy, leading to
higher spatial resolution of PET images. A robust radiomethylation procedure for compound 22
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 74
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
could not be established in the short timeꢀframe of the project, and further work to progress this
candidate was thus stopped.
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
Scheme 7: Synthesis of carbon-11 and fluorine-18 tracers
a
11
1
8
2
[ C]7
b
1
1
3
[ C]8
c
a
d
1
8
4
9
[ F]9
1
1
35
[ C]34
1
8
38
[ F]37
a
11
11
Reagents and conditions: (a) [ C]MeI, NaH, DMF, 80 °C, 3 min; (b) [ C]MeI, 6N NaOH,
1
8
18
DMSO, 80 °C, 3 min; (c) [ F]KF, Kryptofix 2.2.2, DMSO, ꢁw 50 Watts, 4 min; (d) [ F]KF,
Kryptofix 2.2.2, DMSO, ꢁw 50 Watts, 80 s.
1
1
The final product [ C]7 was isolated in an average nonꢀdecay corrected yield of 26% (n = 3)
in 33 minutes. The final product average specific radioactivity was 15987 mCi/µmole (592
GBq/µmole) with a radiochemical purity of 97%. The average concentration of carrier 7 and
precursor was 0.132 µg/mL and 0.009 µg/mL, respectively. Initial developmental radiosyntheses
11
showed [ C]7 was susceptible to radiolysis during the standard semipreparative HPLC and solid
ACS Paragon Plus Environment

Page 25 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
phase extraction (SPE) formulation. By changing the HPLC solvent from an
acetonitrile/triethylamine buffer to an ethanol/triethylamine buffer eluent system, the radiolysis
during the semipreparative HPLC was reduced dramatically. Also, a significant reduction in
radiolysis was observed by including ascorbic acid during the SPE isolation and formulation
procedure. No significant radiolysis was observed during the validation 40ꢀminute stability tests.
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
1
The final product [ C]8 was isolated in an average nonꢀdecay corrected yield of 6.3% (n = 16)
in 33 minutes. The final product had an average specific radioactivity of 14080 mCi/µmole (521
GBq/µmole) and radiochemical purity of 95.5%. The average concentration of carrier product
and precursor was 0.099 µg/mL and 0.009 µg/mL, respectively. As this radiotracer product was
also susceptible to radiolysis, ascorbic acid was used in the formulation to eliminate this
degradation.
18
Using the microwave radiofluorination module, [ F]9 was isolated in an average nonꢀdecay
corrected yield of 20.3% (n = 42) in 59 minutes. The final product had average specific
radioactivity of 41901 mCi/µmole (1550 GBq/µmole) and radiochemical purity of 100%. The
average concentration of carrier and precursor was 0.199 µg/mL and 0.010 µg/mL, respectively.
Since many PET radiotracer programs do not have access to microwave heating for
18
radiofluorinations, we selected to test the synthesis of [ F]9 using an easily programmed,
1
8
flexible radiosynthesizer platform (Sofie ELIXYS). Using the ELIXYS, [ F]9 was isolated in an
average nonꢀdecay corrected yield of 30.8% (n = 5) in 70 minutes. The final product had an
average specific radioactivity of 141720 mCi/µmole (5243 GBq/µmole) and radiochemical
purity of 100%. The average concentration of carrier and precursor was 0.0342 µg/mL and
0.0314 µg/mL, respectively.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 74
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
11
[
C]34 had an average nonꢀdecay corrected yield of 7.5% (n = 4) in 33 minutes with an
average specific radioactivity of 16848 mCi/µmole (610 GBq/µmole) and radiochemical purity
of 97.7 %. The average concentration of carrier and precursor 0.059 µg/mL and 0.003 µg/mL,
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
11
respectively. In the initial radiosynthesis of [ C]34, analytical HPLC showed the final product
1
1
with less than 90% radiochemical purity. As was done with [ C]8, the addition of ascorbic acid
in the product collection flask of the SPE formulation system increased the radiochemical and
chemical purity of the final radiotracer product. Minimal change in radiochemical purity was
observed in the validation 40ꢀminute stability test.
18
[
F]37 was isolated in an average nonꢀdecay corrected yield of 25% (n = 2) in 40 minutes.
The final product had an average specific radioactivity of 49750 mCi/µmole (1841 GBq/µmole)
and radiochemical purity of 96%. The average concentration of carrier was 0.057µg/mL and no
residual precursor was observed.
PET imaging in non-human primates.
PET studies in tauꢀnaïve baboons demonstrated favorable pharmacokinetic properties for three
1
1
11
18
out of five tracer candidates. [ C]7, [ C]8 and [ F]9 were characterized by good uptake in the
baboon brain with peak SUVs between 1.1 and 2.0 immediately after tracer injection and scan
18
initiation and rapid clearance of radioactivity from the baboon brain with [ F]9 showing the
fastest washout (Figure 5). All three tracer candidates were characterized by an homogeneous
distribution of the radioactivity in different brain regions, as expected in the absence of tau
1
1
aggregates in the animals. Tracer candidate [ C]34 displayed moderate brain uptake (peak SUV
1
.2) but the whole brain timeꢀactivity curve slightly increased at time points later than 20 min,
suggesting the generation and accumulation of a bloodꢀbrain barrier penetrating radiometabolite
ACS Paragon Plus Environment

Page 27 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
18
which counteracts the washout of the parent compound. The timeꢀactivity curve for [ F]37
showed minimal brain uptake, followed by a steady increase of radioactivity uptake in the brain,
which clearly points to the generation of a lipophilic radiometabolite that penetrates the bloodꢀ
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
11
brain barrier and slowly accumulates in the brain. The unfavorable kinetic profile of [ C]34 and
18
[ F]37 led to the exclusion of these two tracer candidates from further development.
Figure 5. Timeꢀactivity curves of the whole gray matter of candidate radioligands as
determined in a baboon PET study.
Conclusions
In summary, a lead optimization program has led to the identification of six promising PET
tracer candidates for imaging of tau aggregates in AD. These novel structures possess high
specificity and affinity for tau aggregates and an excellent selectivity against Aβ plaques. Further
11
11
18
characterization of the compounds indicated the three tracers [ C]7, [ C]8 and [ F]9 to be of
particular interest. In addition to the highly favorable binding properties to pathological tau
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 74
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
aggregates, these three radioligands possess appropriated pharmacokinetic properties and good
brain uptake in rodents and nonꢀhuman primates, followed by fast systemic clearance. Based on
the favorable profile, the three radiotracers were progressed into clinical validation studies in
Alzheimer’s disease patients (ClinicalTrials.gov Identifier: NCT02187627). The results of this
work will be communicated in due course.
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
EXPERIMENTAL SECTION
Chemical synthesis of cold compounds.
General methods. Unless otherwise noted, all reagents and chemicals were obtained from
commercial suppliers and used without further purification. Nonꢀaqueous reactions were
carried out under an inert atmosphere of Ar or N . Microwave heating of reactions was carried
2
out on a Biotage Initiator apparatus. Reactions were monitored by TLC (TLC plates F254,
Merck) or LC−MS (liquid chromatography−mass spectrometry) analysis. Flash column
chromatography was carried out either using cartridges packed with silica gel (Isolute columns,
Telos flash columns) on an ISCO machine or on glass columns on silica gel 60 (32−60 mesh, 60
Å). The purity of final compounds was determined by HPLC and was at least above 95%. LC
highꢀresolution MS spectra were recorded with an Agilent LC system consisting of an Agilent
1290 highꢀpressure system, a CTC PAL autosampler, and an Agilent 6520 QTOF. The
separation was achieved on a Zorbax Eclipse Plus C18 1.7 ꢁm, 2.1 × 50 mm column at 55 °C (A
= 0.01% formic acid in water; B = 0.01% formic acid in acetonitrile) at a flow of 1 mL/min with
1
the following gradient: 0 min, 5% B; 0.3 min, 5% B; 4.5 min, 99% B; 5 min, 99% B. HꢀNMR
spectra were measured on a Bruker 300 MHz instrument, a Bruker 600 MHz instrument in a 5
mm TCI cryoprobe at 298 K, or a Bruker Avance 400 MHz. A TMS internal standard was used
ACS Paragon Plus Environment

Page 29 of 74
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for experiments done in CDCl . The deuterated DMSOꢀd6 solvent signal was used as the
3
reference with 2.50 ppm.
2ꢀ(4ꢀMethoxyphenyl)imidazo[1,2ꢀa]pyridinꢀ7ꢀamine (7). A vial was charged with pyridineꢀ2,4ꢀ
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
diamine (11) (400 mg, 3.67 mmol), 2ꢀbromoꢀ1ꢀ(4ꢀmethoxyphenyl)ethanone (14) (882 mg, 3.85
mmol), NaHCO (329 mg, 3.92 mmol) and MeOH (3.5 mL). The reaction mixture was stirred at
3
reflux for 4 h. After cooling to rt, the mixture was diluted with water and EtOAc, sonicated and
stirred at room temperature for 15 min. The solids were collected by filtration, then rinsed with
H O and EtOAc. The resulting pale yellow solid was dried under high vacuum to afford 2ꢀ(4ꢀ
2
methoxyphenyl)imidazo[1,2ꢀa]pyridinꢀ7ꢀamine (7) hydrobromide. This material was suspended
in 5 mL saturated aq. NaHCO , sonicated, filtered and washed with H O. The residue was
3
2
suspended in 5 mL aq. 2 M NaOH, sonicated, filtered and rinsed with water. The resulting
residue was put under high vacuum to afford 2ꢀ(4ꢀmethoxyphenyl)imidazo[1,2ꢀa]pyridinꢀ7ꢀ
1
amine (7) as a light brown solid (310 mg, 35 %). H NMR (300 MHz, DMSOꢀd ) δ 8.09 (d, J =
6
7.07 Hz, 1H), 7.83 (s, 1H), 7.77 (d, J = 8.68 Hz, 2H), 6.94 (d, J = 8.68 Hz, 2H), 6.31ꢀ6.36 (m,
+
+
2
H), 5.60 (s, 2H), 3.77 (s, 3H); LCꢀHRMS (m/z): [M+H] calcd for C H N O+H : 240.1132,
14 13 3
found: 240.1134, Diff 0.2 mDa.
NꢀMethylꢀ2ꢀ(3ꢀmethylphenyl)imidazo[1,2ꢀa]pyrimidinꢀ7ꢀamine (8). N4ꢀMethylpyrimidineꢀ2,4ꢀ
diamine (25) (900 mg, 7.25 mmol) and 2ꢀbromoꢀ1ꢀmꢀtolylethanone (1.62 g, 7.61 mmol) were
combined with acetone (14 mL) under Ar and stirred overnight at 65 °C. The reaction mixture
was cooled to rt and extracted with DCM and saturated aq. NaHCO . The aq. layer was backꢀ
3
extracted four times with DCM adding each time a small amount of MeOH. The solid still
present in the aq. layer was filtered off. The organic layers were combined, dried over Na SO ,
2
4
filtered, combined with the filtered solid and concentrated. The residue was absorbed on 12 g
ACS Paragon Plus Environment