A Kinome-Wide Selective Radiolabeled TrkB/C Inhibitor for in Vitro and in Vivo Neuroimaging: Synthesis, Preclinical Evaluation, and First-in-Human

Article abstract of DOI:10.1021/acs.jmedchem.7b00396

The proto-oncogenes NTRK1/2/3 encode the tropomyosin receptor kinases TrkA/B/C which play pivotal roles in neurobiology and cancer. We describe herein the discovery of [¹¹C]-(R)-3 ([¹¹C]-(R)-IPMICF16), a first-in-class positron emission tomography (PET) TrkB/C-targeting radiolabeled kinase inhibitor lead. Relying on extensive human kinome vetting, we show that (R)-3 is the most potent and most selective TrkB/C inhibitor characterized to date. It is demonstrated that [¹¹C]-(R)-3 readily crosses the blood-brain barrier (BBB) in rodents and selectively binds to TrkB/C receptors in vivo, as evidenced by entrectinib blocking studies. Substantial TrkB/C-specific binding in human brain tissue is observed in vitro, with specific reduction in the hippocampus of Alzheimer's disease (AD) versus healthy brains. We additionally provide preliminary translational data regarding the brain disposition of [¹¹C]-(R)-3 in primates including first-in-human assessment. These results illustrate for the first time the use of a kinome-wide selective radioactive chemical probe for endogenous kinase PET neuroimaging in human.

Full text of DOI:10.1021/acs.jmedchem.7b00396

Subscriber access provided by AUSTRALIAN NATIONAL UNIV
Article
A Kinome-Wide Selective Radiolabeled TrkB/C Inhibitor for in Vitro and in
Vivo Neuroimaging: Synthesis, Preclinical Evaluation and First-in-Human
Vadim Bernard-Gauthier, Justin J. Bailey, Andrew V. Mossine, Simon Klaus Lindner, Lena
Vomacka, Arturo Aliaga, Xia Shao, Carole A. Quesada, Phillip S. Sherman, Anne Mahringer, Alexey
Kostikov, Marilyn Grand'Maison, Pedro Rosa-Neto, Jean-Paul Soucy, Alexander Thiel, David R.
Kaplan, Gert Fricker, Björn Wängler, Peter Bartenstein, Ralf Schirrmacher, and Peter J. H. Scott
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00396 • Publication Date (Web): 11 Jul 2017
Downloaded from http://pubs.acs.org on July 11, 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 55
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 55
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 KinomeꢀWide Selective Radiolabeled TrkB/C Inhibitor for in Vitro
and in Vivo Neuroimaging: Synthesis, Preclinical Evaluation and Firstꢀ
inꢀHuman
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,1*
a
b
c
Vadim Bernard-Gauthier , Justin J. Bailey , Andrew V. Mossine , Simon Lindner , Lena
c
d
b
b
b
Vomacka , Arturo Aliaga , Xia Shao , Carole A. Quesada , Phillip Sherman , Anne
e
f
g
d
h
Mahringer , Alexey Kostikov , Marilyn Grand’Maison , Pedro Rosa-Neto , Jean-Paul Soucy ,
f,i
j,k
e
l
c
Alexander Thiel , David R. Kaplan , Gert Fricker , Björn Wängler, Peter Bartenstein , Ralf
a,1,2*
b,m,2
Schirrmacher
, Peter J. H. Scott
a
Department of Oncology, Division of Oncological Imaging, University of Alberta, Edmonton, AB,
T6G 2R3, Canada.
b
Division of Nuclear Medicine, Department of Radiology, The University of Michigan Medical
School, Ann Arbor, MI, 48109, United States.
c
Department of Nuclear Medicine, LudwigꢀMaximiliansꢀUniversity of Munich, Marchioninistr. 15,
Munich, 81377, Germany.
d
Translational Neuroimaging Laboratory, McGill Centre for Studies in Aging, Douglas Mental
Health University Institute, 6875 Boulevard LaSalle, Montreal, QC, H4H 1R3, Canada.
e
Institute of Pharmacy and Molecular Biotechnology, University of Heidelberg, Heidelberg
69120, Germany.
f
McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, 3801
University Street, Montreal, QC, H3A 2B4, Canada.
g
Biospective Inc., 6100 Avenue Royalmount, Montreal, QC H4P 2R2, Canada.
h
Department of Neurology and Neurosurgery, Montreal Neurological Institute, McGill University,
3801 University Street, Montreal, QC, H3A 2B4, Canada.
ACS Paragon Plus Environment

Page 3 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
i
j
Jewish General Hospital, Lady Davis Institute, Montreal, QC, HT3 1E2, Canada.
Program in Neurosciences and Mental Health, Hospital for Sick Children, Toronto, ON, M5G
0
A4, Canada.
k
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
Department of Molecular Genetics, University of Toronto, ON, M5S1A8, Canada
l
Molecular Imaging and Radiochemistry, Department of Clinical Radiology and Nuclear
Medicine, Medical Faculty Mannheim of Heidelberg University, TheodorꢀKutzerꢀUfer 1ꢀ3,
Mannheim 68167, Germany.
m
The Interdepartmental Program in Medicinal Chemistry, University of Michigan, Ann Arbor, MI
48109, United States.
1
Corresponding authors:
2
P.J.H.S. and R.S. contributed equally to this work.
Keywords: Tropomyosin receptor kinases, Trk, positron emission tomography, PET,
neuroimaging, tyrosine kinase inhibitor, Alzheimer’s disease, NTRK.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 55
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
ABSTRACT
The protoꢀoncogenes NTRK1/2/3 encode the tropomyosin receptor kinases TrkA/B/C which
11
play pivotal roles in neurobiology and cancer. We describe herein the discovery of [ C]ꢀ(R)ꢀ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
11
([ C]ꢀ(R)ꢀIPMICF16), a firstꢀinꢀclass positron emission tomography (PET) TrkB/Cꢀtargeting
radiolabeled kinase inhibitor lead. Relying on extensive human kinome vetting, we show that
(
R)ꢀ3 is the most potent and most selective TrkB/C inhibitor characterized to date. It is
11
demonstrated that [ C]ꢀ(R)ꢀ3 readily crosses the bloodꢀbrain barrier (BBB) in rodents and
selectively binds to TrkB/C receptors in vivo, as evidenced by entrectinib blocking studies.
Substantial TrkB/Cꢀspecific binding in human brain tissue is observed in vitro, with specific
reduction in the hippocampus of Alzheimer’s disease (AD) versus healthy brains. We
11
additionally provide preliminary translational data regarding the brain disposition of [ C]ꢀ(R)ꢀ3 in
primates including firstꢀinꢀhuman assessment. These results illustrate for the first time the use of
a kinomeꢀwide selective radioactive chemical probe for endogenous kinase PET neuroimaging
in human.
ACS Paragon Plus Environment

Page 5 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
Specific interaction between the fullꢀlength catalytic tropomyosin receptor kinases TrkA/B/C
encoded by NTRK1-3 respectively) and their cognate neurotrophin ligands (nerve growth factor
(
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
(NGF), brainꢀderived neurotrophic factor (BDNF) and neurotrophinꢀ3 (NTꢀ3), respectively)
crucially supports differentiation, growth and survival within distinct neuronal populations of the
mammalian central (CNS) and peripheral nervous systems (PNS) as well as nonꢀneural tissues
1ꢀ3
during development and adult life. Reduced expression or impaired signaling of Trk tyrosine
kinase (TK) receptors in the CNS is found in an extensive spectrum of disorders and
pathologies including ischemic brain injury, schizophrenia, Rett syndrome, depression and
2ꢀ14
various neurodegenerative diseases.
In addition to their central roles in neurobiology, all
three members of the Trk family are protoꢀoncoproteins with wellꢀdefined oncogenic potential in
15ꢀ19
human, both in neural and nonꢀneural neoplasms.
So far, the validation of Trk changes
during aging and neurodegenerative diseases has solely originated from studies relying on
11ꢀ13,
immunohistochemical (IHC) detection and in situ hybridization (ISH) of postꢀmortem tissue
20ꢀ24
while preclinical assessment of brain exposure for anticancer Trk inhibitors (Figure 1 and
25ꢀ26
Table S1) has been circumscribed to the use of destructive methods
a nonꢀinvasive approach to assess and quantify Trk levels in vivo.
due to the absence of
Quantitative neuroimaging of proteins in humans using positron emission tomography
27
(
PET) is a robust in vivo approach which relies on shortꢀlived radiolabeled small molecules.
Most clinical CNS radiotracers validated to date target G proteinꢀcoupled receptors (GPCR), ion
28
channels, transporters and abnormally conformed proteins. Clinical neuroimaging of CNS
endogenous kinases has not yet been achieved despite key realisations in a few other enzyme
28ꢀ30
classes.
Given the importance of Trk (especially TrkB/C, which are most abundant in the
CNS and expressed in overlapping neuron populations), we sought to develop a radiolabeled
targeted kinase inhibitor for PET imaging to enable the visualisation and quantification of TrkB/C
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 55
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
receptor density in healthy and diseased brains. An important advantage of the selection of a
radiolabeled tyrosine kinase inhibitor (TKI) rather than an extracellular binding compound
resides in the intrinsic capacity of such probes to dissect proꢀsurvival neuronal fullꢀlength TrkB/C
receptors from catalyticallyꢀincompetent truncated isoforms (most notably TrkB.T1) largely
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
31ꢀ32
present in glia.
11
Herein, we describe our radiotracer development effort in the assessment of [ C]ꢀ(R)ꢀ3
11
([ C]ꢀ(R)ꢀIPMICF16) as a firstꢀinꢀclass brain penetrating TrkB/Cꢀtargeted lead for PET
neuroimaging (Figure 1). The work presented includes the radiotracer structureꢀactivity
11
relationship (SAR) characterization and comparative analysis leading to [ C]ꢀ(R)ꢀ3, the
enantioselective synthesis of (R)ꢀ3 and labeling precursor thereof, as well as the entire in vivo
imaging translational work in four species from early evaluation in rats and mice, preclinical nonꢀ
human primates as well as firstꢀinꢀhuman clinical PET imaging. We concomitantly provide proofꢀ
11
ofꢀprinciple of in vivo TrkB/C target engagement in the living brain for [ C]ꢀ(R)ꢀ3 and TrkB/C
receptor occupancy with pharmacological preꢀdosing of the phase II clinical lead entrectinib
using imaging experiments in mice. Particularly, we also demonstrate that the radiotracer
displays high TrkB/Cꢀspecific binding in human brain tissue and observes a distinct regional
binding pattern in Alzheimer’s disease (AD) versus healthy control (HC) in vitro.
ACS Paragon Plus Environment

Page 7 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RESULTS
Chemistry, biochemical evaluations and radiochemistry. Optimal compound triage for the
development of neuroimaging PET radioligands entails the alignment of favorable
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
physicochemical properties (molecular weight (MW) < 500 Da, topological polar surface area
2
(TPSA) ≤ 80 Å , hydrogen bond donor (HBD) ≤ 1, 2 ≤ clogD < 4, 3 ≤ clogP < 5, pK ≤ 8, rotatable
a
bonds (RBs) < 5, flexibility for radiolabeling with carbonꢀ11 (t1/2 = 20.3 min) or fluorineꢀ18 (t1/2
=
109.8 min)) and pharmacological parameters (low nano or picomolar affinity, Bmax/K > 10, > 30ꢀ
D
28, 33ꢀ35
100ꢀfold target(s) selectivity).
We applied those principles in our selection process for radiotracer development from a
recently reported library of 6ꢀ(2ꢀ(3ꢀfluorophenyl)pyrrolidinꢀ1ꢀyl)imidazo[1,2ꢀb]pyridazine panꢀTrk
36
11
inhibitors. The validation of [ C]ꢀ(R)ꢀ3 as a potential clinical radiotracer emerged from the
preclinical validation of a series of racemic radiolabeled derivatives with picomolar potencies for
1
1
TrkB/C which included [ C]ꢀ(±)ꢀ3 (Figure 1, Figure S1). A concomitant comparative advanced
11
11
37
preclinical evaluation of the radiolabeled panꢀTrk inhibitor [ C]5 ([ C]GW441756) (Figure 1)
is provided. Following preclinical validation in rodents and nonꢀhuman primates (vide infra),
11
radiotracer [ C]ꢀ(±)ꢀ3 was selected for structural refinement and the enantioselective synthesis
36, 38
of the putatively fitting Rꢀenantiomer was undertaken (Scheme 1, Figure 2a,b).
The key
(R)ꢀ2ꢀ(3ꢀfluorophenyl)pyrrolidine ((R)ꢀ11) intermediate required for securing (R)ꢀ3 was
synthesized via a Ellman sulfinamide route in place of the previously described enantioselective
lithiation as the key step due to the limited availability of (ꢀ)ꢀsparteine required using the latter
39
technique (Scheme 1). Diastereoselective Grignard addition to γꢀchloro Nꢀsulfinyl aldimine
40
(R )ꢀ8 delivered the diastereomer (R ,R)ꢀ9 in 72% yield as the unique addition product
S
S
(structure confirmed by Xꢀray crystallography, Scheme 1). Baseꢀpromoted ring closure followed
by sulfinyl group cleavage cleanly afforded pyrrolidine (R)ꢀ11 on gram scale providing an
efficient and mild alternative to the known and less scalable secꢀBuLi/lithiation sequence. With
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 55
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
(
R)ꢀ11 in hand, the carboxylic acid (R)ꢀ12 (ee > 99.6%; chiral SFC/MS, Supplementary
Information Section 3) was obtained and derivatized with the required anilines for the
synthesis of the nonꢀradioactive standard and norꢀprecursor (R)ꢀ14.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The nonꢀradioactive inhibitor (R)ꢀ3 displays halfꢀmaximal inhibitory concentrations (IC s)
50
33
of 4.0, 0.2 and 0.1 nM for human TrkA, TrkB and TrkC, respectively, in [γꢀ P]ATPꢀbased
enzymatic assays (Figure 2c). Inhibitory measurements at Km ATP and conversion using
41
ChengꢀPrusoff analysis for competitive inhibition gave inhibitory constants (K) values of 2.80 ±
i
0
.16 nM, 0.050 ± 0.005 nM and 0.021 ± 0.011 nM for TrkA, TrkB and TrkC respectively (n = 3,
42
s.d.) (~200,000ꢀ500,000ꢀfold over K ATP for Trk) . (R)ꢀ3 therefore exhibits notable intraꢀTrk
m
isoform selectivities of 56ꢀfold for TrkB and 133ꢀfold for TrkC with regard to TrkA. In order to
evaluate the broader selectivity profile of our lead, comprehensive kinome screening on a panel
of 369 human kinases was conducted at 0.2 ꢀM cutꢀoff (Reaction Biology Corp., full wild type
kinase panel). This assay revealed that (R)ꢀ3 exhibits >1000/2000ꢀfold TrkB/C selectivity for
9
9% of targets tested, with only four kinases (BMX, TXK, ACK1 and ROS1) at or above their
IC50 values for the tested concentration (Figure 2e, Supplementary Information Section 5).
Aside from BMX, which was inhibited approximately at IC50 with (R)ꢀ3 0.2 ꢀM, individual dose
response curves were generated for TXK, ACK1 and ROS1 revealing IC s of 106 nM, 62.6 nM
50
and 5.96 nM respectively (Figure 2c,e). (R)ꢀ3 exhibits ~30ꢀ and 60ꢀfold selectivity for TrkB and
TrkC respectively versus ROS1 – the closest offꢀtarget identified. As anticipated, (S)ꢀ3 displayed
still acceptable TrkB/C activities albeit in the nanomolar rather than the picomolar range (Figure
36, 38
S2).
11
Overall, for the purpose of this study, the radiosynthesis of [ C]ꢀ(R)ꢀ3 was implemented
at four distinct production sites including a Good Manufacturing Practice (GMP)ꢀcompliant
11
automated synthesis for firstꢀinꢀhuman PET imaging. For human PET studies, [ C]ꢀ(R)ꢀ3 was
11
prepared from (R)ꢀ14 using [ C]CH I in the presence of K CO in DMF with radiochemical
3
2
3
ACS Paragon Plus Environment

Page 9 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
yields (RCYs) of 13.0 ± 2.0 % (s.d., decayꢀcorrected RCYs at endꢀofꢀsynthesis based upon
11
[
C]CO , n = 3, > 99% radiochemical purity), and molar activity (A ) of 118 GBq/ꢀmol (Scheme
2
m
11
1). Throughout the translational imaging study, the radiosynthesis of [ C]ꢀ(R)ꢀ3 was highly
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
reproducible, robust and flexible to protocol adaptation (n > 40). In all cases, from preclinical to
11
clinical PET, the identity of all radiotracers including [ C]ꢀ(R)ꢀ3, was confirmed by nonꢀ
radioactive standard coꢀinjection using highꢀperformance liquid chromatography (HPLC)
(
Scheme 1, Supplementary Information Section 6).
11
11
In vivo PET imaging of [ C]-(R)-3 in rodent. Initial PET imaging experiments with [ C]ꢀ(±)ꢀ3
in Sprague Dawley rats as part of the racemic series screening confirmed this scaffold as a lead
based on overall brain penetration, brainꢀtoꢀblood ratio and stability (Figure S3aꢀc). The
11
radioactivity uptake of [ C]ꢀ(R)ꢀ3 in the FVB mouse brain (wildtype mice) was rapid and peak
brain uptake (SUVmax) of 1.33 ± 0.19 (s.d.) was reached within 1 min following intravenous
injection (i.v.) (SUV, standardized uptake value ꢀ normalization for injected dose and body
weight) as illustrated by the corresponding wholeꢀbrain timeꢀactivity curve (TAC) (Figure 3f).
The radiotracer was found extensively distributed within the CNS in accordance with reported
43
TrkB expression (Figure 3c,h). Peak brain radioactivity was followed by moderate washout
throughout the remainder of the 60 min scan (0.40 ± 0.02 SUV at 60 min post injection, s.d.,
ꢁSUVmaxꢀ60min = ꢀ 69%). Biodistribution imaging studies highlighted extensive liver uptake
suggesting hepatobiliary clearance as the primary route of excretion (Figure 5.2g).
11
Despite similar overall brain kinetics profile, with [ C]ꢀ(±)ꢀ3 lower SUV (1.04 ± 0.12 at
max
0.5 min, s.d.) were reached as compared to the enantiopure radiotracer (Figure 3a,b,e). These
11
results demonstrate that [ C]ꢀ(R)ꢀ3 readily diffuses through the bloodꢀbrain barrier (BBB) and
persists moderately in the rodent brain while suggesting faster elimination of the racemate.
11
Comparative analysis with [ C]5, the only available panꢀTrk radiotracer lead previously reported
11
was performed (Figure 3j). In vivo PET of [ C]5 in FVB mice revealed substantial and rapid
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 55
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 penetration (SUVmax of 4.14 ± 0.55 at 1 min, s.d.) which was followed by prompt and
extensive washout (0.14 ± 0.02 SUV at 60 min post injection, s.d., ꢁSUVmaxꢀ60min = ꢀ97%, Figure
37
3l) on par with previously reported data in rats. Renal clearance was more prominent with
11
11
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
[
C]5 compared to [ C]ꢀ(R)ꢀ3 (Figure 3m). With the expectation that Trk CNS binding should
be associated with tracer retention in the brain to allow for eventual quantification, the kinetics of
11
11
[
C]5, in contrast to [ C]ꢀ(R)ꢀ3, likely indicate limited target engagement in rodents due to rapid
subsequent washout, despite extensive initial permeation of the BBB. This observation can be
11
11
rationalized in part by the significantly higher potency of [ C]ꢀ(R)ꢀ3 compared to [ C]5 (34ꢀ46ꢀ
1
1
fold potency difference for TrkB/C). Metabolite analysis revealed [ C]ꢀ(R)ꢀ3 to be particularly
robust in vivo, with > 95% intact tracer 30 min post injection in mouse blood plasma (Figure 3i).
In order to investigate potential efflux, baseline scans were then conducted in
(−/−)
(−/−)
11
11
11
Mdr1a/b Bcrp1
knockout mice with both [ C]ꢀ(±)ꢀ3 and [ C]ꢀ(R)ꢀ3 (as well as [ C]ꢀ(±)ꢀ4 –
data presented in Figure S5.3 and Figure S1dꢀf). Increased wholeꢀbrain uptake was observed
for all tracers (Figure 3aꢀf) indicating that primary ABC transporter components Pꢀglycoprotein
(Pꢀgp –Mdr1a/b) and breast cancer resistance protein (BCRP –Bcrp1) at the BBB play a role in
elimination of the tracer from the brain in mice. In the case of 3 tracers (racemic and R), this
effect was most pronounced with the racemate compared to the R-enantiomer with SUV0ꢀ60min
(−/−)
(−/−)
11
11
(
(
FVB)/SUV0ꢀ60min (Mdr1a/b Bcrp1 ) ratios of 0.56 for [ C]ꢀ(R)ꢀ3 compared to 0.39 for [ C]ꢀ
±)ꢀ3 (Figure 3e,f). This result was however paralleled by opposite variations in blood
radioactivity which may in part explain interꢀtracer variations (Figure 3g, Figure S4). Whole
brain TACs displayed kinetics consistent with results obtained in FVB mice. CalceinꢀAM cellular
assay in Pꢀgp and BCRP overexpressing MadinꢀDarby Canine Kidney (MDCKII) cells and
LineweaverꢀBurk plot analysis confirmed (R)ꢀ3 as a Pꢀgp/BCRP substrate with moderate affinity
(
EC50 of 2.9 ± 1.3 ꢀM and 2.23 ± 1.1ꢀM respectively, s.d., Figure S5). Interestingly, while no
11
difference was observed in the brain distribution of [ C]5 under microdosing PET imaging
ACS Paragon Plus Environment

Page 11 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(−/−)
(−/−)
conditions between FVB and Mdr1a/b Bcrp1
mice groups (Figure 3jꢀl), the calceinꢀAM
assay outlined Pꢀgp interaction of 5 at higher doses (EC of 79.2 ± 8.3 ꢀM, s.d.) (Figure S5e).
50
In contrast to (R)ꢀ3, this disparity is consistent with the possibility that 5 acts as a weak Pꢀgp
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
inhibitor rather than a substrate.
Supported by recent nextꢀgeneration sequencing (NGS) efforts, oncogenic Trk fusions
have been described in over twenty tumor types, reinforcing the view that Trk chimeric proteins
2
are lowꢀfrequency but recurrent oncogenic drivers in human cancer . Yet, concerns over
neurotoxicity arising from sustained CNS Trk blockade upon treatment for peripheral conditions
15ꢀ16, 44ꢀ45
have been described
Current clinical trial efforts include genomicallyꢀdriven phase II
trials for NTRK fusionꢀpositive patient subpopulations based on promising phase I results with
4
6ꢀ47
entrectinib (Figure 3n, Figure 1 and Table S1).
In order to provide unambiguous evidence
for TrkB/C specific binding, we conducted Trk receptor heterologous blocking through
11
pharmacological challenge in vivo using this inhibitor. [ C]ꢀ(R)ꢀ3 brain uptake following
entrectinib administration (30 min intraperitoneal predosing) was investigated, simultaneously
providing confirmation of brain penetration and target engagement of a currently advanced panꢀ
Trk inhibitor clinical lead using a nonꢀinvasive technique. Pretreatment doses were selected
under and above a threshold of 240 mg/kg which was previously shown to achieve 0.43 brainꢀ
toꢀplasma ratio in mice with sustained treatment regimen (240 mg/kg/day in continuous 2 weeks
25
administration). We observed doseꢀdependent whole brain radioactivity reduction (SUV0ꢀ60min
)
of ꢀ7% and ꢀ40% with 50 mg/kg and 350 mg/kg pretreated mice (single doses) compared to
vehicleꢀtreated control respectively (Figure 3oꢀr). At the time of SUVmax (1 min post injection) in
the vehicleꢀtreated animals, brain concentration was reduced in animals treated with entrectinib
at pharmacological doses (50 and 350 mg/kg respectively) by 29% and 88% respectively. This
is consistent with a rapid brain penetration and target binding followed by a tracer washout. As
expected, trends in brainꢀtoꢀblood K ratios during dynamic scanning did not differ from SUV
p
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 55
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
measurements as entrectinib pretreatment did not affect blood radioactivity upon [ C]ꢀ(R)ꢀ3
injection (as measured at 50 mg/kg). Data from a CalceinꢀAM cellular assay with MDCKII cells
included herein also demonstrate that entrectinib is not a prominent substrate for Pꢀgp or BCRP
1
1
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 S5fꢀh). At this point, [ C]ꢀ(R)ꢀ3 emerged as a unique lead on the basis of favorable
pharmacology and imaging properties in FVB mice. In spite of its efflux liability at microdoses
(−/−)
(−/−)
11
described in Mdr1a/b Bcrp1
mice, the choice of [ C]ꢀ(R)ꢀ3 was supported by documented
interꢀspecies variation in intracerebral disposition of actively effluxed PET tracers between
rodents and higher species which was in line with our concurrently collected data in nonꢀhuman
33, 48ꢀ49
primates (vide infra).
11
Evaluation of [ C]-(R)-3 in the non-human primate brain. Radioactivity brain uptake
11
following [ C]ꢀ(R)ꢀ3 administration was higher in nonꢀhuman primates compared to rodents and
displayed distinctively slow brain kinetics characterized by steady radioactivity increase without
washout from brain tissue for the entire duration of the 60 min scan (Figure 4a, Figure S6e).
Based on volumetric region of interest (ROI) analysis, the uptake was shown to be
heterogeneously distributed, ubiquitous across gray matter regions and most pronounced in the
thalamus (~0.8 SUV at 60 min post injection) while white matter displayed comparatively low
accumulation (~0.3 SUV at 60 min post injection) matching known TrkB/C receptor distribution
(
thalamus ≥ cerebellum ≥ cortex >>> white matter) (Figure 4a, Figure S6e). Binding kinetics of
11
11
[
C]ꢀ(R)ꢀ3 were different from [ C]ꢀ(±)ꢀ3, exhibiting higher gray matter uptake compared to the
11
racemic tracer 60 min post injection (Figure 4b, Figure S6d,e). [ C]ꢀ(R)ꢀ3 also presented
11
significantly higher testꢀretest reproducibility of SUVs compared to [ C]ꢀ(±)ꢀ3. Brain
11
11
pharmacokinetics of [ C]5 contrasted with [ C]ꢀ(R)ꢀ3, reaching SUV
of 2.41 SUV in
max
cerebellum at 4 min post injection (2.17 SUV in thalamus at 2.5 min) followed by sustained
washout (Figure 4c and Figure S6f). Whole brain washout in primates was however reduced
11
11
compared to mice (0.62 SUV at 60 min post injection, ꢁSUVmaxꢀ60min = ꢀ65%). [ C]5 and [ C]ꢀ(R)ꢀ
ACS Paragon Plus Environment

Page 13 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
showed identical brain concentration 60 min post injection despite opposite kinetic trends.
Based on known CNS TrkB/C full length levels, regional grayꢀtoꢀwhite matter SUV ratios
SUVgray matter/SUVwhite matter) were used as an estimate for specific binding and tracer
prioritization. As illustrated in Figure 4d, specific signal approximations using summed SUV
(
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
11
ratios for 30ꢀ60 min post injection were markedly higher for [ C]ꢀ(R)ꢀ3 compared to [ C]ꢀ(±)ꢀ3
11
and [ C]5 in all gray matter regions with representative average values of 2.3 and 2.1 in the
thalamus and cerebellum respectively (similar trends were observed in 0ꢀ60 min summed data,
11
11
Figure S7). Despite [ C]5 displaying higher uptake than [ C]ꢀ(R)ꢀ3, the use of this tracer is
impeded by lower specific binding, inferior target selectivity and intrinsic contamination with an
11
37
inseparable [ C]ꢀ(E)ꢀ5 inactive photoisomer, which constitute major hurdles to clinical
11
translation. Taken together, the results obtained with [ C]ꢀ(R)ꢀ3 in primates are in contrast to
the efflux kinetic profile in rodents and are indicative of a moderate and continuous brain uptake
as well as favorable TrkB/C specific signal properties in the primate brain.
11
[
C]-(R)-3 binding in human post-mortem brain tissue. To further assess the binding
specificity and selectivity of our lead radiotracer in brain tissue as well as to provide proofꢀofꢀ
principle evidence of its potential use in the context of neurodegenerative diseases, in vitro
autoradiography experiments using postꢀmortem tissue from human healthy control (HC) and
11
agedꢀmatched AD patient groups were conducted. Binding properties of [ C]ꢀ(R)ꢀ3 were
measured in four brain regions (20 ꢀM cryosections) including the hippocampus which has been
1
3ꢀ14
shown to undergo drastic TrkB full length reduction in AD brains
as well as unaffected
50
regions with regards to TrkB expression (e.g. parietal cortex and cerebellum). In an initial
11
series of baseline experiments using HC human brains (n = 12ꢀ15), [ C]ꢀ(R)ꢀ3 displayed
extensive, yet heterogeneous total binding in all regions mirroring the nearꢀubiquitous
125
expression of TrkB/C in neurons of the mammalian brain as characterized using [ I]BDNF,
125
125
43
[
I]NTꢀ3 and [ I]NTꢀ4/5 and mRNA hybridization (Figure 5 and Figure S8). In order to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 55
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
evaluate the extent of specific binding (impact of brain lipids and proteins nonꢀspecific binding)
and TrkB/C selectivity (impact of offꢀtarget receptorꢀmediated interactions), competition blocking
experiments using the chemically distinct selective panꢀTrk type I inhibitor 5 (10 ꢀM) were
conducted in HC brains. Heterologous blocking was prominent in all analyzed ROIs (P ≤
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
0.0001) with the most pronounced effects observed in the prefrontal cortex (ꢁ ~ ꢀ71%, n = 15)
and hippocampus (ꢁ ~ ꢀ69%, n = 12), while significant albeit less pronounced blocking was
observed in the inferior parietal cortex (ꢁ ~ ꢀ43%, n = 15) and cerebellum (ꢁ ~ ꢀ54%, n = 15),
11
revealing high specific and TrkB/Cꢀselective binding of [ C]ꢀ(R)ꢀ3 throughout the brain (Figure
). Total or specific/selective binding interꢀindividual variability did not correlate with gender, age
or postꢀmortem delay in HC samples. Having demonstrated excellent TrkB/Cꢀbinding in human
5
11
healthy brains, measurements of total and specific/selective binding of [ C]ꢀ(R)ꢀ3 in postꢀ
mortem tissue from AD patients in the corresponding ROIs (n = 11ꢀ12) were performed. Total
tracer binding and specific/selectiveꢀTrkB/C binding were found to be significantly reduced in the
hippocampus of AD brains compared to HC (ꢁ ~ ꢀ35%, P ≤ 0.001) as reported using IHC and
13ꢀ14
ISH techniques.
With the exception of this effect, nonꢀspecific/nonꢀselective levels upon
blocking with 5 in AD brains were similar to the observations from HC brains suggesting strong
TrkB/C interaction in the AD group as well (P ≤ 0.0001). No additional noticeable differences
between AD and HC groups in total tracer binding or specific/selective binding were observed in
the other regions analyzed. Again, interꢀindividual binding differences within the AD brain
samples could not be ascribed to differences in gender, age or postꢀmortem delay.
11
Imaging TrkB/C in the human brain using [ C]-(R)-3. These preclinical data prompted the
11
evaluation of [ C]ꢀ(R)ꢀ3 as a radiotracer for use in humans in vivo. Firstꢀinꢀhuman brain uptake
11
measurements of [ C]ꢀ(R)ꢀ3 were performed in one healthy male volunteer (41 years of age).
Radiotracer brain uptake was prompt with SUVmax reached in all analyzed ROIs at 25 sec post
injection (including white matter) (1.5 SUV in the thalamus) followed by rapid stabilization and
ACS Paragon Plus Environment

Page 15 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
steady retention with negligible decrease during the 60 min scan. The uptake was
heterogeneous but prevalent within the gray matter regions with the highest uptake seen in the
thalamus (0.7 SUV) and lowest in white matter (0.4 SUV) with overall distribution aligning with
nonꢀhuman primate results and known TrkB/C expression (thalamus ≥ cerebellum ≥ cortex >>>
white matter) (Figure S9a and Figure 6).Uptake in gray matter regions was superior to white
matter at all time points. There were no adverse events or clinically detectable pharmacologic
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
effects associated with the injection of [ C]ꢀ(R)ꢀ3.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 55
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
DISCUSSION
The development of the first lead radiotracer for nonꢀinvasive PET imaging of TrkB/C receptors
11
in the living brain and postꢀmortem tissue is described. Human PET imaging with [ C]ꢀ(R)ꢀ3
was translated from preclinical research and demonstrated moderate brain uptake in agreement
with regional TrkB/C distribution and favorable grayꢀtoꢀwhite matter ratios. It was shown that
most obstacles commonly associated with the development of orthosteric probes for intracellular
in vivo neuroimaging of protein kinases such as i) the combination of favorable physicochemical
properties for rapid membrane diffusion (see result section), ii) the achievement of adequate
affinity in light of high intracellular ATP concentration (hence low available receptor density,
Bavail), as well as iii) the realization of sufficient target selectivity within the human protein kinome
which presents highly conserved ATP/inhibitorꢀbinding site can be overcome by thorough SAR
screening. Our biochemical examination showed that (R)ꢀ3 displays intrinsic affinity in the low
picomolar range for TrkB/C. Moreover, we demonstrated that (R)ꢀ3 displays exceptional kinome
selectivity in comprehensive screenings (including notably TrkA) well beyond the 30ꢀ100ꢀfold
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
28, 33
threshold considered necessary for CNS radiotracers.
Importantly, as in vivo selectivity
depends both on the affinity of the probe and the target density (Bmax), the low TrkA expression
compared to TrkB/C in the mammalian CNS combined with the 50ꢀ100ꢀfold selectivity in favor of
11
TrkB/C against TrkA infer that [ C]ꢀ(R)ꢀ3 in vivo binding should be highly restricted to TrkB/C
43, 51
without detectable contribution from TrkA receptor binding as intended.
Similar, unwanted
interaction with ROS1, the only target for which (R)ꢀ3 shows less than 100ꢀfold selectivity
identified to date, is not an impediment foreseen due to the lack of detectable ROS1 levels in
52
normal human brain tissue. Based on available data, (R)ꢀ3 displays both the highest affinity for
TrkB/C and the best kinome selectivity of all panꢀTrk inhibitors characterized to date and
therefore constitutes a prime lead for radiotracer development.
ACS Paragon Plus Environment

Page 17 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A central objective of the present report is to provide the most comprehensive overview
11
of the development of [ C]ꢀ(R)ꢀ3 from radiotracer design to human use. At its core, the work
presented here was enabled by the efficient enantioselective synthesis of the required nonꢀ
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
radioactive standard and norꢀprecursor as well as by the reliable radiosynthesis of [ C]ꢀ(R)ꢀ3. In
1
1
particular, we showed that [ C]ꢀ(R)ꢀ3 can be obtained in high molar activity, excellent
radiochemical purity as well as sufficient isolated RCYs both in manual radiosyntheses as well
as in fully GMPꢀcompliant automated radiosyntheses for use in clinical studies. In addition, with
extensive metabolism often emerging as a detrimental factor precluding radiotracer
11
development, we investigated the stability of [ C]ꢀ(R)ꢀ3 in vivo. Only < 5% of emerging polar
metabolites were detected during in vivo radioꢀTLC metabolism studies. No lipophilic
metabolites that could pass the BBB were observed. Our report also highlights the importance
of multiꢀtracer/multiꢀspecies translational studies in the context of PET radiotracer development.
11
To provide additional evidence of TrkB/C specificity and selectivity of [ C]ꢀ(R)ꢀ3 in brain
tissue, we determined radiotracer binding properties in autoradiography experiments using
healthy human brain tissue. In a first set of data collected using HC tissue, we showed
extensive specific and selective binding, using the selective panꢀTrk inhibitor 5 as a second
structurally distinct heterologous pharmacological competitor. Anomalies in neurotrophin/Trk
receptor systems have been described in various CNS diseases and conditions and are best
characterised in the context of AD. For instance, genome association analysis has suggested a
possible role for polymorphism of the genes for both BDNF and NTRK2 (TrkB) as risk factors in
53
54
AD, although this remains a subject of debate. Direct evidence of involvement of the
BDNF/TrkB system in AD comes from studies indicating that TrkB full length levels are
13
profoundly decreased in the hippocampus of patients with AD, although apparently not in the
50
parietal cortex. Moreover, progressive loss of TrkB signal (as well as of the TrkA/C variants) in
11
basal forebrain cholinergic nuclei is well correlated with the clinical progression of the disease.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 55
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
Treatment with agonists of the BDNF/TrkB system in transgenic mice models of AD results in an
increase of dendritic spines in the hippocampus and cortex, an inhibition of neuronal apoptosis
55ꢀ57
and neurodegeneration, and improves spatial memory performance.
We therefore used AD
11
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
as a proofꢀofꢀconcept for the application of [ C]ꢀ(R)ꢀ3 as quantitative radiotracer, and
11
determined a [ C]ꢀ(R)ꢀ3 specific binding profile in AD brains as a parallel study to our HC
experiments. Considering known trends previously measured using IHC, we show that
11
hippocampal [ C]ꢀ(R)ꢀ3 binding is significantly decreased in AD versus HC. Those results
indicate the potential use for Trk PET radiotracers in the study and diagnosis of
neurodegenerative diseases. While our study design did not allow for pharmacological blocking
in nonꢀhuman primates due to the lack of an approved Trk inhibitor (ideally also BBB
permeable) or other suitable tool inhibitors with prominent Trk activity and established safety
profile, multiple lines of evidence are provided illustrating a displaceable specific binding of
11
[
C]ꢀ(R)ꢀ3 both in vitro and in vivo together with concordant regional CNS distributions in the
human and nonꢀhuman primate brain. We recognize that the absolute brain uptake of our lead
radiotracer is moderate which imply that further structural optimization will be required to reach
optimal clinical utility. The absence of a goldꢀstandard radiotracer for comparison however
precludes the determination of what constitutes a standard brain uptake in this context.
Currently, the only other radiolabeled kinase inhibitor shown to engage its target in vivo is the
11
58
glycogen synthase kinaseꢀ3 (GSK3) tracer [ C]PFꢀ367. Notably, in a preclinical nonꢀhuman
11
primate study, this tracer displayed brain uptake comparable to [ C]ꢀ(R)ꢀ3 while not being
susceptible to Pꢀgp efflux. It is, at the moment, not possible to discern whether the observed
11
brain disposition of [ C]ꢀ(R)ꢀ3 in higher species is primarily related to active efflux transport or
other factors such as TrkB/C.FL Bmax which has not been defined in human tissue yet.
ACS Paragon Plus Environment

Page 19 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CONCLUSION
In summary, our study provides a molecular imaging tool tracer for the in vitro and in vivo nonꢀ
invasive study of neuronal signal transduction at the interface of neurology and oncology which
is as of yet unexplored. Importantly, the data described herein delineates, to the best of our
knowledge, the first use of a small molecule kinase inhibitor radiotracer as an in vivo probe to
explore endogenous kinase densities using PET neuroimaging in human. Using portꢀmortem
tissue, it is anticipated that our molecular probe should provide a rapid mean to clarify TrkB/C
expression levels in a wide range of neurological diseases wherein TrkB/C dysregulation has
long been suspected to play a prominent pathogenic role. In particular, when taking into account
the intrinsic in vivo translatability of our approach, the finding that the primary radiotracer
identified herein enables the discrimination of AD from HC brains could ultimately have
substantial implications for AD diagnostic beyond conventionally sought biomarkers. Given also
that the primary safety concern in the current efforts towards the development of Trk TKIs as
novel antineoplastic and antinociceptive drug classes remain the prospect of unwanted CNS
neurotoxicity via TrkB/C inhibition, our approach provides an immediate tool to screen CNS
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
45, 59ꢀ60
occupancy nonꢀinvasively in preclinical settings as shown for entrectinib here.
Our efforts
currently revolve around conducting iterative structural optimization in order to mitigate efflux
liabilities, improve brain pharmacokinetics, and identify second generation leads based on the
(
R)ꢀ2ꢀ(3ꢀfluorophenyl)pyrrolidinꢀ1ꢀyl)imidazo[1,2ꢀb]pyridazine scaffold to clarify the exact
determinants presiding to the observed brain uptake of in human and nonꢀhuman primates.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 55
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
EXPERIMENTAL SECTION
General Methods for Chemistry. All moisture sensitive reactions were carried out in ovenꢀ
dried flasks under nitrogen atmosphere with dry solvents. Reagents and solvents were
purchased at the highest commercial quality from Fisher, SigmaꢀAldrich, AlfaꢀAesar, Synthonix
or Oakwood Products and were used without further purification unless specified otherwise.
Compounds used for blocking PET studies, 5 hydrochloride and entrectinib, were purchased
from Aldrich (G3420) and Cedarlane (HYꢀ1267) respectively. Organic solutions were
concentrated under reduced pressure on a Heidolph rotary evaporator. In general, reactions
were magnetically stirred and monitored by TLC performed on preꢀcoated glassꢀbacked TLC
plates (Analtech, 250 microns) and chromatographic purification of products was accomplished
using flash chromatography on AlfaꢀAesar silica gel (230ꢀ450 mesh). TLC visualization 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
13
performed by fluorescence quenching, KMnO or ninhydrin. H NMR and C NMR spectra were
4
recorded on
a
Agilent/Varian DD2 MR two channel 400 MHz spectrometer,
a
Agilent/Varian VNMRS twoꢀchannel 500 MHz spectrometer or a Agilent/Varian Inova fourꢀ
channel 500 MHz spectrometer in CDCl or d ꢀDMSO and peak positions are given in parts per
3
6
million using TMS as internal standard. Peaks are reported as: s = singlet, d = doublet, t =
triplet, q = quartet, p = quintet, m = multiplet, b = broad; coupling constant(s) in Hz; integration.
High Resolution Mass Spectra (HRMS) analysis was obtained from the Mass Spectrometry
Facility of the Chemistry Department of the University of Alberta (Agilent Technologies 6220
oaTOF) or from the Regional Center for Mass Spectrometry of The Chemistry Department of
the Université de Montréal (LCꢀMSDꢀTOF Agilent). Compounds tested for biological evaluation
were >95% pure (HPLC). The chiral SFC/MS analysis was performed at the Regional Mass
Spectrometry Center of the department of chemistry of the Université de Montréal.
Crystallographic analysis was performed by the XꢀRay Diffraction Laboratory of the department
of chemistry of the Université de Montréal.
ACS Paragon Plus Environment

Page 21 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The syntheses of (±)ꢀ15, (±)ꢀ4 ((±)ꢀIPMICF22), (±)ꢀ2 ((±)ꢀIPMICF10), (±)ꢀ1 ((±)ꢀIPMICF6), (±)ꢀ3
36
and (±)ꢀ12 (Figure S2) have been reported in the literature. The synthetic sequences leading
to the radiolabeling precursors (±)ꢀ14 and (R)ꢀ14 as well as the compounds (R)ꢀ3/(S)ꢀ3 are
40
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
described here. The synthesis of compound (R )ꢀ8 has been described in the literature. The
S
absolute stereochemistry of (R ,R)ꢀ9 was confirmed by Xꢀray crystallography analysis.
S
Compounds (R)ꢀ12, (R)ꢀ13), (R)ꢀ14 and (R)ꢀ3 were synthesized in accordance with the methods
described for the corresponding racemic compounds. All R /NMR/MS data obtained for those
f
compounds were in full agreement with the corresponding known racemates. The synthesis of
(
S)ꢀ3 was carried out following the procedure described for (R)ꢀ3 with (S )ꢀ8 as starting material.
S
3
The synthesis of compound (S )ꢀ8 has been described in the literature . All R /NMR/MS data
S
f
obtained for compounds (S ,S)ꢀ9 and (S ,S)ꢀ10 were in full agreement with the corresponding
S
S
(
R ,R)ꢀenantiomers. The enantiopurities of the pyrrolidine C2 centers in both the (R)ꢀ and (S)ꢀ
S
synthetic sequences were further confirmed at the stage of the key intermediates (R)ꢀ12 and
1
13
(
S)ꢀ12 using chiral SFC/MS analysis (ee > 99.6%). H and C spectra of new compounds and
chiral SFC analysis are provided in SI Appendix.
Chemistry.
4-((tert-Butyldimethylsilyl)oxy)-3-fluoroaniline (17). tertꢀButyldimethylsilyl chloride (791 mg, 5.25
mmol, 1.05 equiv) was added to a solution of 4ꢀaminoꢀ2ꢀfluorophenol (636 mg, 5.0 mmol, 1.0
equiv) and imidazole (851 mg, 12.5 mmol, 2.5 equiv) in anhydrous acetonitrile (25 mL) at 23°C.
The reaction mixture was stirred at this temperature for 12 h and then diluted with water (50 mL)
and ethyl acetate (50 mL). The organic layer was separated and extracted with ethyl acetate (3
X 50 mL). The organic layers were combined, washed with brine (50 mL) and dried over
anhydrous sodium sulfate, filtered and concentrated under reduced pressure. The crude product
was purified by flash column chromatography (25→50% ethyl acetate in hexane) to afford 712
mg of the title compound (59%). Physical State: red oil. R : 0.57 (1:1 hexanes/EtOAc, UV light).
f
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
+
1
HRMS (ESI+): m/z calc. for C H FNOSi (M + H) : 242.1371, found 242.1368. H NMR (400
12
21
MHz, CHLOROFORMꢀd): δ = 6.71 (dd, J = 8.6, 9.3 Hz, 1H), 6.42 (dd, J = 2.8, 12.1 Hz, 1H),
.31 (ddd, J = 1.3, 2.8, 8.5 Hz, 1H), 3.61 ꢀ 3.30 (m, 2H), 1.04 ꢀ 0.94 (m, 9H), 0.15 (d, J = 1.0 Hz,
6
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
6H). C NMR (101 MHz, CHLOROFORMꢀd): δ =154.4 (d, J = 242.8 Hz, 1C), 141.1 (d, J = 9.1
Hz, 1C), 135.3 (d, J = 12.6 Hz, 1C), 122.7 (br d, J = 3.1 Hz, 1C), 110.7 (d, J = 3.1 Hz, 1C),
1
03.9 (br d, J = 22.2 Hz, 1C), 25.6, 18.2, ꢀ4.8 (br d, J = 1.7 Hz, 1C).
N-(4-((tert-Butyldimethylsilyl)oxy)-3-fluorophenyl)-6-(2-(3-fluorophenyl)pyrrolidin-1-
yl)imidazo[1,2-b]pyridazine-3-carboxamide ((±)ꢀ(13). N,NꢀDiisopropylethylamine (0.12 mL, 0.69
mmol, 2.5 equiv) was added to a solution of 6ꢀ(2ꢀ(3ꢀfluorophenyl)pyrrolidinꢀ1ꢀyl)imidazo[1,2ꢀ
b]pyridazineꢀ3ꢀcarboxylic acid ((±)ꢀ12, 90 mg, 0.28 mmol, 1.0 equiv) in N,Nꢀdimethylformamide
(3 mL). HATU (107 mg, 0.28 mmol, 1.0 equiv) was then added in one portion and the reaction
mixture was stirred at 23°C for 5 min. A solution of 4ꢀ((tertꢀbutyldimethylsilyl)oxy)ꢀ3ꢀfluoroaniline
(
17) 81 mg. 0.34 mmol, 1.2 equiv) in N,Nꢀdimethylformamide (1 mL) was added dropwise and
the reaction mixture was stirred at 23°C for 12 h. The reaction mixture was diluted with ethyl
acetate (50 mL), washed with water (25 mL) and brine (25 mL), dried over anhydrous sodium
sulfate, filtered and concentrated under reduced pressure. The crude product was purified by
flash column chromatography (1→5% methanol in dichloromethane) to afford 96 mg of the title
compound (62%). Physical State: white solid. R : 0.13 (1:99 MeOH/CH Cl , UV light). HRMS
f
2
2
+
1
(
ESI+): m/z calc. for C H F N O Si (M + H) : 550.2444, found 550.2458. H NMR (498 MHz,
29
34
2
5
2
CHLOROFORMꢀd): δ = 10.46 (br s, 1H), 8.29 (s, 1H), 7.72 (br d, J = 9.9 Hz, 1H), 7.51 (br s,
H), 7.34 ꢀ 7.29 (m, 1H), 7.05 ꢀ 6.95 (m, 3H), 6.93 ꢀ 6.86 (m, 2H), 6.60 ꢀ 6.53 (m, 1H), 5.11 (br d,
J = 7.9 Hz, 1H), 4.03 ꢀ 3.95 (m, 1H), 3.84 ꢀ 3.76 (m, 1H), 2.63 ꢀ 2.54 (m, 1H), 2.23 ꢀ 2.10 (m,
1
13
3H), 1.04 (s, 9H), 0.23 (s, 6H). C NMR (125 MHz, CHLOROFORMꢀd): δ = 163.2 (d, J = 247.5
Hz, 1C), 157.0, 153.8 (br d, J = 244.1 Hz, 1C), 152.1, 145.1 ꢀ 144.9 (m, 1C), 144.8 ꢀ 144.6 (m,
ACS Paragon Plus Environment

Page 23 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
C), 139.8 (d, J = 12.6 Hz, 1C), 138.0, 132.0 (br s, 1C), 130.7 (br d, J = 8.3 Hz, 1C), 127.2,
22.1, 121.1 (br d, J = 2.6 Hz, 1C), 116.1 (br s, 1C), 115.9 ꢀ 115.7 (m, 1C), 114.6 (br d, J = 21.4
1
Hz, 1C), 112.6 (br d, J = 22.2 Hz, 1C), 110.9 (br s, 1C), 109.4 (br s, 1C), 62.1, 48.7, 35.9, 25.6,
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
22.8, 18.3, ꢀ4.7.
N-(3-Fluoro-4-hydroxyphenyl)-6-(2-(3-fluorophenyl)pyrrolidin-1-yl)imidazo[1,2-b]pyridazine-3-
carboxamide ((±)ꢀ(14). Tetrabutylammonium fluoride (1.0 M in tetrahydrofuran, 0.23 mL, 0.23
mmol, 1.5 equiv) was added dropwise to a solution of Nꢀ(4ꢀ((tertꢀbutyldimethylsilyl)oxy)ꢀ3ꢀ
fluorophenyl)ꢀ6ꢀ(2ꢀ(3ꢀfluorophenyl)pyrrolidinꢀ1ꢀyl)imidazo[1,2ꢀb]pyridazineꢀ3ꢀcarboxamide ((±)ꢀ7,
85 mg, 0.15 mmol, 1.0 equiv) in tetrahydrofuran (5 mL) at 23°C. After 2 h, the reaction mixture
was quenched with water (25 mL) and diluted with ethyl acetate (50 mL). The organic layer was
separated and the aqueous layer was washed with ethyl acetate (3 X 50 mL). The organic
layers were combined, washed with brine (50 mL) and dried over anhydrous sodium sulfate,
filtered and concentrated under reduced pressure. The crude product was purified by flash
column chromatography (1→10% methanol in dichloromethane) to afford 65 mg of the title
compound (91%). Physical State: tan solid. R : 0.45 (1:9 MeOH/CH Cl , UV light). HRMS
f
2
2
+
1
(ESI+): m/z calc. for C H F N O (M + H) : 436.1580, found 436.1596. H NMR (400 MHz,
23
20
2
5
2
DMSOꢀd ): δ = 7.98 (s, 1H), 7.77 (d, J = 9.9 Hz, 1H), 7.42 (br dd, J = 1.9, 12.9 Hz, 1H), 7.27 ꢀ
6
7.16 (m, 1H), 7.01 (d, J = 7.8 Hz, 1H), 6.96 (td, J = 1.9, 10.0 Hz, 1H), 6.92 ꢀ 6.83 (m, 3H), 6.80
(
d, J = 10.0 Hz, 1H), 5.19 (dd, J = 1.5, 8.0 Hz, 1H), 4.01 ꢀ 3.90 (m, 1H), 3.76 ꢀ 3.65 (m, 1H), 2.07
13
ꢀ
1.88 (m, 3H), 1.59 ꢀ 1.48 (m, 1H). C NMR (101 MHz, DMSOꢀd ): δ = 163.0 (d, J = 245.0 Hz,
6
1C), 156.8, 152.6, 150.9 (d, J = 240.0 Hz, 1C), 145.9 (d, J = 6.2 Hz, 1C), 141.7 (d, J = 12.5 Hz,
1C), 136.6, 130.6 (d, J = 8.3 Hz, 1C), 129.9 (d, J = 9.1 Hz, 1C), 126.6, 124.7, 122.1, 121.7 (d, J
= 2.7 Hz, 1C), 117.9 (d, J = 3.3 Hz, 1C), 117.0 (br s, 1C), 113.9 (d, J = 21.4 Hz, 1C), 112.8,
112.5 (br d, J = 8.3 Hz, 1C), 109.7, 61.8, 48.6, 35.6, 22.6.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 55
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
(
R)-N-((R)-4-Chloro-1-(3-fluorophenyl)butyl)-2-methylpropane-2-sulfinamide ((R ,R)ꢀ(9). To a
S
solution of (R )ꢀ8 (900 mg, 4.29 mmol, 1.0 equiv) in CH Cl (22 mL) at ꢀ78 °C under nitrogen
S
2
2
was added 3ꢀfluorophenylmagnesium bromide (5.15 mL of 1.0 M THF solution, 5.15 mmol, 1.2
equiv). The reaction mixture was stirred at ꢀ78 °C for 4 h, brought to room temperature and
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
quenched with saturated NH Cl (25 mL). The organic layer was washed with saturated NaHCO3
4
(
25 mL), brine (25 mL), dried over anhydrous sodium sulfate, filtered and concentrated under
reduced pressure. The crude product was purified by flash column chromatography (20→100%
ethyl acetate in hexane) to afford 942 mg of the title compound (72%). For Xꢀray crystallography
analysis, the title compound was recrystallized from EtOAc/hexanes yielding white needles.
Physical State: White solid. R : 0.17 (1:4 EtOAc/hexanes, UV light/KMnO ). HRMS (ESI+): m/z
f
4
+
1
calc. for C H ClFNOS (M + H) : 306.1089, found 306.1084. H NMR (498 MHz,
14
22
CHLOROFORMꢀd) δ = 7.32 (dt, J = 5.9, 7.9 Hz, 1H), 7.11 (d, J = 7.7 Hz, 1H), 7.04 (td, J = 2.1,
.7 Hz, 1H), 7.01 ꢀ 6.97 (m, 1H), 4.40 ꢀ 4.34 (m, 1H), 3.49 (t, J = 6.5 Hz, 2H), 3.45 (d, J = 3.9
Hz, 1H), 2.19 ꢀ 2.09 (m, 1H), 1.95 ꢀ 1.85 (m, 1H), 1.82 ꢀ 1.71 (m, 1H), 1.67 ꢀ 1.57 (m, 1H), 1.23
9
13
(s, 9H). C NMR (125 MHz, CHLOROFORMꢀd) δ = 163.0 (d, J = 247.0 Hz, 1C), 144.6 (d, J =
6.7 Hz, 1C), 130.4 (br d, J = 8.3 Hz, 1C), 122.9 (br d, J = 2.8 Hz, 1C), 115.0 (d, J = 21.2 Hz,
1
C), 113.9 (d, J = 21.9 Hz, 1C), 58.1, 55.9, 44.5, 33.9, 28.7, 22.6.
(
R)-1-((R)-tert-Butylsulfinyl)-2-(3-fluorophenyl)pyrrolidine ((R ,R)ꢀ(10). To a solution of (R ,R)ꢀ9
S S
(
3.25 g, 11.9 mmol, 1.0 equiv) in THF (50 mL) at 23°C was added LiHMDS (17.9 mL of 1.0 M
THF solution, 17.9 mmol, 1.5 equiv). The reaction mixture was stirred at 23°C for 1 h and
quenched with NH Cl (25 mL) and extracted ethyl acetate (3 X 50 mL). The organic layers were
4
combined, washed with brine (50 mL) and dried over anhydrous sodium sulfate, filtered and
concentrated under reduced pressure. The crude product was purified by flash column
chromatography (20→40% ethyl acetate in hexane) to afford 2.29 g of the title compound
ACS Paragon Plus Environment

Page 25 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
71%). Physical State: white solid. R : 0.38 (2:3 EtOAc/hexanes, UV light). HRMS (ESI+): m/z
f
+
1
calc. for C H FNOS (M + H) : 270.1328, found 270.1322. H NMR (498 MHz,
14
22
CHLOROFORMꢀd) δ = 7.27 (dt, J = 6.0, 7.9 Hz, 1H), 7.03 (td, J = 0.7, 7.7 Hz, 1H), 6.96 (td, J =
2.1, 10.0 Hz, 1H), 6.93 ꢀ 6.87 (m, 1H), 5.07 (dd, J = 2.8, 8.2 Hz, 1H), 3.68 ꢀ 3.61 (m, 1H), 3.60 ꢀ
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
.53 (m, 1H), 2.21 ꢀ 2.12 (m, 1H), 1.94 ꢀ 1.86 (m, 1H), 1.85 ꢀ 1.78 (m, 1H), 1.77 ꢀ 1.71 (m, 1H),
13
1.06 (s, 9H). C NMR (125 MHz, CHLOROFORMꢀd) δ = 162.9 (d, J = 245.9 Hz, 1C), 147.6 (d,
J = 6.5 Hz, 1C), 129.9 (d, J = 8.3 Hz, 1C), 122.1 (br d, J = 2.8 Hz, 1C), 113.5 (br d, J 5 .9 Hz,
C), 113.3 (br d, J = 6.7 Hz, 1C), 57.5, 56.8, 55.2, 36.5, 24.2, 23.0.
1
(
R)-2-(3-fluorophenyl)pyrrolidine ((R)ꢀ(11). To a solution of (R)ꢀ10 (660 mg, 2.45 mmol, 1.0
equiv) in MeOH (10 mL) was added HCl (2.0 mL, 4.0 M solution in dioxane). The reaction
mixture was stirred 45 min at 23°C and then concentrated to dryness. The crude residue was
dissolved in water (10 mL) and neutralized to pH ~ 12 using a 6 N NaOH solution. The aqueous
phase was extracted with ethyl acetate (5 X 10 mL). The organic layers were combined, washed
with brine (50 mL) and dried over anhydrous sodium sulfate, filtered and concentrated under
reduced pressure to afford 400 mg of the crude title compound (99%, crude) which was used
directly in the following step without further purification.
In Vitro Biological Evaluations. The potencies of (R)ꢀ3 and (S)ꢀ3 and the K values and
i
33
selectivity profile (369 kinases) for (R)ꢀ3 were obtained through [γꢀ P]ATPꢀbased enzymatic
assays (Reaction Biology Corporation, Malvern, PA). Complete description of the assays can be
found in the SI Appendix (Supplementary Information Section 6).
Docking analyses. Molecular docking simulations of (R)ꢀ3 were performed using the Xꢀray coꢀ
crystal structure of the TrkBꢀcpdn5 complex (PDB code: 4AT3), the TrkAꢀN4ꢀ(4ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 55
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
morpholinophenyl)ꢀN6ꢀ(pyridinꢀ3ꢀylmethyl)pyridoꢀ[3,2ꢀd]pyrimidineꢀ4,6ꢀdiamine complex (PDB
code: 4PMT) using the FITTED 3.5 program (FORECASTER platform). Docking structures and
figures were prepared using Pymol.
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
P-gp/BCRP Assays and Lineweaver-Burk plot analysis. Cell Culture. Native MDCKII cells or
stably transfected with human Pꢀgp or wildꢀtype BCRP (482A) were seeded at a density of
2
1
56.250 cells/cm using Dulbecco’s modified Eagle’s medium (DMEM) with 10% FCS, 4.0 mM
LꢀGlutamine, 100 U/ml Penicillin, 100 ꢁg/ml Streptomycin and 100 ꢁg/ml Kanamycin (Biochrom,
Munich, Germany) and split after reaching cell confluency to 70ꢀ80%. Cell passages from 7ꢀ25
have been used throughout the experiments. Calcein-AM Uptake-Assay. The assay for
61
quantifying Pꢀgp efflux activity was performed as described previously. Briefly, human Pꢀgp
overexpressing MDCKII cells were washed two times with KrebsꢀRinger Buffer (142 mM NaCl, 3
mM KCl, 1.5 mM K HPO *3 H O, 10 mM HEPES, 4 mM DꢀGlucose, 1.2 mM MgCl , 1.4 mM
2
4
2
2
CaCl , pH 7.4) and preincubated with 100 ꢁl of doubleꢀconcentrated modulator for 15 min that
2
was followed by addition of 100 ꢁl 4°C cold 2 ꢁM CalceinꢀAM (SigmaꢀAldrich, Taufkirchen,
Germany) solution for 30 min. Plates were incubated at 37°C at 200 rpm. Control cells were
exposed to CalceinꢀAM in absence of a transport modulator. Subsequently, MDCKII cells were
washed with iceꢀcold KrebsꢀRinger Buffer to stop transporter activity and intracellular
fluorescence was released after incubation with 0.1% Triton X for 20 min at 37°C. The amount
of intracellular fluorescence was recorded using a fluorescence plate reader (Tecan Infinite
F200 Pro, λ(excitation) = 485 nm and λ(emission) = 520 nm). Background fluorescene was subtraced
from each signal and intracellularly accumulated fluorescence was related to control cells. EC50
values were calculated via nonꢀlinear regression using the 4ꢀparameter logistic equation with
®
variable Hillꢀslope (GraphPad Prism version 6.01) to generate doseꢀresponse curves.
Lineweaver-Burk Plot. First, MichaelisꢀMenten Kinetics were calculated from efflux experiments
measuring extracellular fluorescence every 90 sec on a Tecan Infinite F200 Pro. The kinetics
ACS Paragon Plus Environment

Page 27 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
were determined by assigning the linear gradient of the efflux curve (RFU (relative fluorescent
unit verse time) as velocity (v=RFU/time.) against the respective CalceinꢀAM concentration
(
[S]=1, 2 and 3 ꢁM) according to the following equation:
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
ꢀꢁꢂꢃ
∗ [ꢄ]
=
ꢀ
ꢅꢁ + [ꢄ]
where v represents the efflux velocity, vmax the maximum reaction velocity, [S] the Calceinꢀ
substrate concentration and K the MichaelisꢀMenten constant. A LineweaverꢀBurkꢀPlot was
m
obtained by plotting 1/v as the ordinate against 1/[S] as the abscissa to estimate the type of
inhibitory mode. BODIPY-Prazosin Uptake-Assay. BCRP efflux activity was analysed in native
TM
and human Bcrp MDCKII cells employing a flow cytometer (FACSCalibur flow cytometer,
Becton–Dickinson). In general, the net uptake of the BCRPꢀspecific fluorescent substrate
BODIPYꢀPrazosin into MDCKII cells in presence or absence of BCRP modulators or nonꢀ
62
modulators was determined. MDCKII cells were grown as monolayer, trypsinised and the
7
density adjusted to 2.5 × 10 cells/ml in DMEM/Ham’s F12 1:1. Hence, cells were washed twice
with 37°C preꢀwarmed medium and increasing concentrations of test compound added to a
volume of 600ꢁl cell suspension. Cells were incubated for 15 min at 37°C and 300 ꢁl BODIPYꢀ
Prazosin in DMEM/Ham’s F12 1:1 (2 ꢁM) was added to a final concentration of 0.5 ꢁM and
®
incubated for 30 min at 37°C, 700 rpm (Thermoshaker Incubator, Thriller , Peqlab).
Subsequently, the incubation was stopped with iceꢀcold medium and suspensions were
centrifuged at 200×g for 5 min and washed with 4°C PBS (2% FCS). Finally, intracellular
TM
BODIPYꢀPrazosin fluorescence was measured using a FACSCalibur flow cytometer equipped
with a 488 nm argon laser and a 530/ 30 bandꢀpass filter. By gating on forward and sideward
scatter, MDCKII cells were selected and dead cells excluded likewise using propidium iodide
staining (1.0 ꢁg/ml). Twenty thousand cells were sorted in one run. Data were analysed with
TM
CellQuest Pro (Franklin Lakes, NJ, USA). The increase in intracellular fluorescence (MF,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 55
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
median fluorescence) caused by a test compound was referred to control fluorescence levels
(
100%) and given as percentage of control.
Radiochemical Syntheses. The complete descriptions of the radiochemical syntheses for
preclinical studies (three sites: manual synthesis at the Edmonton PET Center (University of
Alberta), automated synthesis at the McConnell Brain Imaging Center (McGill University) and
automated synthesis at the University of Michigan PET Center (University of Michigan)) and
firstꢀinꢀhuman study (one site: LudwigꢀMaximiliansꢀUniversity of Munich) are provided in SI
Appendix (Supplementary Information Section 6). The manual and automated
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
11
radiosyntheses of [ C]ꢀ(±)ꢀ3 and [ C]ꢀ(R)ꢀ3 proceeded from the desmethyl phenolic precursors
11
and the manual and automated radiosyntheses of [ C]ꢀ(±)ꢀ4 proceeded from the corresponding
18
18
primary amide as described in SI Appendix. The manual radiosyntheses of [ F]ꢀ(±)ꢀ1, [ F]ꢀ(±)ꢀ2
11
36,38
11
and [ C]5 were performed as previously reported
. Briefly, for human study, [ C]CO was
2
14
11
produced via the N(p,α) C reaction by proton irradiation of a N /1% O target of a GE
2
2
11
PETtrace cyclotron and converted into [ C]CH I by reduction with H /Ni and iodination with I
2
3
2
11
using a GE FX C Pro automated module. [ C]CH I was bubbled into a solution of the precursor
3
(R)ꢀ14 (1.7 mg in 0.3 mL DMF) and K CO (10 ꢂL of a saturated aq. solution) at ꢀ20°C. The
2
3
crude mixture was heated to 80°C for 3 min, diluted with HPLC eluent (1.5 mL,) and purified via
semipreparative HPLC (Phenomenex LUNA C18 column, 100 Å, 250 × 10 mm, 10 ꢂm, 75%
11
MeCN/25% H O (v/v) isocratic; flow rate: 3 mL/min). The pure [ C]ꢀ(R)ꢀ3 was diluted with 10
2
mL H O, trapped on a preconditioned C18 cartridge (Waters SepPak C18 light), washed with
2
water (3 mL) and then eluted with EtOH (1 mL) followed by sterile filtration and formulation with
11
PBS. [ C]ꢀ(R)ꢀ3 was obtained in 13 ± 2 % decayꢀcorrected radiochemical yield at endꢀofꢀ
11
synthesis based upon [ C]CO , n = 3, radiochemical purity > 99%, S = 118 GBq/ꢀmol.
2
A
Rodent PET Imaging Studies. Animals. All animal studies were carried out according to the
guidelines of the Canadian Council on Animal Care (CCAC) and approved by the Cross Cancer
ACS Paragon Plus Environment

Page 29 of 55
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Institute AnimalꢀCare Committee (AC14214). In vivo studies were done using female Sprague
Dawley rats (body weight: 320ꢀ420 g, University of Alberta, Edmonton, AB, Canada, 6ꢀ12
months of age), normal female FVB mice (body weight: 24ꢀ30 g, CharlesꢀRiver Laboratories,
(−/−)
(−/−)
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
Quebec, Canada, 3ꢀ9 months of age) and female Mdr1a/b Bcrp1
mice (body weight: 24ꢀ30
g, Taconic, Hudson, NY, USA, 3ꢀ9 months of age). Imaging procedure and analysis. PET
18
18
11
11
11
11
imaging scans of [ F]ꢀ(±)ꢀ1, [ F]ꢀ(±)ꢀ2, [ C]ꢀ(±)ꢀ3, [ C]ꢀ(±)ꢀ4, [ C]ꢀ(R)ꢀ3 and [ C]5 were

performed on an INVEON PET scanner (Siemens Preclinical Solutions, Knoxville, TN, U.S.A.).
PET imaging experiments followed the same procedure for all 6 radiotracers: Prior to radiotracer
injection, rodentswere anesthetized through inhalation of isoflurane in 40% oxygen/ 60%
nitrogen (gas flow 1 L/min), and body temperature was kept constant at 37 °C. Mice or rats were
placed in a prone position into the center of the field of view. A transmission scan for attenuation
correction was only acquired for rat experiments, not for mice. Mice were injected with 10ꢀ50
11
MBq of Cꢀlabeled radiotracer in 100ꢀ150 ꢂL of isotonic saline solution (0.9% w/v of NaCl)
11
through a tail vein catheter. Rats were injected with 15ꢀ72 MBq of Cꢀlabeled or 11ꢀ21 MBq of
18
Fꢀlabeled radiotracer in 200ꢀ600 ꢂL of saline solution (0.9% w/v of NaCl). Data acquisition was
performed over 60 min in 3D list mode. The dynamic list mode data were sorted into sinograms
with 54 time frames (10 × 2, 8 × 5, 6 × 10, 6 × 20, 8 × 60, 10 × 120, 6 × 300 s). The frames
were reconstructed using maximum a posteriori (MAP) as reconstruction mode. No correction
for partial volume effects was applied. The image files were processed using the ROVER
v2.0.51 software (ABX GmbH, Radeberg, Germany). Masks defining 3D regions of interest
(ROI) were set, and the ROIs were defined by 50% thresholding. Mean standardized uptake
values [SUVmean = (activity/mL tissue)/(injected activity/body weight), mL/g] were calculated for
each ROI, and time−activity curves (TAC) were generated. All semiꢀquantified PET data are
presented as means ± SD.
ACS Paragon Plus Environment