Lessons from (S)-6-(1-(6-(1-methyl-1H-pyrazol-4-yl)-[1,2,4]triazolo[4,3- b ]pyridazin-3-yl)ethyl)quinoline (PF-04254644), an inhibitor of receptor tyrosine kinase c-met with high protein kinase selectivity but broad phosphodiesterase family inhibition lea

Article abstract of DOI:10.1021/jm400926x

The hepatocyte growth factor (HGF)/c-Met signaling axis is deregulated in many cancers and plays important roles in tumor invasive growth and metastasis. An exclusively selective c-Met inhibitor (S)-6-(1-(6-(1-methyl-1H-pyrazol-4-yl)- [1,2,4]triazolo[4,3-

Full text of DOI:10.1021/jm400926x

Subscriber access provided by ANDREWS UNIV
Article
Lessons from (S)-6-(1-(6-(1-Methyl-1H-pyrazol-4-yl)-
[
1,2,4]triazolo[4,3-b]pyridazin-3-yl)ethyl)quinoline (PF-04254644),
an Inhibitor of Receptor Tyrosine Kinase c-Met with High
Protein Kinase Selectivity, but Broad Phosphodiesterase
Family Inhibition Leading to Myocardial Degeneration in Rats
Jean Cui, Hong Shen, Michelle Tran-Dubé, Mitchell Nambu, Michele A McTigue, Neil Grodsky, Kevin
Ryan, Shinji Yamazaki, Shirley Aguirre, Max Parker, Qiuhua Li, Helen Y Zou, and James G. Christensen
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm400926x • Publication Date (Web): 14 Aug 2013
Downloaded from http://pubs.acs.org on August 20, 2013
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 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Lessons from (S)ꢀ6ꢀ(1ꢀ(6ꢀ(1ꢀMethylꢀ1Hꢀpyrazolꢀ4ꢀ
yl)ꢀ[1,2,4]triazolo[4,3ꢀb]pyridazinꢀ3ꢀ
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
yl)ethyl)quinoline (PFꢀ04254644), an Inhibitor of
Receptor Tyrosine Kinase cꢀMet with High Protein
Kinase Selectivity, but Broad Phosphodiesterase
Family Inhibition Leading to Myocardial
Degeneration in Rats
J. Jean Cui,* Hong Shen, Michelle Tran-Dubé, Mitchell Nambu, Michele McTigue, Neil
Grodsky, Kevin Ryan, Shinji Yamazaki, Shirley Aguirre, Max Parker, Qiuhua Li, Helen Zou, and
James Christensen
La Jolla Laboratories, Pfizer Worldwide Research and Development, 10770 Science Center
Drive, San Diego, CA 92121
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 57
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 HGF/cꢀMet signaling axis is deregulated in many cancers and plays important roles in
tumor invasive growth and metastasis. Exclusively selective cꢀMet inhibitor (S)ꢀ6ꢀ(1ꢀ(6ꢀ(1ꢀ
methylꢀ1Hꢀpyrazolꢀ4ꢀyl)ꢀ[1,2,4]triazolo[4,3ꢀb]pyridazinꢀ3ꢀyl)ethyl)quinoline (8) was discovered
from a highly selective HTS hit via structure based drug design and medicinal chemistry lead
optimization. 8 had many attractive properties meriting preclinical evaluation. Broad offꢀtarget
screens identified 8 as a panꢀPDE family inhibitor which was implicated in a sustained increase
in heart rate, increased cardiac output, and decreased contractility indices, as well as myocardial
degeneration in in vivo safety evaluations in rats. 8 was terminated as a preclinical candidate
because of a narrow therapeutic window in cardioꢀrelated safety. The learning from multiꢀ
parameter lead optimization and strategies to avoid the toxicity attrition at the late stage of drug
discovery are discussed.
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
INTRODUCTION
cꢀMet, also called hepatocyte growth factor receptor (HGFR), belongs to a unique subfamily of
receptor tyrosine kinases including macrophage stimulating 1 receptor RON. Hepatocyte growth
factor (HGF), also known as scatter factor (SF), is the highꢀaffinity natural ligand of cꢀMet. The
HGF/cꢀMet signaling pathway plays important roles in invasive growth during embryo
development and postnatal organ regeneration, and is only fully active in adults for wound
1
healing and tissue regeneration processes. However, the HGF/cꢀMet axis is frequently hijacked
1
,2
by cancer cells for tumorigenesis, invasive growth, and metastasis.
HGF and/or cꢀMet are
expressed at abnormally high levels in a large variety of solid tumors including liver, breast,
pancreas, lung, kidney, bladder, ovary, brain, prostate, gallbladder myeloma and many others,
3
that are frequently associated with a metastatic phenotype and poor prognosis. Various cꢀMet
ACS Paragon Plus Environment
2

Page 3 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mutations have been identified in many tumors, including hereditary and sporadic human
papillary renal carcinomas, ovarian cancer, childhood hepatocellular carcinomas, gastric cancer,
3
and lung cancer. The activation of HGF/cꢀMet signaling plays important roles in resistance to
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
EGFR and BRAF kinase inhibitors. MET amplification has been detected in up to 20% of
NSCLC patients with EGFR mutations and who acquired resistance to gefitinib or erlotinib
4
treatment. Growth factorꢀdriven resistance from tumor microenvironment represents another
5
potential mechanism to anticancer kinase inhibitors. HGF is present in melanoma patient’s
6
stromal cells and correlates with a poor response to the BRAF inhibitor vemurafenib treatment.
The upregulation of stromal HGF confers resistance to vemurafenib in BRAFꢀmutant melanoma
6
cells.
Because of the role of aberrant HGF/cꢀMet signaling in human oncogenesis,
invasion/metastasis, and resistance, the inhibition of this pathway has great potential in cancer
7
therapy. There are three approaches in modulating HGF/cꢀMet signaling currently in human
clinical studies for oncology applications including the antiꢀHGF antibody (e. g. AMG 102,
8
9
rilotumumab ), cꢀMet monoclonal antibody (e. g. MetMab ), and small molecule cꢀMet
10
11
12
inhibitors (e. g. crizotinib , cabozantinib , and tivantinib ). Preliminary clinical benefits from
1
3
the inhibition of HGF/cꢀMet signaling have been reported. The cꢀMet monoclonal antibody
MetMab in combination with erlotinib for nonꢀsmall cell lung cancer significantly improved
progressionꢀfree survival and overall survival, resulting in a near 3ꢀfold reduction in the risk of
1
4
death. One patient with a de novo highly METꢀamplified NSCLC achieved confirmed partial
15
response with crizotinib.
While existing data have shown the potential of inhibiting the
HGF/cꢀMet pathway, challenges remain in identifying specific patient populations as well as
developing the right combination strategies (targeted agents, classic chemotherapy, or
13
radiotherapy).
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Potent and specific kinase inhibitors are highly desired to validate the pharmacology of the
proposed target and allow for effective combination with other agents for the maximum efficacy
of cancer treatment. A highly specific cꢀMet HTS hit 1 (Chart 1) was discovered at legacy
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
16
SUGEN/Pharmacia. 1 was a highly selective, and ATP competitive cꢀMet inhibitor with high
biochemical ligand efficiency (LE = ꢀRTLogK /the number of heavy atoms = 0.42) and lipophilic
i
efficiency (lipE = ꢀLogK −cLogD = 4.80). Optimization of 1 by Koenig et al. afforded analogs
i
with improved enzymatic and cellular potencies, as illustrated with compound 2 that has an
enzymatic K of 1.3 nM and an IC of 7 nM for the inhibition of cꢀMet autoꢀphosphorylation in
i
50
17
A549 cell line. The cocrystal structure of 2 with unphosphorylated cꢀMet kinase domain (PDB
ID 3zze) revealed an unexpected binding mode, having the phenol residue as a hinge binder and
oxindole hydrazide interacting with the Aꢀloop to stabilize the cꢀMet protein in an autoinhibitory
16,18
inactive conformation.
Vojkovsky et al. designed the tetracyclic aromatic scaffold,
1
9
exemplified by 3, which proved to be an effective mimic of original oxindole hydrazide series.
The electron deficient tetracyclic ring retained the strong πꢀπ stacking interaction with electron
rich Tyrꢀ1230, and could form hydrogen bonds with NꢀH of Aspꢀ1226 and C=O of Argꢀ1208.
Zhang et al. further chopped the nonꢀdruggable tetracyclic scaffold to the bicyclic
triazolotriazine scaffold, exemplified by 4, that revealed an even better ligand efficiency against
cꢀMet than the tetracyclic scaffold despite the loss of the hydrogen bond with the carbonyl group
2
0
of Argꢀ1208. The electron deficiency of the bicyclic aromatic ring governs the strength of the
interaction with Tyroꢀ1230 at Aꢀloop and is a determining factor for the potency and selectivity.
2
1
Therefore, the triazolopyrazine scaffold, exemplified by 5 and the triazolopyridazine scaffold,
exemplified by 6, are less potent in general.
Chart 1. The discovery of highly selective cꢀMet inhibitors.
ACS Paragon Plus Environment
4

Page 5 of 57
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
The initial chemistry efforts on discovering the unique cꢀMet binding mode and the critical
chemical scaffolds for exclusively selective and highly potent cꢀMet inhibitors (Chart 1) play an
important role for the discovery and development of a class of exquisitely selective cꢀMet
inhibitors into human clinical trials. The phenol hinge binder in the original chemistry series was
successfully replaced with the conventional kinase hinge binder quinoline leading to 7 (Chart
1
6
16
2
). However, the triazolotriazine scaffold suffered with high nonꢀP450 hepatocyte clearance.
Optimization of the less potent triazolopyrazine series produced the clinical compound 9 (PFꢀ
1
6
0
4217903). The cocrystal structure of 9 with unphosphorylated cꢀMet kinase domain revealed
1
6
the similar binding environment as 2, and 9 was exquisitely selective over 208 kinases. In
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 57
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
22
23
addition, the highly potent and selective cꢀMet inhibitors 10 (SGXꢀ523) , 11 (JNJꢀ38877605) ,
2
4
25
1
2 (INCBꢀ28060) , and 13 (AMGꢀ208) were reported and have entered human clinical trials
for potential oncology indications (Chart 2).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Chart 2. Exquisitely selective cꢀMet inhibitors.
Herein, we report the medicinal chemistry lead optimization of the less potent
triazolopyridazine series leading to the discovery of (S)ꢀ6ꢀ(1ꢀ(6ꢀ(1ꢀmethylꢀ1Hꢀpyrazolꢀ4ꢀyl)ꢀ
[1,2,4]triazolo[4,3ꢀb]pyridazinꢀ3ꢀyl)ethyl)quinoline (8). The effort was conducted in parallel
with the optimization of the triazolopyrazine series leading to the discovery of the clinical
1
6
candidate 9. 8 and 9 were potent and exquisitely selective ATP competitive cꢀMet inhibitor,
and demonstrated low nM potency against cꢀMet in both in vitro cell assays and in vivo target
modulation studies. Both compounds showed effective tumor growth inhibition with good oral
PK properties. However, 8 was also a potent phosphodiesterase family inhibitor, which proved
to be responsible for a sustained increase in heart rate, increased cardiac output, and decreased
ACS Paragon Plus Environment
6

Page 7 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
contractility indices, as well as myocardial degeneration in rats. 8 was terminated as a
preclinical candidate because of a narrow therapeutic window in cardioꢀrelated safety. 9, which
has a higher LipE value for cꢀMet than 8, has a clean kinase selectivity profile, and also has a
cleaner profile in a broad offꢀtarget screen. This favorable profile supported a large therapeutic
window in in vivo animal tolerance studies. 9 was selected as preclinical candidate and entered
human clinical evaluations for cancer treatment.
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
Toxicity and serious adverse events in late stage drug development are the major causes of
drug failure. The adverse toxicologic effects are further classified as onꢀtarget, and offꢀtarget
effects. The onꢀtarget toxicity reflects the exaggerated and adverse pharmacologic effects of the
interested target which may lead to the termination of a drug development based on the target.
The offꢀtarget adverse effects are a result of modulation of biologically unrelated targets which
should be avoided at the early drug discovery stage. Although kinase selectivity has been a
major challenge in kinase drug development because of the conservation of the ATP binding
pocket among kinases, the offꢀtargets outside the protein kinase family deserve equal attentions
at the early drug discovery stage. Both 8 and 9 are highly selective cꢀMet RTK inhibitors.
However, the offꢀtarget effects from the inhibition of PDE family leading to a narrow therapeutic
window in cardioꢀrelated safety result in the terminate of 8 as a preclinical candidate. The
learning from multiꢀparameter lead optimization and strategies to avoid the late stage toxicity
attrition will be discussed here.
RESULTS AND DISCUSSION
The triazolotriazine chemical series, represented with 7 (Chart 2), was successfully designed as
1
6
a class of potent and kinaseꢀselective cꢀMet inhibitors evolving from the HTS hit 1.
The
quinoline group functions as a hinge binder and the triazolotriazine group plays an important role
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
in both cꢀMet cellular potency and kinase selectivity via the unique interaction with the Aꢀloop
residue Tyrꢀ1230 (Figure 1).
The unique autoinhibitory position of the Aꢀloop in
2
6
unphosphorylated cꢀMet has been documented in cꢀMet crystal structures, including apoꢀcꢀMet.
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
In particular, many cꢀMet inhibitors form interactions with Tyrꢀ1230, and the interactions vary in
1
6
strength and contribute to different affinities and selectivity.
Although it had an attractive
potency and selectivity profile, the triazolotriazine series suffered high clearance in human
16
hepatocyte, and was deꢀprioritized. Alternatively, the less potent triazolopyrazine (represented
with 5) and triazolopyridazine (represented with 6) series were explored. The detailed medicinal
chemistry lead optimization of the triazolopyrazine series leading to the clinic candidate 9 was
1
6
reported previously , and some of the data are cited here as direct comparisons with the
triazolopyridazine series for illustration of the learning. We previously reported the selection of
quinoline as a suitable hinge binder with good potency and metabolic stability in the
1
6
triazolopyrazine series in the replacement of phenol hinge binder in 1. Here, the quinoline
group was directly used as the desired hinge binder in the triazolopyridazine series for the further
lead optimization at the 6ꢀposition. With the replacement of Nꢀ7 in the triazolotriazine core with
a carbon atom, the triazolopyridazine bicyclic aromatic ring is more electron rich, resulting in a
less effective interaction with Tyrꢀ1230. As a result, weaker potencies were generally observed
in the triazolopyridazine series (Aꢀseries) as demonstrated with 14A. 14A was about 10ꢀfold less
potent in both enzymatic and cellular assays relative to 7. The structure activity relationship at
the 6ꢀposition of the [1,2,4]triazolo[4,3ꢀb]pyridazinꢀ3ꢀylmethylꢀquinoline series was investigated
(Table 1) for the potential improvement of potency and ADME properties. Overall, 6ꢀsubstituted
[
1,2,4]triazolo[4,3ꢀb]pyridazinꢀ3ꢀylmethylꢀquinolines were metabolically stable and had good
permeability. However, a flat SAR in potency was observed for 6ꢀaryl compounds 14A−16A
ACS Paragon Plus Environment
8

Page 9 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and 8A. In general, the Aꢀseries was about 10ꢀfold less potent than the Bꢀseries (the Bꢀseries
1
6
was reported previously and cited here for the comparison), as demonstrated with 8A and 8B.
The electron donating amino and dimethylamino groups in 17A and 18A provided even weaker
potencies. Consistently, 17A was much less potent than the triazolopyrazine analogue 17B.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Cocrystal structure of 7 with the unphosphorylated cꢀMet kinase domain (PDB 3zbx).
Table 1. Structure activity relationship of 6ꢀsubstituted [1,2,4]triazolo[4,3ꢀb]pyridazinꢀ3ꢀ
ylmethylꢀquinolines.
N
N
R
N
R
N
N
N
N
N
N
N
N
A
B
a
Compd R
cLogD
cꢀMet Ki cꢀMet
LipE
HLM
Solub.
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 57
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
b
c
d
e
(
ꢀM)
Cell IC50 (IC50
)
CLint, app
(ꢀM)
b
(ꢀM)
22
7
1
2.03
3.44
0.006
0.044
0.005
0.048
6.27
2.88
63.1
13.4
ND
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
4A
11.5
1
5A
6A
3.65
2.90
0.013
0.024
0.019
0.044
4.07
4.46
22.7
ND
19.1
ND
1
8
A
B
1.73
0.68
0.074
0.006
0.044
0.001
5.63
8.32
ND
ND
8
5.95
37.8
9
0.05
0.54
0.005
0.379
0.004
2.5
8.35
5.06
<7.6
<7.6
22.2
453
1
7A
7B
1
ꢀ0.95
1.93
0.015
0.25
0.034
0.349
8.42
4.53
<7.6
35
ND
134
1
8A
a
Calculated logarithm of the octanol/water distribution coefficient at pH 7.4 using ACD
pchbat version 9.3. Inhibition constants (K ) and cell IC were determined as described under
Experimental Section. The coefficients of variance were typically less than 20% (n = 2). A549
human lung carcinoma cell line was used for the evaluation of the inhibition of
LipE (IC ) = pIC – cLogD.
intrinsic clearance (ꢀL/min/mg). Kinetic solubility at pH 6.5. ND, not determined.
b
i
50
c
d
autophosphorylation of cꢀMet.
Human liver microsomal
5
0
50
e
ACS Paragon Plus Environment
10

Page 11 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The triazolopyridazine series has a cLogD value more than one log unit higher than the
triazolopyrazine series, and coupled with the weaker inhibition, this leads to lipE values more
than three log units lower than the triazolopyrazine series. Considering LipE, there is little
advantage to continue working on the triazolopyridazine series. However, comprehensive
analyses of the ADME properties of both series indicated that the triazolopyridazine series has
much higher intrinsic human liver microsome metabolic stability relative to the triazolopyrazine
series (Figure 2, log scale of human liver microsomal intrinsic clearance HLM CLint.,app vs
cLogD). Overall, triazolopyridazines (blue square) were more metabolically stable than
triazolopyrazines (red square) at higher cLogD value. Therefore, the triazolopyridazines could
potentially provide a better chemistry space for multiꢀparameter optimizations on metabolic
stability, permeability, and potency.
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
2
1
0
200
0
00
100
8
0
80
6
0
60
4
0
40
2
0
20
1
0
10
8
8
6
6
44
-
-
2
2
-1
-1
0
1
1
2
2
3
3
0
cLogD @ pH = 7.4
cLogD (pH 7.4)
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 2. Metabolic clearance in human liver microsomes vs cLogD (red color for the
triazolopyrazine series, and blue color for the triazolopyridazine series).
Because of the attractive ADME properties, the triazolopyridazine series was explored further
to improve the cꢀMet potency. The cocrystal structure of 7 with the unphosphorylated cꢀMet
kinase domain revealed a small hydrophobic pocket for the potential improvement of the cꢀMet
potency (Figure 1). The hinge binding quinoline substituent was linked with the triazolotriazine
group via a methylene linker. The methylene group was adjacent to a hydrophobic pocket
surrounding with the side hydrophobic chains of Lysꢀ1110, Valꢀ1092, Leuꢀ1157 and Alaꢀ1226.
The modeling indicated that an (S)ꢀmethyl group could fit into the pocket well and potentially
boost the potency. A series of compounds with an alphaꢀmethyl group having the Sꢀ
configuration were synthesized, and screened for the potency against cꢀMet (Table 2).
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
Indeed, the (S)ꢀmethyl group improved potency significantly as demonstrated with 14A and
14C (10ꢀfold improvement for the enzymatic activity and 4ꢀfold for the cellular potency).
Although the introduction of a methyl group increased the lipophilicity (cLogD) by 0.35 unit, the
dramatic improvement on potency caused the cell LipE of 14C to increase by 1.25 unit relative
to 14A, which justified the introduction of a hydrophobic group, and also implied that the methyl
group had a strong van der Waals interaction with the protein as originally designed. A broad set
of R groups were investigated at the 6ꢀposition in Cꢀseries and Dꢀseries (Table 2) in order to
identify the potential candidates with comprehensively acceptable properties. Overall, 6ꢀaryl
substituents provided good potency against cꢀMet. The substituents at ortho, meta, and paraꢀ
positions of 6ꢀphenyl group were well tolerated as demonstrated with compounds 14C−26C.
The R group in both the Cꢀ and Dꢀseries extended into the solventꢀexposed area, and as
expected, the substituents of both polar and hydrophobic groups didn’t change the potency
ACS Paragon Plus Environment
12

Page 13 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
significantly, however, the LipE and metabolic stability varied significantly according to the
nature of the R groups. The N or Oꢀalkyl substituents on 6ꢀaryl group are highly cleared in
human liver microsomes as demonstrated with compounds 19C, 20C, 23C and 27C. 21C with a
methyl group also showed a high metabolic clearance. The polar substituents including cyano
group in 15C, 16C, and 22C, amide group in 24C, and sulfonyl group in 25C and 26C made the
compounds potent and metabolically stable. In general, the 6ꢀphenyl compounds in Cꢀseries had
low solubility. 8 with Nꢀmethyl pyrazolꢀ4ꢀyl group had a lower cLogD of 2.07, and comparable
potency in comparison with 6ꢀphenyl compounds, which led to a much improved LipE (6.15),
good metabolic stability and solubility. Consistently, the Dꢀseries was about 10 fold more potent
than the Cꢀseries as shown with 8 and 8D. Again, the (S)ꢀmethyl group improved the potency of
6ꢀNꢀsubstituted compounds as exemplified by 28C, which demonstrated an overall good profile,
especially in solubility (433 ꢀM). The contribution of the magic (S)ꢀmethyl group to the potency
was further observed with the direct comparison of 9 and 9D. Although the Dꢀseries was much
more potent, the Cꢀseries had better solubility as illustrated with 8 (60.4 ꢀM) and 8D (16.3 ꢀM).
Table 2. Structure activity relationship of 6ꢀsubstituted (S)ꢀ6ꢀ(1ꢀ[1,2,4]triazolo[4,3ꢀb]pyridazinꢀ
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
3ꢀyl)ethyl)quinolines and (S)ꢀ6ꢀ(1ꢀ1Hꢀ[1,2,3]triazolo[4,5ꢀb]pyrazinꢀ1ꢀyl)ethyl)quinolines.
cꢀMet
cꢀMet Ki
LipE
HLM
Solub.
a
Compd
R
cLogD
Cell IC50
b
c
d
e
(ꢀM)
(IC )
CLint, app
(ꢀM)
5
0
b
(
ꢀM)
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 57
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
4C
9C
3.79
3.71
0.004
0.004
0.012
0.008
4.13
4.39
13.4
281
8.73
10.9
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
2
2
0C
1C
4.2
0.003
0.005
0.009
0.011
3.85
3.57
304
1.77
19.0
4.39
81.8
*
1
1
2
2
5C
6C
2C
3C
4.0
0.004
0.008
0.006
0.004
0.004
0.012
0.005
0.006
4.40
4.68
5.12
4.88
17.9
12.1
23.1
304
32.5
15.5
213
3.24
3.18
3.34
6.08
24C
2.46
2.63
0.007
0.024
0.015
0.036
5.36
4.81
12.2
10.4
11.0
136
25C
26C
2.1
0.012
0.008
0.024
0.025
5.52
4.31
<7.6
232
44.9
16.2
27C
3.29
ACS Paragon Plus Environment
14

Page 15 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8
8
2.07
1.02
0.0103
0.0014
0.006
6.15
8.13
<7.60
<7.6
60.4
16.3
D
0.0007
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
9
2
D
0.40
2.42
0.0028
0.007
0.0005
0.023
8.90
5.22
<7.6
20.2
17.3
433
8C
a
Calculated logarithm of the octanol/water distribution coefficient at pH 7.4 using ACD
pchbat version 9.3. Inhibition constants (K ) and cell IC were determined as described under
Experimental Section. The coefficients of variance were typically less than 20% (n = 2). A549
human lung carcinoma cell line was used for the evaluation of the inhibition of
autophosphorylation of cꢀMet.
b
i
50
c
d
LipE (IC ) = pIC – cLogD.
Human liver microsomal
5
0
50
e
intrinsic clearance (ꢀL/min/mg). Kinetic solubility at pH 6.5.
To test whether the effect of the alphaꢀmethyl group is universal for the different hinge
binders, a subseries with azaindole hinge binder was investigated (Table 3). As a direct
comparison, 8F with (S)ꢀmethyl group was 10 fold more potent than 8E. 8F−29F with the
azaindole hinge binder had a similar potency as the analogues in Cꢀseries having the quinoline
hinge binder, however, 8F−29F provided better solubility profiles. The metabolite identification
study of 8F indicated the presence of the GSHꢀadducts at azaindole and methylene positions
(
Figure S1), and the Fꢀseries was deꢀprioritized.
Table 3. Structure activity relationship of 6ꢀsubstituted 3ꢀ(1ꢀ(1Hꢀpyrrolo[2,3ꢀb]pyridinꢀ3ꢀ
yl)ethyl)ꢀ[1,2,4]triazolo[4,3ꢀb]pyridazines
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 57
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
cꢀMet
cꢀMet Ki
LipE
HLM
Solub.
a
Compd R
cLogD
Cell IC50
b
c
d
e
(
ꢀM)
(IC )
CLint, app (ꢀM)
5
0
b
(
ꢀM)
8
E
F
1.46
1.81
0.0897
0.016
0.209
5.22
5.94
<7.6
<7.6
111
150
8
0.0176
16F
2.98
0.0149
0.023
4.66
28
ND
2
2
8F
2.16
1.81
0.0101
0.0167
0.019
5.56
5.78
25
18
475
426
9F
0.0254
a
Calculated logarithm of the octanol/water distribution coefficient at pH 7.4 using ACD
pchbat version 9.3. Inhibition constants (K ) and cell IC were determined as described under
Experimental Section. The coefficients of variance were typically less than 20% (n = 2). A549
human lung carcinoma cell line was used for the evaluation of the inhibition of
LipE (IC ) = pIC – cLogD.
intrinsic clearance (ꢀL/min/mg). Kinetic solubility at pH 6.5. ND, not determined.
b
i
50
c
d
autophosphorylation of cꢀMet.
Human liver microsomal
5
0
50
e
Collectively, a number of compounds achieved good potency and in vitro ADME profiles, and
were selected for in vivo rat PK studies (Table 4) and in vitro toxicity evaluations. The rat in
1
6
vivo PK data of 9 was cited as a reference for the comparison. 8, 14C and 15C demonstrated
moderate plasma clearance and volume distribution.
bioavailability, and 14C and 15C have good half life.
Table 4. in vivo Rat pharmacokinetic properties.
8 and 14C showed good oral
ACS Paragon Plus Environment
16

Page 17 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CLplasma
Vss
T1/2
(h)
3.7
1.9
3.6
4.8
cꢀMet Cell
a
b
c
Compd MW pKa cLogD eLogD
Foral %
d
(mL/min/kg) (L/kg)
eLipE
9
372
355
387
376
4.66 1.68
4.79 2.07
4.79 3.79
4.79 4.00
1.38
2.05
3.07
2.80
8.2
21
31
22
1.6
3.5
3.6
4.9
71
58
52
13
7.02
6.17
4.75
5.60
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
8
14C
15C
a
b
Calculated ionization constant of a molecule using ACD pchbat version 9.3.
logarithm of the octanol/water distribution coefficient at pH 7.4 using ACD pchbat version 9.3.
Experimentally measured biꢀlayer participating coefficient at pH 7.4. eLipE = −pIC (cell)−
Calculatedc
d
5
0
eLogD.
The tumor growth inhibition (TGI) and the relationship to inhibition of cꢀMet autoꢀ
phosphorylation in vivo were evaluated in parallel for 8 and 9 in cꢀMet amplified GTLꢀ16
xenograft tumor model. No weight loss was observed at all dose levels in the efficacy studies for
both 8 and 9. The marked efficacy results and the potential mechanisms for the tumor growth
1
6,27
inhibition of 9 were reported previously.
Similar to 9, 8 demonstrated doseꢀdependent
inhibition of cꢀMet phosphorylation and tumor growth in the cꢀMet amplified GTLꢀ16 xenograft
tumor model in mice (Figure 3). Tumor regression ( ꢀ34% TGI) was observed with 100 mg/kg
QD group, and tumor stasis (100% TGI) was achieved with 30 mg/kg QD group. Near complete
inhibition of cꢀMet activity for 24 hours is consistent with tumor regression and tumor stasis.
The potent inhibition of cꢀMet activity for only a portion of the dosing schedule is consistent
with the observed significant but submaximal antitumor efficacy at lower doses (10 & 3 mg/kg
QD, 84% to 75% TGI). These results suggest that nearꢀcomplete inhibition of cꢀMet
autophosphorylation (>90% inhibition) for the duration of the administration schedule is
necessary for maximal efficacy. The total cꢀMet level was not changed in the study as reported
27
previously in the same studies of 9.
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 57
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
40
20
1200
0
.5h
1h
8h
24h
Control
100mg/kg QD
1
Values = Mean + SEM, n = 8-12
1000
30mg/kg QD
100
10mg/kg QD
3mg/kg QD
800
600
400
8
6
0
0
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
Values = Mean + SEM
n = 8-12
*
75%Inh.
40
*
84%Inh.
100% Inh.
34%Reg.
2
0
0
200
0
*
*
16 days of Dosing
3
mg
10mg
30mg
100mg
8
12
16
20
24
Compound 8 Dose (mg/kg/day, Day-15)
Study Days (Tumor Implantation at Day 0)
Figure 3. Inhibition of cꢀMet phosphorylation and tumor growth by 8 in the GTLꢀ16 xenograft
tumor model in mice.
Based on the encouraging antiꢀcꢀMet and antitumor activities in vitro and in vivo, and the
desired predicted human PK properties, 8 and 9, representing two different series were chosen
for kinase selectivity and preclinical in vitro and in vivo toxicity studies. Similar to 9 which
2
8
demonstrated an exquisite cꢀMet kinase selectivity , 8 is a highly kinaseꢀselective cꢀMet
inhibitor (>1000 selectivity to 208 protein kinases). The cocrystal structure of 8 with
unphosphorylated cꢀMet kinase domain (Figure 4) resembles the original cocrystal structure of 7
with cꢀMet. As expected, the quinoline functions as a hinge binder hydrogen bound with the Nꢀ
H of Metꢀ1160. Nꢀ2 in [1,2,4]triazolo[4,3ꢀb]pyridazine forms a strong hydrogen bond with NꢀH
of Aspꢀ1222. [1,2,4]Triazolo[4,3ꢀb]pyridazine has the cꢀMet characteristic πꢀπ stacking
interaction with Tyrꢀ1230, and the 6ꢀpyrazolꢀ4ꢀyl group forms a coplanar with
[1,2,4]triazolo[4,3ꢀb]pyridazine ring to strengthen the πꢀπ stacking interaction. CꢀH at Cꢀ5
position of 6ꢀpyrazolꢀ4yl group has a close interaction with the conserved C=O of Argꢀ1208, and
Nꢀ2 interacts with a water molecule. As originally designed, the (S)ꢀmethyl group on methylene
bridge is surrounded by a group of hydrophobic residues. 8 and 9 have a near completely
ACS Paragon Plus Environment
18

Page 19 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
overlay of cocrystal structures with cꢀMet in both ligands and proteins (Figure 5), which supports
the exclusive kinase selectivity for both 8 and 9.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Cocrystal structure of 8 with cꢀMet unphosphorylated kinase domain (PDB 3zc5).
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 5. Overlay of cocrystal structures of 8 (green color) (PDB 3zc5) and 9 (pink color) (PDB
3zxz) with cꢀMet unphosphorylated kinase domain.
The interaction of the [1,2,4]triazolo[4,3ꢀb]pyridazine core with the Aꢀloop anchors the whole
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
molecule at the ATP binding site of cꢀMet, which tolerates different hinge binders and
substituents at 6ꢀposition. The cocrystal structure of 28F superimposes well with 8 (Figure 6) at
the [1,2,4]triazolo[4,3ꢀb]pyridazine position even though 28F has an azaindole hinge binder and
an isopropylamino group at 6ꢀposition. As expected, the azaindole forms hydrogen bonds with
the C=O of proꢀ1158 and the NꢀH of Metꢀ1160 at the hinge. The alphaꢀmethyl groups from 28F
and 8 are positioned closely in the small hydrophobic pocket. The isopropyl group in 28F is
closer to the glycine rich loop, and the NꢀH from the 6ꢀisopropylamino function forms a water
bridged hydrogen bond with the C=O of Argꢀ1208. Taken together, the cocrystal structures of cꢀ
Met reveal the specific binding characteristics of unphosphorylated cꢀMet protein, which allows
for the design of highly selective and druggable cꢀMet inhibitors.
ACS Paragon Plus Environment
20

Page 21 of 57
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
Figure 6. Superimposition of 28F (golden color) (PDB 3zcl) with 8 (green color) (PDB 3zc5) in
cꢀMet cocrystal structures.
Albeit the excellent selectivity to protein kinases, 8 and 9 were submitted to CEREP BioPrint
(Poitiers, France) for enzymatic binding assays (68 targets) to assess the potential offꢀtarget
pharmacology from inhibition of other enzyme classes (Table 5 for targets with >50% inhibition
at 10 ꢀM). Out of 68 targets, 8 had 4 hits with >50% inhibition at 10 ꢀM, and the IC s were
5
0
obtained for these targets. 8 potently inhibited GABA A BZD (IC = 0.180 ꢀM), and PDE3
5
0
(
IC = 0.170 ꢀM). 9 was a cleaner compound with only moderate inhibition against PDE3 (IC
5
0
50
= 1.30 ꢀM).
Table 5. Off target hits of 8 and 9 in CEREP BioPrint screen.
%
inhibition at 10 ꢀM
IC (ꢀM)
5
0
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 57
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
Target
GABA A BZD
Muscarinic 2
PDE3 (h)
Compd 8
Compd 9
Compd 8
0.180
7.00
Compd 9
97
59
98
63
16
33
71
44
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
0.170
3.70
1.50
PDE4 (h)
Due to the offꢀtarget activity identified in CEREP BioPrint, 8 and 9 were further evaluated in
an internal Pfizer PDE assay panel including PDE1A, PDE1B, PDE1C, PDE2, PDE3A, PDE3B,
PDE4A, PDE4B, PDE4C, PDE4D, PDE5, PDE6, PDE7A, PDE7B, PDE8A, PDE8B, PDE9,
PDE10, and PDE11 (Table 6 for PDEs hits having IC <10 ꢀM). 9 had weak enzymatic
5
0
activities against PDE3B (5.36 ꢀM), PDE5 (8.06 ꢀM) and PDE10 (2.7 ꢀM). However, 8 was
unexpectedly a panꢀPDE inhibitor with more potent inhibition against PDE3B (0.15 ꢀM) and
PDE10 (0.085 ꢀM). 14C and 15C from the same chemistry scaffold as 8 demonstrated a similar
panꢀPDE inhibition profile. The introduction of the alphaꢀmethyl group into [1,2,3]triazolo[4,5ꢀ
b]pyrazine scaffold (Dꢀseries) increased the potency against cꢀMet and also the interactions with
PDEs, as demonstrated with 8D and 9D (Table 6) in comparison with 9.
Table 6. PDE family inhibition.
PDE IC (ꢀM)
5
0
Compd
1A
>10
0.58
1C
>10
1.4
2
3A
3B
5.36
0.15
0.107
4C
>10
6.7
4D
>10
9.61
2.23
5
10
11
9
8
>10
4.56
>10
>10
3.53
2.87
8.06
4.89
1.75
2.7
>10
3.03
2.44
0.085
0.403
14C
0.452 0.963
2.91
ACS Paragon Plus Environment
22

Page 23 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
5C
0.462
>10
1.48
>10
>10
>10
>10
>10
1.8
1.25
0.589
0.088
0.699
0.42
1.97
2.08
3.26
2.06
2.5
0.287 0.432 0.963
8
D
>10
>10
>10
>10
>10
>10
9
D
>10
2.12
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
The cardiovascular system controls blood pressure and beating of cardiac myocytes to
accurately pump blood out of the heart to other parts of the body. These key physiological
events are produced by contraction and relaxation of vascular smooth muscle and cardiac
myocytes, that are regulated by intracellular cyclic nucleotide concentrations of second
29
messengers, cꢀAMP and cꢀGMP. The degradation of cyclic nucleotides is catalyzed by 3,5ꢀ
2
9
cyclic nucleotide phosphodiesterases (PDEs).
Multiple PDEs function as a particular
2
9
modulator of each cardiovascular function and regulate physiological homeostasis. Therefore,
inhibition of PDEs may imply a severe liability in cardiovascular safety. To further evaluate the
potential cardio toxicity and other liabilities, 8 and 9 were studied in rats for finding the
maximum tolerance levels. 8 in 2ꢀ, 6ꢀ or 7ꢀday single and repeat dose rat studies resulted in a
sustained increase in heart rate (HR), increased cardiac output (CO) and decreased contractility
indices, as well as myocardial degeneration at dose levels ≥40 mg/kg which was equivalent to or
3
0,31
below the predicted efficacious dose in human.
Further mechanism investigation revealed
that the broad inhibition of PDE family members especially PDE3B and the activation of cꢀAMP
31
may contribute to the observed cardio toxicity with 8. Consistently, the intracellular calcium
and oxidative stress signaling pathways were perturbed with a doseꢀdependent increase in
3
1
intracellular calcium following the activation of cꢀAMP by the treatment with 8. 8, a panꢀPDE
inhibitor that was proved in vivo with cardio toxicity, was terminated as a preꢀclinical candidate.
As expected, the cardio toxicity was not observed for 9 at 42ꢀ150 fold over the predicted free
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 57
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
human Ceff of 36 nM because of a much cleaner offꢀtarget profile of 9 with only a weak
inhibition against PDE3B (IC 5.36 ꢀM).
5
0
8
and 9 from two different chemistry series, however, having the same binding mode with cꢀ
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
Met, were evaluated in parallel for the selection of a potential preclinical candidate. The ovrall
profiles of 8 and 9 were summarized in Table S1. Both compounds demonstrated attractive
potency in vitro and in vivo, exclusive kinase selectivity, good in vitro human ADME properties
and in vivo rodent and dog pharmacokinetic profiles. Unexpectedly, 8 was shown to be a panꢀ
PDE inhibitor, which might be linked with the observed in vivo cardio toxicity of 8 in rat.
Although the protein expression levels of PDEs in in vivo GTLꢀ16 tumor model were not
available, a similar PK/PD relationship of 8 and 9 implied the inhibition of PDEs by 8 may not
contribute significantly to the antitumor efficacy in GTLꢀ16 cꢀMetꢀdriven tumor model. Based
on the overall profile, 9 was selected as a preclinical candidate for an early development
1
6
evaluation, and eventually went into human phase I clinical study.
CHEMISTRY
The syntheses of [1,2,4]triazolo[4,3ꢀb]pyridazine compounds are summarized in Scheme I. 2ꢀ
(Quinolinꢀ6ꢀyl)acetic acid (31) was synthesized from 2ꢀ(4ꢀaminophenyl)acetic acid (30) using
the Skraup quinoline synthesis conditions with a yield of 17.8%. After esterification, alkylation
with methyl iodide in the presence of the base LDA, and hydrolysis, 32 was obtained in a
combined yield of 65%. 34 was obtained in 97% yield after Bocꢀprotection was introduced on
33. Following methylation, removal of the Bocꢀgroup, and hydrolysis, 35 was obtained in a
combined yield of 88%. The [1,2,4]triazolo[4,3ꢀb]pyridazine ring was built up via the coupling
reaction of 36 with the activated carboxylic acid, and followed by heating under an acidic
ACS Paragon Plus Environment
24

Page 25 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
condition to afford 38. A variety of substituents were introduced at 6ꢀposition of 38 via the
conventional Suzuki coupling reaction or replacement of chloride with amino groups.
a
Scheme I . Syntheses of [1,2,4]triazolo[4,3ꢀb]pyridazine compounds.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: (a) glycerol/FeSO /conc.H SO , nitrobenzene, reflux, 5 h. (b)
4
2
4
o
SOCl , Methanol, reflux. (c) LDA/MeI, anhydrous THF, ꢀ78 C to ambient temperature,
overnight. (d) LiOH, MeOH/H O, 65 C, 4 h. (e) (Boc) O/DMAP, THF, overnight. (f)
HCl/Dioxane (4N), CH Cl . (g) EDC or HATU, DMF, overnight. (h) SOCl , DMF, 100 C, 30
min. (i) Acetic acid, reflux, 2 h. (j) Pd(dppf)Cl ꢁCH Cl , Cs CO , DME/H O, 80 C, overnight.
2
o
2
2
o
2
2
2
o
2
2
2
2
3
2
o
(
k) amine, nꢀButanol, microwave at 125 C, ~1 h.
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 57
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
The syntheses of (S)ꢀ6ꢀ(1ꢀ(1Hꢀ[1,2,3]triazolo[4,5ꢀb]pyrazinꢀ1ꢀyl)ethyl)quinoline compounds
are summarized in Scheme II. 47 was prepared with a fourꢀstep procedure. Quinolineꢀ6ꢀ
carboxylic acid 41 reacted with N,Oꢀdimethyl hydroxylamine in the presence of the activation
reagent CDI in DMF to afford 42 in 97% yield, which reacted with MeMgCl to provide the
ketone 43 in 97% yield, and followed by amination and hydrogenation to provide 45 in 97%
yield. The thermal replacement of 2ꢀbromo in 46 with 45 were accomplished in nꢀbutanol in 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
o
presence of DIPEA after microwaving at 225 C for 1 h to afford 47 in 66% yield. The
o
[1,2,3]triazolo[4,5ꢀb]pyrazine ring was constructed with isoamyl nitrile in DMF at 0 C and then
o
at 70 C for 1 h to afford 48 in 72% yield. The racemic 48 was purified on a chiral SFC column
using MeOH and liquid CO as elution system to provide 49 and 50. The absolute configuration
2
of 49 was determined by the xꢀray cocrystal structure with cꢀMet. A variety of compounds were
made using the chiral intermediate 49 via the conventional Suzuki coupling reaction.
a
Scheme II . Syntheses of (S)ꢀ6ꢀ(1ꢀ(1Hꢀ[1,2,3]triazolo[4,5ꢀb]pyrazinꢀ1ꢀyl)ethyl)quinoline
compounds
ACS Paragon Plus Environment
26

Page 27 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: (a) MeNHOMe/CDI, DMF, ambient temperature, overnight. (b)
o
MeMgBr, anhydrous THF, 0 C to ambient temperature, 16 h. (c) NH OHꢁHCl/NaOH, ethanol,
6 h. (d) 7N NH in methanol, EtOH, H , Raney nickel, 16 h. (e) DIPEA, nꢀBuOH, microwave at
25 C, 1 h. (f) isoamyl nitrile, DMF, 70 C, 1 h. (g) chiral separation on a chiral SFC column
using MeOH and liquid CO as elution system. (h) Pd(dppf)Cl ꢁCH Cl , Cs CO , DME/H O, 80
2
1
2
3 2
o
o
2
2
2
2
2
3
2
o
C, overnight.
CONCLUSIONS
Receptor tyrosine kinases (RTKs) play fundamental roles in cell proliferation, migration,
metabolism, differentiation, and survival. Alterations in the normal functions of protein kinases
are linked with the pathological changes within cell that leads to diseases. The constitutively
enhanced RTK activities have been implicated in the development and progression of many
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 57
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
32
types of cancer.
Since the first approval of ABL protein kinase inhibitor Gleevec for the
treatment of chronic myeloid leukemia patients with BCR-ABL abnormal gene in 2001, a
number of RTK inhibitors have been successfully discovered and developed against the aberrant
RTK signaling in various cancers, including sunitinib and axitinib for targeting the VHLꢀ
dependent VEGF pathway in renal cell carcinoma, erlotinib and gefitinib targeting EGFR in nonꢀ
small cell lung cancer (NSCLC) with mutant EGFR, and crizotinib targeting cꢀMet, Alk and Ros
in NSCLC. The discovery and development of highly selective kinase inhibitors have been
challenging because of the similarity of the protein kinase ATP binding sites. Most of the kinase
inhibitor drugs approved for cancer treatments are multiꢀtarget kinase inhibitors and suffer side
effects which limit their utility. From a highly specific HTS hit, we have discovered two cꢀMet
inhibitors 8 and 9 exhibiting exquisite kinase selectivity as well as many attractive properties.
During the broad ligand offꢀtarget screens, it was discovered that 8 was a panꢀPDE family
inhibitor, which was correlated to a sustained increase in heart rate, increased cardiac output, and
decreased contractility indices, as well as myocardial degeneration in in vivo safety evaluations
in rats. 8 was terminated as a preclinical candidate because of narrow therapeutic window in
cardioꢀrelated safety. 9, which has a higher LipE value for cꢀMet compared to 8 (7.02 vs 6.17)
demonstrated a more selective profile in the broad ligand offꢀtarget screen, and a large
therapeutic window in in vivo animal tolerance studies. 9 was selected as a preclinical candidate
and later entered human clinical evaluations for cancer treatment. These findings reinforce the
need for broad preclinical evaluations to identify a more complete range of affected
pharmacology.
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 attrition rate for drug development is rising, and late phase attrition remains high, which
33
contributes to an unsustainable increase in R&D spending. Although there are several factors
ACS Paragon Plus Environment
28

Page 29 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
that contribute to attrition, one of the major reasons is toxicity. 40% of NCEs that begin
3
4
preclinical safety studies in animals fail due to toxicity. It is important to establish in vitro
toxicity models to predict human in vivo toxicity, and identify issues at early drug discovery
stages. Broad offꢀtarget screening is one effective way to profile compounds at various stages
and alert chemists to potential issues that can be addressed during optimization. Although both 8
and 9 have exquisite kinase selectivity for cꢀMet, and desired in vitro ADME and in vivo PK
properties, the offꢀtarget inhibition of members of the PDE family by compound 8 caused severe
cardio toxicity in rats. 8 was terminated at the early stage. It is challenging to design
compounds with a completely clean profile in the offꢀtarget screens. LipE has been proposed as
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
5
an index that may have some correlation with a good toxicity profile. Indeed, 9 with a higher
cꢀMet cell LipE value than 8 (7.02 vs 6.17, respectively) demonstrated a much better profile in
broad ligand screens leading to good safety profile in in vivo preclinical animal studies.
Therefore, 9 was selected for early development. One of the major challenges in drug discovery
is to identify a compound with a good balanced profile in biological functions and
physicochemical properties. It is important at the early stage of drug discovery to establish an
effective screen cascade for multiꢀparameter optimizations, and identify the key ADME and
safety issues associated with the selected chemistry scaffolds. Optimization of multiple
chemistry series, or at least two chemical series, is highly recommended to avoid late attrition of
the project because of toxicity.
Parallel optimization of both triazolopyrazine and
triazolopyridazine series overcame the preclinical cardio toxicity from the triazolopyridazine
series and successfully selected 9 from the triazolopyrazine series into human clinical trials
without a significant delay of the project progression.
EXPERIMENTAL SECTION
ACS Paragon Plus Environment
29