Discovery of N -(4-(3-(2-Aminopyrimidin-4-yl)pyridin-2-yloxy)phenyl)-4-(4-methylthiophen-2-yl)phthalazin-1-amine (AMG 900), A Highly Selective, Orally Bioavailable Inhibitor of Aurora Kinases with Activity against Multidrug-Resistant Cancer Cell Lines

Article abstract of DOI:10.1021/acs.jmedchem.5b00183

Efforts to improve upon the physical properties and metabolic stability of Aurora kinase inhibitor 14a revealed that potency against multidrug-resistant cell lines was compromised by increased polarity. Despite its high in vitro metabolic intrinsic cleara

Full text of DOI:10.1021/acs.jmedchem.5b00183

Subscriber access provided by NEW YORK UNIV
Drug Annotation
Discovery of N-(4-(3-(2-Aminopyrimidin-4-yl)pyridin-2-
yloxy)phenyl)-4-(4-methylthiophen-2-yl)phthalazin-1-amine (AMG
00), a highly selective, orally bioavailable inhibitor of Aurora
9
kinases with activity against multidrug-resistant cancer cell lines
Stephanie D. Geuns-Meyer, Victor John Cee, Holly L. Deak, Bingfan Du, Brian L. Hodous,
Hanh Nho Nguyen, Philip R. Olivieri, Laurie B Schenkel, Karina R. Vaida, Paul Andrews,
Annette Bak, Xuhai Be, Pedro J. Beltran, Tammy L. Bush, Mary Katherine Chaves, Grace
Chung, Yang Dai, Patrick Eden, Kelly Hanestad, Liyue Huang, Min-Hwa Jasmine Lin, Jin
Tang, Beth Ziegler, Robert Radinsky, Richard Kendall, Vinod F. Patel, and Marc Payton
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b00183 • Publication Date (Web): 13 May 2015
Downloaded from http://pubs.acs.org on May 17, 2015
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 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of Nꢀ(4ꢀ(3ꢀ(2ꢀAminopyrimidinꢀ4ꢀ
yl)pyridinꢀ2ꢀyloxy)phenyl)ꢀ4ꢀ(4ꢀmethylthiophenꢀ2ꢀ
yl)phthalazinꢀ1ꢀamine (AMG 900), a highly
selective, orally bioavailable inhibitor of Aurora
kinases with activity against multidrugꢀresistant
cancer cell lines
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
Stephanie GeunsꢀMeyer,* Victor J. Cee, Holly L. Deak, Bingfan Du, Brian L. Hodous,⊥, ξ
,
⊥
⊥
⊥
⊥,◊
Hanh Nho Nguyen,⊥, η Philip R. Olivieri, Laurie B. Schenkel, Karina R. Vaida, Paul Andrews,
⊥
⊥
⊥
ꢁ
Annette Bak, ║,θ Xuhai Be, Pedro J. Beltran, Tammy L. Bush, Mary K. Chaves, Grace
Ψ
∇
∇, σ
║
∇
Ψ, φ
∇
∇
Ψ
Ψ
Chung, Yang Dai, Patrick Eden, Kelly Hanestad, Liyue Huang, MinꢀHwa Jasmine Lin,
Jin Tang, Beth Ziegler, Robert Radinsky,∇,σ Richard Kendall, Vinod F. Patel, Marc
Ω
∇,ϕ
∇
⊥,∞
Payton∇
⊥
║
Departments of Medicinal Chemistry, Pharmaceutical Research and Development
Ψ
ꢁ
∇
Pharmacokinetics and Drug Metabolism, Molecular Structure, and Oncology Research,
Amgen Inc., 360 Binney Street, Cambridge, Massachusetts 02142 and Amgen Inc., One Amgen
Center Drive, Thousand Oaks, California 91320.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 58
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: Efforts to improve upon the physical properties and metabolic stability of
Aurora kinase inhibitor 14a revealed that potency against multidrugꢀresistant cell lines was
compromised by increased polarity. Despite its high in vitro metabolic intrinsic clearance, 23r
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
(AMG 900) showed acceptable pharmacokinetic properties and robust pharmacodynamic
activity. Projecting from in vitro data to in vivo target coverage was not practical due to
disjunctions between enzyme and cell data, complex and apparently contradictory indicators of
binding kinetics, and unmeasurable free fraction in plasma. In contrast, it was straightforward to
relate pharmacokinetics to pharmacodynamics and efficacy by following the time above a
threshold concentration. Based on its oral route of administration, a selectivity profile that favors
Auroraꢀdriven pharmacology, and its activity against multidrugꢀresistant cell lines, 23r was
identified as a potential bestꢀinꢀclass Aurora kinase inhibitor. In phase 1 dose expansion studies
with GꢀCSF support, 23r has shown promising single agent activity.
INTRODUCTION
The serine/threonine Aurora kinases A and B (AurA and AurB) play essential and distinct
1
roles in mitosis. Aurora kinases have been implicated in tumorigenesis, with their
overexpression levels correlating to clinical staging of cancers and poor prognosis. Thus, these
mitotic kinases have been the subject of much interest as targets for anticancer therapy, and
2
many Aurora kinase inhibitors (AKIs) have entered clinical trials.
3
AurA is involved in centrosome maturation and the development of bipolar spindles.
Inhibition of AurA results in abnormal spindles, increased G /M cells, delay in mitotic
2
progression, and an increased degree of aneuploidy. AurB activity is necessary for maintaining
the spindle assembly checkpoint (SAC), ensuring proper chromosome alignment and
4
segregation, and enabling cytokinesis. Inhibition of AurB leads to abrogation of the SAC,
2
ACS Paragon Plus Environment

Page 3 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
thereby inducing cells to exit mitosis without dividing. This aborted cytokinesis is a unique
mechanism of blocking mitotic progression relative to other antimitotic mechanisms, including
AurA inhibition. In cancer cells, inhibition of AurB can induce apoptosis at the resulting 4N G1
stage; alternatively, cells may continue to progress through successive cell cycles without
division, resulting in a lethal accumulation of DNA. Interestingly, abrogation of the SAC by
inhibition of AurB overrides the phenotype induced by inhibition of AurA by itself so that the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
AurB inhibitory phenotype is dominant with an inhibitor of both kinases.
We have previously described the discovery of a series of pyridinylꢀpyrimidine phthalazine
AKIs with an AurB inhibitory phenotype, exemplified by the highly selective, orally bioavailable
6
compounds 14a and 14b (Table 2). Our target profile for the clinical candidate that would
evolve from this class of leads was framed by three AurB phenotype inhibitors in Table 1 that
7
8
were undergoing phase 1 clinical trials at the time: 1 (MKꢀ0457), 2 (PHAꢀ739358), and 3a
9
(
AZD1152). Given the increasingly crowded clinical arena, a compound developed from 14a/b
would need to possess potential advantages over known clinical compounds or be better suited to
test important clinical hypotheses about the effects of AK inhibition.
High selectivity against other kinases was a key attribute of the phthalazines we intended to
maintain. For example, 14b provided IC values <25 nM in AurB enzyme and cellular assays
5
0
and >58ꢀfold enzyme selectivity against a panel of 19 diverse kinases; 14b showed the strongest
nonꢀAK activity against Tieꢀ2 (872 nM) and no significant inhibition of 17 of the other kinases at
6
concentrations up to 25 ꢀM. Inhibition of offꢀtarget kinases may provide anticancer activities
through mechanisms other than AK inhibition; for example, 1 and 2 have demonstrated clinical
1
0
efficacy against T315I BcrꢀAblꢀmutated CML. Offꢀtarget activities may also lead to doseꢀ
limiting adverse events unrelated to AK inhibition, for example, hypertension observed with the
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 58
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
antiꢀangiogenic AurA inhibitor 7 (ENMDꢀ2076). Our aim was to test unconfounded clinical
effects of AK inhibition; hence a clinical candidate would require at least comparable kinase
selectivity to 14b for reasonable confidence that efficacy as well as doseꢀlimiting toxicity would
be onꢀmechanism.
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
Another feature of the phthalazine compounds that was deemed essential to preserve was their
potency against MDR cell lines that overexpress Pꢀglycoprotein (Pꢀgp). In contrast to
9
12
compounds 1, 2, and 3b (AZD1152ꢀHQPA ), many of the phthalazines possessed comparable
potency against the doxorubicinꢀselected Pꢀgpꢀoverexpressing uterine sarcoma cell line MESꢀSA
13
Dx5 relative to their potency against non Pꢀgpꢀexpressing cell lines such as the parental
MESꢀSA or HeLa cells. For example, compound 14a suppressed phosphorylation of the AurB
substrate Histone H3 (pꢀHH3) with a 24 h IC50 of 11 nM in MESꢀSA Dx5 cells, which is
6
comparable to its average EC (15 nM) for induction of polyploidy in HeLa cells. In transport
50
studies using MDR1ꢀLLCꢀPK1 cells, the AKIs 2 and 3 showed poor passive permeability, and 1
1
4
had a high efflux ratio (>43). It is plausible that the strongly basic amines of these AKIs
15
promote transportꢀmediated efflux and a resulting loss of potency in MDR cells.
Table 1. Aurora kinase inhibitors with known structures that have entered clinical trials
4
ACS Paragon Plus Environment

Page 5 of 58
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
b
ID
AK
AK cell assay (nM)
Most potentlyꢀinhibited nonꢀAKs
(nM) ; [Total kinases in panel]
Admin
route
c
Compound name
1
(MKꢀ0457/
A/B
A/B
B
MDAꢀMBꢀ231
43
49
16
FLT3 (6), PLK4 (9), ABL (13), MLCK
IV
IV
IV
IV
IV
7
12
16
VXꢀ680/tozasertib)
pꢀHH3
(15), RET (28); [317]
2
(PHAꢀ739358/
MDAꢀMBꢀ231
ABL (25), RET (31), TrkA (31),
8
12
17
danusertib)
pꢀHH3
FGFR1 (47); [35]
3a (AZD1152/
MDAꢀMBꢀ231
FLT3 (8), cKIT (17), PDGFRA (38);
d,9
12
16
barasertib)
pꢀHH3
PDGFRB (41), RET (80); [317]
18
4
(AT9283)
A/B
A/B
HCTꢀ116
DNA ploidy
∼30
JAK2 (1), JAK3 (1) Abl (T315I) (4), 9
others ≤ 10 nM; [230]
19
5 (SNSꢀ314)
HCTꢀ116
DNA ploidy
9
TrkB (5), TrkA (12), FLT4 (14), Fms
(15), DDR2 (82); Axl (84); [219]
2
0
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 58
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
2
1
6
7
8
9
(GSK1070916)
B
A
HCTꢀ116
pꢀHH3
20
130
480
21
FLT1 (42), TIE2 (59), SIK (70), FLT4
(74), FGFR (78); [328]
IV
PO
PO
PO
2
2
22
23
(ENMDꢀ2076)
HCTꢀ116
pꢀAurA
FLT3 (2), RET (10), FLT4 (16), SRC
(20) TrkA (24) Fms (25); [100]
24
(CYC116)
(ABTꢀ348)
A/B
A/B
A/B
A/B
A
A549
pꢀHH3
VEGFR2 (44), FLT3 (44), CDK2
(390); [23]
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
25
HCTꢀ116
pꢀHH3
VEGFR1 (1), FLT3 (1), VEGFR2 (2),
CSFꢀ1R (3), PDGFRꢀα (11) [128]
10 (AS703569/
R763)
A549
pꢀHH3
14
Cellꢀbased assays: VEGFR2 (11), FLT3 PO
(27), AMPK (201); [10]
2
6
27
1
1 (PFꢀ03814735)
MDAꢀMBꢀ231
pꢀHH3
∼ 50
FLT1 (10), FAK (22), TrkA (30), 17
others ≥ 90% inh@100 nM; [220]
PO
PO
PO
PO
2
8
12 (MKꢀ5108)
HeLa S3 < 1,000
pꢀHH3+ cells
HCT116
TrkA (2), ABL (8), FLT4 (12), TrkB
(13), VEGFR2 (30); [233]
↑
29
13a (MLN8054)
A
34
7
DRAK2 (8), BLK (68), DRAK1 (190),
1
6
pꢀAurA
FGR (220); [317]
13b (MLN8237/
A
HeLa
pꢀAurA
%inh@1 ꢀM: EphA2 (111), FGR (97),
CAMK2A (95), EphA4 (94); [220]
3
0
alisertib)
a
AK = Aurora kinase family member(s) inhibited (AurA and/or AurB; AurC potency not
listed.) Cell line; substrate or phenotype detected. Kinase activities of greatest potency listed in
published literature. Listed enzyme and cellular potency data is for 3b, the parent of prodrug 3a.
b
c
d
Cytotoxic agents are commonly dosed intravenously. However, we hypothesized that the
ability to dose on multiple consecutive days may be important for clinical efficacy of a
reversiblyꢀbinding agent that can target cells only when they are in the mitotic phase of the cell
cycle. Toward that end, the good oral bioavailability of the phthalazine AKIs would be another
characteristic to retain.
Despite the promising features of compounds such as 14a and 14b, we were concerned that their
marginal solubility (< 8 ꢀg/mL in PBS, pH 7.4) and fast in vitro metabolism [>80% turnover
upon 30 min incubation with human or rat liver microsomes (HLM or RLM)] could present a
challenge for achieving sufficient exposure in the clinic. On the other hand, for a cytotoxic agent
a moderate in vivo halfꢀlife (< 4 h) may be an advantage, providing good control over the
duration of exposure. Given uncertainty about the optimal level and duration of AurB inhibition,
flexibility was key. Perhaps it would be desirable to have 24ꢀhour coverage of the target with a
6
ACS Paragon Plus Environment

Page 7 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
single dose in humans, but with a schedule involving dosing over many days it could also turn
3
1
out to be useful to enable pulsatile coverage (AurB inhibition for a fraction of each day).
Therefore, we did not seek a long halfꢀlife in order to obtain extended inhibition of the aurora
kinases. Rather, reduction of the dose required to achieve significant daily target coverage was
pursued through a variety of means (such as increasing potency and reducing metabolism) in
order to minimize formulation/pharmacokinetic risk related to the marginal solubility of these
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
RESULTS AND DISCUSSION
32
Chemistry. Enaminone 15 was condensed with guanidine to provide 2ꢀchloropyridine
derivative 16, which was then reacted with 4ꢀaminophenol to produce aniline intermediate 20a.
The respective aminopyridine intermediate 20b was best prepared from S Ar reaction of
N
5ꢀaminopyridinꢀ2ꢀol with the more electrophilic 2ꢀfluoropyridine reagent 19, which itself was
obtained by Suzuki coupling of 4ꢀchloropyrimidinꢀ2ꢀamine 17 and 2ꢀfluoropyridineꢀ3ꢀboronic
acid 18. 20a was then reacted with 3ꢀchloroꢀ6ꢀphenylpyridazine to produce pyridazine
compound 21; similarly, 20a or 20b underwent S Ar with a variety of chlorophthalazines 22 to
N
produce phthalazinamines 23. Alternatively, Suzuki reaction of chlorophthalazine 23b with an
appropriate boronic acid provided 23k, 23p and 23q (Scheme I).
Scheme 1. Preparation of substituted pyridazine and phthalazine pyridinyl pyrimidine AKIs
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 58
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
N
R
H N
N
2
H N
N
H N
N
N
2
2
N
N
N
N
Cl
O
O
X
N
R
Cl
b
O
X
N
a
22
Cl
N
N
N
N
N
f
NH2
H
23
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
20a: X = C
20b: X = N
1
5
d
23a: X = C; R = H
2
2
2
2
3b: X = C; R = Cl
3c: X = C; R = phenyl
3d: X = N; R = phenyl
H N
2
N
N
H2N
N
N
3e: X = C; R = pyridin-2-yl
1
7 Cl
c
F
e
23f: X = N; R = pyridin-2-yl
g
N
23g: X = N;R = 5-methylpyridin-2-yl
23h: X = C;R = 5-methylpyridin-2-yl
23i: X = N; R = 6-methylpyridin-2-yl
+
HO
OH
19
B
F
H N
N
23j: X = C; R = 6-methylpyridin-2-yl
2
2
2
3k: X = C; R = 4-fluorophenyl
3l: X = C; R = 4-chlorophenyl
N
N
18
O
N
23m: X = C; R = 3-chlorophenyl
3n: X = C; R = benzo[d][1,3]dioxol-5-yl
N
2
N
N
23o: X = C; R = thiophen-3-yl
23p: X = C; R = thiophen-2-yl
H
21
2
2
3q: X = C; R = 5-methylthiophen-2-yl
3r: X = C; R = 4-methylthiophen-2-yl
Reagents and conditions: (a) GuanidineꢀHCl, NaOMe, MeOH, 50 ºC, 24 h (77%); (b) 4ꢀ
aminophenol or 5ꢀaminopyridinꢀ2ꢀol , Cs CO , DMSO, microwave, 150 ºC (77%); (c) PdCl (tꢀ
Bu PhP) , CH COOK, ACN/H O, 85 ºC, 15 h (91%); (d) 5ꢀaminopyridinꢀ2ꢀol hydrochloride,
Cs CO , DMF, 100 ºC, 24 h (56%); (e) 3ꢀchloroꢀ6ꢀphenylpyridazine, PPTS, sꢀBuOH, 100 ºC, 1
2
3
2
2
2
3
2
2
3
h (43%); (f) sꢀBuOH, tꢀBuOH, or DMSO, ~100 ºC 1ꢀ16 h (14ꢀ93%); (g) RꢀB(OH) , PdCl (dppf),
2
2
Na CO , H O, dioxane, 100 ºC, 1.5 h (10ꢀ30%).
2
3
2
Iodophenyl phthalazinamine reagent 25 was generated by reacting 4ꢀiodoaniline with 1ꢀchloroꢀ
33
4
ꢀphenylphthalazine 22c. Suzuki reaction with boronic ester 26 or Sonogashira coupling with
3
4
alkyne 28 provided quinazoline 27 and alkyne 29, respectively (Scheme 2).
Scheme 2. Synthesis of truncated N,4ꢀdiphenylphthalazinꢀ1ꢀamines
8
ACS Paragon Plus Environment

Page 9 of 58
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
Reagents and conditions: (a) 1ꢀchloroꢀ4ꢀphenylphthalazine, 110 ºC, overnight (88%) (b)
PdCl (dppf), Na CO , H O, dioxane, 90 ºC, overnight (52%); (c) CuI, (PPh ) PdCl , ACN/TEA,
2
2
3
2
3 2
2
90 ºC, overnight 52%)
3
5
The syntheses of chlorophthalazines 22g, 22i, 22l, and 22r have been described previously,
and 22e was obtained by the threeꢀstep procedure from N,Nꢀdimethylaminophthalamide that is
36
described in this work (Scheme 3). A process chemistry route to 22r has also been published.
2m, 22n, and 22o were provided by Suzuki reaction of 1,4ꢀdichlorophthalazine with an
2
appropriate boronic ester (Scheme 3).
Scheme 3. Synthesis of 4ꢀsubstituted 1ꢀchlorophthalazines
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 58
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
N
OH
N
O
N
b
R
N
a
O
O
31
30
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
R
HN
22e: R = pyridin-2-yl
22g: R = 5-methylpyridin-2-yl
22i: R = 6-methylpyridin-2-yl
22l: R = 4-chlorophenyl
O
32
22r: R = 4-methylthiophen-2-yl
N
R
c
N
Cl
N
Cl
d
N
22m: R = 3-chlorophenyl
22n: R = benzo[d][1,3]dioxol-5-yl
22o: R = thiophen-3-yl
Cl
33
Reagents and conditions: (a) RꢀBr, nꢀBuLi, THF, ꢀ78 ºC → rt; or 3ꢀmethylthiophene, nꢀBuLi,
TMEDA, Et O, reflux → rt (65ꢀ86%); (b) H NꢀNH , EtOH, reflux, overnight (81ꢀ98%); (c)
POCl , reflux, 18 h (32ꢀ99%); (d) RꢀB(OH) , PdCl (dppf), Na CO , H O, dioxane, 90 ꢀ 100 ºC,
2
2
2
3
2
2
2
3
2
1
ꢀ16 h (21ꢀ68%).
DISCUSSION
6
Compound 14a (Table 2) is relatively large (MW 498); therefore, SAR studies began by
assessing the potency contributions of various portions of this scaffold in order to remove any
atoms that were not critical for potency. The first modification was productive: removal of the
methyl group on the aminopyrimidine provided a 5ꢀfold improvement in cellular potency
34
(
14a→23c). Next, we pursued other structural modifications to reduce MW. Truncation of the
kinase hingeꢀbinding region of pyridyl pyrimidine 14a produced quinazoline 27 and alkynyl
pyrimidine 29, both of which maintained singleꢀdigit AurB enzyme potency (Table 2). However,
in contrast to compound 23c, the smaller variants 27 and 29 did not induce an AurB inhibitory
10
ACS Paragon Plus Environment

Page 11 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
phenotype (polyploidy) in a cellular assay at concentrations up to 1,200 nM (Table 2). Other
truncations aimed at reducing MW provided compounds 21 and 23a; while a 2ꢀ to 3ꢀfold loss of
AurB enzyme activity relative to 23c was observed, cellular AurB phenotype activity was
attenuated by 43ꢀ and 160ꢀfold, respectively, indicating that the fused and pendant aryl rings of
the phthalazine were also necessary for high cellular 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
Compounds 23c, 27, and 29 demonstrated good passive permeability and low efflux in MDR1ꢀ
LLCꢀPK1 cells, which suggested that the absence of cellular potency for the latter two
compounds was not related to poor cell penetration. In fact, AurB “false positives”—compounds
that showed strong inhibition of AurB in the enzyme assay and little or no potency in an AurB
cellular assay (despite independent evidence of good cell penetration)—were frequently
encountered in the original AurB HTS screen, in early lead optimization efforts, and also in
counterscreening efforts for other kinase inhibitor programs. A plausible explanation for this
phenomenon is that AK enzyme assays may fail to capture the intrinsic potency of compounds in
their cellular context, where they exist in multiprotein complexes that include the activating
37
38
partners TPX2 (AurA) and INCENP (AurB) . Although 27 and 29 showed AurB enzymatic
potency similar to 23c, both showed marginal potency relative to 23c in an AurA enzyme assay
in which the kinase is complexed with a peptide portion of its activating partner protein (TPX2)
(Table 2). However, for this set of compounds the apparent selectivity in favor of AurB vs. AurA
1
6
remained when TPX2 was removed as a variable (competitive binding K values ), thus leaving
d
the “real” selectivity unclear, as well as the extent to which the AurA assay was improved by use
of partner protein peptide. Other AKI drug discovery programs have relied on Aurora kinases
9,21
complexed to partner peptides in order to more closely approximate the cellular context. The
29
medicinal chemistry effort that culminated in 13a (MLN8054) was guided primarily by high
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 58
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
39
throughput cellular assays. Because of the high “false positive” rate using enzyme data we
found that it was crucial to focus on cellular data in order to interpret SAR. For simplicity AK
enzyme data is omitted from the tables that follow but is available in the Supporting Information.
Table 2. Efforts to reduce molecular weight
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
HeLa DNA
a
b
c
c
AurA/TPX2
(IC , nM) (IC , nM)
AurB
50
ploidy (EC ,
Papp
A→B
Efflux
ratio
AurA; AurB
5
0
a
ꢀ6
16
Cmpd
4a
14b
3c
MW
498
516
484
441
414
433
407
nM)
15±10
24±25
3±3
(x 10 cm/s)
(Ambit K , nM)
5
0
d
d
1
150±310
41±13
22±29
15±14
5±5
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
11
8.5
10
ꢀꢀ
2
32±24
1.5
1.6
0.9
ꢀꢀ
6; 4
250; 12
710; 33
ꢀꢀ
27
160±3
4±5
>1,200
>1,200
130
29
590±220
280±3
8±2
21
15±1
12±2
23a
1,100±11
490
ꢀꢀ
ꢀꢀ
ꢀꢀ
a
All kinase assays were run at the Km for ATP and represent an average of at least two
determinations. HeLa cellular assay with 10% FBS, Cellomics, ≥ 4N DNA content at 24 h.
EC50 values <25 nM represent an average of at least three determinations; the others are single
determinations. Apparent permeability, apical to basolateral velocity and efflux ratio in hMDR1ꢀ
LLCꢀPK1 cells. ꢀꢀ: Not determined.
b
c
d
12
ACS Paragon Plus Environment

Page 13 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In addition to the strategy of decreasing MW, we sought to decrease lipophilicity toward the
40
aim of reducing metabolism and improving physical properties, and therefore investigated
whether phenyl rings could be replaced by pyridines (Table 3). In HeLa cells, a nitrogen in the
central ring (23d) at position X provided comparable potency to carbon (23c), whereas a
nitrogen in the terminal ring (at Y) sacrificed potency by 9ꢀfold (23c→23e) or 15ꢀfold
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
(
23d→23f). However, potency could be restored by introduction of a metaꢀ or paraꢀ methyl
group on the terminal pyridine ring (compounds 23g-23j).
Despite potency in HeLa cells, nitrogen substitution had moderate costs with regard to potency
13
in an MDR cell line (MESꢀSA Dx5) relative to parental cells (MESꢀSA) that do not
overexpress Pꢀgp. Dose response curves for inhibition of pꢀHH3 in the two cell lines were
determined in parallel on a set of compounds that included 23g-c. The experiment was
performed twice in duplicate, resulting in IC s that for all compounds in Table 3 were
50
significantly higher in the MDR cells relative to parental cells (p <0.02, Students tꢀtest, twoꢀ
2
tailed). Introduction of one N increased PSA from 112 to 125 Å and provided IC values 1.9ꢀ to
5
0
2.5ꢀfold higher in MESꢀSA Dx5 cells relative to MESꢀSA cells (23d, 23e, 23h, 23j). For
2
compounds having two substitutions of N for C the PSA increased to 138 Å (23f, 23g, and 23i)
and the potency shifts from parental line to MESꢀSA Dx5 were 2.6ꢀ to 5.3ꢀfold. These
compounds were likely Pꢀgp substrates, with efflux ratios ≥16 in hꢀMDR1ꢀtransfected LLCꢀPK1
ꢀ6
cells. It is plausible that their high average permeabilities (≥ 25 x 10 cm/s) contributed to
keeping potency shifts in a moderate range. However, the quantitative relationship between
transporterꢀmediated efflux in cancer cell lines and loss of clinical efficacy remains unclear.
However, the quantitative relationship between transporterꢀmediated efflux in cancer cell lines
4
1
and loss of clinical efficacy remains unclear. Our aim was to remove Pꢀgpꢀmediated efflux as a
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 58
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
potential clinical variable by advancing a compound for which a potency shift into MESꢀSA Dx5
cells was less than 2ꢀfold.
Still, several compounds were further characterized in vitro and in vivo to gain an
understanding of the SAR of metabolism and plasma protein binding (PPB) and to assess
whether any of them were acceptable for evaluation in an in vivo pharmacodynamic (PD) assay.
There was a clear impact of increased polarity on protein binding: in contrast to 23c, compounds
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
with PSA of 125 Å had measurable fraction unbound (Fu) in human plasma (0.6ꢀ0.8%), and
2
compounds with PSA of 138 Å had human plasma Fu values of 3 to 4%. With regard to PK
properties, nitrogen at position Y reduced intrinsic CL in RLM (compounds 23e-j relative to
23c&d). However, in vitro metabolic clearance did not correlate well with in vivo rat CL for the
chemical series.
Compound 23j had measurable plasma Fu (assuming comparable human and rat values) and
reduced in vivo CL. Therefore its unbound clearance (CL/Fu) was significantly lower than that of
23c, albeit by a factor that cannot be quantified given the unmeasurably low plasma Fu of 23c.
On the other hand, since the cell assays (HeLa/MESꢀSA) were run in complete media containing
10% fetal bovine serum (FBS), it is likely that 23j was considerably more free under the assay
conditions than 23c; thus the ratio of the “true” IC s (23j IC _ / 23c IC50_unbound) is
50 unbound
5
0
42
probably higher than the ratio of the respective measured IC s. Assessment in a PD assay
5
0
would be required to determine whether the reduced free in vivo CL of 23j would translate into
an advantage that could override its potency disadvantage, including its moderate potency shift
(2.5ꢀfold) in MESꢀSA Dx5 vs. parental cells.
Table 3. Effects of nitrogen substitution on protein binding, cellular potency, and
pharmacokinetic parameters
14
ACS Paragon Plus Environment

Page 15 of 58
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
Cellular potency
(nM)
hMDR1ꢀ
LLCꢀPK1
Protein binding/
metabolism
Properties
PSA
Rat PK
MESꢀSA
HeLa
parental→
Turnover
d
DNA b Dx5
Papp
(x 10
Human (RLM;
A→B
c
ꢀ6
e
f
g
g
h
ploidy
pꢀHH3
PPB
HLM ,
CL
MRT
F
2
a
ID
(Å ) LogP
(EC50
)
(IC50
)
cm/s); [ER] (%)
%)
(L/h/kg) (h)
(%)
65
25
ꢀꢀ
2
2
2
3c 112 3.9
3d 125 3.4
3e 125 3.0
3±3
2.4→3.5
1.8→3.9
34→64
75→400
5.0→25
4.3→11
10→26
11; [1.5]
7.7; [6.6]
15; [2.8]
3.0; [16]
3.2; [19]
18; [3.3]
3.5; [19]
12; [2.2]
>99.9
99.4
99.2
96.0
96.8
99.4
97.0
99.4
93; 94
92; 18
46; 32
38; 13
63; 32
51; 36
50; 15
64; 20
1.0
2.2
ꢀꢀ
0.5
0.7
ꢀꢀ
4±3
27±1
23f
138 2.3
59±25
2
ꢀꢀ
ꢀꢀ
ꢀꢀ
23g 138 3.0
4.6
0.68
1.1
0.35
0.3
0.6
0.4
0.7
ꢀꢀ
2
2
2
3h 125 3.4
4
0
3i
3j
138 2.9
125 3.4
5±4
6±2
ꢀꢀ
6.0→15
33
a
43 b
Measured LogP based on LC/UV/MS. HeLa cellular assay with 10% FBS, Cellomics, ≥ 4N DNA
content at 24 h. Each IC represents the average of at least two determinations except for 23g & 23h
5
0
(
n=1). Values returned as < x are excluded from averages and are listed in the Supporting Information.
MESꢀSA [parental and Dx5 (doxorubicinꢀselected)] cellular assay with 10% FBS, Cellomics, pꢀHH3 at
4 h. Each IC represents the average of two independent experiments run in duplicate (n=4). For
c
2
5
0
d
standard deviations see the Supporting Information. Apparent permeability, apical to basolateral
e
f
velocity; efflux ratio in brackets. Human protein binding, equilibrium dialysis. Rat and human LM, %
percent substrate disappearance at 30 min. In vivo experiments were carried out with male Spragueꢀ
=
g
dawley rats (n=3). iv, 0.25 mg/kg for 23c, 0.5 mg/kg for all others; vehicles: 10% EtOH, 20% PEG400,
0% propylene glycol, 50% water for 23d; 100% DMSO for all others. po, 2 mg/kg, vehicle: 1%
Tween 80/2% HMPC, pH 2.2.
h
2
In parallel work, various nonꢀpyridine substituents on the phthalazine were explored in an effort to
increase cellular potency (Table 4). Following on the observation that lipophilic substituents provided
potency gains for pyridine compounds 23g-j, the terminal phenyl of compound 23c was substituted with
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 58
halogen atoms. No significant increase in potency was observed, and the Clꢀsubstituted compounds 23l
and 23m demonstrated somewhat reduced solubility at pH 2 relative to compound 23c (0.10 and 0.15
mg/mL versus >0.20 mg/mL) and reduced oral exposure in rats (F% = 16 & 27), respectively.
Interestingly, despite poor solubility, benzodioxolane compound 23n showed good oral absorption.
While the low in vivo CL and high bioavailability of 23n were attractive, the compound was a potent
CYP3A4 inhibitor (IC = 210 nM) and therefore not a good clinical candidate. Thiophenes also were
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
5
0
explored as replacements for the terminal phenyl ring (23o-23r). Methylꢀsubstituted thiophenes 23q and
23r provided greater potency than unsubstituted analogs 23o and 23p. Despite its high HLM CL,
4
ꢀMethylꢀ2ꢀthienyl compound 23r showed acceptable in vivo pharmacokinetics in rats and was selected
for in vivo assessment based on its potency against MESꢀSA Dx5 cells.
Table 4. Phthalazine substituent SAR with regard to cellular potency and PK parameters in rats
Microsomal
metabolism
Cellular potency (nM)
Rat PK
HeLa
Solubility
0.01 N
MESꢀSA
DNA a
Turnover
(RLM;
HLM, % )
d
d
e
CL
MRT
(h)
F
ID
R
parental→Dx5
b
ploidy
c
(L/h/kg)
(%) HCl
(mg/mL)
pꢀHH3 (IC )
44
5
0
(
EC )
50
2
2
2
2
2
3k
3±2
2.5→4.6
2.6→4.5
ꢀꢀ
90; 34
81; 43
93; 45
82; 56
95; 63
0.70
0.70
1.3
1.3
0.7
0.4
1.6
0.8
57
16
27
79
75
>0.2
3l
< 2
5
0.10
0.15
0.08
>0.2
3m
3n
3o
< 2
1.4→2.1
ꢀꢀ
0.15
1.1
15±12
ACS Paragon Plus Environment
16

Page 17 of 58
Journal of Medicinal Chemistry
1
2
2
2
3p
7
ꢀꢀ
93; 49
92; 82
82; 51
0.2
0.6
0.9
0.5
24
ꢀꢀ
0.16
ꢀꢀ
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
3q
2±1
3±2
ꢀꢀ
1.1
1.0
3r
1.9→2.3
73
>0.2
a
HeLa cellular assay with 10% FBS, Cellomics, ≥ 4N DNA content/cell at 24 h. Each IC represents
the average of at least two determinations except for 23n and 23 p (n=1). Values returned as < x are
excluded from averages and are listed in the Supporting Information. MESꢀSA [parental and Dx5
doxorubicinꢀselected)] cellular assay with 10% FBS, Cellomics, pꢀHH3 at 24 h. Each IC represents
the average of two independent experiments run in duplicate (n=4). For standard deviations see the
Supporting Information. Rat and human LM, % = percent substrate disappearance at 30 min. In vivo
experiments were carried out with male Spragueꢀdawley rats (n=3). iv, 0.5 mg/kg; vehicles: 10%
EtOH, 20% PEG400, 20% propylene glycol, 50% water for 23r, DMSO for all others, po, 2 mg/kg
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
50
b
(
50
c
d
e
(
2
23l, 23m, 23n) or 5 mg/kg (23k, 23o, 23p, 23r); vehicles: 1% Tween 80/2%HMPC pH 2.2 (23l, 23m,
3n, 23r) or OraPlus pH 2 (23k, 23o, 23p).
Phosphorylation of histone H3 in bone marrow cells was measured in the PD model as a surrogate
tissue indicator of in vivo activity against AurB. Mice were administered compounds 23c and 23r orally
at 3, 10, and 30 mg/kg. Six hours after a single dose of 23r or vehicle control, the percentage of bone
marrow cells with 4N DNA content and the level of pꢀHH3 in this cell subset were quantified by flow
cytometry. Compound 23c showed no significant suppression of pꢀHH3 (Figure 1A). In contrast,
compound 23r demonstrated significant (p < 0.02) inhibition of pꢀHH3 at a 10 mg/kg dose and
complete suppression (p < 0.001) with a 30 mg/kg dose (Figure 1A), with a good dose/exposure
relationship and plasma concentrations that were higher than for 23c at the same doses. Both
compounds exhibited a doseꢀdependent increase in the percentage of bone marrow cells with 4N DNA
content (Figure 1B), which for compound 23c is evidence that inhibition of AurB was likely achieved at
45
timepoints earlier than six hours postꢀdose.
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 58
Dose (mg/kg)
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
Vehicle
3
10
30
A
23c
23r
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
Plasma Conc (ꢀM)
Plasma Conc (ꢀM)
B
23c
23r
Plasma Conc (ꢀM)
Plasma Conc (ꢀM)
Figure 1. Pharmacodynamic response of compounds 23c and 23r 6 h after a single oral dose in female
mice (individual animal data shown, 3, 10, and 30 mg/kg; n=3 each group). Bone marrow cells were
collected and processed for analysis by flow cytometry. (A) Suppression of pꢀHH3 in bone marrow cells
gated for 4N DNA content versus plasma concentration. (B) Percentage of cells containing 4N DNA
versus plasma concentration.
As expected based on its lack of nitrogens in the central and terminal rings, compound 23r exhibited
high PPB (>99.9% in mouse, rat, dog, monkey, and human). Compound 23j, with its halfꢀpercent free
fraction and lower rat IV CL, was judged to have the potential to provide greater potency in vivo. The
two compounds were compared in a 6 h bone marrow PD study (23j with a highest dose of 15 mg/kg,
2
3r at 15 mg/kg,). In this experiment 23r provided 96% suppression of pꢀHH in 4N cells 6 h after
3
ACS Paragon Plus Environment
18

Page 19 of 58
Journal of Medicinal Chemistry
dosing, with an average plasma total concentration of 3.0 ꢀM at that time point; but 23j at 15 mg/kg
provided a 6ꢀh total plasma concentration of only 0.65 ꢀM, and at that level failed to show significant
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
46
evidence of AurB inhibition biochemically (pꢀHH3) or phenotypically (percentage of 4N cells).
Profiling of 23r
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
Robust and consistent performance of 23r in the PD assay prompted us to characterize it in greater
depth. As described above, two features of 14a and 14b that were deemed essential to preserve in a
clinical candidate were high selectivity against other kinases and potency against MDR cell lines. Data
for 23r demonstrating these attributes has been published: in a panel of 26 kinases only p38α and TYK2
4
7
12
enzymes returned IC values <500 nM (TYK2: 220 nM, p38α : 53 nM). Upon further screening at 1
50
ꢀM in Ambit’s competitive binding assays, 340 out of 353 nonmutant kinases returned % of control
4
8
values >20 [a cutꢀoff value corresponding to estimated Kd ∼250 nM [Conc * (POC) / (1 ꢀ POC)]]. Of
the remaining kinases, nonꢀAur kinases that returned followꢀup K values < 100 nM were: DDR1 (2
d
12
nM), LTK (13 nM), DDR2 (31 nM), FES (61), and Tieꢀ1 (97 nM).
With regard to cellular activity, remarkably consistent potency was observed across 26 tumor cell
1
2
lines in a cell count imaging assay, including several MDR cells. Table 5 compares a smaller set of
potency data for 23r to EC values for benchmark clinical AKIs 1, 2, and 3b. In contrast to 23r, the
50
benchmarks show significantly less potency in the MDR cells than in the sensitive cell line CAL51.
Similar results were obtained for 23r and the three benchmark compounds in 24 h assays measuring
12
pꢀHH3 inhibition or DNA ploidy in cell lines that overexpress Pꢀgp or BCRP.
a
Table 5. Potency of 23r and three clinical AurB inhibitors against sensitive and MDR cell lines
Cell Count, EC at 72 h, nM
5
0
Compound
CAL51
Breast)
MESꢀSA Dx5 HCT15
SNU449
(Liver)
769P
(Renal)
(
(Uterine)
(Colon)
2
1
2
3r
2
1
2
2
2
21
58
>625
>625
>625
>1250
>625
>1250
242
>625
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 58
3
b
54
>313
>313
>313
>313
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
Fluorescenceꢀbased cell count imaging assay (ArrayScan VTi); cells in complete media (with 10%
FBS) were treated for 24 h, washed twice, and incubated in complete media lacking inhibitor for 48 h.
The assay was run in duplicate.
23r thus demonstrated a desirable in vitro potency and selectivity profile in addition to promising in
vivo potency. However, given that its free fraction in plasma was unquantifiably low, it was crucial to
investigate the extent to which any species differences in free fraction might impact projections of
human in vivo potency and pharmacokinetics. To that end, suppression of pꢀHH3 in Jurkat cells 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
49
measured 3 h after incubation with compound 23r and serum (5% and 25%, Table 6). 23r potency
values with 5% and 25% serum were in good agreement across the species, which provided confidence
that PPB differences would not be a significant confounding variable in predicting human
pharmacokinetics, and also that a given total plasma concentration of 23r would provide comparable
AurB inhibition in mice and humans.
Table 6. Crossꢀspecies effect of serum addition on in vitro cellular potency of compound 23r
3
h pꢀHH3 suppression (Jurkat, IC , nM)
50
Species
5%
25%
Serum
Potency
shift
Serum
Mouse
Rat
42
189
4.5x
3.7x
4.6x
4.6x
3.9x
48
176
Dog
59
273
Monkey
Human
24 ± <1
54 ± 13
111 ± 4
208 ± 61
23r was preincubated with serum for 60 min and then Jurkat cells were exposed to 23r/serum for 3 h.
pꢀHH3 IC values were determined using a Cellomics ArrayScan VTI. Monkey and human serum were
evaluated in two independent experiments.
5
0
It is notable that the IC values in the serum shift assay (Table 6) are high relative to the 2ꢀ3 nM
5
0
potency of 23r in phenotypic and biochemical cellular assays (Table 4). This perhaps suggests slow
binding kinetics, given that the serum shift was a 3 h assay rather than 24 h as in the other cellular
ACS Paragon Plus Environment
20

Page 21 of 58
Journal of Medicinal Chemistry
assays. Indeed, a timeꢀdependence for apparent cellular potency of 23r is captured in a time course
experiment measuring the suppression of pꢀHH3 in Jurkat cells (Figure 2); the 24 h IC50 for AurB
inhibition by 23r was only 2ꢀfold lower than the 8 h IC , but it was 100ꢀfold lower than the 1 h IC .
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
5
0
50
The observed slow equilibration to maximal pꢀHH3 suppression could be the result of low
permeability, slow binding to AurB (low k ), or slow offꢀrates (low k ) for binding to the experimental
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
on
off
ꢀ6
apparatus or proteins in the media. The good passive permeability of 23r (∼10 x 10 cm/s in WT
1
4
LLCꢀPK1 cells) is evidence against the first possible reason. Data below suggests that slow binding to
AurB in cells is also unlikely (Figure 3).
Time (h)
pꢀHH3
1
3
8
24
119
9.6
2.4
1.2
(
IC50, nM) ± 8
± 2
± 0.2
± 0.03
Figure 2. Timeꢀdependent cellular potency of 23r. Jurkat cells in complete media with 10% FBS were
treated with 23r for the indicated time intervals and suppression of pꢀHH3 was measured by
fluorescenceꢀbased imaging assay (ArrayScan VTi, duplicate experiments, error bars: ±SD)
Regardless of cause, the timeꢀdependence of AK inhibition made it necessary to assess relative
cellular potencies of 23r against AurA versus AurB using the same duration of exposure. In experiments
that were designed to maximize the percentage of cells in mitosis for optimal AurA activity signal,
semiꢀsynchronized cells released from G ꢀS block 2 or 3 h prior were treated for 6 h with a
1
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 58
2
88
concentration range of 23r, and IC values for suppression of pꢀHH3 and pꢀAuroraꢀA Thr (pꢀAurA)
5
0
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
were determined (Table 7). In both HeLa and HCT116 cells, 23r inhibited biochemical activity of AurA
and AurB with equal potency, providing 6 h IC values (4ꢀ8 nM) that were between the 3 and 8 h IC s
5
0
50
of the Jurkat timeꢀcourse experiment in Figure 2.
Table 7. Equivalent cellular inhibition of AurA and AurB by 23r
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
Substrate
detected
Time
(h)
IC50
(nM)
Cell line
HeLa
HeLa
pꢀAurA
pꢀHH3
6
6
6
6
7 ± 1
8 ± 1
HCT116 pꢀAurA
HCT116 pꢀHH3
4 ± <1
4 ± <1
Cells semiꢀsynchronized with thymidine were released from G ꢀS phase and treated with 23r in
1
complete media with 10% FBS. Immunofluorescence imaging assay (ArrayScan VTi; Cellomics) was
used to determine the mean fluorescence intensity of individual mitotic objects, independent
experiments (n=2)
Slow binding kinetics of 23r to AurB was observed in an enzyme binding assay: the dissociation halfꢀ
life was observed to be ≥ 2.6 h by surface plasmon resonance (Table 8). The resulting K was in
d
agreement with the AurB K determined in Ambit’s competition binding assay (1.5 nM) as well as the
d
AurB enzyme IC (4 nM).
5
0
Table 8. AurB enzyme binding kinetics for 23r
a
b
kon
(
k
Kd
(nM)
t1/2
(h)
Rmax
(RU)
off
ꢀ1
ꢀ1 ꢀ1
M s ) (s )
4
ꢀ5
3
.6 x 10 ≤ 7.3 x 10
≤ 2.0
≥2.6 10.0
a
b
Enzyme kinetics were determined by surface plasmon resonance using a Biacore instrument. A
single determination provided the listed, statistically valid koff value; multiple other experiments
provided higher values that were flagged as unreliable by the Biacore T200 Evaluation software.
On the other hand, in a washꢀout experiment with Jurkat cells, 23r did not appear to have a slow AurB
enzyme offꢀrate: the percentage of cells that were pꢀHH3 positive returned to baseline levels (Figure 3)
two hours after removal of 23r, which is moderately longer than the time to restore pꢀHH3 levels after
ACS Paragon Plus Environment
22

Page 23 of 58
Journal of Medicinal Chemistry
washout of AKIs 1 and 2. In contrast, AurBꢀselective inhibitor 3b continued to suppress pꢀHH3 signal
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
several hours after washout; inhibition of AurB activity by this AKI may be limited by the halfꢀlife of
5
0
AurB itself, which is estimated to be only a few hours in cell culture. AurB inhibitor 6
21
(
GSK1070916) has been reported to have a very long residence time in AurB enzyme assays and to
5
1
wash out of cells slowly. The authors note that extension of an AKI into the hydrophobic pocket
adjacent to the ATP binding site is necessary but not sufficient for the slowꢀbinding kinetics observed
with 6 and related compounds. It is plausible that this hydrophobic pocket is also occupied by AKIs 5
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
9
25
(SNSꢀ314), 9 (ABTꢀ348), 3a, and 23r. In any case, after 23r was washed out of the cells, the AurBꢀ
inhibitory activity decreased faster than may be expected based on the slow offꢀrate observed in the
enzyme binding assay. This apparent discrepancy may be another example of an AKI enzyme assay
failing to capture the kinetics as well as thermodynamics of binding to the multiprotein complex that
exists in cells. Also, as mentioned above, the lack of a slow AurB offꢀrate in cells suggests that a slow
onꢀrate is unlikely to explain the timeꢀdependent cellular potency of 23r.
Figure 3. Jurkat cells in complete media with 10% FBS were treated for 3 h with compound at
concentrations previously determined by Western blot to provide maximal inhibition of pꢀHH3+ signal:
2
3r (250 nM), 1 (500 nM), 2 (1000 nM), 3b (250 nM). After drug washout, cells were harvested at the
indicated time intervals and assessed by flow cytometry for the percentage of pꢀHH3+ cells. Time zero
is baseline level of pꢀHH3 suppression at 3 h postꢀtreatment (arrow). DMSO pꢀHH3 level is indicated at
0
and 24 h timepoints. Data represents the mean pꢀHH3+ cells from two independent experiments.
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 58
The moderate residence time of 23r after washꢀout avoids one potential complication in the
development of an understanding of PK/PD/efficacy/toxicity relationships. However, complexities
around potential timeꢀdependence of binding to the AKs, plasma proteins, or assay surfaces show that it
is not straightforward to estimate expected in vivo duration and intensity of target coverage by
extrapolating from in vitro determinations of 23r potency. It would be important, then, to further
characterize the inhibition of AurB in vivo. To that end, mice bearing established COLO205 tumors
were administered a single oral dose of vehicle alone or 23r at 1.5, 3.5, 7.5, 15, and 30 mg/kg. Tumor
and bone marrow samples that were taken 3 and 6 h after dosing were assessed for the percentage of
pꢀHH3+ cells relative to that in control animals. As shown in Figure 4, the in vivo IC for pꢀHH3
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
5
0
suppression in COLO205 tumors was determined to be about 0.44 ꢀM. The in vivo IC in bone marrow
50
cells was somewhat higher, but the confidence intervals between the two experiments overlapped. This
and other PD experiments quantified the duration of time for which a given dose of 23r in mice
provides plasma concentrations that maintain at least 50% suppression of pꢀHH3. For example, from a
PK/PD study examining pꢀHH3 inhibition in HCT116 tumor xenografts by fluorescenceꢀbased
microscopy, a single 3.75 mg/kg dose of 23r was estimated to provide >50% inhibition of pꢀHH3 for
3ꢀ4 h, whereas a 7.5 mg/kg dose provided 6ꢀ8 h of target coverage, and 15 mg/kg inhibited pꢀHH3 for
12
>9 h.
ACS Paragon Plus Environment
24

Page 25 of 58
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 4. Athymic nude mice bearing human COLO205 tumors were treated with a single dose of 23r
at 1.5, 3.5, 7.5, 15, and 30 mg/kg or vehicle control alone (n=3 animals per dose group). Tumor and
bone marrow samples were collected at 3 h (ꢀ, black) and 6 h (ꢀ, red) after treatment. IC values were
5
0
determined by plotting observed pꢀHH3+ cells in G2/M as a POC (yꢀaxis) vs. 23r concentration in
plasma (xꢀaxis) from individual animals. The vehicle control treatment group was used as the control
[pꢀHH3+; bone marrow (mean 19.8, SD±2.3), COLO205 tumor (mean 10.8, SD±2.6)]. The predicted
bone marrow and tumor pꢀHH3 IC values in mouse plasma were 0.87 and 0.44µM, respectively.
5
0
These PD studies provided context for estimating the duration of AurB inhibition required for
suppression of tumor growth in vivo. For example, in the study illustrated in Figure 5, mice bearing
established HCT116 tumors were orally administered vehicle alone or 23r with three distinct doses and
schedules (n=10 animals for each group). Based on pharmacokinetic data for the lowest dose in the
above study, daily dosing of 23r at 3.0 mg/kg BID would provide >50% suppression of pꢀHH3 for
roughly 6ꢀ8 h/day, target coverage that resulted in 71% tumor growth inhibition (TGI). Comparable TGI
was observed with BID dosing of 15 mg/kg 23r on two consecutive days of every week, a dosing
scheme that is estimated to provide >18 h/day with plasma concentrations above the in vivo IC of 0.44
5
0
ꢀM. A 60 mg/kg QD dose on days 1 and 4 of every week provided 61% TGI. Terminal PK of this group
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 58
showed that 24 h after dosing a 60 mg/kg dose provided only 15 nM plasma concentration; therefore the
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
duration of target coverage above the in vivo IC provided by that dose was significantly less than 24 h.
50
All dosing regimens with 23r were wellꢀtolerated with no adverse effects on body weight, except in the
1
5 mg/kg 2ꢀconsecutiveꢀdays group (average of 5%, which includes one mouse with >20% body weight
loss). In all groups treated with 23r, a decrease in neutrophil counts (25ꢀ50% of the vehicleꢀcontrol
group) was observed at the end of the study. In other studies neutropenia induced by 23r has been
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
12
shown to be transient and reversible.
Figure 5. 23r inhibits the growth of HCT116 tumor xenografts in vivo with daily or intermittent dosing
schedules
1
2
The experiment above and tumor xenograft dose response studies reported earlier demonstrate that
preclinical efficacy is achieved from pulsatile inhibition of AurB activity by 23r, and that this inhibition
can occur with a variety of distinct schedules. Clearly 23r is a suitable candidate compound for studying
the pharmacodynamic/efficacy/toxicity relationship of AurB inhibition in mice. It is also wellꢀsuited for
preꢀclinical combination studies, for example with PEGylatedꢀgranulocyte colonyꢀstimulating factor to
1
2
52
53
attenuate neutropenia and microtubuleꢀtargeting agents or HDAC inhibitors to enhance efficacy.
43
Despite its high LogP (4.4), limited solubility (1 ꢀg/mL in PBS at pH 7 and 360 ꢀg/mL in the acidic
formulation listed in Table 4) and high in vitro metabolic intrinsic clearance, 23r exhibited good
ACS Paragon Plus Environment
26

Page 27 of 58
Journal of Medicinal Chemistry
pharmacokinetic properties in preclinical species and was predicted to have low CL (0.036 L/h/kg) and
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
4
good oral absorption in humans. Therefore we anticipated that a suitable salt form could provide
sufficient drug exposure to achieve efficacy in humans. Based on its potency, selectivity, dosing
12,52
flexibility, and efficacy against MDR tumor cells and tumor xenografts,
development candidate for clinical trials.
CONCLUSION
23r was selected as a
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
We have described our path toward identifying an AurB phenotype inhibitor clinical candidate from
advanced leads. These pyridinylꢀpyrimidine phthalazines had features we intended to retain (oral
bioavailability, high selectivity against other kinases, and relative potency against MDR cell lines) and
some areas for improvement (potency, solubility, PPB, and metabolism). Efforts to reduce LogP and
increase polarity by replacement of carbon with nitrogen did provide lower protein binding and better
microsomal stability, but at the expense of reduced potency in MDR cells. It was, however, possible to
achieve improved potency, and 23r was determined to be 5ꢀ to 8ꢀfold more potent than advanced leads
14a/b. Despite high in vitro metabolism, low solubility, and high PPB, 23r proved to have bestꢀinꢀseries
in vivo target coverage, efficacy in multiple tumor xenograft models at reasonable doses, and predicted
human pharmacokinetics that suggested adequate target coverage could be achieved. 23r is currently in
phase 1 trials for evaluation in adult patients with advanced solid and hematological cancers. Human PK
results have been consistent with predictions, and in dose expansion studies with GꢀCSF support, 23r
(AMG 900) has shown promising single agent activity in heavilyꢀpretreated patients with taxaneꢀ
5
4
resistant ovarian cancer as well as activity against other female reproductive cancers. Based on strong
5
5
preclinical and biomarker data, combination studies in patients with these tumor types are planned.
EXPERIMENTAL SECTION
Unless otherwise noted, all materials were obtained from commercial suppliers and used without
further purification. Anhydrous solvents were obtained from EMD or Aldrich and used directly. All
reactions involving airꢀ or moistureꢀsensitive reagents were performed under a nitrogen or argon
atmosphere. Microwaveꢀassisted reactions were conducted with a Smith synthesizer from Personal
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 58
Chemistry (Uppsala, Sweden). Silica gel chromatography was performed using either glass columns
packed with silica gel (200ꢀ400 mesh, Aldrich chemical) or medium pressure liquid chromatography
(MPLC) on a CombiFlash® Companion® (Teledyne Isco) with RediSep® normalꢀphase silica gel (35–
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
0 micron) columns and UV detection at 254 nm. Preparative reversedꢀphase HPLC was performed on a
Gilson (215 liquid handler), YMCꢀPack Pro C18, 150 x 30 mm I.D. column, eluting with a binary
solvent system A and B using a gradient elusion (A: H O with 0.1% TFA; B: CH CN with 0.1% TFA;
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
3
gradient elution) with UV detection at 254 nm. All final compounds were purified to ≥95% purity (94%
for 23g) as determined by Agilent 1100 Series high performance liquid chromatography (HPLC) with
UV detection at 254 nm using one of the following methods: (Method A: Zorbax SBꢀC8, 4.6 x 150 mm,
1
5 min; 1.5 mL/min flow rate; 0 to 100% 0.1% TFA in CH CN/0.1% TFA in H O; Method B:
3 2
Phenomenex Synergi, 2 x 50 mm, 3 min, 1.0 mL/min flow rate, 5 to 95% 0.1% formic acid in CH CN
3
/
0.1% formic acid in H O). Lowꢀresolution mass spectral (MS) data were determined on an Agilent
2
1
100 Series LCMS with UV detection at 254 nm and a low resonance electrospray mode (ESI). NMR
spectra were determined with either a Bruker AVꢀ400 (400 MHz) spectrometer or a Varian 400 or 300
MHz spectrometer at ambient temperature. Chemical shifts are reported in ppm from the solvent
resonance (DMSOꢀd 2.49 ppm). Data are reported as follows: chemical shift, number of protons,
6
multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m = multiplet), and coupling
constants. Differential scanning calorimetry (DSC) was performed on a TA Instruments Q100
calorimeter at 10 ⁰C/min from 30 to 300 ⁰C in a crimped aluminum pan under dry nitrogen at 50
mL/min.
4
-(2-Chloropyridin-3-yl)pyrimidin-2-amine (16). Sodium pellets (2.8 g, 123 mmol) were rinsed
with hexanes and placed in a 500 mL roundꢀbottomed flask fitted with an argon inlet and outlet. The
flask was cooled to 0 °C, 100 mL anhyd MeOH was added, and the reaction was stirred under argon.
After most of the sodium had been consumed, the reaction was warmed to rt and an additional 30 mL
anhydrous MeOH was added. After 15 min, guanidine hydrochloride (15 g, 152 mmol) was added and
the mixture was stirred for 30 min. (E)ꢀ1ꢀ(2ꢀChloropyridinꢀ3ꢀyl)ꢀ3ꢀ(dimethylamino)propꢀ2ꢀenꢀ1ꢀone
ACS Paragon Plus Environment
28

Page 29 of 58
Journal of Medicinal Chemistry
(
10.0 g, 47 mmol) was added in one portion and the yellow mixture was allowed to stir overnight. The
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
reaction was fitted with an airꢀcooled reflux condenser and was heated to 50 °C for 16 h. The mixture
was cooled and concd to a slurry, which was filtered, and the solid was rinsed with saturated NaHCO3
and water. The solid was dried in vacuo to give the title compound (8.3 g, 85% yield) as a white solid.
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
H NMR (Bruker, 400 MHz, DMSOꢀd ) δ: 8.51 (d, J = 4.4 Hz, 1H), 8.36 (d, J = 5.5 Hz, 1H), 8.01 (d, J
6
+
+
=
7.3 Hz, 1H), 7.55 (m, 1H), 6.67 (s, 3H). MS Calcd for C H ClN : [M] = 206. Found: [M+H] = 207.
9 7 4
4
-(2-Fluoropyridin-3-yl)pyrimidin-2-amine (19). A pressure bottle was charged with bis(diꢀtertꢀ
butyl(phenyl)phosphine)palladium dichloride (0.288 g, 0.463 mmol), potassium acetate (4.55 g, 46.3
mmol), 2ꢀfluoropyridineꢀ3ꢀboronic acid hydrate (4.89 g, 34.7 mmol), 4ꢀchloropyrimidinꢀ2ꢀamine (3.00
g, 23.2 mmol) and was purged with nitrogen for several minutes. Water (12.5 mL, 695 mmol) and
acetonitrile (93 mL) were added, the bottle was sealed, and the reaction mixture was heated at 85 ºC for
1
5 h. The mixture was diluted with water and DCM, and the layers were separated. The organic portion
was washed with saturated NaHCO , dried with Na SO , filtered, and concd. The resulting solid was
3
2
4
triturated with ether and filtered to provide 4ꢀ(2ꢀfluoropyridinꢀ3ꢀyl)pyrimidinꢀ2ꢀamine (4.02 g, 91%
1
yield) as a light yellow solid. H NMR (400 MHz, DMSOꢀd ) δ ppm 8.50 (1H, t, J = 8.0 Hz), 8.38ꢀ8.35
6
+
(2H, m), 7.55ꢀ7.52 (1H, m), 7.05ꢀ7.02 (1H, m), 6.83 (2H, br s). MS Calcd for C H FN : [M] = 190;
9
7
4
+
Found: [M+H] = 191.
4-(2-(4-Aminophenoxy)pyridin-3-yl)pyrimidin-2-amine (20a). 4ꢀ(2ꢀChloropyridinꢀ3ꢀyl)pyrimidinꢀ
ꢀamine (15, 6.0 g, 29.0 mmol), 4ꢀaminophenol (3.17 g, 29.0 mmol) and Cs CO (18.9 g, 58.1 mmol)
2
2
3
were suspended in 50 mL DMSO under nitrogen in a resealable pressure vessel, and the mixture was
heated to 130 ºC overnight. After cooling to room temperature (rt) and diluting with water, the resulting
precipitate was isolated by filtration, washed with water and diethyl ether, and dried under high vacuum
1
to provide 4ꢀ(2ꢀ(4ꢀaminophenoxy)pyridinꢀ3ꢀyl)pyrimidinꢀ2ꢀamine (7.5 g, 92%). H NMR (Bruker, 400
MHz, CDCl ) δ: 8.33ꢀ8.14 (m, 3H), 7.27 (d, J = 5.1 Hz, 1H), 7.20 (t, J = 4.8 Hz, 1H), 6.87ꢀ6.73 (m,
3
ACS Paragon Plus Environment
29