Novel 6-aminofuro[3,2-c]pyridines as potent, orally efficacious inhibitors of cMET and RON kinases

Article abstract of DOI:10.1016/j.bmcl.2013.05.074

A series of novel 6-aminofuro[3,2-c]pyridines as kinase inhibitors is described, most notably, OSI-296 (6). We discuss our exploration of structure-activity relationships and optimization leading to OSI-296 and disclose its pharmacological activity against cMET and RON in cellular assays. OSI-296 is a potent and selective inhibitor of cMET and RON kinases that shows in vivo efficacy in tumor xenografts models upon oral dosing and is well tolerated.

Full text of DOI:10.1016/j.bmcl.2013.05.074

Bioorganic & Medicinal Chemistry Letters 23 (2013) 4381–4387
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel 6-aminofuro[3,2-c]pyridines as potent, orally efficacious
inhibitors of cMET and RON kinases
a,
a
a
a
a
a
Arno G. Steinig a, , An-Hu Li , Jing Wang , Xin Chen , Hanqing Dong , Caterina Ferraro , Meizhong Jin ,
Mridula Kadalbajoo a, Andrew Kleinberg a, Kathryn M. Stolz a, Paula A. Tavares-Greco a, Ti Wang a,
Mark R. Albertella a, Yue Peng a, Linda Crew a, Jennifer Kahler a, Julie Kan a, Ryan Schulz a, Andy Cooke b,
Mark Bittner b, Roy W. Turton b, Maryland Franklin b, Prafulla Gokhale b, Darla Landfair b,
⇑
⇑
Christine Mantis b, Jen Workman b, Robert Wild b, Jonathan Pachter a, David Epstein a, Mark J. Mulvihill a
a OSI Pharmaceuticals, LLC, A Wholly-Owned Subsidiary of Astellas US, 1 Bioscience Park Drive, Farmingdale, NY 11735, USA
b OSI Pharmaceuticals, LLC, A Wholly-Owned Subsidiary of Astellas US, 2860 Wilderness Place, Boulder, CO 80301, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of novel 6-aminofuro[3,2-c]pyridines as kinase inhibitors is described, most notably, OSI-296 (6).
We discuss our exploration of structure–activity relationships and optimization leading to OSI-296 and
disclose its pharmacological activity against cMET and RON in cellular assays. OSI-296 is a potent and
selective inhibitor of cMET and RON kinases that shows in vivo efficacy in tumor xenografts models upon
oral dosing and is well tolerated.
Received 27 March 2013
Revised 20 May 2013
Accepted 21 May 2013
Available online 30 May 2013
Ó 2013 Elsevier Ltd. All rights reserved.
Keywords:
Furopyridines
Tyrosine kinase inhibitor
Tumor growth inhibition
cMET
RON
2-Aminopyridines 1 are a common chemotype for inhibitors of
kinases. Examples include the recently approved ALK/cMET inhib-
itor Crizotinib (PF-02341066, 2),1 AXL/cMET inhibitor 3 first dis-
closed by AstraZeneca,2 and the CHK2 inhibitor 4.3 Typically, the
pyridine nitrogen acts as hydrogen bond acceptor for a backbone
NH of the hinge region, while the 2-amino group donates a hydro-
gen bond to the interior hinge carbonyl,4 thus interacting with the
same residues as ATP in a mutually exclusive way. The C3 substi-
tuent then reaches towards the interior of the kinase, while the
C5 substituent projects through a channel towards the solvent-ex-
posed region.5 Fusing an additional (hetero)aromatic ring onto the
2-aminopyridine core via its 4- and 5-positions would retain the
hinge binding moiety while providing different vectors for substit-
uents to interact with the protein. Herein, we demonstrate this
concept with 6-aminofuro[3,2-c]pyridines 5 and disclose the opti-
mization of this series leading to the potent cMET/RON inhibitor
OSI-296 (6) (Fig. 1).
homology.6 They are activated by their respective ligands HGF
(Hepatocyte Growth Factor) and MSP (Macrophage Stimulating
Protein) and signal via the PI3K/AKT and MAPK pathways. Signal-
ing through the cMET/HGF system can be deregulated in cancer
by mechanisms such as HGF-dependent autocrine activation, cMET
gene amplification, cMET transcriptional upregulation, chromo-
somal rearrangements, and/or the presence of activating muta-
tions.7 Approaches to abrogate aberrant cMET signaling that have
led to agents in clinical trials include inhibiting the kinase function
of cMET with small molecules and preventing ligand–receptor
binding with HGF or cMET antibodies.8 While activating mutations
or gene amplification of RON in cell lines or tumor samples have
not been reported, constitutively active RON variants generated
by alternative splicing9 or by methylation-dependent promoter
usage [short-form RON (sfRON)]10 have been identified. Cells
expressing these RON proteins show greater scatter activity, focus
formation, anchorage-independent growth, and tumor formation
in nude mice compared with cells expressing wild-type RON, sug-
gesting that activation of RON may contribute to the invasive phe-
notype of various human carcinomas.9,10
cMET and RON are receptor tyrosine kinases of the MET proto-
oncogene family that share significant sequence and structural
The cellular and in vivo component of our cMET drug discovery
cascade was built upon the MKN45 human gastric carcinoma cell
line. This cancer cell line displays a high level of amplification of
⇑
Corresponding authors. Tel.: +1 845 490 8365 (A.G.S.).
E-mail addresses: arnosteinig@gmail.com (A.G. Steinig), anhu_li@yahoo.com (A.-
H. Li).
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.05.074

4382
A. G. Steinig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4381–4387
1
Cl
R
O
O
3
5
R"
N
N
R'
a
N
F
13
N
N
O
NH
Cl
O
H2N
N
H2N
N
B
N
2
O
N
H
1
5a–c
NBoc
14
H2N
N
2
O
Cl
O
S
Cl
O
N
S
H2N
Cl
N
NH
F
O
b
F
O
N
3
Cl
O
R
Cl
a
O
Br
H2N
N
3
H2N
N
4
H2N
N
H2N
N
5o–t
(R)-13a
2
O
O
B
F
O
R'
7
R"
NBoc
c
3
O
O
Cl
O
N
15
H2N
N
Cl
Cl
6
NH2
H2N
N
5
6 (OSI-296)
F
O
F
O
O
Cl
O
Cl
O
NH
Figure 1. 2-Aminopyridines and 6-aminofuro[3,2-c]pyridines as kinase inhibitors.
N
H
H2N
N
H2N
N
O
O
O
5w
5u
b
a
c
HO
Br
d
e
O2N
N
N
O
O
N
Cl
Cl
O
H
9
8
7
1
1
1
R
O
O
F
F
R
O
R
O
O
O
N
O
e
O
Br
d
Cl
O
Cl
Br
N
NH
6
R
H2N
N
13
H2N
N
H2N
O2N
N
11
O2N
N
12
5v
5x–aa, 6
Scheme 2. Synthesis of the furo[3,2-c]pyridines 5 and 6. Reagents and conditions:
(a) (i) Pd(PPh3)4, K2CO3, dioxane/water or DME/water 4:1, 100 °C; (ii) 4 M HCl in
dioxane, rt; (b) (pinacolato)B–R3 or (HO)2B–R3, Pd(PPh3)4, K2CO3, dioxane/water or
DME/water 4:1, 100 °C; or Bu3Sn–R3, Pd(PPh3)4, KF, dioxane, 100 °C; (c) (i) CO,
Pd(PPh3)4, DIPEA, MeOH, 68 °C, overnight, (ii) 1 M HCl aq, 100 °C, 7 h, (iii) TBTU,
MeNH2ꢀHCl, DIPEA, DMF, rt, 30 min; (d) R6C(@O)Cl or R6–NCO or TMS–NCO; (e) H2,
Pd/C, EtOAc, rt, 1 h.
Scheme 1. Synthesis of the furo[3,2-c]pyridine intermediates 12 and 13. Reagents
and conditions: (a) (i) NBS, DMF; (ii) POBr3, toluene; (iii) sodium formate, Pd(PPh3)4,
DMF, 100 °C, overnight; (b) (i) Pd2dba3, 2-di-tert-butylphosphino-3,4,5,6-tetra-
methyl-20,40,60-triisopropyl-1,10-biphenyl, KOH, dioxane/water 1:1, 100 °C, over-
night, (ii) concd HNO3, 0 °C, 30 min; (c) R1-OH (10), DIAD, PPh3, THF, 0 °C to rt; (d)
(i) Br2, CCl4, 50 °C, overnight, (ii) DBU, THF 0 °C, 1 h; (e) Fe powder, cat. HCl, EtOH,
95 °C, 1 h.
Cl
cMET and constitutive activation of cMET and can be used for
xenografts in immunocompromised mice. Treatment of this cell
line with a selective cMET inhibitor was reported to result in
induction of apoptosis and inhibition of proliferation, whereas
non-MET-amplified cell lines were not affected.11 This cell line is
thus ‘driven’ by c-MET, and antiproliferative effects correlate with
the inhibition of cMET phosphorylation so that the cell prolifera-
tion IC50 values12 can be used as surrogate for the cMET cell mech-
anistic IC50 values.13,14 Cellular activity against RON was assessed
in HeLa cells transiently transfected with sfRON, monitoring the
endogenous pRON at the autophosphorylation site Y1238/1239.15
The synthesis of the key furo[3,2-c]pyridine intermediates 12
and 13 is shown in Scheme 1. The furopyridone 716 was converted
to the known 7-bromofuro[3,2-c]pyridine (8) in three steps by
reaction with NBS, treatment with POBr3, and selective reduction
of the 4-Br substituent.17 Conversion of the 7-Br group to 7-OH
and subsequent nitration at C6 provided 9, onto which the R1 sub-
stituent was introduced by Mitsunobu reaction with the alcohols
10. Addition of bromine to the furan ring, followed by elimination
of HBr with DBU gave intermediate 12. Subsequent reduction of
the nitro group yielded 13.
O
N
N
F
O
HO
a
Cl
O
Br
O2N
N
O2N
N
NBoc
16
12a
1
R
O
O
N
N
b
H2N
N
N
5a,d,e,g–n
H
Scheme 3. Alternative synthesis of the furo[3,2-c]pyridines 5. Reagents and
conditions: (a) (i) 14, Pd(PPh3)4, K2CO3, dioxane/water 4:1, 100 °C, (ii) 48% aq
HBr, 60 °C, overnight, (iii) Boc2O, NEt(iPr)2, DCM, 0 °C, 1–3 h; (b) (i) R1–OH (10),
DIAD, PPh3, THF, 0 °C to rt, (ii) Fe powder, cat. HCl, EtOH, 95 °C, 1 h; (iii) 4 M HCl in
dioxane, rt.
hetaryl, and heterocyclyl substituents could be introduced by Su-
zuki or Stille coupling, followed in the case of Boc-protected boro-
nates 141a and 15 by deprotection with HCl (5a–c,o–u). The
tetrahydropyridine 5u thus obtained was subsequently acylated
with acid chlorides, chloroformates, carbamoyl chlorides, isocya-
Compound 13 served as a versatile intermediate for incorporat-
ing substituents at the 3-position of the furo[3,2-c]pyridine core,
providing a second point of diversity (Scheme 2). A variety of aryl,

A. G. Steinig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4381–4387
4383
R 5
Table 2
H
O
b
a
SAR of the phenyl ring in R1
R
R
O
R
OH
18
10
17
R
O
Scheme 4. Synthesis of the alcohols 10. Reagents and conditions: (a) R5MgBr or
N
N
O
R5MgCl, THF; (b) NaBH4, EtOH.
H2N
N
Table 1
N
H
cMET and RON potencies of 5a
Cl
Compda
R
cMET biochem. IC50
M)
MKN45 prolif. IC50
(lM)
(l
F
O
5a
5b
5c
5d
5e
5fb
2,6-Di-Cl-3-F
2,6-Di-Cl
6-Cl-3-F
0.086
0.22
0.29
0.36
0.11
0.33
0.57
2.0
6.4
5.1
1.3
1.6
N
N
Cl
O
H2N
N
2-Cl-3-F
N
H
2,3,6-Tri-Cl
2,6-Di-Cl-3-
MeO
2,6-Di-Cl-3,5-
di-F
5a
Cmpd cMET
biochem. IC50
M)18
cMET cell
mech. IC50
M)
MKN45
proliferation
sfRON cell
mech. IC50
5g
0.042
0.48
(l
(l
IC50
(l
M)
(l
M)14
5h
5i
2-Cl-3-F-6-Br
2,4,6-tri-Cl
0.089
0.13
0.51
3.6
5a
0.086
0.56
0.57
0.13
a
All compounds are racemic except for 5f [(R) enantiomer].
Prepared from Boc-protected (R)-5a by treatment with sodium methoxide in
b
DMSO followed by removal of the Boc group.
nates, or sulfamide to afford compounds 5x–ab and 6. Hydrogena-
tion of 5u gave the piperidine analog 5v. In addition, carboxyl
derivatives could also be prepared, exemplified by the synthesis
of the amide 5w from (R)-13a by palladium-catalyzed carbonyl-
ation with CO gas in methanol, hydrolysis of the resulting methyl
ester to the acid, and amide formation with methylamine.
While the route outlined in Scheme 2 allowed rapid exploration
of the C3 substituent, the R1 substituent was introduced fairly
early. Therefore, an alternative synthesis was used for the explora-
tion of R1 that took advantage of the advanced intermediate 12a
(Scheme 3). After Suzuki coupling with boronate 14, the 1-(2,6-di-
chloro-3-fluorophenyl)ethyl moiety was removed cleanly with 48%
aq HBr. Reprotecting the piperidine nitrogen as Boc carbamate
gave the key intermediate 16. Using similar conditions as for the
corresponding steps in the other route, 16 was reacted with the
alcohols 10 via Mitsunobu reaction, followed by reduction of the
nitro group and removal of the Boc group to give the final com-
pounds 5.
Table 3
SAR of the benzylic position R5 in R1
Cl
5
R
F
O
N
N
Cl
O
H2N
N
N
H
Compd R5
cMET biochem
IC50 M)
cMET cell mech.
MKN45 prolif.
IC50 (lM)
(l
IC50
(
l
M)
(R)-5a
(S)-5a
5j
5k
5l
(R)-CH3
(S)-CH3
H
0.047
0.66
0.43
0.41
7.4
>10
2.2
Not det.
2.1
0.56
0.30
4.5
8.6
3.6
2.7
rac-Et
rac-nPr
rac-Allyl
rac-
0.092
0.13
0.11
The benzylic alcohols 10 were prepared from the corresponding
aldehydes 17 by addition of the Grignard reagents R5MgBr or
R5MgCl or by reduction of the acetophenones 18 with NaBH4
(Scheme 4). The racemic mixtures were separated into the enanti-
omers either at the alcohol stage or as final compound.
5m
5n
Not det.
0.36
0.070
Propargyl
potency (5i vs 5b). While 5a, 5g, and 5h were comparable in cMET
potency, the substitution pattern of 5a was preferred because of its
lower molecular weight and better synthetic accessibility and was
therefore held constant during further optimization.
Initial proof-of-concept efforts targeted the direct furopyridine
analog 5a of compound 2 (Table 1). This racemic compound
showed sub-lM cellular activity against cMET and sfRON, as well
as good metabolic stability as measured by extraction ratios
(ERs) of 0.47 and 0.30 in mouse and human liver microsomes,
respectively, and was considered a viable starting point. Its 1-
(piperidin-4-yl)-1H-pyrazol-4-yl moiety was chosen as C3 substi-
tuent group for SAR exploration and optimization of the R1 group,
focusing on cellular potency against cMET.
Data for exploration of the substitution pattern of the phenyl
ring in R1 are shown in Table 2. The 3-fluoro substituent was
important, as removing it (5b) or replacing it with a chloro (5e)
or methoxy (5f) group resulted in less potent compounds. An addi-
tional fluorine at the 5-position (5g) improved the cMET biochem-
ical potency; however, this improvement did not translate to the
cellular context. Removing either of the ortho-chloro substituents
(5c, 5d) imparted a 3- to 11-fold potency loss, while replacing
the 6-chloro with a bromo group (5h) retained cMET potency.
Incorporating a chloro substituent at C-4 did not improve the
Table 3 shows the SAR at the benzylic position in R1. The (R)-
methyl substitution was preferred over its (S) enantiomer by 14-
to 20-fold [(R)-5a vs (S)-5a]. The methyl group proved to be supe-
rior in terms of potency to hydrogen (5j), other alkyl groups (5k–
5l), or allyl (5m), especially at the cellular level. While a propargyl
group (5n) had comparable cMET potency to a methyl group, the
compound showed significantly increased inhibition of CYP3A4
(IC50: (R)-5a, >20 lM; 5n, 1.6 lM) and was therefore deprioritized.
Based on this exploration, the 2,6-di-chloro-3-fluoro substitution
pattern on the phenyl ring and (R)-CH3 substitution at the benzylic
position in R1 were preferred.19
An X-ray structure of (R)-5a bound to cMET confirmed the ex-
pected binding mode with the amino acid backbone of the hinge
region (Fig. 2): the 6-amino group interacts with the backbone
carbonyl of Pro1158, and the core nitrogen forms a hydrogen bond

4384
A. G. Steinig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4381–4387
Table 5
SAR of the R3 substituent
Cl
F
O
3
Cl
O
R
H2N
N
Compda R3
cMET biochem MKN45
IC50 M) proliferation IC50
M)
(
l
(l
N
N
(R)-5a
0.047
0.30
NH
N
N
O
5o
0.95
0.67
3.3
11
4.6
N
5p
5q
5r
5s
5t
Pyrazol-4-yl
0.19
0.59
1.6
0.52
0.26
Phenyl
Pyridin-2-yl
Pyridin-3-yl
Pyridin-4-yl
2.7
0.90
5u
NH
0.030
0.45
5v
NH
0.20
0.59
3.1
3.5
Figure 2. X-ray structure of (R)-5a/cMET (PDB code 4KNB). Shown are Tyr1230
from the activation loop and the backbone of the residues in the hinge region
(Pro1158, Tyr1159, Met1160, Lys1161).
5w
CONHCH3
O
O
5x
5v
0.32
0.78
0.33
0.85
N
N
Table 4
Mouse PK data for (R)-5aa
O
O
iv Dosing
po Dosingb
N
N
5z
0.58
0.28
0.51
0.19
Dose (mg/kg)
Vss (L/kg)
Cl (mL/min/kg)
2
12
227
Dose (mg/kg)
Cmax (lM)
AUC0–last (ng h/mL)
%F
20
0.15
509
35
N
O
t
(h)
0.6
½
5aa
a
PK parameters calculated by non-compartmental modeling using the median
concentration at each timepoint (three animals per timepoint).
N
H
O
b
Dosed as solution of bis-HCl salt in 1:1 PEG400-water.
N
N
6
0.14
1.0
0.13
NH2
with the backbone NH of Met1160.20 In addition, the 2,6-dichloro-
3-fluorophenyl ring stacks tightly against the phenyl ring of
Tyr1230 from the activation loop with a closest distance of
3.26 Å. Based on the structure, the improved potency of the 3-flu-
oro compound 5a compared to 5b, 5e, or 5f may be explained by a
O
S
O
5ab
0.55b
NH2
a
All compounds are the (R) enantiomers.
cMET mechanistic IC50
b
.
strengthened p–p interaction with the Tyr1230 phenyl ring, an en-
hanced pseudo H-bond between the adjacent CH and the carbonyl
of Arg1208 (distance 3.78 Å), and by direct interactions with the b-
was suspected that the high basicity of the piperidyl moiety in
(R)-5a, while beneficial for solubility, might contribute to the high
clearance and Vss as well as low permeability (as indicated by the
PAMPA assay). Thus, the emphasis was placed on compounds with
reduced basicity (Table 5).
Acetylation of the piperidine (compound 5o) improved the per-
meability (PAMPA: Pe at pH 7.4 = 318 ꢁ 10ꢂ6 cm/s) but also de-
creased the potency. Similar results were obtained with urea,
carbamate, and sulfonamide analogs (data not shown). Addition-
ally, the molecular weight of these analogs was >500, a less desir-
able feature. Rather than further modifying the piperidine ring, we
truncated the R3 substituent and explored various proximal rings
(compounds 5p–5v). Among those, the tetrahydropyridine analog
5u stood out in terms of potency. The corresponding piperidine
5v was about sevenfold less potent, indicating that planarity with
the furopyridine core was important. A set of amides, represented
by 5w, was also prepared that aimed to form a hydrogen bond
carbon of Ala1221, the sulfur of Met1211, and the
Met1211 (distances 3.48–3.73 Å).
c-carbon of
Compound (R)-5a, emerging as the best compound from this
first round of optimization, showed metabolic stability as mea-
sured by extraction rations (ERs) of 60.7 in both mouse and human
liver microsomes (human, 0.56; mouse, 0.19) and was progressed
to single-dose PK in female CD-1 mice. Very high clearance and
high volume of distribution indicated extensive tissue distribution,
resulting in low plasma levels and a short iv half-life. In addition,
the overall exposure was low (Table 4). No significant metabolites
were identified in plasma samples, indicating that the low expo-
sure and rapid clearance were likely not due to metabolic
liabilities.
With the R1 substituent considered optimized, efforts shifted to
the R3 moiety to improve the pharmacokinetic properties of this
series, as this group entered into the solvent-exposed region. It

A. G. Steinig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4381–4387
4385
cMET in the presence of mouse plasma.21 The biochemical potency
against two cMET Y1230 mutants was reduced by 3–4ꢁ compared
to wild-type cMET.22 To establish the biochemical selectivity pro-
file, 6 was tested against a panel of 96 kinases.23 Besides cMET,
Aurora, Arg, and Abl were the only kinases that showed >50% inhi-
bition at 1 lM of 6; RON was not present in this screening panel.
The potent biochemical Aurora B activity in that assay did not
translate to the cellular setting with a mechanistic IC50 in the pres-
Table 6
Rodent PK of 6 (OSI-296)a
6 (OSI-296)
Mouse
Rat
iv Dosing
Dose (mg/kg)
Cl (mL/min/kg)
Vss (L/kg)
2
51
3.8
2.9
2
39
6.8
2.0
t
(h)
½
po Dosing
Dose (mg/kg)
Salt form
ence of mouse plasma of >10
impact on polyploidy in a FACS assay up to 10
l
M; in addition, 6 had no significant
M.24 It showed
20
20
l
Mesylate saltb
Free baseb
2.5
minimal activity against KDR in biochemical and cellular assays,
thus achieving differentiation from cMET/KDR inhibitors such as
XL880 and XL184.8 Compound 6 was less potent against ALK than
against cMET and RON and showed no significant activity in the
ALK-driven Karpas-299 xenograft model at doses of up to
200 mg/kg qd.25 While 6 inhibited proliferation of cMET-driven cell
lines such as MKN45 at concentrations near the respective cell
mechanistic IC50, no antiproliferative activity was observed in
non-driven lines, for example, HCT-116 and MDA-MB-231 (IC50
Cmax
(
lM)
4.8
5425
83
AUC0–last (ng h/mL)
%F
6204
72
a
PK parameters calculated by non-compartmental modeling using the median
concentration at each timepoint (three animals per timepoint).
b
Both formulated as solution in 40% HPCD in 0.01 M HCl.
Table 7
Selectivity data for 6 (OSI-296)a
>2 lM). This indicates that the antiproliferative effects of 6 are
Kinase
IC50
(
lM)
Kinase
IC50 (lM)
based on inhibition of the intended targets and not general cyto-
toxicity. Thus, OSI-296 (6) is best described as a selective cMET/
RON inhibitor.
The favorable pharmacokinetic profile displayed by 6 in the
mouse, with good exposure and oral bioavailability, led to its eval-
uation in pharmacodynamic and efficacy studies.
Pharmacodynamic (PD) studies were performed in the MKN45
xenograft model, evaluating effects on phosphorylation of the
cMET receptor in relation to plasma drug levels.
Following a 100 mg/kg single oral dose of 6, a >80% sustained
inhibition of MET auto-phosphorylation (pY1234/1235) was ob-
cMET biochem.
cMET Y1230C
cMET Y1230H
cMET cell mech.
cMET mech. + mpb
RON biochem.
sfRON cell mech.
0.026
0.072
0.11
0.042
0.60
ALK biochem.
0.040
0.50
1.3
6.1
0.011
1.0
ALK cell mech.
KDR biochem.
KDR cell mech.
AurB biochem.
AurB cell mech.
AurB mech. + mpb
0.025
0.20
>10
a
Cell mechanistic assays monitored the autophosphorylation sites of cMET
(MKN45), RON (HeLa-sfRON), ALK (Karpas-299), and KDR (HUVEC) or p-Ser10 of
histone H3 for Aurora B (HT-29). Biochemical IC50 values were determined using
the Invitrogen Omnia format at an ATP concentration equal to Km
.
b
Cell mechanistic assay in the presence of 50% mouse plasma.
served up to 8 h, corresponding to plasma levels of >3 lM. This cor-
related with the plasma concentration being above the cMET
mech. + mouse plasma IC50 value for the entire 8 h time period.
Drug levels dropped to 0.52 lM by 16 h, with slow recovery of
pMET content (45% inhibition at 24 h) (Fig. 3). Similar effects were
observed for pY1349 cMET (site essential for creating docking site
for downstream signaling; data not shown).
The in vivo efficacy of 6 dosed orally on a once-daily schedule
for 14 days was initially evaluated in the MKN45 xenograft model
in female nu/nu CD-1 mice (Fig. 4). The compound exhibited a clear
dose response, showing an ED50 of 50 mg/kg and reaching 100% tu-
mor growth inhibition (TGI) with regressions (27%) at 100 mg/kg
qd. Analysis of cMET pY1234/1235 phosphorylation status in tu-
mor samples collected at the end of dosing indicated that >90%
inhibition of pMET for 24 h correlated with complete inhibition
from the amide NH with the hinge backbone carbonyl of Met1160.
However, sub-micromolar cellular potencies could not be achieved
with these compounds.
From among these proximal rings, the tetrahydropyridine was
selected for further optimization (compounds 5x–5ab, 6). Gratify-
ingly, cellular potency was retained upon acetylation (5x), and the
PAMPA assay indicated good permeability (Pe at pH
7.4 = 199 ꢁ 10ꢂ6 cm/s). While the carbamate 5v and dimethylurea
5z had reduced potencies, the methyl urea 5aa and primary urea 6
showed comparable or better cell potencies than (R)-5a. Based on
the structural information, this may be explained by a hydrogen
bond between the urea NH and the backbone carbonyl of
Lys1161 at the tip of the hinge region. The sulfamide analog 5ab,
in contrast, was less potent.
Compound 6 (OSI-296) met the in vitro criteria of cellular cMET
potencies equal to or better than (R)-5a, significant permeability in
the PAMPA assay (Pe at pH 7.4 = 359 ꢁ 10ꢂ6 cm/s), and metabolic
stability as measured by extraction ratios (ERs) of 0.7 in both
mouse and human liver microsomes and was progressed to sin-
gle-dose mouse PK. It showed a significantly improved PK profile
Plasma Drug Conc
% pMET Content
10
8
120
100
80
60
40
20
0
6
4
compared to (R)-5a, achieving a Cmax of 2.4 lM with moderate
2
clearance and good oral exposure when dosed as an HCl salt in
PEG400/water. With an optimized formulation suitable for repeat
dosing in efficacy and toxicology studies [methanesulfonate (mes-
0
0
6
12
Time (h )
18
24
ylate) salt of 6 in 40% w/v HPCD (TrappsolÒ) in 0.01 M HCl], Cmax
,
oral exposure, and bioavailability were further improved (Table 6).
Good PK was also observed in female Sprague–Dawley rats. The
Figure 3. Correlation of inhibition of cMET phosphorylation and plasma drug
exposure in MKN45 tumor xenografts following oral dosing with 6 (OSI-296). The
blue line corresponds to median plasma concentrations of 6 at indicated time
following a 100 mg/kg qd dose. The orange line represents median pMET (pY1234/
1235) content normalized with total cMET in MKN45 tumors and expressed as a
percentage of control pMET content from vehicle-treated animals. The dotted blue
line indicates the cMET cell mechanistic IC50 in the presence of mouse plasma
high solubiliy of 6 at low pH (pH 3, >100
lM; pH 4.5, 11 lM) de-
creased to approx. 1 M around physiological pH.
l
Biochemical and cellular IC50 values are summarized in Table 7.
Compound 6 showed potent inhibition of cMET and RON in bio-
chemical and cellular assays with submicromolar activity against
(0.60 lM).

4386
A. G. Steinig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4381–4387
once-daily translated to 100% TGI. The compound was negative in
700
600
500
400
300
200
100
0
the AMES test. Its overall activity profile as cMET/RON inhibitor
differentiates it from cMET/KDR inhibitors, such as XL880 and
XL184, and from the ALK/cMET inhibitor Crizotinib (2).26 Taken to-
gether, these data supported further evaluation of OSI-296 in RON-
driven models, which will be detailed in subsequent publications.
Control vehicle
6, 25 mg/kg (QD1-14)
6, 50 mg/kg (QD1-14)
6, 100 mg/kg (QD1-14)
6, 200 mg/kg (QD1-14)
Acknowledgments
We would like to thank Ms. Viorica M. Lazarescu and Dr. Yingjie
Li for analytical support and Mr. Paul Maresca, Mr. Peter Meyn, Ms.
Brianna Tokar, Dr. Earl May, and Dr. Brenda Petronella for conduct-
ing in vitro ADMET studies and kinase selectivity profiling.
1
6
11
Days
16
Supplementary data
Supplementary data (detailed procedures of MKN45 prolifera-
tion, cMET and sfRON cell mechanistic, and cMET biochemical as-
says; descriptions of metabolic stability, CYP3A4, and PAMPA
assays; SAR at the 2-position of the furo[3,2-c]pyridine core;
waterfall plot of biochemical kinase selectivity data; analytical
data for compound 6 as free base and mesylate) associated with
this article can be found, in the online version, at http://
dx.doi.org/10.1016/j.bmcl.2013.05.074.
Figure 4. Tumor growth inhibition (TGI) data for 6 (OSI-296) in MKN45 tumor
xenografts. The compound was dosed orally as mesylate salt in 40% HPCD in 0.01 M
HCl.
Table 8
In vivo efficacy data for 6 (OSI-296)
Tumor line Dose and schedule
Avg.% TGI Max.% regression % BWL
MKN45a
25 mg/kg qd (1–14)
50 mg/kg qd (1–14)
100 mg/kg qd (1–14)
200 mg/kg qd (1–14)
13
51
100
100
0
0
27
49
6
2
3
3
References and notes
1. (a) Cui, J.; Tran-Dubé, M.; Shen, H.; Nambu, M.; Kung, P.-P.; Pairish, M.; Lei, J.;
Meng, J.; Funk, L.; Botrous, I.; McTigue, M. A.; Grodsky, N.; Ryan, K.; Padrique,
E.; Alton, G.; Timofeevski, S. L.; Yamazaki, S.; Li, Q.; Zou, H. Y.; Christensen, J. G.;
Mroczkowski, B.; Bender, S.; Kania, R. S.; Edwards, M. P. J. Med. Chem. 2011, 54,
6342; (b) Rodig, S. J.; Shapiro, G. I. Curr. Opin. Investig. Drugs 2010, 11, 1477.
2. Barlaam, B. C.; Bower, J. F.; Delouvrie, B.; Fairley, G.; Harris, C. S.; Lambert, C.;
Ouvry, G.; Winter, J. J. G. WO 2009/053737; Chem. Abstr. 2009, 150, 472757.
3. Hilton, S.; Naud, S.; Caldwell, J. J.; Boxall, K.; Burns, S.; Anderson, V. E.; Antoni,
L.; Allen, C. E.; Pearl, L. H.; Oliver, A. W.; Aherne, G. W.; Garrett, M. D.; Collins, I.
Bioorg. Med. Chem. 2010, 18, 707.
4. Ghose, A. K.; Herbertz, T.; Pippin, D. A.; Salvino, J. M.; Mallamo, J. P. J. Med.
Chem. 2008, 51, 5149.
5. This binding mode has been confirmed for 2 by crystallography of a complex
with cMET; see: Timofeevski, S. L.; McTigue, M. A.; Ryan, K.; Cui, J.; Zou, H. Y.;
Zhu, J. X.; Chau, F.; Alton, G.; Karlicek, S.; Christensen, J. G.; Murray, B. W.
Biochemistry 2009, 48, 5339.
SNU-5b
50 mg/kg qd (1–20)
100 mg/kg qd (1–20)
200 mg/kg qd (1–20)
100
100
100
31
64
80
4
2
4
U87MGb
50 mg/kg qd (1–14)
100 mg/kg qd (1–14)
200 mg/kg qd (1–14)
50 mg/kg bid (1–14)
100 mg/kg bid (1–14)
58
79
94
79
97
0
0
8
0
22
0
0
0
5
0
a
Dosed as solution of mesylate salt in 40% HPCD in 0.01 M HCl.
Dosed as solution of free base in 60:40 PEG400/25 mM tartaric acid.
b
of tumor growth (100% TGI). Significant tumor growth inhibition
was also observed in other cMET-driven models such as SNU-5
(gastric carcinoma) and U87MG (glioblastoma). The study in the
U87MG model also compared dosing on a once-daily versus
twice-daily schedule. The data suggest that for this model there
is no advantage in administration of 6 as two divided doses com-
pared to a cumulative dose on a once-daily schedule. In all models,
the compound was very well tolerated with minimal body weight
loss (BWL; 0–6%), even at the highest doses of 200 mg/kg qd and
100 mg/kg bid (Table 8).
6. Maggiora, P.; Gambarotta, G.; Olivero, M.; Giordano, S.; Di Renzo, M. F.;
Comoglio, P. M. J. Cell Physiol. 1997, 173, 183.
7. Comoglio, P. M.; Giordano, S.; Trusolino, L. Nat. Rev. Drug Disc. 2008, 7, 504.
8. Eder, J. P.; Vande Woude, G. F.; Boerner, S. A.; LoRusso, P. M. Clin. Cancer Res.
2009, 15, 2207.
9. Wang, M. H.; Wang, D.; Chen, Y. Q. Carcinogenesis 2003, 24, 1291.
10. Angeloni, D.; Danilkovitch-Miagkova, A.; Ivanova, T.; Braga, E.; Zabarovsky, E.;
Lerman, M. I. Oncogene 2007, 26, 4499.
11. Smolen, G. A.; Sordella, R.; Muir, B.; Mohapatra, G.; Barmettler, A.; Archibald,
H.; Kim, W. J.; Okimoto, R. A.; Bell, D. W.; Sgroi, D. C.; Christensen, J. G.;
Settleman, J.; Haber, D. A. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 2316.
12. MKN45 proliferation assay: 3-day incubation, Cell Titer-Glo used for readout.
For details, please see the Supplementary data.
The potential for mutagenicity of 6 was evaluated in an AMES
test using the TA98 and TA100 strains; it was negative for the
induction of reverse mutations with and without an S9 metabolic
activation system.
13. cMET cell mechanistic assay: MKN45 cells, phosphorylated MET levels
measured by incubating with
a rabbit polyclonal antibody against MET
[pYpYpY1230/1234/1235] followed by an anti-rabbit antibody conjugated to
HRP. For details, please see the Supplementary data.
14. Under the assay conditions used herein, the IC50 values between the MKN45
proliferation and cMET cell mechanistic assays correlated nearly 1:1.
15. sfRON cell mechanistic assay: HeLa cells transfected with sfRON-pcDNA,
phosphorylated RON levels measured by incubating with a goat polyclonal
antibody against phosphorylated RON ([pYpY1238/1239]) followed by an anti-
goat antibody conjugated to HRP. For details, please see the Supplementary
data.
In summary, we designed and synthesized 6-aminofuro[3,2-
c]pyridines 5 and demonstrated their usefulness as kinase inhibi-
tors by optimizing this series towards inhibition of cMET and
RON, leading to OSI-296 (6). This compound exhibited selectivity
in a panel of 96 kinases with potent activity against cMET, includ-
ing common Y1230 mutants, and RON. OSI-296 showed a PK pro-
file in rodents suitable for oral dosing with >70% bioavailability. In
multiple cMET-driven xenograft models in mice, significant tumor
growth inhibition was observed at once-daily doses of 50 mg/kg or
less, with regression at higher doses. OSI-296 was very well toler-
ated with little body weight loss and no adverse effect even at the
highest tested dose of 200 mg/kg po once-daily. A PK/PD/TGI corre-
lation has been established wherein >90% inhibition of cMET phos-
phorylation sustained over 24 h by OSI-296 dosed at 100 mg/kg po
16. Eloy, F.; Deryckere, A. Helv. Chim. Acta 1970, 53, 645.
17. Engler, T. A.; Malhotra, S.; Burkholder, T. P.; Henry, J. R.; Mendel, D.; Porter, W. J.;
Furness, K.; Diefenbacher, C.; Marquart, A.; Reel, J. K.; Li, Y.; Clayton, J.;
Cunningham, B.; McLean, J.; O’Toole, J. C.; Brozinick, J.; Hawkins, E.; Misener, E.;
Briere, D.; Brier, R. A.; Wagner, J. R.; Campbell, R. M.; Anderson, B. D.; Vaughn, R.;
Bennett, D. B.; Meier, T. I.; Cook, J. A. Bioorg. Med. Chem. Lett. 2005, 15, 899.
18. cMET biochemical assay: AlphaScreen in 384-well plates using biotinylated
poly(Glu, Tyr) as substrate. For details, please see the Supplementary data.
19. We also briefly explored substitution at the 2-position of the furo[3,2-
c]pyridine core. However, none of the synthesized compounds improved
cMET potency compared to (R)-5a (please see Supplementary data, Table S1,
for details).

A. G. Steinig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4381–4387
4387
20. Crystallographic information: resolution 2.40 Å, space group P21, four
molecules in the asymmetric unit. Crystallization and solution of structure
performed by Proteros Biostructures GmbH, Am Klopferspitz 19, D-82152,
Martinsried, Germany, www.proteros.com. Crystallographic data for the
structure have been deposited with the PDB (pdb code 4KNB).
21. In addition to the ‘shift assays’, the plasma protein binding of 6 was also
determined by equilibrium dialysis (human plasma, 98.1% bound; mouse
plasma, 98.2% bound).
23. The mobility shift assay technology from Caliper LifeSciences (ProfilerPro™
Kinase Selectivity Assay Kits 1–4) was used.
24. Polyploidy is a phenotypic readout for inhibition of Aurora B, see for example:
Ditchfield, C.; Johnson, V. L.; Tighe, A.; Ellston, R.; Haworth, C.; Johnson, T.;
Mortlock, A.; Keen, N.; Taylor, S. S. J. Cell. Biol. 2003, 161, 267.
25. In contrast, Crizotinib (2) was reported to give 96% TGI at 50 mg/kg qd in this
model, see: Christensen, J. G.; Zou, H. Y.; Arango, M. E.; Li, Q.; Lee, J. H.;
McDonnell, S. R.; Yamazaki, S.; Alton, G. R.; Mroczkowski, B.; Los, G. Mol. Cancer
Ther. 2007, 6, 3314.
22. The cMET inhibitors PF-04217903 and PF-02341066 (2) showed a biochemical
potency loss against these two mutants of P600-fold and 8- to 9-fold,
respectively; see Ref.5
26. Dussault, I.; Bellon, S. F. Anticancer Agents Med. Chem. 2009, 9, 221.