4-Quinazolinyloxy-diaryl ureas as novel BRAFV600E inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.07.019

Aryl phenyl ureas with a 4-quinazolinoxy substituent at the meta-position of the phenyl ring are potent inhibitors of mutant and wild type BRAF kinase. Compound 7 (1-(5-tert-butylisoxazol-3-yl)-3-(3-(6,7-dimethoxyquinazolin-4-yloxy) phenyl)urea hydrochloride) exhibits good pharmacokinetic properties in rat and mouse and is efficacious in a mouse tumor xenograft model following oral dosing.

Full text of DOI:10.1016/j.bmcl.2011.07.019

Bioorganic & Medicinal Chemistry Letters 21 (2011) 5342–5346
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
4-Quinazolinyloxy-diaryl ureas as novel BRAFV600E inhibitors
Mark W. Holladay a, , Brian T. Campbell a, Martin W. Rowbottom a, Qi Chao a, Kelly G. Sprankle a,
⇑
Andiliy G. Lai a, Sunny Abraham a, Eduardo Setti b, Raffaella Faraoni a, Lan Tran a, Robert C. Armstrong a,
Ruwanthi N. Gunawardane a, Michael F. Gardner a, Merryl D. Cramer a, Dana Gitnick a, Mark A. Ator b,
Bruce D. Dorsey b, Bruce R. Ruggeri b, Michael Williams b, Shripad S. Bhagwat a, Joyce James a
a Ambit Biosciences Corporation, 4215 Sorrento Valley Boulevard, San Diego, CA 92121, USA
b Cephalon, Inc., 145 Brandywine Parkway, West Chester, PA 19380, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Aryl phenyl ureas with a 4-quinazolinoxy substituent at the meta-position of the phenyl ring are potent
inhibitors of mutant and wild type BRAF kinase. Compound 7 (1-(5-tert-butylisoxazol-3-yl)-3-(3-(6,7-
dimethoxyquinazolin-4-yloxy)phenyl)urea hydrochloride) exhibits good pharmacokinetic properties in
rat and mouse and is efficacious in a mouse tumor xenograft model following oral dosing.
Ó 2011 Elsevier Ltd. All rights reserved.
Received 28 April 2011
Revised 30 June 2011
Accepted 6 July 2011
Available online 14 July 2011
Keywords:
BRAF kinase
Xenograft model
The RAS-RAF-MEK-ERK pathway plays a central role in transduc-
tion of mitogen and growth factor-initiated signals from the cell sur-
face to the nucleus, thereby modulating key cellular functions
including cell growth, proliferation, motility, and survival.1 The
RAF proteins (ARAF, BRAF, and RAF1) are serine/threonine kinases
that catalyze phosphorylation of MEK, a downstream serine/threo-
nine kinase in this pathway. Point mutation of the BRAF Val-600 res-
idue toGlu (BRAFV600E) resultsin constitutive activation ofBRAF, and
hence constitutive activation of the RAS-RAF-MEK-ERK pathway.2
Mutated BRAF is found in 55–68% of melanomas, 29–83% of papil-
lary thyroid tumors, and 4–16% of colorectal cancers, among others,
and is a key factor in the uncontrolled proliferation of these tu-
mors.3–5 Drug discovery efforts directed toward inhibitors of BRAF
and related family members have resulted in the identification of
sorafenib (Nexavar™, 1, Fig. 1), a marketed agent for renal cell and
hepatocellular carcinomas. Sorafenib exhibits potent pan-VEGF-R
activity, and it is hypothesized that this mechanism of action sub-
stantially contributes to the observed clinical effects. Other agents
targeting BRAFV600E including RAF-265 and PLX-4032 are undergo-
ing clinical evaluation,1,6 with significant clinical benefit having
been recently reported for PLX-4032.7
We have described previously a series of amide-substituted
5-t-butyl-3-isoxazolyl phenyl ureas 2 (Fig. 1) as potent FLT3 inhib-
itors.8 Through application of Ambit’s KINOMEscan™ platform,9
activity against BRAFV600E was recognized in some analogs of the
series represented by 2. Further elaboration around 2 included
Figure 1. Structure of sorafenib (1) and evolution of the 4-quinazolinyloxy-diaryl
urea series of BRAF inhibitors.
incorporation of the amide carbonyl oxygen (as nitrogen) into a
ring fused to the adjacent aryl moiety (cf. structure 3), leading
eventually to quinazolines 4. Certain meta-substituted variants of
2 and 4 were observed to show superior activity against BRAF,
⇑
Corresponding author. Tel.: +1 858 334 2141.
E-mail address: mholladay@ambitbio.com (M.W. Holladay).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.07.019

M. W. Holladay et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5342–5346
5343
Table 1
Table 3
Effect of linker X on properties of BRAF inhibitors
Effect of central phenyl ring substitutions on properties of BRAF inhibitors
R1
N
N
4
4
6
O
N
N
O
X
O
O
N
N
3
N
H
N
H
N
H
N
O
2
OMe
OMe
H
OMe
OMe
Cmpd
X
BRAFV600E Kd (nM)a
A375 pMEK IC50 (nM)a
S10
%Fb
Cmpd R1
BRAFV600E Kd (nM)a A375 pMEK IC50 (nM)a S10
%Fb
1
6
7
8
9
—
284
6599
0.16
0.10
0.29
0.18
0.30
—
17
18
19
20
21
2-Me
4-F
4-Cl
6-F
4,6-di-
F
38 (2)
56 (2)
99 (2)
72 (2)
46 (2)
179 (2)
113 (2)
229 (2)
62 (4)
0.30
0.38
0.24
—
—
—
3-NH
3-O
3-S
40 (4)
73 (20)
61 (4)
478 (4)
82 (5)
121 (20)
118 (6)
—
1.9
49
24
<1
0.31 22
4-NH
113 (2)
0.37 14c
a
Average value, with number of determinations shown in parentheses.
Dosed as hydrochloride salt in rats unless otherwise indicated.
b
a,b
Same as Table 1.
Dosed as free base.
c
leading to the series represented by 5. Herein we describe the SAR
around structure with respect to binding affinity (Kd) to
BRAFV600E and potency (IC50 to inhibit BRAFV600E-dependent
MEK phosphorylation in A375 cells. A previous letter from these
laboratories described a selectivity score (S) as a preliminary indi-
cator of specificity across the human kinome.9 In the present letter,
an S10 score represents the fraction of 290 kinases tested that show
5
inferior PK properties for meta-NH-linked compounds persisted
across numerous analogs that were prepared with all three (NH-,
O-, and S-) linkers (data not shown). The remainder of this letter
will focus on analogs having a meta-O-linker between the quinaz-
oline moiety and the central phenyl ring, with compound 7 as the
prototype.
)
Table 2 shows the properties of compounds with Ar1 = phenyl
or substituted phenyl. The unsubstituted phenyl analog 10 was
markedly less potent than 7 in both binding and cell assays,
whereas a number of phenyl substitution patterns afforded analogs
with binding affinities comparable to that of 7. On the other hand,
there was a trend toward modest to significantly reduced cell po-
tency across this series of analogs. For example, compound 16
(which has the same phenyl substitution pattern as sorafenib 1)
showed similar binding affinity to that of 7 but nearly six-fold low-
er cell potency. These analogs also generally exhibited higher ki-
nome specificity than that of 7.
The effects of conservative substitutions on the central phenyl
ring of compound 7 are shown in Table 3. In general, these modi-
fications were well tolerated, but did not appear to afford any par-
ticular advantage to the profiles of these compounds.
Tables 4 and 5 show, respectively, the effects of replacing the
methoxy group on either the 6- or 7-position of the quinazoline
ring with hydroxy, ethoxy, methoxyethoxy, or an aminoalkoxy
group. In general, these modifications were well tolerated with re-
spect to binding affinity, consistent with a hypothesis that this re-
gion of the molecule extends outside the catalytic cleft toward
solvent. These changes also are consistent with potent cell activity,
<10% activity (>90% inhibition) compared to control at a 10 lM
screening concentration. Pharmacokinetic assessments played a
significant role in guiding the SAR in this series, and therefore oral
bioavailability in rat is reported for selected analogs. Finally, effi-
cacy in a mouse tumor xenograft model is disclosed for three lead-
ing analogs identified from these studies.
Table 1 shows the effect of various linkers X on BRAFV600E inhi-
bition and other key properties. Among compounds linked to qui-
nazoline through the meta-position of the central phenylene ring,
NH-, O-, and S-linked compounds (6–8) all showed potent binding
affinity (40–73 nM) and cell activity (82–121 nM). The S10 scores of
these analogs followed the order NH<S<O, that is, the meta-O ana-
log showed the least kinome specificity among this set. Consistent
with the SAR described above, compound 9, having the para-NH-
linker, showed markedly lower binding affinity for BRAFV600E than
the corresponding meta-NH linked compound 6. Compound 9 also
showed substantially lower kinome specificity (higher S10 score)
than 6.
Although the binding affinity, cell potency, and kinome specific-
ity of compound 6 were attractive, this analog showed inferior
pharmacokinetic (PK) properties in rats compared to the corre-
sponding meta-O- and meta-S-linked compounds. This pattern of
Table 2
Effect of Ar1 moiety on properties of BRAF inhibitors
O
N
N
Ar1
N
H
N
H
O
OMe
OMe
Cmpd
Ar1
BRAFV600E Kd (nM)a
A375 pMEK IC50 (nM)a
S10
%F
10
11
12
13
14
15
16
Ph
370 (2)
46 (2)
36 (2)
90 (2)
30 (2)
96 (2)
98 (2)
>15000 (2)
1664 (2)
333 (2)
ca. 10000 (2)
1850 (2)
290 (2)
0.09
0.18
0.18
0.10
0.15
0.17
0.16
—
—
—
—
3-CF3-Ph
3-t-Bu-Ph
3-MeO-Ph
4-Cl-Ph
4-t-Bu-Ph
3-CF3-4-Cl-Ph
—
15b
—
696 (2)
a
Same as Table 1.
Dosed as free base.
b

5344
M. W. Holladay et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5342–5346
Table 4
Effect of modifications at the 6-position of the quinazoline ring on properties of BRAF inhibitors
O
N
N
O
O
N
N
H
N
H
O
OMe
R2
6
Cmpd
R2
BRAFV600E Kd (nM)a
A375 pMEK IC50 (nM)a
S10
%Fb
22
23
24
25
26
27
H
Et
46 (2)
67 (4)
48 (2)
38 (2)
29 (2)
96 (2)
86 (4)
80 (4)
59 (2)
52 (4)
36 (4)
293 (2)
0.24
0.28
0.26
0.28
0.30
0.25
—
30 (2)
13
-(CH2)2OMe
-(CH2)2-4-morpholinyl
-(CH2)3-4-morpholinyl
-(CH2)2-1-piperidinyl
11c
20
—
a,b
Same as Table 1.
Dosed as free base.
c
Table 5
Effect of modifications at the 7-position of the quinazoline ring on properties of BRAF inhibitors
O
N
N
O
N
N
H
N
H
O
7
OR3
OMe
Cmpd
R3
BRAFV600E Kd (nM)a
A375 pMEK IC50 (nM)a
S10
%Fb
28
29
30
31
32
H
Et
57 (2)
82 (2)
57 (8)
34 (2)
56 (2)
110 (4)
173 (4)
100 (12)
165 (8)
96 (4)
0.20
0.29
0.31
0.27
0.35
<1c
16
-(CH2)2OMe
-(CH2)2-4-morpholinyl
-(CH2)3-4-morpholinyl
56
52c
18c
a,b
Same as Table 1.
Dosed as free base.
c
with a number of variants exhibiting cellular potency in the
mid-double digit nanomolar range. In general, these modifications
did not result in meaningful alterations in the kinome specificities
of these molecules. In many cases, bioavailabilities were less favor-
able after incorporation of solubilizing groups, although com-
pounds 30 and 31 showed bioavailabilities comparable to that of
compound 7.
When plasma clearance was determined from iv studies, values
for most of the analogs in Tables 4 and 5 bearing solubilizing
groups were higher than that of 7, consistent with the likelihood
of a significant first pass clearance, although compounds 30 and
31 showed lower iv clearance values than that of 7.
Compounds 7, 8, and 31 were further evaluated in pharmacoki-
netic and antitumor efficacy studies. Good plasma exposure was
observed for all three compounds upon oral administration to mice
[cmpd; dose (mg/kg); plasma AUC (
lM Â h)]: [7; 30; 12.5]; [8; 30;
22]; [31; 100; 48]. Oral administration of 7 to mice bearing Colo-
205 tumors at 100 mg/kg bid resulted in sustained tumor stasis
throughout the dosing period (Fig. 2), which included a 55% inci-
dence of tumor regressions. Compound 8 at 55 mg/kg or 100 mg/
kg showed tumor stasis early in the dosing period, accompanied
by significant tumor regrowth before the end of the dosing period
(see Supplementary data). Interestingly, by day 8, efficacy was
notably inferior at the higher dose of compound 8. Compound 31
at 100 mg/kg bid exhibited a maximum 65% tumor growth inhibi-
tion, also accompanied by tumor regrowth before the end of the
dosing period (see Supplementary data). No body weight loss
Figure 2. Antitumor efficacy of orally administered compound 7 in the mouse Colo-
205 tumor xenograft model. Statistical significance relative to vehicle controls:
30 mg/kg bid group: NS; 100 mg/kg bid group: days 5–8, p < 0.0004; days 10–12
(termination of dosing), p < 0.0002.
was observed for 8 or 31, whereas 7 showed 8% maximum body
weight loss that was reversible following cessation of compound
administration. On day 5 of dosing, all three compounds at
100 mg/kg bid showed >50% reduction in levels of phosphorylated

M. W. Holladay et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5342–5346
5345
Scheme 1. Reagents and conditions: (a) (R = NH2) Cs2CO3, THF, 50 °C; (b) (R = NO2)
(i) KOtBu, DMA, 150 °C; (ii) H2, Pd-C, MeOH; (c) ArNCO, THF or DMF, rt – 80 °C; (d)
Ar-NHCO2Ph, THF, NEtiPr2, DMAP, 50–80 °C.
Scheme 3. Reagents and conditions: (a) alkyl halide, KI, K2CO3, DMF, 90–95 °C; (b)
Na2S2O4, MeOH, H2O, 55 °C; (c) HCONH2, HOAc, 130 °C; (d) POCl3, 110–115 °C; (e)
3-aminophenol, Cs2CO3, THF, 75 °C, 25 h; (f) H2, Pd-C, EtOH/THF, 50–55 °C; (g) Ar-
NHCO2Ph, DMF, 60 °C.
Scheme 2. Reagents and conditions: (a) ArNCO, THF or PhMe or DMF, optional
NEt3, 50–70 °C; (b) ArNHCO2Ph, CH3CN, DBU, 50 °C; (c) (X = 3-O) KOtBu or Cs2CO3,
optional K2CO3 or NEt3, THF or DMF, rt – 80 °C; (d) (X = 3-S) NaH, THF, rt – 50 °C; (e)
(X = 4-NH) n-BuOH, reflux.
MEK (pMEK) in tumor tissue at 10 h post-dose (see Supplementary
data). Given comparable cellular potencies among 7, 8, and 31, to-
gether with similar or better single dose exposure and pMEK sup-
pression for 8 and 31 as compared with 7, the observation of
overall inferior antitumor activity observed for 8 and 31 merits
comment. We note that the greatest differences in antitumor
behavior among the compounds first become apparent about
8 days after the start of dosing, which is several days after the
above pharmacokinetic and pharmacodynamic readings were ta-
ken. We speculate that differential changes in metabolism or dis-
tribution of the compounds, for example as
a result of
differential induction of metabolizing enzymes or transporter mol-
ecules, may have become manifest by this later time. The reverse
dose-response relationship observed for 8 in vivo might further-
more be a consequence of a propensity for this compound to in-
duce such effects. We do not believe that potential differences in
kinome specificity among the compounds offers a likely explana-
tion for the disparate in vivo behaviors observed, since review of
the kinase inhibition profiles indicated that any additional kinases
significantly inhibited by 7 were similarly inhibited by at least one
of, and usually both of, 8 and 31.
Scheme 4. Reagents and conditions: (a) Br(CH2)nCH2Cl, K2CO3, DMF, rt – 60 °C; (b)
HNO3, Ac2O, HOAc, 50 °C; (c) H2, Pd-C, EtOAc; (d) formamidineÁHCl, EtOH, sealed
tube 130 °C; (e) POCl3, NEt3, 100 °C; (f) Ar-NH(CO)NH-C6H4-m-OH, KOtBu or
Cs2CO3, optional K2CO3 or NEt3, THF or DMF, rt – 80 °C; (g) 2° amine, n-Bu4NI,
NEtiPr2, DMF, 60 °C, 2–3 d.
Schemes 1–4 illustrate representative synthetic routes to key
compounds described in this letter. Compounds 28–32 (quinazo-
line 7-position modifications, Table 5) were prepared using meth-
Acknowledgments
We are grateful to the Ambit KINOMEscan™ team for generat-
ing BRAFV600E Kd values and S10 scores and to Ianina Valenta for
her contributions to generating data from cellular assays.
ods analogous to those described in Schemes
3 and 4 for
quinazoline 6-position modifications (Table 4) using the appropri-
ate regioisomeric starting materials.
In summary, quinazolinoxy-diaryl ureas and structurally re-
lated analogs were synthesized and evaluated as inhibitors of
BRAF, and analogs with favorable potency, kinome specificity,
and pharmacokinetic profiles were identified. In a mouse xenograft
model, compound 7 exhibited sustained antitumor efficacy (stasis
with regressions) accompanied by a pharmacodynamic readout for
BRAF inhibition at a dose that induced minimal body weight loss.
Further optimization of this series will be reported in due course.
Supplementary data
Supplementary data (synthetic details for preparation of com-
pounds 7, 8, and 31, a table summarizing antitumor and pharma-
codynamic data from mouse tumor xenograft studies for
compounds 7, 8, and 31, and the antitumor efficacy results in
graphical form for compounds 8 and 31 are included) associated

5346
M. W. Holladay et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5342–5346
6. Hersey, P.; Bastholt, L.; Chiarion-Sileni, V.; Cinat, G.; Dummer, R.; Eggermont, A.
M. M.; Espinosa, E.; Hauschild, A.; Quirt, I.; Robert, C.; Schadendorf, D. Ann.
Oncol. 2009, 20, vi35.
with this article can be found, in the online version, at doi:10.1016/
j.bmcl.2011.07.019.
7. Garber, K. Science 2009, 326, 619.
8. Patel, H. K.; Grotzfeld, R. M.; Lai, A. G.; Mehta, S. A.; Milanov, Z. V.; Chao, Q.;
Sprankle, K. G.; Carter, T. A.; Velasco, A. M.; Fabian, M. A.; James, J.; Treiber, D. K.;
Lockhart, D. J.; Zarrinkar, P. P.; Bhagwat, S. S. Bioorg. Med. Chem. Lett. 2009, 5182.
9. Karaman, M. W.; Herrgard, S.; Treiber, D. K.; Gallant, P.; Atteridge, C. E.;
Campbell, B. T.; Chan, K. W.; Ciceri, P.; Davis, M. I.; Edeen, P. T.; Faraoni, R.; Floyd,
M.; Hunt, J. P.; Lockhart, D. J.; Milanov, Z. V.; Morrison, M. J.; Pallares, G.; K Patel,
H.; Pritchard, S.; Wodicka, L. M.; P Zarrinkar, P. Nat. Biotechnol. 2008, 26, 27.
References and notes
1. Roberts, P. J.; Der, C. J. Oncogene 2007, 26, 3291.
2. Michaloglou, C.; Vredeveld, L. C.; Mooi, W. J.; Peeper, D. S. Oncogene 2007, 27, 1.
3. Beeram, M.; Patnaik, A.; Rowinsky, E. K. J. Clin. Oncol. 2005, 23, 6771.
4. Haluska, F. G.; Tsao, H.; Wu, H.; Haluska, F. S.; Lazar, A.; Goel, V. Clin. Cancer Res.
2006, 12, 2301s.
5. Ouyang, B.; Knauf, J. A.; Smith, E. P.; Zhang, L.; Ramsey, T.; Yusuff, N.; Batt, D.;
Fagin, J. A. Clin. Cancer Res. 2006, 12, 1785.