Indazolylamino quinazolines and pyridopyrimidines as inhibitors of the EGFr and c-erbB-2

Article abstract of DOI:10.1016/S0960-894X(01)00219-0

Described herein is the design and synthesis of indazolylaminopyridopyrimidines and quinazolines as inhibitors of the class 1 tyrosine kinase receptor family. Data is presented for N⁴-(1-benzyl-1H-indazol-5-yl)-N⁶,N⁶- dime

Full text of DOI:10.1016/S0960-894X(01)00219-0

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1401–1405
Indazolylamino Quinazolines and Pyridopyrimidines
as Inhibitors of the EGFr and C-erbB-2
Stuart Cockerill,* Colin Stubberfield, Jeremy Stables, Malcolm Carter, Stephen Guntrip,
Kathryn Smith, Steve McKeown, Robert Shaw, Peter Topley, Lindy Thomsen,
Karen Affleck, Amanda Jowett, David Hayes, Malcolm Willson, Patrick Woollard and
David Spalding
Enzyme Chemistry 1, Oncology Unit, Enzyme Pharmacology, Research Biomet. Glaxo Wellcome
Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK
Received 4 December 2000; revised 16 February 2001; accepted 29 March 2001
Abstract—Described herein is the design and synthesis of indazolylaminopyridopyrimidines and quinazolines as inhibitors of the
class 1 tyrosine kinase receptor family. Data is presented for N4-(1-benzyl-1H-indazol-5-yl)-N6,N6-dimethylpyrido[3,4-d]pyrimidine-
4,6-diamine 3B. This compound inhibited EGFr and c-erbB-2 enzymes selectively over other kinases. It inhibited the proliferation
of a range of tumour cell lines in vitro and the growth of BT474 xenografts in SCID mice. # 2001 Elsevier Science Ltd. All rights
reserved.
One in 12 women will develop breast cancer at some-
time in their lifetime. In Europe and the US, it is one of
the three most prevalent forms of cancer, with over
200,000 cases occurring in 1994.1a Current chemotherapy
regimes involve the use of cytotoxics and antiestrogens,
both with inherent selectivity disadvantages. Increased
levels of c-erbB-2 and epidermal growth factor receptor
(EGFr) each occur in a significant range of cancers, for
example breast and non-small cell lung cancers. Their
overexpression correlates with shorter time to relapse,
resistance to hormone, chemotherapy and reduced
survival.1b
dependent proliferation would reduce the likelihood of
severe adverse effects and allow longer-term, effective
treatment. Herceptin (trastuzumab), Genentech’s
humanised monoclonal antibody, is approved in
patients with metastatic breast cancer overexpressing
the c-erbB-2 growth factor receptor.2 Small molecule
inhibitors of EGFr, Iressa and OSI-774, have now pro-
gressed through clinical trial without serious adverse
events.2b Inhibition of the tyrosine kinase activity of this
receptor would appear to represent a tractable target
for a small molecule inhibitor. The epidemiological
evidence implicating both EGFr and c-erbB-2 in a
diverse range of tumour types would suggest a clinical
benefit in inhibiting the kinase domains of both recep-
tors.1
c-erbB-2 and EGFr are members of the epidermal
growth factor receptor (EGFr) subfamily of protein
tyrosine kinases. Their overexpression in cells causes
transformation, such that the cells will grow in the
absence of growth factors and in anchorage-indepen-
dent conditions. A kinase-inactive receptor does not
have these effects. An inhibitor of this family of kinases
should block signalling from the receptor, subsequent
transformation and inappropriate growth. The specifi-
city of such a compound for c-erbB-2 and EGFr
*Corresponding author. Fax: +44-208-652-9620; e-mail: scockerill@
arrowt.co.uk
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00219-0

1402
S. Cockerill et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1401–1405
Optimisation of 4557W
one can speculate upon the potential H-bonding effect
of Lys53 in Figure 1 (the equivalent residue in c-erbB-2 is
Lys753) compensating for a subsequent conformational
alteration in the hydrophobic region of the binding
pocket.
Screening identified quinazoline 4557W as a potent
inhibitor of both c-erbB-2 and EGFr [c-erbB-2
0.079 mM, EGFr 0.020 mM (isolated enzyme), 2.0 mM in
HB4aC5.2 cells, 1.2 mM in BT474, both overexpressing
c-erbB-2; 2.5 mM in HN5 cells overexpressing EGFr].
This combined EGFr/c-erbB-2 potency was in direct
contrast to smaller, previously reported anilinoquinazo-
lines like CAQ3,4 and subsequent analogues which have
demonstrated selectivity for EGFr over c-erbB-2.6À8
A
binding hypothesis was constructed for this compound.
We utilised our knowledge of P38 binding (since repor-
ted5) based upon the principle of quinazoline binding
through N1 to Thr798, the residue equivalent to Met109
in P38 as shown in Figure 1.
Figure 1. A model of 4557W binding to a protein kinase (P38). The
complex between P38 and the quinazoline GI2616345 was used as a
starting point for modelling. The quinazoline 4-amino bond was rota-
ted ꢀ180ꢁ from the conformation seen in that structure in order to
accommodate the increased bulk of the benzyl group in the back of the
hydrophobic pocket. The complex was then energy minimised keeping
the protein and the quinazoline rings fixed. The hydrogen bond
between quinazoline N1 and the backbone NH of Met109 is shown as
a dotted magenta line. Nearby P38 side chains are shown and labelled
(P38 numbering) where they are conserved in c-erbB-2.
This binding hypothesis has the benzyloxyaniline
accommodated in the back of the hydrophobic pocket
with the 6,7-dimethoxy groups pointing towards the lip
of the ATP binding cleft. We decided to investigate the
effect of this proposed hydrophobic binding interaction
by imposing a conformational restriction into the ani-
line fragment. A bicyclic nucleus was chosen to replace
the aniline system and reduce the number of rotatable
bonds in this region from 3 to 2. Subsequently a range
of bicyclics A–G were combined with quinazoline and
equivalent ‘bottom halves’ 1, 2 and 3 (Fig. 2).
Figure 2. Top half aniline X and quinazoline equivalent Q combina-
tions. Data for selected examples are shown in Table 1.
Of particular interest from this evaluation were the
indazolyl systems exemplified by 1B, 2B and 3B. T he
enhanced cell potency of compound 3B (GW974, in the
c-erbB-2 overexpressing HB4a.c5.2 cell line) over other
analogues shown in Table 1, made it worthy of further
evaluation. This compound was taken forward into
pharmacokinetic and further in vitro evaluation.
Several key structural features became apparent in this
exercise. Greater intrinsic potency was observed for 6,7-
dimethoxyquinazolines and pyridopyrimidines, in line
with those previously reported for EGFr inhibition.6,7
In general, 1-substitution of the bicyclic anilino system
was very potent, for example in the case of indole and
indazole. However in the case of benzimidazole, 2-sub-
stitution was clearly the preferred regioisomer with
potency observed with all three bottom halves.
Although the precise reason for this remains unclear,
GW974 Evaluation In Vitro
GW974 was evaluated against a range of isolated kina-
ses (Table 2) and found to be selective for the EGFr
subfamily. This was a common trait of this class with
routine selectivity observed in the range 100–10,000-
fold.8 The compound also exhibited potency across a

S. Cockerill et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1401–1405
1403
range of cell types, including several (CaLu3, HN5 and
BT474) derived from clinical extracts. Selectivity was
also observed against a non-oncogenic MRC5 cell line.
Data are shown in Table 2.
of distribution and moderate clearances were character-
istic of all three species.
GW974 Evaluation In Vivo
GW974 was evaluated alongside close analogues (2B
and 3A) in the fast growing N87 gastric tumour xeno-
graft in SCID mice. At doses of 10 mg/kg po, bid a
50% inhibition of tumour growth was observed over the
20-day dosing period. Analogues 2B and 3A showed 20
and 0% inhibition, respectively, a result perhaps asso-
ciated with their poorer oral absorption profiles. In the
BT474 breast tumour xenograft model, GW974 gave
complete inhibition of tumour growth 10 mg/kg po bid
over the 20-day dosing period. A longer-term study with
initial dosing at 30 mg/kg po bid produced tumour
shrinkage. A maintenance dose of 10 mg/kg for 15 days
followed. Upon cessation of dosing (at day 80), no
tumour regrowth was observed (Fig. 3). The precise
explanation of this effect is uncertain but may be a
demonstration of the apoptosis inducing potential of
c-erbB-2 inhibition, reported for anti-HER2 mono-
clonal antibodies.10
GW974 Pharmacokinetic Evaluation
In pharmacokinetic evaluation, GW974 exhibited an
improved profile over its quinazoline counterparts. It
exhibited improved bioavailability (24%) over its dime-
thoxyquinazoline analogue 2B (7%) in initial studies in
the rat. Further PK data in the mouse and dog for this
compound are shown in Table 3. Comparable PK was
observed in mouse and dog. Low to moderate volumes
Table 1. Enzyme potency (c-erbB-2), cell potency and selectivity
figures in the rat for selected compounds
Q
X
c-erbB-2 inhibition
IC50 (mM)a
Cell potency/selectivity
IC50 (mM)b
1
1
1
1
1
2
2
2
2
3
3
3
A
B
D
E
F
A
B
C
E
A
B
E
0.09
0.01
6.6
0.56
1.3/10
NT
27/39
1.3/36
21/27
4.4
0.049
0.001
0.290
0.01
0.026
0.027
0.049
0.35/12.5
0.29/>5.6
3.1/8.9
1.1/19
NTc
0.097/>5.6
1.86/18.5
aValues are means of at least 2 experiments. Assays are run at Km
concentrations of ATP and substrate.
bPotencies and selectivities relate to inhibition of proliferation of
HB4a cells overexpressing c-erbB-2 relative to a mutant RAS HB4a
control.
cNT, not tested.
Table 2. In vitro evaluation of GW974. Enzyme potency for several
kinases and cell potencies
Kinase
enzyme
IC50 (mM)a
Cell line (overexpressingb)
IC50 (mM)
c-erbB-2
EGFr
c-erbB-4
CDK1
CDK2
Mek
0.018
0.001
0.02
>8.0
>8.0
9.7
HB4a.c5.2 (erbB2)
HB4a.R4.2 (RAS)
CaLu3 (erbB2)
HN5(egfr)
BT474 (erbB2)
MRC5 (non trans.)
—
0.097
30.0
0.39
0.21
0.07
>50.0
—
Figure 3. Dosing of 3B (Gw974) in the BT474 c-erbB-2 overexpressing
xenograft model in SCID mice. compound was dosed at 30 mg/kg for
5 days, at 10 mg/kg for 15 days. Dosing ceased at day 80. All doses
were well tolerated. Only one control animal remained alive at day 100.
Src
Raf
>50.0
49.0
—
—
aValues are means of at least 2 experiments. Assays are run at Km
concentrations of ATP and substrate.
bBrackets indicate class 1 RTK overexpressed in each cell type.
In addition, inhibition of the growth of EGFr over-
expressing xenografts was observed with this com-
pound. For example, 80% inhibition of growth of an
HN5 EGFr cell based xenograft was demonstrated in
SCID mice at 10 mg/kg, bid, oral dosing for 20 days.
Table 3. Pharmacokinetic data for compound 3B (GW974)a
PK parameter
Mouse
Rat
Dog
Cl (mL/min/kg)
Vdss
iv T1/2 (min)
F (%)
>27
>0.15
20
19
0.19
18
34
0.87
44
>23
24
82
Synthetic Chemistry
aCompound administered in b-cyclodextrin sulphate, DMSO water
suspensions at 1 mg/kg iv, 10 mg/kg po.
Quinazoline analogues were prepared via previously
described procedures.9 Pyridopyrimidines were prepared

1404
S. Cockerill et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1401–1405
from 6-chloronicotinic acid (Scheme 1). Curtius rearrange-
ment and ortholithiation of the Boc-amine 4 allowed a
facile preparation of the anthranilic acid equivalent 5.
Reaction with formamidine acetate furnished the chloro-
pyridopyrimidinone 6. Chlorination and aniline intro-
duction in acetonitrile to give 8 could be followed by
reaction with dimethylamine to give 9 (GW974 where
NHAr is 12). Alternatively, in order to allow efficient
parallel synthesis of the analogues described herein,
chloropyridopyrimidinone 6 could be reacted with
dimethylamine to furnish the dimethylaminopyridopyr-
imidinone 7. Subsequent chlorination could then be
followed by parallel introduction of a variety of anilines.
Synthesis of aniline top halves was straightforward as
exemplified in Scheme 2 for the N-benzyl indazolyl-
amine 12. Alkylation of the nitro bicycle (5-nitroindazole
shown) was followed by hydrogenation. In the cases of
anilines required for B, C, D and F in Figure 3, these
were prepared as mixtures and separated by flash chro-
matography. Indazole aniline B could also be isolated
free of the 2 regioisomer by an acetone–water recry-
stallisation in 30% yield from the 5-nitroindazole. 2-
Benzyl benzimidazole aniline for E in Figure 3 was pre-
pared by known methods.11
Summary
Described herein is the identification and evaluation of
GW974, a potent inhibitor of the tyrosine kinase activ-
ity of EGFr/c-erbB-2. This in vitro potency transcribed
to in vivo potency in xenografts overexpressing both
receptor kinases. At higher doses in the BT474 xeno-
graft, tumour shrinkage was observed and found to be
irreversible, that is no regrowth of tumours was
observed upon cessation of dosing. GW974 represents
valuable progress in the identification of combined
EGFr/c-erbB-2 inhibitors as an approach to cancer
chemotherapy with low toxicity.
References and Notes
1. (a) ONKOS reports. Based on seven country prevalence/
incidence data (France, Germany, Spain, Italy, Japan, UK,
USA). (b) Slamon, D. J.; Clark, G. M.; Wong, S. G. Science
1987, 235, 177. Macias, A.; Azavedo, E.; Hagerstrom, T.
Anticancer Res. 1987, 7, 459.
Scheme 1. (1) (PhO)2PON3, Et3N, tert-BuOH, Á, 83%; (2) (i) n-BuLi,
TMEDA, toluene,À60 to À15; (ii) CO2, 65%; (iii) aq NaOH, 85%;
(3) formamidine acetate, AcOH, 79%; (4) aq Me2NH, EtOH, MeCN,
90%; (5) (i) POCl3, Et3N, 90%; (ii) ArNH2, MeCN, Á, 60–90%;
(6) aq. Me2NH, EtOH, MeCN, 90%; (7) (i) POCl3, Et3N, 90%;
(ii) ArNH2, MeCN, Á, 60–90%.
2. (a) For a recent review of progress in this area, see:
Noonberg, S. B.; Benz, C. C. Drugs 2000, 59, 753. (b)
Bonomi, P.; Perez-Soler, R.; Chachoua, A.; Huberman, M.;
Karp, D.; Rigas, J.; Hammond, L.; Rowinsky, E.; Preston,
G.; Ferrante, K. J.; Allen, L. F.; Proceedings of the 11th
NCI-EORTC-AACR Symposium, Amsterdam, 2000. Goss,
G.; Lorimer, L.; Miller, W.; Hirte, H.; Stewart, D.; Batiste,
G.; Mathews, S.; Averbuch, S.; Seymour, L.; Proceedings of the
11th NCI-EORTC-AACR Symposium, Amsterdam, 2000.
3. Barker, A. J. Davies, J. H. Eur. Patent, 520,722, 1994.
4. Woodburn, J. R.; Barker, A. J.; Gibson, K. H. Abstract
4251, Proceedings of the 88th Annual Meeting of the Amer-
ican Association for Cancer Research, San Diego CA, Apr
12–16, 1997.
5. Shewchuk, L.; Hassell, A.; Wisely, B.; Rocque, W.;
Holmes, W.; Veal, J.; Kuyper, L. F. J. Med. Chem. 2000, 43,
133.
6. Bridges, J.; Zhou, H.; Cody, D. R.; Rewcastle, G. W.;
McMichael, A.; Showalter, H. D. H.; Fry, D. W.; Kraker,
A. J.; Denny, W. A. J. Med. Chem. 1996, 39, 267.
7. Rewcastle, G. W.; Palmer, B. D.; Thompson, A. M.;
Bridges, A. J.; Cody, D. R.; Zhou, H.; Fry, D. W.; McMichael,
A.; Kraker, A. J.; Denny, W. A. J. Med. Chem. 1996, 39, 1823.
8. 7-Substituted quinazolines and pyridopyrimidines have
Scheme 2. (1) K2CO3, benzyl bromide followed by aq acetone recry-
stallisation, 30%; (2) H2, Pd/C, 95%.

S. Cockerill et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1401–1405
1405
been reported to inhibit c-erbB-2 autophosphorylation in cells
(MDA-MB-453). Smaill, J. B.; Rewcastle, G. W.; Loo, J. A.;
Greis, K. D.; Chan, O. H.; Reyner, E. I.; Lipka, E.; Showalter,
H. D.; Vincent, P. W.; Elliott, W. L.; Denny, W. A. J. Med.
Chem. 2000, 43, 1380. Smaill, J. B.; Palmer, B. D.; Denny,
W. A.; McNamara, D. J.; Dobrusin, A. M.; Bridges, A. J.;
Zhou, H.; Showalter, H. D.; Winters, R. T.; Leopold, W. R.;
Fry, D. W.; Nelson, J. M.; Slintak, V.; Elliott, W. L.; Roberts, B.;
Vincent, P. W.; Patmore, S. J. J. Med. Chem. 2000, 43, 1380.
9. Cockerill, G. S.; Guntrip, S. B.; Mckeown, S. C; Page, M.
J.; Smith, K. J.; Vile, S.; Hudson, A. T.; Barraclough, P.;
Franzmann, K. W. WO Patent, 9,713,771, 1997.
10. Kita, Y.; Tseng, J.; Horan, T.; Wen, J.; Philo, J.; Chang,
D.; Ratzkin, B.; Pacifici, R.; Brankow, D. Biochem. Biophys.
Res. Commun. 1996, 226, 59.
11. Ram, S.; Wise, D. S.; Townsend, L. B. J. Heterocycl.
Chem. 1986, 23, 1109.