Discovery and biological evaluation of potent dual ErbB-2/EGFR tyrosine kinase inhibitors: 6-Thiazolylquinazolines

Article abstract of DOI:10.1016/S0960-894X(02)01047-8

We have identified a novel class of 6-thiazolylquinazolines as potent and selective inhibitors of both ErbB-2 and EGFR tyrosine kinase activity, with IC₅₀ values in the nanomolar range. These compounds inhibited the growth of both EGFR (HN5) and ErbB-2 (BT474) over-expressing human tumor cell lines in vitro. Using xenograft models of the same cell lines, we found that the compounds given orally inhibited in vivo tumor growth significantly compared with control animals.

Full text of DOI:10.1016/S0960-894X(02)01047-8

Bioorganic & Medicinal ChemistryLetters 13 (2003) 637–640
Discovery and Biological Evaluation of Potent Dual
ErbB-2/EGFR Tyrosine Kinase Inhibitors:
6-Thiazolylquinazolines
Micheal D. Gaul,y Yu Guo, Karen Affleck, G. Stuart Cockerill,{ Tona M. Gilmer,
Robert J. Griffin, Stephen Guntrip, BarryR. Keith, Wilson B. Knight,
Robert J. Mullin, Doris M. Murray, David W. Rusnak, Kathryn Smith,
Sarva Tadepalli, Edgar R. Wood and Karen Lackey*
GlaxoSmithKline, 5 Moore Drive, Research Triangle Park, NC 27709, USA
Received 20 August 2002; accepted 25 November 2002
Abstract—We have identified a novel class of 6-thiazolylquinazolines as potent and selective inhibitors of both ErbB-2 and EGFR
tyrosine kinase activity, with IC50 values in the nanomolar range. These compounds inhibited the growth of both EGFR (HN5) and
ErbB-2 (BT474) over-expressing human tumor cell lines in vitro. Using xenograft models of the same cell lines, we found that the
compounds given orallyinhibited in vivo tumor growth significantlycompared with control animals.
# 2003 Elsevier Science Ltd. All rights reserved.
Epidermal growth factor receptor (EGFR) and ErbB-2
are members of the Type I receptor tyrosine kinase
(TK) family(also known as the HER or ErbB family).
Over-expression of these receptors is found in a number
of cancers (e.g., breast, ovarian, colon, prostate) and
has been associated with poor prognosis in patients.1,2
Recent success in the clinical evaluation of TK inhibi-
tors (Fig. 1) stronglysuggests that these targets repre-
sent drug intervention opportunities.3 Excellent
descriptions of the involvement of erbB familyproteins
in cell physiology and disease applications have been
published byAdams and Yarden. 4,5 There are also recent
reviews that cover Type I Receptor TK inhibitors.6,7
Figure 1. Examples of ErbB familyTK inhibitors currentlyin clinical
trials for anti-cancer therapy.
Type I receptors function as ligand receptors in a matrix
of hetero- and homo-dimers. The greatest over amplifi-
cation of signal is observed with erbB-2 heterodimers
and this heterodimerization of the erbB-2 receptor may
be the preferred initiating event for signaling with this
receptor.2,5 There appears to be a strong hierarchyof
receptor dimerization, with those involving erbB-2 as
the most stable and preferred.2,5 For these reasons, we
designed molecules to specificallytarget both erbB-2
and EGFR TK inhibition.
*Corresponding author. Fax:+1-919-315-1092; e-mail: karen.e.lackey@
gsk.com
yCurrent address: 3-Dimensional Pharmaceuticals, 8 Clarke Drive,
Cranbury, NJ 08512, USA.
Because small substitution changes in kinase inhibitors
greatlyaffect the kinase inhibition and drug properties,
a series of compounds were generated to investigate the
{Current address: Arrow Therapeutics, Britannia House, 7 Trinity
Street, London SE1 1DS, UK.
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(02)01047-8

638
M. D. Gaul et al. / Bioorg. Med. Chem. Lett. 13 (2003) 637–640
structure–activityrelationships (SARs) of the 6-
thiazolylquinazoline series. We report several dual erbB-
2/EGFR TK inhibitors that possess efficacyin pre-clin-
ical models for cancer and a facile synthetic route to
obtain these compounds.
quantitative yield. Oxidation with potassium peroxy-
monosulfate produced the sulfonyl compound, which
was in turn converted to the thioamide bypassing
through hydrogen sulfide in the presence of triethyl-
amine. This thioamide was then utilized in step 4 of the
synthetic route illustrated in Figure 2.
A straightforward five step synthetic route was used to
generate 6-thiazolylquinazoline derivatives and is
depicted in Figure 2. Representative 4-anilino groups
used in step 1 are listed in Table 1 and the synthesis of
the thioamide reagent in step 4 is shown in Figure 3.
The quinazoline scaffold provides the necessarybinding
properties for inhibition of the erbB familyof tyrosine
kinases.7 The IC50 values for enzyme activity were gen-
erated bymeasuring the inhibition of the phosphoryl-
ation of
a
peptide substrate.11 Smaller aniline
Treatment of 4-chloro-6-iodoquinazoline8 with the
appropriatelysubstituted anilines in warm isopropanol
afforded the 4-anilino-6-iodoquinazoline in high yields
(>80%) using a simple filtration isolation. Stille coupl-
ing of the 4-anilino substituted iodoquinazoline with
1-ethoxy-1-vinyltributyltin gave the desired 6-(1-ethoxy-
vinyl)-quinazoline in 75% yield. Treatment of the
6-ethoxyvinylquinazoline with N-bromosuccinimide in
dichloromethane followed byan immediate Hantzsch
reaction9 of the unstable a-bromoketone intermediate
with the appropriate thioamide produced the sub-
stituted 6-(2,4-thiazole)-quinazoline in moderate
(15–40%) yields. Hydrolysis of the trifluoroacetamide
from the aminomethyl thiazole side chain with NaOH/
MeOH and subsequent treatment with anhydrous HCl
provided the desired final products as hydrochloride
salts in good (70–80%) yields.
Table 1. Catalytic enzyme assay results of 6-thiazolylquinazoline
aniline derivatives
EntryAniline group
1
ErbB-2
EGFR
IC50, mMa
IC50, mMa
0.2
0.035
ꢀ0.11
ꢀ0.022
2
3
0.048
ꢀ0.015
0.082
ꢀ0.053
The tolerance for substitution in the 6-position of the
quinazoline was generallylarge for potent enzyme inhib-
ition, so we used cellular activityto develop the SAR to
determine the preferred substitutions. Work in our
laboratoryhad demonstrated that one of the best
6-position side chain substitutions for providing optimal
cellular activitywas achieved when the 2-(methylsulfo-
nyl)ethyl amino group was the preferred side chain
substitution linked via a methylene unit to a hetero-
cyclic ring (Fig. 1, GW572016).10 The synthetic route to
the fullyfunctionalized thioamide for use in the con-
densation reaction to generate the thiazole is shown in
Figure 3. Beginning with the commerciallyavailable
methylthioethylamine, an alkylation with chloro-
acetonitrile followed byacetylation with trifluoroacetic
anhydride afforded the trifluoroacetamide in near
0.028
ꢀ0.011
0.071
ꢀ0.030
4
0.55
ꢀ0.11
0.54
ꢀ0.32
5
6
0.014
ꢀ0.0016
0.008
ꢀ0.0010
0.014
ꢀ0.0002
0.01
ꢀ0.0003
aAverage values nꢁ3.
Figure 2. Synthetic route to 6-thiazolylquinazoline derivatives.
Figure 3. Synthetic route to the thioamide side chain.

M. D. Gaul et al. / Bioorg. Med. Chem. Lett. 13 (2003) 637–640
639
substitutions generallyprovide compounds with potent
EGFR TK inhibition, while larger aniline substitutions
generallyconfer greater dual erbB-2 and EGFR tyro-
sine kinase inhibition. Using a mix and match strategy
for lead optimization, the ‘best’ anilines for dual inhibi-
tion were combined with the optimized side chain. The
results are listed in Table 1.
Table 3. Averaged enzyme values, average tumor cell line values, and
selectivityratio
EntryEGFR and
ErbB-2 TK
Tumor
cell lines
average
IC50, mM
HFF
average
IC50, mM
Selectivitya
X-fold
difference
dual inhibition
average
IC50, mM
1
2
3
4
5
6
0.054
0.028
0.012
0.32
0.011
0.012
2.45
1.32
0.089
7.4
0.29
0.11
>10
>30
12
>30
>30
>30
>4
>14
135
>4
>105
>210
The choice of cell lines for evaluating these compounds as
dual inhibitors was veryimportant. The control cell line
that was used was HFF, a normal human foreskin fibro-
blast and the desired compound profile should not inhibit
the proliferation of these cells. A head and neck carcinoma
cell line, HN5, which overexpresses EGFR, was used to
determine the effectiveness of the EGFR TK inhibitory
properties. Both N87, a gastric carcinoma cell line, and
BT474, a breast carcinoma cell line, overexpress erbB-2
and should be potentlyinhibited bya dual erbB-2/EGFR
TK inhibitor. Further, we used the same tumor cell lines
in subcutaneous (sc) human xenograft murine models to
evaluate the pre-clinical anti-tumor activity.
aRatio of the growth inhibitoryIC values of normal cell line (HFF)
to the average of the three tumor cell lines.
50
Table 4. Representative murine in vivo data
No.
IV
AUC
PO
AUC
F
(%)
%TIa
(HN5)b
%TIa
(BT474)c
1
2
3
5
6
nd
11,308
2489
14,722
19,040
nd
7534
274
2439
7520
nd
66.6
11
16.6
39.5
55
43
65
84
81
nd
33
0
59
83
A good correlation is observed between the catalytic
enzyme activity profile and the cellular efficacy of these
derivatives (Table 2). For example, 1 is a potent EGFR
TK inhibitor and it is relativelyselective for the HN5
tumor line over the normal cell line and the lines over-
expressing erbB-2 (N87 and BT474). Excellent growth
inhibitoryactivityis observed in cell lines over-expres-
sing EGFR (HN5) and erbB-2 (BT474 and N87) with
dual inhibitors such as 3, 5, and 6. The best overall cellular
activityis observed with compounds 3 (N-1-benzyl-
indazole) and 6 (4-[3-fluorobenzyloxy]-3-chloroaniline).
nd, no data.
a%TI=tumor inhibition.
bHead and neck tumor line.
cBreast tumor line.
IV or oral dose of compound at 10 mg/kg. The data are
compiled in Table 4 and demonstrate a range of plasma
exposures for these compounds with oral bioavailability
ranging from 11 to 67%. The compounds were also
administered orallyat 100 mg/kg BID for 21 days in the
sc xenograft studies of both an erbB-2 and EGFR over-
expressing human tumor cell lines. The tumor inhibi-
tion, which was recorded for the final dayof the study,
is listed in Table 4.12 The data illustrate that compound
6 was the most efficacious compound in this 6-thiazoyl-
quinazoline series, displaying approximately 80% tumor
inhibition in both the HN5 and BT474 xenograft mod-
els as further illustrated in Figures 4 and 5. It is note-
worthythat the 6-thiazole substitution generally
afforded better mouse oral exposures than the 6-furyl
Compounds were progressed for further studyprovided
that the IC50 values for cellular efficacywere below 300
nM and the selectivityratio exceeded 20-fold tumor
versus normal cells. The tabulated average dual enzyme
inhibition values and averaged tumor cell line values are
illustrated in Table 3. Since 1 would not be expected to
function as a dual inhibitor, it does not meet the criteria for
progression. However, the most promising 6-thiazolyl-
quinazoline compounds (3, 5, 6) exceed 100-fold selec-
tivityfor tumor cells over normal cells.
13
series that led to the discoveryof GW572016.
Pharmacokinetic parameters were generated in female
CD-1 mice (n=2 per time point), treated with a single
Table 2. Cellular activityresults for a representative normal line
(HFF) and tumor lines
a
EntryHFF
HN5b
N87c
BT474d
1
2
>10
>18
12
>30
>30
>23
0.57ꢀ0.14
1.6ꢀ0.4
0.11ꢀ0.01
18
4.4ꢀ0.26
1.5ꢀ0.3
0.09ꢀ0.01
7.4
2.4ꢀ3.3
0.86ꢀ0.5
0.06ꢀ0.00
15
3
4e
5
0.23ꢀ0.02
0.36ꢀ0.12
0.28ꢀ0.11
6
0.08ꢀ0.01
0.13ꢀ0.03
0.11ꢀ0.04
IC50 values reported as mM concentrations.
aHuman foreskin fibroblasts.
bHead and neck tumor line over-expressing EGFR.
cGastric tumor line over-expressing erbB-2.
dBreast tumor line over-expressing erbB-2.
en=1.
Figure 4. Inhibition of HN5 xenograft growth by 6.

640
M. D. Gaul et al. / Bioorg. Med. Chem. Lett. 13 (2003) 637–640
dual inhibitors of ErbB-2/EGFR tyrosine kinases have
the potential for providing therapeutic benefit across a
broad range of tumors.
Acknowledgements
We would like to thank KimberlyPetrov for synthesiz-
ing several unique anilines and Stephen Frye for his
guidance and support.
References and Notes
1. Slichenmeyer, W. J.; Elliott, W. L.; Fry, D. W. Semin.
Oncol. 2001, 28 (Suppl. 16), 67.
2. Rowinsky, E. K. Horizons in Cancer Therapeutics: From
Bench to Bedside 2001, 2, 2.
3. Cockerill, S. G.; Lackey, K. E. Curr. Topics Med. Chem.
2002, 2, 1001.
4. Adams, J. A. Chem. Rev. 2001, 101, 2271.
Figure 5. Inhibition of BT474 xenograft growth by 6.
5. Yarden, Y.; Sliwkowski, M. X. Nat. Rev. 2001, 2, 127.
6. Dumas, J. Curr. Opin. Drug Discov. Develop. 2001, 4, 378.
7. Bridges, A. J. Chem. Rev. 2001, 101, 2541.
The data for tumor weights measured over the treat-
ment period are shown graphicallyfor 6, one of the
most efficacious compounds in this series, in both the
erbB-2 and EGFR over-expressing lines. The control
line represents a vehicle treated group of animals and
approximatelytwo tumor size doublings occur during
the treatment period. A dose response can be seen in
both models for a 0, 30, 100 mg/kg BID po of com-
pound 6. Lack of significant bodyweight loss (data not
shown) during treatment with 6 in the subcutaneous in
vivo xenograft mouse model suggests a good thera-
peutic index.
8. Hudson, A. T.; Vile, S.; Barraclough, P.; Franzmann,
K.W.; McKeown, S. C.; Page, M. J. WO 9609294 A1, 1996.
9. Schwarz, G. Org. Synth. Coll., 1955; Vol. III, p 332.
10. Affleck, K.; Allen, P.; Alligood, K. J.; Cai, Z.; Carter,
M. C.; Cockerill, G. S.; Crosby, R. M.; Dickerson, S.; Frye,
S. V.; Gaul, M.; Gauthier, C.; Gilmer, T. M.; Glennon, K.;
Griffin, R.; Guntrip, S. B.; Guo, Y.; Johnson, N. W.; Keith,
B. R.; Knight, W. B.; Luzzio, M.; Mook, R. A.; Mullin, R. J.;
Murray, D. M.; Rusnak, D. W.; Tadepalli, S. M.; Sinhababu,
A. K.; Smith, K. J.; Wood, E.; Zhang, Y.; Lackey, K. In Proc.
Am. Assoc. Cancer Res. 92nd Annual Meeting, New Orleans,
LA, 2001.
11. Brignola, P. S.; Lackey, K.; Kadwell, S. H.; Hoffman, C.;
Horne, E.; Carter, H. L.; Stuart, J. D.; Blackburn, K.; Moyer,
M. B.; Alligood, K. J.; Knight, W. B.; Wood, E. R. J. Biol.
Chem. 2002, 277, 1576.
The quinazoline scaffold provides the necessarybinding
properties for inhibition of the ErbB familyof tyrosine
kinases. A series of 6-thiazolyl-quinazolines were dis-
cussed and the results of extensive SAR investigations
were summarized using representative examples. The
compounds possessed excellent in vitro tyrosine kinase
inhibition as well as selective tumor cell-based activity.
Furthermore, the compounds provided an acceptable
pharmacokinetic profile to allow for oral dosing in a
series of subcutaneous in vivo xenograft models. All of
these studies, taken together, resulted in the discoveryof
compound 6, which showed good anti-tumor activity
with a lack of significant bodyweight loss during treat-
ment, suggesting a good therapeutic index for this com-
pound. Dual inhibition of EGFR and ErbB-2 mayoffer
increased activityover agents which target onlyone of
these receptor kinases. Our results suggest that these
12. Cells were cultured in RPMI 1640 + 10% fetal bovine
serum, sodium pyruvate and l-glutamine, at 37 ꢂC in a 95/5
air/CO2 atmosphere. HN5 tumors were initiated byinjection
of the cell suspension subcutaneouslyin the axillaryregion.
BT474 tumors were initiated byinjection of the tumor frag-
ments subcutaneouslyin the axillaryregion. Solid tumors were
measured byelectronic caliper measurement. Tumor weights
were calculated with the equation Tumor weight (mg)=(tumor
length (mm)ꢃtumor width (mm)2/2. Compounds were admini-
stered po using a 10% sulfobutyl ether beta cyclodextrin for-
mulation. These studies were performed under under a
protocol approved bythe institutional animal care committee.
13. Rusnak, D. W.; Lackey, K.; Affleck, K.; Wood, E. R.;
Alligood, K. J.; Rhodes, N.; Keith, B.; Murray, D. M.; Mul-
lin, R. J.; Knight, W. B.; Gilmer, T. M. Mol. Cancer Thera.
2001, 1, 85.