Tyrosine kinase inhibitors. 8. An unusually steep structure-activity relationship for analogues of 4-(3-bromoanilino)-6,7-dimethoxyquinazoline (PD 153035), a potent inhibitor of the epidermal growth factor receptor

Article abstract of DOI:10.1021/jm9503613

4-(3-Bromoanilino)-6,7-dimethoxyquinazoline (32, PD 153035) is a very potent inhibitor (IC₅₀ 0.025 nM) of the tyrosine kinase activity of the epidermal growth factor receptor (EGFR), binding competitively at the ATP site. Structure-activity relationships for close analogues of 32 are very steep. Some derivatives have IC₅₀s up to 80-fold better than predicted from simple additive binding energy arguments, yet analogues possessing combinations of similar phenyl and quinazoline substituents do not show this 'supra-additive' effect. Because some substituents which are mildly deactivating by themselves can be strongly activating when used in the correct combinations, it is proposed that certain substituted analogues possess the ability to induce a change in the conformation of the receptor when they bind. There is some bulk tolerance for substitution in the 6- and 7-positions of the quinazoline, so that 32 is not the optimal inhibitor for the induced conformation. The diethoxy derivative 56 [4-(3-bromoanilino)- 6,7-diethoxy-quinazoline] shows an IC₅₀ of 0.006 nM, making it the most potent inhibitor of the tyrosine kinase activity of the EGFR yet reported.

Full text of DOI:10.1021/jm9503613

J . Med. Chem. 1996, 39, 267-276
267
Tyr osin e Kin a se In h ibitor s. 8. An Un u su a lly Steep Str u ctu r e-Activity
Rela tion sh ip for An a logu es of 4-(3-Br om oa n ilin o)-6,7-d im eth oxyqu in a zolin e (P D
153035), a P oten t In h ibitor of th e Ep id er m a l Gr ow th F a ctor Recep tor
Alexander J . Bridges,*,† Hairong Zhou,† Donna R. Cody,† Gordon W. Rewcastle,‡ Amy McMichael,†
H. D. Hollis Showalter,† David W. Fry,† Alan J . Kraker,† and William A. Denny*,‡
Parke-Davis Pharmaceutical Research, Division of Warner-Lambert Company, 2800 Plymouth Road,
Ann Arbor, Michigan 48106-1047, and Cancer Research Laboratory, University of Auckland School of Medicine,
Private Bag 92019, Auckland, New Zealand
Received May 18, 1995X
4-(3-Bromoanilino)-6,7-dimethoxyquinazoline (32, PD 153035) is a very potent inhibitor (IC50
0.025 nM) of the tyrosine kinase activity of the epidermal growth factor receptor (EGFR), binding
competitively at the ATP site. Structure-activity relationships for close analogues of 32 are
very steep. Some derivatives have IC50s up to 80-fold better than predicted from simple additive
binding energy arguments, yet analogues possessing combinations of similar phenyl and
quinazoline substituents do not show this “supra-additive” effect. Because some substituents
which are mildly deactivating by themselves can be strongly activating when used in the correct
combinations, it is proposed that certain substituted analogues possess the ability to induce a
change in the conformation of the receptor when they bind. There is some bulk tolerance for
substitution in the 6- and 7-positions of the quinazoline, so that 32 is not the optimal inhibitor
for the induced conformation. The diethoxy derivative 56 [4-(3-bromoanilino)-6,7-diethoxy-
quinazoline] shows an IC50 of 0.006 nM, making it the most potent inhibitor of the tyrosine
kinase activity of the EGFR yet reported.
In tr od u ction
the unsubstituted 4-(3-bromoanilino)quinazoline parent
compound 4. This leaves the 6,7-dimethoxy derivative
32 another 2 orders of magnitude more potent than
would be expected from completely additive effects of
the two substituents. In the present paper we focus on
some closely related analogues of 4, delineate where in
these SAR “supra-additive” substituent effects can be
found, and speculate on their origin.
Many of the tyrosine kinase enzymes which are early
components of the growth signal transduction pathway
in mammalian cells1 are encoded by proto-oncogenes,
and their transformation or overexpression has been
shown to occur in a large percentage of clinical cancers.2-5
These tyrosine kinase enzymes, particularly the recep-
tors for growth factors such as EGF, bFGF, and PDGF,
have thus become important targets for drug design.1,6
Several classes of inhibitors have been described,6-9 but
most have proven to be of limited use in cellular assays
and more advanced models. Following up on leads from
our mass screening program, and concomitant patent
publications from Zeneca,10-12 we have recently re-
ported13,14 on the inhibitory properties of 4-anilino-
quinazolines against the epidermal growth factor re-
ceptor (EGFR). Despite binding at the ATP site of
EGFR (a site of close homology among different ki-
nases),12,15 members of this class are very potent and
selective inhibitors of the EGFR. In particular, we have
shown that 4-(3-bromoanilino)-6,7-dimethoxyquinazo-
line (32) has an IC50 of 25 pM (Ki 6 pM) for this
enzyme.13
In a previous publication14 we delineated the initial
structure-activity relationship (SAR) for this series.
The pyrimidine ring is mandatory, and a free NH linker
is clearly optimal. Electron-withdrawing lipophilic sub-
stituents on the 3-position of the aniline are favorable,
with chlorine and bromine optimal, as reported previ-
ously by others.12 Electron-donating groups at the 6-
and 7-positions (but not the 5- and 8-positions) of the
quinazoline are preferred. However, the best of these
only produce an increase in potency of about 10-fold over
Ch em istr y
The compounds were prepared by coupling the ap-
propriate 4-chloroquinazolines with substituted anilines
in refluxing 2-propanol as described previously.14 The
6-nitro-7-(methylamino), -(dimethylamino), and -meth-
oxy derivatives 49, 50, and 52 were prepared from the
known16 7-chloro-6-nitroquinazolinone 70. This was
first converted to 4,7-dichloro-6-nitroquinazoline (71),
which underwent selective displacement of the 4-chloro
group with 3-bromoaniline to give 53, followed by
substitution of the 7-chloro group to give 49 and 50
(Scheme 1). Conversion of 70 to 72 with methoxide
followed by chlorination to 73 and reaction of this with
3-bromoaniline gave 52. The 6-NHMe derivative 35 was
prepared by LiAlH4 reduction of the methyl carbamate
37, and the 7-NRR compounds 40-42 were prepared
by fluoride displacement from the 7-fluoro analogue 75
(Scheme 2). The 7-substituted-6-NH2 compounds 44-
47 were prepared by reduction of the corresponding
† Parke-Davis Pharmaceutical Research.
‡ University of Auckland School of Medicine.
X Abstract published in Advance ACS Abstracts, December 1, 1995.
0022-2623/96/1839-0267$12.00/0 © 1996 American Chemical Society

268 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1
Bridges et al.
Sch em e 1a
Sch em e 4a
a
(i) RCN/HCl and then NaHCO3; (ii) POCl3/reflux and then
3-bromoaniline/2-methoxyethanol/HCl.
a
(i) SOCl2/DMF/reflux/1 h; (ii) 3-bromoaniline/HCl (trace)/
iPrOH/reflux/30 min; (iii) R2NH or RNH2/EtOH/100 °C (pressure
vessel)/1 h; (iv) MeOH/MeONa/100 °C (pressure vessel)/18 h; (v)
POCl3/reflux/30 min.
conditions competitive loss of the anilino side chain
became appreciable. Conversion of 55 to the corre-
sponding higher alkyl ethers 56-58 was accomplished
by treatment with an excess of the corresponding alkyl
iodide and K2CO3 or Cs2CO3 in DMF at 25 °C, followed
by preparative TLC (Scheme 3).
Sch em e 2a
The 2-substituted analogues 61 and 62 were prepared
as shown in Scheme 4. Reaction of methyl 4,5-
dimethoxyanthranilate (82) with either acetonitrile or
cyanamide using HCl gas catalysis gave the quinazoli-
nones 83 and 84, respectively. In both cases the
intermediate amidines cyclized spontaneously upon
neutralization of the reaction mixture. The quinazoli-
nones were converted to 61 and 62 by usual 4-chlorina-
tion (POCl3) and displacement with 3-bromoaniline in
2-methoxyethanol at 90 °C.
a
(i) POCl3/reflux/1 h and then iPrOH/3-bromoaniline/HCl (trace)/
reflux/30 min; (ii) R2NH or RNH2/N-Me-pyrrolidone/130 °C/4 days.
Sch em e 3a
Resu lts a n d Discu ssion
The structures, physicochemical properties, and EGFR
inhibitory potencies (IC50 in nanomolar) are given in
Table 1. The IC50 values are the average of at least two
independent determinations of full dose-response curves.
The assay used initially was a purified full length
EGFR, stimulated by EGF, phosphorylating a tyrosine-
containing peptide based upon the major EGFR phos-
phorylation site on PLCγ.13 Similar results were seen
with representative compounds when the phosphopep-
tide substrate was substituted by a Glu-Tyr copolymer.
The conditions under which this assay was used,
described in detail elsewhere,13 were critical. The use
of less purified enzyme preparations gave unreliable
results with these ATP competitive compounds, while
performing reliably with literature standards which are
probably not inhibitors of the same mechanistic class,
such as several tyrphostins,17 a sulfonylbenzoylnitrosty-
rene,18 a thioindole,9 and erbstatin.19
Table 2 compares, for a limited set of quinazoline
substituents (first column), the effect on activity of
changes in the 3-halogen in the aniline side chain. For
each quinazoline substituent pattern, the first row in
Table 2 gives the compound number from Table 1 and
the second row the actual IC50 value for the specified
compound. The third row is the ratio of the IC50 value
of the parent anilinoquinazoline (1, 7, 9, 12, 15, 17, 23,
29) over the IC50 values for the corresponding 3-substi-
tuted side chain analogues [344/IC50(3′-X)] and is there-
fore a direct measure of the effect of the 3′-H/3′-X
substitution. The entry in the fourth row is the notional
IC50 one would calculate for that compound assuming
that the quinazoline and aniline substituents act com-
a
(i) Formamidine hydrochloride/210 °C/15 min (R ) H) or
triazine/piperidine/MeOH/reflux (R ) OMe); (ii) (COCl)2/DMF (R
) H) or POCl3/reflux (R ) OMe); (iii) 3-bromoaniline/iPrOH/25
°C/18 h or reflux/1 h; (iv) pyridine hydrochloride/205 °C; (v) RI/
K2CO3 or RI/Cs2CO3/DMF/25 °C.
6-NO2 compounds 49, 50, 52, and 53 with Fe dust in
65% aqueous EtOH.
The normal cyclization of anthranilic acids with
formamide at elevated temperatures (170-195 °C) was
not efficient in forming the quinazolinone precursors for
compounds 32 and 64. In the former case, the quinazoli-
none 78 was formed by cyclizing the anthranilic acid
76 with an intimately ground mixture of formamidine
hydrochloride at 210 °C for 15 min, and in the latter
the trimethoxyquinazolinone 79 was formed from 77 by
refluxing with triazine and a catalytic amount of pip-
eridine in methanol (Scheme 3). The 6,7-dimethox-
yquinazolinone 78 was converted in good yield to the
4-chloro compound 80 with oxalyl chloride/DMF (Vils-
meier procedure), while the 5,6,7-trimethoxy derivative
79 was converted to 81 using POCl3 in 1,2-dichloroet-
hane buffered with Hunig’s base. Demethylation of 32
to 55 was achieved by fusion with pyridinium hydro-
chloride at 200 °C. Extended reaction times (>2 h) were
required for complete demethylation, but under these

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1 269
Ta ble 1. Physicochemical and EGFR Tyrosine Kinase Inhibition Data for 4-Anilinoquinazolines
a
no.
R1
R2
X
mp (°C)
ref 14
formula
C14Hl0N3F‚HCl
anal.
IC50 (nM)
1
2b
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
344
56
F
257-260
C,H,N
3b,c
4b
Cl
Br
I
CF3
H
Br
H
CF3
Br
H
Br
CF3
H
Br
H
F
Cl
Br
I
CF3
H
F
Cl
Br
I
CF3
H
F
ref 14
23
ref 14
27
5b
ref 14
80
6b
ref 14
577
55
7
OMe
OMe
NH2
NH2
NH2
NO2
NO2
NO2
H
260-261
C
15H13N3O‚HCl
C,H,N
8
ref 14
30
9
195-196
180-181.5
ref 14
C14H12N4
C15H11F3N4
C,H,N
C,H,N
770
574
0.78
10d
11
12
13
14d
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29d
30d
31d
32
33
34d
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
61
62
63
64
65
66
67b
68
69
ref 14
ref 14
5000
900
>104
120
10
209-211
265-267
ref 14
199-201
271-274
311-313
ref 14
259-261
271-274
ref 14
253-254
311-313
ref 14
C15H9F3N4O2
C15Hl3N3O‚HCl
C,H,N
C,H,N
MeO
MeO
NH2
NH2
NH2
NH2
NH2
NH2
NO2
NO2
NO2
NO2
NO2
NO2
OMe
OMe
OMe
OMe
OMe
OMe
H
H
H
C14H12N4
C,H,N
C,H,N
C,H,N
100
H
C14H11N4F‚HCl‚0.5C3H8O
2.0
H
C14H11N4Cl‚2HCl
0.25
H
0.1
H
C14H11IN4‚0.4AcOH‚0.5H2O
C15H11N4F3‚HCl‚C3H8O
C,H,N
C,H,N
0.35
H
3.3
H
1.2 × 104
H
C14H9N4FO2‚HCl
C14H9N4ClO2‚HCl
C,H,N
C,H,N
6100
H
810
1000
540
H
H
248-249 dec
240-241 dec
>250
C14H9N4IO2‚HCl
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
H
C15H9N4F3O2‚HCl
C16H15N3O2‚HCl
>104
29
OMe
OMe
OMe
OMe
OMe
OMe
NHMe
NMe2
NHCO2Me
H
253-254
261-262
ref 13
C16Hl4N3FO2‚HCl‚H2O
C16Hl4N3ClO2‚HCl
3.8
Cl
Br
I
0.31
0.025
0.89
0.24
4
273
C
16H14N3O2I‚HCl
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
HRMS
C,H,N
CF3
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
274-276 dec
141-144
248-249
197-198.5
290 dec
351-353
198-199
160-161.5
239-240
ref 14
C17Hl4N3F3O2‚HCl
C
15H13BrN4
H
H
C16H16BrN4
84
12
4.7
C16H13BrN4O2
OH
C
C
C
C
14H10BrN3O
16H13BrN4O
15H13BrN4
16H15BrN4
H
NHAc
NHMe
NHEt
NMe2
NH2
NHMe
NMe2
OMe
Cl
40
7
12
11
H
H
H
C16H15BrN4
NH2
NH2
NH2
NH2
NH2
NO2
NO2
NO2
NO2
NO2
NO2
0.12
0.69
159
3.8
6.5
238-240
174-175
210-211.5
218.5-220.5
ref 14
253-255
240.5-243.5
ref 14
206-207
257-259
278.5-280
ref 14
155-167
145-146
123-124
159-160
250-252
264-265
262-263
218-219
130-132
212-213
230-231
275
C15H14BrN5
C16H16BrN5
C15H13BrN4O
C14H10BrClN4
C,H,N
C,H,N
C,H,N
C,H,N,Cl
NH2
NHMe
NMe2
NHAc
OMe
Cl
53
C15H12BrN5O2
C16H14BrN5O2
C,H,N
C,H,N
68
2000
28
C15H11BrN4O3
C14H8BrClN4O2
C15H10BrN3O2
C,H,N
C,H,N
C,H,N
15
25
OCH2O
15
OH
OH
0.17
0.006
0.17
OEt
OPr
OBu
OEt
OPr
OBu
C
C
C
C
C
18Hl8N3BrO2‚H2O
20H22BrN3O2‚1.5H2O
22H26BrN3O2
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
105
5,6-diOMe
7,8-diOMe
16H14BrN3O2
1367
>104
>104
463
16H14BrN3O2
C17H16BrN3O2‚HCl
2-Me
3′-Br
2-NH2
3′-Br
C16H15BrN4O2‚3HCl‚0.25H2O
C17H16BrN3O2‚HCl‚0.25C3H8O
N4-Me
3′-Br
152
5-OMe
3′-Br
C
C
17H16BrN3O3
17H16BrN3O3
0.67
8-OMe
3′-Br
>104
H
H
H
H
2′-Br
C16H14BrN3O2‚HCl
128
4′-Br
C16H14N3BrO2‚HCl
0.96
0.072
113
3′,4′-Br
3′,5′-diBr
254-264
>250
C
C
16H13Br2N3O2‚HCl
16H13Br2N3O2‚0.5H2O
a
IC50 ) concentration of drug (nM) to inhibit the phosphorylation of a 14-residue fragment of phospholipase Cγ1 by EGFR (prepared
from human A431 carcinoma cell vesicles by immunoaffinity chromatography). See the Experimental Section for details. Values are the
averages from at least two independent dose-response curves; variation was generally (15%. Reference 10. c Reference 12. Reference
b
d
11.

270 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1
Bridges et al.
Ta ble 2. Additivity of 3′-Halo and 6,7-Quinazoline Substituents in 4-Anilinoquinazolines
substituent
3′-H
3′-F
3′-Cl
3′-Br
3′-I
3′-CF3
compd no.
1
2
56
6.1
1.1
3
23
15
1.6
4
27
12.7
1.5
5
80
4.3
0.9
6
577
0.60
-0.30
6,7-diH
344a
IC50 ratiob
∆(ΣE) (kcal/mol)c
1
0
compd no.
6-H, 7-NH2
IC50 ratio
calcd IC50
∆(ΣE) (kcal/mol)
17
100
1 (3.44)
100
0 (0.74)
18
2.0
50
16
1.24
19
0.25
400
6.7
20
0.10
1000
21
0.35
286
23
2.49
22
3.3
30
167
2.33
d
7.9
2.6
1.95
compd no.
6-NH2, 7-H
IC50 ratio
calcd IC50
∆(ΣE) (kcal/mol)
9
770
1 (0.45)
770
0 (-0.48)
11
0.78
990
61
2.6
10
574
1.35
1283
0.48
compd no.
6,7-diMeO
IC50 ratio
calcd IC50
∆(ΣE) (kcal/mol)
29
29
1 (11.9)
29
0 (1.47)
30
3.8
7.6
4.8
0.13
31
0.31
94
1.93
1.1
32
0.025
1160
2.28
2.7
33
0.89
33
6.7
1.2
34
0.24
121
48.4
3.16
compd no.
6-NO2, 7-H
IC50 ratio
calcd IC50
∆(ΣE) (kcal/mol)
12
5000
1 (0.069)
5000
0 (-1.59)
13
900
5.6
390
-0.50
14
>10 000
<0.5
compd no.
6-H, 7-NO2
IC50 ratio
calcd IC50
∆(ΣE) (kcal/mol)
23
12 000
1 (0.029)
12 000
0 (-2.1)
24
6100
1.97
1970
-0.65
25
810
14.8
800
0
26
1000
12
945
0
27
540
22
28
>10 000
e1.2
2790
1/1.67
∼0
0.98
compd no.
7
8
6-MeO, 7-H
IC50 ratio
calcd IC50
55
30
1.8
4.3
-1.15
1 (6.3)
55
0 (1.1)
∆(ΣE) (kcal/mol)
compd no.
6-H, 7-MeO
IC50 ratio
calcd IC50
∆(ΣE) (kcal/mol)
15
120
1 (2.87)
120
0 (0.63)
16
10
12
9.4
∼0
a
b
Experimental IC50 value (nM) for the compound specified by the first row and column. Ratio of IC50 values of the parent unsubstituted
compound (1, 344 nM) over the IC50 values for the 3-substituted side chain analogues [)344/IC50(aniline-3′-X)]. Figures in parentheses
are ratios of the IC50 value of 1 over the IC50 values for the quinazoline-substituted compounds [)344/IC50(quinazoline-R)]. c Difference
in binding energy [∆(ΣE) in kcal/mol] between the experimental and calculated IC50s (see text). Compounds which bind more strongly
d
than anticipated show a positive value, those that bind less well than expected a negative value. Calculated IC50, assuming additive
effects of quinazoline and aniline substituents to binding energy (see text).
pletely independently of each other in their contribu-
tions to the binding constant (energetically additive).
Finally, the fifth entry is the difference in binding
energy [∆(ΣE) in kcal/mol] between the experimental
and calculated IC50s, calculated from the Boltzmann
equation (∆G ) -RT ln K) by making the reasonable
assumption that enzyme inhibition is directly propor-
tional to the binding energy of the inhibitor to the
enzyme, so that IC50 differences are directly proportional
to ∆(ΣE). Compounds which inhibit more strongly than
anticipated show a positive “excess” binding energy;
those which inhibit less well than expected show a
negative value. For those where the quinazoline and
anilino substituents act completely independently of
each other, this value will be close to zero. The figures
in parentheses in the first column are the ratios of the
IC50 value of the parent unsubstituted compound (1)
(344 nM) over the IC50 values for the compounds
substituted only on the quinazoline ring [344/IC50-
(quinazoline-R)], and the corresponding ∆(ΣE) values,
and are used to calculate “expected” binding energies.
From Table 1 it is apparent that the compounds,
which vary little in structure, show an enormous range
in potency (>106-fold, IC50s from 12 000 to 0.006 nM).
The large range in potencies among 4-anilinoquinazo-
lines as EGFR inhibitors was also shown in the earlier
SAR analysis,14 but this was with a more structurally
heterogeneous data set. The present investigation of
the shape of the SAR around the anilinoquinazoline 4
fell into four distinct parts. The first part investigates
the effects of individual substituents on either the
quinazoline or anilino ring and establishes “base-line”
values for them. The second part looks at the effects of
combining certain quinazoline substituents with a series
of halogens in the 3′-position. In the third part, the
anilino substituent is held constant (as 3′-bromo) and
the quinazoline substituents are altered. In the last
part, compound 32 is used as a template and the effects
of adding a substituent to the various available quinazo-
line positions are investigated.
Gen er a l SAR for Qu in a zolin e a n d An ilin e Su b-
stitu en ts. We initially investigated the effect of a series
of substitutions either on the aniline ring or at the 6-
and/or 7-position of the quinazoline rings (compounds
1-7, 12, 15, 17, 23, and 29 of Table 1). These com-
pounds, many of which have been reported previously

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1 271
by ourselves and others,10-14 provided base-line data
for the effects of single substitutions on inhibitory
potency. The data for these compounds in Tables 1 and
2 show that the effect of individual substitutions (top
row and first column of Table 2) is generally not very
great. All the 3′-halogens in the aniline ring are 4-15-
fold activating, representing an increase of binding
energy to the enzyme of 0.9-1.6 kcal. The similar CF3
group appeared to be slightly detrimental, but the effect
is small. For quinazoline substituents, nitro has a large
deactivating effect (over a 2 kcal/mol loss of binding
energy for 7-NO2). However, as noted previously,14
single electron-donating substituents at either the 6- or
7-position are usually moderately activating. The most
potent compound in this group is the 6,7-diOMe (29),
where the effects of the two methoxy substituents are
almost additive, leading to a reasonable 12-fold increase
in potency. Although these are significant substituent
effects, they do not compare to what can happen when
the two sets of substituents are combined.
3′-Br substituent and varied the substituents on the
quinazoline ring to determine the extent of the supra-
additive effect (compounds 35-59). Because the cor-
responding debromo derivatives were not made in this
study, we cannot definitely say where the supra-additive
effect was present, but we can make the implicit
assumption that it is for the very potent compounds.
None of the other single quinazoline substituents ex-
amined produced the potentiating effect of the 7-NH2.
Perhaps most surprising was the fact that both of the
the NMe2 derivatives 36 and 42 were around 100-fold
less potent than the corresponding amino compounds
11 and 20. The NHMe analogues 35 and 40 were of
intermediate potency, making it hard to determine the
status of these compounds. In line with the results
noted above for nitro substituents, the less electron-
donating NHAc and NHCOOMe substituents gave
compounds (37 and 39) of modest activity.
In the 6,7-disubstituted compounds, the 6,7-diNH2
derivative 43 had essentially the same potency as the
7-NH2 derivative 20, suggesting either that the electron
donation of the 7-NH2 is large enough to give the full
effect available or that unfavorable steric or H-bonding
effects occur between the o-amino groups (see below).
The relatively low potency of the 6-amino-7-methoxy
compound 44 makes it unlikely that a supra-additive
effect is present in this case. Compounds 47 and 53
suggest that a 7-chloro substituent is moderately dis-
advantageous (8-fold) in conjunction with 6-NH2 but
highly beneficial (200-fold) in conjunction with 6-NO2.
In the other 6-NO2-7-substituted compounds 48-53, all
of the electron-donating 7-groups (except the bulky
dimethylamino) strongly mitigate (15-67-fold) the dys-
therapeutic effect of the 6-NO2 substituent.
Effects of Va r yin g th e 3′-Ha logen on th e An ilin e
Sid e Ch a in . The second part of the investigation
examined the effects of varying the 3′-halogen on the
aniline for three series of quinazoline-substituted ana-
logues (7-NH2, 7-NO2, and 6,7-diOMe) (compounds 17-
34). The 7-NH2 and 6,7-diOMe series were selected
because these substituent patterns provided the most
potent single- and double-substituted compounds found
in the earlier study.14 In a few other cases (6-NH2,
6-NO2, 6-OMe, and 7-OMe) just the 3′-Br and unsub-
stituted anilino compounds were compared. The bind-
ing data in rows 5-8 of Table 2 (for the 6- and 7-NO2
and 6- and 7-OMe quinazoline substituents) suggest
that the substituent groups are by and large indepen-
dent of one another. Thus, four of these compounds (17,
25, 26, and 28) show the IC50s predicted by additivity
arguments, three (8, 13, and 24) are 3-17 times less
potent, and one (27) is 5-fold more potent than pre-
dicted. This is not an unusual amount of scatter, and
as this data set includes the strongly deactivating nitro
substituents, and the most strongly activating single
substituent (6-OMe) when attached to 1, the results
might be expected to be broadly applicable.
In the 6,7-diOR-substituted compounds 32 and 54-
58, combining the two OMe substituents of 32 into a
methylenedioxy ring (54) resulted in enormous (600-
fold) loss of potency. In compound 54, the covalently
bonded 5-membered ring does not place the oxygen lone
pairs in an optimal orientation for overlap with the
aromatic π system. Conversion of both methoxy sub-
stituents to hydroxyls was moderately deleterious, with
diol 55 being 6-fold less potent than 32. Conversely,
making the alkyl side chains a little larger led to the
the diethoxy analogue 56 which, with an IC50 of 0.006
nM, is the most potent quinazoline inhibitor of the
EGFR tyrosine kinase activity we have found. The
dipropoxy side chain (57, IC50 0.17 nM) gave potency
similar to diol 55, and the dibutoxy compound 58 (IC50
105 nM) appears too large to be tolerated well.
However, the results for the 6-NH2, 7-NH2, and 6,7-
diOMe substituents (rows 2-4 of Table 2) show a quite
different pattern. An example of the large differences
between closely related compounds is the 100-fold
greater potency of 20 (3′-Br,7-NH2) compared with 16
(3′-Br,7-OMe). Whereas the latter has exactly the IC50
predicted by simple additivity, the former is 80-fold
more potent than expected (Table 2). Similar “supra-
additive” effects are seen for the 3-bromoanilino deriva-
tives bearing 6-NH2 and 6,7-diOMe substituents (11 and
32). This effect is not restricted to bromine, being seen
(where studied in the 7-NH2 and 6,7-diOMe series) with
the other three halogens. On average, the 3′-Br com-
pounds show 2.6 kcal/mol of “excess” binding energy,
the 3′-Cl and 3′-I compounds 1.6 kcal/mol, and the 3′-F
compounds 0.7 kcal/mol. Most interesting is the 3′-CF3
substituent, which is mildly deactivating in the parent
compound 6, is at least as activating as bromine in the
7-amino and 6,7-dimethoxy series, but only weakly
activating in conjunction with the 6-NH2 substituent.
These results suggest that that there is significant
bulk tolerance available in the region of the 6- and
7-positions of the quinazoline for further design. The
limits of this area of bulk tolerance are demonstrated
by compounds 59 and 60, where moving the gem-diOMe
substituents to either of the other allowed positions in
the B-ring (5,6- or 7,8-positions) is clearly highly
detrimental.
Ad d ition a l Su bstitu en ts on 4-(3-Br om oa n ilin o)-
6,7-d im eth oxyqu in a zolin e (32). In the fourth part
of the SAR study, we looked briefly at a small series of
compounds, based on 32 but either being positional
isomers (66 and 67) or possessing additional substitu-
ents at various positions (61-65, 68, and 69). Placing
an additional OMe group at the 5-position gave a
subnanomolar inhibitor (64, IC50 0.67 nM). By com-
SAR for Qu in a zolin e Su bstitu en ts w ith a 3′-
Br om oa n ilin e Sid e Ch a in . In the third part of the
study, we kept the anilino side chain with the optimal

272 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1
Bridges et al.
parison with the 5,6-diOMe derivative 59, adding a
7-OMe substituent led to a 2000-fold improvement in
IC50. This is very similar to the improvement seen on
going from the 6-OMe to the 6,7-diOMe analogues (7
and 32), indicating that while a 5-OMe cannot act in
concert with the 6-OMe to produce the supra-additive
effect, it cannot prevent its occurrence (due to the 6,7-
diOMe combination). However, in either case the
5-OMe substituent is about 25-fold detrimental. The
8-OMe substituent is intrinsically much more detrimen-
tal (>105-fold loss in affinity between 32 and 65). Thus
the 6,7,8-triOMe derivative 65 and the 7,8-diOMe
analogue 60 both have very low potency (the assay was
not carried to high enough inhibitor concentrations to
establish the separation quantitatively).
Introduction of a methyl group to the 2-position (61)
was enormously deactivating, with at least a 400 000-
fold loss of potency. The corresponding amino com-
pound 62 was still over 18 000-fold less active than 32,
so even small substituents at this position are highly
disfavored in the quinazolines. This is consistent with
our previous results with phthalazines and cinnolines,
and especially with benzotriazines,14 where the loss of
potency upon replacing the 2-position ring methine with
nitrogen was similarly enormous. Methylation of the
N4-linking nitrogen led to a 6000-fold drop in activity
(63, IC50 152 nM). In the parent anilinoquinazoline, this
substitution was also highly disfavored, leading to a 300-
fold loss of affinity.14
F igu r e 1. Double-reciprocal plot for inhibition of EGFR
tyrosine kinase by compound 4. Enzyme activity was deter-
mined as described in the Experimental Section but with the
indicated concentrations of drug and ATP. Lines were deter-
mined by least-squares fitting.
Finally, we looked briefly at the effects of Br substitu-
tion at different positions in the aniline ring. The 4′-
Br substituent is reasonably activating on 6,7-dimethox-
yquinazoline (30-fold increase in potency between 29
and 67) and only moderately deactivating when com-
bined with 3′-Br on the same nucleus (3-fold decrease
in potency between 32 and 68). In contrast both 2′-Br
and 3′,5′-diBr substituents are about 5-fold deactivating
on the same nucleus (compounds 66 and 69 compared
with 29), suggesting limits to the steric tolerance of the
aniline-binding site.
Kin etic a n d Cellu la r Da ta . Figure 1 shows a
double-reciprocal plot for inhibition of the EGFR by 4
with respect to ATP concentration. The converging line
on the y axis, as well as best-fit analysis by nonlinear
regression to Cleland’s equations,21 indicates that 4
behaves as a pure competitive inhibitor to ATP. Similar
experiments with the peptide as the varying substrate
(data not shown) indicated noncompetitive inhibition.
Since tight-binding inhibitors do not conform to the
equations derived for steady-state enzyme kinetics, a
similar analysis could not be performed for 32. A direct
demonstration that this derivative is competitive with
ATP was therefore not feasible, and other mechanisms
cannot be ruled out, but it seems reasonable that the
close analogues 4 and 32 have a similar mechanism.
F igu r e 2. Time course for reversal of inhibition of EGFR
autophosphorylation. Cells were treated with 1 µM 4 or 32
for 2 h and then washed free of drug. The cells were then
exposed to EGF (100 ng/mL) for 5 min at varying times after
the wash, and EGFR autophosphorylation was quantified by
Western blotting as described in the Experimental Section.
compound, and then exposed to EGF for 5 min and the
level of EGFR autophosphorylation quantitated at vary-
ing times. As would be expected for reversible inhibi-
tors, there is a slow return of autophosphorylation
activity, but this has a half-life of ca. 30 min following
inhibition by 4 but becomes much longer (ca. 4 h)
following inhibition by 32. Data to be reported else-
where (Mark Kunkel, personal communication) dem-
onstrates that this reappearance of autophosphorylating
activity is due to recovery of the original receptor
population rather than synthesis of new protein. These
data demonstrate that, while 32 is still a reversible
inhibitor as would be expected from its structure, it has
a very slow off-rate (consistent with very tight binding).
Or igin s of th e Su p r a -a d d itive P oten tia tion Seen
w ith Som e Su bstitu en t P a tter n s. Although devia-
tions from strict additivity for substituent effects are
Compounds 4 and 32 inhibited EGF-stimulated re-
ceptor autophosphorylation in A431 human epidermoid
carcinoma cells, with IC50s of 126 and 14 nM, respec-
tively. Time course experiments indicated that inhibi-
tion of receptor autophosphorylation occurs immediately
after the cells are exposed to either compound (data not
shown), but there is a marked difference in the recovery
of this property after the compound is washed out.
Figure 2 shows an experiment in which cells were
exposed to either 4 or 32 (1 µM) for 2 h, washed free of

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1 273
not uncommon (e.g., when substituents of two different
parts of the molecule both bind in the same receptor
pocket), the enormous potentiation shown by a subset
of the compounds in this SAR is unusual. The most
obvious explanation here is that particular combinations
of substituents produce some dramatic steric or elec-
tronic change in the compounds in question. However,
while there is probably good electronic communication
between the quinazoline and aniline rings, as previously
shown for 9-anilinoacridines,22 a comparison of com-
pounds which do and do not show the supra-additive
effect makes any kind of unusual electronic “cross talk”
between the two rings most unlikely. For example,
electronic differences between the 6-OMe and 6-NH2
3-bromoanilino compounds 8 and 11 are likely to be
minimal. It is also difficult to see how steric interac-
tions could play any part in this effect. Diarylamines
such as these have only two rotatable bonds and will
maintain a large dihedral angle between the rings.
Additionally, the effect occurs with substituents which
are well away from any possible interactions with the
other ring, and, as discussed above, in the most likely
case where a steric effect between the two rings could
occur (compounds 59 and 64), the supra-additive effect
of the favorable groupings appears to still be present.
inhibitor.24,25 In the present example, all of the data
(see also below) can be explained by this concept, where
the excess binding energy for the supra-additive com-
pounds in effect comes from the protein, not the inhibi-
tor. Under this hypothesis, the reason that the SARs
are obviously discontinuous is that the inhibitors are
binding to two different species and that intercompari-
sons are therefore inappropriate. Although no direct
evidence is available as to the nature of the induced
conformation changes with EGFR, the known X-ray
structures of the closely homologous enzymes MAP
kinase26 and the insulin receptor kinase23 suggest a
possible scenario. In unactivated form, both of these
show open conformations with the ATP-binding site
completely exposed to solvent on one face of the N-
terminal lobe, many angstroms from the catalytic
machinery in the C-terminal lobe. In contrast, the X-ray
structure of protein kinase A in active conformation,
bound to both a substrate analogue (PK1) and ATP,
shows that the lobes have fully closed down around the
hinge and the domain surfaces have slid with respect
to one another in the axis perpendicular to the major
axis, trapping ATP completely inside the enzyme.27 We
suggest that the 4-anilinoquinazoline inhibitors dis-
cussed here differ in their ability to induce these kinds
of motion in the EGFR protein. The very slow off-rates
of the most potent inhibitors might then be a result of
hydrophobic collapse of the enzyme in its most stable
active conformation around an optimally shaped inhibi-
tor, which now becomes a component of the hydrophobic
core of the enzyme/inhibitor complex.
Another possible explanation is that the supra-addi-
tive combinations of substituents produce a change in
binding mode of the inhibitor on the enzyme. Recent
crystallographic data for the catalytic domain of the
closely related insulin receptor show that the ATP-
binding site is situated on the N-terminal lobe of the
protein, where there is sufficient space to accommodate
different inhibitor-binding modes.23 However, while
this hypothesis might explain why two activating sub-
stituents can combine together supra-additively on the
enzyme, by binding to alternative acceptor sites, it is
not adequate to account for the fact that two deactivat-
ing substituents can participate supra-additively with
some of the already defined activating substituents (e.g.,
11, 22, and 34). If, for example, there are two different
binding modes available to accommodate the 3′-CF3 or
6-NH2 group of the monosubstituted compounds 6 and
9, the inhibitor should adopt the more favorable of these
if it is capable of doing so. Normally this could only be
prevented by an unfavorable steric interaction else-
where in the molecule. However, here we see low
activity (due to putative inhibition of occupancy) for the
two monosubstituted compounds, before steric bulk (the
other substituent in either case) is added. Therefore,
we feel that this argument is also untenable and that
one can only conclude that the reason that the parent
3′-CF3 and 6-NH2 compounds do not utilize a more
favourable binding mode on the enzyme is that it is not
there for those compounds.
Con clu sion s
The above results for a large number of close ana-
logues of the potent EGFR tyrosine kinase inhibitor 32
(PD 153035) show that the high potency of 32 is not
unique, with derivatives possessing appropriate com-
binations of phenyl and quinazoline substitutions hav-
ing similar IC50s for inhibition of the isolated enzyme.
Generally these highly potent analogues have IC50s up
to about 80-fold better than predicted from simple
additivity arguments. However, other very similar
compounds do not show this “supra-additive” effect
when combined together on the anilinoquinazoline
nucleus, and some substituents which are mildly deac-
tivating by themselves can be strongly activating when
used in the correct combinations. To explain these
results, we propose that certain appropriately disubsti-
tuted analogues possess the ability to induce a change
in the conformation of the tyrosine kinase domain when
they bind. Using this hypothesis, some of the results
can be interpreted to indicate differences between the
“native” and “induced” binding sites. For example, there
appears to be greater bulk tolerance at the quinazoline
6- and 7-positions at the induced binding site but less
at the N4 site. Because of the former property, 32 is
not the optimal inhibitor in the induced conformation,
with the diethoxy analogue 56 being superior. This
compound, 4-(3-bromoanilino)-6,7-diethoxyquinazoline,
has an IC50 of 0.006 nM, making it the most potent
quinazoline inhibitor (and, to our knowledge, the most
potent inhibitor of any type yet reported) of the tyrosine
kinase activity of the EGFR.
In our view the most reasonable explanation is that
the enzyme itself undergoes a conformational change
when binding the inhibitors which contain supra-
additive combinations of substituents. This concept of
“induced fit” of ligands to protein substrates has been
well-validated and is dramatically illustrated with the
binding of the experimental antiviral drug disoxiril to
the capsid proteins of picornavirus. Here, the X-ray
crystallographic structure of the native virus shows no
binding site for disoxiril, but the drug still binds, with
the X-ray structure of the complexed virus showing how
the protein alters conformation to accommodate the
Exp er im en ta l Section
Analyses were performed by the Microchemical Laboratory,
University of Otago, Dunedin, NZ, or Parke-Davis Pharma-
ceutical Research Analytical Department. Melting points were

274 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1
Bridges et al.
determined using an Electrothermal Model 9200 or Gallen-
kamp digital melting point apparatus and are as read. NMR
spectra were measured on Bruker AC-200 or AM-400 or Varian
Unity 400 MHz spectrometers and referenced to Me4Si. Mass
spectra were recorded on a Varian VG 7070 spectrometer at
nominal 5000 resolution or a Finnigan MAT 900Q spectrom-
eter. Unless otherwise noted (below), the compounds of Table
1 were prepared from known28 quinazolin-4(3H)-ones by
conversion to the corresponding 4-chloroquinazolines with
POCl3, followed by reaction with substituted anilines in
2-propanol.14
4-[(3-Br om op h en yl)a m in o]-7-ch lor o-6-n it r oq u in a zo-
lin e (53): Gen er a l Exa m p le of Cou p lin g P r oced u r e
(Sch em e 1). A solution of 7-chloro-6-nitro-4-quinazolin-4(3H)-
one (70)16 (7.6 g, 33.7 mmol) in SOCl2 (100 mL) containing 2
drops of DMF was heated under reflux for 1 h. Excess SOCl2
was removed under reduced pressure, and the residue was
dissolved in CH2Cl2 and washed with aqueous Na2CO3 solu-
tion. Removal of the solvent and recrystallization from light
petroleum gave 4,7-dichloro-6-nitroquinazoline (71) (8.2 g, 99%
yield): mp 197-210 °C (semimelt at 139-141.5 °C); 1H NMR
[(CD3)2SO] δ 8.60 (s, 1 H, ArH), 8.27 (s, 1 H, ArH), 7.85 (s, 1
H, ArH); 13C NMR δ 159.0, 150.3, 144.8, 131.1, 121.4 (C), 149.2,
129.4, 124.2 (CH). Anal. (C8H3Cl2N3O3) C, H, N, Cl.
A mixture of 71 (7.9 g, 32 mmol) and 3-bromoaniline (7.9
mL, 70 mmol) in iPrOH (250 mL) was heated to reflux and
treated with 3 drops of concentrated HCl. Heating was
continued for 30 min, sufficient Et3N was added to basify the
solution, and the solvent was then concentrated to give 4-[(3-
bromophenyl)amino]-7-chloro-6-nitroquinazoline (53) (11.1 g,
91% yield): mp (EtOH) 257-259 °C dec; 1H NMR [(CD3)2SO]
δ 10.36 (s, 1 H, NH), 9.43 (s, 1 H, H-5), 8.77 (s, 1 H, H-2), 8.16
(br s, 1 H, H-2′), 8.11 (s, 1 H, H-8), 7.86 (br d, J ) 7.1 Hz, 1 H,
H-6′), 7.42-7.36 (m, 2 H, H-4′,5′). Anal. (C14H8BrClN4O2) C,
H, N.
one29 (74) was reacted sequentially with POCl3 and
3-bromoaniline as above to give 4-[(3-bromophenyl)amino]-7-
fluoroquinazoline (75): mp (MeOH) 201-203 °C; 1H NMR
[(CD3)2SO] δ 9.96 (br s, 1 H, NH), 8.68-8.64 (m, 1 H, H-5),
8.66 (s, 1 H, H-2), 8.21 (br s, 1 H, H-2′), 7.89 (d, J ) 7.9 Hz,
H-6′), 7.63-7.55 (m, 2 H, H-6,8), 7.37 (t, J ) 7.9 Hz, 1 H, H-5′),
7.32 (br d, J ) 7.9 Hz, 1 H, H-4′). Anal. (C14H9BrFN3) C, H,
N.
A mixture of 75 (1.0 g, 3.1 mmol), methylamine hydrochlo-
ride (20 g, 0.3 mol), and Et3N (30 mL, 0.3 mol) in N-
methylpyrrolidone (50 mL) was heated at 130 °C in a sealed
pressure vessel for 4 days, and the resulting mixture was
partitioned between EtOAc and water. The organic layer was
dried, the solvent was removed, and the residue was chro-
matographed on alumina gel, eluting with CH2Cl2, to give
4-[(3-bromophenyl)amino]-7-(methylamino)quinazoline (40) (0.54
g, 52%): mp (CH2Cl2) 198-199 °C; 1H NMR [(CD3)2SO] δ 9.45
(s, 1 H, NH), 8.43 (s, 1 H, H-2), 8.22 (br s, 1 H, H-2′), 8.19 (d,
J ) 9.1 Hz, 1 H, H-5), 7.86 (br d, J ) 8.2 Hz, 1 H, H-6′), 7.31
(t, J ) 8.1 Hz, 1 H, H-5′), 7.23 (br d, J ) 7.9 Hz, 1 H, H-4′),
6.96 (dd, J ) 9.1, 2.3 Hz, 1 H, H-6), 6.67 (t, J ) 4.9 Hz, 1 H,
NHMe), 6.57 (d, J ) 2.3 Hz, 1 H, H-8), 2.81 (d, J ) 4.9 Hz, 3
H, NCH3). Anal. (C15H13BrN4) C, H, N.
Similarly prepared were 41 and 42.
7-Acet a m id o-4-[(3-b r om op h en yl)a m in o]q u in a zolin e
(39). Treatment of 30 with Ac2O/pyridine gave 7-acetamido-
4-[(3-bromophenyl)amino]quinazoline (39): mp (EtOH) 351-
353 °C; 1H NMR [(CD3)2SO] δ 10.42 (s, 1 H, NH), 9.76 (s, 1 H,
NH), 8.59 (s, 1 H, H-2), 8.46 (d, J ) 9.0 Hz, 1 H, H-5), 8.24 (s,
1 H, H-8), 8.15 (br s, 1 H, H-2′), 7.90 (d, J ) 7.8 Hz, 1 H, H-6′),
7.74 (d, J ) 8.5 Hz, 1 H, H-6), 7.35 (t, J ) 8.0 Hz, 1 H, H-5′),
7.29 (d, J ) 7.87 Hz, 1 H, H-4′), 2.14 (s, 3 H, CH3). Anal.
(C16H13BrN4O) C, H, N.
4-[(3-Br om op h e n yl)a m in o]-6,7-d im e t h oxyq u in a zo-
lin e Hydr och lor ide (32) (Sch em e 3). 2-Amino-4,5-dimethox-
ybenzoic acid (76) (29.58 g, 0.15 mol) and formamidine
hydrochloride (20.20 g, 0.251 mol) were intimately ground
together and then spread in an even layer around the bottom
of a 1 L round bottom flask. This was blanketed with N2 under
an air condenser and rapidly heated to 210 °C on an oil bath.
During heating a circular melting zone followed by resolidi-
fication passed from the outside to the center of the flask. After
15 min at 210 °C, the reaction mixture was allowed to cool to
80 °C and then sonicated with dilute NaOH solution (0.33 M,
150 mL). The resulting purple-gray solid was collected, rinsed
with water (3 × 50 mL), and dried under vacuum at 70 °C to
give 6,7-dimethoxyquinazolin-4(3H)-one (78) (24.50 g, 79%),
mp 295-298 °C, which was used directly: 1H NMR [(CD3)2-
SO] δ 12.05 (br s, 1 H, NH), 7.99 (s, 1 H, H-2), 7.44 (s, 1 H,
H-5), 7.12 (s, 1 H, H8), 3.90, 3.87 (2s, 2 × 3 H, 2 × OCH3).
DMF (13.16 g, 0.18 mol) was added dropwise over 20 min
to a stirred solution of oxalyl chloride (22.85 g, 0.18 mol) in
1,2-dichloroethane (120 mL) under N2 at 25 °C, resulting in
an exotherm and gas evolution. When gas evolution ceased,
78 (24.76 g, 0.12 mol) was added with mechanical agitation,
the mixture was heated to reflux for 4 h and cooled to 25 °C,
then the reaction was quenched with dilute aqueous Na2HPO4
solution (0.5 M, 250 mL). The resulting mixture was stirred
on an ice bath for 2 h, and the solid was collected, rinsed with
water (2 × 50 mL), and dried at 50 °C under vacuum to give
4-chloro-6,7-dimethoxyquinazoline (80) (22.12 g, 82%) which
was used directly: 1H NMR (CDCl3) δ 8.86 (s, 1 H, H-2), 7.38
(s, 1 H, H-5), 7.33 (s, 1 H, H-8), 4.07 (s, 6 H, OCH3).
Reaction of 80 (22.12 g, 98 mmol) and 3-bromoaniline (16.34
g, 95 mmol) in stirred 2-propanol (1 L), under N2 for 18 h gave
4-[(3-bromophenyl)amino]-6,7-dimethoxyquinazoline hydro-
chloride (32) (35.95 g, 92%): mp 264-266 °C; 1H NMR [(CD3)2-
SO] δ 11.61 (br s, 1 H, NH), 8.88 (s, 1 H, H-2), 8.43 (s, 1 H,
H-5), 8.04 (s, 1 H, H-2′), 7.80 (br d, J ) 7.4 Hz, 1 H, H-4′),
7.45 (br d, J ) 8 Hz, 1 H, H-6′), 7.44 (br t, J ) 7.5 Hz, 1 H,
H-5′), 7.38 (s, 1 H, H-8), 4.03, 4.00 (2s, 2 × 3 H, 2 × OCH3).
Anal. (C16H14BrN3O2‚HCl) C, H, N.
4-[(3-Br om op h en yl)a m in o]-5,6,7-t r im et h oxyq u in a zo-
lin e (64) (Sch em e 3). 2,3,4-Trimethoxy-6-nitrobenzoic acid30
(7.72 g, 30 mmol) in MeOH (150 mL) was hydrogenated over
Pd/C (600 mg). After 17 h, TLC showed 70% conversion, a
further 200 mg of catalyst was added, and the reaction was
4-[(3-Br om op h en yl)a m in o]-7-(d im eth yla m in o)-6-n itr o-
qu in a zolin e (50): Gen er a l Exa m p le of Ch lor o Disp la ce-
m en t. A solution of 4-[(3-bromophenyl)amino]-7-chloro-6-
nitroquinazoline (53) (1.0 g, 26 mmol) in EtOH (100 mL) was
heated at 100 °C with 40% aqueous dimethylamine (20 mL)
in a closed pressure vessel for 1 h to give, on cooling, 4-[(3-
bromophenyl)amino]-7-(dimethylamino)-6-nitroquinazoline (50)
(0.75 g, 74% yield): mp (EtOH) 240.5-243.5 °C; 1H NMR
[(CD3)2SO] δ 10.01 (s, 1 H, NH), 9.12 (s, 1 H, H-5), 8.58 (s, 1
H, H-2), 8.19 (s, 1 H, H-2′), 7.89 (d, J ) 8.0 Hz, 1 H, H-6′),
7.36 (m, 2 H, H-4′,5′), 7.18 (s, 1 H, H-8), 2.93 (s, 6 H, NMe2).
Anal. (C16H14BrN5O2) C, H, N.
Similar reaction of 53 with methylamine gave compound
49.
4-[(3-Br om op h en yl)a m in o]-7-m eth oxy-6-n itr oqu in a zo-
lin e (52). A solution of 70 (1.0 g, 4.4 mmol) in MeOH (150
mL) containing dissolved sodium metal (1.01 g, 44 mmol) was
heated at 100 °C in a sealed pressure vessel for 18 h. The
solution was neutralized with AcOH and diluted with water
to give 7-methoxy-6-nitroquinazolin-4(3H)-one (72) (0.95 g,
97%): mp (EtOH) 287 °C dec; 1H NMR [(CD3)2SO] δ 12.51 (s,
1 H, NH), 8.53 (s, 1 H), 8.24 (s, 1 H), 7.43 (s, 1 H), 4.05 (s, 3 H,
OCH3). Anal. (C9H7N3O4) C, H, N.
A suspension of 72 (0.85 g, 3.8 mmol) in POCl3 (50 mL) was
heated under reflux for 30 min, when a clear solution was
obtained. The POCl3 was removed under reduced presssure,
and the residue was dissolved in a mixture of CH2Cl2 and
aqueous Na2CO3. The organic layer was dried and the solvent
removed to give 4-chloro-7-methoxy-6-nitroquinazoline (73)
1
(0.86 g, 93%): mp (hexane) 172-173 °C; H NMR (CDCl3) δ
9.06 (s, 1 H, H-5), 8.65 (s, 1 H, H-2), 7.56 (s, 1 H, H-8), 4.13 (s,
3 H, OMe). Anal. (C9H6ClN3O3) C, H, N, Cl.
Reaction of 73 with with 3-bromoaniline in iPrOH as above
gave 4-[(3-bromophenyl)amino]-7-methoxy-6-nitroquinazoline
(52) (86% overall yield): mp (EtOAc/hexane) 206-207.5 °C;
1H NMR [(CD3)2SO] δ 10.13 (s, 1 H, NH), 9.25 (s, 1 H, H-5),
8.70 (s, 1 H, H-2), 8.18 (br s, 1 H, H-2′), 7.88 (d, J ) 7.7 Hz, 1
H, H-6′), 7.49 (s, 1 H, H-8), 7.40-7.33 (m, 2 H, H-4′,5′), 4.07
(s, 3 H, OCH3). Anal. (C15H11BrN4O3) C, H, N.
4-[(3-Br om op h en yl)a m in o]-7-(m et h yla m in o)qu in a zo-
lin e (40): Exa m p le of F lu or o Disp la cem en t fr om Un a c-
tiva ted Qu in a zolin e (Sch em e 2). 7-Fluoroquinazolin-4(3H)-

Tyrosine Kinase Inhibitors
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1 275
continued for a further 3.5 h. The suspension was filtered over
Celite, and the filtrate was concentrated to near dryness. The
resulting crude air-sensitive 6-amino-2,3,4-trimethoxybenzoic
acid (77) was used immediately, by dissolving in MeOH (125
mL) and treating with s-triazine (2.505 g, 30 mmol) and
piperidine (2.1 mL) according to the procedure of Kreutzberger
and Uzbek.31 The solution was heated under N2 at 65 °C for
2.5 h and then concentrated to a solid residue that was purified
by flash silica gel column chromatography. Elution with CH2-
Cl2 and then CH2Cl2/MeOH (9:1), followed by pooling and
evaporation of appropriate fractions, resulted in a solid which
was triturated in hot EtOAc to give 5,6,7-trimethoxyquinazo-
lin-4(3H)-one (79) (5.00 g, 72%): mp 187-189° C; 1H NMR
(CDCl3) δ 8.15 (br s, exchanges with D2O, 1 H, NH), 7.27 (s, 1
H), 7.06 (s, 1 H), 4.02 (s, 3 H), 4.00 (s, 3 H), 3.95 (s, 3 H); CIMS
(1% NH3 in CH4) m/ z 221 (18), 237 (100, MH+). Anal.
(C11H12N2O4) C, H, N.
A suspension of 79 (2.36 g, 10 mmol) and POCl3 (0.98 mL,
10.5 mmol) in 1,2-dichloroethane (20 mL) was treated dropwise
with diisopropylethylamine (4.4 mL, 25 mmol). The resulting
solution was heated at 80 °C under N2 for 1.5 h and then cooled
and concentrated to a solid that was diluted with CH2Cl2. The
solution was washed with 5% aqueous NaHCO3, dried (Mg-
SO4), and then filtered through a short column of silica gel,
washing with CH2Cl2 and then CH2Cl2/EtOAc (3:1), to give
4-chloro-5,6,7-trimethoxyquinazoline (81) (1.29 g, 51%): mp
(tert-butyl methyl ether) 115-116 °C; 1H NMR (CDCl3) δ 8.84
(s, 1 H), 7.23 (s, 1 H), 4.05 (s, 3 H), 4.03 (s, 3 H), 4.00 (s, 3 H);
CIMS (1% NH3 in CH4) m/ z 250 (31), 251 (100), 279 (12). Anal.
(C11H11ClN2O3) C, H, Cl, N.
CHCl3/MeOH, and volatiles were removed under reduced
pressure to give 4-[(3-bromophenyl)amino]-6,7-diethoxyquinazo-
line monohydrate (56) (43 mg, 42%) as a glass: mp 155-167
1
°C; H NMR [(CD3)2SO] δ 9.75 (br s, 1 H, NH), 8.56 (s, 1 H,
H-2), 8.12 (m, 1 H, H-2′), 7.87 (s, 1 H, H-5), 7.86 (d, J ) 8.0
Hz, H-6′), 7.37 (t, J ) 8.0 Hz, H-5′), 7.31 (d, J ) 8.4 Hz, 1 H,
H-4′), 7.19 (s, 1 H, H-8), 4.24, 4.22 (2q, J ) 6.7 Hz, 2 H, OCH2),
1.45, 1.42 (2t, J ) 6.7 Hz, 3 H, CH3); CIMS m/ z 388 (100,
79BrMH+). Anal. (C18Hl8N3BrO2‚H2O) C, H, N.
4-[(3-Br om op h e n yl)a m in o]-6,7-d im e t h oxy-2-m e t h -
ylqu in a zolin e Hyd r och lor id e (61) (Sch em e 4). HCl gas
was bubbled through
a solution of methyl 2-amino-4,5-
dimethoxybenzoate (82) (0.210 g, 0.99 mmol) in CH3CN (10
mL) for 1 h, and the resulting mixture was heated under reflux
for 8 h and then cooled to room temperature. The resulting
solid was collected and dissolved in water (10 mL), and the
solution was neutralized with saturated aqueous NaHCO3
solution to pH 7. Filtration and air-drying of the resulting
solid gave 6,7-dimethoxy-2-methylquinazolin-4(3H)-one (83)
(152 mg, 70%), which was used directly: 1H NMR [(CD3)2SO]
δ 7.39 (s, 1 H, H-5), 7.05 (s, 1 H, H-8), 3.87 (s, 3 H, OCH3),
3.84 (s, 3 H, OCH3), 2.30 (s, 3 H, CH3).
Reaction of 83 (152 mg, 0.70 mmol) and POCl3 (3 mL), under
reflux for 8 h, followed by removal of the excess reagent POCl3
under reduced pressure and partition of the residue between
CHCl3 and water gave 4-chloro-6,7-dimethoxy-2-methylquinazo-
line (85) (93 mg, 56%) as a white solid which was used
directly: 1H NMR [(CD3)2SO] δ 7.38 (s, 1 H, H-5), 7.36 (s, 1 H,
H-8), 3.99 (s, 3 H, OCH3), 3.97 (s, 3 H, OCH3), 2.67 (s, 3 H,
CH3).
Reaction of this with 3-bromoaniline in 2-methoxyethanol
at 90 °C for 30 min gave 4-[(3-bromophenyl)amino]-6,7-
dimethoxy-2-methylquinazoline hydrochloride (61) (129 mg,
80%): mp 264-265 °C; 1H NMR [(CD3)2SO] δ 8.08 (br s, 1 H,
H-5), 8.00 (br s, 1 H, H-8), 7.81 (d, J ) 6.6 Hz, 1 H, H-6′),
7.48-7.44 (m, 2 H, H-4′,5′), 7.18 (s, 1 H, H-2′), 4.0 (s, 3 H,
OCH3), 4.0 (s, 3 H, OCH3), 2.61 (s, 3 H, CH3); CIMS m/ z 373
(100, 79BrMH+). Anal. (C17H16BrN3O2‚HCl) C, H, N.
Reaction of 81 with 3-bromoaniline in 2-propanol under
reflux for 1 h as above gave 4-[(3-bromophenyl)amino]-5,6,7-
1
trimethoxyquinazoline (64) (82%): mp 130-132 °C; H NMR
(CDCl3) δ 9.92 (br s, exchanges with D2O, 1 H), 8.61 (s, 1 H),
8.11 (s, 1 H), 7.71-7.67 (m, 1 H, H-5′), 7.26-7.24 (m, 2 H,
H-4′,6′), 7.08 (s, 1 H, H-8), 4.17 (s, 3 H), 4.01 (s, 3 H), 3.96 (s,
3 H); CIMS (1% NH3 in CH4) m/ z 310 (65), 312 (21), 390 (100,
MH+), 392 (83), 418 (19), 420 (18). Anal. (C17H16BrN3O3) C,
H, N.
2-Am in o-4-[(3-b r om op h e n yl)a m in o]-6,7-d im e t h ox-
yqu in a zolin e Tr ih yd r och lor id e (62) (Sch em e 4). Con-
centrated HCl (0.05 mL) was added dropwise to a solution of
methyl 2-amino-4,5-dimethoxybenzoate (82) (0.217 g, 1.0
mmol) and cyanamide (67 mg, 1.6 mmol) in dioxane (10 mL)
at 25 °C. The mixture was heated at 80 °C for 7.5 h and then
cooled to 25 °C, and the resulting solid was collected and
treated as above to give 2-amino-6,7-dimethoxyquinazolin-
4(3H)-one (84) (174 mg, 79%), which was used directly: 1H
NMR [(CD3)2SO] δ 7.23 (s, 1 H, H-5), 6.67 (s, 1 H, H-8), 3.82
(s, 3 H, OCH3), 3.77 (s, 3 H, OCH3), 6.14 (br, 2 H, NH2).
Reaction of 84 with POCl3 as above gave 2-amino-4-chloro-
6,7-dimethoxyquinazoline (86) (38%) as a yellow solid which
was used directly: 1H NMR [(CD3)2SO] δ 7.14 (s, 1 H, H-5),
6.89 (s, 1 H, H-8), 3.90 (s, 3 H, OCH3), 3.86 (s, 3 H, OCH3).
Reaction of this with 3-bromoaniline as above gave 2-amino-
4-[(3-bromophenyl)amino]-6,7-dimethoxyquinazoline trihydro-
chloride (62) (7%): mp 262-263 °C; 1H NMR [(CD3)2SO] δ 9.44
(br s, 1 H, NH), 8.03 (d, J ) 2.9 Hz, 2 H, H-2′,6′), 7.70 (s, 1 H,
H-5), 7.35-7.27 (m, 2 H, H-4′,5′), 6.82 (s, 1 H, H-8), 6.56 (br s,
2 H, NH2), 3.88 (s, 3 H, OCH3), 3.86 (s, 3 H, OCH3); CIMS
m/ z 374 (100, 79BrMH+). Anal. (C16H15BrN4O2‚3HCl‚0.25H2O)
C, H, N.
En zym e Assa y. EGFR was prepared from human A431
carcinoma cell-shed membrane vesicles by immunoaffinity
chromatography as previously described,32 and the assays were
carried out as reported previously.13 The substrate used was
based on a portion of phospholipase Cγ1, having the sequence
Lys-His-Lys-Lys-Leu-Ala-Glu-Gly-Ser-Ala-Tyr472-Glu-Glu-
Val. The reaction was allowed to proceed for 10 min at room
temperature and then stopped by the addition of 2 mL of 75
mM phosphoric acid. The solution was then passed through
a 2.5 cm phosphocellulose disk which bound the peptide. This
filter was washed with 75 mM phosphoric acid (5×), and
incorporated label was assessed by scintillation counting in
an aqueous fluor. Control activity (no drug) gave a count of
ca. 100 000 cpm. At least two independent dose-response
curves were done and the IC50 values computed. The reported
values are averages; variation was generally (15%.
4-[(3-Br om op h e n yl)a m in o]-6,7-d ih yd r oxyq u in a zo-
lin e Hyd r och lor id e (55) (Sch em e 3). A mixture of 32 (198
mg, 0.5 mmol) and pyridinium hydrochloride (1.15 g, 10 mmol)
was heated together under N2 with stirring at 205 °C for 1 h.
Upon cooling, the fused melt was sonicated with water (20 mL),
and the residual solid was collected to give 4-[(3-bromophenyl)-
amino]-6,7-dihydroxyquinazoline hydrochloride monohydrate14
(55) (25 mg, 13%): mp (EtOH at 0 °C) >320 °C.
4-[(3-Br om op h en yl)a m in o]-6-(m et h yla m in o)qu in a zo-
lin e (35). A solution of 6-amino-4-[(3-bromophenyl)amino]-
quinazoline14 (11) (2.25 g, 7.14 mmol) in pyridine (20 mL) at
0 °C was treated dropwise with MeOCOCl (1.0 mL, 13 mmol),
and the mixture was allowed to warm to room temperature
over 2 h. Addition of water gave a precipitate of 4-[(3-
bromophenyl)amino]-6-[(methoxycarbonyl)amino]quinazo-
line (37) (2.34 g, 98%): mp (EtOH) 197-198.5 °C; 1H NMR
[(CD3)2SO] δ 10.00 (s, 1 H, NH), 9.88 (s, 1 H, NH), 8.56 (s, 1
H, H-2), 8.54 (br s, 1 H, H-5), 8.15 (br s, 1 H, H-2′), 7.85 (br d,
J ) 7.9 Hz, 1 H, H-6′), 7.77 (br s, 2 H, H-7,8), 7.34 (t, J ) 8.0
Hz, 1 H, H-5,5′), 7.28 (br d, J ) 8.2 Hz, 1 H, H-4′), 3.74 (s, 3
H, OCH3). Anal. (C16H13BrN4O2) C, H, N.
A solution of 37 (1.0 g, 3 mmol) in dry THF (30 mL) was
reacted with an excess of LiAlH4 at reflux for 1 h. The mixture
was neutralized with H2O, the solvent was removed, and the
residue was extracted into EtOAc. Chromatography on Al2O3,
eluting with CH2Cl2, gave 4-[(3-bromophenyl)amino]-6-(me-
thylamino)quinazoline (35) (0.24 g, 24%): mp (MeOH/H2O)
141-144 °C; 1H NMR [(CD3)2SO] δ 9.41 (s, 1 H, NH), 8.39 (s,
1 H, H-2), 8.19 (br s, 1 H, H-2′), 7.93 (br d, J ) 6.9 Hz, 1 H,
H-6′), 7.55 (d, J ) 9.0 Hz, 1 H, H-8), 7.35 (t, J ) 8.0 Hz, 1 H,
H-5′), 7.28-7.25 (m, 2 H, H-7,4′), 7.14 (d, J ) 2.3 Hz, 1 H,
H-5), 6.32 (q, J ) 5.0 Hz, 1 H, NHMe), 2.86 (d, J ) 5.0 Hz, 3
H, CH3). Anal. (C15H13BrN4) C, H, N.
4-[(3-Br om op h en yl)a m in o]-6,7-d iet h oxyq u in a zolin e
(56). EtI (80 mL, 1 mmol) was added to a stirred suspension
of crude 55 (81 mg, 0.25 mmol) and powdered K2CO3 (138 mg,
1.0 mmol) in DMSO (1 mL) under N2 at 25 °C. After 4 h, the
reaction mixture was put on a vacuum line overnight to remove
all volatiles, and the residue was slurried in EtOH. Prepara-
tive TLC on silica gel, eluting twice with 4% CHCl3/MeOH
(19:1), gave a major band (Rf 0.53) that was extracted with

276 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 1
Bridges et al.
(16) Leonard, N. J .; Morrice, A. G.; Sprecker, M. A. Linear benzoad-
enine. A stretched-out analog of adenine. J . Org. Chem. 1975,
40, 356-363.
(17) Yoneda, T.; Lyall, R. M.; Alsina, M. M.; Persons, P. E.; Spada,
A. P.; Levitski, A.; Zilberstein, A.; Mundy, G. R. The antiprolif-
erative effects of tyrosine kinase inhibitors tyrphostins on a
human squamous cell carcinoma in vitro and in nude mice.
Cancer Res. 1991, 51, 4430-4435.
(18) Traxler, P. M.; Wacker, O.; Bach, H. L.; Geissler, J . F.; Kump,
W.; Meyer, T.; Regenass, U.; Roesel, J . L.; Lydon, N.
Sulfonylbenzoyl-nitrostyrenes: potential bisubstrate type inhibi-
tors of the EGF-receptor tyrosine protein kinase. J . Med. Chem.
1991, 34, 2328-2337.
(19) Umezawa, H.; Imoto, M.; Sawa, T.; Isshiki, K.; Matsuda, N.;
Uchida, T.; Iinuma, H.; Hamada, M.; Takeuchi, T. Studies on a
new epidermal growth factor-receptor kinase inhibitor, erbstatin,
produced by MH435-hF3. J . Antibiot. 1986, 39, 170-173.
(20) Morley, J . S.; Simpson, J . C. E. The chemistry of simple
heterocyclic systems. Part 1. Reactions of 6- and 7-nitro-4-
hydroxyquinazoline and their derivatives. J . Chem. Soc. 1948,
360-366.
(21) Cleland, W. W. Statistical analysis of enzyme kinetic data.
Methods Emzymol. 1979, 103-138.
(22) Denny, W. A.; Atwell, G. J .; Cain, B. F. Potential Antitumor
Agents. 26. Anionic congeners of the 9-anilinoacridines. J . Med.
Chem. 1978, 21, 5-10.
(23) Hubbard, S. R.; Wei, L.; Ellis, L.; Hendrickson, W. A. Crystal
structure of the tyrosine kinase domain of the human insulin
receptor. Nature 1994, 372, 746-754.
(24) Rossmann, M. G.; Arnold, E.; Erickson, J . W.; Frankenberger,
E. A.; Griffith, J . P.; Hecht, H.-J .; J ohnson, J . E.; Kamer, G.;
Luo, M.; Mosser, A. G.; Rueckert, R. R.; Sherry, B.; Vriend, G.
Structure of a human common cold virus and functional rela-
tionship to other picornaviruses. Nature 1985, 317, 145-153.
(25) Smith, T. J .; Kremer, M. J .; Luo, M.; Vriend, G.; Arnold, E.;
Kamer, G.; Rossmann, M. G.; McKinlay, M. A.; Diana, G. D.;
Otto, M. J . The site of attachment in human rhinovirus 14 for
antiviral agents that inhibit uncoating. Science 1986, 233, 1286-
1293.
EGF R Au top h osp h or yla tion in Hu m a n A431 Ep id er -
m oid Ca r cin om a Cells. Cells were grown to confluency in
6-well plates (35 mm diameter), exposed to serum-free medium
for 18 h, and then treated with 4 or 32 for 2 h, followed by
EGF (100 ng/mL) for 5 min. The monolayers were lysed in
0.2 mL of boiling Laemlli buffer (2% sodium dodecyl sulfate,
5% 2-mercaptoethanol, 10% glycerol, and 50 mM Tris, pH 6.8),
and the lysates were heated to 100 °C for 5 min. Proteins in
the lysate were separated by polyacrylamide gel electrophore-
sis and electrophoretically transferred to nitrocellulose. The
membrane was washed once in 10 mM Tris, pH 7.2, 150 mM
NaCl, and 0.01% azide (TNA) and blocked overnight in TNA
containing 3% bovine serum albumin and 1% ovalbumin. The
membrane was blotted for 2 h with antiphosphotyrosine
antibody (UBI, 1 µg/mL in blocking buffer) and then washed
twice in TNA, once in TNA containing 0.05% Tween-20 and
0.05% nonidet P-40, and twice in TNA. The membranes were
then incubated for 2 h in blocking buffer containing 0.1 µCi/
mmol [125I]protein and then washed again as above. After the
blots were dry they were loaded into a film cassette and
exposed to X-AR X-ray film for 1-7 days. Band intensities
were determined with a Molecular Graphics laser densitom-
eter.
Ack n ow led gm en t. This work was supported in part
by the Auckland Division of the Cancer Society of New
Zealand.
Su p p or tin g In for m a tion Ava ila ble: Melting points,
crystallization solvents, and 1H NMR data for the new
compounds of Table 1 (8 pages). Ordering information is given
on any current masthead page.
Refer en ces
(1) Powis, G.; Workman, P. Signalling targets for the development
of cancer drugs. Anti-Cancer Drug Des. 1994, 9, 263-277.
(2) El-Zayat, A. A. E.; Pingree, T. F.; Mock, P. M.; Clark, G. M.;
Otto, R. A.; Von Hoff, D. D. Epidermal growth factor receptor
amplification in head and neck cancer. Cancer J . 1991, 4, 375-
380.
(26) Zhang, F.; Strand, A.; Robbins, D.; Cobb, M. H.; Goldsmith, E.
J . Atomic structure of the MAP kinase ERK2 at 2.3 Å resolution.
Nature 1994, 367, 704-711.
(3) Morishige, K.; Kurachi, H.; Amemiya, K.; Fujita, Y.; Yamamoto,
T.; Miyake, A.; Tanizawa, O. Evidence for the involvement of
transforming growth factor
a and epidermal growth factor
(27) (a) Zheng, J .; Knighton, D. R.; Ten Eyck, L. F.; Karlsson, R.;
Xuong, N.-H.; Taylor, S. S.; Sowadski, J . M. Crystal structure
of the catalytic subunit of cAMP-dependent protein kinase
complexed with MgATP and peptide inhibitor. Biochemistry
1993, 32, 2154-2161. (b) Bossemeyer, D.; Engh, R. A.; Kinzel,
V.; Ponstingl, H.; Huber, R. Phosphotransferase and substrate
binding mechanism of the cAMP-dependent protein kinase
catalytic subunit from porcine heart as deduced from the 2.0 Å
structure of the complex with Mn2+ adenylyl imidodiphosphate
and inhibitor peptide PKI(5-24). EMBO J . 1993, 12, 849-859.
(28) (a) Maillard, J .; Benard, M.; Vincent, M.; Vo-Van-Tri; J olly, R.;
Morin, R.; Benharkarte, Mrs.; Menillet, C. 3H-Quinazolin-4-one
derivatives with antiinflammatory activity. II. Derivatives sub-
stituted in the aromatic nucleus, and related compounds. Chim.
Ther. 1967, 2, 231-239. (b) Chien, P.-L.; Cheng, C. C. Structural
modification of febrifugine. Some methylenedioxy analogs. J .
Med. Chem. 1970, 13, 867-870. (c) Simpson, W. R. J . Bi-4-[1-
(quinazolinyl-4)piperidyls] and bis 4-[1-(quinazolinyl-4)piperidyl-]
alkanes. U.S. Patent 4,011,323, 1977; Chem. Abstr. 1977, 87,
6013z. (d) Takase, Y.; Saeki, T.; Watanabe, N.; Adachi, H.;
Souda, S.; Saito, I. Cyclic GMP phosphodiesterase inhibitors. 2.
Requirement of 6-substitution of quinazoline derivatives for
potent and selective inhibitory activity. J . Med. Chem. 1994, 37,
2106-2111.
(29) Armarego, W. L. F.; Smith, J . I. C. Quinazolines. Part IX.
Covalent hydration in the neutral species of substituted quinazo-
lines. J . Chem. Soc. B 1967, 449-454.
(30) Coppola, G. M.; Schuster, H. F. The chemistry of 2H-3,1-
benzoxazine-2,4(1H)-dione (isatoic anhydride). 21. A mild process
for the preparation of 10-alkyl-9-acridanones and it’s application
to the synthesis of acridone alkaloids. J . Heterocycl. Chem. 1989,
26, 957-964.
(31) Kreutzberger, A.; Uzbek, M. U. Ringschluss zum chinazolon-
(4)-system durch aminomethinylierung. (Synthesis of the 4(3H)-
quinazolinone ring system through aminomethenylation.) Arch.
Pharmacol. 1972, 305, 171-175.
(32) Gill, G. N.; Weber, W. Purification of functionally active epider-
mal growth factor receptor protein using a competitive antago-
nist monoclonal antibody and competitive elution with epidermal
growth factor. Methods Enzymol. 1987, 146, 82-88.
receptor autocrine growth mechanism in primary human ovarian
cancers in vitro. Cancer Res. 1991, 51, 5322-5328.
(4) J ardines, L.; Weiss, M.; Fowble, B.; Greene, M. Neu (c-erbB-2/
HER2) and the epidermal growth factor receptor (EGFR) in
breast cancer. Pathobiology 1993, 61, 268-282.
(5) Hickey, K.; Grehan, D.; Reid, I. M.; O’Brian, S.; Walsh, T. N.;
Hennessy, T. P. J . Expression of epidermal growth factor
receptor and proliferating cell nuclear antigen predicts response
of esophageal squamous cell carcinoma to chemoradiotherapy.
Cancer 1994, 74, 1693-1698.
(6) Dobrusin, E. M.; Fry, D. W. Protein tyrosine kinases and cancer.
Annu. Rep. Med. Chem. 1992, 27, 169-178.
(7) Chang, C.-J .; Geahlen, R. L. Protein-tyrosine kinase inhibition:
mechanism-based discovery of antitumor agents. J . Nat. Prod.
1992, 55, 1529-1560.
(8) Levitski, A.; Gazit, A. Tyrosine kinase inhibition; An approach
to drug development. Science 1995, 267, 1782-1788.
(9) Palmer, B. D.; Rewcastle, G. W.; Thompson, A. M.; Boyd, M.;
Showalter, H. D. H.; Sercel, A. D.; Fry, D. W.; Kraker, A. J .;
Denny, W. A. Tyrosine kinase inhibitors. 4. Structure-activity
relationships among N- and 3-substituted 2,2′-dithiobis(1H-
indoles) for in vitro inhibition of receptor and nonreceptor protein
tyrosine kinases. J . Med. Chem. 1995, 38, 58-67.
(10) Barker, A. J .; Davies, D. H. Eur. Patent Appl. No. 0 520 722
A1, 1992.
(11) Barker, A. J . Eur. Patent Appl. No. 0 566 226 A1, 1993.
(12) Ward, W. H. J .; Cook, P. N.; Slater, A. M.; Davies, D. H.;
Holdgate, G. A.; Green, L. R. Epidermal growth factor receptor.
Investigation of catalytic mechanism, structure-based searching
and discovery of a potent inhibitor. Biochem. Pharmacol. 1994,
48, 659-666.
(13) Fry, D. W.; Kraker, A. J .; McMichael, A.; Ambroso, L. A.; Nelson,
J . M.; Leopold, W. R.; Connors, R. W.; Bridges, A. J . A specific
inhibitor of the epidermal growth factor receptor tyrosine kinase.
Science 1994, 265, 1093-1095.
(14) Rewcastle, G. W.; Denny, W. A.; Bridges, A. J .; Zhou, H.; Cody,
D. R.; McMichael, A.; Fry, D. W. Tyrosine kinase inhibitors. 5.
Synthesis and structure-activity relationships for 4-(phenylm-
ethyl)amino- and 4-phenylaminoquinazolines as potent adenos-
ine 5′-triphosphate binding site inhibitors of the tyrosine kinase
domain of the epidermal growth factor receptor. J . Med. Chem.
1995, 38, 3482-3487.
(15) Hanks, S. K.; Quinn, A. M.; Hunter, T. The protein kinase family;
conserved features and deduced phylogeny of the catalytic
domains. Science 1988, 241, 42-52.
J M9503613