Discovery of a potent and selective EGFR inhibitor (AZD9291) of both sensitizing and T790M resistance mutations that spares the wild type form of the receptor

Article abstract of DOI:10.1021/jm500973a

Epidermal growth factor receptor (EGFR) inhibitors have been used clinically in the treatment of non-small-cell lung cancer (NSCLC) patients harboring sensitizing (or activating) mutations for a number of years. Despite encouraging clinical efficacy with these agents, in many patients resistance develops leading to disease progression. In most cases, this resistance is in the form of the T790M mutation. In addition, EGFR wild type receptor inhibition inherent with these agents can lead to dose limiting toxicities of rash and diarrhea. We describe herein the evolution of an early, mutant selective lead to the clinical candidate AZD9291, an irreversible inhibitor of both EGFR sensitizing (EGFRm+) and T790M resistance mutations with selectivity over the wild type form of the receptor. Following observations of significant tumor inhibition in preclinical models, the clinical candidate was administered clinically to patients with T790M positive EGFR-TKI resistant NSCLC and early efficacy has been observed, accompanied by an encouraging safety profile.

Full text of DOI:10.1021/jm500973a

Drug Annotation
pubs.acs.org/jmc
Discovery of a Potent and Selective EGFR Inhibitor (AZD9291)
of Both Sensitizing and T790M Resistance Mutations That Spares the
Wild Type Form of the Receptor
M. Raymond V. Finlay,*,† Mark Anderton,‡ Susan Ashton,† Peter Ballard,† Paul A. Bethel,†
Matthew R. Box,† Robert H. Bradbury,† Simon J. Brown,† Sam Butterworth,† Andrew Campbell,†
Christopher Chorley,† Nicola Colclough,‡ Darren A. E. Cross,† Gordon S. Currie,† Matthew Grist,†
Lorraine Hassall,† George B. Hill,† Daniel James,† Michael James,† Paul Kemmitt,† Teresa Klinowska,†
Gillian Lamont,† Scott G. Lamont,† Nathaniel Martin,† Heather L. McFarland,§ Martine J. Mellor,†
Jonathon P. Orme,∥ David Perkins,† Paula Perkins,† Graham Richmond,† Peter Smith,† Richard A. Ward,†
Michael J. Waring,† David Whittaker,† Stuart Wells,† and Gail L. Wrigley†
‡
§
∥
†Oncology Innovative Medicines, Drug Safety and Metabolism, Global Medicines Development, and Discovery Sciences,
AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, U.K.
S
* Supporting Information
ABSTRACT: Epidermal growth factor receptor (EGFR)
inhibitors have been used clinically in the treatment of non-
small-cell lung cancer (NSCLC) patients harboring sensitizing
(or activating) mutations for a number of years. Despite en-
couraging clinical efficacy with these agents, in many patients
resistance develops leading to disease progression. In most
cases, this resistance is in the form of the T790M mutation. In
addition, EGFR wild type receptor inhibition inherent with
these agents can lead to dose limiting toxicities of rash and
diarrhea. We describe herein the evolution of an early, mutant
selective lead to the clinical candidate AZD9291, an irreversible inhibitor of both EGFR sensitizing (EGFRm+) and T790M
resistance mutations with selectivity over the wild type form of the receptor. Following observations of significant tumor
inhibition in preclinical models, the clinical candidate was administered clinically to patients with T790M positive EGFR-TKI
resistant NSCLC and early efficacy has been observed, accompanied by an encouraging safety profile.
phenyl]amino]pyrimidin-4-yl]oxyphenyl]prop-2-enamide (WZ-
4002), 6, Gatekeeper Pharmaceuticals Inc.; N-[3-[[2-[[4-(4-
acetylpiperazin-1-yl)-2-methoxyphenyl]amino]-5-chloropyrimidin-
4-yl]amino]phenyl]prop-2-enamide (CO-1686), 7, Clovis Oncology)
have described efforts to identify mutant selective inhibitors that
target both the sensitizing mutations and the T790M resistance
mutation while also sparing the wild type form of the receptor,
inhibition of which can lead to dose limiting toxicities including
skin rash and diarrhea. Herein, we describe some of our work that
has led to the identification of the clinical candidate AZD9291
(8), a potent inhibitor of both sensitizing and double mutant
(sensitizing and T790M resistance) forms of EGFR with selec-
tivity over the wild type form (Figure 2).
INTRODUCTION
■
The role of the epidermal growth factor receptor (EGFR) in non-
small-cell lung cancer is well-known, and substantial therapeutic
progress in the treatment of this disease has been made over the
past 10 years through the exploitation of this insight.1 Inhibition
of the kinase domain of EGFR and resultant oncogenic cell
signaling disruption by small molecule inhibitors such as gefitinib
1 and erlotinib 2 (Figure 1) have been shown to be particularly
beneficial in those patients carrying the so-called “sensitizing
mutations” such as L858R and the exon-19 deletion.2,3 However,
emergence of resistance to these targeted therapies continues to
be a cause for concern, with the T790M mutation in particular
being highlighted as a key therapeutic target.4 The subsequent
identification of irreversible EGFR inhibitors such as dacomitinib
3 and afatinib 45 (Figure 1) that inhibit both mutants de-
scribed above as well as the wild type receptor potentially offers
therapeutic options for T790M positive patients. However, the
high wild-type potency and accompanying toxicities associated
with such activity may well limit their utility in this setting.
Recently, both ourselves6 (compound 5, Figure 2) and others7,8
(N-[3-[5-chloro-2-[[2-methoxy-4-(4-methylpiperazin-1-yl)-
RESULTS AND DISCUSSION
■
Previously, we had described our efforts in this area that
culminated in the identification of 5, a potent and selective EGFR
inhibitor that had demonstrated in vivo efficacy in both activating
Received: June 27, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
pyrimidine 5-position appears to offer a potentially favorable
interaction with the methionine in both IGF1R and the double
mutant ATP binding sites, and we reasoned that kinetic selec-
tivity might be achievable because of irreversible binding to
the double mutant, whereas the IGF1R interaction would be
reversible. While IGF1R inhibitors have been employed in a
clinical setting, their lack of selectivity with respect to inhibition
of insulin receptor signaling has in some instances led to dose
limiting toxicity and monitoring these effects can complicate
study designs.10 An additional driver for favoring 10 was that in
contrast to 9, it did offer modest exposure when dosed orally at
50 mg/kg in SCID mice. The aqueous solubility and hERG
inhibition profile of 10 were also viewed as offering a good
starting point, and we focused our initial efforts on further
evolution of this lead.
Chemistry. A body of work focused on preparing amides of
the distal piperazine nitrogen of 10 was first instigated. By use
of chemistry similar to that described above (Scheme 1), fluoro
intermediate 116 was exposed to various piperazine amides
bearing protected amino/hydroxyl groups. The nitro group was
then reduced, and the resulting aniline was acylated with acryloyl
chloride using the chemistry described previously. Protecting
groups (either BOC or TBDMS) were then removed to afford
the required compounds (Table 2).
Figure 1. Structures of reversible and irreversible EGFR inhibitors.
A variety of alternative diamine reagents were available in
house, and guided by a calculated lipophilicity threshold (log D
of no more than 3 to try and maintain good physicochemical
properties), we targeted the examples shown in Table 3. We
employed our previously described chemistry to access these com-
pounds, using late-stage fluoro intermediate 11 (Scheme 2).6
Introduction of the amine side chain was achieved via SNAr
reaction; subsequent reduction of the nitro group and reaction
with acryloyl chloride furnished the desired final compounds.
To access indole compounds 23−27 (Table 8), commercial
pyrimidines with R2 = H, Me and Cl were used. Introduction of
the indole was achieved by deprotonation with methylmagne-
sium bromide (Scheme 3) and subsequent SNAR in adequate
yield. Where R1 = Me, deprotonation of the indole NH and
methylation with methyl iodide proceeded in good yield.
Introduction of the aniline side chain and further elaboration
into the desired compounds were achieved in a similar manner to
the compounds detailed in Scheme 1.
Figure 2. Mutant selective EGFR inhibitors.
and double mutant human xenograft disease models while also
displaying minimal activity in a wild type setting.6
During the course of these studies, we had also identified 9 and
10 as potent and selective mutant EGFR inhibitors, where the
pendent basic center, normally resident on the terminal position
of the acrylamide, had been migrated to the aryl ring adjacent to
the amide functionality. Our interest was piqued by this archi-
tectural modification for a number of reasons, not least of which
was the improved potency offered by 10 for no apparent increase
in lipophilicity. In our previous work,6 we had encountered
a log D/cell potency correlation whereby although cellular
potency could be increased, it was generally accompanied by an
increase in lipopilicity of the inhibitor compounds, and this
tended to both increase their plasma protein binding and reduce
their aqueous solubility. While compound 9 showed broadly
similar double mutant potency to 5, its reduced IGF1R activity
and increased activating mutant activity again with no increase in
lipophilicity were of great interest. Kinase selectivity for these
inhibitors, as judged by profiling in a kinase selectivity screen
(Dundee panel),9a appeared acceptable with only a small number
of significant hits (Table 1). Compound 10 appeared to offer
superior selectivity for the mutant forms of the receptor,
although it should be noted that 9, as mentioned above, did carry
significantly less activity against the methionine gatekeeper
bearing receptor tyrosine kinase IGF1R. A lipophilic group at the
In the case where R2 was cyano, this group was introduced at
the penultimate step (Scheme 3). Although isolated in modest
yield, this approach yielded sufficient material to deliver the
desired compound.
As can be seen from Table 2, it became apparent that while
initial compounds such as 12 and 13 offered reasonable potency
profiles, all subsequent attempts to achieve further improve-
ments provided less potent analogs, particularly against the
activating mutant cell line. The enantiomeric pair of compounds
14 and 15 showed virtually identical cellular inhibition profiles,
whereas in contrast, the enantiomeric pair of 13 and 19 did
appear to favor the S-enantiomer. Removal of a methyl group
from 12 to give 16 slightly reduced activity, and replacement of
the NH2 group of 13 with a hydroxyl to give 17 again reduced
activity. The urea analogue 18 also led to modest levels of activity
in cells. We concluded (particularly in the light of the emerging
data for the basic side chain anlaogues; vide infra) that amides
such as these did not offer the levels of potency to justify
additional investigation, and these compounds were not pursued
further.
B
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Table 1. Structures and Selected Properties of Lead Optimization Start Points
5
9
10
a
DM cell IC50 (μM)
0.091
0.063
0.015
1.2
0.019
a
AM cell IC50 (μM)
WT cell IC50 (μM)
log D7.4
0.388
0.14
12
a
>20
3.6
3.4
3.6
b
LLE (DM cell)
3.4
3.8
4.1
nude mouse po 50 mg/kg: Cmax (μM), AUC (μM·h)
mouse Heps Clint [(μL/min)/(106 cells)]
5.8, 67
0, 0
53
2.7, 13.4
35
c
solubility, pH 7.4 (μM)
1
828
9.1
>973
hERG IC50 (μM)
4.2
7.1
IGF1R (enz) IC50 (μM)
0.004 (n = 1)
CHK2 (90%)
CAMKKβ (94%)
IGF1R (98%)
INSR (97%)
0.813 (n = 1)
0.002
d
kinase panel hits (>90% inhibition at 1 μM)
not tested
CAMKKβ (92%)
IGF1R (97%)
a
Inhibition of EGFR phosphorylation. All cell IC50 measurements reported are the geometric mean of at least two independent measurements unless
otherwise indicated. The standard error of the mean (SEM) is not shown, as all three assays were found to be highly reproducible on repeat testing
(pIC50 SEM typically <0.2; see Supporting Information for more detail). Because of the irreversible nature of inhibition of the compounds employed
in this study, IC50 values are time dependent and are quoted as measured following an incubation period of 2 h. Double mutant (DM) cell line:
b
H1975. Activating mutant exon 19 del (AM) cell line: PC9. Wild type (WT) cell line: LoVo. Ligand lipophilicity efficiency (LLE) is calculated from
c
d
DM cell pIC50 − log D7.4. Aqueous solubility measurement was performed under thermodynamic conditions using solid material. MRC protein
phosphorylation unit at the University of Dundee.9a
a
Scheme 1. General Synthesis of Piperazine Amides 12−19
a
Reagents and conditions: (i) piperazine amide, DIPEA, DMA, 140 °C (microwave); (ii) iron, NH4Cl, EtOH, water, 100 °C; (iii) acryloyl chloride,
DIPEA, DCM, 0 °C; (iv) functional group deprotection if required; TFA, DCM, rt; or TBAF, THF, rt.
A number of other carbon and oxygen linked basic side chains
were also examined, but they offered little advantage in terms of
potency, physical properties, or oral exposure when compared
to the nitrogen linked systems described above. In contrast,
the compounds where the piperazine group had been replaced
with certain particular basic side chains produced a number of
examples with exquisite cellular potency, in some cases being
below 1 nM (Table 3). Of particular interest was the observa-
tion that the increase in potency observed did not come with
increasing lipophilicity (as also noted for 9 and 10), and in
contrast to the previously described series with the pendent base
on the acrylamide, the log D/cell potency correlation appeared to
be weaker. Intrigued by these findings, we elected to further
profile a variety of the most potent examples from this set. As can
be seen from Table 3, the compounds showed excellent cellular
potency against both mutants in phosphorylation and pheno-
typic antiproliferation assays with selectivity over wild type
EGFR, log D values of around 3, modest aqueous solubility in pH
7.4 buffer, and acceptable free fractions in rat and human plasma.
Table 3 also highlights the observation that the compounds
again showed moderate affinity for the hERG channel and potent
IGF1R activity in a similar fashion to 9 and 10. Nevertheless,
additional profiling in a broad kinase selectivity panel (Dundee
panel,9a 1 μM) did show that the compounds possessed good
overall kinase selectivity and had improved selectivity relative
to 10 and 5. We had previously6 employed a glutathione (GSH)
reaction assay to determine the relative reactivity of inhibitor
compounds,11−14 and the GSH reaction half-lives of these novel
analogues fell within our earlier defined suitable “reactivity
window”. Interestingly, a small dependence of GSH half-life on
the pKa of the basic group is seen, suggestive of intramolecular
base catalysis. The intrinsic clearance of the compounds was
relatively high, but encouraged by the overall profile of the series,
we elected to investigate the exposure of the compounds when
C
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Table 2. Structures and Selected Properties of Piperazine Amides 12−19
dosed orally in the SCID mouse. Compounds were prioritized
for in vivo work based on cell potency, free fraction, and aqueous
solubility.
significant level of biomarker modulation. Given the clear
evidence of in vivo effects on the target protein, the three com-
pounds were selected for further evaluation in efficacy studies.
For this work, in vivo xenograft models representing the drug-
resistant double mutant (H1975), sensitizing mutant (PC9), and
wild type (A431) were selected. Compounds were dosed orally
once a day for 7 days (Figure 6).
We selected 20 as our initial compound to be administered
(the opposite enantiomer of 20 showed 2- to 3-fold less potency
in cells), and as in our previous work,6 we employed a 50 mg/kg
once daily oral dose (Figure 3). The compound was well
tolerated at this dose, and excellent exposure was observed. As
can be seen from Table 4, the plasma levels achieved were
substantial and well in excess of the concentrations needed to
achieve free cover over the cellular IC50 values. As a consequence,
for subsequent compounds in the series, we moved to 10 mg/kg
as our initial dose. At this lower dose, multiples of free cover were
still achieved for compounds 21 and 22 (although more modest
in the latter, Table 4), and these inhibitors were selected for
evaluation in a pharmacodynamic (PD) study.
A single dose of compounds 20, 21, and 22 were dosed orally
in SCID mice at 10 mg/kg, and the percentage knockdown
of EGFR phosphorylation in H1975 (double mutant) and
PC9 (activating mutant) xenografts was measured compared to
control (Tables 5 and 6 and Figures 4 and 5). All three
compounds showed a profound effect on the PD biomarker. In
the PC9 tumor cell line, 22 and 21 showed broadly similar effects
with maximal inhibition (85% and 83% respectively) at the
6 h time point. Compound 20 showed slightly less biomarker
modulation at this point, but in contrast to the other two
compounds where p-EGFR levels began to return toward normal
when measured at subsequent points, 20 continued to show
sustained inhibition at the 24 h time point. In contrast, in the
H1975 PD study, all three compounds showed sustained bio-
marker inhibition out to 24 h, although 21 now showed the most
marked effect. By the 30 h time point, 22 inhibitory effects
seemed to have been reduced whereas 21 still displayed a
From examination of Figure 6, it can be seen that at 10 mg kg−1
day−1, each of the compounds showed pronounced tumor
growth inhibition in the two mutant models, with significant
tumor regression being achieved in each case while having
relatively little impact on the A431 wild type model. In addition,
these data allowed us to build on the observation that approximately
6 h cover over the free cell IC50 translated into both meaningful and
prolonged in vivo PD biomarker modulation as well as substantial
disease model efficacy (based on a range of additional compounds,
data not shown).
With these highly encouraging in vivo data in hand, we elected
to more fully characterize the pharmacokinetic properties of these
compounds in species other than mice, and this information is
captured in Table 7. In vivo clearance in rodents was moderate to
high, and at low doses (5 mg/kg) oral bioavailability was low.
Upon increasing dose however (>50 mg/kg), substantial oral
levels could be obtained, and we postulated that this was probably
occurring because of saturation of clearance mechanisms. How-
ever, dog pharmacokinetic profiling revealed medium clearance
and generally good oral bioavailability.
Clearly the incorporation of the three basic side chains
described above in 20, 21, and 22 combined with the migration
of the base from the acrylamide had brought significant benefits
relative to the start point 5, and we proposed that this advance in
SAR understanding might be applicable to some degree to the
previously described indole analogues such as 9. Although we
D
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Table 3. Structures and Selected Properties of Compounds 20, 21, and 22
20
21
22
DM cell IC50 (nM)
AM cell IC50 (nM)
WT cell IC50 (nM)
2
0.6
2
4
16
21
357
145
3
938
a
DM antiproliferative cell GI50 (nM)
7
17
a
AM antiproliferative cell GI50 (nM)
10
4
19
a
WT antiproliferative cell GI50 (nM)
384
180
2.8
6.4
846
log D7.4
3.1
3.3
b
LLE (DM cell)
5.6
5.1
min sol. pH 7.4 (μM)
28 (Xtl)
222 (Xtl)
43 (Xtl)
b
rat % free
7.1
6.4
5.1
human % free
2.7
3.2
2.5
nude mouse % free
7.8
3.2
5.0
rat Heps Clint [(μL/min)/(106 cells)]
human Heps Clint [(μL/min)/(106 cells)]
mouse Heps Clint [(μL/min)/(106 cells)]
hERG IC50 (μM)
20
59
27
<3
14
5.2
19
60
22
5.6
6.7
4.0
IGF1R (enz) IC50 (μM)
0.007
0.006
0.026
GSH mean t1/2 (min)
140
90
257
pKa (measured)
kinase panel9a hits (>90% inhibition at 1 μM)
8.1
9
7.2
IGF1R (98%)
INSR (98%)
IGF1R (97%)
INSR (99%)
IGF1R (97%)
INSR (98%)
a
b
Double mutant (DM) cell line: H1975. Activating mutant (AM) cell line: PC9. Wild type (WT) cell line: Calu3. Plasma protein binding was
assessed by equilibrium dialysis in the appropriate species at 37 °C in 10% plasma and extrapolated to100% with a single site binding model. Free
and bound concentrations were quantified by LCMS.
a
Scheme 2. General Synthesis of Basic Side Chain Compounds 20, 21, and 22
a
Reagents and conditions: (i) R1R2NH, DIPEA, DMA, 85−100 °C; (ii) iron, NH4Cl, EtOH, water, 100 °C; (iii) acryloyl chloride, DIPEA,
DCM, −15 to 0 °C.
had previously moved away from the indole headgroup because
of its more modest oral PK properties (vide supra) and smaller
wild type margin relative to the pyrazolopyridine,6 the enhanced
potency available from this novel basic side chain SAR combined
with the synthetic diversity handle offered by the indole N-H
prompted us to re-examine this subseries. Thus, we decided to
explore a matrix of compounds where we would incorporate
favored groups at the 4 and 5 positions of the pyrimidine core
with the newly identified basic side chain SAR (vide supra).
Leading examples from this work are shown in Table 8. It should
be noted that in addition to varying R1 and R2 as depicted in
Table 8, additional examples were also prepared where the basic
side chain was replaced with amines such as those in 22 and 20,
but none of these offered the levels of potency afforded by the
N,N,N′-trimethylethylenediamine.
It can be seen that incorporation of the newly identified
basic side chains delivered high cellular potency across all the
examples shown (Table 8). The presence or indeed absence of a
substituent at the 5-position of the pyrimidine ring also had a
differential effect on double mutant potency, IGF1R potency,
and hERG binding.
From Table 8 and the matched pair analysis in Figure 7, it is
apparent that a 5-substituent has a beneficial impact on double
mutant cell potency, possibly driven by interaction with the
methionine gatekeeper residue, although in the case of the more
polar cyano 5-substituent, the conformational effect on the
indole group may also play a role. A similar trend is observed for
IGF1R potency, again perhaps driven by this kinase’s methionine
gatekeeper, although it should be noted that the potency increase
imparted by the 5-Cl is of a greater magnitude for IGF1R than
any of the EGFR mutants or wild type receptor. Potency against
E
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
a
Scheme 3. General Synthesis of Indole Compounds 23−27
a
Reagents and conditions: (i) MeMgBr (1 equiv, 3.2 M in 2-methyl THF), indole (1 equiv), THF, 0 °C → 60 °C; (ii) sodium hydride (1.05 equiv),
methyl iodide (1.05 equiv), THF, 0 °C; (iii) 4-fluoro-2-methoxy-5-nitroaniline (1.05 equiv), tosic acid (1.1 equiv), 2-pentanol, 125 °C; (iv) N,N,N′-
trimethylethane-1,2-diamine (2.2 equiv), DMA, 140 °C; (v) iron (3 equiv), ammonium chloride (0.7 equiv), EtOH, water, 100 °C; (vi) Zn(CN)2
(0.6 equiv), zinc (0.1 equiv), tris(dibenzylideneacetone)dipalladium(0) (0.1 equiv), dicyclohexyl(2′,4′,6′-triisopropylbiphenyl-2-yl)phosphine
(XPhos) (0.2 equiv), DMA, 95 °C; (vii) acryloyl chloride (1 M, THF, 1 equiv), DIPEA (1.1 equiv), THF, 0 °C.
Table 5. % Inhibition of p-EGFR (Measured by R&D Systems
ELISA Kits, Not Site Specific and Expressed as Ratio of
Phosphorylated/Total) in Mice Bearing PC9 Tumors after
a
10 mg/kg Oral Dose of Compounds 20, 21, and 22
compd
1 h
6 h
76*
83***
85**
16 h
24 h
30 h
10
20
21
22
68*
38
30
57**
60**
52*
19
a
Statistical significance (p values) was calculated by comparing com-
pound treated and vehicle groups using a two-tailed t-test, unequal
variance. p values are indicated as follows: (∗) <0.05, (∗∗) <0.01,
(∗∗∗) <0.001. See Supporting Information for further detail.
Table 6. % Inhibition of p-EGFR in Mice Bearing H1975
Tumors after 10 mg/kg Oral Dose of Compounds 20, 21,
a
and 22
Figure 3. Mouse plasma concentration data following a 50 mg/kg oral
dose of compound 20, a 10 mg/kg oral dose of compound 21, and a
10 mg/kg oral dose of compound 22. Data points shown are concentrations
from individual animals with mean values being represented by lines.
compd
1 h
6 h
16 h
24 h
30 h
20
21
22
94***
100***
99***
88***
93**
79***
89***
37
94**
82***
80***
50**
a
p values are indicated as follows: (∗) <0.05, (∗∗) <0.01, (∗∗∗) <0.001.
Table 4. Mouse Plasma Cmax Concentration Data after a
50 mg/kg Oral Dose of Compound 20 and a 10 mg/kg dose of
21 and 22
on the indole 1-position nitrogen also had interesting effects.
Again, examination of Table 8 and matched pair analysis shown
in Figure 8 reveals that the removal of the methyl group does
appear to drive increased cell potency against all three EGF
receptors (only double mutant shown graphically), with hERG
potency having no discernible trend. The picture for IGF1R
potency is more straightforward, with the NH indole analogues
showing increased IGF1R potency. The SAR for wild type
selectivity is also shown, with double mutant selectivity over wild
type showing a statistically significant trend to be greater in the
N-methylated compounds, whereas the N-Me indole examples
when examined as a whole showed no statistically significant
Cmax free Cmax free cover over free cover over free cover over
compd (μM)
(μM)
DM cell IC50
AM cell IC50
WT cell IC50
20
21
22
5.1
0.40
200
42
25
12
3
1.1
0.78
1.3
0.025
0.065
0.17
0.06
16
the activating mutant (data not shown) and wild type forms of
the receptor was also increased, and the hERG potency for the
compounds also followed a similar trend, although this is more
difficult to rationalize. The presence or absence of a methyl group
F
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Table 7. In Vivo Pharmacokinetic Parameters for 20, 21,
a
and 22
20
21
22
b
rat iv CL (mL min−1 kg−1
)
22
13
77
31
16
67
60
10
17
13
25
30
59
24
106
57
7
rat HW po F (%)
mouse iv CL (mL min−1 kg−1
)
mouse po F (%)
dog iv CLp (mL min−1 kg−1
)
dog po F (%)
56
a
Rat oral bioavailability is from a 5 mg/kg dose. Mouse oral
bioavailability is from a 10 mg/kg dose. Dog oral bioavailability is from
b
a 5 mg/kg dose. Han Wistar strain.
reducing IGF1R and INSR activity, we had also maintained the
high selectivity of our above-described inhibitors.
Figure 4. p-EGFR inhibition time course profile in PC9 (sensitizing
mutant) xenograft bearing mice following 10 mg/kg oral dose of
compounds 20, 21, and 22.
By examination of the data for this matrix of compounds in
more detail, it can be seen that while very attractive, the profiles
of most of the compounds were broadly similar to the earlier
examples 20, 21, and 22 described above. Given the wealth of
available attractive options, additional candidate compounds
would need to offer something quite different from those already
selected to proceed further. Compound 8 demonstrated
improved rat PK, reduced hERG affinity, and improved IGF1R
margins relative to the previously described compounds, and
so this compound was selected for further investigation.
Compound 8 also offered an additional degree of broader
chemical and profile diversity when compared to the previously
described lead compounds. Upon dosing 8 in our three efficacy
models, we observed comparable efficacy at relatively low doses
(10 mg kg−1 day−1, Figure 9). Intriguingly, excellent efficacy was
also observed when 8 was dosed at 5 mg kg−1 day−1 and this
unprecedented level of activity combined with its previously
mentioned attractive attributes added to our interest in this
compound, despite the slightly lower wild type selectivity relative
to 20, 21, and 22. Given our earlier free cover model, it seems
difficult to explain how this level of efficacy could be achieved
Figure 5. p-EGFR inhibition time course profile in H1975 (double
mutant) xenograft bearing mice following 10 mg/kg oral dose of
compounds 20, 21, and 22.
activating mutant selectivity over the wild type form of the
receptor when compared with their NH matched pairs. When
profiled in the Millipore kinase selectivity panel,9b it became
evident that for the two 5-H compounds tested, in addition to
with the relatively modest exposure achieved with 8 (free Cmax
=
13 nM, 0.9-fold free cover over DM cell IC50), but further under-
standing did become apparent with additional studies (vide infra).
Figure 6. In vivo efficacy of 22, 21, and 20 when dosed orally in tumor bearing immune compromised mice at 10 mg kg−1 day−1 q.d. H1975 efficacy is
represented in blue, with PC9 data in green and A431 in red. Statistical significance (p values) was calculated by comparing compound treated and
vehicle groups using a one-tailed t test. “NS” indicates % TGI was not significant from vehicle.
G
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Table 8. Structure and Selected Data for Indole Compounds 23−27
23
24
25
H/Cl
26
8
27
R1/R2
Me/CN
Me/Cl
Me/Me
Me/H
15
H/H
2
DM cell IC50 (nM)
0.9
2
0.2
1
AM cell IC50 (nM)
1
2
0.6
2
17
2
WT cell IC50 (nM)
46
58
11
71
480
24
33
DM antiproliferative cell GI50 (nM)
AM antiproliferative cell GI50 (nM)
WT antiproliferative cell GI50 (nM)
IGF1R IC50 (nM)
2 (n = 1)
0.8 (n = 1)
101 (n = 1)
38 (n = 1)
4.3
4 (n = 1)
3 (n = 1)
192 (n = 1)
40 (n = 1)
7.4
2 (n = 1)
2 (n = 1)
144 (n = 1)
7 (n = 1)
14.8
4 (n = 1)
2 (n = 1)
970 (n = 1)
196 (n = 1)
15.6
16
23
21
264
2900
16.2
3.4
58
263
17.5
2.9
79
hERG IC50 (μM)
log D7.4
2.7
3.3
3.3
2.8
min sol. pH 7.4 (μM)
rat HW % free
57
759
75
576
7
11
7.5
3.2
16
human % free
8.9
2.3
3.6
7.5
5.6
2.9
7.5
13.4
27
mouse % free
3.4
rat Heps Clint [(μL/min)/(106cells)]
human Heps Clint [(μL/min)/(106 cells)]
88
27
10
10
<3
3.4
<3
rat HW iv Cl (mL min−1 kg−1
)
62
27
12
45
rat HW po F (%)
6
7
1
45
mouse AUC/Cmax (μM·h/μM), 10 mg/kg po
GSH mean t1/2 (min)
0.28/0.11
1.3/0.31
3.3/1.2
136
0/0
123
1.4/0.38
121
0.203/0.047
191
With a selection of potent, wild type selective and efficacious
compounds in hand, we investigated the in vivo toxicological
profile of the most favored examples. The in vitro kinase selec-
tivity testing had revealed a desirable selectivity profile albeit
with inhibition of IGF1R and INSR as common hits. With this
observation in mind, we designed bespoke rat toxicology studies
to explore whether reversible inhibition of IGF1R-INSR would
lead to hyperglycemia or impaired insulin signaling in vivo. Rats
received a single oral dose of 200 mg/kg of each compound with
assessment of blood glucose and insulin concentrations over a
24 h period (see Supporting Information). While all compounds
were tolerated well, a single oral dose of 21 and 20 led to marked
hyperglycemia and hyperinsulinemia (Figures 10 and 11). At the
same dose, 22 did not cause significant changes in blood glucose
levels, but insulin levels were clearly increased, consistent with
insulin being a more sensitive marker for perturbation of the
glucose/insulin axis. In contrast, 8 did not impact measured
levels of glucose or insulin. This is consistent with the
observation that 8 was the only compound out of the four
leading candidates that did not inhibit IGF1R (Table 8) or INSR
(Table 9) with high potency.
Assessment of the toxicokinetics of the four compounds at
200 mg/kg in this study revealed that the free exposure of 8 was
lower than the other three compounds (Table 9). Compounds
20, 21, and 22 were subsequently administered orally at the
lower dose of 50 mg/kg, and despite a similar free exposure to
compound 8, they were still associated with hyperinsulinemia.
The observed hyperinsulinemia with 20, 21, and 22 at both doses
tested was consistent with the free exposure (Cave) giving cover
over the INSR IC50 (Table 9). In contrast, compound 8 did not
provide cover over the INSR IC50, and furthermore at a dose of
200 mg/kg the INSR IC50 was approximately 50-fold higher than
the free exposure (Cave). For compound 8, pharmacological
cover over the INSR IC50 and perturbation of the insulin−
glucose axis was considered extremely unlikely.
Although IGF1R inhibitors have been studied clinically in
an oncology setting, hyperglycemia can be dose limiting.10
Inhibition of IGF1R and INSR signaling and its associated
pharmacology had the potential to limit maximal tolerated doses
in the NSCLC EGFR sensitive patient setting. Furthermore, it
could affect patient exclusion criteria and require additional
monitoring during clinical trials. In light of these findings it was
decided to focus remaining efforts on 8 and 22 which had no or
minimal effects on the IGF1R−INSR axis, respectively.
We subsequently performed a preliminary cardiovascular
safety evaluation of the final two lead compounds. The com-
pounds were assessed in an in vivo guinea pig model15 to
explore the effect of the compounds on hemodynamic, ECG, and
contractility paramaters. Comparison of the cardiovascular
profile revealed a major difference in the compounds with
respect to QT prolongation. In contrast to 8, 22 was associated
with a pronounced QT prolongation at free Cmax concentrations
less than 30-fold predicted human free Cmax. These data were
consistent with the in vitro data for both compounds with 22
clearly having a higher affinity for the hERG ion channel than 8
(Ionworks hERG IC50 for 22 and 8 is 4.3 and 16.2 μM, respec-
tively). In view of the data package as a whole, 8 was considered
to have a very desirable profile and was ultimately selected as a
clinical candidate. The selectivity profile of 8 as determined in the
Millipore kinase panel16 of 270 kinases is shown in Figure 12
(complete data are available in the Supporting Information), and
it can be seen that the compound displays a good overall profile
with the majority of more active hits being other tyrosine kinases.
Compound 8 both exploits and illustrates the SAR we have discussed
H
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Figure 7. Pyrimidine 5-substituent analysis. N-H indoles appear in red, with N-Me indoles in blue.
above. The addition of the methyl group to the indole increases WT
selectivity and reduces IGF1R potency (Figure 8). Removing the 5-Cl
group from the monoanilinopyrimidine scaffold reduces the IGF1R
potency, WT selectivity, and to a lesser extent DM potency, but the
last can be recovered by employing the inherently more potent indole
rather than pyrazolopyridine headgroup, and indole N−H alkylation
can be used to drive WT selectivity.
Although we are yet to obtain an X-ray crystal structure of 8
bound to the EGFR T790M mutant, we were able to employ a
published structure (PDB code 3IKA)7 to model the binding
mode of the inhibitor, and this is shown in Figure 13. Notable
features include the covalent bond formed to C797, the hinge
interaction at M793, and the orientation of the indole.
Shown in Figure 15 are the total levels for 8, 27, and 28 over a
12 h period following a 25 mg/kg oral dose in the nude mouse. As
can be seen from Table 10, 28 appeared to offer a broadly similar
potency and selectivity profile to the parent compound. In con-
trast, we had previously prepared and profiled 27 (see Table 8),
but despite its impressive double and activating mutant potency,
we had not progressed the compound further, partly because of
its modest in vitro wild type margin. In addition, oral dosing
in the mouse led to only modest levels of exposure and low levels
of free cover over the cellular IC50 values. Nevertheless, we were
now faced with the intriguing prospect that a degree of the
observed efficacy in vivo for 8 was actually driven by the
metabolite 27.
Having selected 8 for further development, we wished to
further understand its metabolic fate in vivo. Previously, when
looking at in vivo metabolism in rat bile samples, little had been
evident apart from glutathione conjugation to the acrylamide
portion of the molecule. In contrast, when in vivo plasma samples
from rat and mouse were examined in more detail, it became
evident that loss of the indole N-methyl group was occurring,
leading to small quantities of the metabolite 27 (Figure 14). In
addition, dealkylation was also seen to occur to a small extent in
the N,N,N′-trimethylethylenediamine side chain of 8, yielding 28
(Figure 14). A small amount of the side chain N-oxide was also
evident at low levels.
While 27 is present at lower concentrations in vivo than 8, its
higher cellular potency and lower protein binding equate to
broadly similar free cover in vivo. A number of experiments were
undertaken to try and establish the in vivo efficacy of 8 in the
absence of 27. These included co-dosing mice with 8 and the pan
cytochrome P-450 inhibitor benzotriazol-1-amine (ABT)17,18 in
an attempt to reduce oxidative metabolism of the indole N-
methyl group to give 27. In addition, we undertook the pre-
paration of an isotopically labeled version of 8, where the indole
N-methyl group was replaced by the 13CD3 equivalent. Our
hypothesis here was that the kinetic isotope effect19 might slow
the deuterium atom extraction process relative to the
I
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Figure 8. Matched pair analyses for N-H-indole and N-Me examples. Each marker represents an individual compound matching the substructure shown.
Lines connect exact structural matches (∗ = positions of variation). Top left mean pIC50 difference is 0.61 (SE = 0.1, n = 11). Top right mean pIC50
difference is 0.06 (SE = 0.078, n = 11). Middle left mean pIC50 difference is 0.77 (SE = 0.09, n = 11). Middle right mean pIC50 difference is 0.072 (SE =
0.02, n = 11). Bottom mean pIC50 difference is 0.04 (SE = 0.02, n = 11).
corresponding hydrogen atom reaction and thus suppress
formation of 27, but neither of these measures were able to
prevent the in vivo formation of 27 to an extent that its
contribution to efficacy could be ignored.
In order to try and understand the implications of the presence
of the metabolite 27 in vivo, we elected to dose 27 in all three
of the previously described efficacy models. Examination of
the outcome from these studies (Figure 16) showed that while
substantial double and activating mutant efficacy is apparent, the
level of wild type efficacy was also significant, reflecting the
narrower in vitro margin as predicted from the in vitro cellular
potency. It is also noteworthy that in the case of 27, only very
modest cover over the free cell IC50 is needed to achieve sig-
nificant efficacy, in contrast to other compounds profiled in this
J
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Figure 9. In vivo xenograft activity of 8 when dosed orally in tumor bearing immune compromised mice at 10 and 5 mg/kg q.d.
Table 9. Exposure, INSR Potency/Fold Cover, and Effects on
Blood Insulin Levels in the Rat after a Single 200 or 50 mg/kg
po Dose of Compounds 20, 21, 22 and a Single 200 mg/kg
Dose of 8
INSR
IC50
(μM)
mean insulin
increase over
24 h
dose
free Cave
free fold cover
over INSR IC50
compd (mg/kg) over 24 h
8
200
200
50
0.0169
0.1101
0.0358
0.0575
0.0277
0.102
0.912
0.021
0.018
5.3
1.7
6.2
3
1.1
20
30.7
13.4
33.7
9.5
21
22
200
50
0.009
0.035
200
50
2.9
1.7
9.5
0.06
3.3
Figure 10. Compound effects of blood glucose levels in the rat following
a single 200 mg/kg oral dose of compounds 20, 21, 22, and 8.
afatinib (4) and that mutant selective inhibitors 8 and 27 offer a
superior efficacy profile than these earlier agents. Clearly, the
extent of formation of 27 in human patients would be a pivotal
piece of information, as while its absence could possibly
compromise the efficacy of the agent, more substantial levels
could potentially lead to diminished margins to EGFRi driven
wild type toxicity. Predicting an efficacious dose in patients for
this agent presented challenges in terms of the irreversible
inhibition mode of action, but the contribution of the metabolite
to PD and efficacy also complicated the picture. The full details
of these studies will be reported elsewhere, but basing human
PK estimations on a metabolite 27 formation rate of 25% of
parent (formation of 28 could not be quantified in vitro),
the predicted efficacious dose in patients based on data from
the PC9 model was approximately 20 and 10 mg based on the
H1975 model.
Figure 11. Compound effects of blood insulin levels in the rat following
a single 200 mg/kg oral dose of compounds 20, 21, 22, and 8.
way. This is also apparent for 8, but the presence of 27 in vivo
makes this observation simpler to rationalize.
In silico modeling of human absorption20 showed that 8
should not be solubility limited at clinically relevant doses with
Fabs > 50% up to 330 mg. Although the free base properties
of 8 were considered to offer a low biopharmaceutical risk,
the mesylate salt was selected for further work because of
manufacturing risks around the physical form of the free base. In
vivo studies established that the pharmacokinetics of 8 mesylate
were largely comparable to those of the free base.
Comparison of the efficacy profiles between the two
compounds revealed significant differences, and while it is likely
that 27 contributes to the observed in vivo efficacy of 8, it does
not appear to be the sole driver. It is apparent from Figure 16 that
the efficacy profiles of 8 and 27 are quite different from both the
reversible inhibitor gefitinib (1) and the irreversible compound
K
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Figure 12. Millipore kinase selectivity profile for compound 8. Colors are indicative of % inhibition at 1 μM as follows: dark green, 0−20% of control;
light green, 20−40% of control; yellow, 40−60% of control; orange, 60−80% of control; red, 80−100% of control. Full % inhibition data are listed in the
Supporting Information including IC50 data for those kinases with an analogous cysteine to C797.
The mesylate salt of 8 was first dosed in March 2013 to
patients with EGFRm+ advanced NSCLC who had disease
progression following treatment with an EGFR TKI (including 1
or 2). On the basis of the work described above, the initial dose
selected for oral administration to patients was set at 20 mg, once
a day. Pharmacokinetic data from this cohort of six patients are
shown below.21 It can be seen that after a single dose, 7-day
washout, and then 8 days of once daily dosing, exposure to 8, 27,
and 28 was observed (Figure 17). It is also evident that the
formation rate of the 27 metabolite is largely consistent with the
predicted levels and that interpatient variability is low. Clearly
these data discharge to some extent the above-mentioned
potential risks around the presence or absence of 27.
Clinical testing with 8 is ongoing, with encouraging initial
efficacy seen in NSCLC patients. Data from two patients have
recently been reported21b,22 and are discussed here. Serial CT
scans from two patients who received 8 at a 20 mg once daily dose
Figure 13. Modeled binding mode of compound (8). See Supporting
Information for further details. Adapted by permission from the
are shown (Figures 18 and 19). Prior to commencing dosing, both
patients’ tumors were known to exhibit T790M mutations
according to local testing, as well as EGFR sensitizing mutations.
The first patient was a 57 year old East Asian female from South
Korea diagnosed with stage IV EGFR mutant (ex 19 del) NSCLC
in May 2011. Tumor shrinkage on treatment with 8 was 39.7%
at RECIST 1.1 (response evaluation criteria in solid tumors)23
American Association for Cancer Research: Cross, D. A. E.; et al.
AZD9291, an irreversible EGFR TKI, overcomes T790M-mediated
resistance to EGFR inhibitors in lung cancer. Cancer Discovery 2014, 4,
1046−1061.21b
scan 1, was 48.3% byscan 2 (Figure 18), remained at 48.3% at both
scan 3 and scan 4, and was 51.7% at scan 5 (data not shown).
L
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Table 10. Compound 28 in Vitro Profile
28
DM cell IC50 (nM)
45
AM cell IC50 (nM)
26
WT cell IC50 (nM)
786
19
DM antiproliferative cell GI50 (nM)
AM antiproliferative cell GI50 (nM)
WT antiproliferative cell GI50 (nM)
IGF1R IC50 (μM)
15
537
1.6 (n = 1)
>33
2.3
hERG IC50 (μM)
log D7.4
min sol. pH 7.4 (μM)
rat HW % free
14
2.3
human % free
3.9
Figure 14. Compound 8 and identified circulating metabolites 27 and 28.
mouse % free
1.05
50
rat Heps Clint [(μL/min)/(106 cells)]
human Heps Clint [(μL/min)/(106 cells)]
<3
The second subject21b was a 57 year old white female from the
United Kingdom, diagnosed with stage IV lung adenocarcinoma
in December 2010. At the cycle 1 day 15 assessments on 8, the
patient reported full resolution of pre-existing persistent
nocturnal cough. Tumor shrinkage was 38% at RECIST 1.1
scan 1, 39.3% by scan 2, 56.7% by scan 3 (Figure 19), 62% by
scan 4, and 59.3% by scan 5 (data not shown). By cycle 7 day 1,
the patient reported significant improvement in pre-existing hair
and eyelash abnormalities which had developed during the
immediately prior gefitinib therapy. Both patients had a duration
of response of approximately 9 months and were progression-
free on 20 mg/day of compound 8 mesylate salt for
approximately 11 months until disease progression by RECIST
1.1. Both patients continue to receive compound 8 treatment on
study as per protocol, as they continue to derive clinical benefit
according to their treating physicians. Consistent with 8 being
less active against wild-type EGFR, in these two cases there
Figure 15. Total levels for 8, 27, and 28 over a 12 h period following a 25
mg/kg oral dose in the nude mouse. The markers are mean values SEM.
Figure 16. In vivo efficacy profiles (% tumor growth inhibition) of 8, 27, afatinib (4), and gefitinib (1).
M
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
were no rash events and only one reported CTCAE (common
terminology criteria for adverse events)24 grade 1 diarrhea. No
significant aberration of blood glucose levels were noted in either
patient during the study.
CONCLUSION
■
We have described herein the evolution of a chemical series from
lead selection to candidate nomination and the initial clinical
findings with this novel agent. During the course of this work, a
number of key challenges were addressed including off-target
kinase selectivity and cardiovascular safety. Compound 8 offers
a number of advantages over earlier EGFR TKI agents
including high cellular potency, superior efficacy in relevant
disease models, and selectivity over the wild type form of the
receptor. The promising data package for 8 strongly supported
its selection as a clinical candidate, and first dose in patients
was achieved in March 2013. To date, encouraging and
reproducible drug exposure has been achieved in humans, and
of particular significance is the early evidence of tumor
responses in patients who had previously progressed on earlier
treatments, accompanied by good tolerability.25 Additional
safety, pharmacokinetic, and clinical data for 8 will be reported
in due course.
Figure 17. Preliminary plasma concentration versus time plots (mean +
SD) of 8, 27, and 28 after 8 days of once daily 20 mg oral doses of 8 to
patients with advanced EGFRm+ NSCLC in AURA phase 1 study
(NCT01802632) (n = 6). Adapted by permission from the American
Association for Cancer Research: Cross, D. A. E.; et al. AZD9291, an
irreversible EGFR TKI, overcomes T790M-mediated resistance to EGFR
inhibitors in lung cancer. Cancer Discovery 2014, 4, 1046−1061.21b
EXPERIMENTAL SECTION
■
General. All solvents and chemical used were reagent grade. Purity
and characterization of compounds were established by a combination
of LCMS and NMR. LCMS spectra were obtained using a Waters liquid
chromatography mass spectrometry system, where purity was
determined by UV absorption at a wavelength of 254 nm, and the
mass ion was determined by electrospray ionization (Micromass
instrument). All test compounds were >95% pure as assessed by LCMS
and 1H NMR. 1H NMR spectra were recorded using a Bruker Avance 400
FT spectrometer or via flow NMR process using an AVANCE 500 FT
spectrometer and using DMSO-d6 or CDCl3 with the data expressed as
chemical shifts in ppm from internal standard TMS. Preparative HPLC
was performed on a Waters or Phenomenex column using decreasingly
polar mixtures of water (containing 1% formic acid or 1% aqueous
NH4OH) and MeCN. Purification by column chromatography (FCC)
was typically performed using silica gel (Merck 7734 grade), and solvent
mixtures and gradients are recorded herein. Water present in compounds
was quantified by Karl Fischer analysis. All reactions were performed
under nitrogen unless otherwise stated. All IC50 data are quoted as
geometric mean values, and statistical analysis is availble in the Supporting
Information.
Figure 18. Serial computed tomography scans of the chest from patients
before and after treatment with 8 in a phase I trial: images from a 57-year
old Korean female patient diagnosed with stage IV non-small-cell lung
cancer in May 2011. Adapted by permission from the American Associa-
tion for Cancer Research: Cross, D. A. E.; et al. AZD9291, an irreversible
EGFR TKI, overcomes T790M-mediated resistance to EGFR inhibitors
in lung cancer. Cancer Discovery 2014, 4, 1046−1061.21b
Synthesis of Representative Key Examples (Compounds 20,
21, 22, and 8). N-[5-[(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyr-
imidin-2-yl)amino]-2-[(3R)-3-dimethylaminopyrrolidin-1-yl]-4-
methoxyphenyl]prop-2-enamide (20). 5-Chloro-N-[4-[(3R)-3-di-
methylaminopyrrolidin-1-yl]-2-methoxy-5-nitrophenyl]-4-pyrazolo-
[1,5-a]pyridin-3-ylpyrimidin-2-amine.
(R)-(+)-3-(Dimethylamino)pyrrolidine dihydrochloride (90 mg, 0.48
mmol) was added to a suspension of 5-chloro-N-(4-fluoro-2-methoxy-5-
nitrophenyl)-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-amine6 (11,
200 mg, 0.48 mmol) and DIPEA (0.250 mL, 1.45 mmol) in DMA
(3 mL). The mixture was heated at 140 °C in a microwave for 0.5 h. The
mixture was diluted with MeOH and absorbed onto an SCX column,
washed with MeOH, and eluted with 1:1 DCM/NH3 in MeOH.
Product-containing fractions were concentrated. The crude product was
purified by flash silica chromatography, eluting with 1.5% 7 N NH3/MeOH
Figure 19. Serial computed tomography scans of the chest from patients
before and after treatment with 8 in a phase I trial: images from a
57-year old British female never smoker diagnosed with stage IV lung
adenocarcinoma in December 2010. Adapted by permission from the
American Association for Cancer Research: Cross, D. A. E.; et al.
AZD9291, an irreversible EGFR TKI, overcomes T790M-mediated
resistance to EGFR inhibitors in lung cancer. Cancer Discovery 2014, 4,
1046−1061.21b
N
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
in DCM. Pure fractions were evaporated to dryness to afford
5-chloro-N-[4-[(3R)-3-dimethylaminopyrrolidin-1-yl]-2-methoxy-
5-nitrophenyl]-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-amine
C = 60.7%, H = 5.6%, N = 21.0%. m/z (ES+) (M + H)+ = 533.5. HRMS:
(M + H)+ = C27H30O2N8Cl = 533.21747, found 533.21747.
N-[5-[(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-
amino]-2-(2-dimethylaminoethylmethylamino)-4-methoxyphenyl]-
prop-2-enamide (21). N-(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyr-
imidin-2-yl)-N′-(2-dimethylaminoethyl)-2-methoxy-N′-methyl-5-ni-
trobenzene-1,4-diamine.
1
(149 mg, 61%) as an orange foam. H NMR (400 MHz, DMSO,
21 °C) δ 1.76−1.89 (1H, m), 2.14−2.25 (7H, m), 2.69−2.84 (1H,
m), 3.12−3.27 (3H, m), 3.41−3.53 (1H, m), 3.89 (3H, s), 6.56
(1H, s), 7.13 (1H, td), 7.26−7.38 (1H, m), 8.06 (1H, s), 8.40−8.43
(2H, m), 8.73 (1H, s), 8.85 (1H, d), 8.95 (1H, s); m/z (ES+)
(M + H)+ = 509.5.
N-(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-4-[(3R)-
3-dimethylaminopyrrolidin-1-yl]-6-methoxybenzene-1,3-diamine.
N,N,N′-Trimethylethylenediamine (0.188 mL, 1.45 mmol) was added
to a suspension of 5-chloro-N-(4-fluoro-2-methoxy-5-nitrophenyl)-4-
pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-amine (11, 500 mg, 1.21 mmol)
and N-ethyldiisopropylamine (0.250 mL, 1.45 mmol) in DMA (5 mL).
The mixture was heated at 140 °C in a microwave for 30 min. The
mixture was diluted with MeOH and absorbed onto an SCX column,
washed with MeOH, and eluted with 1:1 DCM/NH3 in MeOH.
Product containing fractions were evaporated to dryness to afford
N-(5-chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-N′-(2-di-
methylaminoethyl)-2-methoxy-N′-methyl-5-nitrobenzene-1,4-dia-
mine (624 mg, 104%) as an orange solid. 1H NMR (400 MHz,
DMSO, 30 °C) δ 2.17 (6H, d), 2.89 (3H, d), 3.87−3.93 (3H, m), 6.84
(1H, s), 7.14 (1H, td), 7.31−7.38 (1H, m), 8.15 (1H, s), 8.39 (1H, d),
8.44 (1H, d), 8.69 (1H, s), 8.85 (1H, d), 8.95 (1H, s); m/z (ES+)
(M + H)+ = 497.
5-Chloro-N-[4-[(3R)-3-dimethylaminopyrrolidin-1-yl]-2-methoxy-5-
nitrophenyl]-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-amine (145 mg,
0.28 mmol), iron (95 mg, 1.71 mmol), and ammonium chloride (11.43 mg,
0.21 mmol) were heated in ethanol (6 mL) and water (2 mL) at reflux
for 1.5 h. The mixture was cooled and concentrated. The residue
was triturated in 10% MeOH/DCM (15 mL) for 15 min and filtered.
The residues were retriturated with 10% MeOH/DCM (15 mL) and
filtered. The combined filtrates were washed with brine, dried (Na2SO4),
and concentrated. The crude product was purified by flash silica
chromatography, eluting with 2% 7 N NH3/MeOH in DCM. Pure
fractions were evaporated to dryness to afford N-(5-chloro-4-pyrazolo-
[1,5-a]pyridin-3-ylpyrimidin-2-yl)-4-[(3R)-3-dimethylaminopyrroli-
din-1-yl]-6-methoxybenzene-1,3-diamine (112 mg, 82%) as a yellow
N4-(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-N1-(2-
dimethylaminoethyl)-5-methoxy-N1-methylbenzene-1,2,4-tria-
mine.
1
gum. H NMR (400 MHz, CDCl3, 21 °C) δ 1.83−1.96 (1H, m),
2.08−2.23 (1H, m), 2.30 (6H, s), 2.82−2.92 (1H, m), 2.99−3.13
(2H, m), 3.17−3.28 (2H, m), 3.65 (2H, s), 3.84 (3H, s), 6.72 (1H, s),
6.96 (1H, td), 7.38 (1H, ddd), 7.52 (1H, s), 7.90 (1H, s), 8.36 (1H, s),
8.55−8.60 (1H, m), 8.65 (1H, dd), 8.94 (1H, s); m/z (ES+) (M + H)+ =
479.5.
N-[5-[(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-
amino]-2-[(3R)-3-dimethylaminopyrrolidin-1-yl]-4-methoxyphenyl]-
prop-2-enamide.
A solution of ammonium chloride (45.2 mg, 0.85 mmol) in water
(10 mL) was added in one portion to a stirred mixture of N-(5-chloro-4-
pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-N′-(2-dimethylaminoethyl)-
2-methoxy-N′-methyl-5-nitrobenzene-1,4-diamine (600 mg, 1.21 mmol)
and iron (405 mg, 7.24 mmol) in ethanol (30 mL). The resulting mixture
was stirred at 105 °C for 3 h. The mixture was evaporated to dryness and
the residue triturated with DMF (10 mL). The crude product was purified
by ion exchange chromatography, using an SCX column. The desired
product was eluted from the column using 0.35 M NH3/MeOH/DCM,
and pure fractions were evaporated to dryness to afford N4-(5-chloro-4-
pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-N1-(2-dimethylaminoethyl)-
5-methoxy-N1-methylbenzene-1,2,4-triamine (530 mg, 94%) as a brown
gum. 1H NMR (400 MHz, DMSO, 22 °C) δ 2.16 (6H, d), 2.38 (2H, t),
2.66 (3H, d), 2.92 (2H, t), 3.66 (3H, s), 4.60 (2H, s), 6.78 (1H, s), 6.92
(1H, s), 7.12 (1H, t), 7.27−7.4 (1H, m), 8.38 (1H, s), 8.43 (1H, d), 8.49
(1H, s), 8.83 (1H, d), 8.95 (1H, s); m/z (ES+) (M + H)+ = 467.
N-[5-[(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-
amino]-2-(2-dimethylaminoethylmethylamino)-4-methoxyphenyl]-
prop-2-enamide.
Acryloyl chloride (0.042 mL, 0.51 mmol) in DCM (1 mL) was
added dropwise to N-(5-chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyr-
imidin-2-yl)-4-[(3R)-3-dimethylaminopyrrolidin-1-yl]-6-methox-
ybenzene-1,3-diamine (245 mg, 0.51 mmol) and DIPEA (0.097 mL,
0.56 mmol) in DCM (10 mL), cooled in an ice/water bath. The
mixture was stirred for 2 h and then washed with brine, dried
(Na2SO4), and concentrated. The crude product was purified by
flash silica chromatography, eluting with 2% 7 N NH3/MeOH in
DCM. Pure fractions were evaporated to dryness to give a foam.
This was triturated with DCM/Et2O and concentrated. The solid
was triturated with Et2O and filtered to afford N-[5-[(5-chloro-4-
pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)amino]-2-[(3R)-3-di-
methylaminopyrrolidin-1-yl]-4-methoxyphenyl]prop-2-enamide
1
(157 mg, 58%) as a yellow solid. H NMR (400 MHz, DMSO,
22 °C) δ 1.68−1.83 (1H, m), 2.05−2.16 (1H, m), 2.18 (6H, s),
2.64−2.76 (1H, m), 3.18−3.29 (3H, m), 3.36−3.47 (1H, m), 3.77
(3H, s), 5.67 (1H, dd), 6.16 (1H, dd), 6.48 (1H, dd), 6.54 (1H, s),
7.12 (1H, t), 7.37 (1H, t), 7.43 (1H, s), 8.28−8.46 (2H, m), 8.55
(1H, s), 8.83 (1H, d), 8.94 (1H, s), 9.37 (1H, s); 13C NMR (176
MHz, DMSO, 22 °C) δ 29.4, 43.8, 49.2, 54.8, 55.6, 65.1, 99.3,
107.1, 113.8, 114.2, 117.7, 119.0, 120.6, 125.5, 125.7, 126.9, 129.3,
131.9, 139.3, 142.8, 143.2, 151.5, 155.4, 157.8, 159.4. CHN analysis,
C27H29O2N8Cl requires C = 60.84%, H = 5.48%, N = 21.02%. Found:
Acryloyl chloride (1.248 mL, 1M, THF, 1.25 mmol) was added
dropwise to N4-(5-chloro-4-(pyrazolo[1,5-a]pyridin-3-yl)pyrimidin-2-
yl)-N1-(2-(dimethylamino)ethyl)-5-methoxy-N1-methylbenzene-
1,2,4-triamine (530 mg, 1.13 mmol) and DIPEA (0.244 mL, 1.36 mmol)
in THF (20 mL) cooled to 0 °C. The resulting mixture was stirred at
0 °C for 2 h. The reaction mixture was evaporated to dryness and
redissolved in DCM (100 mL) and washed sequentially with saturated
O
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
NaHCO3 (25 mL), water (25 mL), and saturated brine (25 mL). The
organic layer was evaporated to afford crude product. The crude product
was purified by flash silica chromatography, with elution gradient
0−20% 2 M NH3/methanol in DCM. This gave product still containing
impurities. The crude product was purified by flash silica chromatog-
raphy, with elution gradient 0−20% MeOH in DCM. Pure fractions
were evaporated to dryness to afford a brown gum. Again this showed
impurities by LCMS. The crude product was purified by flash silica
chromatography, with elution gradient 0−20% MeOH in DCM to yield
N-[5-[(5-chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)amino]-
2-(2-dimethylaminoethylmethylamino)-4-methoxyphenyl]prop-2-en-
amide (191 mg, 32%) as a brown gum. Lyophilization from MeOH/
water gave a brown semisolid. Trituration of the semisolid with ether
followed by evaporation of the ether gave N-[5-[(5-chloro-4-pyrazolo-
[1,5-a]pyridin-3-ylpyrimidin-2-yl)amino]-2-(2-dimethylaminoethyl-
methylamino)-4-methoxyphenyl]prop-2-enamide (191 mg, 32%) as a
pale yellow solid. 1H NMR (400 MHz, CDCl3, 22 °C) δ 2.28 (6H, s),
2.32 (2H, t), 2.71 (3H, s), 2.84−2.92 (2H, m), 3.87 (3H, s), 5.67 (1H,
dd), 6.29 (1H, dd), 6.37 (1H, dd), 6.80 (1H, s), 6.89 (1H, td), 7.23−
7.33 (1H, m), 7.46 (1H, s), 8.45 (1H, s), 8.52 (1H, d), 8.56 (1H, d), 8.94
(1H, s), 9.39 (1H, s), 10.09 (1H, s); 13C NMR (176 MHz, DMSO,
22 °C) δ 29.4, 41.8, 44.9, 45.0, 55.7, 56.7, 105.8, 106.9, 113.8, 114.3,
117.7, 120.3, 124.0, 126.2, 127.1, 129.0, 132.0, 139.0, 140.0, 143.0, 148.9,
155.3, 157.5, 158.9. CHN analysis, C26H29 N8O2Cl requires C = 59.94%,
H = 5.61%, N = 21.51%. Found: C = 60.2%, H = 5.7%, N = 21.7%; m/z
(ES+) (M + H)+ = 521.29; HRMS: (M + H)+ = C26H30 N8O2Cl =
521.217 47, found 521.217 59.
0.54 mmol), iron (179 mg, 3.21 mmol), and ammonium chloride
(20.05 mg, 0.37 mmol) were heated in ethanol (6 mL) and water (2 mL)
at reflux for 1 h. The crude product was purified by ion exchange
chromatography, using an SCX column. The desired product was eluted
from the column using 7 M NH3/MeOH and evaporated to dryness.
Analysis showed clean product, N-(5-chloro-4-pyrazolo[1,5-a]pyridin-3-
ylpyrimidin-2-yl)-4-(3-dimethylaminoazetidin-1-yl)-6-methoxybenzene-
1,3-diamine (235 mg, 94%) as a yellow solid which was used directly with
no further purification. 1H NMR (400 MHz, DMSO, 30 °C) δ 2.13 (6H, s),
3.07 (1H, s), 3.50 (2H, t), 3.66 (3H, s), 4.00 (3H, t), 4.05 (2H, s), 6.28 (1H,
s), 6.79 (1H, s), 7.10 (1H, t), 7.3−7.39 (1H, m), 8.33 (1H, s), 8.37 (1H, s),
8.80 (1H, d), 8.93 (1H, s); m/z (ES+) (M + H)+ = 465.25.
N-[5-[(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-
amino]-2-(3-dimethylaminoazetidin-1-yl)-4-methoxyphenyl]prop-
2-enamide.
Acryloyl chloride (0.038 mL, 0.47 mmol) in DCM (1 mL) was added
dropwise to a stirred solution of N-(5-chloro-4-pyrazolo[1,5-a]pyridin-
3-ylpyrimidin-2-yl)-4-(3-dimethylaminoazetidin-1-yl)-6-methoxyben-
zene-1,3-diamine (220 mg, 0.47 mmol) and DIPEA (0.090 mL,
0.52 mmol) in DCM (5 mL), cooled in an ice/water bath. The mixture
was stirred for 3 h and then washed with brine, dried (Na2SO4), and
concentrated. The crude product was purified by flash silica
chromatography, eluting with 0−5% 7 N NH3/MeOH in DCM. Pure
fractions were evaporated to dryness to afford N-[5-[(5-chloro-4-
pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)amino]-2-(3-dimethylami-
noazetidin-1-yl)-4-methoxyphenyl]prop-2-enamide (221 mg, 90%) as a
yellow solid. 1H NMR (400 MHz, DMSO, 30 °C) δ 2.09 (6H, s), 3.08 (1H,
p), 3.55−3.62 (2H, m), 3.76 (3H, s), 3.97 (2H, t), 5.66 (1H, dd), 6.16 (1H,
dd), 6.25 (1H, s), 6.45 (1H, dd), 7.10 (1H, dd), 7.35 (1H, s), 7.39 (1H, dd),
8.25−8.4 (1H, m), 8.35 (1H, s), 8.45 (1H, s), 8.81 (1H, d), 8.92 (1H, s),
9.24 (1H, s); 13C NMR (176 MHz, DMSO, 22 °C) δ 41.1, 55.5, 55.5,
56.9, 56.9, 97.2, 106.9, 113.8, 113.8, 115.8, 118.5, 120.2, 124.9, 125.1,
126.5, 128.9, 131.7, 139.1, 142.8, 144.4, 151.4, 155.2, 157.4, 159.2,
163.5. CHN analysis, C26H27ClN8O2 requires C = 60.17%, H = 5.24%,
N = 21.59%. Found: C = 60.3%, H = 5.3%, N = 21.7%. m/z (ES+)
(M + H)+ = 519.56. HRMS: (M + H)+ = C26H28 N8O2Cl = 519.201 82,
found 519.201 90.
N-[5-[(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-
amino]-2-(3-dimethylaminoazetidin-1-yl)-4-methoxyphenyl]-
prop-2-enamide (22). 5-Chloro-N-[4-(3-dimethylaminoazetidin-1-
yl)-2-methoxy-5-nitrophenyl]-4-pyrazolo[1,5-a]pyridin-3-ylpyrimi-
din-2-amine.
DIPEA (0.341 mL, 1.96 mmol) was added to a mixture of 5-chloro-N-
(4-fluoro-2-methoxy-5-nitrophenyl)-4-(pyrazolo[1,5-a]pyridin-3-yl)-
pyrimidin-2-amine (11, 254 mg, 0.61 mmol) and N,N-dimethylazetidin-
3-amine dihydrochloride (106 mg, 0.61 mmol) in DMA (4 mL). The
mixture was heated at 100 °C for 30 min. The reaction was incomplete
and further N,N-dimethylazetidin-3-amine (35 mg, 0.19 mmol) was
added. The mixture was stirred at 100 °C for a further 2 h and then sat at
room temperature overnight.
The reaction mixture was purified directly by ion exchange
chromatography, using an SCX column. The desired product was
eluted from the column using 7 M NH3/MeOH and evaporated onto
silica for purification by flash silica chromatography. The crude product
was purified by flash silica chromatography, with elution gradient 0−4%
7 N NH3/MeOH in DCM. Pure fractions were evaporated to dryness to
afford 5-chloro-N-[4-(3-dimethylaminoazetidin-1-yl)-2-methoxy-5-ni-
trophenyl]-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-amine (310 mg,
102%) as an orange solid. 1H NMR (400 MHz, DMSO, 30 °C) δ 2.14
(6H, s), 3.08−3.18 (1H, m), 3.76 (2H, dd), 3.89 (3H, s), 4.02−4.11
(2H, m), 6.28 (1H, s), 7.12 (1H, td), 7.3−7.39 (1H, m), 8.12 (1H, s),
8.37 (1H, br s), 8.42 (1H, s), 8.68 (1H, s), 8.83 (1H, d), 8.94 (1H, d);
m/z (ES+) (M + H)+ = 495.56.
N-[2-(2-Dimethylaminoethylmethylamino)-4-methoxy-5-[[4-
(1-methylindol-3-yl)pyrimidin-2-yl]amino]phenyl]prop-2-en-
amide (8). 3-(2-Chloropyrimidin-4-yl)-1H-indole.
Methylmagnesium bromide (3 M in diethyl ether) (22.68 mL, 68.03
mmol) was added dropwise over a period of 10 min to a stirred solution
of 1H-indole (7.97 g, 68.03 mmol) in 1,2-dichloroethane (250 mL) at
0 °C under nitrogen. The resulting solution was stirred for 15 min.
2,4-Dichloropyrimidine (15 g, 100.69 mmol) was added in one portion.
The resulting solution was allowed to warm to ambient temperature and
stirred for 16 h. The reaction was quenched by the addition of MeOH
(25 mL). The mixture was evaporated to dryness, absorbed onto silica,
and purified by flash silica chromatography, with elution gradient
0−20% MeOH in DCM. Pure fractions were evaporated to dryness to
afford 3-(2-chloropyrimidin-4-yl)-1H-indole (7.17 g, 46%) as a yellow
solid. 1H NMR (400 MHz, DMSO, 30 °C) δ 7.2−7.28 (2H, m), 7.49−
7.53 (1H, m), 7.91 (1H, d), 8.42 (1H, dd), 8.50 (1H, d), 8.53 (1H, d),
12.06 (1H, s); m/z (ES+) (M + H)+ = 230.
N-(5-Chloro-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-yl)-4-(3-di-
methylaminoazetidin-1-yl)-6-methoxybenzene-1,3-diamine.
5-Chloro-N-[4-(3-dimethylaminoazetidin-1-yl)-2-methoxy-5-nitro-
phenyl]-4-pyrazolo[1,5-a]pyridin-3-ylpyrimidin-2-amine (265 mg,
P
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
3-(2-Chloropyrimidin-4-yl)-1-methylindole.
(1H, t), 8.35 (1H, d), 8.61 (1H, s), 8.85 (1H, d), 9.46 (1H, s); m/z
(ES−) M− = 392.
N′-(2-Dimethylaminoethyl)-2-methoxy-N′-methyl-N-[4-(1-meth-
ylindol-3-yl)pyrimidin-2-yl]-5-nitrobenzene-1,4-diamine.
Sodium hydride (1.707 g, 42.68 mmol) was added portionwise to 3-(2-
chloropyrimidin-4-yl)-1H-indole (8.1684 g, 35.57 mmol) in THF (250 mL)
cooled to 0 °C. The resulting mixture was stirred at 0 °C for 30 min before
iodomethane (2.67 mL, 42.68 mmol) was added. The mixture was stirred at
0 °C for 3 h. The reaction was quenched by the addition of saturated aqueous
NaHCO3 solution (25 mL). The reaction mixture was diluted with EtOAc
(100 mL) and washed sequentially with saturated NaHCO3 (50 mL), water
(50 mL), and saturated brine (50 mL). The organic layer was evaporated to
afford crude product. The crude product was purified by flash silica
chromatography, with elution gradient 0−20% MeOH in DCM. Pure frac-
tions were evaporated to dryness to afford 3-(2-chloropyrimidin-4-yl)-1-
methylindole (8.35 g, 96%) as a pale yellow solid. 1H NMR (400 MHz,
DMSO, 30 °C) δ 3.90 (3H, s), 7.30 (2H, pd), 7.54−7.6 (1H, m), 7.82
(1H, d), 8.38−8.44 (1H, m), 8.49 (1H, s), 8.53 (1H, d); m/z (ES+)
(M + H)+ = 244.
N1,N1,N2-Trimethylethane-1,2-diamine (80 mg, 0.79 mmol) was
added to a suspension of N-(4-fluoro-2-methoxy-5-nitrophenyl)-4-(1-
methylindol-3-yl)pyrimidin-2-amine (350 mg, 0.79 mmol) and DIPEA
(0.342 mL, 1.97 mmol) in 2,2,2-trifluoroethanol (5 mL). The mixture
was heated in a microwave at 140 °C for 60 min.
The cooled reaction mixture was purified by ion exchange
chromatography, using an SCX column. The desired product was
eluted from the column using 7 M NH3/MeOH and evaporated onto
silica for silica chromatography. The crude product was purified by flash
silica chromatography, with elution gradient 0−4% 7 N NH3/MeOH
in DCM. Pure fractions were evaporated to dryness to afford N′-(2-
dimethylaminoethyl)-2-methoxy-N′-methyl-N-[4-(1-methylindol-3-
yl)pyrimidin-2-yl]-5-nitrobenzene-1,4-diamine (230 mg, 62%) as an
orange solid. 1H NMR (400 MHz, DMSO, 22 °C) δ 2.16 (6H, s), 2.45−
2.49 (2H, t), 2.86 (3H, s), 3.26 (2H, t), 3.87 (3H, s), 3.95 (3H, s),
6.85 (1H, s), 7.11 (1H, t), 7.21 (1H, d), 7.25 (1H, t), 7.52 (1H, d), 8.10
(1H, s), 8.31 (1H, d), 8.33 (1H, s), 8.36 (1H, d), 8.62 (1H, s); m/z
(ES+) (M + H)+ = 476.40.
Alternative Synthesis of 3-(2-Chloropyrimidin-4-yl)-1-methylin-
dole.
N1-(2-Dimethylaminoethyl)-5-methoxy-N1-methyl-N4-[4-(1-
methylindol-3-yl)pyrimidin-2-yl]benzene-1,2,4-triamine.
A suspension of 2,4-dichloropyrimidine (0.5 g, 3.36 mmol) and
aluminum chloride (0.183 mL, 3.36 mmol) in DME (5 mL) was stirred
at ambient temperature for 5 min. To this was added 1-methylindole
(0.429 mL, 3.36 mmol), and the mixture was heated to 80 °C for 2 h.
The cool reaction mixture was added dropwise to vigorously stirring
water (50 mL) over 5 min. Upon complete addition the mixture was
stirred for 30 min, filtered and the solid washed with water (50 mL).
The crude product was purified by flash silica chromatography,
eluting with DCM. Pure fractions were evaporated to dryness to afford
3-(2-chloropyrimidin-4-yl)-1-methylindole (0.515 g, 63%) as a white
N′-(2-Dimethylaminoethyl)-2-methoxy-N′-methyl-N-[4-(1-methylin-
dol-3-yl)pyrimidin-2-yl]-5-nitrobenzene-1,4-diamine (220 mg, 0.46 mmol),
iron (155 mg, 2.78 mmol), and ammonium chloride (17.32 mg,
0.32 mmol) were heated in ethanol (12 mL) and water (4 mL) at reflux
for 2 h. The crude product was purified by ion exchange chro-
matography, using an SCX column. The desired product was eluted
from the column using 7 M NH3/MeOH and evaporated onto silica for
chromatographic purification.
1
solid. H NMR (400 MHz, DMSO, 30 °C) δ 3.96 (3H, s), 7.36 (2H,
dqd), 7.64 (1H, dd), 7.88 (1H, d), 8.45−8.5 (1H, m), 8.56 (1H, s), 8.59
(1H, d); m/z (ES+) (M + H)+ = 244.54.
N-(4-Fluoro-2-methoxy-5-nitrophenyl)-4-(1-methylindol-3-yl)-
pyrimidin-2-amine.
The crude product was purified by flash silica chromatography, with
elution gradient 0−5% 7N NH3/MeOH in DCM. Pure fractions were
evaporated to dryness to afford N1-(2-dimethylaminoethyl)-5-methoxy-
N1-methyl-N4-[4-(1-methylindol-3-yl)pyrimidin-2-yl]benzene-1,2,4-
triamine (175 mg, 85%) as a beige foam. 1H NMR (400 MHz, DMSO,
22 °C) δ 2.17 (6H, s), 2.36 (2H, t), 2.63 (3H, s), 2.88 (2H, t), 3.74 (3H,
s), 3.88 (3H, s), 4.58 (2H, br s), 6.76 (1H, s), 7.12−7.19 (2H, m), 7.21−
7.27 (1H, m), 7.48 (1H, s), 7.51 (1H, d), 7.78 (1H, s), 8.27 (1H, d), 8.30
(1H, s), 8.42 (1H, d); m/z (ES+) (M + H)+ = 446.32.
N-[2-(2-Dimethylaminoethylmethylamino)-4-methoxy-5-[[4-(1-
4-Methylbenzenesulfonic acid hydrate (22.73 g, 119.50 mmol) was
added in one portion to 3-(2-chloropyrimidin-4-yl)-1-methylindole
(24.2679 g, 99.58 mmol) and 4-fluoro-2-methoxy-5-nitroaniline6 (18.54 g,
99.58 mmol) in 2-pentanol (500 mL). The resulting mixture was stirred
at 105 °C for 2.5 h. The mixture was cooled to room temperature. The
precipitate was collected by filtration, washed with 2-pentanol (50 mL),
and dried under vacuum to afford N-(4-fluoro-2-methoxy-5-nitro-
phenyl)-4-(1-methylindol-3-yl)pyrimidin-2-amine as a yellow solid. The
filtrate was cooled, and the precipitate was collected by filtration and
washed with 2-pentanol (10 mL). The two crops were combined and
triturated with MeCN to give a solid which was collected by filtration
and dried under vacuum to give N-(4-fluoro-2-methoxy-5-nitrophenyl)-
4-(1-methylindol-3-yl)pyrimidin-2-amine (37.4 g, 95%) as a yellow
solid. 1H NMR (400 MHz, DMSO, 30 °C) δ 3.92 (3H, s), 4.01 (3H, s),
7.13 (1H, dd), 7.27−7.36 (1H, m), 7.4−7.51 (2H, m), 7.59 (1H, d), 8.26
methylindol-3-yl)pyrimidin-2-yl]amino]phenyl]prop-2-enamide.
Acryloyl chloride (34.5 mg, 0.38 mmol) in DCM (1 mL) was added
dropwise to a stirred solution of N1-(2-dimethylaminoethyl)-5-
methoxy-N1-methyl-N4-[4-(1-methylindol-3-yl)pyrimidin-2-yl]-
benzene-1,2,4-triamine (170 mg, 0.38 mmol) and DIPEA (0.073 mL,
0.42 mmol) in DCM (5 mL), cooled in an ice/water bath. The mixture
was stirred for 90 min and then diluted with DCM (25 mL) and washed
with saturated aqueous NaHCO3 solution (50 mL). The organics were
Q
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
removed, and the aqueous portion was further extracted with DCM
(2 × 25 mL). The combined organics were dried over MgSO4, filtered,
and concentrated onto silica for purification. The crude product was
purified by flash silica chromatography, with elution gradient 0−4% 7 N
NH3/MeOH in DCM. Pure fractions were evaporated to dryness and
triturated with ether to afford N-[2-(2-dimethylaminoethylmethylamino)-
4-methoxy-5-[[4-(1-methylindol-3-yl)pyrimidin-2-yl]amino]phenyl]prop-
2-enamide (75 mg, 39%) as a cream solid. 1H NMR (400 MHz, DMSO,
22 °C) δ 2.21 (6H, s), 2.29 (2H, t), 2.72 (3H, s), 2.89 (2H, t), 3.86 (3H,
s), 3.92 (3H, s), 5.77 (1H, dd), 6.27 (1H, dd), 6.43 (1H, dd), 7.04 (1H, s),
7.15 (1H, t), 7.2−7.27 (2H, m), 7.53 (1H, d), 7.91 (1H, s), 8.24 (1H, d),
8.33 (1H, d), 8.68 (1H, s), 9.14 (1H, s), 10.22 (1H, s); 13C NMR (176
MHz, DMSO, 22 °C) δ 32.8, 42.6, 45.1, 55.7, 56.0, 56.8, 105.3, 107.1,
110.4, 112.4, 113.3, 120.8, 121.2, 121.9, 125.3, 125.5, 125.9, 127.7, 132.4,
133.8, 137.3, 137.7, 145.8, 157.6, 158.9, 159.8, 161.5, 162.3. CHN analysis,
C28H33N7O2·H2O requires C = 64.97%, H = 6.82%, N = 18.94%. Found:
C = 64.6%, H = 6.7%, N = 18.8%. m/z (ES+) (M + H)+ = 500.42. HRMS:
(M + H)+ = C28H34N7O2 = 500.276 84, found 500.276 86.
AUTHOR INFORMATION
Corresponding Author
*E-mail: ray.finlay@astrazeneca.com. Phone: +44(0)1625510379.
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Didier Danillion, Gareth Hughes, Cath
Eberlein, and Scott Martin for expert technical assistance. We are
also grateful to Anne Galer and Paula Daunt for project team
leadership. We also acknowldege Susan Galbraith for her support
of the AZD9291 program. Jamie Scott, Jason Kettle, and Kevin
Foote are also gratefully acknowledged for their helpful
comments during the preparation of this manuscript.
ABBREVIATIONS USED
N-[2-(2-Dimethylaminoethylmethylamino)-4-methoxy-5-[[4-(1-
methylindol-3-yl)pyrimidin-2-yl]amino]phenyl]prop-2-enamide Me-
sylate Salt.
■
EGFR, epidermal growth factor receptor; NSCLC, non-small-
cell lung cancer; TKI, tyrosine kinase inhibitor; GSH,
glutathione; PK, pharmacokinetics; WT, wild-type; AM,
activating mutant; DM, double mutant; LLE, ligand lipophilicity
efficiency; LCMS, liquid chromatography−mass spectrometry;
SAR, structure−activity relationship; HPLC, high-performance
liquid chromatography; DCM, dichloromethane; DMA, N,N-
dimethylacetamide; DMF, N,N-dimethylformamide; DIPEA,
N,N-diisopropylethylamine; DMSO, dimethyl sulfoxide;
TBAF, tetra-N-butylammonium fluoride; TFA, trifluoroacetic
acid; THF, tetrahydrofuran
To a stirred solution of N-[2-(2-dimethylaminoethylmethylamino)-4-
methoxy-5-[[4-(1-methylindol-3-yl)pyrimidin-2-yl]amino]phenyl]-
prop-2-enamide (5g, 9.11 mmol) in acetone (45.5 mL) and water
(4.55 mL) at 50 °C was added methanesulfonic acid (0.893g, 9.11
mmol) as a solution in acetone (4.55 mL). The resulting mixture was
stirred for 1.5 h. The resulting solid was collected by filtration and
dried at 80 °C under vacuum to give the title salt (4.9 g, 94%) as a solid
(mp(DSC) = 246−259 °C). 1H NMR (400 MHz, acetone-d6, 22 °C) δ 2.72
(3H, s), 2.96 (3H, s), 3.01 (6H, s), 3.58 (3H, t), 3.87−3.90 (7H, m), 5.76
(1H, dd), 6.38−6.53 (2H, m), 7.12 (1H, t), 7.20 (1H, t), 7.29 (1H, s), 7.40
(2H, t), 8.07−8.16 (3H, m), 8.56 (1H, s), 9.30 (1H, s), 9.60 (1H, s), 9.66
(1H, s); 13C NMR (101 MHz, acetone-d6, 22 °C) δ 34.3, 39.9, 44, 45.5, 50.1,
55.2, 57, 105.9, 107.7, 112, 113.4, 123.4, 123.9, 124.4, 124.9, 125.1, 125.2,
126.9, 129.4, 131.5, 139.9, 140.2, 142.3, 144, 152.8, 153.7, 167.1, 169.3. CHN
analysis, C29H37O5N7S (0.1% w/w H2O) requires C = 58.41%, H = 6.27%,
N = 16.44%, S = 5.38%. Found: C = 58.6%, H = 6.3%, N = 16.5%, S = 5.2%.
m/z (ES+) (M + H)+ = 500.26. HRMS: (M + H)+ = C28H34N7O2 =
500.27684, found 500.2759.
REFERENCES
■
(1) Ohashi, K.; Maruvka, Y. E.; Michor, F.; Pao, W. Epidermal growth
factor receptor tyrosine kinase inhibitor-resistant disease. J. Clin. Oncol
2013, 31, 1070−1080.
(2) Pao, W.; Chmielecki, J. Rational, biologically based treatment of
EGFR-mutant non-small-cell lung cancer. Nat. Rev. Cancer 2010, 10,
760−774.
(3) Maemondo, M.; Inoue, A.; Kobayashi, K.; Sugawara, S.; Oizumi, S.;
Isobe, H.; Gemma, A.; Harada, M.; Yoshizawa, H.; Kinoshita, I.; Fujita,
Y.; Okinaga, S.; Hirano, H.; Yoshimori, K.; Harada, T.; Ogura, T.; Ando,
M.; Miyazawa, H.; Tanaka, T.; Saijo, Y.; Hagiwara, K.; Morita, S.;
Nukiwa, T. Gefitinib or chemotherapy for non-small-cell lung cancer
with mutated EGFR. N. Engl. J. Med. 2010, 362, 2380−2388.
(4) Kobayashi, S.; Boggon, T. J.; Dayaram, T.; Jaenne, P. A.; Kocher,
O.; Meyerson, M.; Johnson, B. E.; Eck, M. J.; Tenen, D. G.; Halmos, B.
EGFR mutation and resistance of non-small-cell lung cancer to gefitinib.
N. Engl. J. Med. 2005, 352, 786−792.
Cellular Assays. Cellular phosphorylation and proliferation assays
were performed as reported previously by ourselves.6
In Vivo Antitumor Efficacy Studies. All in vivo efficacy studies
were performed as reported previously by ourselves.6
Rat in Vivo Toxicology Studies. Studies were carried out in
accordance with U.K. Home Office legislation (Animals [Scientific
Procedures] Act 1986) and AstraZeneca’s institutional policies. The
animals used were 10-week-old male RccHan:WIST rats obtained from
Harlan, U.K. Animals (n = 3/compound) received a single oral dose of
compound as a suspension in 0.5% w/v HPMC/0.1% w/v Tween in
deionized water at a concentration of 20 mg/mL. Blood glucose levels
were measured using an Accuchek Active meter (ACCU-CHEK Active;
Roche, Basel, Switzerland). Serum insulin concentrations were
determined using a commercial rat ELISA kit (Mercodia, Upsala,
Sweden). Water and food were available ad libitum.
(5) Ou, S.-H. I. Second-generation irreversible epidermal growth factor
receptor (EGFR) tyrosine kinase inhibitors (TKIs): A better mouse-
trap? A review of the clinical evidence. Crit. Rev. Oncol. Hematol. 2012,
83, 407−421.
(6) Ward, R. A.; Anderton, M. J.; Ashton, S.; Bethel, P. A.; Box, M.;
Butterworth, S.; Colclough, N.; Chorley, C. G.; Chuaqui, C.; Cross, D.
A.; Dakin, L. A.; Debreczeni, J. E.; Eberlein, C.; Finlay, M. R. V.; Hill, G.
B.; Grist, M.; Klinowska, T. C.; Lane, C.; Martin, S.; Orme, J. P.; Smith,
P.; Wang, F.; Waring, M. J. Structure- and reactivity-based development
of covalent inhibitors of the activating and gatekeeper mutant forms of
the epidermal growth factor receptor (EGFR). J. Med. Chem. 2013, 56,
7025−7048.
(7) Zhou, W.; Ercan, D.; Chen, L.; Yun, C.-H.; Li, D.; Capelletti, M.;
Cortot, A. B.; Chirieac, L.; Iacob, R. E.; Padera, R.; Engen, J. R.; Wong,
K.-K.; Eck, M. J.; Gray, N. S.; Janne, P. A. Novel mutant-selective EGFR
kinase inhibitors against EGFR T790M. Nature 2009, 462, 1070−1074.
(8) Walter, A. O.; Sjin, R. T. T.; Haringsma, H. J.; Ohashi, K.; Sun, J.;
Lee, K.; Dubrovskiy, A.; Labenski, M.; Zhu, Z.; Wang, Z.; Sheets, M.; St
Martin, T.; Karp, R.; van Kalken, D.; Chaturvedi, P.; Niu, D.; Nacht, M.;
Petter, R. C.; Westlin, W.; Lin, K.; Jaw-Tsai, S.; Raponi, M.;
ASSOCIATED CONTENT
■
S
* Supporting Information
Preparation and characterization for additional final compounds,
cell assay statistical analyses, pharmacodynamic assay procedure,
kinase selectivty data, binding mode computational modeling,
and enzyme/cell correlations. This material is available free of
charge via the Internet at http://pubs.acs.org.
R
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Van Dyke, T.; Etter, J.; Weaver, Z.; Pao, W.; Singh, J.; Simmons, A. D.;
Harding, T. C.; Allen, A. Discovery of a mutant-selective covalent
inhibitor of EGFR that overcomes T790M-mediated resistance in
NSCLC. Cancer Discovery 2013, 3, 1404−1415.
(9) (a) International Centre for Kinase Profiling. http://www.kinase-
screen.mrc.ac.uk/ (accessed August 12, 2014). (b) Compounds were
screened in the Millipore kinase panel. Full data are described in ref 20.
(10) Yee, D. Insulin-like growth factor receptor inhibitors: baby or the
bathwater? J. Natl. Cancer Inst. 2012, 104, 975−981.
(11) Clarke, E. D.; Greenhow, D. T.; Adams, D. Metabolism-related
assays and their application to agrochemical research: reactivity of
pesticides with glutathione and glutathione transferases. Pestic. Sci. 1998,
54, 385−393.
(12) MacFaul, P. A.; Morley, A. D.; Crawford, J. J. A simple in vitro
assay for assessing the reactivity of nitrile containing compounds. Bioorg.
Med. Chem. Lett. 2009, 19, 1136−1138.
(23) Eisenhauer, E. A.; Therasse, P.; Bogaerts, J.; Schwartz, L. H.;
Sargent, D.; Ford, R.; Dancey, J.; Arbuck, S.; Gwyther, S.; Mooney, M.;
Rubinstein, L.; Shankar, L.; Dodd, L.; Kaplan, R.; Lacombe, D.; Verweij,
J. New response evaluation criteria in solid tumours: Revised RECIST
guideline (version 1.1). Eur. J. Cancer 2009, 45, 228−247.
(24) Common Terminology Criteria for Adverse Events (CTCAE).
http://evs.nci.nih.gov/ftp1/CTCAE/CTCAE_4.03_2010-06-14_
QuickReference_5x7.pdf (accessed August 12, 2014).
(25) Janne, P. A.; Ramaligam, S. S.; Yang, J. C. H.; Ahn, M.-J.; Kim, D.-
W.; Kim, S.-W.; Planchard, D.; Ohe, Y.; Felip, E.; Watkins, C.; Cantarini,
M.; Ghiorghiu, S.; Ranson, M. Clinical activity of the mutant-selective
EGFR inhibitor AZD9291 in patients (pts) with EGFR inhibitor−
resistant non-small-cell lung cancer (NSCLC). J. Clin Oncol 2014, 32
(Suppl. 5), 8009.
(13) Schwobel, J. A. H.; Madden, J. C.; Cronin, M. T. D. Examination
of Michael addition reactivity towards glutathione by transition-state
calculations. SAR QSAR Environ. Res. 2010, 21, 693−710.
(14) Mah, R.; Thomas, J. R.; Shafer, C. M. Drug discovery
considerations in the development of covalent inhibitors. Bioorg. Med.
Chem. Lett. 2014, 24, 33−39.
́
(15) Marks, L.; Borland, S.; Philip, K.; Ewart, L.; Lainee, P.; Skinner,
M.; Kirk, S.; Valentin, J. P. The role of the anaesthetised guinea-pig in the
preclinical cardiac safety evaluation of drug candidate compounds.
Toxicol. Appl. Pharmacol. 2012, 263, 171−83.
(16) Eurofins. Discovery Series. KinaseProfiler Services. http://www.
eurofins.com/pharma-services/pharma-discovery-services/eurofins-
discovery-services/kinases-phosphatases/kinaseprofiler-services.aspx
(accessed August 12, 2014).
(17) Tang, W.; Stearns, R. A.; Chen, Q.; Bleasby, K.; Teffera, Y.;
Colletti, A.; Hafey, M.; Evers, R.; Dean, D. C.; Magriotis, P. A.
Importance of mechanistic drug metabolism studies in support of drug
discovery: a case study with N-sulfonylated dipeptide VLA-4 antagonists
in rats. Xenobiotica 2008, 38, 223−237.
(18) El-Kattan, A. F.; Poe, J.; Buchholz, L.; Thomas, H. V.;
Brodfuehrer, J.; Clark, A. The use of 1-aminobenzotriazole in
differentiating the role of CYP-mediated first pass metabolism and
absorption in limiting drug oral bioavailability: a case study. Drug Metab.
Lett. 2008, 2, 120−124.
(19) Manley, P. W.; Blasco, F.; Mestan, J.; Aicholz, R. The kinetic
deuterium isotope effect as applied to metabolic deactivation of imatinib
to the des-methyl metabolite CGP74588. Bioorg. Med. Chem. 2013, 21,
3231−3239.
(20) Sjogren, E.; Westergren, J.; Grant, I.; Hanish, G.; Lindfors, L.;
Lennernas, H.; Abrahamsson, B.; Tannergren, C. In silico predictions of
gastrointestinal drug absorption in pharmaceutical product develop-
ment: application of the mechanistic absorption model GI-Sim. Eur. J.
Pharm. Sci. 2013, 49, 679−698.
(21) (a) Ballard, P.; Ashton, S.; Cross, D.; Dimelow, R.; Yates, J.
Integrating the pre-clinical pharmacokinetic, pharmacodynamic and
efficacy data for AZD9291, an oral, irreversible inhibitor of EGFR
activating (EGFRm+) and resistant (EGFRm+/T790M) mutations and
an active metabolite to predict the human pharmacokinetics and
potential efficacious dose in patients. Mol. Cancer Ther 2013, 12 (11
Suppl.), B212. (b) Cross, D. A. E.; Ashton, S.; Ghiorghiu, S.; Eberlein,
C.; Nebhan, C.; Spitzler, P.; Orme, J.; Finlay, M. R. V.; Ward, R. A.;
Mellor, M.; Hughes, G.; Rahi, A.; Jacobs, V.; Red Brewer, M.; Ichihara,
E.; Sun, J.; Jin, H.; Ballard, P.; Al-Kadhimi, K.; Rowlinson, R.; Klinowska,
T.; Richmond, G.; Cantarini, M.; Kim, D.-W.; Ranson, M.; Pao, W.
AZD9291, an irreversible EGFR TKI, overcomes T790M-mediated
resistance to EGFR inhibitors in lung cancer. Cancer Discovery 2014, 4,
1046−1061.
(22) Ranson, M.; Pao, W.; Planchard, D.; Ohe, Y.; Kim, S.-W.; Kim, D.
W.; Felip, E.; Ghiorghiu, S.; Cantarini, M.; Janne, P. Preliminary results
from a phase I study with AZD9291: an irreversible inhibitor of
epidermal growth factor receptor (EGFR) activating and resistance
mutations in non-small-cell lung cancer (NSCLC). Eur. J. Cancer 2013,
49 (Suppl. 3), S15 (Abstract LBA33).
S
dx.doi.org/10.1021/jm500973a | J. Med. Chem. XXXX, XXX, XXX−XXX