Refinement of covalent EGFR inhibitor AZD9291 to eliminate off-target activity

Article abstract of DOI:10.1016/j.tetlet.2021.153178

Non-small-cell lung cancer (NSCLC) is a major disease that accounts for 85% of all lung cancer cases which claimed around 1.8 billion lives worldwide in 2020. Tyrosine kinase inhibitors (TKIs) that target EGFR have been used for the treatment of NSCLC, but often develop drug resistance, and the covalent inhibitor AZD9291 has been developed to tackle the problem of drug resistance mediated by the T790M EGFR mutation; however, there is a side effect of hyperglycemia that may be due to off-target activity. This study examines analogs of AZD9291 by chemical proteomics, identifying analogs that maintain T790M-EGFR engagement while showing reduced cross-reactivity with off-targets.

Full text of DOI:10.1016/j.tetlet.2021.153178

Tetrahedron Letters xxx (xxxx) xxx
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Refinement of covalent EGFR inhibitor AZD9291 to eliminate off-target
activity
⇑
⇑
Elise Bouffard, Balyn W. Zaro, Melissa M. Dix, Benjamin Cravatt , Chi-Huey Wong
Department of Chemistry, The Scripps Research Institute, 10550 N. Torrey Pines Rd., La Jolla, CA 92037, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Non-small-cell lung cancer (NSCLC) is a major disease that accounts for 85% of all lung cancer cases which
claimed around 1.8 billion lives worldwide in 2020. Tyrosine kinase inhibitors (TKIs) that target EGFR
have been used for the treatment of NSCLC, but often develop drug resistance, and the covalent inhibitor
AZD9291 has been developed to tackle the problem of drug resistance mediated by the T790M EGFR
mutation; however, there is a side effect of hyperglycemia that may be due to off-target activity. This
study examines analogs of AZD9291 by chemical proteomics, identifying analogs that maintain
T790M-EGFR engagement while showing reduced cross-reactivity with off-targets.
Received 29 March 2021
Revised 3 May 2021
Accepted 7 May 2021
Available online xxxx
Keywords:
EGFR tyrosine kinase inhibitor
Chemical proteomic analysis
Proteomic reactivity
Lysosome
Ó 2021 Elsevier Ltd. All rights reserved.
Covalent inhibitors
Analogs synthesis
Introduction
for off-target activity with the cysteine-dependent cathepsin pro-
teases [4].
The tyrosine kinase domain of EGFR is a well-known target for
the treatment of non-small-cell lung cancer (NSCLC) and several
generations of inhibitors have been developed. However, resis-
tance against these therapies has emerged, especially due to the
T790M mutation. To overcome this problem, new irreversible inhi-
bitors have been developed to target an active-site cysteine resi-
due, and the compound AZD9291 developed by AstraZeneca
belongs to this class of inhibitors [1] (Fig. 1). This orally active,
potent T790M-EGFR inhibitor has been approved by the FDA as a
second line treatment for patients with T790M EGFR NSCLC. How-
ever, in the phase I/II clinical trials, this compound has shown an
unusual side effect of hyperglycemia [2], suggesting that
AZD9291 has off-target inhibition activity.
Since cysteines perform many functions in diverse protein
classes, irreversible EGFR inhibitors may affect the activity of pro-
teins outside of the kinome, potentially leading to target-indepen-
dent side effects. Indeed, the Cravatt laboratory has demonstrated
that some irreversible EGFR inhibitors, including AZD9291, have
off-target reactivity across the broader proteome [3,4]. AZD9291,
in particular, accumulates in the lysosome, which may account
In an attempt to improve the selectivity of AZD9291, we modi-
fied key groups in the structure of the compound and evaluated the
proteomic reactivity of the analogs in NSCLC cells expressing
T790M-EGFR. Our goal was to identify hot spots for further refine-
ment of the compound. While other groups have studied the
impact of changing the reactive group on the AZD9291 scaffold
[5], we chose to retain the same Michael acceptor and instead
modify other sites. Based on AstraZeneca’s studies on T790M-EGFR
inhibitors that led to the discovery of AZD9291 [6], as well as the
propensity of compounds with basic amines to accumulate in the
lysosome (the organelle that harbors a major set of cathepsin off-
targets of AZD9291), we decided to explore the consequences of
modifying the diamino group, as well as other groups on the com-
pound (the methoxy and methylindole moieties) (Fig. 1). The
pyrimidine group is essential for binding to T790M-EGFR so we
chose to keep this region intact, along with the acrylamide respon-
sible for covalent interaction with C797 in the EGFR active site.
Results and discussion
Analogs with 1-methylindole or 1-H indole as R1 were obtained
by reaction of either a 1-methylindole or 1-H indole with 2,4-
dichloropyrimidine. The former was introduced by a Friedel-Craft
reaction using aluminum trichloride whereas the latter was intro-
duced by an SNAr reaction of the indole, deprotonated with methyl
⇑
Corresponding authors.
E-mail addresses: cravatt@scripps.edu (B. Cravatt), wong@scripps.edu
(C.-H. Wong).
https://doi.org/10.1016/j.tetlet.2021.153178
0040-4039/Ó 2021 Elsevier Ltd. All rights reserved.
Please cite this article as: E. Bouffard, B.W. Zaro, M.M. Dix et al., Refinement of covalent EGFR inhibitor AZD9291 to eliminate off-target activity, Tetrahe-
dron Letters, https://doi.org/10.1016/j.tetlet.2021.153178

E. Bouffard, B.W. Zaro, M.M. Dix et al.
Tetrahedron Letters xxx (xxxx) xxx
N
R1
N
N
NH
R3
N
NH
N
N
MeO
R2
O
O
N
H
N
H
N
AZD9291
NH
N
R1 =
R2 = H, OMe
R3 = H or
N
N
N
N
O
N
N
N
O
N
Fig. 1. Structure of AZD9291 and library compounds.
magnesium bromide, on the 4-position of the 2,4-dichloropyrimi-
dine. The second chlorine was then substituted by various
nitroanilines bearing either a hydrogen or a methoxy group as R2
and either a hydrogen or a fluorine in the 4 position. In the latter
case, the fluorine was substituted by different amines as R3 groups.
The nitro group was reduced into an amine, using hydrogen and
palladium on carbon or iron powder and ammonium chloride, in
order to introduce the Michael acceptor moiety (Fig. 2).
acceptor. After elimination of the chlorine with triethylamine in
acetonitrile, the final compounds were obtained (Fig. 3f).
Using this synthetic route, the following compounds were
obtained (Fig. 4).
The in situ activity of the various analogs was evaluated using
the NSCLC lung cancer line H1975, which harbors a T790M-EGFR
variant, by competitive activity-based protein profiling (ABPP)
[3]. Cells were first treated with an inhibitor analog or DMSO vehi-
cle (1 and 10 mM). After 2 h, the cells were treated with the alkyne-
bearing AZD9291 probe (Fig. 5). After 4 h, cells were lysed, and
membrane and soluble proteomes isolated and subject to treat-
ment with rhodamine-azide by copper(I)-catalyzed azide-alkyne
cycloaddition [7], as described previously [4]. Probe-labeled pro-
teins were then separated by SDS-PAGE and visualized by in-gel
fluorescence scanning (Fig. 6). In this competitive ABPP experi-
ment, a compound is considered inhibitory against T790M-EGFR
or off-targets if it blocks the AZD9291 probe labeling of these pro-
teins at 1 or 10 mM test concentrations (reflected in a loss of in-gel
fluorescence signal for these proteins).
These studies demonstrated that replacing the diamino group
with other basic amines had variable impact on EGFR and cathep-
sin reactivity, but did not furnish analogs with superior selectivity.
In contrast, an analog where the diamino group was removed along
with removal of the N-methyl of the indole unit (EB037) main-
tained T790M-EGFR engagement at 1 and 10 mM while showing
reduced reactivity with cathepsins. In contrast, analog EB017,
which also lacked the diamino group, but maintained the N-methyl
indole lost interactions with both EGFR and cathepsins. We should
also note that other off-targets of AZD9291 were retained with
EB037 (e.g., ~100 kDa soluble protein).
The introduction of the acrylamide group in the final step
proved to be more tedious than expected for some of the analogs.
With only one equivalent of acryloyl chloride, the reaction was
incomplete and the starting material and the product could not
be separated (Fig. 3a). The use of a slight excess of acryloyl chloride
lead to a mixture of product and a by-product bearing 2 acrylamide
groups, also non-separable (Fig. 3b). The use of acrylic acid and a
coupling agent, T3P, gave the product with non-separable
impurities. The same issue occurred when using acrylic acid and
triphenylphosphite (Fig. 3c). To avoid the formation of the bis-
acrylamide compound, the synthesis route was modified to protect
the secondary amine with a Boc group. However, during the depro-
tection of the Boc group, the acrylamide was partially cleaved lead-
ing to a mixture of products that were not separable (Fig. 3d). We
also used tert-butyl acrylate and ethyl acrylate in an attempt to
achieve selective introduction on the primary amine, but no reac-
tion was observed (Fig. 3e). We then decided to use 2 equivalents
of acryloyl chloride to introduce the acrylamide and hydrolyze the
acrylamide on the guanidine-like nitrogen. Several acidic
conditions were tested and a solution of HCl 3 M in AcOEt was
determined to be optimal, despite the formation of another
by-product resulting from the addition of a chlorine on the Michael
2

E. Bouffard, B.W. Zaro, M.M. Dix et al.
Cl
Tetrahedron Letters xxx (xxxx) xxx
R1
NO2
R2
R1
X
N
N
N
AlCl3, DME,
N
N
N
Cl
Cl
H2N
80 °C, 16 h
NO2
R2
82%
X
Cl
N
N
APTS, 2-pentanol,
105 °C 18 h 30
80-95%
NH
MeMgBr, DCE,
N
N
N
Cl
H
N
RT, 16 h
70%
1a
1b
R
R
1 = H
1 = H, R2 = H, X = H
2a
R
R
R
R
2 = Me
1 = Me, R2 = OMe, X = H
2b
2c
2d
1 = Me, R2 = H, X = F
1 = Me, R2 = OMe, X = F
R3-NH, DIEA,
DMA, 140 °C, 3 h
if X = F
49-90%
R1
R1
R1
N
N
N
O
NO2
R2
NH2
R2
NH
H2, Pd/C,
R3
R3
R3
N
N
R
N
MeOH, RT, 16 h
42-94%
N
N
N
N
H
N
N
H
H
R2
N
N
1 = Me, R2 = H, R3
=
=
3c
3d
3e
3f
4a
4b
R
R
1 = H, R2 = H, R3 = H
1 = Me, R2 = OMe, R3 = H
N
N
R
1 = Me, R2 = H, R3
N
N
N
R
1 = Me, R2 = OMe, R3
1 = Me, R2 = OMe, R3
=
=
N
N
4c
4d
4e
4f
R
1 = Me, R2 = H, R3
=
=
N
R
R
1 = Me, R2 = H, R3
N
N
O
1 = Me, R2 = OMe, R3
1 = Me, R2 = OMe, R3
=
=
3g
3h
R
R
1 = Me, R2 = OMe, R3
=
N
N
O
N
R
R
1 = Me, R2 = OMe, R3
=
N
N
N
O
4g
4h
R
1 = Me, R2 = OMe, R3
1 = Me, R2 = OMe, R3
=
=
N
R
N
O
Fig. 2. Synthetic route for the variation of R1, R2 and R3.
propensity to enter the lysosome and cross-react with cathepsins.
Further work is required to identify compounds with optimal
selectivity for T790M-EGFR over cathepsins, as well as to further
improve selectivity against other off-targets in the proteome. Work
continues to develop next generation covalent inhibitors against
T790M EGFR with reduced off-target reactivity and to explore
the lysosomotropic properties of the new compounds using estab-
lished assays [8].
Conclusion
In summary, this study revealed that modification of the
AZD9291 scaffold can reduce off-targets of the parent compound,
while maintaining reactivity with T790M-EGFR. In particular, our
findings indicate that the basic amine in AZD9291 is not required
for engagement of T790M-EGFR and replacement of this group
with suitable neutral analogs may yield compounds with lower
3

E. Bouffard, B.W. Zaro, M.M. Dix et al.
Tetrahedron Letters xxx (xxxx) xxx
R1
N
R1
R1
R1
R1
R1
R1
N
N
N
N
N
N
N
N
N
N
O
N
N
N
HCl 3M in AcOEt
RT, 16 h
NEt3, ACN,
+
+
+
N
NH
R3
N
NH
N
NH
N
N
N
NH
N
NH
N
NH
R3
R2
80 °C, 16 h
8-54% over 3 steps
R2
R2
R2
R2
R2
R2
O
O
O
O
O
O
N
H
NH2
N
H
N
H
N
H
N
H
N
H
R3
R3
R3
R3
R3
Cl
a) Acryloyl chloride 1 eq,
Acryloyl chloride 2 eq,
DIPEA, THF, RT, 16 h
f)
DIPEA, THF, RT, 16 h
R1
R1
R1
R1
N
N
N
N
e)
ethyl acrylate
or
b) Acryloyl chloride excess,
tert-butylacrylate
DIPEA, THF
DIPEA, THF, RT, 16 h
N
N
O
N
N
+
N
NH
N
N
N
NH
N
NH
R3
R2
R2
R2
R2
O
O
O
N
H
N
H
NH2
N
H
d) 1) Boc2O, DMAP,
THF, reflux, 20 h
37-67%
R3
R3
R3
c)
Acrylic acid, T3P,
DIPEA, THF
or
2) Pd/C, H2
,
Acetic acid,
MeOH, RT
44%
Acrylic acid, P(OPh)3
Pyridine
,
R1
N
R1
R1
N
R1
R1
N
N
N
TFA, CH2Cl2
,
N
N
O
N
O
N
N
or
Acryloyl chloride,
HCl 3M in AcOEt
N
NH
R3
N
N
O
N
N
O
N
NH
R3
N
NH
R3
+
R2
R2
R2
R2
R2
DIPEA, THF
58%
O
O
O
N
NH2
N
H
NH2
N
H
H
R3
R3
Fig. 3. Different strategies attempted to introduce the acrylamide.
N
N
N
NH
N
N
N
N
NH
NH
N
NH
N
N
MeO
N
N
NH
N
MeO
O
O
O
O
N
H
N
H
N
N
H
H
N
EB037
EB017
EB126
EB113
N
N
N
N
N
N
N
N
N
NH
NH
NH
N
MeO
N
MeO
N
MeO
NH
N
N
MeO
O
O
O
O
N
N
H
N
H
N
H
H
N
N
N
O
N
O
N
EB138
EB159
EB104
EB137
Fig. 4. Library of biologically tested compounds.
4

E. Bouffard, B.W. Zaro, M.M. Dix et al.
Tetrahedron Letters xxx (xxxx) xxx
Fig. 5. Structure of AZD9291 probe and Rhodamine azide for competitive activity-based protein profiling of AZD9291 analogs.
Fig. 6. Gel-based competitive activity-based protein profiling (ABPP) of analogs of AZD9291. H1975 cells were treated at indicated concentrations of inhibitor (AZ = AZD9291)
or DMSO vehicle (2 h) prior to incubation with the alkyne AZD9291 probe (1 M, 4 h). Cells were then lysed and their membrane and soluble proteomes subject to Cu(I)-
catalyzed azide-alkyne cycloaddition with rhodamine-azide, SDS-PAGE, and in-gel fluorescence analysis.
l
Declaration of Competing Interest
References
[1] M.R.V. Finlay, M. Anderton, S. Ashton, P. Ballard, P.A. Bethel, M.R. Box, R.H.
Bradbury, S.J. Brown, S. Butterworth, A. Campbell, C. Chorley, N. Colclough, D.A.
E. Cross, G.S. Currie, M. Grist, L. Hassall, G.B. Hill, D. James, M. James, P. Kemmitt,
T. Klinowska, G. Lamont, S.G. Lamont, N. Martin, H.L. McFarland, M.J. Mellor, J.P.
Orme, D. Perkins, P. Perkins, G. Richmond, P. Smith, R.A. Ward, M.J. Waring, D.
Whittaker, S. Wells, G.L. Wrigley, J. Med. Chem. 57 (2014) 8249–8267.
[2] I. Sullivan, D. Planchard, Ther Adv Respir Dis. 10 (2016) 549–565.
[3] B.R. Lanning, L.R. Whitby, M.M. Dix, J. Douhan, A.M. Gilbert, E.C. Hett, T.O.
Johnson, C. Joslyn, J.C. Kath, S. Niessen, L.R. Roberts, M.E. Schnute, C. Wang, J.J.
Hulce, B. Wei, L.O. Whiteley, M.M. Hayward, B.F. Cravatt, Nat. Chem. Biol. 10
(2014) 760–767.
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
Acknowledgement
This work was supported by Kwang-Hwa Foundation and the
National Institutes of Health (CA231991).
[4] S. Niessen, M.M. Dix, S. Barbas, Z.E. Potter, S. Lu, O. Brodsky, S. Planken, D.
Behenna, C. Almaden, K.S. Gajiwala, K. Ryan, R. Ferre, M.R. Lazear, M.M.
Hayward, J.C. Kath, B.F. Cravatt, Cell Chem. Biol. 24 (2017) 1388–1400.e7.
[5] M. Sato, H. Fuchida, N. Shindo, K. Kuwata, K. Tokunaga, G. Xiao-Lin, R. Inamori,
K. Hosokawa, K. Watari, T. Shibata, N. Matsunaga, S. Koyanagi, S. Ohdo, M. Ono,
A. Ojida, ACS Med. Chem. Lett. 11 (2020) 1137–1144.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.tetlet.2021.153178.
5

E. Bouffard, B.W. Zaro, M.M. Dix et al.
Tetrahedron Letters xxx (xxxx) xxx
[6] R.A. Ward, M.J. Anderton, S. Ashton, P.A. Bethel, M. Box, S. Butterworth, N.
Colclough, C.G. Chorley, C. Chuaqui, D.A.E. Cross, L.A. Dakin, J.É. Debreczeni, C.
Eberlein, M.R.V. Finlay, G.B. Hill, M. Grist, T.C.M. Klinowska, C. Lane, S. Martin, J.
P. Orme, P. Smith, F. Wang, M.J. Waring, J. Med. Chem. 56 (2013) 7025–7048.
[7] V.V. Rostovtsev, L.G. Green, V.V. Fokin, K.B. Sharpless, Angew. Chem. 114 (2002)
2708–2711.
[8] S. Nadanaciva, S. Lu, D.F. Gebhard, B.A. Jessen, W.D. Pennie, Y. Will, Toxicol. In
Vitro 25 (2011) 715–723.
6