Surmounting the resistance against EGFR inhibitors through the development of thieno[2,3-d]pyrimidine-based dual EGFR/HER2 inhibitors

Article abstract of DOI:10.1016/j.ejmech.2018.06.011

In light of the emergence of resistance against the currently available EGFR inhibitors, our study focuses on tackling this problem through the development of dual EGFR/HER2 inhibitors with improved enzymatic affinities. Guided by the binding mode of the

Full text of DOI:10.1016/j.ejmech.2018.06.011

Accepted Manuscript
Surmounting the resistance against EGFR inhibitors through the development of
thieno[2,3-d]pyrimidine-based dual EGFR/HER2 inhibitors
Sandra N. Milik, Amal Kamal Abdel-Aziz, Deena S. Lasheen, Rabah A.T. Serya,
Saverio Minucci, Khaled A.M. Abouzid
PII:
S0223-5234(18)30502-6
10.1016/j.ejmech.2018.06.011
EJMECH 10481
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 20 April 2018
Revised Date: 11 May 2018
Accepted Date: 4 June 2018
Please cite this article as: S.N. Milik, A.K. Abdel-Aziz, D.S. Lasheen, R.A.T. Serya, S. Minucci, K.A.M.
Abouzid, Surmounting the resistance against EGFR inhibitors through the development of thieno[2,3-
d]pyrimidine-based dual EGFR/HER2 inhibitors, European Journal of Medicinal Chemistry (2018), doi:
10.1016/j.ejmech.2018.06.011.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Surmounting the resistance against EGFR inhibitors through
the development of thieno[2,3-d]pyrimidine-based dual
EGFR/HER2 inhibitors
1
** †
2,3 †
1
Sandra N. Milik , Amal Kamal Abdel-Aziz , Deena S. Lasheen , Rabah
1
3,4
1*
A. T. Serya , Saverio Minucci , Khaled A. M. Abouzid
1
Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Ain
Shams University, Abbassia, Cairo 11566, Egypt
2
Department of Pharmacology and Toxicology, Faculty of Pharmacy, Ain
Shams University, Abbassia, Cairo 11566, Egypt
3
Department of Experimental Oncology, European Institute of Oncology,
Via Adamello 16, 20139, Milan, Italy
4
Department of Biosciences, University of Milan, Milan 20100, Italy
†
These authors contributed equally to this work
*
*
Correspondence: khaled.abouzid@pharma.asu.edu.eg
* Correspondence: Sandra.milik3@pharma.asu.edu.eg
Page 1 of 49

ACCEPTED MANUSCRIPT
1
ABSTRACT
In light of the emergence of resistance against the currently available EGFR
inhibitors, our study focuses on tackling this problem through the development of dual
EGFR/HER2 inhibitors with improved enzymatic affinities. Guided by the binding mode
of the marketed dual EGFR/HER2 inhibitor, Lapatinib, we proposed the design of dual
EGFR/HER2 inhibitors based on the 6-phenylthieno[2,3-d]pyrimidine as a core scaffold
and hinge binder. After two cycles of screening aiming to identify the optimum aniline
headgroup and solubilizing group, we eventually identified 27b as a dual EGFR/HER2
inhibitor with IC50 values of 91.7 nM and 1.2 µM, respectively. Notably, 27b
dramatically reduced the viability of various patient-derived cancer cells preferentially
overexpressing EGFR/HER2 (A431, MDA-MBA-361 and SKBr3 with IC50 values of
1
.45, 3.5 and 4.83 µM, respectively). Additionally, 27b efficiently thwarted the
proliferation of lapatinib-resistant human non-small lung carcinoma (NCI-H1975) cells,
harboring T790M mutation, with IC of 4.2 µM. Consistently, 27b significantly blocked
5
0
EGF-induced EGFR activation and inactivated its downstream AKT/mTOR/S6 signaling
pathway triggering apoptotic cell death in NCI-H1975 cells. The present study presents a
promising candidate for further design and development of novel EGFR/HER2 inhibitors
capable of overcoming EGFR TKIs resistance.
Keywords: Dual EGFR/HER2 inhibitors, EGFR inhibitors Resistance, Thieno[2,3-
d]pyrimidine
Abbreviations: Akt, Protein kinase B (PKB), also known as Akt; ALogP, Atomic logP; BSA, bovine
serum albumin; CDOCKER, CHARMm-based docker; CHARMm, Chemistry at Harvard
Macromolecular Mechanics (force field); DCM, dichloromethane; DMAP, 4-(Dimethylamino)pyridine;
DMF, Dimethylformamide; DMF-DMA, N,N-dimethylformamide dimethyl acetal; DMSO, Dimethyl
sulfoxide; EDCI, N-Ethyl-N′-(3-dimethylaminopropyl)carbodiimide; EGFR, Epidermal Growth Factor
Receptor; FDA, Food and Drug Administration; HER2, human epidermal growth factor receptor 2;
NSCLC, non-small cell lung cancer; PDB, protein data bank; RTK, receptor tyrosine kinase; SAR,
structure-activity relationship; SRB, sulforhodamine B; THF, Tetrahydrofuran; TKI, tyrosine kinase
inhibitor; TPSA, Topological polar surface area.
Page 2 of 49

ACCEPTED MANUSCRIPT
1
. Introduction
The Epidermal Growth Factor Receptor (EGFR) family belongs to receptor tyrosine
kinases (RTKs) which acts as a key regulator of diverse critical cellular functions
pertaining to cell proliferation and differentiation [1]. The EGFR family comprises four
structurally-related RTKs: EGFR (HER1), HER2, HER3 and HER4 [2–4]. Notably,
EGFR dysregulation through either receptors overexpression or activating gain-of-
function mutations [5,6], typically prompts proliferation upsurge, alters cell
differentiation, confers apoptosis resistance and assists neoangiogenesis and metastasis,
ultimately leading to tumorigenesis and progression (Supplementary Figure S1) [7,8].
Not surprisingly, EGFR dysregulation was reportedly associated with a number of the
most lethal cancers worldwide such as non-small cell lung cancer (NSCLC), colorectal
cancer, breast cancer and pancreatic cancer [5,8,9]. Consequently, promoted by the
tremendous efforts invested in developing small molecule EGFR tyrosine kinase
inhibitors (EGFR TKIs), first-generation EGFR TKIs (as gefitinib and erlotinib) are
currently used as the front-line standard therapy in treating NSCLC patients with
sensitizing EGFR mutations (Figure 1) [10]. Nevertheless, despite achieving initial
promising clinical responses, most EGFR TKIs-treated patients eventually relapse and
progress secondary or acquired resistance.
Cancer cells can develop resistance to EGFR TKIs through adopting several
mechanisms that can be classified into three categories; (1) resistance mechanisms
involving alterations in the primary drug target (EGFR), (2) alterations in the downstream
signaling pathways, (3) or bypass mechanisms employing other RTKs to compensate for
EGFR blockade [11]. The most prevalent acquired resistance mechanism is the EGFR
mutation T790M, accounting for 50-60% of the cases [12]. The second most frequent
mechanism is HER2 amplification as a bypass mechanism that accounts for around 12%
of the cases [13]. Hence, there is ultimate need for developing novel EGFR inhibitors to
overcome the untiring resistance emerged against the currently used ones. Premised on
these findings, we proposed that developing inhibitors that can dually inhibit EGFR and
HER2 and can overcome EGFR-T790M mutation as well might overwhelm the
development of resistance, and consequently provide better treatment options.
Herein, we present the development of thieno[2,3-d]pyrimidine-based dual
EGFR/HER2 inhibitors with improved anticancer activities. Two screening cycles led to
the identification of 27b with potent dual EGFR/HER2 inhibitory activity and improved
cytotoxicity against diverse patient-derived cancer cells preferentially overexpressing
EGFR or HER2 (A431, MDA-MB361 and SKBr3) as well as lapatinib-resistant human
NSCLC (NCI-H1975) cell line harboring the T790M mutation.
Page 3 of 49

ACCEPTED MANUSCRIPT
Figure 1 FDA-approved EGFR inhibitors
2
. Rationale and Design
In addition to overcoming the bypass resistance mechanism of HER2 amplification,
dual EGFR/HER2 inhibition is proposed to potentiate EGFR pathway blockade and to
enhance the anticancer activity for the following reasons [14];
(1) HER2, which has no known ligands, is the favored dimerization partner for the
activation of other members of the EGFR family. [15]
(
2) HER2-containing heterodimers are more stable with prolonged and enhanced
signaling potential. Those heterodimers demonstrate sluggish ligand dissociation,
delayed endocytosis, diminished EGFR degradation and rapid receptor recycling
[15–17], therefore HER2 overexpression usually predicts poor prognosis [5,16].
Based on these benefits, lapatinib (Figure 1) was developed to be the first FDA-
approved reversible dual EGFR/HER2 inhibitor indicated for advanced and metastatic
breast cancer [18], however, resistance to lapatinib soon developed [19]. Subsequently,
Afatinib (Figure 1), an irreversible dual EGFR/HER2 inhibitor, was approved for
metastatic NSCLC [20], which also encountered acquired resistance through the
progression of T790M mutation [21–23], hence the need to develop novel dual
EGFR/HER2 inhibitors that can combat resistance.
Our design of dual EGFR/HER2 inhibitors was based on the binding mode of
lapatinib in the ATP binding site of EGFR (Figure 2). It is assumed that lapatinib adopts a
very similar binding mode in HER2 kinase domain owing to the high sequence homology
of the two kinase domains (≈80%) [24,25] and the binding mode observed for other dual
Page 4 of 49

ACCEPTED MANUSCRIPT
inhibitors in both binding sites [26], however, small, yet critical, differences in the
sequence of the amino acids in the αC-helix of both enzymes attenuate the stability of the
active conformation of HER2 kinase domain compared to that of EGFR, which in turn
render the αC-helix of HER2 more flexible with higher accessibility to the back pocket
[26], therefore, dual EGFR/HER2 inhibitory activity mostly requires bulky 4-anilino
derivatives that can extend into the back pocket [27,28].
Based on the analysis of the binding site (Figure 2 and Figure 3), our approach to
design dual EGFR/HER2 inhibitors with the potential to overcome resistance was as
follows;
(
1) The main mechanism by which T790M mutation causes resistance to the EGFR
inhibitors is through increasing the receptor's affinity to ATP with respect to the
tyrosine kinase inhibitors [29,30]. Therefore, we proposed increasing the affinity
of the inhibitors to the binding site by using the 6-phenylthieno[2,3-d]pyrimidine
as a core scaffold and a hinge binder. The phenylthiophene is assumed to bind in
the hydrophobic pocket II fulfilling the hydrophobic interactions with the
surrounding residues (Figure
4), the increased hydrophobicity of the
phenylthiophene compared to the corresponding phenylfuran in lapatinib is
hypothesized to increase the ligand affinity through strengthening the
hydrophobic interactions contributing to the enthalpic component of the binding
free energy, in addition to an entropic contribution through its favorable
desolvation effect [31]. Meanwhile, the N-1 of the pyrimidine is proposed to
interact with Met793/EGFR or Met801/HER2 of the hinge region through
hydrogen bonds, and N-3 is expected to form a water-mediated hydrogen bond
with Thr854/EGFR or Thr862/HER2 (Figure 4).
(
2) In order to identify the optimum 4-anilino derivative that can fit into the back
pocket of EGFR and HER2 and attain desirable dual inhibitory activity in vitro
and in vivo, we constructed and screened a small targeted library of aniline
derivatives (Figure 5) that were selected based on the review of the literature of the
reported dual EGFR/HER2 inhibitors [11].
(3) Finally, with the aim of optimizing the dual inhibitory activity and the
pharmacokinetic profile of the elected compound from the first screening cycle,
we assembled and screened a small library of solubilizing groups (Figure 5
)
employing variable functionalities (such as amide, sulfonamide, carbamate, urea
and Michael acceptor moieties).
Page 5 of 49

ACCEPTED MANUSCRIPT
Back Pocket
Hydrophobic
Pocket I
Hinge
region
Hydrophobic
Pocket II
Solvent
accessible
Front pocket
Figure 2 Binding mode of lapatinib in the ATP binding site in the kinase domain of
EGFR (PDB: 1XKK). Interactions are illustrated as follows; red arrow= hydrogen bond
acceptor, green arrow= hydrogen bond donor, yellow spheres= hydrophobic interactions.
The pharmacophore model was generated by LigandScout 4.2.
Figure 3 2D interaction diagram of lapatinib in the ATP binding site in the kinase
domain of EGFR (PDB: 1XKK) showing the interacting amino acids. Interactions are
illustrated as follows; red arrow= hydrogen bond acceptor, green arrow= hydrogen bond
donor, yellow highlights= hydrophobic interactions. The diagram was generated by
LigandScout 4.2.
Page 6 of 49

ACCEPTED MANUSCRIPT
Figure 4 Design of thieno[2,3-
d
]pyrimidine-based dual EGFR/HER2 inhibitors
1
Figure 5 Screening approach for the identification of dual inhibitors. R = Library of
2
selected anilines. R = Library of selected solubilizing groups. Groups in red have
Michael acceptor functionality for the design of irreversible inhibitors.
Page 7 of 49

ACCEPTED MANUSCRIPT
3
. Results and Discussion
3
.1. Chemistry
1
The construction of the targeted library of aniline derivatives (R ) is outlined in
scheme S1 (see supplementary). The benzyloxyaniline derivatives (2a-d) (Scheme
S1a), the phenoxyaniline derivatives (4a,b) (Scheme S1b) and the 1-benzyl-1 -indazol-
-amine (6) (Scheme S1c) were prepared following similar previously reported
H
5
procedures (2a-d [32,33], 4a,b [34,35], 6 [36]). The 4-(benzo[d]thiazol-2-ylthio)-3-
chloroaniline (8) was prepared through a nucleophilic aromatic substitution (S Ar)
N
between 2-chloro-1-fluoro-4-nitrobenzene and benzo[d]thiazole-2-thiol by heating in
presence of potassium carbonate in dimethylformamide, followed by the reduction of the
nitro derivative (7) with stannous chloride dihydrate in ethyl acetate.
The synthetic route for the preparation of the target compounds (13-20) for the first
screening cycle is shown in scheme 1. Firstly, the core scaffold, 4-chloro-6-(4-
nitrophenyl)thieno[2,3-d]pyrimidine (12), was synthesized starting with a Batcho-
Leimgruber-like condensation between 4-nitrotoluene and N,N-dimethylformamide
dimethyl acetal in dimethylformamide affording the enamine (9) [37], followed by
Gewald's 2-aminothiophene synthesis through a one-pot reaction between the enamine
(
9), ethyl cyanoacetate, elemental sulfur and morpholine in ethanol giving the 2-
aminothiophene (10) [38] which was then cyclized by heating with formamide into the
thieno[2,3- ]pyrimidinone (11), eventually, (11) was activated for nucleophilic
substitution by chlorination with phosphorus oxychloride producing the activated 4-
chloro-6-(4-nitrophenyl)thieno[2,3- ]pyrimidine (12). Subsequently, the target
compounds (13-18) were obtained by nucleophilic substitution of the 4-chloro derivative
d
d
1
(
12) with varied aniline derivatives (R ) in i-propanol or n-butanol.
With the aim of further optimization of the solubilizing group, compounds 14a, 17a
and 18a were reduced into their corresponding aniline derivatives 21, 19 and 20,
respectively, using iron and ammonium chloride in ethanol/water.
2
Finally, varied solubilizing groups (R ) were installed in compounds 21-27 following
the synthetic pathways illustrated in scheme 2.
Page 8 of 49

ACCEPTED MANUSCRIPT
Scheme 1. Synthesis of Compounds 13-20
Reagents and Conditions: (a) DMF-DMA, DMF, 110 C, 48 h, 70%, (b) Ethyl cyanoacetate, Sulfur powder, morpholine,
EtOH, reflux 24 h, 85%, (c) Formamide, 200 C, 5 h, 90%, (d) POCl , reflux 24 h, 50%, (e) aniline derivatives, i-propanol,
o
o
3
reflux 24-72h, 20-80%, (f) 4a,b, n-butanol, reflux 24 h, 32-54%, (g) 6, i-propanol, reflux 48 h, 40%, (h) 8, i-propanol, reflux
Page 9 of 49
7
2 h, 43%, (i) amine, i-propanol, reflux 24-48 h, 53-80%, (j) 2a-d, n-butanol, reflux 24-48 h, 25-41%, (k) Fe, NH Cl, EtOH,
4
2 4 2
H O, reflux 3 h, 20%, (l) Fe, NH Cl, EtOH, H O, reflux 2 h, 100%.

ACCEPTED MANUSCRIPT
O
CF3
b
HN
N
Cl
O
HN
H C
3
N
S
22
O
CF3
O
CF3
HN
N
Cl
c
d
e
HN
Cl
N
a
O2N
N
S
HN
S
H3C
O
S
N
O
14a
2
3
O
CF3
HN
N
Cl
Et
O
O
CF3
O
N
HN
HN
N
Cl
S
N
2
4
H2N
S
21
O
CF3
HN
N
Cl
O
N
HN
S
25
O
CF3
HN
N
Cl
f
N
HN
S
O
O
2
6
HO
R
O
CF3
g
h
HN
N
Cl
O
HN
N
HN
2
2
7a: R=
7b: R=
S
2
7a,b
O
N
Scheme 2. Synthesis of Compounds 21-27
Reagents and Conditions: (a) Fe, NH Cl, EtOH, H O, reflux 2 h, 100%, (b) Acetic
4
2
anhydride, THF, rt, 24 h, 35%, (c) Methanesulfonyl chloride, pyridine, rt, 24 h, 35%, (d)
Ethyl chloroformate, pyridine, rt, 24 h, 40%, (e) 4-methyl-pent-2-enoic acid, EDC.HCl,
DMAP, DMF, rt, 24 h, 10%, (f) Maleic anhydride, THF, rt, 24 h, 59%, (g) Triphosgene,
THF, reflux 4 h, (h) amine, THF, reflux 24 h, 35%.
Page 10 of 49

ACCEPTED MANUSCRIPT
3
.2. Screening cycle (1): identifying the aniline derivative
3
.2.1. In vitro EGFR/HER2 kinase inhibitory activity at 10
In order to identify the aniline derivative that can afford dual inhibitory activity,
compounds 13-18 – bearing variable 4-anilino derivatives – were preliminary screened
for their dual inhibitory activity through determining their % inhibition at 10 M against
µM
µ
EGFR and HER2. Based on the results shown in Table 1, the compounds were
categorized into four groups that reflect certain structure-activity relationships (SAR) for
1
the R moiety:
(a) Compounds with good EGFR inhibition (80-86%) & good HER2 inhibition (63-
7
8
6
2%): compounds 14a, 15, 18a and 18d showed good dual inhibitory activities of
0%, 85%, 86% and 82% against EGFR, respectively, and 67%, 63%, 72% and
5% against HER2, respectively. It can be observed that all these compounds are
1
bearing bulky aniline headgroups (R ), that are capable of extending into the back
pocket such as phenoxyaniline (14a),
benzyloxyanilines (18a,d), conforming to the reported SAR of dual inhibitors
27,28].
N-benzyl indazole (15) and
[
(
b) Compounds with moderate EGFR inhibition (67-81%) & moderate HER2
inhibition (46-56%): compounds 14b, 17a, 17b and 18b showed moderate dual
inhibitory activities of 81%, 76%, 67% and 70% against EGFR, respectively, and
5
1%, 56%, 46% and 43% against HER2, respectively. Compounds 14b and 18b,
though they are structurally-related to compounds 14a and 18a with the same
bulky headgroups, showed weaker dual inhibitory activities than their respective
analogues. This discrepancy in activities can be attributed to the only structural
variability which is the substituent at the 3' position. It can be inferred that the
presence of a fluoro substituent at the 3' position – 3'-CF in 14a opposite to 3'-
3
3
CH in 14b and 3'-F in 18a opposite to 3'-Cl in 18b – greatly enhanced the dual
inhibitory activity which might be attributed to the extra hydrogen bonding that
fluoro can establish with the gatekeeper in the back pocket (Thr790/EGFR or
Thr798/HER2) (Figure 3) [39]. As for compounds 17a and 17b bearing the
aromatic amines (benzylamine for 17a and phenethylamine for 17b), they also
demonstrated moderate dual inhibitory activities despite having much less bulkier
headgroups.
(c) Compounds with good EGFR inhibition (68-87%) & poor HER2 inhibition (19-
3
5%): compounds 13a, 13c and 13d showed selectivity towards EGFR with good
EGFR inhibitory activities of 86%, 87% and 68%, respectively, while showing
poor HER2 inhibitory activities of 28%, 35% and 19%, respectively. Also
conforming to the SAR of dual inhibitors, these compounds bearing simple
monocyclic anilines failed to maintain activity against HER2, emphasizing the
importance of back pocket-binding in order to attain HER2 activity.
Page 11 of 49

ACCEPTED MANUSCRIPT
(d) Compounds with poor EGFR inhibition (18-21%) & poor HER2 inhibition (8-
1
6%): compounds 13b, 16 and 18c showed poor activities towards both enzymes
of 18%, 18% and 21% against EGFR, respectively, and 8%, 10% and 16% against
HER2, respectively.
Firstly, compound 13b showed very poor activities albeit carrying a simple
monocyclic aniline like 13a,c,d. Attempting to explain its inactivity would require
us to draw a comparison between the structures of the four structurally-related
compounds (13a-d). The main difference between the four anilines is the
substituent at the 3 position. The three EGFR-active compounds have 3-Br (13a),
3
3
-ethynyl (13c) and 3-Cl (13d), while the inactive one has a 3-CF (13b). A
common factor between the three active moieties which doesn’t apply to the
inactive one is their abilities to form halogen bonds. Br and Cl (especially when
associated with an electron-withdrawing group like the 4-F in 13d) are capable of
forming halogen bonds while F can't [40]. The ethynyl group (as a
‘supramolecular halogen’ [41]) is capable of forming a hydrogen bond equivalent
to the halogen bond [42]. Based on these assumptions, we can hypothesize that
the activity against EGFR requires the formation of a halogen bond in the
hydrophobic pocket I.
Compound 16 with the 4-(benzo[d]thiazol-2-ylthio)-3-chloroaniline also showed
very poor activity against both enzymes despite having a bulky aniline resembling
the active phenoxyaniline derivatives (14a,b). This might be attributed to the
difference between the biaryl ethers (14a,b) and the thioether (16) in the bond
angles between the linker atom and the biaryl substituents. C-S-C angle
approaches 90°, while the C-O-C angle is approximately 110° [43], therefore, the
conformation of the benzothiazolylthioaniline must be different from that of the
phenoxyaniline, and the sulfur linker might be bringing the second ring into clash
with the gatekeeper residue hindering it from properly extending into the back
pocket.
Finally, the inactivity of the benzyloxyaniline-bearing compound 18c was quite
surprising compared to the activities of the other benzyloxyaniline analogues
1
8a,b,d. It seems that the additional 4'-Cl abolished the activity which can be
ascribed to a steric clash with the binding site arose by the extra 4'-substituent.
Based on these results, compounds 14a, 15, 18a and 17a were selected for further
anticancer activity testing. These were the most active representatives from the four
classes of dually active compounds determined during this screening (14a for
phenoxyanilines, 15 for
N-benzyl indazole, 18a for benzyloxyanilines and 17a for
aromatic amines).
Page 12 of 49

ACCEPTED MANUSCRIPT
3
.2.2. In vitro antiproliferative activity against A431 and MDA-MB-361 cancer cell
lines
The anticancer activities of these compounds were further tested using the SRB
cytotoxicity assay on patient-derived epidermoid (A431) and breast (MDA-MB-361)
cancer cell lines which overexpress EGFR and HER2, respectively [44]. Lapatinib was
used as a reference compound. As shown in Table 1, compound 14a evidently reduced the
viability of A431 cells with IC of 1.72
µ
M which was comparable to that of lapatinib
M). Also, in alignment with the inhibitory effects on HER2, compounds 14a and
8a exhibited more potent cytotoxicity against MDA-MB-361 cells (with IC values of
5
0
(1.69 µ
1
5
0
3
.65 and 7.5 µM, respectively) compared to lapatinib (IC50 = 13.73 µM). Compound 15
showed minimal potency as an antiproliferative agent against A431 and MDA-MB-361
cells though it exhibited appreciated inhibitory effects on EGFR and HER2 assay in vitro
.
This could be secondary to reduced cellular uptake of this compound. Compound 17a on
the other hand demonstrated more notable cytotoxicity against A431 cells than 18a,
though not as potent as 14a, while showed minimal activity against MDA-MB-361 cells.
2
3
.2.3. Reduction of the nitro group (R )
Emerging from the previous results, we selected compounds 14a, 17a and 18a for
further optimization. With the aim of improving the kinase inhibitory activity of the
2
compounds, the reduction of the nitro group at the solubilizing group position (R ) into
2
the corresponding NH derivative was sought. The nitro group, albeit highly polar, is
considered to be "hydroneutral" rather than being hydrophilic [45], and since the
solubilizing group extends into the solvent accessible front pocket, we proposed that a
more hydrophilic moiety (such as the NH group) would offer a better physicochemical
2
complementarity.
2
As presented in Table 1, transforming the R from NO (in compounds 14a, 17a and
2
1
2
8a) into the more hydrophilic NH (compounds 21, 19 and 20, respectively)
dramatically improved the dual kinase inhibitory activities of the compounds with a more
pronounced effect on their HER2 activities, for instance, 19 exhibited superior dual
inhibitory activities of 99% (EGFR) and 96% (HER2) compared to its NO parent, 17a,
2
that displayed activities of 76% (EGFR) and 56% (HER2), and similar patterns were
observed for 20 (HER2; 90% versus 72% for 18a) and 21 (97% and 96% versus 80% and
6
7% of 14a against EGFR and HER2, respectively). These observations accentuate the
influence of optimizing the solubilizing group on the kinase inhibitory activity.
Moreover, compounds 18a (NSC: D-793249/1) and its aniline derivative 20 (NSC: D-
7
93254/1) were submitted to the Developmental Therapeutics Program (DTP) of the
National Cancer Institute (NCI) for screening against the NCI-60 panel of human cancer
cell lines. The two compounds were tested at an initial 1-dose (10 M) assay determining
µ
the growth percent of the full NCI-60 panel. The results of the assay (Supplementary
Figures S2, S3) illustrate an overall enhancement of cytotoxicity upon reducing the NO2
Page 13 of 49

ACCEPTED MANUSCRIPT
(
18a) into NH (20) subsiding from a mean growth percent of 79% (18a) to 42% (20).
2
Focusing on the lapatinib-sensitive cell lines [46] (listed in Table 2), we can notice a
consistent improvement in the cytotoxicity of 20 against these particular cell lines
compared to 18a (Table 2), again affirming the significance of optimizing the
hydrophilicity of the solubilizing group for dual inhibitory activity.
Finally, the cytotoxicity of the three compounds (19, 20 and 21) against A431 and
MDA-MB-361 cell lines was determined. In contrast to their kinase inhibitory effects,
compounds 19 and 21 exhibited anticancer activities inferior to their respective NO2
parents (17a and 14a). This highlights the need to further optimize the solubilizing group
–
balancing its hydrophilicity and the overall polarity of the compound – in order to attain
optimum cellular activities. While compounds 20 and 21 showed comparable cytotoxic
activities, compound 21 was selected for optimization through the second screening cycle
since its aniline headgroup (3-chloro-4-(3-(trifluoromethyl)phenoxy)aniline) was
previously identified through a fragment-based drug design approach as being effective
against EGFR-T790M [47].
Table 1 % inhibitory effect on EGFR/HER2 kinase activities at 10
activity against A431 and MDA-MB-361 of compounds 13-21
µM and anticancer
%
inhibition at 10
µ
M
Cytotoxicity (IC )* – µM
50
1
2
Comp
3a
R
R
EGFR
HER2
A431
MDA-MB-361
1
NO2
NO2
NO2
NO2
NO2
86
28
---
---
1
3b
18
87
68
80
8
---
---
1
3c
35
19
67
---
---
---
---
1
3d
1
.72 ± 0.03
3.65 ± 1.62
1
4a
1
4b
NO2
81
51
---
---
Page 14 of 49

ACCEPTED MANUSCRIPT
%
inhibition at 10
µ
M
Cytotoxicity (IC )* – µM
50
1
2
Comp
R
R
EGFR
HER2
A431
MDA-MB-361
>
20
>20
1
1
5
6
NO2
NO2
NO2
85
63
18
76
10
56
---
---
1
7a
4.87 ± 0.17
---
>20
1
7b
NO2
NO2
NO2
NO2
NO2
NH2
67
86
70
21
82
99
46
72
43
16
65
96
---
1
8a
11.00 ± 0.98
7.50 ± 2.94
1
8b
---
---
---
---
1
8c
1
8d
---
---
1
9
>20
>20
2
0
1
NH
2
87
97
90
96
9.76 ± 1.18
5.70 ± 0.53
5.85 ± 0.90
2
NH2
11.37 ± 2.20
Page 15 of 49

ACCEPTED MANUSCRIPT
%
inhibition at 10
µ
M
Cytotoxicity (IC )* – µM
50
1
2
Comp
R
R
EGFR
86%
HER2
94%
A431
---
MDA-MB-361
---
Staurosporine (1
µM)
Lapatinib di-p-toluenesulfonate
---
---
1.69 ± 0.26
13.73 ± 2.32
*
IC50 values are expressed as mean ± SD and were obtained using Graphpad prism
software based on the data obtained from triplicates of at least two independent
experiments
Table 2 % inhibition of 18a and 20 at 10
µM against selected NCI-60 cancer cell lines
Lapatinib-sensitive
cancer cell lines [46]
Lung EKVX
% inhibition* of 18a at 10 % inhibition* of 20 at 10 µM
µ
M
against the selected cell lines
against the selected cell lines
20%
15%
12%
0%
23%
30%
40%
80%
75%
60%
100%
70%
70%
85%
Lung NCI-H322M
Ovarian IGROV1
Ovarian SK-OV-3
Renal ACHN
Renal TK-10
Breast MDA-MB-468
*
Approximated values calculated from the growth percent values shown in the one dose
mean graph provided by NCI-developmental therapeutics program (Supplementary
Figures S2, S3)
3
.3. Screening cycle (2): optimizing the solubilizing group
3
.3.1. In vitro EGFR/HER2 kinase inhibitory activity at 10
µM
During the second screening cycle, we were on a quest to find the solubilizing group
that could afford the desirable balance between the required hydrophilicity for optimum
kinase activity and the hydrophobicity needed for cellular permeability. Deriving from
compound 21, a series of compounds (22-27) decorated with varied solubilizing groups
2
(
R ) was synthesized and their kinase inhibitory activities at 10
µM were determined
(Table 3). Compounds 26 and 27b maintained inhibitory activities comparable to 21,
while 22 and 23 only exhibited similarly potent EGFR inhibition (98% and 92%,
Page 16 of 49

ACCEPTED MANUSCRIPT
respectively) and much less potent HER2 inhibition (74% and 65%, respectively).
Finally, 24, 25 and 27a showed reduced dual inhibitory activities compared to 21 with a
more prominent decline in HER2 inhibition (52%, 46% and 44%, respectively). One
factor that might be contributing to this pattern of activity is the hydrophilicity of the
solubilizing group. Evidently the solubilizing groups of compounds 24, 25 and 27a have
the highest logP values in this series (Supplementary Table S1), and so they showed the
least potent HER2 activities, while 26 and 27b have lower logP values closer to 21, hence
their potent HER2 inhibitory activities. Apparently HER2 inhibition, and accordingly
dual inhibition, requires certain hydrophilicity in the solubilizing group.
3
.3.2. In vitro antiproliferative activity against A431 and MDA-MB-361 cancer cell
lines
Like the first series of compounds, the anticancer activities of the second series of
compounds were also determined using SRB cytotoxicity assay on A431 and MDA-MB-
3
61 cancer cell lines. As demonstrated in Table 3, compounds 27a and 27b diminished
A431 viability with IC50 values of 2.34 and 1.45 M, respectively, which were
µ
comparable to that of lapatinib. Additionally, they exhibited superior antiproliferative
activities on MDA-MB-361 cells as compared to lapatinib with IC values of 3.83 and
5
0
3
.5 µM, respectively. The strong anticancer activities of compound 27a on A431 and
MDA-MB-361 cells are inconsistent with its kinase assay data. This could be due to the
inhibitory effects of compound 27a on other crucial prosurvival signalling kinases
involved in A431 and MDA-MB-361 proliferation, since the cyclopropyl urea moiety is
reported to be encountered in other kinase inhibitors (like the multikinase inhibitor
lenvatinib, or the Aurora kinase inhibitor AT9283) [48]. Not only did the cytotoxicity of
2
7b conform to its kinase inhibition, but also in alignment with its evident
antiproliferative activity on MDA-MB-361 cells, 27b induced apoptotic cell death in
those cells in a concentration-dependent pattern as illustrated by Annexin V/PI staining
(
Figure 6). Interestingly, compound 25 demonstrated remarkable antiproliferative activity
on A431 cells compared to lapatinib with IC value of 0.23 M, while showed minimal
M). Compounds 22 and 23 showed dual
cellular inhibition though with lower potencies on A431 compared to lapatinib with IC50
values of 10.5 and 5.25 M, respectively. Consistently, compounds 21, 22 and 27b
significantly decreased the proliferation of SKBr3 - another human breast cancer cell line
which overexpresses HER2 [44] - with IC50 values of 6, 4.7 and 4.83 M respectively
Table 5). Despite its potent kinase inhibitory activities, compound 26 showed negligible
µ
5
0
activity against MDA-MB-361 cells (IC50 >20
µ
µ
µ
(
antiproliferative activities against A431 and MDA-MB-361 cells, this might be attributed
to its high polarity hampering its cell permeability. Finally, compound 24 showed
minimal cytotoxicity against both cancer cell lines even though it exhibited appreciated
Page 17 of 49

ACCEPTED MANUSCRIPT
inhibitory effects on EGFR and HER2 which could also be ascribed to the reduced
cellular uptake of this compound.
Live cells
1
00
Early apoptotic cells
Late apoptotic and necrotic
cells
5
0
0
Figure 6 Compound 27b induces apoptosis in human breast cancer (MDA-MB-361)
cells. Twenty four hours after treatment with 27b, MDA-MB-361 cells were subjected to
Annexin V/PI staining and percent of live cells, early as well as late apoptotic and
necrotic cells were analysed using flow cytometry.
3
.3.3. Potential EGFR/HER2 kinase inhibitory activity (IC ) determination
50
Hinging upon the previous assays, five compounds (21, 22, 23, 27a and 27b) that
maintained dual inhibitory activities both enzymatically and cellular were selected for
kinase IC determination. Their dose-related EGFR/HER2 inhibition at 5 concentrations
5
0
(1 nM, 10 nM, 100 nM, 1 µM and 10 µM) were measured in order to determine their IC50
values. The five compounds showed EGFR IC50 in the nanomolar range, and HER2 IC50
in low to sub-micromolar range (Table 4), except for the two compounds 23 and 27a that
exhibited less potent HER2 inhibition with IC50 values >20 µM. Eventually, the three
compounds 21, 22 and 27b were identified as potent dual EGFR/HER2 inhibitors with
good cytotoxic activities against cancer cell lines overexpressing EGFR and/or HER2,
representing promising candidates.
3
.3.4. Inhibitory activities of tested compounds on EGFR/HER2 signaling in human
cancer cells
In addition to the kinase inhibitory and anticancer activities of the tested compounds,
we further evaluated whether these compounds could inhibit EGFR/HER2 signalling
pathways in A431 and SKBr3 cells which overexpress EGFR and HER2, respectively.
Compounds 21, 22 and 27b were selected for further experiments. A431 cells were serum
starved overnight and then incubated with varying concentrations of 21 and 27b for 3h
Page 18 of 49

ACCEPTED MANUSCRIPT
before being stimulated with EGF. Potent and concentration-dependent inhibition of
phosphorylation of EGFR was observed after treatment with 21 and 27b (Figure 7).
Moreover, 21 and 27b evidently inhibited EGF-induced activation of downstream
signalling effectors as shown by reduced phosphorylation of AKT, p70 S6 kinase and
ribosomal S6 (Figure 7). Furthermore, 22 inactivated AKT/mTOR/S6 signalling in HER2
overexpressing SKBr3 cells (Figure 8).
Figure 7 Effect of compounds 21 and 27b on EGF-induced EGFR phosphorylation
in EGFR overexpressing human epidermoid carcinoma (A431) cells. A431 cells were
serum-starved overnight and then treated with the indicated concentrations of compounds
2
1 and 27b for 3h before being stimulated with EGF for 10 min. Representative western
blot analysis of inhibition of EGFR and its downstream targets by 21 and 27b in A431
cell line. β-actin served as the loading control.
Page 19 of 49

ACCEPTED MANUSCRIPT
Figure 8 Effect of compound 22 on AKT/mTOR/S6 signalling pathway in HER2
overexpressing human breast cancer (SKBr3) cells. Representative western blot
analysis of AKT/mTOR/S6 signalling in cell lysates of SKBr3 treated with the indicated
concentrations of 22 for 72h. β-actin served as the loading control.
3
.3.5. Inhibition of EGFR signaling in lapatinib-resistant non-small lung cancer
NCI-H1975) cells harboring T790M mutation
(
Drug resistance is a critical and frequently encountered issue in cancer therapy, and
T790M mutation in EGFR has been reported to contribute to lapatinib, gefitinib and
erlotinib resistance [29]. Hence, development of novel EGFR inhibitors is ultimately
needed. We then questioned whether these compounds could inhibit the growth of human
non-small lung cancer cells (NCI-H1975) which harbour EGFR T790M mutation.
Intriguingly, compound 27b dramatically impeded NCI-H1975 proliferation with IC50
value of 4.2 µM (Figure 9A and Table 5) and increased % of apoptotic and necrotic cells
in NCI-H1975-treated cells (Figure 9B).
We then examined the ability of compound 27b to inhibit EGFR T790M signalling
pathway in NCI-H1975 cells. Indeed, compound 27b robustly inactivated AKT/S6K/S6
signalling in treated NCI-H1975 cells in a concentration-dependent pattern (Figure 9C). In
addition, compound 27b thwarted EGF-induced activation of EGFR and its downstream
signalling molecules including AKT, p70 S6 kinase and ribosomal S6 in NCI-H1975
cells which were serum starved overnight and then incubated with compound 27b for 2.5
h followed by EGF stimulation (Supplementary Figure S4).
Page 20 of 49

ACCEPTED MANUSCRIPT
Figure 9 Apoptotic effects of compound 27b in human non-small cell lung
carcinoma (NCI-H1975) cells harbouring T790M mutation are associated with
inhibition of EGFR downstream signalling effectors (AKT/S6K/S6). A) Anticancer
activity of compound 27b against NCI-1975 cells following 72h of treatment with the
indicated concentrations were assessed using SRB cytotoxicity assay. B) Seventy two
hours after treatment with 27b, NCI-H1975 cells were subjected to Annexin V/PI
staining and percent of live cells, early as well as late apoptotic and necrotic cells were
analysed using flow cytometry. C) Representative western blot analysis of AKT/S6K/S6
signalling in cell lysates of NCI-H1975 treated with the indicated concentrations of 27b
for 72h. β-actin was used as the loading control.
Page 21 of 49

ACCEPTED MANUSCRIPT
Table 3 % inhibitory effect on EGFR/HER2 kinase activities at 10 µM and anticancer
activity against A431 and MDA-MB-361 of compounds 22-27
%
inhibition at 10
µ
M
Cytotoxicity (IC )* – µM
50
Comp
R
EGFR
HER2
A431
MDA-MB-361
O
1
0.50 ± 1.41
10.45 ± 3.06
2
2
2
2
3
4
98
74
N
H
O
H
N
5.25 ± 0.35
7.00 ± 0.98
9.86 ± 2.74
>20
O S
92
85
65
52
H
O
N
1
O
H
N
0
.23 ± 0.04
>20
2
2
5
6
81
92
76
46
98
44
O
O O
H
>20
>20
HO
N
H
N
H
N
2
1
.34 ± 0.17
3.83 ± 0.23
2
7a
O
O
N
HN
H
.45 ± 0.21
---
3.50 ± 0.73
2
7b
100
97
N
O
Staurosporine (10
µ
M)
100%
100%
---
---
Lapatinib di-p-
toluenesulfonate
---
1.69 ± 0.26
13.73 ± 2.32
*
IC50 values are expressed as mean ± SD and were obtained using Graphpad prism
software based on the data obtained from triplicates of at least two independent
experiments
Page 22 of 49

ACCEPTED MANUSCRIPT
Table 4 IC values of compounds 21, 22, 23, 27a and 27b against EGFR and HER2
5
0
Compound ID
EGFR IC50 (nM)
HER2 IC50
1.5 µM
0.879
(µM)
2
2
2
1
2
3
21.4 nM
47.7 nM
47.1 nM
112 nM
91.7 nM
47 nM
µ
M
19.2 µM
37.1 µM
1.2 µM
2
7a
7b
Staurosporine
2
60.8 nM
Table 5 Anticancer activity of tested compounds (21, 22 and 27b) against SKBr3 and
NCI-H1975 cancer cell lines
Cytotoxicity (IC )* – µM
5
0
Compound ID
SKBr3
NCI-H1975
13.90 ± 2.40
12.69 ± 3.98
4.20 ± 0.19
11.46 ± 2.45
2
1
2
6.00 ± 1.10
4.70 ± 2.08
4.83 ± 1.06
0.17 ± 0.04
2
27b
Lapatinib di-p-toluenesulfonate
*
IC50 values are expressed as mean ± SD and were obtained using Graphpad prism
software based on the data obtained from triplicates of at least two independent
experiments
3
.4. Molecular modeling studies
3
.4.1. Molecular docking study
Molecular docking was performed using CDOCKER protocol in Accelrys Discovery
®
Studio 2.5. This docking study was performed to gain insight into the SAR of the target
compounds, and as an attempt to interpret the observed enzymatic activities of the tested
compounds on the basis of the ligand-protein interactions. The enzymatic activity of
kinase inhibitors depends on the ability of the compound to properly dock into the
binding site and to establish strong enough interactions with the key amino acids (ligand
affinity) in order to compete with the ATP for the binding site. Accordingly, the active
compounds in this study should attain the same binding mode observed for lapatinib
(Figure 2 and Figure 3) and kinase inhibitors in general, and to establish the key
Page 23 of 49

ACCEPTED MANUSCRIPT
interactions with the binding site, while the inactive compounds would fail either to
achieve the correct pose and/or to establish the desired interactions. The results of the
docking study (Supplementary Table S2) will be discussed in order of the activity of the
compounds as follows:
•
Compounds with good EGFR inhibition (80-86%) & good HER2 inhibition (63-
2%): compounds 14a, 15 and 18a showed the same binding modes as lapatinib in
7
both EGFR (PDB: 3POZ) and HER2 (PDB: 3RCD) (exemplified by the binding
modes of compound 21 in Figure 10 and Figure 11), that are also conforming to the
proposed binding mode (Figure 4), where the hinge binder is correctly positioned in
the hinge region properly directing the aniline headgroup into the back pocket and
conveniently extending the solubilizing group towards the solvent accessible region.
Additionally, the three compounds successfully established all the key interactions
with the binding site of EGFR, such as the HB of N1 with the hinge region (Met793),
and the HB of N3 through the water bridge, also the π-+ interaction of the aniline
ring with the Lys745 that interrupts the salt bridge with the Glu762 and hampers the
active protein conformation, in addition to a π-σ interaction between the thiophene
ring and Leu718 (hydrophobic region II), also the nitro group formed a charge
interaction with the Asp800 in the solvent accessible region. Similar binding modes
and interactions were observed in HER2, in addition to the extra HB of the CF (14a)
3
and 3-F (18a) with the gatekeeper residue Thr798 in the back pocket, which explains
their higher activities on HER2 compared to other analogues.
•
Compounds with moderate EGFR inhibition (67-81%) & moderate HER2
inhibition (46-56%): represented by compound 17a, the compound showed the same
binding mode and interactions in EGFR but it was missing the key HB with Met793,
which might explain its lower EGFR activity, while the docking protocol could only
generate a flipped pose for the compound in the binding site of HER2 with the amine
and the solubilizing groups exchanging positions, which could also be a probable
explanation for the lower HER2 inhibition.
•
•
Compounds with good EGFR inhibition (68-87%) & poor HER2 inhibition (19-
3
5%): compounds 13a, 13c and 13d achieved similar binding modes and interactions
like lapatinib with both EGFR and HER2. Their poor activities on HER2 couldn’t be
explained through their docking results which seem to be about right, however, in
our experience, HER2 inhibitors usually have aniline headgroups extending into the
back pocket which they don’t have.
Compounds with poor EGFR inhibition (18-21%) & poor HER2 inhibition (8-
1
6%): Unexpectedly, compound 13b showed the correct pose and the correct
interactions in both EGFR and HER2 comparable to the active compounds, however,
the docked pose is of a high energy conformer (CHARMm Energy= 9.58482)
Page 24 of 49

ACCEPTED MANUSCRIPT
(compared to its global minimum conformer with CHARMm Energy of 4.41552),
hence its minimal inhibition. Compound 16 showed a skewed pose in EGFR due to
the improper fitting of the benzothiazole in the back pocket, so it failed to establish
any of the essential interactions, while in HER2 it failed to achieve the correct pose
also due to the lower flexibility of the benzothiazolylthioaniline that hinders the
proper positioning of the hinge binder and so the remaining moieties. Again,
compound 18c showed the correct pose and all the proper interactions in EGFR
though it is inactive on EGFR which might be attributed to that the docked pose is of
a high energy conformer (CHARMm Energy= 32.1588) compared to the global
minimum with CHARMm Energy of 14.1322. On the other hand, 18c showed a
flipped pose in HER2, failing to achieve the proper pose and interactions.
Figure 10 Docking pose of compound 21 in EGFR (PDB: 3POZ) along with its 2D
interaction diagram illustrating the interacting amino acids. Interactions are illustrated as
follows; red arrow= hydrogen bond acceptor, yellow spheres= hydrophobic interactions.
The diagrams were generated using LigandScout 4.2.
Page 25 of 49

ACCEPTED MANUSCRIPT
Figure 11 Docking pose of compound 21 in HER2 (PDB: 3RCD) along with its 2D
interaction diagram illustrating the interacting amino acids. Interactions are illustrated as
follows; red arrow= hydrogen bond acceptor, yellow spheres= hydrophobic interactions.
The diagrams were generated using LigandScout 4.2.
3
.4.2. Molecular properties affecting the cellular activity
The cellular activity of an enzymatically active compound is affected by other factors
governing the cellular permeability of the compound. The balance between the
hydrophobicity/hydrophilicity of the compound is one of the main factors affecting its
cellular permeability. The hydrophobicity/ hydrophilicity of a compound can be
described in terms of different molecular properties like, for instance, the LogP. In order
to identify the molecular properties affecting the cellular activity of our target
compounds, the correlation between the cellular activity (Log(A431_IC )) and different
5
0
molecular properties was calculated and a correlation matrix was generated using
Accelrys Discovery Studio 2.5 (Supplementary Figure S5). The correlation matrix
shows that the molecular properties of highest correlation with the cellular activity are the
LogP and the Polar Surface Area (PSA), where both properties directly reflect the degree
of hydrophobicity/hydrophilicity of the compounds, assuring the influence of the
hydrophobicity on the cellular permeability and hence the cellular activity. The two
properties (Supplementary Table S3) were plotted against Log cellular activity. The
plots show a positive correlation between the cellular activity and the ALogP (Figure 12),
Page 26 of 49

ACCEPTED MANUSCRIPT
and a negative correlation between the cellular activity and the PSA (Figure 13), meaning
that increasing the ALogP and/or decreasing the PSA of the compounds would increase
the cellular activity.
Figure 12 The correlation between the cellular activity and ALogP of the compounds
Figure 13 The correlation between the cellular activity and PSA of the compounds
Page 27 of 49

ACCEPTED MANUSCRIPT
4
. Conclusion
We developed a series of thieno[2,3-
with the aim of overcoming the resistance against EGFR inhibitors. Building on the 6-
phenylthieno[2,3- ]pyrimidine as a core scaffold and a hinge binder, we screened a small
d]pyrimidine-based dual EGFR/HER2 inhibitors
d
targeted library of aniline derivatives in order to identify the aniline headgroup that can
attain optimum dual inhibition. Compounds 14a and 18a, bearing the bulky aniline
headgroups
(3-chloro-4-(3-(trifluoromethyl)phenoxy)aniline
and
3-chloro-4-((3-
fluorobenzyl)oxy)aniline, respectively) that can bind efficiently in the back pocket (as
proposed through the docking study), showed the best dual inhibitory activities both
enzymatically (with % inhibition at 10 µM against EGFR of 80% and 86%, respectively,
and against HER2 of 67% and 72%, respectively) and cellularly (IC against A431 of
5
0
1
.72 µM and 11 µM, respectively, and against MDA-MB-361 of 3.65 µM and 7.5 µM,
respectively). The impact of increasing the hydrophilicity of the solubilizing group on the
dual inhibitory activities of the compounds was demonstrated through the improved
activity of the aniline analogue 21 (97% and 96% inhibition against EGFR and HER2 at
1
0 µM, respectively) compared to its NO2 parent 14a, however, its cellular
antiproliferative activity was hampered by the increased hydrophilicity (IC50 of 11.37 µM
and 5.85 µM against A431 and MDA-MB-361, respectively) with respect to its NO2
parent. Therefore, a small library of diverse solubilizing groups was screened in order to
optimize the enzymatic and cellular activities of 21. Throughout a series of enzymatic
and cell-based assays, 27b maintained potent dual EGFR/HER2 inhibitory activity with
consistently robust anticancer cellular activities. By virtue of its solubilizing group (1-(2-
morpholinoethyl)urea), 27b provided the desirable balance between the degree of
hydrophilicity of the solubilizing group required for optimum dual enzymatic activity,
and the overall extent of polarity of the compound needed to attain optimum cellular
activities. 27b showed potent dual EGFR/HER2 inhibition with IC of 91.7 nM and 1.2
5
0
µM, respectively. Additionally, 27b induced apoptotic cell death in MDA-MB-361 cells
with IC50 of 3.5 µM and an IC50 of 1.45 µM against A431, exhibiting superior anti-
proliferative activities compared to lapatinib against these cell lines. Moreover, 27b
potently inhibited the EGF-induced activation of EGFR as well as its downstream
signaling effectors (AKT/S6K/S6) in A431 cells in a concentration-dependent manner.
Finally, in accordance with our proposed design to tackle the T790M mutation, 27b
dramatically impeded the proliferation of NCI-H1975 cell line, harboring the T790M
mutation, with IC50 of 4.2 µM. Evidently, 27b inhibited the EGFR T790M signaling
pathway in NCI-H1975 cells as illustrated by western blot analysis. On these grounds,
2
7b represents a promising candidate for further development of dual EGFR/HER2
inhibitors that can overcome the resistance against currently used EGFR inhibitors.
Page 28 of 49