Synthesis, biological evaluation, QSAR and molecular dynamics simulation studies of potential fibroblast growth factor receptor 1 inhibitors for the treatment of gastric cancer

Article abstract of DOI:10.1016/j.ejmech.2016.10.066

Accumulating evidence suggests that fibroblast growth factor receptor 1 (FGFR1) is an attractive target in gastric cancer therapy. Based on our previous discovery of two non-ATP competitive FGFR1 inhibitors, A114 and A117, we designed and screened a series of compounds with the framework of bisaryl-1,4-dien-3-one. Among them, D12 and D15 exhibited the most potent FGFR1 inhibitory activity, which was ATP-independent. Furthermore, a quantitative structure-activity relationship analysis of 41 analogs demonstrated that the specific structural substitutions alter their bioactivities. Molecular docking and dynamics simulation analysis indicated the hydrophobic interaction at the FGFR1-D12/D15 interaction was dominant. Evaluation for anti-gastric cancer efficacy of D12 and D15 indicated effective inhibition of cell proliferation, apoptosis induction and cell cycle arrest. Thus, these two FGFR1 inhibitors have therapeutic potential in the treatment of gastric cancer, and this study provides will contribute to the rational design of novel non-ATP competitive FGFR1 inhibitors.

Full text of DOI:10.1016/j.ejmech.2016.10.066

Accepted Manuscript
Synthesis, biological evaluation, QSAR and molecular dynamics simulation studies
of potential fibroblast growth factor receptor 1 inhibitors for the treatment of gastric
cancer
Shilong Ying, Xiaojing Du, Weitao Fu, Di Yun, Liping Chen, Yuepiao Cai, Qing Xu,
Jianzhang Wu, Wulan Li, Guang Liang
PII:
S0223-5234(16)30937-0
10.1016/j.ejmech.2016.10.066
EJMECH 9031
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 27 August 2016
Revised Date: 12 October 2016
Accepted Date: 31 October 2016
Please cite this article as: S. Ying, X. Du, W. Fu, D. Yun, L. Chen, Y. Cai, Q. Xu, J. Wu, W. Li, G. Liang,
Synthesis, biological evaluation, QSAR and molecular dynamics simulation studies of potential fibroblast
growth factor receptor 1 inhibitors for the treatment of gastric cancer, European Journal of Medicinal
Chemistry (2016), doi: 10.1016/j.ejmech.2016.10.066.
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
Synthesis, biological evaluation, QSAR and molecular dynamics
simulation studies of potential Fibroblast Growth Factor Receptor 1
inhibitors for the treatment of gastric cancer
Shilong Yinga1, Xiaojing Duab1, Weitao Fua1, Di Yuna, Liping Chena, Yuepiao Caia,
Qing Xuc, Jianzhang Wua*, Wulan Liad*, Guang Lianga
Graphical abstract

ACCEPTED MANUSCRIPT
Synthesis, biological evaluation, QSAR and molecular dynamics simulation
studies of potential Fibroblast Growth Factor Receptor 1 inhibitors for the
treatment of gastric cancer
Shilong Yinga1, Xiaojing Duab1, Weitao Fua1, Di Yuna, Liping Chena, Yuepiao Caia, Qing Xuc,
Jianzhang Wua*, Wulan Liad*, Guang Lianga
1 These authors contribute to this work equally.
(aChemical Biology Research Center, School of Pharmaceutical Sciences, Wenzhou
Medical Universtiy, Wenzhou, Zhejiang 325035, China )
(bDepartment of Digestive Diseases, the First Affiliated Hospital of Wenzhou Medical
University, Wenzhou, Zhejiang 323000, China. )
(cCollege of Chemistry and Materials Engineering, Wenzhou University , Wenzhou,
Zhejiang 325035, China)
(dCollege of Information Science and Computer Engineering, Wenzhou Medical
University, Wenzhou, Zhejiang 325035, China)
*Correspondence authors: Jianzhang Wu
Address: Chemical Biology Research Center, School of Pharmaceutical Sciences, Wenzhou Medical
Universtiy, Wenzhou, Zhejiang 325035, China
E-mail: wjzwzmu@163.com
*Correspondence authors: Wulan Li
Address: Chemical Biology Research Center, School of Pharmaceutical Sciences, Wenzhou Medical
Universtiy, Wenzhou, Zhejiang 325035, China
E-mail: lwlwzmu@163.com
Abstract
Accumulating evidence suggests that fibroblast growth factor receptor 1 (FGFR1) is an
attractive target in gastric cancer therapy. Based on our previous discovery of two non-ATP
competitive FGFR1 inhibitors, A114 and A117, we designed and screened a series of
compounds with the framework of bisaryl-1,4-dien-3-one. Among them, D12 and D15 exhibited
the most potent FGFR1 inhibitory activity, which was ATP-independent. Furthermore, a
quantitative structure-activity relationship analysis of 41 analogs demonstrated that the specific
structural substitutions alter their bioactivities. Molecular docking and dynamics simulation
analysis indicated the hydrophobic interaction at the FGFR1-D12/D15 interaction was dominant.
Evaluation for anti-gastric cancer efficacy of D12 and D15 indicated effective inhibition of cell
proliferation, apoptosis induction and cell cycle arrest. Thus, these two FGFR1 inhibitors have
therapeutic potential in the treatment of gastric cancer, and this study provides will contribute to
the rational design of novel non-ATP competitive FGFR1 inhibitors.

ACCEPTED MANUSCRIPT
Keyword: Design; Fibroblast growth factor receptor 1; Gastric cancer; Quantitative structure-
activity relationship
1. Introduction
Gastric cancer (GC) is the fifth most common malignancy and the third leading cause of
cancer-related death worldwide [1, 2]. Although surgery procedures and chemotherapy have
improved greatly in recent years, the clinical outcomes with treatment of gastric cancer remain
poor due to absence of early detection of the cancer [3]. For advanced gastric cancer patients, a
combination of surgery and traditional chemotherapy is still the mainstream remedy [4, 5].
However, the severe and irreversible side-effects of chemotherapy limit their therapeutic use [6,
7]. Thus, development of newly targeted therapy for gastric cancer with minimal side-effects is
highly desirable. At present, the only small molecule inhibitor for GC therapy available on the
market is Apatnib, an inhibitor of vascular endothelial growth factor receptor (VEGFR) which
was approved by the China Food and Drug Administration (CFDA) in 2014 [8, 9]. In recent
years, an increasing number of signaling molecules are considered as novel targets against cancer
[10]. Of these, Fibroblast growth factor receptor 1 (FGFR1) is considered to be an important
potential target for treating GC [3, 11, 12].
FGFR1, belonging to the superfamily of receptor tyrosine kinases (RTK), has been
intensively studied as a drug target for cancer [13, 14]. The binding of fibroblast growth factor
(FGF) and FGFR triggers the dimerization of the extracellular receptor domains, trans-
phosphorylation of the intracellular kinase domain, and the subsequent activation of downstream
signaling pathways, the mitogen activated protein kinase (MAPK) and phosphoinositide-3-
kinase (PI3K)/Akt [15, 16]. Increased activity of FGFR1 is closely linked with malignant
tumors, such as lung cancer [17], gastric adenocarcinoma [12], renal cell carcinoma [18], and
breast cancer [19]. Aberrant FGFR1 activity as a result of mutations, gene amplifications, or
chromosomal translocations are reported as a key driver of tumor progression by promoting
proliferation, angiogenesis, survival, migration and invasion of cancer cells [20, 21]. Recently,
Oki et al. found that FGFR1 expression is upregulated in GC tissue specimens [22], as are
several GC cell lines (MKN-74, MKN-45, BGC-823, MGC-803, and SGC-7901) [23].
At present, several small molecule inhibitors of FGFR1 are being evaluated in pre-clinical
or clinical trials as anti-gastric cancer agents [13, 24, 25]. In one study, NVP-BGJ398, a pan-
FGFR inhibitor showed strong inhibition of proliferation of KKLS cells which have high FGFR1
expression, and built with low inhibition on TMK-1 cells expressing low levels of FGFR1 [25].
Another potent FGFR1 inhibitor, AZD4547, has been completed a phase 2 clinical trial assessing
its efficacy and safety in advanced gastric or gastro-oesophageal junction cancer patients
(NCT01795768). However, the development of current FGFR inhibitors as anti-cancer drugs
remains limited. Most of these inhibitors belong to ATP-competitive inhibitors, which bind to
and inhibit the relatively conserved ATP binding domain of RTKs. Such inhibition of FGFR
often results in significant side-effects and toxicity in clinical and pre-clinical studies [26-28].

ACCEPTED MANUSCRIPT
Thus, the development of non-ATP-competitive FGFR inhibitors can be a more selective
approach to overcome the above difficulties.
Previously, we identified two bisaryl-1,4-dien-3-one-containing non-ATP competitive
FGFR1 inhibitors, A114 and A117, and characterized their therapeutic potency in non-small cell
lung cancer in vitro and in vivo studies [29]. To identify FGFR1 inhibitors with greater efficacy
for gastric cancer, we analyze the structure-activity relationship of analogs of bisaryl-1,4-dien-3-
one-containing FGFR1 inhibitors. We developed four series of analogs by a rational modified
strategy based on the inhibitory activity screening results (Figure 1). We also established a
QSAR model to identify the pharmacophore. Firstly, we altered the cyclopentanone ring in the
structure of A114 and A117 into various ketones (the A series compounds) to investigate
whether the middle-linking chain was essential. Moreover, further detailed modifications were
made on optimization of one of the phenyl substituents to develop the B series compounds. The
latter was the unilateral 4-nitrogen atom-containing moiety, which significantly enhanced the
activity of the compound. To confirm the effect of this interesting active fragment, the C series
compounds, characterized by a common para-position nitrogen atom-containing heterocyclic
group, was produced. By removing the propionyl or isobutyryl fragment to produce an exposed
hydroxyl, the D series compound was designed, which was expected to effectively elevate the
activity by forming hydrogen bond donors and acceptor.
In this study, we successfully screened and found two active FGFR1 inhibitors, D12 and
D15, by kinase inhibition assay and confirmed their non-ATP competitive inhibitory
characteristic. Biological assessment indicated effective suppression of GC proliferation and
induction of apoptosis. This finding constructs an appropriate QSAR model of non-ATP
competitive FGFR1 inhibitors for the first time, and provides two potential targeted therapeutic
agents against the GC.
(Insert Figure 1 here)
2. Results and discussion
2.1. Chemistry
The synthetic routes of the 41 derivatives of A114 and A117 were shown as Scheme 1. The
synthesis of A1-A7 started from corresponding ketones 1 (1a: tetrahydro-4H-pyran-4-one, 1b:
cyclohexanone, 1c: acetone, 1d: 4-Piperidone hydrochlorides hydrate). The intermediates 3a-3g
were obtained through a step of the classic Claisen-Schmidt condensation reaction, which
applied HCl gas as catalyst, and then reacted with propionyl or isobutyryl chloride under Et3N
catalyzing to prepare the compounds A1-A7. The procedures to prepare B1-B14, C1-C5 and D1-
D13 were similar, in which cyclopentanone 1e or cyclohexanone 1b was activated with
morpholine to generate the enamines 5e or 5b. The ethanol solution of 5e or 5b with vanillin was
o
reacted under 78 C, adjusted the pH to acidic conditions to obtain the key intermediates 7e or
7b. Next, the solutions were condensed with different substitutional aromatic aldehyde for
preparing the desired D1-D13, 8e1-8e14 and 8b1-8b5. Further, acylation of 8e1-8e14 and 8b1-

ACCEPTED MANUSCRIPT
8b5 with propionyl or isobutyruyl in the presence of Et3N produced the B1-B14 or C1-C5. The
general process for synthesizing D14-D15 was by a three-step reaction. Protection of Vanillin
with 3,4-dihydro-2H-pyran (THP) yielded the corresponding benzaldehyde 10. Through a
unilateral aldol condensation between 10 with acetone in a catalyst of 20% NaOH achieved the
conversion to unilateral unsaturated ketone 11. Finally, D14-D15 was obtained by Claisen-
Schmidt condensation between 11 and corresponding benzaldehydes in the presence of 40%
NaOH and the deprotection of THP in an acidic environment. All compounds were purified by
column chromatography. The structures of A-D series compounds were shown in Scheme 1, and
confirmed by spectral analysis by high-resolution mass spectrometry (HRMS), nuclear magnetic
resonance (1H-NMR and 13C-NMR) spectroscopy.
(Insert Scheme1 here)
2.2. The inhibitory activity of synthetic compounds on FGFR1 kinase
The synthesized bisaryl-1,4-dien-3-one-containing compounds were initially screened for
inhibition of FGFR kinase activity at a concentration of 20 ꢀM by a Mobility Shift Assay. As
shown as in the Scheme 1, compound B14 exhibited the most potent inhibitory activity of
FGFR1 kinase, and approximately half of the compounds (18/41) exhibited inhibitory ratio
greater than 50%. These compounds were chosen for the further determination of half-maximal
inhibitory concentration (IC50).
Based on the inhibitory potency of all compounds (at 20 ꢀM), a primary structure-activity
relationship (SAR) was analyzed. For the A series compounds, the middle aliphatic ketone linker
exerted a crucial influence on the inhibitory activity of compounds, changing the middle ketone
into either pyrone or piperidone, abolished the potent action of A114 and A117. Based on our
previous molecular docking study [29], the middle connecting ketone group of A114 and A117
was likely sandwiched in a hydrophobic cleft of FGFR1 kinase domain. An aliphatic ketone
linker can greatly stabilize the inactive conformation of kinase domain, as well as provide the
selective inhibition on FGFR1.
It’s noteworthy that compound (B1), which adopted a cross position exchange between the
propionyloxyl and methoxyl group, or a replacement of the propionyloxyl group with a methoxyl
group (B2), resulted in a complete activity loss, which demonstrating that the inhibitory effect of
A114 and A117 was closely related to the substituents on their benzene rings. Whether
introducing electron-donating (methoxyl, ethyoxyl) or electron-withdrawing substituents (F
atom) in one benzene ring of A114 and A117, either contributed little to activity, other than with
the 4-nitrogen atom-containing substituents. Significantly, results indicated that compounds
bearing 4-pyrrolidyl or 4-morpholinyl (B7, B13 and B14) exhibited outstanding potency among
the B series compounds, providing an attractive active fragment for the design of novel structural
FGFR1 inhibitors based on the bisaryl-1,4-dien-3-one scaffold. The screening for the asymmetric
derivatives (B, C, and D series) revealed that introducing the 4-nitrogen atom-containing
substituents could be useful to enhance the inhibition efficacy. The C series compounds

ACCEPTED MANUSCRIPT
displayed a high level of inhibitory activity, with inhibitory ratio of IR % and a range of 84.3%
to 94.0% at a compound concentration of 20 ꢀM. Moreover, compounds with the 4-nitrogen
atom-containing group on the unilateral benzene ring or a 3-substituted indole moiety effectively
suppressed FGFR1 kinase activity (D9-D15, IR %: 50.7% - 87.1%, 20 ꢀM).
2.3. Quantitative evaluation of structure-activity relationship (QSAR)
A QSAR model was constructed to further investigate the SAR of the derivatives of
A114 and A117. Small molecules can be represented as molecular descriptors including
constitutional descriptors, topological descriptors, auto-correlation descriptors, charge
descriptors, molecular properties, which capture and magnifies distinct aspects of chemical
structures. Scatter plot of the predicted activity versus experimental values was displayed in
Figure 2. Relative high regression coefficient was obtained from the ChemoPy descriptor
calculation program and R program. It had a high adjusted squared regression coefficient (R2
=
adj
0.8196) and described more than 83.31% of variances in the experimental activity. Equation
listed in the figure showed that MoRSEC27 (weighted by atomic charge) and bcute5 (weighted
by atomic Sanderson electronegativities) created the major influence in the model. The inhibitory
activities of molecules were negatively regulated by MoRSEC27, and positively regulated by
bcute5, while S35 (Sum of E-State of atom type: dO) showed little relevance. Consequently, the
electronegativity may exert a critical influence on the FGFR1 inhibitory activity of compounds.
The combined SAR and QSAR results may provide useful information for the future
development of novel FGFR1 kinase inhibitors.
(Insert Figure 2 here)
2.4. D12 and D15 inhibit FGFR1 through an ATP-independent mechanism
In our early research, we identified the non-ATP competitive inhibitory mode of A114
and A117. Thus, the effect of ATP concentration on the potency of inhibition of FGFR1
activation by D12 and D15 was evaluated by a Caliper Mobility Shift Assay to determine
whether the inhibitory mechanism was the same as with the parent compound. With increasing
ATP concentration, the rate of phosphorylation of FGFR1 substrate gradually increased,
reaching a plateau with ATP at 1048 ꢀM. In the presence of D12 up to 20 ꢀM, most of
phosphorylation was abolished, and further increase in ATP concentration, up to 4192 ꢀM, had
no effect on the inhibitory potency of D12 (Figure 3), The finding supported a likely non-ATP
competitive inhibition mechanism. We previously reported that the inhibition effect of
PD173074, an ATP-competitive FGFR1 inhibitor, markedly reduced with increasing ATP
concentration, indicating competition between ATP with the inhibitor [29]. Similarly, we
evaluated D15 using the ATP-competitive assay and found ATP-independent mechanism of
inhibition. Based on these results, we conclude that the mechanism of action of D12 and D15
was in an ATP-independent, and consistent with their leads A114 and A117.

ACCEPTED MANUSCRIPT
(Insert Figure 3 here)
2.5. Molecular docking and molecular dynamics simulation
We investigated interactions between the compounds and the kinase domain of FGFR1
by molecular docking analysis of the active compound D12/D15. We docked D12/D15 into the
inactive conformation of FGFR1 kinase using our well-established molecular docking method
which is based on the FGFR1-ARQ069 complex structure for reference (PDB Code: 3RHX)
[28]. From an overall determination of root mean square deviations (RMSD) and estimated free
energy, three poses of docking results were chosen as initial structures for molecular dynamics
(MD) simulation. For D12, 93% poses with RMSD less than 2 Å, and the top ranking pose was
chosen. For D15, there were two types of binding modes, one occupied 23% and others were
72%. Hence, two poses of D15 were chosen as initial structures for MD simulation.
The RMSD of the receptor in FGFR1/D12 and FGFR1/D15 complexes was converged after
10-20 ns in 50 ns MD simulations, and the RMSD of backbone atoms was ≈ 1-2 Å and ≈ 4-5 Å,
respectively (Figure 4A-B). Binding energies for D12/D15 were calculated by MM-GBSA using
500 snapshots from 40-50 ns stable trajectories. Two poses of D15 were carried out for MD
simulation. One showed -43.8067 ± 2.6494 kcal/mol binding energy, another was -26.9956 ±
3.6117 kcal/mol. Therefore, D12 and the significantly lower binding energy (high binding
affinity) pose of D15 was chosen for further energy decomposition. In Figure 4C and 4D, the
results showed that D12 and D15 shared seven identical key residues: PHE498, LEU630,
ALA640, MET535, ASP641, VAL492 and ILE545. Interestingly, except ASP641, all these
residues were hydrophobic amino acids, indicating that D12 and D15 likely interact primarily
with non-polar residues in FGFR1 to exert their inhibitory activity. Kinases can exist as active
and/or inactive states in cells. The hydrophobic residues in kinases interact to form “hydrophobic
clusters” that stabilize the inactive conformation and interfere with ATP binding [26, 33].
Targeting the auto-inhibited conformation (inactive conformation) of the kinase by the inhibitors
represents an attractive approach due to the possibility of discovering more selective inhibitor
[33]. For example, ARQ069, previously identified to be a non-ATP-competitive inhibitor of
FGFR1/2, binds and inhibits the auto-inhibited conformation of FGFR1, and displays favorable
kinase selectivity [28].
Figure 5A shows the whole molecule of D12 was sandwiched in a hydrophobic cleft
between the residues PHE489, ALA640, LEU630, VAL492, VAL561, IEL545, and PHE642.
The phenyl ring of indolyl in D12 created p-p stacking with the benzene side chain of PHE489 to
stabilize the binding conformation, and another phenyl ring occupied the main hydrophobic
pocket by making hydrophobic contacts with residues ILE545, MET535, PHE642, and ALA640.
The results indicated that D12 formed one hydrogen bond with the side chain of PHE642. We
compared modeling of the binding of D12 into FGFR1 with the FGFR1-ARQ069 co-crystal
complex structure. D12 was observed to possess highly similar interactions with ARQ069, while
ARQ069 only made two hydrogen bond interactions with the kinase domain, the major binding
feature of ARQ069 was hydrophobic interactions, other than hydrogen bond interactions. The

ACCEPTED MANUSCRIPT
simulation results of D15 was almost identical to D12 (Figure 5B). The left phenyl ring of D15
formed p-p stacking with PHE489 as well. In addition, the right phenyl ring was wrapped in a
hydrophobic pocket comprised of VAL561, ILE545, ALA640, and MET535. An extraordinary
phenomenon was that no hydrogen bond existed between D15 with the kinase domain,
suggesting that D15 interacted with FGFR1 kinase by only forming stable hydrophobic
interactions. In summary, a critical hydrophobic interaction between D12 (and D15) with FGFR1
kinase domain was predicted by a rational method combined with molecular docking and
molecular dynamics simulation. These methods clarified the interaction mechanisms of D12 and
D15 with FGFR1. The further verification of co-crystallization of FGFR1-D12 and FGFR1-D15
is in progress.
(Insert Figure 4 here)
(Insert Figure 5 here)
2.6. D12 and D15 suppress the growth of GC cell lines
Aberrant activation of FGFR1 is correlated with tumor growth and development in a number
of different cancers, including gastric cancer. The most potent compounds for evaluation of GC
growth were selected based on their IC50 values. Most compounds showed favorable inhibitory
activities using single concentration screening, with only four compounds showing relative low
kinase IC50 (<10 ꢀM) (Figure 6). The effect of the four FGFR1 kinase inhibitors (C5, D12, D14,
and D15) on three GC cell lines was determined using the MTT assay. The GC cell lines were
treated with each of the compounds, or with A114, A117, AZD4547 and Nordihydroguaiaretic
acid (NDGA), a non-ATP competitive FGFR1 inhibitor, as the positive controls for 72 hours.
Among all tested compounds, D12 and D15 exhibited the greatest inhibitory activity of the
MGC803 and BGC823 cell lines, with IC50 values of D12 and D15 lower than 10ꢀM in growth
inhibition, and the anti-proliferative efficacies greater than their leads (Figure 6).
(Insert Figure 6 here)
2.7. D12 and D15 inhibit the phosphorylation of FGFR1 and the downstream signaling
FGF binding to FGFR1 induces the phosphorylation of FGFR1 and activates downstream
pathways, including extracellular signal-regulated kinase (ERK). ERK is a member of mitogen
activated protein kinase (MAPK) and crucial for cellular proliferation and differentiation [34].
We determined the inhibitory effects of D12 and D15 on the phosphorylation of FGFR1 and
ERK by Western blot analysis. Figure 7 showed that D12 and D15 significantly suppressed the
FGF2-induced phosphorylation of the threonine (T653/654) of FGFR1 and ERK activation in a
dose-dependent manner.
(Insert Figure 7 here)

ACCEPTED MANUSCRIPT
2.8. D12 and D15 arrest the cell cycle at G0/G1 phase
We further examined whether D12 and D15 possessed cell cycle arrest potential. The
SGC7901 cells were treated with different concentrations of D12 and D15 (5 ꢀM, 10 ꢀM, 20ꢀM)
for 8 h, and cell cycle population distribution was measured by flow cytometry. Results indicated
that D12 and D15 both induced cell cycle G0/G1 phase arrest in a concentration-dependent
manner, and dramatically reduced the cell population of the G2/M phase (Figure 8A). The up-
regulation of cyclin D1 can be an oncogene driving the cell-cycle regulation [35], which has
been reported to overexpressed in gastric cancer cells [36]. Therefore, we determined the
expression level of cyclin D1 by Western blot analysis. Results indicated that both D12 and D15
down-regulated cyclin D1 expression in a dose-dependent manner (Figure 8B). In summary,
D12 and D15 induced cell cycle arrest of gastric cancer cells, which was attributed to at least in
part by reducing cyclin D expression.
(Insert Figure 8 here)
2.9. D12 and D15 induce apoptosis of GC cell lines
We determined whether the GC cell death induced by D12 and D15 was attributed to
apoptosis through the mitochondria-mediated apoptotic pathway. The morphological changes of
cell nuclei was determined by Hoechst staining, and the expression of apoptosis-related
molecules, i.e., cleaved-PARP and anti-apoptotic factor Bcl-2, was detected by Western blot
analysis. Following 12 h treatment of GC cells with compounds, Hoechst staining showed
chromatin condensation (strong blue fluorescence) and nucleus fragmentation (Figure 9A). At
increasing higher concentrations of the compounds concentration, chromatin condensation
developed in a concentration-dependent manner. For Western blot analysis, SGC7901 cells
were incubated with different concentrations of D12, D15 (5 ꢀM, 10 ꢀM, 20 ꢀM) or 10 ꢀM of
A117. Results indicated that the expression of cleaved-PARP was increased, which was
accompanied by decreased expression of Bcl-2, both in a dose-dependent manner (Figure 9B).
Note that as low as 5ꢀM of D12 and D15 was sufficient to induce changes of the protein
expression. Compared with A117, D12 and D15 achieved the similar effects at 10ꢀM (Figure
9B). These results indicated that D12 and D15 were effective inducers of GC cells apoptosis.
(Insert Figure 9 here)
3. Conclusion
In summary, with A114 and A117 as lead compounds, four series of derivatives were
designed and synthesized. Among them, D12 and D15 were screened by kinase inhibition assay,
and identified as non-ATP competitive inhibitors of the kinase domain of FGFR1. QSAR
analysis confirmed that the inhibition potency was highly correlated with the skeleton structure

ACCEPTED MANUSCRIPT
and special substituents in the synthetic compounds. Molecular docking and molecular dynamics
simulations analyses demonstrated that the binding mode of D12 and D15 with the kinase
domain of FGFR1 was most likely hydrophobic interactions. Moreover, biological assays of anti-
gastric cancer efficacy of D12 and D15 indicated effective inhibition of cell proliferation,
apoptosis induction and cell cycle arrest. Based on these findings, we conclude that D12 and D15
are potential agents for treatment of gastric cancer. This work provided a reference for the
discovery of novel FGFR1 inhibitors.
4. Experimental section
4.1. Chemistry
Reagents and solvents for the synthesis were commercially available and obtained from
Sigma-Aldrich (St Louis, Missouri, USA) and Aladdin (Shanghai, China), which were used
without further purification. Silica gel (GF254) for column chromatography (200–300 mesh) was
obtained from Aladdin. Melting points were measured on a Fisher-Johns melting apparatus and
1
were uncorrected. The H-NMR and 13C-NMR spectra data were recorded on a 600 MHz
spectrometer (Bruker Corporation, Switzerland) with TMS as an internal standard. Part of the
synthetic compounds had been screened for their anti-inflammatory properties, and the structural
spectral data reported in our previous papers [30-32]. In this article, we only listed the spectral
data of unreported compounds (B7, B14, C1-C5, and D13-D15). High-resolution mass spectra
(m/z) were recorded on a micrOTOF-Q II instrument. The purity of all compounds was detected
by HPLC (column: Agilent Eclipse XDB C18 5 ꢀm 4.6 mm × 150 mm, flow: 1 mL/min,
detected wavelength (DW): 420 nm or 450 nm, condition: methanol/water from 50/50 to 90/10).
The retained time (TR) and purities of unreported compounds in this article were collected.
4.1.1. General procedure for preparation of the intermediates 3a-3d
A mixture of Vanillin (2, 10 mmol) and appropriate ketone (5 mmol, 1a: tetrahydro-4H-
pyran-4-one, 1b: cyclohexanone, 1c: acetone) in ethanol (20 mL) was stirred at room
temperature. HCl (gas) was bubbled into the solution to catalyze the reaction until it was
completed. 4-Piperidone hydrochloride hydrate (1d) and Vanillin were dissolved in the mixture
solvent of ethanol and water (10:1) different with others. The resulting mixture was cooled and
poured into cold water (20 mL) to precipitate the product. The filter residue was purified by
silica gel chromatography using PE/EA as eluent to afford the intermediates 3a-3d.
4.1.2. General procedure for synthesis of compounds A1-A7
A solution of propionyl or isobutyryl chloride (3 mmol) in THF (2 mL) was added dropwise
into a solution of 3a-3d (1 mmol) and Et3N (0.25 mL) in THF (10 mL) under the condition of
o
ice-water bath (5-8 C). After being stirred overnight at room temperature, the solvent was
removed under vacuum, water and CH2Cl2 were added. The mixture was extracted with

ACCEPTED MANUSCRIPT
dichloromethane twice and dried over MgSO4. The organic layer was concentrated in vacuum.
The crude was refined by chromatography.
4.1.3. Synthesis of the intermediates 5e and 5b
A solution of Cyclopentanone (1e, 20 mmol) or cyclohexanone (1b), morpholine (4, 30
mmol) and 4-methylbenzenesulfonic acid (200 mg) in cyclohexane (20 mL) was heated to reflux
at 90 °C for 4 h. After cooling to room temperature, the mixture was washed with water, dried,
and concentrated to obtain the enamine intermediates 5e or 5b as a brown oil, which was directly
used for the next reaction.
4.1.4. Synthesis of the intermediates 7e and 7b
A mixture of 5e or 5b (10 mmol) and vanillin (10 mmol) was dissolved in ethanol (20 mL)
and the resulting solution was stirred at 90 °C for 2 h. The residue was concentrated under
vacuum and purified by column chromatography to give an orange powder. The powder was re-
dissolved in ethanol (10 mL), and 10% HCl solution (4 mL) was added. After being stirred at
room temperature for 3 h, distilled water (20 mL) was poured into the reaction flask. A light
yellow precipitate of 7e and 7b was collected and washed with water.
4.1.5. General procedure for synthesis of compounds 8e1-8e14 and 8b1-8b5
Different aromatic aldehydes (2 mmol) were dissolved into a suspension of the intermediates
7e and 7b (2 mmol) in ethanol (10 mL). Subsequently, HCl gas was bubbled into the mixture for
o
30 min, and the resulting mixture was stirred at 50-70 C for 2-3 h. After removal of ethanol,
water (10 mL) and ethyl ester (20 mL) were poured into the reaction solution. The organic
solvent was removed under reduced pressure and the residue was purified by chromatography to
obtain the products 8e1-8e14 and 8b1-8b5.
4.1.6. General procedure for synthesis of compounds B1-B14
A mixture of compounds 8e1-8e14 (1 mmol) and Et3N (0.25 mL) dissolved in THF (10 mL)
was cooled to 0 °C. A solution of propionyl chloride or isobutyryl chloride (3 mmol) was added
dropwise. The mixture was allowed to warm at room temperature and stirred overnight. After
evaporation of the solvent, the mixture was extracted with CH2Cl2, washed with water, and
concentrated under vacuum. The desired compounds B1-B14 were eluted with petroleum ether
and ethyl acetate by column chromatography, with the final product as various color powders.
4.1.6.1. 2-Methoxy-4-{(E)-{(E)-2-oxo-3-[4-(pyrrolidin-1-yl)benzylidene]cyclopentylidene}me-
thyl}phenyl propionate (B7). Orange yellow powder, 6.5% yield, mp: 200.8-202.9 °C. HPLC
1
purity: 98.756% (TR = 8.789 min, DW = 420nm). H-NMR (500MHz, CDCl3) δ: 7.595 (s, 1H,
β’-H), 7.528 (d, J = 8.5 Hz, 2H, H-2’, H-6’), 7.492 (s, 1H, β-H), 7.207 (d, J = 8.5 Hz, 1H, H-6),
7.171 (s, 1H, H-2), 7.084 (d, J = 8.0 Hz, 1H, H-5), 6.593 (d, J = 8.5 Hz, 2H, H-3’, H-5’), 3.872
(s, 3H, 3-OCH3), 3.370 (t, J = 7.0 Hz, 4H, CH2-N-CH2), 3.082 (s, 4H, 3’’-CH2, 4’’-CH2), 2.326

ACCEPTED MANUSCRIPT
(t, J = 7.5 Hz, 2H, -COCH2CH3), 2.038 (t, J = 6.5 Hz, 2H, N-CH2-CH2×2), 1.111 (t, J = 7.0 Hz,
3H, -COCH2CH3). 13C-NMR (125MHz, CDCl3) δ: 195.837, 173.063, 172.415, 151.146,
148.687, 140.369, 138.547, 135.776, 135.198, 133.140, 131.510, 123.034, 122.993, 114.543,
111.842, 55.918, 47.559, 29.688, 26.263, 25.4533, 9.631. HRMS (ESI): calcd for C27H29NO4
[M+H]+: 432.2175, found: 432.2183.
4.1.6.2. 2-Methoxy-4-{(E)-{(E)-2-oxo-3-[4-(pyrrolidin-1-yl)benzylidene]cyclopentylidene}me-
thyl}phenyl isobutyrate (B14). Orange yellow powder, 12.0% yield, mp: 205.8-206.9 °C. HPLC
1
purity: 94.308% (TR = 11.258 min, DW = 420nm). H-NMR (500MHz, CDCl3) δ: 7.575 (s, 1H,
β’-H), 7.512 (d, J = 9.0 Hz, 2H, H-2’, H-6’), 7.488 (s, 1H, β-H), 7.204 (d, J = 8.0 Hz, 1H, H-6),
7.161 (s, 1H, H-2), 7.069 (d, J = 8.0 Hz, 1H, H-5), 6.697 (d, J = 8.5 Hz, 2H, H-3’, H-5’), 3.860
(s, 3H, 3-OCH3), 3.424 (t, J = 7.0 Hz, 4H, CH2-N-CH2), 3.075 (s, 4H, 3’’-CH2, 4’’-CH2), 2.823-
2.879 (m, 1H, -COCH(CH3)2), 2.000-2.020 (m, 4H, N-CH2-CH2×2), 1.211 (d, J = 7.0 Hz, 3H, -
COCH(CH3)2). 13C-NMR (125MHz, CDCl3) δ: 184.589, 169.867, 145.598, 142,875, 134.630,
133.366, 131.531, 129.792, 129.683, 127.811, 125.712, 117.734, 117.243, 109.147, 105.723,
50.640, 38.114, 28.707, 23.520, 23.045, 17.777. HRMS (ESI): calcd for C28H31NO4 [M+H]+:
446.2330, found: 446.2330.
4.1.7. General procedure for synthesis of compounds C1-C5
With the intermediates 8b1-8b5 as starting material, the target compounds C1-C5 were
synthesized by applying the same method described for synthesis of B1-B14.
4.1.7.1. 2-Methoxy-4-{(E)-[(E)-3-(4-morpholinobenzylidene)-2-oxocyclohexylidene]methyl}ph-
enyl propionate (C1). Yellow powder, 5.4% yield, mp: 186.1-189.4°C. HPLC purity: 96.795%
1
(TR = 8.568 min, DW = 420nm). H-NMR (500MHz, CDCl3) δ: 7.760 (s, 2H, β-H, β’-H), 7.473
(d, J = 8.5 Hz, 2H, H-2’, H-6’), 7.279 (s, 1H, H-2), 7.056-7.073 (m, 2H, H-5, H-6), 6.939 (d, J =
8.5 Hz, 2H, H-3’, H-5’), 3.883 (t, J = 4.5 Hz, 4H, CH2-O-CH2), 3.846 (s, 3H, 3-OCH3), 3.270 (t,
J = 3.0 Hz, 4H, CH2-N-CH2), 2.948 (d, J = 2.0 Hz, 4H, 3’’-CH2, 5’’-CH2), 2.646 (q, J = 2.5 Hz,
2H, -COCH2CH3), 1.823 (m, 2H, 4’’-CH2), 1.297 (t, J = 2.5 Hz, 3H, -COCH2CH3). 13C-NMR
(125MHz, CDCl3) δ: 184.74, 167.17, 145.61, 134.64, 132.15, 131.23, 130.31, 129.65, 127.02,
117.44, 117.37, 109.26, 109.18, 61.36, 50.65, 43.00, 23.37, 23.05, 22.10, 17.69. HRMS (ESI):
calcd for C28H31NO5 [M+H]+: 462.2278, found: 462.2278.
4.1.7.2. 2-methoxy-4-((E)-((E)-2-oxo-3-(4-(piperidin-1-yl)benzylidene)cyclohexylidene)methyl)
phenyl propionate (C2) Orange yellow powder, 19.4% yield, mp: 126.2-129.0 oC. HPLC purity:
1
97.949% (TR = 10.122 min, DW = 420nm). H-NMR (500MHz, CDCl3) δ: 7.758 (s, 1H, β’-H),
7.732 (s, 1H, β-H), 7.437 (d, J = 9.0 Hz, 2H, H-2’, H-6’), 7.051 (s, 1H, H-2), 7.041 (d, J = 7.5
Hz, 1H, H-6), 7.031 (d, J = 7.5 Hz, 1H, H-5), 6.905 (d, J = 9.0 Hz, 2H, H-3’, H-5’), 3.843 (s, 3H,
3-OCH3), 3.289 (t, J = 6.0 Hz, 4H, CH2-N-CH2), 2.946 (t, J = 6.0 Hz, 2H, 3’’-CH2), 2.905 (t, J =
5.0 Hz, 2H, 5’’-CH2), 2.627 (q, J = 7.5 Hz, 2H, -COCH2CH3), 1.802 (t, J = 5.5 Hz, 2H, 4’’-CH2),

ACCEPTED MANUSCRIPT
1.693 (t, J = 4.5 Hz, 4H, N-CH2-CH2×2), 1.631 (t, J = 5.0 Hz, 2H, N-CH2-CH2-CH2), 1.283 (t, J
= 7.5 Hz, 3H, -COCH2CH3). 13C-NMR (125MHz, CDCl3) δ: 189.996, 172.467, 151.749,
150.889, 139.875, 138.032, 136.707, 135.275, 135.048, 132.484, 125.710, 122.701, 122.626,
114.656, 114.466, 55.940, 49.182, 28.712, 28.340, 27.385, 25.501, 24.338, 23.303, 9.148.
HRMS (ESI): calcd for C29H33NO4 [M+H]+: 460.2488, found: 460.2490.
4.1.7.3. 4-{(E)-{(E)-3-[4-(Diethylamino)benzylidene]-2-oxocyclohexylidene}methyl}-2-meth-
oxyphenyl propionate (C3). Orange yellow powder, 31.6% yield, mp: 141.7-144.2°C. HPLC
1
purity: 97.569% (TR = 8.019 min, DW = 420nm). H-NMR (500MHz, CDCl3) δ: 7.776 (s, 1H,
β’-H), 7.733 (s, 1H, β-H), 7.448 (d, J = 8.5 Hz, 2H, H-2’, H-6’), 7.048 (s, 1H, H-2), 7.040 (d, J =
7.5 Hz, 1H, H-6), 7.033 (d, J = 7.5 Hz, 1H, H-5), 6.680 (d, J = 8.5 Hz, 2H, H-3’, H-5’), 3.844 (s,
3H, 3-OCH3), 3.408 (q, J = 7.0 Hz, 4H, CH2-N-CH2), 2.951 (t, J = 5.5 Hz, 2H, 3’’-CH2), 2.898
(t, J = 5.5 Hz, 2H, 5’’-CH2), 2.2628 (q, J = 7.5 Hz, 2H, -COCH2CH3), 1.805 (t, J = 6.0 Hz, 2H,
4’’-CH2), 1.283 (t, J = 7.5 Hz, 3H, -COCH2CH3), 1.199 (t, J = 7.0 Hz, 6H, N--CH2-CH3×2). 13C-
NMR (125MHz, CDCl3) δ: 184.623, 167.213, 145.555, 134.470, 133.447, 131.573, 129.874,
129.670, 127.831, 117.373, 117.300, 109.133, 105.742, 50.638, 39.132, 23.518, 23.042, 22.097,
17.771. HRMS (ESI): calcd for C28H33NO4 [M+H]+: 448.2488, found: 448.2506.
4.1.7.4. 2-Methoxy-4-{(E)-[(E)-3-(4-morpholinobenzylidene)-2-oxocy-clohexylidene]methyl}
phenylisobutyrate (C4). Yellow powder, 8.1% yield, mp: 165.7-168.1°C. HPLC purity: 93.093%
(TR = 8.459 min, DW = 420nm). 1H-NMR (500MHz, CDCl3) δ: 7.756 (s, 1H, β’-H), 7.738 (s,
1H, β-H), 7.463 (d, J = 9.0 Hz, 2H, H-2’, H-6’), 7.028-7.047 (m, 2H, H-2, H-5), 7.038 (t, 1H, H-
6), 6.906 (d, J = 8.5 Hz, 2H, H-3’, H-5’), 3.873 (t, J = 5.0 Hz, 4H, CH2-O-CH2), 3.837 (s, 3H, 3-
OCH3), 3.255 (t, J = 5.0 Hz, 4H, CH2-N-CH2), 2.902-2.953 (m, 4H, 3’’-CH2, 5’’-CH2), 2.821-
2.877 (m, 1H, -COCH(CH3)2), 1.783-1.830 (m, 2H, 4’’-CH2), 1.334 (d, J = 7.0 Hz, 6H, -
COCH(CH3)2). 13C-NMR (125MHz, CDCl3) δ: 184.743, 169.843, 145.655, 143.809, 132.116,
131.181, 130.360, 129.563, 128.077, 127.008, 117.446, 117.315, 109.256, 109.197, 61.365,
50.651, 43.000, 28.708, 23.367, 23.050, 17.696, 13.739, 13.718, 13.688. HRMS (ESI): calcd for
C29H33NO5 [M+H]+: 462.2644, found: 462.2646.
4.1.7.5.
2-Methoxy-4-{(E)-{(E)-2-oxo-3-[4-(piperidin-1-
yl)benzylidene]cyclohexylidene}methyl} phenyl isobutyrate (C5). Orange yellow powder,
26.7% yield, mp: 158.5-161.6 °C. HPLC purity: 93.671% (TR = 12.264 min, DW = 420nm). 1H-
NMR (500MHz, CDCl3) δ: 7.758 (s, 1H, β’-H), 7.733 (s, 1H, β-H), 7.437 (d, J = 9.0 Hz, 2H, H-
2’, H-6’), 7.042-7.063 (m, 2H, H-2, H-5), 7.026 (s, 1H, H-6), 6.906 (d, J = 8.5 Hz, 2H, H-3’, H-
5’), 3.834 (s, 3H, 3-OCH3), 3.288 (t, J = 5.5 Hz, 4H, CH2-N-CH2), 2.944 (t, J = 5.5 Hz, 2H, 3’’-
CH2), 2.902 (t, J = 5.5 Hz, 2H, 5’’-CH2), 2.819-2.875 (m, 1H, -COCH(CH3)2), 1.801 (t, J = 6.0
Hz, 2H, 4’’-CH2), 1.684-1.693 (m, 4H, N-CH2-CH2×2), 1.626 (d, J = 4.5 Hz, 2H, N-CH2-CH2-
CH2), 1.333 (d, J = 7.0 Hz, 6H, -COCH(CH3)2). 13C-NMR (125MHz, CDCl3) δ: 184.685,
169.844, 146.441, 145.627, 134.726, 132.693, 131.350, 130.031, 129.673, 127.192, 120.407,

ACCEPTED MANUSCRIPT
117.416, 117.280, 109.356, 109.171, 50.646, 43.885, 28.710, 23.430, 23.057, 20.213, 19.050,
17.733, 13.750, 13.696. HRMS (ESI): calcd for C30H35NO4 [M+H]+: 474.2644, found: 474.2654.
4.1.8. General procedure for synthesis of compounds D1-D13
The preparation procedure of compounds D1-D13 was the same as compounds 8e1-8e14,
which used the different substituted aromatic aldehyde for the aldol condensation reaction.
4.1.8.1. (2E,6E)-2-(4-hydroxy-3-methoxybenzylidene)-6-(4-(pyrrolidi-1-yl)benzylidene)cyclo-
hexanone (D13). Orange yellow powder, 15.2% yield, mp: 203.2-206.1°C. HPLC purity:
1
99.267% (TR = 6.230 min, DW = 450nm). H-NMR (500MHz, d6-DMSO) δ: 7.558 (d, J = 7.5
Hz, 2H, H-2’, H-6’), 7.436 (s, 1H, β’-H), 7.419 (s, 1H, β-H), 7.100 (s, 1H, H-2), 7.013 (d, J = 8.5
Hz, 1H, H-6), 6.851 (d, J = 7.0 Hz, 1H, H-5), 6.595 (d, J = 8.5 Hz, 2H, H-3’, H-5’), 3.817 (s, 3H,
3-OCH3), 3.292 (t, J = 6.5 Hz, 4H, CH2-N-CH2), 2.874 (d, J = 5.0 Hz, 4H, 3’’-CH2, 5’’-CH2),
1.968 (t, J = 6.5 Hz, 4H, N-CH2-CH2×2), 1.728 (t, J = 5.5 Hz, 2H, 4’’-CH2). 13C-NMR
(125MHz, d6-DMSO) δ: 188.028, 147.901, 147.612, 147.388, 137.017, 135.397, 133.754,
132.536, 130.769, 127.043, 124.032, 122.370, 115.490, 114.655, 111.547, 55.617, 47.171,
28.229, 27.919, 24.919, 22.631. HRMS (ESI): calcd for C25H27NO3 [M+H]+: 390.2069, found:
390.2078.
4.1.9. Preparation of the protected Vanillin 10
A solution of Vanillin (2, 10 mmol), 3,4-2H-dihydropyran (9, 40 mmol) and pyridinium 4-
toluenesulfonate (2.51 mmol) in CH2Cl2 (30 mL) was stirred at room temperature for 3 h. The
resulting mixture was extracted with CH2Cl2 twice and concentrated in vacuum. The protected
product 10 was purified by silica gel column chromatography as colorless oil.
4.1.10. Preparation of the unilateral substituted intermediate 11
To a solution of compound 10 (5 mmol) in acetone (20 mL) a solution of 20% NaOH (1 mL)
was added dropwise slowly at room temperature. TLC monitored the endpoint and removed the
solvent. The residue was extracted with ethyl ester (20 mL) and washed water (10 mL) twice.
After evaporating the ethyl ester, the mixture was purified by column chromatography to obtain
the intermediate 11 as white powder.
4.1.11. General procedure for synthesis of compounds D14-D15
To a mixture of dry powdered reactant 10 (1 mmol) and corresponding substituted
benzaldehyde (1 mmol), a suspension of NaOH in dioxane (5g NaOH dispersed in 20 mL
dioxane) was added dropwise until the reaction color turned wine red. The solution was stirred at
room temperature for 4 h. The residue was concentrated and extracted with ethyl ester (20 mL).
After purification by column chromatography, the colored powder was dissolved in ethanol (10
mL), and de-protected by 10% HCl solution (4 mL). The formed precipitates was filtered and
purified by column chromatography to give the desired products D14-D15.

ACCEPTED MANUSCRIPT
4.1.11.1. (1E,4E)-1-(4-Hydroxy-3-methoxyphenyl)-5-[4-(pyrrolidin-1-yl)phenyl]penta-1,4-dien-
3-one (D14). Red powder, 3.0% yield, mp: 204.3-207.4°C. HPLC purity: 97.934% (TR = 7.402
1
min, DW = 450nm). H-NMR (500MHz, d6-DMSO) δ: 7.559 (d, J = 7.0 Hz, 2H, H-2’, H-6’),
7.547 (d, J = 15.5 Hz, 1H, β’-H), 7.539 (d, J = 15.5 Hz, 1H, β-H), 7.148 (s, 1H, H-2), 7.059 (d, J
= 8.0 Hz, 1H, H-6), 6.960 (d, J = 16.0 Hz, 1H, α’-H), 6.829 (d, J = 15.0 Hz, 1H, α-H), 6.579 (d, J
= 8.0 Hz, 2H, H-3’, H-5’), 6.504 (d, J = 7.0 Hz, 1H, H-5), 3.769 (s, 3H, 3-OCH3), 3.308 (d, J =
6.5 Hz, 4H, CH2-N-CH2), 1.974 (s, 4H, N-CH2-CH2×2). 13C-NMR (125MHz, d6-DMSO) δ:
187.574, 149.248, 149.199, 147.938, 143.235, 141.853, 130.341, 130.196, 126.513, 123.319,
123.078, 121.503, 120.195, 115.646, 111.762, 111.450, 55.723, 47.423, 24.911. HRMS (ESI):
calcd for C22H23NO3 [M+H]+: 350.1756, found: 350.1760.
4.1.11.2. (1E,4E)-1-[4-(Diethylamino)phenyl]-5-(4-hydroxy-3-methoxyphenyl)penta-1,4-dien-3-
one (D15). Red powder, 7.8% yield, mp: 134.7-135 °C. HPLC purity: 99.900% (TR = 6.311 min,
DW = 450nm). 1H-NMR (500MHz, d6-DMSO) δ: 7.624 (d, J = 15.5 Hz, 1H, β’-H), 7.592 (d, J =
15.5 Hz, 1H, β-H), 7.567 (d, J = 9.5 Hz, 2H, H-2’, H-6’), 7.352 (s, 1H, H-2), 7.182 (d, J = 8.0
Hz, 1H, H-6), 7.119 (d, J = 15.5 Hz, 1H, α’-H), 6.977 (d, J = 15.5 Hz, 1H, α-H), 6.812 (d, J = 8.0
Hz, 1H, H-5), 6.709 (d, J = 8.0 Hz, 2H, H-3’, H-5’), 3.848 (s, 3H, 3-OCH3), 3.414 (d, J = 6.5 Hz,
13
4H, CH2-N-CH2), 1.124 (t, J = 6.5 Hz, 6H, N-CH2-CH3×2). C-NMR (125MHz, d6-DMSO) δ:
187.586, 149.264, 149.216, 147.942, 143.006, 141.864, 130.527, 126.502, 123.295, 123.084,
121.242, 120.244, 115.651, 111.472, 111.224, 55.727, 43.759, 12.437. HRMS (ESI): calcd for
C22H25NO3 [M+H]+: 352.1912, found: 352.1928.
4.2. In vitro kinase assay
Based on microfluidics chip technology, the kinase activity of FGFR1 was determined by
Mobility Shift Assay on EZ Reader (Caliper Life Sciences, MA). As the experimental condition
was constrained, the kinase activity screening and ATP competitive assay both were authorized
to ChemPartner (Shanghai, China). The specific protocol is described briefly as follows:
The enzyme solution was prepared in 1.25× kinase base buffer (62.5 mM HEPES,
0.001875% Brij-35, 12.5 mM MgCl2, 2.5 mM DTT), stop buffer (100 mM HEPES, 0.015%
Brij-35, 0.2% coating reagent No. 3, 50 mM EDTA) The compound was dissolved in DMSO,
and diluted to the specific concentration with water in 384-well plate. 5 ꢀL of compound solution
was diluted with 10 ꢀL of 2.5× enzyme solution. 10 mM EDTA was used as the blank control.
The mixture was incubated at room temperature for 10 min. 10 ꢀL of peptide solution (2.5×,
FAM-P22, add FAM-labeled peptide and ATP in the 1.25× kinase base buffer) was added to
each well of the 384-well assay plate. After incubation at 28 oC for 1 h, 25 ꢀL of stop buffer was
added into the mixture to stop the reaction. Conversion data were collected on Caliper EZ reader
(Hopkinton, MA), the percent inhibition of kinase activity was calculated based on the formula
below:
Inhibition % = (max-conversion) / (max-min) × 100%,

ACCEPTED MANUSCRIPT
“Max” stands for DMSO control; “Min” stands for low control.
To determine the half-maximal inhibitory concentration of the compounds relative to kinase
activity, 10 gradient concentrations of compounds (100, 33.33, 11.11, 3.70, 1.23, 0.41, 0.14,
0.046, 0.015, and 0.005 ꢀM) were set up. The inhibition ratios of different concentrations
relative to kinase were calculated and the concentration-inhibition rate curve was fitted by
GraphPad Prism.
4.3. Quantitative structure-activity relationships analysis
4.3.1. Materials and descriptors selection
All products were drawn and saved as SDF format. The compounds were calculated by the
ChemoPy descriptor calculation program [37, 38]. After calculation of molecular descriptors,
about 1000 molecular descriptors based on molecular structure were obtained. In this study, the
descriptors include the constitutional descriptors, physicochemical descriptors, topological
descriptors, geometrical descriptors, charge (electronic) descriptors, etc. After calculation of the
molecular descriptors, those that were constant descriptors for all molecules were eliminated and
pairs of variables with a correlation coefficient greater than 0.75 were classified as inter-
correlated and one in each correlated pair was deleted.
4.3.2. Multiple linear regression (MLR) analysis
Multiple linear regression (MLR) analysis is a statistical technique that uses several
explanatory variables to predict the outcome of a response variable. The goal of MLR is to
model the relationship between the explanatory and response variables. In our present study,
MLR was performed using R program, a widely used tool for statistical computing and graphics,
to derive QSAR models. The biological data used in this study were the FGFR1-inhibitory rates
compared with negative group. Compounds with negative values were abandoned because of
their anti-cancer activities. The inhibition rates against FGFR1 were used as dependent variables
in the linearization procedure. Subsequently, stepwise Multiple Linear Regression (Stepwise-
MLR) was used to select the significant descriptors. The most relevant descriptors were used as
independent variables.
4.3.3. Validation of the models
To confirm that the lineal model was reliable, testing the predictive ability and generalizing
the methods by cross-validation is required. The leave-one-out (LOO) procedure was employed.
When a data point was removed from the analyzed set, the regression was recalculated and the
predicted value for that point was compared to its actual value. This process was repeated until
each datum had been omitted once and the sum of squares of these deletion residuals could be
used to calculate q2, an equivalent statistic to R2.
4.4. ATP competitive inhibition assay

ACCEPTED MANUSCRIPT
With the use of kinase inhibitory activity screening assay, four gradient concentrations of
compounds to be measured were set up. At each concentration of different compounds, eight
concentrations of ATP (4192, 2096, 1048, 524, 262, 131, 66, and 33 ꢀM) were established to
determine the conversion of substrate peptide catalyzed by FGFR1 kinase within 1 h. The
inhibition curve was fitted by GraphPad Prism.
4.5. Molecular docking
The binding pose of compound D12 and D15 in FGFR1’s binding site was predicted by the
software Autodock (version 4.2.6) [39]. The crystallographic co-ordinates for human wild type
FGFR1 (PDB ID: 3RHX) was retrieved from the Protein Data Bank (PDB). Prior to docking,
protein structure was prepared by removing water molecules and other ligands using PyMol
software [40]. Individual ligand files were prepared for docking using the prepare_ligand4.py
script, then charges and hydrogen atoms of FGFR1 protein were added using the
prepare_receptor4.py script from last version of AutoDockTools 1.5.6. A grid box size of 60 ×
60 × 60 dimensions, with a spacing of 0.375 Å between the grid points, was implemented and
covered almost the entire FGFR1 binding site. The grid parameter files were created by setting
up of map files directly. The Lamarckian genetic algorithm (LGA) was applied to deal with the
protein and inhibitors interactions. The docking parameters were as follows: trials of 100
dockings, the number of individuals in population was set to 300, the crossover rate was 0.8, and
the local search rate was 0.06 and 25 million of maximum number of energy evaluations. Other
settings were set to default. AutoDockTools and PyMol was used to analyze the docking results.
4.6.1. Molecular dynamics (MD) simulations
Molecular dynamics (MD) simulations were performed to verify whether the docking results
are stable enough. Molecular dynamics (MD) simulations of D12/FGFR1 complex system and
D15/FGFR1 system were carried out. The initial complexes came from the top ranked binding
pose, cluster by the docking results with the root mean square deviations (RMSD =2) of each
pose. The partial charges of the D12 and D15 were derived by using the restrained electrostatic
potential (RESP) fitting procedure based on the electrostatic potentials calculated by Hartree-
Fock (HF) method with 6-31G (d) basis set in the Gaussian09 Package [41]. Molecular
mechanics parameters, from the ff99SB and GAFF force fields using the LEaP module of
AMBER 11 software packages, were assigned to the protein and the ligand respectively [42, 43].
All the water molecules and the specified ligand were deleted. The two systems were all solvated
in a box of TIP3P water molecules with a hydration shell of 10 Å. Additionally, an appropriate
number of sodium ions was used to neutralize these two systems.
The AMBER 11 software package was employed, using the same protocol, for all
simulations [42]. We carried out an equilibration protocol consisting of an initial minimization of
water box of 5000 steps, 2500 steps for the steepest descent and 2500 steps in the conjugate
gradient. Then, the TIP3P water box was heated at constant volume until 300 K using a time
constant for the heat bath with a coupling time of 100 ps. Equilibration was at 300 K under

ACCEPTED MANUSCRIPT
constant pressure of 100 ps. Before the production of MD, the whole system was equilibrated at
100 ps at a constant pressure of 1 bar, and the Langevin temperature scaling was turned on with a
collision frequency of 2 ps [44]. The last step used 50 ns of MD simulation without any
restraints. In addition, short-range non-bonded interactions were calculated based on a cutoff of
8.0 Å. Periodic Boundary Conditions were turned on in every step of the whole process. Long-
range electrostatic interactions were handled by the Particle Mesh Ewald (PME) algorithm [45].
The SHAKE method was used to constrain hydrogen atoms and the time step was set to 2 fs
[46]. The coordinates were saved every 10 ps for the subsequent analysis.
4.6.2. Binding free energy calculations and decomposition analysis
The binding free energies were determined by the molecular mechanics/ generalized Born
surface area (MM-GBSA) method as implemented in AMBER 11. As to each system, 500
snapshots from the last 10 ns MD trajectory were chosen for the calculations. The binding free
energy was calculated according to the equation:
△
△
△
△
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃꢄ
=
=
△
Gꢅꢆꢇꢈꢉꢊꢋ – (
△
Gꢌꢊꢅꢊꢈꢍꢆꢌ
+
△ Gꢉꢎꢏꢐꢑꢒ
)
(1)
(2)
(3)
(4)
△ H T △ S ≈ △ Eꢏꢐꢓ
–
+
△
Gꢓꢆꢉ T △ S)
−
Gꢏꢐꢓ =△ Gꢔꢕꢖ +△ Gꢊꢉꢊꢅ
Gꢓꢆꢉ =△ Gꢗꢘ +△ Gꢙꢚ
In the equation (3), △Ggas between protein and ligand is the gas-phase interaction, including
the van der Waals and the electrostatic energies; △GGB and △GSA are the polar and nonpolar
components of the desolvation free energy, respectively; T△S, at temperature T, represents the
change of conformational entropy upon ligand binding.
4.7. Biological activity experiments
4.7.1. Cell culture
SGC7901 was obtained from Cell Bank of Chinese Academy of Sciences (Wuhan, China)
,
and MGC823, BGC803 were bought from Shanghai Institute of Biosciences and Cell Resources
Center (Chinese Academy of Sciences, Shanghai, China). All cells were grown in RMPI-1640
media (Gibco) supplemented with 10% FBS (Gibco), 1% Penicilin Streptomycin (Gibco) in a
humidified ThermoForma (Thermo Fisher Scientific) containing 5% CO2 at 37 °C.
4.7.2. MTT Cytotoxicity Assay
Cells (4 × 103) were seeded in a 96-well cell culture plate and treated with compounds for 72
h. Assay was made by adding 20 ꢀL of methylthiazoletetrazolium (MTT; 0.5 mg/mL; M5655;

ACCEPTED MANUSCRIPT
Sigma) to the culture medium and incubated at 37 °C for 4 h. At the end of incubation, the
produced crystal was dissolved in 150 ꢀL dimethyl sulfoxide (sigma) and quantified at 490 nm.
4.7.3. Western blot analysis
Treated cells were washed with PBS and harvested using RIPA Buffer (AR0103; BOSTER,
Wuhan, China) with 1% PMSF (100 mM; ST506; Beyotime). Protein quantification was
performed using the Bradford 1 × Dye Reagent (#5000205; Bio-Rad). Cell extracts were
separated by 10% SDS-PAGE and transferred to PVDF membrane (Millipore). Then the
membrane was blocked for 90 min at room temperature in TBS with 0.1% Tween and 5% milk
powder. The membrane was incubated with the following primary antibodies: Phospho-Flg
(Tyr653/654; 1:300; sc-30262; Santa cruz Biotechnology, lnc.), p-ERK (1:300; sc-7383; Santa
cruz Biotechnology, lnc.), cleaved PARP-1(194C1439; 1:300; sc-56196; Santa cruz
Biotechnology, lnc.), cyclin D1 (1:300; sc-753 Santa cruz Biotechnology, lnc.), Bcl-2 (1:300; sc-
492; Santa cruz Biotechnology, lnc.), or GAPDH (1:300; sc-25778; Santa cruz Biotechnology,
lnc.). After washing, the membranes were incubated with the relevant secondary antibodies,
visualization of bands was by enhanced chemiluminescence (ECL; Bio-Rad), and densitometric
analysis made using Image J software.
4.7.4. Hoechst 33342 staining
SGC7901 cells at logarithmic growth were seeded in 6-well cell culture plate and treated
with DMSO at different concentrations of D12 or D15 or 10 ꢀM A117 for 12 h. Following
fixation with 4% paraformaldelyde for 10 min at room temperature (RT), cells were washed and
stained with Hoechst 33342 (Beyotime) for 10 min. Cells were observed under a fluorescence
microscope (Nikon, Tokyo, Japan) using appropriate filters for blue fluorescence.
4.7.5. Cell cycle analysis
Cells (3 × 105) were treated with agents for 8 h, washed with PBS, collected, and fixed in
75% ice-cold ethanol for 4 h. After washing with PBS, the cells were stained with 500 ꢀL of
propidium iodide (PI) (Becton, Dickinson and Company) for 10 min at 4 °C in the dark. The cell
suspension was filtered with 200 - mesh gauze and subjected to a FACSCalibur instrument (BD
Biosciences Clontech). Data were analyzed using the ModFit DNA analysis program.
4.7.6. Statistical analysis
The results were presented as the mean ± standard error (SEMs). The statistics were
performed using Student’s t-test in GraphPad Pro (GraphPad, San Diego, CA, USA). A value of
P less than 0.05 (p < 0.05) was considered statistically significant. All experiments were repeated
a minimum of three times.

ACCEPTED MANUSCRIPT
Conflict of interest
The authors have declared no conflict of interest.
Acknowledgment
This work was supported by the National Natural Science Foundation of China (Grant Nos.
81402839, 81272462, 81473242), the Natural Science Foundation of Zhejiang Province of China
(LY17H160059), and the Technology Foundation for Medical Science of Zhejiang Province
(Grant No. 2012KYA129).
References
[1] J. Ferlay, I. Soerjomataram, R. Dikshit, S. Eser, C. Mathers, M. Rebelo, D.M. Parkin, D.
Forman, F. Bray, Cancer incidence and mortality worldwide: sources, methods and major
patterns in GLOBOCAN 2012, International Journal of Cancer 136 (2015) E359-386.
[2] R.L. Siegel, K.D. Miller, A. Jemal, Cancer statistics, 2015, CA-A Cancer Journal for
Clinicians 65 (2015) 5-29.
[3] T. Oshima, M. Masuda, Molecular targeted agents for gastric and gastroesophageal junction
cancer, Surgery Today 42 (2012) 313-327.
[4] J. Wang, R. Xu, J. Li, Y. Bai, T. Liu, S. Jiao, G. Dai, J. Xu, Y. Liu, N. Fan, Y. Shu, Y. Ba,
D. Ma, S. Qin, L. Zheng, W. Chen, L. Shen, Randomized multicenter phase III study of a
modified docetaxel and cisplatin plus fluorouracil regimen compared with cisplatin and
fluorouracil as first-line therapy for advanced or locally recurrent gastric cancer, Gastric
Cancer 19 (2016) 234-244.
[5] S.S. Khakoo, A. Georgiou, J.S. Waters, A restrospective analysis of toxicities encountered
with palliative epirubicin, oxaliplatin, and capecitabine (EOX) chemotherapy for advanced
esophagogastric cancer, Journal of Clinical Oncology 32 (2014).
[6] E.C. Smyth, D. Cunningham, Targeted therapy for gastric cancer, Current Treatment
Options in Oncology 13 (2012) 377-389.
[7] H. Arai, S. Hironaka, T. Suzuki, K. Sudo, K. Nakamura, K. Minashi, T. Hara, T. Denda, T.
Yamaguchi, A retrospective study of early toxicity of weekly paclitaxel as second line
chemotherapy for advanced gastric cancer, Annals of Oncology 26 (2015) 132-132.
[8] T. Aoyama, T. Yoshikawa, Targeted therapy: Apatinib - new third-line option for refractory
gastric or GEJ cancer, Nature Reviews Clinical Oncology 13 (2016) 268-270.
[9] V. Brower, Apatinib in treatment of refractory gastric cancer, Lancet Oncol 17 (2016) e137.
[10] J. Zhang, P.L. Yang, N.S. Gray, Targeting cancer with small molecule kinase inhibitors,
Nature Reviews Cancer 9 (2009) 28-39.
[11] C. Hierro, J. Rodon, J. Tabernero, Fibroblast Growth Factor (FGF) Receptor/FGF
Inhibitors: Novel Targets and Strategies for Optimization of Response of Solid Tumors,
Seminars in Oncology 42 (2015) 801-819.
[12] M.H. Schafer, P. Lingohr, A. Strasser, N.C. Lehnen, M. Braun, S. Perner, T. Holler, G.
Kristiansen, J.C. Kalff, I. Gutgemann, Fibroblast growth factor receptor 1 gene
amplification in gastric adenocarcinoma, Human Pathology 46 (2015) 1488-1495.
[13] M.V. Dieci, M. Arnedos, F. Andre, J.C. Soria, Fibroblast growth factor receptor inhibitors
as a cancer treatment: from a biologic rationale to medical perspectives, Cancer Discovery 3
(2013) 264-279.

ACCEPTED MANUSCRIPT
[14] M. Touat, E. Ileana, S. Postel-Vinay, F. Andre, J.C. Soria, Targeting FGFR Signaling in
Cancer, Clinical Cancer Research 21 (2015) 2684-2694.
[15] E.P. Carter, A.E. Fearon, R.P. Grose, Careless talk costs lives: fibroblast growth factor
receptor signalling and the consequences of pathway malfunction, Trends in Cell Biology
25 (2015) 221-233.
[16] G. Liang, Z. Liu, J. Wu, Y. Cai, X. Li, Anticancer molecules targeting fibroblast growth
factor receptors, Trends in Pharmacological Sciences 33 (2012) 531-541.
[17] A. Dutt, A.H. Ramos, P.S. Hammerman, C. Mermel, J. Cho, T. Sharifnia, A. Chande, K.E.
Tanaka, N. Stransky, H. Greulich, N.S. Gray, M. Meyerson, Inhibitor-sensitive FGFR1
amplification in human non-small cell lung cancer, PLoS One 6 (2011) e20351.
[18] I. Tsimafeyeu, L. Demidov, E. Stepanova, N. Wynn, H. Ta, Overexpression of fibroblast
growth factor receptors FGFR1 and FGFR2 in renal cell carcinoma, Scandinavian Journal
of Urology and Nephrology 45 (2011) 190-195.
[19] V.K. Jain, N.C. Turner, Challenges and opportunities in the targeting of fibroblast growth
factor receptors in breast cancer, Breast Cancer Research 14 (2012) 208.
[20] N. Turner, R. Grose, Fibroblast growth factor signalling: from development to cancer,
Nature Reviews Cancer 10 (2010) 116-129.
[21] G. Daniele, J. Corral, L.R. Molife, J.S. de Bono, FGF receptor inhibitors: role in cancer
therapy, Current Oncology Reports 14 (2012) 111-119.
[22] N. Miyamoto, H. Yamamoto, C. Miyamoto, T. Maehata, K. Nosho, H. Taniguchi, K.
Yamashita, Y. Adachi, Y. Arimura, F. Itoh, K. Imai, Y. Shinomura, Overexpression of the
receptor tyrosine kinase epha4 plays an important role in the progression of human gastric
cancers, Gastroenterology 134 (2008) A611-A612.
[23] D. Wen, S. Li, F. Ji, H. Cao, W. Jiang, J. Zhu, X. Fang, miR-133b acts as a tumor
suppressor and negatively regulates FGFR1 in gastric cancer, Tumor Biology 34 (2013)
793-803.
[24] L. Xie, X. Su, L. Zhang, X. Yin, L. Tang, X. Zhang, Y. Xu, Z. Gao, K. Liu, M. Zhou, B.
Gao, D. Shen, L. Zhang, J. Ji, P.R. Gavine, J. Zhang, E. Kilgour, X. Zhang, Q. Ji, FGFR2
gene amplification in gastric cancer predicts sensitivity to the selective FGFR inhibitor
AZD4547, Clinical Cancer Research 19 (2013) 2572-2583.
[25] K. Schmidt, C. Moser, C. Hellerbrand, D. Zieker, C. Wagner, J. Redekopf, H.J. Schlitt, E.K.
Geissler, S.A. Lang, Targeting Fibroblast Growth Factor Receptor (FGFR) with BGJ398 in
a Gastric Cancer Model, Anticancer Research 35 (2015) 6655-6665.
[26] Z. Zhao, H. Wu, L. Wang, Y. Liu, S. Knapp, Q. Liu, N.S. Gray, Exploration of type II
binding mode: A privileged approach for kinase inhibitor focused drug discovery?, ACS
Chemical Biology 9 (2014) 1230-1241.
[27] R.A. Norman, D. Toader, A.D. Ferguson, Structural approaches to obtain kinase selectivity,
Trends in Pharmacological Sciences 33 (2012) 273-278.
[28] S. Eathiraj, R. Palma, M. Hirschi, E. Volckova, E. Nakuci, J. Castro, C.R. Chen, T.C. Chan,
D.S. France, M.A. Ashwell, A novel mode of protein kinase inhibition exploiting
hydrophobic motifs of autoinhibited kinases: discovery of ATP-independent inhibitors of
fibroblast growth factor receptor, Journal of Biological Chemistry 286 (2011) 20677-20687.
[29] Y. Wang, Y. Cai, J. Ji, Z. Liu, C. Zhao, Y. Zhao, T. Wei, X. Shen, X. Zhang, X. Li, G.
Liang, Discovery and identification of new non-ATP competitive FGFR1 inhibitors with
therapeutic potential on non-small-cell lung cancer, Cancer Letters 344 (2014) 82-89.

ACCEPTED MANUSCRIPT
[30] Y. Zhang, L. Zhao, J. Wu, X. Jiang, L. Dong, F. Xu, P. Zou, Y. Dai, X. Shan, S. Yang, G.
Liang, Synthesis and evaluation of a series of novel asymmetrical curcumin analogs for the
treatment of inflammation, Molecules 19 (2014) 7287-7307.
[31] Y.L. Zhang, X. Jiang, K.S. Peng, C.W. Chen, L.L. Fu, Z. Wang, J.P. Feng, Z.G. Liu, H.J.
Zhang, G. Liang, Z. Pan, Discovery and evaluation of novel anti-inflammatory derivatives
of natural bioactive curcumin, Drug Design Development and Therapy 8 (2014) 2161-2171.
[32] L.L. Dong, S.Q. Zheng, Y.L. Zhang, X. Jiang, J.Z. Wu, X.Q. Zhang, X.O. Shan, D.D.
Liang, S.L. Ying, J.P. Feng, G. Liang, Design, synthesis, and evaluation of semi-
conservative mono-carbonyl analogs of curcumin as anti-inflammatory agents against
lipopolysaccharide-induced acute lung injury, Medicinal Chemistry Communications 6
(2015) 1544-1553.
[33] L.N. Johnson, Protein kinase inhibitors: contributions from structure to clinical compounds,
Quarterly Reviews of Biophysics 42 (2009) 1-40.
[34] J.L. Qu, X.J. Qu, M.F. Zhao, Y.E. Teng, Y. Zhang, K.Z. Hou, Y.H. Jiang, X.H. Yang, Y.P.
Liu, Gastric cancer exosomes promote tumour cell proliferation through PI3K/Akt and
MAPK/ERK activation, Digestive and Liver Disease 41 (2009) 875-880.
[35] M. Fu, C. Wang, Z. Li, T. Sakamaki, R.G. Pestell, Minireview: Cyclin D1: normal and
abnormal functions, Endocrinology 145 (2004) 5439-5447.
[36] W.K. Wu, C.H. Cho, C.W. Lee, D. Fan, K. Wu, J. Yu, J.J. Sung, Dysregulation of cellular
signaling in gastric cancer, Cancer Letters 295 (2010) 144-153.
[37] D.S. Cao, S. Liu, L. Fan, Y.Z. Liang, QSAR analysis of the effects of OATP1B1 transporter
by structurally diverse natural products using a particle swarm optimization-combined
multiple linear regression approach, Chemometrics and Intelligent Laboratory Systems 130
(2014) 84-90.
[38] D.S. Cao, Q.S. Xu, Q.N. Hu, Y.Z. Liang, ChemoPy: freely available python package for
computational biology and chemoinformatics, Bioinformatics 29 (2013) 1092-1094.
[39] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell, A.J. Olson,
AutoDock4 and AutoDockTools4: Automated Docking with Selective Receptor Flexibility,
Journal of Computational Chemistry 30 (2009) 2785-2791.
[40] W.L. DeLano, PyMOL molecular viewer: Updates and refinements, Abstracts of Papers of
the American Chemical Society 238 (2009).
[41] M. Frisch, G. Trucks, H.B. Schlegel, G. Scuseria, M. Robb, J. Cheeseman, G. Scalmani, V.
Barone, B. Mennucci, G.e. Petersson, Gaussian 09. Gaussian, Inc. Wallingford, CT: 2009.
[42] D.A. Pearlman, D.A. Case, J.W. Caldwell, W.S. Ross, T.E. Cheatham, S. DeBolt, D.
Ferguson, G. Seibel, P. Kollman, AMBER, a package of computer programs for applying
molecular mechanics, normal mode analysis, molecular dynamics and free energy
calculations to simulate the structural and energetic properties of molecules, Computer
Physics Communications 91 (1995) 1-41.
[43] J. Wang, R.M. Wolf, J.W. Caldwell, P.A. Kollman, D.A. Case, Development and testing of
a general amber force field, Journal of Computational Chemistry 25 (2004) 1157-1174.
[44] J.A. Izaguirre, D.P. Catarello, J.M. Wozniak, R.D. Skeel, Langevin stabilization of
molecular dynamics, Journal of Chemical Physics 114 (2001) 2090-2098.
[45] C. Sagui, T.A. Darden, Molecular dynamics simulations of biomolecules: long-range
electrostatic effects, Annual Review of Biophysics and Biomolecular Structure 28 (1999)
155-179.

ACCEPTED MANUSCRIPT
[46] A.G. Bailey, C.P. Lowe, MILCH SHAKE: an efficient method for constraint dynamics
applied to alkanes, Journal of Computational Chemistry 30 (2009) 2485-2493.

ACCEPTED MANUSCRIPT
A list of captions
Scheme 1. General synthetic route for the target compounds.
Figure 1. Chemical structure of A114, A117 and design strategy of four series derivatives.
Figure 2. Plots of predicted activity against the corresponding experimental activity on FGFR1
inhibition.
Figure 3. D12 and D15 inhibit FGFR1 through a mechanism that is independent of the
concentration of ATP.
Figure 4. Backbone RMSDs are shown as a function of time for FGFR1/D12 and FGFR1/D15
complexes structures at 50 ns.
Figure 5. Molecular dynamics analysis of D12/D15 to the activity cavity of FGFR1.
Figure 6. Active compounds (C5, D12, D14 and D15) effectively inhibited the FGFR1 kinase
activity and the proliferation of three GC cells (SGC7901, MGC803, and BGC823).
Figure 7. Compounds D12 and D15 suppressed bFGF-induced phosphorylation of FGFR1 and
the downstream ERK1/2 in a concentration-dependent manner.
Figure 8. Compound D12 and D15 induced cell cycle arrest at the G0/G1 phase in SGC7901
cells.
Figure 9. (A) Morphological changes and Hoechst staining were observed in SGC7901 cells
cultured with D12, D15 (5, 10, and 20 ꢀM) for 12 h (200×). (B) Effect of D12 and D15 on the
activation of PARP and inhibition of Bcl-2 in SGC7901 cells.

ACCEPTED MANUSCRIPT
Figure 1. Chemical structure of A114, A117 and design strategy of four series
derivatives.

ACCEPTED MANUSCRIPT
Scheme 1. General synthetic route for the target compounds.
O
O
H3CO
CHO
H3CO
j
k
OCH3
OH
i
h
O
H3CO
HO
CHO
R'
+
O
O
O
O
11
D14-D15
2
10
9
Reagents and conditions: (a) HCl gas, EtOH, 40%–50%; (b) THF, propionyl, or
isobutyryl chloride, Et3N, rt, 8%–31%; (c) p-TSA, cyclohexane, reflux, 50%; (d)
EtOH, 90 °C, 30%; (e) 10% HCl solution, rt, 50%; (f1, f2) differently aldehydes, HCl
gas, EtOH, 50–70 °C, 10%–70%; (g1, g2) THF, propionyl or isobutyryl chloride,
Et3N, rt, 5%–54%; (h) PPTS, CH2Cl2, rt, 50%; (i) 20% NaOH solution, EtOH, rt, 50%;
(j) substituted aldehydes, NaOH/dioxane suspension, 10%–70%; (k) 10% HCl
solution, rt, 50%;
Abbreviation: HCl, hydrochloric acid; EtOH, ethanol; THF, tetrahydrofuran; Et3N,
triethylamine; rt, room temperature; p-TSA,para-toluenesulfonic acid; IR%, inhibitory
ratio at a concentration of 20µM.

ACCEPTED MANUSCRIPT
Equation
IRFGFR1=-701.6195(±60.5745)-1163.8396
(±235.7815)MoRSEC27-4.2567(±0.4952)S35
+268.1683(±21.7818)bcute5
N=41, R2=0.8331, Radj2= 0.8196,
s=15.97, F3,37=61.58, p=1.865e-14.
Figure 2. Plots of predicted activity against the corresponding experimental activity
on FGFR1 inhibition.
Abbreviation: MoRSEC27, 3D-MoRSE - signal 27 / weighted by atomic charge;
S35, Sum of E-State of atom type: dO; bcute5, Highest eigenvaluen.5 of Burden
matrix/weighted by atomic Sanderson electronegativities; N, the number of
compounds taken into account in the regression; R2, the multiple correlation
coefficient; Radj2, adjusted multiple correlation coefficient; s, residual standard error;
and the F value is related to the F-statistic analysis (Fischer test). The numbers in
parentheses mean the standard deviation of the coefficients. p means the significance
of the variables in the model.

ACCEPTED MANUSCRIPT
Global fitting of D12
Global fitting of D15
[D12] (µM)
[D15] (µM)
1.5
1.0
0.5
0.0
1.5
1.0
0.5
0.0
0.00
1.30
5.20
20.80
0.00
1.25
5.00
20.00
0
1000
2000
3000
4000
5000
0
1000
2000
3000
4000
5000
[ATP] (µM)
[ATP] (µM)
Figure 3. D12 and D15 inhibit FGFR1 through a mechanism that is independent of
the concentration of ATP. Selective ATP-competitive kinase assay of compounds D12
and D15 with FGFR1 through Caliper Mobility Shift Assay. The conversion data were
fitted with Graphpad for global fitting.

ACCEPTED MANUSCRIPT
Figure 4. Backbone RMSDs are shown as a function of time for FGFR1/D12 and
FGFR1/D15 complexes structures at 50 ns. A) Time evolution of the RMSD of
FGFR1 and D12 are shown with green lines and blue lines, respectively. (B) Time
evolution of the RMSD of FGFR1 and D15 are shown with green lines and blue lines,
respectively. (C, D) Per-residue of top 10 contribution to the binding effective energy
of D12/D15. Per-residue contributions were calculated by the MM-GBSA
decomposition method.