Experience-based discovery (EBD) of aryl hydrazines as new scaffolds for the development of LSD1/KDM1A inhibitors

Article abstract of DOI:10.1016/j.ejmech.2019.01.075

Phenelzine was first employed to design new aryl hydrazine-based LSD1 inhibitors based on the experience-based discovery (EBD) strategy. Among these compounds, D8 potently inhibited LSD1 (IC₅₀ = 882.30 nM) in a reversible manner. Compound D8 was selective to LSD1 over MAO-A/B and showed H3K4me2 competitive binding to LSD1. The interaction between H3K4me2 and LSD1 was also confirmed by the Co-IP assay. In LSD1 overexpressed A549 cells, compound D8 dose-dependently induced accumulation of LSD1 substrates H3K4me1/2 and H3K9me1/2, showed cellular target engagement to LSD1 and significantly inhibited cell migration of A549 cells. Docking studies suggested that compound D8 occupied the peptide binding region and therefore blocked the access of the peptide substrate to the FAD, finally leading to the demethylase activity inhibition of LSD1. The findings indicate that aryl hydrazines are new scaffolds for the design of LSD1 inhibitors, the identification of D8 provides further evidence for our previously proposed general principle that fused heterocycles with an amine group are potentially active toward LSD1 by competitive binding to LSD1 with H3 peptide substrates.

Full text of DOI:10.1016/j.ejmech.2019.01.075

Accepted Manuscript
Experience-based discovery (EBD) of aryl hydrazines as new scaffolds for the
development of LSD1/KDM1A inhibitors
Zhong-Rui Li, Shuai Wang, Linlin Yang, Xiao-Han Yuan, Feng-Zhi Suo, Bin Yu, Hong-
Min Liu
PII:
S0223-5234(19)30095-9
DOI:
https://doi.org/10.1016/j.ejmech.2019.01.075
EJMECH 11082
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 13 November 2018
Revised Date: 1 January 2019
Accepted Date: 29 January 2019
Please cite this article as: Z.-R. Li, S. Wang, L. Yang, X.-H. Yuan, F.-Z. Suo, B. Yu, H.-M. Liu,
Experience-based discovery (EBD) of aryl hydrazines as new scaffolds for the development of LSD1/
KDM1A inhibitors, European Journal of Medicinal Chemistry (2019), doi: https://doi.org/10.1016/
j.ejmech.2019.01.075.
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
Experience-based discovery (EBD) of aryl hydrazines as new scaffolds for the
development of LSD1/KDM1A inhibitors
Zhong-Rui Li a,c,d,#, Shuai Wang a,c,d,#, Linlin Yang e, Xiao-Han Yuan a,c,d, Feng-Zhi Suo
a,c,d, Bin Yu a,b,c,d,*, and Hong-Min Liu a,c,d,*
School of Pharmaceutical Sciences, Zhengzhou University, Zhengzhou 450001,
.
a
China;
b State key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Nanjing
210023, Jiangsu, People’s Republic of China.
c
Key Laboratory of Advanced Drug Preparation Technologies (Ministry of
Education), Zhengzhou University, Zhengzhou 450001, China;
d
Co-Innovation Center of Henan Province for New Drug R & D and Preclinical
Safety, Zhengzhou University, Zhengzhou 450001, China;
e
Department of Pharmacology, School of Basic Medical Sciences, Zhengzhou
University, Zhengzhou 450001, China;
* Corresponding authors:
Dr. Bin Yu (zzuyubin@hotmail.com)
Prof. Hong-Min Liu (liuhm@zzu.edu.cn)
# Zhong-Rui Li and Shuai Wang contributed equally to this work.
1 / 36

ACCEPTED MANUSCRIPT
Abstract: Phenelzine was first employed to design new aryl hydrazine-based LSD1
inhibitors based on the experience-based discovery (EBD) strategy. Among these
compounds, D8 potently inhibited LSD1 (IC50 = 882.30 nM) in a reversible manner.
Compound D8 was selective to LSD1 over MAO-A/B and showed H3K4me2
competitive binding to LSD1. The interaction between H3K4me2 and LSD1 was also
confirmed by the Co-IP assay. In LSD1 overexpressed A549 cells, compound D8
dose-dependently induced accumulation of LSD1 substrates H3K4me1/2 and
H3K9me1/2, showed cellular garget engagement to LSD1 and significantly inhibited
cell migration of A549 cells. Docking studies suggested that compound D8 occupied
the peptide binding region and therefore blocked the access of the peptide substrate to
the FAD, finally leading to the demethylase activity inhibition of LSD1. The findings
indicate that aryl hydrazines are new scaffolds for the design of LSD1 inhibitors, the
identification of D8 provides further evidence for our previously proposed general
principle that fused heterocycles with an amine group are potentially active toward
LSD1 by competitive binding to LSD1 with H3 peptide substrates.
Keywords: Epigenetic regulation; Aryl hydrazines; [1,2,4]triazolo[1,5-a]pyrimidine;
LSD1 inhibitors; Experience-based discovery.
2 / 36

ACCEPTED MANUSCRIPT
1. Introduction
Histone methylation plays a key role in regulating gene expression and genetic
stability, dysregulation of this process has been observed in various cancers [1].
Methyltransferases and demethylases are two important classes of enzymes in this
process, the methyltransferases are responsible for establishing methylation patterns,
and the lysine demethylases are able to catalyze the demethylation of histone specific
lysine residues [2]. The demethylases can be divided into two subgroups based on
their catalytic mechanisms: the Flavin Adenine Dinucleotide (FAD)-dependent LSD1
(also known as KDM1A)/LSD2 and JmjC domain containing JMJD family [3-6].
Since its first identification of LSD1 in 2004 by Professor Shi Yang [7], LSD1 has
been found to be overexpressed in many tumor cells [8, 9], overexpression of LSD1
promotes tumorigenesis [10] and inhibits p53 function [11], thus affecting cancer cell
cycle [12], tumor growth, invasion and metastasis [13-15]. Therefore, targeting LSD1
has proven to be a promising therapeutic strategy for cancer therapy [16]. To date,
numerous LSD1 inhibitors including metal complexes have been identified [17-24],
of which tranylcypromine (TCP), GSK-2879552 [25, 26], ORY-1001 [27], IMG-7289,
CC-90011, INCB059872 [28, 29] and ORY-2001 alone or in combination with other
therapeutic agents are currently being investigated in clinical trials. However, it is
worth noting that all these LSD1 inhibitors in clinical trials are TCP-based irreversible
inhibitors [30, 31], suggesting that TCP is a privileged scaffold for the design of
LSD1 inhibitors. In contrast, another two irreversible LSD1 inhibitors phenelzine (Fig.
1) [32-34] and pargyline [35-38] and further structural modifications are less studied.
Following our previous work on the discovery of LSD1 inhibitors [39-49], herein we
reported the design of aryl hydrazines as new LSD1 inhibitors based on the
experience-based discovery (EBD) strategy, of which D8 inhibited LSD1 potently
(IC50 = 882.3 nM), exhibited selectivity to LSD1 over MAO-A/B, and showed
H3K4me2 competitive binding to LSD1. In A549 cells, D8 showed cellular target
engagement to LSD1, increased expression levels of H3K4me1/2 and H3K9me1/2,
and concentration-dependently suppressed migration and evasion. Our findings may
3 / 36

ACCEPTED MANUSCRIPT
suggest that the aryl hydrazine would be a new scaffold for the development of LSD1
inhibitors.
Figure 1. Experience-based discovery (EBD) of aryl hydrazines as new LSD1
Inhibitors
2. Results and Discussion
2.1. Chemistry
The general synthetic route of compounds D1-D31 is shown in Scheme 1.
3-Amino-5-mercapto-1,2,4-triazole reacted with different chlorides under reflux in the
presence of Na2CO3 in acetone to generate compounds A1-A20, which were then
subjected to the cyclization reactions with different β-ketoesters in acetic acid under
reflux, affording compounds B1-B22. Compounds B1-B22 reacted with POCl3 to
yield compounds C1-C22, which then reacted with different amines and hydrazine
hydrate to afford D1-D29 (Scheme 1A). Compound B23 was synthesized from A1
and methyl 2-oxocyclopentane-1-carboxylate in acetic acid under reflux. Subsequent
chlorination of B23 in POCl3 generated the corresponding product C23. Treatment of
C23 with hydrazine hydrate afforded D30 (Scheme 1B). Following the similar route
for the synthesis of D1-D29, D31 was also efficiently synthesized from
4H-1,2,4-triazol-3-amine (Scheme 1C).
4 / 36

ACCEPTED MANUSCRIPT
Scheme 1. Synthesis of compounds D1-D31. Reagents and conditions: (a) Na2CO3,
acetone, 60 oC, 3-5h, reflux; (b) Acetic acid, 120 oC, 2-15 h, reflux; (c) POCl3, 90 oC,
3-5 h, reflux; (d) Appropriate amine, Et3N, EtOH, rt, 5h or hydrazine hydrate (65％),
EtOH, rt, 5h.
2.2. Experience-based discovery (EBD) of aryl hydrazines as new LSD1 inhibitors
To date, there are loads of information about the 3D structures of therapeutically
relevant targets and their ligand structures (successful or inactive) and properties, the
knowledge are particularly valuable for hit identification and further optimization
[50-53]. The experience or knowledge-based design strategy has witnessed success in
drug discovery [50, 51]. Based on reported X-ray crystal structures of LSD1 and
related inhibitors, we first proposed a general principle that fused heterocycles
installed with a terminal basic group are potentially active toward LSD1 [8]. Such
compounds are able to occupy the H3 peptide substrate-binding region and form
electrostatic interactions with the negatively charged residues through the basic
moieties, thereby inhibiting the demethylase activity of LSD1. Based on the
5 / 36

ACCEPTED MANUSCRIPT
experience and the scaffold hopping/replacement strategy, we have successfully
designed two series of thiosemicarbazide containing heterocycles that inhibit LSD1
with the IC50 values as low as 154 nM [44, 54]. The heterocycle fused pyrimidine
derivatives have exhibited promising antiproliferative activity [55-60], extensive
structural modifications based on the heterocyclic scaffolds have led to the discovery
of new LSD1 inhibitors in our group [43, 44, 46, 54]. To further validate our
previously proposed general principle for designing LSD1 inhibitors, herein we
employed the [1,2,4]triazolo[1,5-a]pyrimidine as the heterocyclic scaffold and the
different amines to design new LSD1 inhibitors (Fig. 2). As shown in Fig. 2,
compounds D1-D7 bearing different amine groups were inactive against LSD1 (IC50 >
50 µM), GSK2879552 was used as the control compound that inhibits LSD1 with an
IC50 value of 24 nM [54]. To our delight, compound D8 bearing a hydrazine group
inhibited LSD1 potently with an IC50 value of 882.3 nM. Conceivably, other aromatic
heterocycles could also be potentially used to design new LSD1 inhibitors when
installed with a hydrazine moiety. The identification of D8 further provides evidence
for our proposed principle for the development of LSD1 inhibitors.
Figure 2. Discovery of [1,2,4]triazolo[1,5-a]pyrimidine-based aryl hydrazine D8 as a
new LSD1 inhibitors
2.3. Structure-activity relationship studies (SARs) of aryl hydrazines
6 / 36

ACCEPTED MANUSCRIPT
The acceptable potency of D8 toward LSD1 promoted us to carry out further SARs,
leading to the generation of compounds D9-D25 and D31 (Fig. 3). Removal of the
benzylthio group of D8 led to the complete loss of the anti-LSD1 activity (D31, IC50 >
50 µM). Replacement of the benzylthio group with several alkyl groups yielded
compounds D9-D14, which exhibited decreased anti-LSD1 activity. D10 and D11
showed moderate inhibitory activity against LSD1 with the IC50 values of 8.21 and
5.13 µM, respectively. The results revealed that the benzylthio group was an essential
structural element for the anti-LSD1 activity. Therefore, subsequent molecular
fine-turning of D8 was performed, generating compounds D13 and D15-D25.
Evidently, molecular fine-turning of D8 did not lead to increase of the activity. Of
these compounds, D16, D17, D21 and D23 bearing the F, Cl, -Me, and -CF3 group,
respectively showed relatively better potency. D23 inhibited LSD1 with an IC50 value
of 2.32 µM, but was about 2.6-fold less potent than D8. However, replacement of the
phenyl ring in D8 with the naphthyl group gave compound D13, which was found to
be devoid of the anti-LSD1 activity (IC50 > 50 µM). We next introduced the
heterocyclic benzothiophene group to the bicyclic aromatic scaffold, yielding D15
that inactivated LSD1 with an IC50 value of 1.61 µM. Besides, replacement of the
methyl group in D8 with steric isopropyl and phenyl groups afforded compounds D28
and D29, which also had decreased anti-LSD1 activity. Compound D30 with a
cyclopentyl group fused to the pyrimidine ring showed comparable inhibitory activity
with D28, but was significantly less potent than D8. The SARs studies indicated that
the benzyl group attached to the core scaffold is preferred over hydrogen, alkyl,
substituted benzyl and bicyclic aromatic groups, the S-linker elongation also led to the
decrease of the activity, and replacement of the methyl group in D8 with steric
substituents (e.g. i-Pr and Ph) caused significant decrease of the activity. The findings
revealed that different from phenelzine, the heteroaryl hydrazine is another new
scaffold for the design of LSD1 inhibitors, but subtle variations of substituents could
lead to decrease of the anti-LSD1 activity.
7 / 36

ACCEPTED MANUSCRIPT
Figure 3. Structure-activity relationship studies of aryl hydrazines.
8 / 36

ACCEPTED MANUSCRIPT
2.4. Selectivity, reversibility and competitive inhibition of compound D8.
In sequence homology, LSD1 is highly similar (~70%) to that of monoamine oxidases
A and B (abbreviated as MAO-A/B) [46, 61]. In light of the acceptable potency of D8
toward LSD1 (Fig. 5A), we next examined the inhibitory activity of compound D8
against MAO-A/B. As shown in Fig. 5B, compound D8 inactivated MAO-A with the
inhibitory rates of 36% and 21% at 10 and 1.0 µM, respectively. Meanwhile, we also
observed that compound D8 inhibited MAO-B with the inhibitory rates of 43% and
15% at 10 and 1 µM, respectively. These results indicated that compound D8 was
selective to LSD1 over MAO-A/B.
Like phenelzine (Fig. 1) [32, 33], compound D8 also possessed a hydrazine group,
whether D8 could inhibit LSD1 irreversibly remains unclear. To evaluate the
reversibility of compound D8 against LSD1, a dilution assay was employed using
GSK2879552 as the control. As shown in Fig. 5C, 80-fold dilution of the LSD1/D8
mixture resulted in recovery of LSD1 activity, while for the group treated with
GSK2879552, 80-fold dilution failed to rescue the LSD1 activity. These results
indicated that D8 inhibited LSD1 in a reversible manner, although D8 also had the
hydrazine group.
The fused heterocycles bearing a basic group could occupy the H3 peptide
substrate-binding region and therefore competitively bind to LSD1 with H3 peptide
substrates. To examine whether D8 could inhibit LSD1 through the H3 substrate
competitive binding manner, we performed the competitive analysis with classic
Lineweaver−Burk plots. As shown in Fig. 5D, D8 inhibited LSD1 in the H3K4me2
competitive binding manner. The results provide further evidence for our proposed
general principle that fused heterocycles with a basic group are potentially active
LSD1 inhibitors [8].
9 / 36

ACCEPTED MANUSCRIPT
Figure 5. Biochemical characterization of compound D8 toward LSD1 activity in
vitro. (A) Inhibition of LSD1 by compound D8 and SP2509 against LSD1. (B)
Selectivity of compound D8 to LSD1 over MAO-A/B. (C) The reversibility of
compound D8 and GSK2879552 (1.6 µM) to LSD1 were determined by the dilution
o
assay after incubation 1 h at 25 C. (D) Lineweaver-Burk plot of competitive
inhibition for compound D8 with indicated concentrations of H3K4me2. (∗∗) P < 0.01
was considered statistically significant.
2.5. Cellular target engagement of D8 to LSD1 in LSD1 overexpressed A549 cells.
In order to validate the cellular target engagement of D8, we chose LSD1
overexpressed human lung cancer cells A549 [25, 54, 62] for the cellular thermal shift
assay (CETSA) [63]. As shown in Fig. 6, treatment with D8 at 10 µM at 46 and 49 oC
increased thermal stability of cellular LSD1. Particularly, compound D8 increased
o
cellular thermal stability of LSD1 in A549 cells at 52 C, while at 52 °C, the cellular
LSD1 was denatured in A549 cells with DMSO treatment. In contrast, after treatment
with the negative control compound D31, the thermal stability of cellular LSD1 was
similar with DMSO group. This data suggested that D8 showed cellular target
engagement to LSD1 in A549 cells.
10 / 36

ACCEPTED MANUSCRIPT
Figure 6. Cellular target engagement to LSD1 in A549 cells. Protein levels of LSD1
in A549 cells were examined by western blotting analysis after treatment with D8 (A)
or D31 (B) at 10 µM or DMSO for 1 h, followed by heating at different temperatures
for 3 min. GAPDH was used as a loading control. (∗) P < 0.05 and (∗∗) P < 0.01 was
considered statistically significant. Data are the mean ± SD. All experiments were
carried out at least three times.
2.6. Compound D8 induced accumulation of methylated H3 peptide substrates in
A549 cells.
LSD1 is capable of removing the methyl groups of H3K4me1/2 and H3K9 me1/me2
in FAD-dependent manner [8], LSD1 inhibitors can cause accumulation of H3 peptide
substrates in cells. Herein the effects of compound D8 on H3K4me1/2 and H3K9
me1/2 in A549 cells were investigated. Like GSK-LSD1, compound D8 induced
accumulation of H3K4me1/2 and H3K9me1/2 in A549 cells in a dose-dependent
manner after treatment with D8 for 4 days (Fig. 7A) while GSK-LSD1 had no effect
on H3K9me1 (Fig. 7B). As expected, the inactive compound D31 has no effect on H3
peptide substrates (Fig. 7C). These results could validate the cellular target
engagement of D8 in A549 cells.
11 / 36

ACCEPTED MANUSCRIPT
Figure 7. Effects on histone methylation in A549 cells The expression levels of
H3K4me1/2 and H3K9me1/2 were then determined after treatment with compound
D8 (A) or GSK-LSD1 (B) or D31 (C) for 96 h by the western blot and quantitative
analysis of protein expression. The histone 3 (H3) was used as the loading control.
Data are the mean ± SD. (∗) P < 0.05 and (∗∗) P < 0.01 were considered statistically
significant. All experiments were carried out at least three times.
2.7. Migration inhibition of compound D8 against A549 cells.
LSD1 is implicated in epithelial-mesenchymal transition (EMT) process and is
responsible for the migration and invasion of cancer cells [64, 65]. Pharmacological
inhibition of LSD1 has been reported to be able to inhibit cell migration in vitro and
12 / 36

ACCEPTED MANUSCRIPT
in vivo through multiple mechanisms [66-68]. Herein we also evaluated the effect of
compound D8 on cell migration of A549 cells by using GSK-LSD1 and D31 as the
positive and negative control, respectively. Microphotographs showed that untreated
A549 cells filled most of the wounded area 2 days later after scratching the cell
monolayer, whereas D8 treatment markedly suppressed the wound healing in a
centration-dependent manner (Fig. 8A). The high content analysis indicated that
compound D8 has a stronger effect on cell migration suppression than GSK-LSD1
(Fig. 8B). To further illustrate the mechanism of D8, we also examined the expression
levels of biomarkers of EMT - E-Cadherin, Claudin-1 and N-Cadherin, Slug and Snail
We found that the epithelial cells biomarkers E-Cadherin and Claudin-1 were
upregulated, whereas N-Cadherin, Slug and Snail was down-regulated after
compound D8 and GSK-LSD1 treatment (Fig. 8C), indicating the inhibitory effect of
EMT. We also found that D8 and GSK-LSD1 had no effect on the expression levels of
LSD1, but both of them inhibited migration of A549 cells. The data may suggest that
the migration inhibition of A549 cells is due to the inhibition of the demethylase
activity of LSD1, not the expression changes of LSD1.
13 / 36

ACCEPTED MANUSCRIPT
Figure 8. Effects of compound D8, GSK-LSD1 and D31 on cell migration in A549
cells. (A) Wound healing assay. (B) Migration ability assay. Treatment of compound
D8 resulted in migration suppression. (D) The expressions of E-Cadherin, Claudin-1,
N-Cadherin, Slug, Snail and LSD1 after treatment with compound D8 for 96 h.
GAPDH was used as the loading control and quantitative analysis of protein
expression. (∗) P < 0.05 and (∗∗) P < 0.01 were considered statistically significant.
Data are the mean ± SD. All experiments were carried out at least three times.
2.8. Co-immunoprecipitation analysis
To evaluate whether D8 can disrupt the interaction between LSD1 and H3K4me2, the
co-immunoprecipitation was performed using LSD1 antibody as a bait. After
treatment with compound D8 for 4 days, the binding of H3K4me2 to LSD1decreased
(Fig. 9). This result indicated that D8 could disrupt the interaction between LSD1 and
H3K4me2 in A549 cells.
14 / 36

ACCEPTED MANUSCRIPT
Figure 9. Immunoprecipitation of LSD1 and H3K4me2 without (control) or with D8
treatment in A549 cells; TCL represents total cell lysate. GAPDH was used as loading
control.
2.9. The effect of D8 on the cell viability of GES-1 cells
In view of the potency of D8 toward LSD1, we next detected the cytotoxity of D8
against normal gastric epithelial cell line GES-1 by the MTT assay. After treatment
with compound D8 for 4 days, D8 did not show remarkable effect on the cell viability
of GES-1 cells with the IC50 value above 50 µM (Fig. 10), suggesting the low toxicity
of D8.
Figure 10. Cytotoxity of compound D8 against normal gastric epithelial cell line
GES-1.
2.10. In silico studies
To explain the observed biological activity of D8, an in silico docking analysis was
performed using the MOE 2015.10. With previous experiment, we characterized that
D8 was a competitive inhibitor over H3K4me2, indicating that D8 may bind to the
active site of LSD1 where H3K4me2 occupies. Thus, our docking study was based on
the X-ray structure of LSD1-CoREST complex with a bound H3K4me2-like peptide
15 / 36

ACCEPTED MANUSCRIPT
inhibitor (PDB code: 2V1D) [69]. After the bound 21-mer pLys4Met (“p” is for
peptide) H3 peptide was removed, D8 was docked into the binding site (Fig. 11).
The binding mode of D8 in the LSD1-CoRERST revealed that D8 binds to the same
region as that occupied by the dimethyl pLys4, pGln5 and pThr6 of the H3K4me2
peptide (Fig. 11B). In this model, the phenyl group in D8 is situated in a hydrophobic
cavity surrounded by residues L693, L677, W695, I356 and V333 and establishes
pi-pi stacking interactions with both W695 and H564 (Fig. 11A). This is the region
that is occupied by the side chain of pThr6 in the predicted H3K4me2 bound model.
Meanwhile, the bicyclic N-heterocyclic ring of D8 lies in proximity to the FAD
cofactor, with the ArNH- moiety forming a hydrogen bond with N5 of the flavin ring
of FAD (Fig. 11A). The position of the bicyclic N-heterocyclic ring in D8 is further
stabilized by a hydrogen bond between the ArNHNH2 and residue Y761 as well as the
pi-pi stacking interaction formed between the bicyclic N-heterocyclic ring and F538,
while compound D1 bearing an amino group (Fig. 2) was inactive toward lSD1,
highlighting the importance of the hydrazine group in D8 to the activity. Interestingly,
the position of the pLys4-N-CH3 group (the site of oxidative demethylation) in the
predicted H3K4me2 bound model is only about 3Å from the reactive N5 atom of the
flavin (Fig. 11B) and the flavin-methylamino group geometry is necessary and
identical to those of the flavodehydrogenases-oxidases [70]. Therefore, the binding of
D8 should block the access of the peptide substrate to the FAD and prevent the
flavin-mediated oxidative demethylation of the N-CH3 group of methyl-Lys4.
Our docking results highlight that the binding site of D8 is exactly the same region as
both H3K4me2 peptide and the dimethyl pLys4. A combination of both polar and
nonpolar interactions position D8 right in front of the FAD cofactor, occupying the
cavity and blocking the access of the peptide to the oxidative demethylation site. This
in silico model not only provides a rational of D8 acting as a competitive inhibitor
over H3K4me2, but also suggests that modifications on D8 would lead to more potent
analogues targeting LSD1.
16 / 36

ACCEPTED MANUSCRIPT
Figure 11. Three-dimensional view of D8 binding mode in LSD1. (A) D8-LSD1
interactions in the binding pocket. D8 is shown in green, FAD yellow and LSD1 cyan.
Residues in LSD1 that are involved in interactions with D8 are shown in sticks.
Hydrogen bonds are shown in red dashed lines. (B) Comparison of binding modes
between D8 and H3K4me2 peptide substrate. The orientation is the same as in panel
A. Model of dimethyl pLys4 (“p” is for peptide) H3 peptide bound in the active site
was obtained by replacing pMet4 of the 21-mer pLys4Met H3 peptide in the PDB
structure 2V1D as described previously [69].
3. Conclusion
In this work, a series of new aryl hydrazine-based LSD1 inhibitors were designed
from phenelzine based on the experience-based discovery (EBD) strategy. Of these
compounds, D8 potently inhibited LSD1 with an IC50 value of 882.30 nM in a
reversible manner. The structure-activity relationship studies (SARs) were also
performed to reveal the essential structural elements for the activity. Compound D8
was selective to LSD1 over MAO-A/B and showed H3K4me2 competitive binding to
LSD1. In LSD1 overexpressed A549 cells, compound D8 dose-dependently induced
accumulation of LSD1 substrates H3K4me1/2 and H3K9me1/2, showed cellular
garget engagement to LSD1 and significantly inhibited cell migration of A549 cells.
Further studies indicated that treatment with D8 led to up-regulation of E-Cadherin
and Claudin-1 and down-regulation of N-Cadherin, Slug and Snail. Docking studies
revealed that D8 occupied the same region as both H3K4me2 peptide and the
dimethyl pLys4 and therefore blocked the access of the peptide substrate to the FAD
and prevented the flavin-mediated oxidative demethylation of the N-CH3 group of
17 / 36

ACCEPTED MANUSCRIPT
methyl-Lys4. The findings indicate that aryl hydrazines are new scaffolds for the
design of LSD1 inhibitors, the identification of D8 provides further evidence for our
previously proposed general principle that fused heterocycles with an amine group are
potentially active toward LSD1 by competitive binding to LSD1 with H3 peptide
substrates.
4. Experimental section
4.1. General Information.
Reagents and solvents were purchased from commercial sources and were used
without further purification. Thin-layer chromatography (TLC) was carried out on
glass plates coated with silica gel (Qingdao Haiyang Chemical Co., G60F-254) and
visualized by UV light (254 nm). The products were purified by column
chromatography over silica gel (Qingdao Haiyang Chemical Co., 200–300 mesh).
Melting points were determined on an X-5 micromelting apparatus and are
uncorrected. All the NMR spectra were recorded with a Bruker DPX 400 MHz
spectrometer with TMS as internal standard in CDCl3 or DMSO-d6. Chemical shifts
are given as δ ppm values relative to TMS. High-resolution mass spectra (HRMS)
were recorded on a Waters Micromass Q-T of Micromass spectrometer by
electrospray ionization (ESI).
4.2. General method for the synthesis of A1-A20.
Compounds A1-A20 were prepared following the previously reported method [43]. To
a solution of 3-amino-5-mercapto-1,2,4-triazole (1.0 g, 8.61 mmol) in acetone (30 mL)
were added sodium carbonate (1.37 g, 12.92 mmol), sodium iodide (129.06 mg, 0.861
mmol) and alkyl halide (9.47 mmol). The reaction mixture was stirred at 60 oC for 3-6
h before cooling to room temperature. Na2CO3 and NaI were removed via filtration.
The residue was concentrated under vacuum and then dissolved in EtOAc. The
organic layer was washed with water (2 × 10 mL) and brine (2 × 20 mL) and then
dried over anhydrous MgSO4. After removal of the solvent, the resulting residue was
subjected to column chromatography, giving the corresponding products A1-A20.
5-(Benzylthio)-4H-1,2,4-triazol-3-amine (A1), white solid, yield: 73 %. 1H NMR
(400 MHz, DMSO-d6) δ 7.41-7.18 (m, 5H), 4.34 (s, 2H). 13C NMR (100 MHz,
18 / 36

ACCEPTED MANUSCRIPT
DMSO-d6) δ 152.19, 147.08, 136.55, 128.79, 128.43, 127.48, 35.41. HRMS (ESI):
m/z calcd for C9H9N4S (M-H)-, 205.0548; found, 205.0548.
4.3. General method for the synthesis of B1-B24.
Compounds B1-B24 were prepared following the previously reported method [43] as
described for B1. A1 (1g, 4.85 mmol) and ethyl acetoacetate (630.94mg, 4.85mmol)
was stirred in AcOH (10 mL) at 120 oC. After 3-6 h, the reaction mixture was cooled
to room temperature and the white precipitate formed was filtered and washed with
water to afford compound B1. 2- (Benzylthio)- 5-methyl- [1,2,4]triazolo[1,5-a]
pyrimidin- 7(4H)-one (B1), white solid, yield: 85 %. m.p.: 240-245 oC. 1H NMR (400
MHz, DMSO-d6) δ 13.17 (s, 1H), 7.44 (d, J = 7.1 Hz, 2H), 7.32 (t, J = 7.3 Hz, 2H),
7.26 (d, J = 7.2 Hz, 1H), 5.80 (s, 1H), 4.43 (s, 2H), 2.29 (s, 3H). 13C NMR (100 MHz,
DMSO-d6) δ 161.92, 154.75, 151.09, 150.63, 137.35, 128.77, 128.38, 127.22, 98.46,
34.48, 18.43. HRMS (ESI): m/z calcd for C13H11N4OS (M-H)-, 271.0654; found,
271.0660.
4.4. General method for the synthesis of C1-C24.
Compounds C1-C24 were prepared following the previously reported method [43] as
described for the synthesis of C1. B1 was stirred in POCl3 (excess) at 90 oC for 3-5h.
The resultant oil was added dropwise to ice cold bath. The resulting aqueous mixture
was extracted with CH2Cl2 (4 × 30 mL) and the organic layer was washed by
saturated NaHCO3 (3×10 mL). The organic layers were dried over MgSO4 and
concentrated under reduced pressure to give the orange solid C1, which was used
directly without further purification. C1 was unstable enough and therefore only
characterized by HRMS. 2- (Benzylthio)- 7- chloro- 5-methyl- [1,2,4]triazolo[1,5-a]
pyrimidine (C1), HRMS (ESI): m/z calcd for C13H12ClN4S (M+H)-, 291.0471; found,
291.0463.
4.5. General method for the synthesis of D1-D31.
Method A (for the synthesis of D1-D7): Compound C1 (150 mg, 1.0 eq.) was
dissolved in ethanol (2 mL), followed by addition of ammonium hydroxide (25％, 1.0
eq.) and triethylamine (78.30 mg, 1.5 eq.). The reaction mixture was stirred at room
temperature for 5 h. The resultant mixture was concentrated under reduced pressure to
19 / 36

ACCEPTED MANUSCRIPT
the residue, which was then purified by column chromatography to give D1. D2-D7
were synthesized following the route as described for that of D1.
Method B (for the synthesis of D8-D31): Compound C1 (150 mg, 1.0 eq.) was
dissolved in ethanol (2 mL), followed by addition of hydrazine hydrate (1.0 eq.). The
reaction mixture was stirred at room temperature for 5 h. The resultant mixture was
concentrated under reduced pressure to the residue, which was then purified by
column chromatography to give D8. D9-D31 were synthesized following the route as
described for that of D8.
1
Compound D1, yellow solid, yield: 89 %. m.p.: 196-201 oC. H NMR (400 MHz,
DMSO-d6) δ 7.93 (s, 2H), 7.55-7.43 (m, 2H), 7.37-7.28 (m, 2H), 7.24 (dd, J = 8.4, 6.1
13
Hz, 1H), 6.12 (s, 1H), 4.49 (s, 2H), 2.37 (s, 3H). C NMR (100 MHz, DMSO-d6) δ
164.29, 162.84, 155.87, 147.50, 138.07, 128.92, 128.38, 127.16, 90.25, 34.36, 24.37.
HRMS (ESI): m/z calcd for C13H12N5S (M-H)-, 270.0813; found, 270.0821.
o
1
Compound D2, white solid, yield: 68 %. m.p.: 168-170 C. H NMR (400 MHz,
CDCl3) δ 7.34 (d, J = 6.4 Hz, 2H), 7.23 – 7.11 (m, 3H), 5.88 – 5.68 (m, 1H), 4.38 (d,
J = 2.7 Hz, 2H), 3.25 (d, J = 2.0 Hz, 2H), 2.99 – 2.88 (m, 2H), 2.39 (d, J = 3.2 Hz,
3H). 13C NMR (100 MHz, CDCl3) δ 169.23, 168.58, 160.82, 151.81, 143.26, 134.15,
133.64, 132.42, 93.28, 49.90, 39.69, 29.92. HRMS (ESI): m/z calcd for C13H12N5S
(M+H)+, 315.1392; found, 315.1390.
o
1
Compound D3, white solid, yield: 72 %. m.p.: 193-196 C. H NMR (400 MHz,
DMSO-d6) δ 7.47 (d, J = 7.3 Hz, 2H), 7.28 (m, 3H), 6.39 (s, 1H), 4.50 (s, 2H), 3.64 (t,
J = 5.4 Hz, 2H), 3.52 (m, 7H), 3.33 (t, J = 5.7 Hz, 3H), 2.62 (t, J = 5.7 Hz, 2H), 2.41
13
(s, 3H). C NMR (100 MHz, DMSO-d6) δ 164.58, 163.88, 156.04, 146.99, 138.49,
129.40, 128.90, 127.69, 88.70, 79.66, 73.03, 70.26, 70.07, 68.97, 41.93, 41.53, 34.91,
25.18. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 403.1916; found, 403.1915.
o
1
Compound D4, white solid, yield: 78 %. m.p.: 138-142 C. H NMR (400 MHz,
CDCl3) δ 7.50 – 7.43 (m, 2H), 7.28 (d, J = 1.3 Hz, 1H), 7.25 – 7.16 (m, 2H), 6.55 (s,
1H), 5.86 (s, 1H), 4.49 (s, 2H), 3.32 (d, J = 6.3 Hz, 2H), 2.67 – 2.58 (m, 2H), 2.49 (s,
3H), 2.28 (s, 6H). 13C NMR (100 MHz, CDCl3) δ 169.34, 168.74, 160.75, 151.44,
20 / 36

ACCEPTED MANUSCRIPT
143.27, 134.16, 133.62, 132.42, 93.25, 62.18, 50.31, 39.69, 29.94. HRMS (ESI): m/z
calcd for C13H12N5S (M+H)+, 343.1705; found, 343.1706.
o
1
Compound D5, white solid, yield: 65 %. m.p.: 151-154 C. H NMR (400 MHz,
CDCl3) δ 7.46 (d, J = 7.0 Hz, 2H), 7.33 – 7.18 (m, 3H), 5.91 (s, 1H), 4.50 (s, 2H),
4.09 (t, J = 6.8 Hz, 2H), 3.20 (s, 3H), 2.56 (t, J = 6.8 Hz, 2H), 2.50 (s, 3H), 2.20 (s,
6H). 13C NMR (100 MHz, CDCl3) δ 165.23, 163.68, 157.89, 148.97, 137.66, 129.01,
128.49, 127.26, 92.53, 57.68, 51.03, 45.75, 39.65, 35.52, 25.05. HRMS (ESI): m/z
calcd for C13H12N5S (M+H)+, 357.1861; found, 357.1863.
o
1
Compound D6, white solid, yield: 85 %. m.p.: 151-153 C. H NMR (400 MHz,
DMSO-d6) δ 7.47 (m, 2H), 7.31 (m, 1H), 7.27 – 7.21 (m, 2H), 7.01 (d, J = 7.0 Hz,
1H), 6.50 (s, 1H), 4.46 (s, 2H), 4.30 (d, J = 11.8 Hz, 1H), 4.13 (d, J = 12.5 Hz, 1H),
3.21 (m, 2H), 2.59 (s, 1H), 2.43 (s, 3H), 1.86 (s, 2H), 1.67 – 1.47 (m, 2H), 1.33 (s,
9H). 13C NMR (100 MHz, DMSO-d6) δ 164.07, 138.34, 129.49, 129.35, 128.94,
128.86, 127.65, 94.87, 78.36, 52.72, 48.47, 34.79, 28.65, 24.94, 23.40. HRMS (ESI):
m/z calcd for C13H12N5S (M+H)+, 455.2229; found, 455.2230.
o
1
Compound D7, white solid, yield: 62 %. m.p.: 192-195 C. H NMR (400 MHz,
DMSO-d6) δ 8.23 (s, 3H), 7.45 (d, J = 7.3 Hz, 2H), 7.31 (t, J = 7.4 Hz, 2H), 7.23 (t, J
= 7.2 Hz, 1H), 6.55 (s, 1H), 4.48 (s, 2H), 3.54 – 3.28 (m, 3H), 2.45 (s, 3H), 2.07 (s,
1H), 1.86 (s, 1H), 1.68 (d, J = 8.1 Hz, 2H). 13C NMR (100 MHz, DMSO-d6) δ 164.58,
163.98, 157.12, 149.02, 138.30, 129.32, 128.89, 127.70, 95.47, 50.38, 48.39, 46.47,
34.86, 28.25, 24.65, 22.52. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 355.1705;
found, 355.1706.
o
1
Compound D8, white solid, yield: 76 %. m.p.: 205-208 C. H NMR (400 MHz,
DMSO-d6) δ 9.36 (s, 1H), 7.52 – 7.45 (m, 2H), 7.34 – 7.28 (m, 2H), 7.25 (m, 1H),
6.47 (s, 1H), 4.81 (s, 2H), 4.48 (s, 2H), 2.41 (s, 3H). 13C NMR (100 MHz, DMSO-d6)
δ 161.92, 154.75, 151.09, 150.63, 137.35, 128.77, 128.38, 127.22, 98.46, 34.48, 18.43.
HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 287.1079; found, 287.1078.
o
1
Compound D9, white solid, yield: 76 %. m.p.: 157-163 C. H NMR (400 MHz,
DMSO-d6) δ 7.18 (s, 3H), 6.45 (s, 1H), 4.78 (s, 2H), 3.17 (t, J = 7.2 Hz, 2H), 2.40 (s,
3H), 1.80 – 1.66 (m, 2H), 0.99 (t, J = 7.3 Hz, 3H). 13C NMR (100 MHz, DMSO-d6) δ
21 / 36

ACCEPTED MANUSCRIPT
165.17, 163.53, 156.18, 149.18, 88.51, 32.90, 25.19, 23.32, 13.58. HRMS (ESI): m/z
calcd for C13H12N5S (M+H)+, 239.1079; found, 239.1078.
1
Compound D10, white solid, yield: 69 %. m.p.: 171-177 oC. H NMR (400 MHz,
DMSO-d6) δ 9.31 (s, 1H), 6.46 (s, 1H), 6.10 – 5.87 (m, 1H), 5.34 (dd, J = 16.9, 1.1 Hz,
1H), 5.10 (d, J = 10.0 Hz, 1H), 4.80 (s, 2H), 3.88 (d, J = 6.9 Hz, 2H), 2.41 (s, 3H). 13C
NMR (100 MHz, DMSO-d6) δ 164.45, 163.63, 156.19, 149.21, 134.49, 118.47, 88.57,
33.65, 25.19. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 357.0922; found,
357.0920.
1
Compound D11, white solid, yield: 88 %. m.p.: 173-180 oC. H NMR (400 MHz,
DMSO-d6) δ 9.25 (s, 1H), 6.49 (s, 1H), 4.81 (s, 2H), 4.08 (d, J = 2.6 Hz, 2H), 3.20 (t,
J = 2.6 Hz, 1H), 2.42 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ 163.32, 163.04,
155.71, 148.78, 88.26, 80.18, 73.81, 24.73, 18.97. HRMS (ESI): m/z calcd for
C13H12N5S (M+H)+, 235.0766; found, 235.0766.
1
Compound D12, white solid, yield: 87 %. m.p.: 153-158 oC. H NMR (400 MHz,
DMSO-d6) δ 6.45 (s, 1H), 4.77 (s, 2H), 3.16 (d, J = 7.2 Hz, 2H), 2.40 (s, 3H), 1.32 –
13
1.16 (m, 1H), 0.61 – 0.49 (m, 2H), 0.32 (q, J = 4.9 Hz, 2H). C NMR (100 MHz,
DMSO-d6) δ 165.32, 163.54, 156.17, 149.17, 88.50, 36.56, 25.20, 11.72, 6.12. HRMS
(ESI): m/z calcd for C13H12N5S (M+H)+, 251.1079; found, 251.1075.
1
Compound D13, white solid, yield: 73 %. m.p.: 169-176 oC. H NMR (400 MHz,
DMSO-d6) δ 9.38 (s, 1H), 8.19 (d, J = 8.3 Hz, 1H), 7.96 (d, J = 7.7 Hz, 1H), 7.87 (d, J
= 8.2 Hz, 1H), 7.72 (d, J = 6.8 Hz, 1H), 7.65 – 7.53 (m, 2H), 7.50 – 7.39 (m, 1H),
6.49 (s, 1H), 5.00 (s, 2H), 4.82 (s, 2H), 2.43 (s, 3H). 13C NMR (100 MHz, DMSO-d6)
δ 164.25, 163.20, 155.72, 148.78, 133.47, 133.16, 130.92, 128.67, 128.17, 127.62,
126.39, 125.93, 125.42, 123.79, 88.18, 32.33, 24.71. HRMS (ESI): m/z calcd for
C13H12N5S (M-H)+, 337.1235; found, 337.1236.
1
Compound D14, white solid, yield: 83 %. m.p.: 207-213 oC. H NMR (400 MHz,
DMSO-d6) δ 9.30 (s, 1H), 8.07 (d, J = 7.6 Hz, 2H), 7.70 (t, J = 7.3 Hz, 1H), 7.58 (t, J
13
= 7.6 Hz, 2H), 6.46 (s, 1H), 4.98 (s, 2H), 4.80 (s, 2H), 2.39 (s, 3H). C NMR (100
MHz, DMSO-d6) δ 194.03, 164.36, 163.66, 156.07, 149.22, 136.05, 134.12, 129.30,
22 / 36

ACCEPTED MANUSCRIPT
128.83, 125.15, 88.64, 25.18. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+,
315.1028; found, 315.1028.
o
1
Compound D15, white solid, yield: 65 %. m.p.: 218-223 C. H NMR (400 MHz,
DMSO-d6) δ 9.35 (s, 1H), 8.10 – 7.95 (m, 3H), 7.42 (m, 1H), 6.48 (s, 1H), 4.76 (s,
3H), 2.41 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ 164.45, 163.75, 156.25, 149.28,
139.64, 138.67, 132.00, 130.00, 128.95, 128.36, 125.19, 125.01, 122.13, 88.66, 28.06,
25.21. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 377.0410; found, 377.0411.
1
Compound D16, white solid, yield: 79 %. m.p.: 168-170 oC. H NMR (400 MHz,
DMSO-d6) δ 9.32 (s, 1H), 7.64 – 7.45 (m, 3H), 7.21 – 7.06 (m, 1H), 6.47 (s, 1H), 4.80
(s, 2H), 4.47 (s, 2H), 2.41 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ 163.93, 148.76,
137.50, 131.74, 130.78, 128.27, 124.50, 88.22, 54.87, 33.46, 24.57. HRMS (ESI): m/z
calcd for C13H12N5S (M+H)+, 305.0985; found, 305.0984.
o
1
Compound D17, white solid, yield: 65 %. m.p.: 196-201 C. H NMR (400 MHz,
DMSO-d6) δ 9.35 (s, 1H), 7.51 (s, 2H), 7.36 (s, 2H), 6.47 (s, 1H), 4.80 (s, 2H), 4.47 (s,
2H), 2.41 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ 163.72, 156.20, 149.25, 131.29,
128.78, 88.68, 33.96, 25.20. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+,
321.0689; found, 321.0689.
1
Compound D18, white solid, yield: 84 %. m.p.: 203-206 oC. H NMR (400 MHz,
DMSO-d6) δ 9.38 (s, 1H), 7.78 – 7.69 (m, 2H), 7.65 (d, J = 7.8 Hz, 1H), 7.32 (t, J =
13
7.2 Hz, 1H), 7.22 (m, 1H), 6.48 (s, 1H), 4.82 (s, 2H), 4.57 (s, 2H), 2.42 (s, 3H). C
NMR (100 MHz, DMSO-d6) δ 164.26, 163.76, 156.24, 149.27, 137.54, 133.18,
131.95, 130.03, 128.33, 124.52, 88.70, 35.56, 25.21. HRMS (ESI): m/z calcd for
C13H12N5S (M+H)+, 365.0184; found, 365.0184.
1
Compound D19, white solid, yield: 78 %. m.p.: 215-219 oC. H NMR (400 MHz,
DMSO-d6) δ 9.36 (s, 1H), 8.15 (d, J = 8.1 Hz, 2H), 7.78 (d, J = 8.1 Hz, 2H), 6.47 (s,
1H), 4.82 (s, 2H), 4.59 (s, 2H), 2.40 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ
164.04, 163.77, 156.21, 149.24, 147.26, 146.99, 130.69, 123.90, 88.73, 33.92, 25.18.
HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 332.0930; found, 332.0932.
o
1
Compound D20, white solid, yield: 82 %. m.p.: 227-232 C. H NMR (400 MHz,
DMSO-d6) δ 9.35 (s, 1H), 7.38 (t, J = 10.1 Hz, 2H), 6.87 (t, J = 8.2 Hz, 2H), 6.47 (s,
23 / 36

ACCEPTED MANUSCRIPT
1H), 4.81 (s, 2H), 4.43 (s, 2H), 3.72 (d, J = 3.8 Hz, 3H), 2.41 (s, 3H). 13C NMR (100
MHz, DMSO-d6) δ 164.82, 163.64, 158.92, 156.19, 149.22, 130.63, 130.33, 114.27,
88.60, 55.52, 34.41, 25.20. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 317.1185;
found, 317.1186.
1
Compound D21, white solid, yield: 71 %. m.p.: 199-206 oC. H NMR (400 MHz,
DMSO-d6) δ 9.34 (s, 1H), 7.34 (t, J = 9.2 Hz, 2H), 7.11 (d, J = 7.7 Hz, 2H), 6.47 (s,
1H), 4.80 (s, 2H), 4.44 (s, 2H), 2.41 (s, 3H), 2.26 (s, 3H). 13C NMR (100 MHz,
DMSO-d6) δ 164.76, 163.66, 156.19, 149.23, 136.85, 135.44, 129.43, 129.33, 88.60,
34.65, 25.21, 21.16. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 301.1235; found,
301.1236.
1
Compound D22, white solid, yield: 59 %. m.p.: 183-186 oC. H NMR (400 MHz,
DMSO-d6) δ 9.36 (s, 1H), 7.54 (d, J = 8.0 Hz, 2H), 7.46 – 7.30 (m, 1H), 6.50 (s, 1H),
4.79 (d, J = 7.4 Hz, 4H), 2.43 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ 148.82,
141.87, 135.05, 132.33, 131.16, 130.43, 128.79, 96.68, 88.31, 31.39, 24.70. HRMS
(ESI): m/z calcd for C13H12N5S (M+H)+, 355.0299; found, 355.0300.
1
Compound D23, white solid, yield: 76 %. m.p.: 176-182 oC. H NMR (400 MHz,
DMSO-d6) δ 9.32 (s, 1H), 7.69 (dd, J = 24.1, 7.6 Hz, 4H), 6.47 (s, 1H), 4.55 (s, 2H),
2.40 (s, 3H), 1.05 (t, J = 6.8 Hz, 2H). 13C NMR (100 MHz, DMSO-d6) δ 163.75,
163.33, 155.70, 148.80, 136.12, 132.87, 131.91, 128.14, 127.11, 126.81, 126.14,
126.08, 125.72, 123.00, 88.29, 55.99, 31.21, 24.69, 18.50. HRMS (ESI): m/z calcd for
C13H12N5S (M+H)+, 355.0953; found, 355.0954.
1
Compound D24, white solid, yield: 85 %. m.p.: 207-213 oC. H NMR (400 MHz,
DMSO-d6) δ 7.84 (d, J = 7.6 Hz, 1H), 7.75 (d, J = 7.8 Hz, 1H), 7.63 (t, J = 7.5 Hz,
1H), 7.51 (t, J = 7.6 Hz, 1H), 6.50 (s, 1H), 4.82 (s, 2H), 4.67 (s, 2H), 2.42 (s, 3H). 13C
NMR (100 MHz, DMSO-d6) δ 164.24, 163.82, 156.19, 149.29, 136.62, 133.38,
132.42, 128.65, 127.60, 127.31, 126.65, 126.59, 126.23, 123.50, 88.79, 31.70, 25.20.
HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 355.0953; found, 355.0953.
1
Compound D25, white solid, yield: 64 %. m.p.: 201-205 oC. H NMR (400 MHz,
DMSO-d6) δ 9.38 (s, 1H), 7.87 (s, 1H), 7.83 (d, J = 7.6 Hz, 1H), 7.60 (d, J = 7.8 Hz,
1H), 7.53 (t, J = 7.7 Hz, 1H), 6.47 (s, 1H), 4.81 (s, 2H), 4.57 (s, 2H), 2.41 (s, 3H). 13C
24 / 36

ACCEPTED MANUSCRIPT
NMR (100 MHz, DMSO-d6) δ 164.25, 163.75, 156.22, 149.25, 140.62, 133.61,
129.86, 129.67, 129.35, 125.98, 125.94, 124.33, 124.30, 123.26, 88.69, 34.04, 25.20.
HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 355.0953; found, 355.0954.
1
Compound D26, white solid, yield: 76 %. m.p.: 198-203 oC. H NMR (400 MHz,
DMSO-d6) δ 9.05 (s, 1H), 7.27 (d, J = 31.9 Hz, 5H), 6.47 (s, 1H), 4.76 (s, 2H), 3.47 –
3.39 (m, 2H), 3.03 (t, J = 7.5 Hz, 2H), 2.42 (s, 3H). 13C NMR (100 MHz, DMSO-d6)
δ 164.95, 163.58, 156.23, 149.23, 140.56, 129.12, 128.80, 126.78, 88.58, 35.94, 32.27,
25.19. HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 301.1235; found, 301.1234.
1
Compound D27, white solid, yield: 69 %. m.p.: 219-223 oC. H NMR (400 MHz,
DMSO-d6) δ 9.25 (s, 1H), 7.28 (dd, J = 10.1, 4.6 Hz, 2H), 7.24 – 7.13 (m, 3H), 6.46 (s,
1H), 4.78 (s, 2H), 3.34 (s, 1H), 3.20 (t, J = 7.2 Hz, 2H), 2.80 – 2.67 (m, 2H), 2.40 (s,
3H), 2.12 – 1.92 (m, 2H). 13C NMR (100 MHz, DMSO-d6) δ 162.72, 155.34, 151.64,
151.11, 141.51, 128.83, 128.78, 126.37, 98.95, 34.43, 31.13, 30.64, 18.98. HRMS
(ESI): m/z calcd for C13H12N5S (M+H)+, 315.1392; found, 315.1393.
1
Compound D28, white solid, yield: 73
%
. m.p.: 176-179 oC. H NMR (400 MHz,
DMSO-d6) δ 9.35 (s, 1H), 7.48 (d, J = 7.2 Hz, 2H), 7.31 (t, J = 7.3 Hz, 2H), 7.28 –
7.22 (m, 1H), 6.50 (s, 1H), 4.82 (s, 2H), 4.50 (s, 2H), 2.95 (m, 1H), 1.24 (d, J = 6.9
13
Hz, 6H). C NMR (100 MHz, DMSO-d6) δ 172.16, 164.85, 156.32, 149.62, 138.62,
129.37, 128.88, 127.65, 86.45, 36.59, 34.82, 22.42. HRMS (ESI): m/z calcd for
C13H12N5S (M+H)+, 315.1392; found, 315.1381.
o
1
Compound D29, white solid, yield: 82 %. m.p.: 185-192 C. H NMR (400 MHz,
DMSO-d6) δ 8.18 – 8.08 (m, 2H), 7.61 – 7.46 (m, 5H), 7.32 (t, J = 7.3 Hz, 2H), 7.29 –
13
7.23 (m, 1H), 7.12 (s, 1H), 4.54 (s, 2H). C NMR (100 MHz, DMSO-d6) δ 155.41,
152.21, 137.21, 132.43, 132.22, 131.72, 131.22, 129.44, 128.89, 128.05, 98.25, 34.26.
HRMS (ESI): m/z calcd for C13H12N5S (M+H)+, 349.1235; found, 349.1236.
o
1
Compound D30, white solid, yield: 51 %. m.p.: 225-228 C. H NMR (400 MHz,
DMSO-d6) δ 9.00 (s, 1H), 7.40 (t, J = 7.6 Hz, 2H), 7.23 (dd, J = 9.9, 4.6 Hz, 2H), 7.17
(dd, J = 8.4, 6.1 Hz, 1H), 4.77 (s, 2H), 4.40 (s, 2H), 3.23 – 3.11 (m, 2H), 2.73 (t, J =
7.8 Hz, 2H), 1.95 (m, 2H). 13C NMR (100 MHz, DMSO-d6) δ 162.15, 154.56, 154.05,
25 / 36

ACCEPTED MANUSCRIPT
151.98, 137.96, 129.34, 128.93, 127.77, 110.67, 35.07, 31.69, 27.29, 22.11. HRMS
(ESI): m/z calcd for C13H12N5S (M+H)+, 313.1235; found, 313.1236.
o
1
Compound D31, white solid, yield: 56 %. m.p.: 162-167 C. H NMR (400 MHz,
13
DMSO-d6) δ 9.36 (s, 1H), 8.35 (s, 1H), 6.50 (s, 1H), 4.80 (s, 2H), 2.44 (s, 3H). C
NMR (100MHz, DMSO-d6) δ 163.51, 155.34, 154.27, 149.60, 87.87, 24.76. HRMS
(ESI): m/z calcd for C13H12N5S (M+H)+, 165.0889; found, 165.0882.
4.6. Cells and cell viability assay.
A549 cell line was purchased from Shanghai Institute of Biochemistry & Cell Biology,
Chinese Academy of Sciences (Shanghai, China). The cells were grown in
RPMI-1640 medium supplemented with 10% fetal bovine serum and maintained in a
humidified atmosphere at 37 °C and 5% CO2. The Cell viability of GES-1 cells was
measured by MTT assay [41]. Cells were plated in 96-well culture plates (3× 103 cells
per well). After incubation overnight, the cells were treated with different
concentrations of D8 for 4 days. DMSO (0.1%) was used as a vehicle. MTT (5
mg/mL) was dissolved in PBS and filter sterilized, then 20 µL of the prepared solution
was added to each well and cells were incubated for 4 h at 37 ℃. The formed
formazan crystals were dissolved in DMSO (150 µL/well) by constant shaking for 10
min. The absorbance was measured at a test wavelength of 570 nm. Cell viability was
calculated by the following formula: % Cell viability = (At/As) × 100%. At and As
denoted the absorbance of the test substances and solvent control, respectively.
4.7. Inhibitory evaluation of D8 against LSD1, MAO-A, and MAO-B.
To evaluate the IC50 of D8 against LSD1, compound D8 was incubated with the
recombinant LSD1 and H3K4me2. After that, the fluorescence was measured at
excitation wavelength 530 nm and emission wavelength 590 nm as reported in order
to evaluate the inhibition rate of D8. Inhibitory effect of D8 against MAO-A and
MAO-B were evaluated with the commercialized kit.
4.8. Reversibility assay.
An amount of 2.5 µg of LSD1 recombinant was incubated with 40 µM of compound
D8, 1.6 µM of GSK2879552 or DMSO. At 1 h later, 1.25 µL aliquots were removed
from all samples and diluted into HRP-assay solution containing substrate (1.25 µL)
26 / 36

ACCEPTED MANUSCRIPT
and coupling reagents to a final volume of 100 µL. These samples were the
after-dilution samples. As for before-dilution samples, all remained mixture were
added into HRP-assay solution containing substrate (10 µL) and coupling reagents to
a final volume of 100 µL, respectively. If the compound inhibits the activity of LSD1
in a reversible manner, LSD1 activity will recovery after dilution. If a compound is an
irreversible inhibitor, LSD1 activity will not be recovered after dilution.
4.9. Competition assay.
For the competitive analysis of compound D8, the demethylase activity of LSD1 was
assessed in the presence of different concentrations of compound D8 (0, 0.31, 0.63,
and 1.25 µM) at a fixed concentration of FAD (2.5 µM) and peptide concentrations
from 0.8 to 106 µM. Assays were performed triplicate, and kinetics values were
obtained using Lineweaver-Burk plots.
4.10. Wound healing assay.
For the wound healing assay, cells were seeded in a 24-wells plate (CORNING, USA)
at 1 × 104 per well. After 24 h incubation, the cell surface was scratched using a 10 µL
pipet tip. Then, the cells were cultured with fresh medium containing 1% FBS and
different concentrations of compound for 48 h, and photographed on an inverted
microscope.
4.11. Transwell assay.
For the migration assay, 400 µL medium without FBS, different concentrations of
compound and 8, 000 cells were added to each upper chamber. In the lower chamber,
600 µL medium with 20% FBS was used as chemoattractant. After incubation for 48
h, both chambers were washed by PBS for three times. After staining with Hoechst
33258 (10 µg/mL) and twice wash, migrated cells were detected and numbered using
high content screening system (ArrayScan XTI, Thermo Fisher Scientific, MA).
4.12. Cellular thermal shift assay.
Cells (5 × 105 per sample) were treated with compound or with DMSO for 1 h,
washed with PBS three times, and dissolved in 50 µL PBS supplemented with a
protease inhibitor, followed by heating at the indicated temperatures. Treated cells
were then subjected to snap-freezing in liquid nitrogen and thawed on ice for 3 cycles.
27 / 36

ACCEPTED MANUSCRIPT
The protein levels of LSD1 in equal amounts of the supernatant were examined by
western blots. GAPDH was used as the control. Results are representative of three
independent experiments.
4.13. Western blot analysis.
Cells were seeded and treated with 0, 1, and 6 µM of compound D8 for 96 h. Cells
were harvested with trypsin and were lysed by radio immunoprecipitation assay
(RIPA) lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.25% sodium
deoxycholate, 0.1% Nonidet P-40, 0.1% Triton X-100) with the complete proteinase
inhibitor cocktail (Roche, Basel, Switzerland) to extract the total proteins. Samples
were kept on ice for 30 min followed by centrifugation at 12,000 g at 4 ℃ for 10
min. Then, the protein concentration of the supernatants was determined using the
BCA protein assay kit (Beyotime, Haimen, China). After added with loading buffer,
cell lyses were boiled for 10 min at 100˚C. Next, aliquots of protein from the total cell
or tumor tissues lysates (40 lg/lane) were separated via sodium dodecyl sulfate (10%
or 12%) polyacrylamide gel electrophoresis (SDS-PAGE, Bio-Rad Laboratories,
Hercules, CA), wet transferred to a PVDF membrane (Bio-Rad Laboratories,
Hercules, CA) and blotted with primary antibodies specific for H3, H3K4me1,
H3K4me2, H3K9me1, H3K9me2, LSD1, E-Cadherin, N-Cadherin, Claudin-1, Slug,
Snail and GAPDH (which was used as the internal standard) overnight. After washing
the membrane with TBST (TBS, 0.05% Tween-20) /TBS three times (5 min per wash),
blots were incubated with the secondary antibody (1:5000) at room temperature for 2
h. Finally, the blots were washed in TBST/TBS. The antibody-reactive bands were
revealed by enhanced chemiluminescence (ECL) and exposed on Kodak radiographic
film.
4.14. Immunoprecipitation assay
Cells were seeded and treated with 0 and 6 µM of compound D8 for 96 h. Then, the
culture medium was carefully removed from cells, and the cells were washed once
with PBS. Cells were harvested with trypsin and lysed by IP Lysis/Wash Buffer with
the complete proteinase inhibitor cocktail (Roche, Basel, Switzerland) to extract the
total proteins. Samples were kept on ice for 5 min followed by centrifugation at
28 / 36