Development and evaluation of selective, reversible LSD1 inhibitors derived from fragments

Article abstract of DOI:10.1039/c3md00226h

Two series of aminothiazoles have been developed as reversible inhibitors of lysine specific demethylase 1 (LSD1) through the expansion of a hit derived from a high concentration biochemical fragment based screen of 2466 compounds. The potency of the initial fragment hit was increased 32-fold through synthesis, with one series of compounds showing clear structure-activity relationships and inhibitory activities in the range of 7 to 187 μM in a biochemical assay. This series also showed selectivity against the related FAD-dependent enzyme mono-amine oxidase A (MAO-A). Although a wide range of irreversible inhibitors of LSD1 have been reported with activities in the low nanomolar range, this work represents one of the first reported examples of a reversible small molecule inhibitor of LSD1 with clear SAR and selectivity against MAO-A, and could provide a platform for the development of more potent reversible inhibitors. Herein, we also report the use of a recently developed cell-based assay for profiling LSD1 inhibitors, and present results on our own compounds as well as a selection of recently described reversible LSD1 inhibitors.

Full text of DOI:10.1039/c3md00226h

MedChemComm
View Article Online
View Journal | View Issue
CONCISE ARTICLE
Development and evaluation of selective, reversible
LSD1 inhibitors derived from fragments†
Cite this: Med. Chem. Commun., 2013,
4, 1513
a
James R. Hitchin, Julian Blagg,b Rosemary Burke,b Samantha Burns,b
*
Mark J. Cockerill,a Emma E. Fairweather,a Colin Hutton,a Allan M. Jordan,a
Craig McAndrew,b Amin Mirza,b Daniel Mould,a Graeme J. Thomson,a Ian Waddella
and Donald J. Ogilviea
Two series of aminothiazoles have been developed as reversible inhibitors of lysine specific demethylase 1
(LSD1) through the expansion of a hit derived from a high concentration biochemical fragment based
screen of 2466 compounds. The potency of the initial fragment hit was increased 32-fold through
synthesis, with one series of compounds showing clear structure–activity relationships and inhibitory
activities in the range of 7 to 187 mM in a biochemical assay. This series also showed selectivity against
the related FAD-dependent enzyme mono-amine oxidase A (MAO-A). Although a wide range of
irreversible inhibitors of LSD1 have been reported with activities in the low nanomolar range, this work
represents one of the first reported examples of a reversible small molecule inhibitor of LSD1 with clear
SAR and selectivity against MAO-A, and could provide a platform for the development of more potent
reversible inhibitors. Herein, we also report the use of a recently developed cell-based assay for profiling
LSD1 inhibitors, and present results on our own compounds as well as a selection of recently described
reversible LSD1 inhibitors.
Received 6th August 2013
Accepted 28th September 2013
DOI: 10.1039/c3md00226h
www.rsc.org/medchemcomm
A variety of different inhibitors of LSD1 have been reported,
and the most potent of these inhibitors are based on tranylcy-
1. Introduction
promine (TCP, 1) (Fig. 1).7 TCP is a clinically validated amine
oxidase inhibitor with potent activities against MAO-A and B,
and was originally developed for use as an antidepressant in the
1960s. Given the high level of sequence homology between the
active sites of MAO-A and B and LSD1, TCP was evaluated as an
inhibitor of LSD1 and showed moderate levels of inhibitory
LSD1 (also known as KDM1A, AOF2, BHC110 or KIAA0601) was
the rst histone demethylase to be discovered and plays
important roles in normal and malignant cells, chiey as a
transcription repressor.1,2 LSD1 belongs to the avin adenine
dinucleotide (FAD)-dependent amine oxidase family of deme-
thylases and catalyses the demethylation of the mono- and di-
methylated lysine residues of histones, specically at the H3K4
and H3K9 positions, as well as other proteins such as TP53 and
DNMT1.1–3 There is currently considerable interest in the
potential of LSD1 as a therapeutic target in human malignan-
cies because its expression correlates with poor prognosis in
prostate cancer, ER-negative breast cancer, non-small cell lung
cancer and neuroblastoma.4–6 Furthermore, high levels of LSD1
expression have been observed in bladder cancer and acute
myeloid leukemia (AML).1 These elevated levels of LSD1 can
result in the epigenetic reprogramming of the genome, with
concomitant reduction in the expression of certain tumour
suppressor genes and a more dedifferentiated phenotype.
aCancer Research UK Drug Discovery Unit, Cancer Research UK Manchester Institute,
The University of Manchester, Manchester, M20 4BX, UK. E-mail: jhitchin@picr.man.
ac.uk; Fax: +44 (0)161 446 3109; Tel: +44 (0)161 918 7160
bCancer Research UK Cancer Therapeutics Unit, Division of Cancer Therapeutics, The
Institute of Cancer Research, London, SM2 5NG, UK
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c3md00226h
Fig. 1 Reported inhibitors of LSD1.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 1513–1522 | 1513

View Article Online
MedChemComm
Concise Article
activity (2 mM).7 Based on this early work, and the emerging oxazoles as novel inhibitors of LSD1 with anti-proliferative
body of literature implicating LSD1 in a number of cancers,1,4–6 properties. A total of four compounds were reported in this
interest in TCP-based inhibitors has increased considerably, paper with IC50 values in the range of 9.5 to >50 mM. Compound
with a number of different research groups working towards the 5 was reported to be the most active of these compounds (IC50
¼
development of functionalized TCP-based inhibitors with 9.5 mM). Interestingly, however, the same compounds also gave
enhanced levels of potency.8–15 Most notably, Oryzon11–13,15 have IC50 values in the range of 1.2 to 2.1 nM in a HeLa cell viability
recently disclosed a series of TCP-based analogues with low assay, representing an apparent increase of 6000 to 23 000-fold
nanomolar potencies for LSD1 and high levels of selectivity in their potency. The signicant disparity in these potencies
against MAO-A and B (>3000-fold). Furthermore, two of the suggests the occurrence of non-specic toxicity effects, and that
these compounds (compounds 2 and 3, Fig. 1) induced cell these compounds are particularly cytotoxic. Furthermore, no
differentiation and signicantly reduced the clonogenic biochemical reversibility experiments were conducted in this
potential of primary human MLL-AF9 AML cells.16 The TCP- study to determine whether these compounds were acting as
based inhibitors are irreversible and deliver their inhibitory reversible or irreversible inhibitors of LSD1.
activity by becoming covalently bound to the FAD cofactor
Tortorici et al.26 recently described a series of short peptides
within the LSD1 active site.17–19 Although some of the more (Mw < 800) as reversible inhibitors of LSD1 with Ki values as low
recently reported compounds belonging to this class have as 0.14 mM. These compounds are particularly interesting
shown high levels of selectivity towards LSD1 over MAO-A and B, because they are the only reversible inhibitors of LSD1 that have
their selectivity against other FAD dependent enzymes has not been reported to date that have been backed up by X-ray
yet been described.
structural data to validate their competitive and reversible
In contrast to the irreversible TCP-based inhibitors, reports inhibition of LSD1.
concerning the development of potent, drug-like and reversible
In our on-going efforts to identify small drug-like molecules
small-molecule inhibitors of LSD1 have been scarce.20–23 There as therapeutic agents against oncology targets, we recently
is a high level of sequence homology between the active site of became interested in LSD1 and the development of reversible
LSD1 and those of N1-acetylpolyamine oxidase (APAO) and inhibitors of this particular enzyme. Herein, we describe the use
spermine oxidase (SMO),2 and several different series of poly- of a high concentration fragment-based biochemical screening
amine compounds related to spermine, bearing guanidine and approach to identify inhibitors of LSD1 and the subsequent
biguanidine moieties, have been reported as inhibitors of development and elaboration of these fragment hits into low
LSD1.20–22 Compounds belonging to this particular class were micromolar reversible inhibitors. Furthermore, we describe the
shown to be noncompetitive inhibitors of LSD1, with respect to application of a recently developed cellular mode of action
FAD, and compounds such as 4, in particular, have been assay27 based on the levels of CD86 expression in THP1 cells to
reported to inhibit the activity of LSD1 by >50% at 1 mM in a measure the inhibitory activities of known and investigational
biochemical assay as well as resulting in a dose-dependent (0.5 LSD1 inhibitors in cells.
to 10 mM) increase in the levels of H3K4Me and H3K4Me2 in
HCT116 human colon carcinoma cells.20 The (bis)urea and (bis)
thiourea analogues of compound 4 have also been synthesized
and reported to inhibit the activity of LSD1 by >50% at 10 mM in
2. Results and discussion
Fragment screen
a biochemical assay.21 Three of these thiourea compounds were A fragment screen was conducted against LSD1 using a frag-
also evaluated in Calu-6 human anaplastic non-small cell lung ment library composed of 2466 compounds (Fig. 2). All of the
carcinoma cells and HCT116 human colon carcinoma cells, compounds were initially screened in a single point in vitro
where they led to increases in global H3K4Me and H3K4Me2 inhibition assay at 300 mM using a biochemical AlphaScreen
levels. Although these compounds demonstrated activity assay (ESI†). One hundred and four compounds showed greater
towards LSD1 in a biochemical assay and delivered an increase than 50% inhibition at this concentration, representing a hit
in global H3K4Me and H3K4Me2 levels in a number of different rate of 4.2%, and were progressed for further evaluation.
cancer cell lines, they are not drug-like in a conventional sense
and it is unlikely that they could be developed into useful 300 mM, with 50 of the original 104 hits being conrmed as
therapeutic agents for clinical use in humans. showing inhibitory activity towards LSD1. An interference assay
An AlphaScreen conrmation assay was conducted at
In a separate publication, Woster et al.24 reported the use of a (ESI†) was then conducted, where the assay was performed in
virtual screening approach to identify a series of low molecular the absence of the LSD1 enzyme. The results of this assay
weight amidoximes. The most potent of these compounds 6 revealed that 43 of the 50 compounds tested gave greater than
exhibited moderate activity towards LSD1 in an biochemical 50% interference. These compounds were subsequently
assay (IC50 ¼ 16.8 mM) as well as producing dramatic changes in removed and the remaining seven fragments were subjected to
methylation at the H3K4 chromatin mark in Calu-6 human a full 10-point IC50 analysis using the AlphaScreen activity and
anaplastic non-small cell lung carcinoma cells and promoting interference protocols. Based on the quality of the IC50 data, as
the re-expression of aberrantly silenced tumour suppressor well as the drug-likeness of the compounds and their perfor-
genes.
mances in the interference and counter-screening assays
More recently, Pal et al.25 reported the synthesis and evalu- against other targets in our portfolio, these seven compounds
ation of a series of 3-amino-guanidine substituted phenyl were ultimately reduced to the two related aminothiazole
1514 | Med. Chem. Commun., 2013, 4, 1513–1522
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
Fig. 4 Structure of dasatinib 9.
a 5-substituted aminothiazole (Fig. 4). With this in mind, we
were encouraged by these aminothiazole fragment hits and
moved forward to explore and expand upon these fragments
through synthesis.
Chemistry. Our initial work on the fragments involved
assessing the importance of their key structural features. With
this in mind, a series of compounds were synthesised (Table 1)
according to the route depicted in Scheme 1. Briey, a mixture
of the a-bromo ketone (10) and thiourea (11) in ethanol was
heated at 60 ꢀC to give the aminothiazole 12 as the free-base
following a basic work-up procedure.
It is noteworthy that the assay format was changed at this
stage from the AlphaScreen protocol used in the initial frag-
ment screen to a homogenous time resolved uorescence
(HTRF) assay (ESI†). The inhibitory activity for the initial frag-
ment hit 7 was found to be consistent across both assay formats
(i.e., 249 vs. 121 mM, for the AlphaScreen and HTRF assays). All
of the inhibition data reported beyond this point were acquired
using the HTRF assay.
Fig. 2 Fragment screening against LSD1.
compounds 7 and 8 (Fig. 3), representing an overall hit rate
of 0.1%.
Recognising potential metabolic liabilities of the ketone
moiety,29 we investigated its role and options for its replace-
ment. Compound 12a, which lacked the carbonyl moiety, was
This low hit rate was not particularly surprising, given the
large size and poorly functionalized nature of the histone
binding site in LSD1. The two fragments 7 and 8 were struc-
turally related and gave IC50 values of 243 and 437 mM,
respectively. The validity of these fragment hits was subse-
quently conrmed using two functional orthogonal assays,
including a Caliper-based assay and an LCMS assay (ESI†).
These compounds were also conrmed to be reversible inhibi-
tors of LSD1 based on the results of a preincubation study using
the AlphaScreen protocol and compound 7 (ESI†).
One of the major issues associated with the use of amino-
thiazole-based compounds in drug discovery is that they can
potentially undergo P450-catalyzed oxidative ring opening
reactions to give the corresponding a-dicarbonyl metabolites
and thioureas, with the latter of these two compounds being
toxic to humans.28 This particular metabolic liability can be
mitigated, however, by the introduction of additional func-
tionality at the 5-position of the thiazole ring, as exemplied by
the clinically validated cancer drug dasatinib 9, which contains
Table 1 Percentage inhibition values for the aminothiazolesd
%Inhibitiona
Cmpd R1
R2
500 mM 125 mM IC50 (mM)b,c
7
8
CH3CO
CH3CO
Et
MeCH(OH) 4-Cl-Ph
PhCO
4-Cl-Ph
2-Cl-Ph
4-Cl-Ph
93
82
9
60
35
0
0
21
0
121 ꢁ 15.9
>188
12a
12b
12c
12d
12e
12f
12g
12h
12i
12j
12k
12l
12m
12n
12o
12p
12q
NDd
NDd
29
6
NDd
4-Cl-Ph
4-Cl-Ph
Ph
3-Cl-Ph
o-Tol
3,4-Cl2-Ph
3,5-Cl2-Ph
2,6-Me2-Ph 49
NDd
p-F-Ph
NDd
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
CH3CO
43
87
63
26
0
2
NDd
61
35
0
123 ꢁ 21.9
>188
NDd
0
NDd
17
74
100
16
35
5
NDd
4-Br-Ph
84
77
63
61
1
26.9 ꢁ 11.5
57 ꢁ 23.2
>188
4-HO-Ph
3-HO-Ph
4-MeO-Ph
4-NC-Ph
4-O2N-Ph
4-pyridyl
>188
NDd
3
22
8
NDd
NDd
NDd
a
All of the reported values represent the geometric mean of at least two
b
independent determinations. All of the reported values represent the
geometric mean of at least two independent determinations ꢁ SD.
c
IC50 values were only determined for compounds showing >50%
inhibition at 500 mM. d ND ¼ Not determined.
Fig. 3 Aminothiazole hits from the fragment screen of LSD1.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 1513–1522 | 1515

View Article Online
MedChemComm
Concise Article
with the corresponding chloro analogue 7. The introduction of a
phenol at the para-position (12l) similarly provided an increase
in the activity (IC50 ¼ 57 mM), although concerns were raised
over this particular derivative, given that para-substituted
phenols are well known pan assay interference compounds
(PAINS)30,31 and this result was treated with some caution.
Interestingly, the corresponding meta-phenol 12m was only
weakly active, with the para-methoxy compound 12n providing
similar results. With the exception of bromine and chlorine,
which are only weakly electron withdrawing through induction,
Scheme 1 Synthesis of the aminothiazoles (12a–q). Reagents and conditions: (a)
(i) a-bromo ketone (1 mmol), thiourea (1 mmol), EtOH, 60 ꢀC 2 h; (ii) saturated
Na2CO3 solution, DCM (52–86%).
found to be completely inactive in a single point HTRF assay, electron withdrawing groups at the para-position were poorly
and highlighted the importance of the ketone moiety to the tolerated, and compounds 12o and 12p, bearing a nitrile or
activity of the initial fragment hit. Reduction of the ketone to nitro group at this position, respectively, were completely
the corresponding alcohol with NaBH4 in methanol gave inactive. The replacement of the phenyl ring with a 4-pyridyl
compound 12b, which was also inactive against LSD1.
moiety (12q) was found to provide only very low levels of
Replacement of the methyl group of the ketone moiety with a inhibition.
phenyl ring (12c) was not tolerated and led to a complete loss of
In summary, the SAR in this particular series was found to be
inhibition, indicating that the phenyl group was too large to be very steep and, in the absence of structural data, it was difficult
accommodated. Replacement of the ketone group with a para- to envisage a strategy that would enable this series to advance
uoro substituted phenyl ring (12d) was not well tolerated towards higher levels of inhibitory activity. In an attempt to
either and resulted in a 3-fold reduction in the inhibitory further elaborate and expand upon the initial fragment hits, a
activity compared with 7. Based on these initial results, it was series of different N-functionalized aminothiazoles 13a–f was
clear that the ketone moiety was critical to the inhibitory activity synthesised incorporating different linker groups between the
of the initial fragment hits, and our attention turned to inves- amino group of the thiazole and the phenyl ring (Table 2).
tigating the impact of changes to other areas of the initial
fragment hit whilst retaining this structural motif.
The introduction of an amide, urea or sulfonamide linker to
the aminothiazole bearing a methyl ketone at the 4-position of
A series of analogues were synthesised bearing different the thiazole ring (compounds 13a–c, respectively) did not have
substituents on the phenyl ring of the phenylaminothiazole. It any discernible impact on the potency (IC50 values > 188 mM in
is noteworthy that the lack of any substitution on the phenyl all cases). In parallel, these linkers were also incorporated into
ring, as in compound 12e, resulted in a signicant reduction in the aminothiazole system bearing a para-chloro-substituted
activity compared with the initial fragment hit 7, and high- phenyl ring at its 4-position to give compounds 13d–e, respec-
lighted the importance to the activity of having some degree of tively. Although the amide and urea containing analogues 13d
substitution on this ring. Varying the regiochemistry of the and 13e, respectively, were inactive, the sulfonamide 13f
substituents around the phenyl ring showed that the presence showed moderate but surprising inhibitory activity. This result
of a chlorine atom at the meta- or para-position of the phenyl was particularly interesting because 13f lacked a ketone at the
ring was well tolerated, with the compounds 12f and 7 providing 4-position of thiazole, and suggested that aminosulfonamides
IC50 values of 123 and 121 mM, respectively. In contrast to bearing a phenyl group at the 4-position of the thiazole may
substitution at the meta- or para-position of the phenyl ring, the
presence of a chlorine atom at the ortho-position (8) was poorly
tolerated and resulted in a reduction in activity. This result was
Table 2 Elaboration of the aminothiazole core
consistent with the original AlphaScreen assay data from the
fragment screen and indicated that the presence of a substit-
uent at the ortho-position was detrimental to the activity likely
because of the loss of co-planarity between the thiazole and
phenyl rings. Compounds bearing multiple chlorine substitu-
ents on the phenyl ring were also synthesised and evaluated.
Unfortunately, however, the introduction of chlorine atoms at
the 3,4- and 3,5-positions of the phenyl ring (compounds 12h
and 12i, respectively) led to a complete loss of inhibition. A
similar result was also observed in the 2,6-dimethylsubstituted
phenyl analogue 12j, suggesting that only a single substituent
could be tolerated on this phenyl ring, preferably at the meta- or
para-position. With this in mind, we proceeded to focus on
para-substituted analogues bearing a range of electron donating
and electron withdrawing groups.
Cmpd
R1
X
IC50 LSD1 (mM)a
13a
13b
13c
13d
13e
13f
CH3CO
CH3CO
CH3CO
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
CO
CONH
SO2
CO
CONH
SO2
>188
>188
>188
>188
>188
85.7 ꢁ 23.7
a
The presence of a bromine atom at the para-position (12k)
led to a 5-fold increase in the activity (IC50 ¼ 25 mM) compared
All of the reported values represent the geometric mean of at least two
independent determinations ꢁ SD.
1516 | Med. Chem. Commun., 2013, 4, 1513–1522
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
represent an alternative series of LSD1 inhibitors, which also
avoid the potential metabolic liability of the ketone function-
ality in the initial fragment hit. Based on this result, we pro-
ceeded to synthesis a library of aminothiazole sulfonamides
from 4-(4-chlorophenyl)thiazol-2-amine (14, R1 ¼ H) to develop
a greater understanding of this series and enhance its potency.
Scheme 2 Synthesis of compounds 16a–y. Reagents and conditions: (a) ami-
nothiazole (1 mmol), sulfonyl chloride (1 mmol), pyridine, 80 ꢀC, 8 h (24–66%).
A
variety of different thiazolesulfonamide compounds
(16a–y) were synthesised (Table 3) according to the route shown
in Scheme 2. Briey, 4-(4-chlorophenyl)thiazol-2-amine (14,
R1 ¼ H) was reacted with a sulfonyl chloride 15 in pyridine at
80 ꢀC to give the desired thiazolesulfonamide 16 in good yield.
Whilst the parent methylsulfonamide 16a provided no
inhibition, the corresponding triuoromethyl sulfonamide 16b
showed improved activity with an IC50 of 44.4 mM together with
an 11-fold selectivity against MAO-A. This result indicated that
the acidity of the sulfonamide N–H may be important to the
inhibitory activity, and this was further conrmed through the
methylation of the sulfonamide to give 16c, which showed no
activity towards LSD1. The unsubstituted phenyl sulfonamide
13f was only weakly active, with an IC50 value of 85.7 mM. The
extension of this vector with the corresponding benzyl sulfon-
amide 16e resulted in a complete loss in activity.
We then proceeded to investigate the decoration of the
phenyl ring with a variety of electron withdrawing and electron
donating substituents. The introduction of electron donating
substituents such as methyl, methoxy and phenyl groups at the
ortho-, meta- or para-position of the phenyl ring led to increases
in the activity compared with the naked phenyl ring, and the
meta-phenyl substituted compound 16k in particular provided
an 11-fold increase in the activity with an IC50 of 7.5 mM. This
compound also showed a moderate level of selectivity against
MAO-A (5.5-fold). The introduction of electron withdrawing
substituents such as chloro, triuoromethyl, and nitro groups
at the ortho-, meta- or para-position of the phenyl ring also led
to a signicant increase in the potency relative to the bare
phenyl ring. The most notable example in this particular group
of compounds was the meta-triuoromethyl substituted
compound 16q, which was 9-fold more potent than the
unsubstituted phenyl sulfonamide 13f. Pleasingly, this
compound was also greater than 50-fold selective for LSD1 over
MAO-A.
A variety of different heteroaromatic groups were also
investigated as replacements for the phenyl ring of the sulfon-
amide. Disappointingly, however, the use of 5-membered
nitrogen-containing heterocycles, such as N-methylated imid-
azoles and pyrazoles led to a complete loss of activity (data not
shown). In contrast, the use of the more lipophilic 2-thiophenyl
sulfonamide in compound 16v gave moderate activity as well as
18-fold selectivity over MAO-A. The introduction of a 3-bromo-5-
Table
3
LSD1 inhibition and MAO-A selectivity data for the
thiazolesulfonamidesb
IC50 data (mM)a
Cmpd
R1
R2
LSD1
MAO-A
13f
H
H
H
Me
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Ph
Me
CF3
p-Tol
Bn
o-Tol
m-Tol
p-Tol
3-MeO-Ph
4-MeO-Ph
3-Ph-Ph
4-Ph-Ph
2-Cl-Ph
3-Cl-Ph
4-Cl-Ph
2-F3C-Ph
3-F3C-Ph
4-F3C-Ph
2-NO2-Ph
3-NO2-Ph
4-NO2-Ph
2-thiophene
85.7 ꢁ 23.7
>188
108 ꢁ 18.9
NDb
16a
16b
16c
16e
16f
16g
16h
16i
44.4 ꢁ 1.7
>500
>188
NDb
>188
NDb
43.2 ꢁ 10.0
17.2 ꢁ 14.2
50.5 ꢁ 6.4
43.6 ꢁ 26.7
76.05 ꢁ 29.8
7.5 ꢁ 1.5
15.8 ꢁ 0.3
40.1 ꢁ 1.9
10.3 ꢁ 3.9
13.2 ꢁ 5.2
18.1 ꢁ 1.6
9.5 ꢁ 3.4
17.3 ꢁ 5.2
49.2 ꢁ 4.6
20.3 ꢁ 0.4
26.1 ꢁ 4.1
27.8 ꢁ 3.5
89.3 ꢁ 6.5
38.3 ꢁ 8.6
20.14 ꢁ 0.6
33.0 ꢁ 16.9
28.4 ꢁ 0.5
40.3 ꢁ 23.5
>500
49.2 ꢁ 5.0
248 ꢁ 69.0
273 ꢁ 122.1
132 ꢁ 69.2
>500
16j
16k
16l
16m
16n
16o
16p
16q
16r
16s
16t
>500
88.9 ꢁ 10.1
207 ꢁ 70.4
>500
pyridyl sulfonamide to give 16w was also well tolerated (IC50
¼
19.3 mM), although the selectivity of this compound for LSD1
over MAO-A was poor. More elaborate decoration of the phenyl
ring with the N-methyl pyrazole and a 2-methyl pyrimidine
derivatives 16x and 16y was also well tolerated, with these
compounds displaying moderate levels of activity with IC50
values of 36.1 and 26.3 mM, respectively, representing a reduc-
tion in potency of at least 3.5-fold compared with the most
potent compound from this series.
16u
16v
>500
16w
H
19.3 ꢁ 3.6
29.3 ꢁ 11.6
9.5 ꢁ 1.3
16x
H
36.1 ꢁ 2.4
Taken together, these results indicated that the introduction
of a substituent at the meta-position of the phenyl ring of the
sulfonamide delivered higher levels of inhibitory activity
towards LSD1 as well as greater levels selectivity over MAO-A.
Furthermore, electron withdrawing substituents provided
higher levels of selectivity for LSD1 over MAO-A. With this in
mind, the sulfonamide was xed as the meta-chloro moiety and
16y
H
26.3 ꢁ 14.1
53.2 ꢁ 22.0
a
All of the reported values represent the geometric mean of at least two
independent determinations ꢁ SD. b ND ¼ Not determined.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 1513–1522 | 1517

View Article Online
MedChemComm
Concise Article
we proceeded to investigate the effect of different substituents
on the phenyl ring at the 4-position of the thiazole ring by
synthesizing a small library of thiazolesulfonamide compound
(19a–p) (Table 4). These derivatives were synthesised according
to the route depicted below (Scheme 3). Briey, a mixture of
a-bromo ketone (17) and thiourea in ethanol was heated at
60 ꢀC to give the aminothiazole 18. Subsequent treatment with
3-chlorobenzene-1-sulfonyl chloride in pyridine at 80 ꢀC gave
the desired thiazolesulfonamide 19 in good yield.
Scheme 3 Synthesis of compounds 19a–p. Reagents and conditions: (a) (i)
a-bromo ketone (1 mmol), thiourea (1 mmol), EtOH, 60 ꢀC 2 h; (ii) saturated
Na2CO3 solution, DCM (52–86%). (b) aminothiazole (1 mmol), 3-chlorobenzene-
1-sulfonyl chloride (1 mmol), pyridine, 80 ꢀC 8 h (25–80%).
The aminothiazole compound bearing an unsubstituted
phenyl ring at the 4-position of the thiazole 19a was some 5-fold co-planarity between the phenyl and thiazole rings, although
less active than the most potent of the corresponding mono- this could also relate to the electronics of the nitro group.
substituted systems. Furthermore, the selectivity of this
compound for LSD1 over MAO-A was only marginal.
The introduction of an electron donating methoxy group at
the ortho-, meta- or para-position of the phenyl ring was gener-
The introduction of electron withdrawing substituents such ally well tolerated. With the exception of the ortho-methoxy
as chloro, nitro and bromo groups to the ortho-, meta- or para- analogue 19j (which was inactive against MAO-A), however, the
position of the phenyl ring generally led to a 5-fold increase in meta- and para-methoxy analogues were generally more selec-
the potency compared with the unsubstituted derivative 19a, tive for MAO-A over LSD1, with the latter of the two analogues
with the most active of these compounds being the meta-chloro 19l being almost 4-fold more active towards MAO-A. These
substituted compound 19c, although this compound was only results indicated that the introduction of further electron
2.5-fold selective for LSD1 over MAO-A. Interestingly, however, donating substituents at the meta- or para-position of the
the corresponding para-chloro compound 19d demonstrated phenyl ring would favour the inhibition of MAO-A over LSD1.
25-fold selectivity for LSD1 over MAO-A whilst maintaining This hypothesis was conrmed by the synthesis of the 2,3-
equivalent potency against LSD1. Furthermore, the ortho-chloro dihydrobenzo[b][1,4]dioxine analogue 19m which effectively
compound 19b was even more selective for LSD1 over MAO-A, as contained electron donating groups at both the meta- and para-
were the ortho-nitro derivative 19d and ortho-methoxy derivative positions of the phenyl ring. This compound was even more
19j. However, the ortho-nitro analogue 19d was some 4.5-fold selective for MAO-A over LSD1 (6-fold), conrming that the
less active against LSD1 than the corresponding ortho-chloro presence of electron donating groups on the phenyl ring had an
analogue, indicating that the presence of a large group at this adverse impact on the selectivity for LSD1 over MAO-A.
position was deleterious to the activity of the compounds
Substitution at the 5-position of the thiazole ring was also
against both LSD1 and MAO-A, likely because of the loss of investigated, because substitution at this position is well-known
to mitigate the P450-catalyzed ring opening of aminothiazoles to
form the corresponding a-dicarbonyl metabolites and thio-
ureas.28 The introduction of a methyl (19n) or phenyl (19o)
substituent at the 5-position of the thiazole ring was well toler-
ated, leading toan increaseinthe potency of3-or4-fold compared
with the corresponding unsubstituted system 19a. These
compounds were also highly selective for LSD1 over MAO-A.
Taken together, these results indicate that substitution at the
ortho-, meta- or para-position of the phenyl ring at the 4-position
of the aminothiazole ring generally leads to a 2 to 5-fold
increase in the potency compared with the unfunctionalized
phenyl ring, except for bulky electron withdrawing groups at the
ortho-position (e.g., 19d). In terms of the selectivity of these
compounds for LSD1 over MAO-A, several clear trends were
observed. Substitution at the ortho-position tended to destroy
any activity towards MAO-A, with the highest level of selectivity
observed in the ortho-chloro substituted analogue, which was at
least 36-fold selective for LSD1. It is noteworthy that the upper
concentration limit of the MAO-A was 500 mM and the actual
selectivity for LSD1 over MAO-A could therefore be much
greater. Substitution at the para- or meta-position with an
electron-donating group led to a reversal in the selectivity and
the methoxy substituted compounds were more selective for
MAO-A over LSD1 (4-fold for the para-methoxy 19l). The intro-
duction of a methyl or phenyl substituent at the 5-position of
the thiazole ring also led to a complete loss of MAO-A activity. In
Table
4
LSD1 inhibition and MAO-A selectivity data for the
thiazolesulfonamides
IC50 data (mM)a
Cmpd
R1
R2
LSD1
MAO-A
16n
19a
19b
19c
19d
19e
19f
19g
19h
19i
4-Cl-Ph
Ph
2-Cl-Ph
3-Cl-Ph
2-NO2-Ph
3-NO2-Ph
4-NO2-Ph
3-EtO2C-Ph
3-Br-Ph
H
H
H
H
H
H
H
H
H
H
H
H
H
10.5 ꢁ 3.9
51.6 ꢁ 0.03
13.7 ꢁ 2.8
10.2 ꢁ 0.4
61.6 ꢁ 1.5
13.7 ꢁ 3.9
17.6 ꢁ 0.4
34.9 ꢁ 8.4
12.9 ꢁ 2.7
16.6 ꢁ 3.6
23.6 ꢁ 3.3
19.0 ꢁ 0.6
26.0 ꢁ 6.9
243.1 ꢁ 69.0
67.3 ꢁ 4.9
>500
25.8 ꢁ 3.5
>500
117.9 ꢁ 23.4
25.3 ꢁ 1.5
11.3 ꢁ 0.5
44.2 ꢁ 23.9
177.5 ꢁ 2.7
>500
4-Br-Ph
19j
19k
19l
2-MeO-Ph
3-MeO-Ph
4-MeO-Ph
14.5 ꢁ 0.8
6.5 ꢁ 3.2
19m
H
29.5 ꢁ 8.5
4.81 ꢁ 3.3
19n
19o
Ph
Ph
Me
Ph
17.5 ꢁ 0.2
13.7 ꢁ 3.8
>500
>500
a
All of the reported values represent the geometric mean of at least two
independent determinations ꢁ SD.
1518 | Med. Chem. Commun., 2013, 4, 1513–1522
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
terms of building selectivity and potency into this series of compounds, the pyrazole compounds were also highly selective
compounds, substitution at the ortho- or para-position of the for LSD1 over MAO-A.
phenyl ring with an electron withdrawing group or the intro-
Reversibility studies. To conrm the reversibility of the
duction of a methyl or phenyl group at the 5-position of the aminothiazole compound series, several representative
thiazole could therefore be used a strategies to provide a high compounds were selected on the basis of their physicochemical
level of potency together with selectivity for LSD1 over MAO-A. and ligand efficiency properties and subjected to a reversibility
Furthermore, the tolerance of these compounds towards study. Rapid dilution experiments (Fig. 5) were conducted to
substitution at the 5-position provides a handle for mitigating determine the reversibility of these inhibitors (ESI†). The
any potential toxicity issues associated with the P450-catalyzed enzyme was incubated for 20 min in the presence of an inhib-
ring opening of aminothiazoles.28
itor, which was added at a concentration 10-fold greater than
Having explored the peripheral functionality of the amino- that of its IC50 concentration. The enzyme and inhibitor were
thiazole series, we turned our attention to the thiazole core then diluted 100-fold into a mixture of the assay buffer and
itself and synthesised a small series of compounds built upon substrate to initiate the reaction, and the rate of the enzymatic
alternative cores whilst retaining some of the key structural reaction was monitored over 60 min. Compound 3, which is an
features identied above in the other SAR studies (Table 5). As irreversible TCP-based inhibitor of LSD1, was also used as a
well as simply exploring the key features of the aminothiazole control. In this particular case, as expected, the enzyme did not
core with regards to their impact on the observed activity, exhibit any discernible activity following the 100-fold dilution.
changing the core was also viewed as an opportunity to move In contrast, the compounds selected from the current amino-
away from some of the potential liabilities associated with the thiazole series all showed good evidence of reversibility, with
use of aminothiazoles in drug discovery in terms of their compounds 7, 12k and 16q demonstrating a re-gain of enzy-
metabolic stability and overall developability.28,32 These matic activity of up to 90%. Following its treatment with
compounds were synthesised from the reactions of the corre- compound 19c, the enzyme also regained its activity, albeit at a
sponding commercially available aminothiadiazole and pyr- slower rate, suggesting that this compound had a slower off-rate
azole compounds with the appropriate sulfonyl chlorides, than the other compounds tested.
using the methods described above.
The thiazole core could be readily replaced with a thiadiazole
core (20 and 21) without any signicant impact on the potency.
Furthermore, both of the thiadiazole compounds synthesised
and evaluated in the current study exhibited high levels of
selectivity for LSD1 over MAO-A, with compound 21 being >30-
fold selective for LSD1. Interestingly, the corresponding ami-
nothiazole compound 16g was only 2-fold selective for LSD1
over MAO-A, indicating that the thiadiazole core provided much
higher levels of selectivity than the thiazole. In stark contrast,
the replacement of the thiazole core for a pyrazole led to a
signicant reduction in the potency with a 6-fold loss in potency
observed for compound 23 compared with the corresponding
Fig. 5 Reversibility studies for compounds 3, 7, 12k, 16q and 19c.
thiazole 16q. In
a similar manner to the thiadiazole
Table 5 Evaluation of alternatives to the aminothiazole core
LSD1 IC50 (mM)a,b
MAO-A IC50 (mM)a
29.6 ꢁ 7.3 (18.7)
16.2 ꢁ 0.4 (17.2)
>500
37.7 ꢁ 29.2
56.9 ꢁ 1.0 (9.5)
>500
>500
>500
a
b
All of reported values represent the geometric mean of at least two independent determinations ꢁ SD. Values in parenthesis are the
corresponding aminothiazole activities.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 1513–1522 | 1519

View Article Online
MedChemComm
Concise Article
Cellular assay. During the course of the current study, we
worked in conjunction with Lynch et al.27 to develop the rst
ever reported in vitro assay for measuring LSD1 activity in cells.
Based on their transcriptome analysis of human acute myeloid
leukemia THP1 cells treated with the irreversible TCP-based
LSD1 inhibitor 3, Lynch et al.27 identied the cell surface marker
CD86 as a sensitive surrogate biomarker of LSD1 inhibition.
Within 24 h of LSD1 inhibition, a substantial dose-dependent
up-regulation of CD86 expression was reliably observed, which
could be used to screen compounds with putative LSD1 inhib-
itory activities in a cellular assay. The most potent of the ami-
nothiazole compounds synthesised in the current paper were
selected for evaluation in the CD86 cellular assay (ESI†),
including compounds 16k, 16q and 19c. Unfortunately, these
derivatives showed no evidence of activity in the cellular assay at
concentrations #50 mM (data not shown). Furthermore, these
compounds were found to be cytotoxic at concentrations greater
than 50 mM, during 24 h exposure (Data not shown).
Whilst this work was in progress, we became aware of two
novel reversible LSD1 inhibitors that were reported to have
submicromolar activities in an LSD1 biochemical assay (Table
6).33,34 The rst of these inhibitors 24 was reported by the
Huntsman Cancer Institute31 and contained a 2-hydroxy-phenyl-
hydrazone moiety. This compound was reported to be highly
active in a biochemical assay with an IC50 value of 0.04 mM.
Furthermore, this compound was reported to show activity
towards PC-3 cells in castration-resistant prostate cancer xeno-
gras. This derivative was synthesised according to the reported
procedure33 and its activity evaluated in our own biochemical
and cellular LSD1 assays. The results of these analyses are
presented in Table 6 and Fig. 6.
Fig. 6 Figure shows a representative example (n ¼ 3) of CD86 expression in
THP1 cells following 24 h of incubation with the indicated compound concen-
trations, as determined by ELISA.
this compound has also been highlighted as a common struc-
tural motif in a number of PAINS compounds.30,31 Taken
together, these results suggested that it was unlikely that the
described in vivo pharmacology of 24 was related to the inhi-
bition of LSD1 and that its reported activity in castration-
resistant prostate cancer xenogras was more likely related to
its observed cytotoxicity rather than its supposed activity
against LSD1.
Similarly, GSK354 (25)34 was recently disclosed at the 2013
AACR meeting in Washington DC (USA), where it was reported
to be a reversible inhibitor of LSD1. The synthesis and subse-
quent evaluation of the activities of this compound in our
biochemical and cellular LSD1 assays revealed that this deriv-
ative displays convincing activity in both of our assays, with IC50
values of 0.09 and 1.4 mM, (Fig. 6), respectively, representing a
16-fold drop in the potency going from the LSD1 biochemical
assay to the CD86 cellular assay. Furthermore, this compound
showed no evidence of cytotoxicity at concentrations #20 mM
and was highly selective against MAO-A (IC50 > 200 mM).
To further conrm the validity of our cellular assay as a
robust screening tool for the evaluation of LSD1 inhibitors, the
known irreversible TCP-based inhibitor 3 was also evaluated,
and gave the highest level of inhibition of the three compounds
tested, with an IC50 value of 0.02 mM (Fig. 6). Furthermore, this
These results conrmed that 24 was active against LSD1 in
our biochemical assay with an IC50 value of 0.04 mM, although
the material was essentially inactive at concentrations below
20 mM in the CD86 cellular assay (Fig. 5). In our hands, however,
compound 24 was found to be cytotoxic in THP1 cells dosed for
24 h at concentrations $2 mM, making meaningful cell activity
difficult to interpret. The 2-hydroxy-phenyl-hydrazone moiety in
Table 6 Biochemical and cellular IC50 data for the reversible LSD1 inhibitors 24 (ref. 33) and 25 (ref. 34) compounds. The irreversible TCP-based compound 3 (ref. 11)
has also been included for comparison
LSD1 Biochem. IC50 (mM)a,b
LSD1 Cellular IC50 (mM)c
Cellular toxicity (mM)
0.07 ꢁ 0.02 (0.25)11
0.02 ꢁ 0.01
>20
0.04 ꢁ 0.01 (0.04)33
0.09 ꢁ 0.01 (0.03)34
>20
$2
1.4 ꢁ 0.3
>20
>200
MAO-A Biochem. IC50 (mM)a
>37.5
>200
a
All of reported values represent the geometric mean of at least two independent determinations ꢁ S.D. b Values in parenthesis are the reported
values. c All of reported values represent the mean of at least two independent determinations ꢁ S.E.
1520 | Med. Chem. Commun., 2013, 4, 1513–1522
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
compound was non-cytotoxic below 20 mM and was also highly
selective against MAO-A (IC50 > 200 mM).
5 S. Lim, A. Janzer, A. Becker, A. Zimmer, R. Schule,
R. Buettner and J. Kirfel, Carcinogenesis, 2010, 31, 512–520.
6 S. Hayami, J. D. Kelly, H. S. Cho, M. Yoshimatsu, M. Unoki,
T. Tsunoda, H. I. Field, D. E. Neal, H. Yamaue, B. A. Ponder,
Y. Nakamura and R. Hamamoto, Int. J. Cancer, 2011, 128,
574–586.
7 M. G. Lee, C. Wynder, D. M. Schmidt, D. G. McCafferty and
R. Shiekhattar, Chem. Biol., 2006, 13, 563–567.
8 D. M. Gooden, D. M. Z. Schmidt, J. A. Polloc, A. M. Kabadi
and D. G. McCafferty, Bioorg. Med. Chem. Lett., 2008, 18,
3047–3051.
3 Conclusion
Two series of aminothiazoles were identied as reversible inhib-
itors of LSD1, through the expansion of a hit derived from a high
concentration fragment based screen of 2466 compounds. The
potency of the initial fragment hit was increased 32-fold through
synthesis and optimisation, with one series of compounds
showing clear SAR and inhibitory activities in the range of 7 to
188 mM in a biochemical assay. This series also showed selectivity
against the homologous amine oxidase enzyme MAO-A.
9 S. Mimasu, N. Umezawa, S. Sato, T. Higuchi, T. Umehara and
S. Yokoyama, Biochemistry, 2010, 49, 6494–6503.
Two recently reported reversible inhibitors of LSD1 were also 10 H. Benelkebir, C. Hodgkinson, P. J. Duriez, A. L. Hayden,
synthesized and tested in a new CD86 cellular assay. In the case
of GSK354, potent reversible enzyme inhibitory activity trans-
R. A. Bullied, S. J. Crabb, G. Pakham and A. Ganesan,
Bioorg. Med. Chem. Lett., 2011, 19, 3709–3716.
lated into potent CD86 cell activity at a non-cytotoxic dose. For 11 N. Guibourt, A. M. Ortega and J. C. P. Laria, Julio from PCT
the Huntsman compound, however, LSD1 inhibitory activity did
Int. Appl., WO 2010084160, 2010.
not translate into cell activity at a non-cytotoxic dose. The most 12 J. C. P Laria, A. M. Ortega, N. Guibourt and J. A. Baker, From
potent of our thiazolesulfonamide compounds (7.7–12 mM) PCT Int. Appl., WO 2011106105 A2 20110901, 2011.
were also evaluated in the CD86 cellular assay, but did not show 13 M. C. T. Fyfe, A. M. Ortega, J. C. P. Laria, M. E. Pedemonte,
any inhibition at less than 50 mM, which was the upper
concentration limit of the cellular assay. Given the 16-fold drop
M. A. Estiarte-Martinez and N. M. Valls, From PCT Int.
Appl., WO 2012013727 A1 20120202, 2012.
in potency observed for the reversible standard GSK354 going 14 N. W. Johnson, J. Kasparec, W. H. Miller, M. B. Rouse,
from the LSD1 biochemical assay to the CD86 cellular assay,
this is perhaps not surprising.
D. Suarez, X. Tian, From PCT Int. Appl., WO 2012135113
A2 20121004, 2012.
Importantly, this report elaborates the use of a cellular proof- 15 A. M. Ortega, M. C. T. Fyfe, M. M. Pedemonte, M. A. Estiarte-
of-mechanism assay appropriate for measuring the effect of
drug-like small molecules on the enzymatic activity of LSD1.
Martinez, N. V. Vidal, G. Kurz and J. C. C. Laria, from PCT
Int. Appl., WO 2013057322 A1 20130425, 2013.
This assay provides a platform for the further optimisation of 16 W. J. Harris, X. Huang, J. T. Lynch, G. J. Spencer,
reversible inhibitors of LSD1 and may lead to the development
of therapeutic agents with clinical utility against several
different tumour types.
J. R. Hitchin, Y. Li, F. Ciceri, J. G. Blaser, B. F. Greystoke,
A. M. Jordan, C. J. Miller, D. J. Ogilvie and
T. C. P. Somervaille, Cancer Cell, 2012, 4, 473–487.
17 D. M. Z. Schmidt and D. G. McCafferty, Biochemistry, 2007,
46, 4408–4416.
18 S. Mimasu, T. Sengoku, S. Fukuzawa, T. Umehara and
S. Yokoyama, Biochem. Biophys. Res. Commun., 2008, 366,
15–22.
Acknowledgements
This work was supported by Cancer Research UK (Grant nos.
C480/A1141 and C309/A8274). JChem for Excel was used for
structure property prediction and calculation, and general data 19 C. Binda, S. Valente, M. Romanenghi, S. Pilotto, R. Cirilli,
handling (JChem for Excel, version 5.4.0.411, 2008–2009,
ChemAxon (http://www.chemaxon.com)).
A. Karytinos, G. Ciossani, O. A. Botrugno, F. Formeris,
M. Tardugno, D. E. Edmonsdsom, S. Minucci, A. Mattevi
and A. Mai, J. Am. Chem. Soc., 2010, 132, 6827–6833.
20 Y. Huang, E. Greene, T. M. Stewart, A. C. Goodwin,
S. B. Baylin, P. M. Wooster and R. A. Casero, Proc. Natl.
Acad. Sci. U. S. A., 2007, 104, 8023–8028.
21 S. K. Sharma, Y. Wu, N. Steinbergs, M. L. Crowley,
A. S. Hanson, R. A. Casero and P. M. Woster, J. Med.
Chem., 2010, 53, 5197–5212.
Notes and references
1 J. T. Lynch, W. J. Harris and T. C. Somervaille, Expert Opin.
Ther. Targets, 2012, 16, 1239–1249.
2 Y. Shi, F. Lan, C. Matson, P. Mulligan, J. R. Whetstine,
P. A. Cole, R. A. Casero and Y. Shi, Cell, 2004, 119, 941–953.
3 E. Metzger, M. Wissmann, N. Yin, J. M. Muller, R. Schneider, 22 J. Wang, F. Lu, Q. Ren, H. Sun, Z. Xu, R. Lan, Y. Liu, D. Ward,
A. H. Peters, T. Gunther, R. Buettner and R. Schule, LSD1
demethylates repressive histone marks to promote
J. Quan, T. Ye and H. Zhang, Cancer Res., 2011, 71, 7238–
7249.
androgen–receptor–dependent transcription, Nature, 2005, 23 D. Willmann Dominica, S. Lim, S. Wetzel, E. Metzger,
437, 436–439.
A. Jandausch, W. Wilk, M. Jung, I. Forne, A. Imhof and
4 P. Kahl, L. Gullotti, L. C. Heukamp, S. Wolf, N. Friedrichs,
A. Janzer, Int. J. Cancer, 2012, 11, 2704–2709.
R. Vorreuther, G. Solleder, P. J. Bastian, J. Ellinger, 24 S. Hazeldine, B. Pachaiyappan, N. Steinbergs, S. Nowotarski,
E. Metzger, R. Schule and R. Buettner, Cancer Res., 2006,
66, 11341–11347.
A. S. Hanson, R. A. Casero Jr and P. M. Woster, J. Med. Chem.,
2012, 55, 7378–7391.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 1513–1522 | 1521

View Article Online
MedChemComm
Concise Article
25 B. Dulla, K. T. Kirla, V. Rathore, G. S. Deora, S. Kavela, 29 H. Ohara, Y. Miyabe, Y. Deyashiki, K. Matsuura and A. Hara,
S. Maddika, K. Chatti, O. Reiser, J. Iqbal and M. Pal, Org.
Biomol. Chem., 2013, 19, 3103–3107.
26 M. Tortorici, M. T. Borrello, M. Tardugno, L. R. Chiarelli,
Biochem. Pharmacol., 1995, 2, 221–227.
30 J. B. Baell and G. A. Holloway, J. Med. Chem., 2010, 53, 2719–
2740.
S. Pilotto, G. Ciossani, N. A. Vellore, S. G. Bailey, J. Cowan, 31 J. B. Baell, Future Med. Chem., 2010, 10, 1529–1546.
M. O'Connell, S. J. Crabb, G. Packham, A. Mai, R. Baron, 32 T. J. Ritchie, S. J. F. Macdonald, S. Peace, S. D. Pickett and
A. Ganesan and A. Mattevi, ACS Chem. Biol., 2013, 8, 1677–
1682.
27 J. T. Lynch, M. J. Cockerill, J. R. Hitchin, D. H. Wiseman
and T. C. P. Somervaille, Anal. Biochem., 2013, 442, 104–
106.
28 D. K. Dalvie, A. S. Kalgutkar, S. C. Khojasteh-Bakht,
R. S. Obach and J. P. O'Donnell, Chem. Res. Toxicol., 2002,
15, 269–299.
C. N. Luscombe, Med. Chem. Commun., 2012, 3, 1062.
33 H. Vankayalapati, V. Sorna, S. L. Warner, D. J. Bearss,
S. Sharma, B. StephensPct Int. Appl., WO 2013025805 A1
20130221, 2013.
34 D. Dhanak, Drugging the cancer epigenome [abstract], in
Proceedings of the 104th Annual Meeting of the American
Association for Cancer Research, AACR, Washington, DC,
Philadelphia (PA), 6–10 April 2013.
1522 | Med. Chem. Commun., 2013, 4, 1513–1522
This journal is ª The Royal Society of Chemistry 2013