Antiproliferative, DNA intercalation and redox cycling activities of dioxonaphtho[2,3-d]imidazolium analogs of YM155: A structure-activity relationship study

Article abstract of DOI:10.1016/j.ejmech.2015.09.026

The anticancer agent YM155 is widely investigated as a specific survivin suppressant. More recently, YM155 was found to induce DNA damage and this has raised doubts as to whether survivin is its primary target. In an effort to assess the contribution of DNA damage to the anticancer activity of YM155, several analogs were prepared and evaluated for antiproliferative activity on malignant cells, participation in DNA intercalation and free radical generation by redox cycling. The intact positively charged scaffold was found to be essential for antiproliferative activity and intercalation but was less critical for redox cycling where the minimal requirement was a pared down bicyclic quinone. Side chain requirements at the N¹ and N³ positions of the scaffold were more alike for redox cycling and intercalation than antiproliferative activity, underscoring yet again, the limited structural overlaps for these activities. Furthermore, antiproliferative activities were poorly correlated to DNA intercalation and redox cycling. Potent antiproliferative activity (IC₅₀ 9-23 nM), exceeding that of YM155, was found for a minimally substituted methyl analog AB7. Like YM155 and other dioxonaphthoimidazoliums, AB7 was a modest DNA intercalator but with weak redox cycling activity. Thus, the capacity of this scaffold to inflict direct DNA damage leading to cell death may not be significant and YM155 should not be routinely classified as a DNA damaging agent.

Full text of DOI:10.1016/j.ejmech.2015.09.026

European Journal of Medicinal Chemistry 104 (2015) 42e56
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Antiproliferative, DNA intercalation and redox cycling activities of
dioxonaphtho[2,3-d]imidazolium analogs of YM155: A
structureeactivity relationship study
a
b
a
a
Si-Han Sherman Ho , Mei-Yi Sim , Wei-Loong Sherman Yee , Tianming Yang ,
b
a, *
Shyi-Peng John Yuen , Mei-Lin Go
Department of Pharmacy, National University of Singapore, 18 Science Drive 4, 117543, Republic of Singapore
Department of Urology, Singapore General Hospital, 20 College Road, 169856, Republic of Singapore
a
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 May 2015
Received in revised form
The anticancer agent YM155 is widely investigated as a specific survivin suppressant. More recently,
YM155 was found to induce DNA damage and this has raised doubts as to whether survivin is its primary
target. In an effort to assess the contribution of DNA damage to the anticancer activity of YM155, several
analogs were prepared and evaluated for antiproliferative activity on malignant cells, participation in
DNA intercalation and free radical generation by redox cycling. The intact positively charged scaffold was
found to be essential for antiproliferative activity and intercalation but was less critical for redox cycling
17 August 2015
Accepted 21 September 2015
Available online 25 September 2015
1
where the minimal requirement was a pared down bicyclic quinone. Side chain requirements at the N
Keywords:
YM155
3
and N positions of the scaffold were more alike for redox cycling and intercalation than antiproliferative
activity, underscoring yet again, the limited structural overlaps for these activities. Furthermore, anti-
proliferative activities were poorly correlated to DNA intercalation and redox cycling. Potent anti-
proliferative activity (IC50 9e23 nM), exceeding that of YM155, was found for a minimally substituted
methyl analog AB7. Like YM155 and other dioxonaphthoimidazoliums, AB7 was a modest DNA inter-
calator but with weak redox cycling activity. Thus, the capacity of this scaffold to inflict direct DNA
damage leading to cell death may not be significant and YM155 should not be routinely classified as a
DNA damaging agent.
DNA intercalation
Redox cycling
Cell viability
Survivin suppressant
Apoptosis
©
2015 Elsevier Masson SAS. All rights reserved.
1
. Introduction
spread, tumor invasiveness and poor prognosis arising from che-
moresistance [9,10]. Survivin is a nodal molecule with global effects
on multiple signaling pathways in cancer cells. Thus, it is widely
regarded as an attractive therapeutic target in cancer [3,4,11]. To
date, only a few small molecules have been proposed as specific
inhibitors of survivin (Fig. 1) [12e14]. Of these, the dioxonaphtho
[2,3-d]imidazolium YM155 (sepantronium bromide) is the most
widely investigated and the only member to be advanced to clinical
trials [15e17].
Survivin is an anti-apoptotic protein that belongs to the family
of Inhibitor of Apoptosis Proteins (IAP). It is selectively expressed in
a plethora of cancers [1e7] but undetectable in most terminally
differentiated adult tissues [8]. Clinically, survivin expression is
correlated to several cancer-related events such as metastatic
YM155 blocks the transcription of the survivin gene by inter-
cepting the binding of transcription factors ILF3/NF110 and SP1 to
the survivin promoter [18,19]. It also induces dissociation of the
ILF3-p54nrb complex which positively regulates survivin expres-
sion [20]. However, there is mounting evidence that YM155-
induced effects are not solely due to the inhibition of survivin
Abbreviations: p54nrb, nuclear RNA-binding protein; ILF3/NF110, interleukin
enhancer binding factor 3/nuclear factor 110; SP1, specificity factor 1; STAT3, signal
transducer and activator of transcription 3; PI3K, phosphoinositide 3-kinase; KAP1,
KRAB-associated protein 1; MCL1, myeloid cell leukemia 1; ERK, extracellular
signal-regulated kinase; GAPDH, glyceraldehyde 3-phosphate dehydrogenase;
0 0
H DCF, 2 ,7 dichlorodihydrofluorescein; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-
2
diphenyltetrazolium bromide; SLC35F2, solute carrier family 35 member F2;
HES1, Hairy and Enhancer of split-1; PER, period circadian protein.
[21]. YM155 reduced the transcription and expression of MCL1,
another anti-apoptotic protein, in various tumors [22]. It also sup-
pressed EGFR signaling, in part by inducing the proteasomal
*
Corresponding author.
E-mail addresses: phagoml@nus.edu.sg, meilin.go@nus.edu.sg (M.-L. Go).
http://dx.doi.org/10.1016/j.ejmech.2015.09.026
223-5234/© 2015 Elsevier Masson SAS. All rights reserved.
0

S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
43
Fig. 1. Small molecules with anticancer activities reportedly linked to suppression of survivin.
degradation of EGFR [23]. The levels of other cancer related mole-
cules like PI3K, ERK, STAT3, securin were also diminished in the
presence of YM155 [23,24]. A transcriptome analysis of Wilms'
tumor cells treated with YM155 revealed down-regulation of
several other genes besides survivin as well as the up-regulation of
the pro-apoptotic genes caspase 9 and DIABLO [25]. More critically,
reports have surfaced that YM155 increased levels of the DNA
YM155 acts primarily as a DNA damage agent. Greater certainty in
the molecular target of YM155 would aid in the selection of patients
and drugs to be used with it in combination chemotherapy [31]. To
this end, a focus library of dioxonaphthoimidazolium analogs was
prepared as part of the structureeactivity relationship (SAR)
exploration of YM155. These compounds were screened for anti-
proliferative activities on two malignant cell types and evaluated by
appropriate assays for the detection of DNA intercalation and redox
cycling by reactive oxygen species.
damage markers gH2AX and phosphorylated KAP1 in malignant
cells [26,27]. This has led to the proposal that YM155 acts primarily
by inducing DNA damage and that any effect on survivin is a sec-
ondary event caused by the general deactivation of gene expression
brought about by DNA damage [26].
2. Chemistry
For an agent to induce DNA damage, it must necessarily interact
reversibly or irreversibly with DNA [27]. Irreversible binding would
involve alkylation of the DNA bases while reversible binding may
occur by chargeecharge interactions with the phosphate backbone,
interactions with the edges of base pairs in either the major or
minor groove or via insertion of the agent between base pairs of the
double helix (intercalation). Reversible binding is followed by a
cascade of events culminating in DNA damage. These events may
involve topoisomerase-induced DNA damage, inhibition of DNA
polymerase activity or the generation of free radicals that cause
DNA strand breaks. Structurally, YM155 is an impropable alkylating
agent as it lacks features associated with a reactive electrophilic
center. On the other hand, the positively charged planar scaffold of
YM155 would favor intercalation with DNA. Intercalation followed
by disruption of DNA winding by topoisomerases is unlikely as
YM155 was not found to inhibit topoisomerase I [28]. On the other
hand, there is a greater likelihood of free radical induced DNA
damage as the embedded quinone moiety in YM155 is an estab-
lished pharmacophore for redox cycling [29] and free radicals
generated during the quinone to semiquinone interconversion
could readily initiate DNA strand breaks.
Four series of YM155 analogs comprising 53 compounds were
synthesized and evaluated on clear cell renal cell carcinoma
(ccRCC) and non-small cell lung carcinoma (NSCLC) cell lines
3
0
(Fig. 3). Series A compounds were modified at the N pyrazin-2 -
1
0
ylmethyl side chain of YM155 with no change at the N 2 -
methoxyethyl side chain. Conversely, series B involved changes to
1
0
3
0
the N 2 -methoxyethyl arm while retaining the N pyrazin-2 -
ylmethyl side chain. In each case, modifications were made to
delineate the importance of the functionality that was being
altered. Hence in series A, we replaced the pyrazine ring with
benzene, cyclohexane, five-membered heteroaromatics (thio-
phene, imidazole), other azines (pyridine, pyrimidine, pyridazine),
alkyl side chains as well as varying the length of the methylene
0
linker between the ring and the scaffold. In series B, the 2 -
methoxyethyl side chain was replaced by various alkyl homologs,
aromatic rings (benzyl, pyrazinylmethyl) and hydroxyl or amino
functionalities in place of methoxy. Series C involved changes to the
dioxonaphthoimidazolium scaffold of YM155. These included
removing the positive charge from the scaffold and stepwise
reduction of the tricyclic scaffold to a single ring. Series AB incor-
porated functionalities that were associated with “optimal” anti-
proliferative activities from series A and B.
The synthesis of compounds in series A, B and AB followed the
general method described in the literature [12]. As shown in
Scheme 1, nucleophilic displacement of chlorine in 2,3-dichloro-
1,4-naphthoquinone was effected by a primary amine in the pres-
ence of triethylamine to yield compounds 4e15. The amino nitro-
gen was then acetylated to give N-acetylamides 16e27, followed by
displacement of the second chlorine by another primary amine to
give compounds 28e70. Finally, ring closure was effected by hy-
drobromic acid to give the final compounds (Scheme 1). In this way,
the compounds in series A, B and AB were successfully prepared in
yields ranging from 6 to 97%. A different route was applied for the
synthesis of A5-1. The N-acetylamide 16 was subjected to a modi-
fied BuchwaldeHartwig coupling reaction with 2-aminopyrazine
to give 71, followed by ring closure in the presence of hydrobro-
mic acid (Scheme 2).
To date, little is known of the structural features of YM155 that
are critical for DNA binding or redox cycling. In fact, the structur-
eeactivity relationship (SAR) for anticancer activity of the dioxo-
naphthoimidazolium scaffold has not been clearly defined even
though nearly 250 YM155 analogs modified at the benzene ring A
and imidazolium ring B were disclosed in the patent on YM155
(Fig. 1) [12]. Antiproliferative activities (IC50) on HelaS3 and mela-
noma A375 cells were cited to be 1 M or less for the entire series
m
but without details on assigned activities or structure-activity
trends [12]. The closest reference, aside from the patent, identi-
fied 1-ethyl-2-methylnaphth[2,3-d]imidazole-4,9-dione (1) as a
cytotoxic candidate with a mean GI50 of 2.3 mM on tumor cell lines
of the NCI panel [30]. Notably, two intermediates (2, 3) involved in
the synthesis of 1 and related compounds had comparable activ-
ities (Fig. 2). In view of the recent claims of DNA damage by YM155,
a comprehensive understanding of the key features of the scaffold
that promote DNA interaction, redox cycling and anti-cancer ac-
tivity would serve to define the contribution of the DNA damage
response to the antiproliferative activity of YM155. Common or
overlapping structural features would strengthen the notion that
Scheme 1 also applied to the series C compounds C1-1, C1-2
and C2-3 with some modifications. To obtain C1-1, propionic
anhydride was used in place of acetic anhydride for the acylation
of 4. The resulting N-propionate 72 was reacted with pyrazin-2-

4
4
S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
Fig. 2. Naphtho[2,3-d]imidazole-4,9-dione 1 and intermediates (2,3) with anticancer activity on NCI cancer cell lines.
Fig. 3. Structures of synthesized compounds. aPositively charged analogs in series C and series AB were synthesized as bromides. bHave been reported in literature [12]. The
remaining 34 analogs are novel.
ylmethylamine to afford 73 which was then cyclized to C1-1.
In the case of C1-2 and C2-3, the same sequence of reactions
were followed starting from 2,3-dichloro-5,6-dimethyl-1,4-
benzoquinone and 2,3-dichloroquinoxaline respectively, except
that the initial displacement of chlorine was effected by pyrazin-
2-ylmethanamine and not 2-methoxyethamine (Scheme 3).

S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
45
Scheme 1. Reagents and conditions: (a) R
1
-amines, EtOH, r.t., 18 h; (b) acetic anhydride, conc. H
SO
2 4
, r.t., 1.5 h; (c) R
2
-amines, toluene, 45 ꢀC, 4 h; (d) 48% HBr (aq), EtOH þ EtOAc,
ꢀ
45
C, 4 h to r.t., overnight.
Scheme 2. Reagents and conditions: (a) 2-Aminopyrazine, Pd(dba)
2
, BINAP, t-BuOK, toluene, 60 ꢀC, 5 h; (b) 48% HBr (aq), EtOH þ EtOAc, 45 ꢀC, 4 h to r.t., overnight.
Scheme 3. Reagents and conditions: (a) Pyrazin-2-ylmethylamine, triethylamine; (b) Acetic anhydride, conc. H
2 4
SO ; (c) 2-Methoxyethylamine, toluene; (d) 48% HBr (aq),
EtOH þ EtOAc, 45 ꢀC, 4 h to r.t., overnight.
The benzimidazolium C2-1 and imidazolium C2-2 were syn-
thesized by N-alkylation with 2-methoxyethyl bromide to give 77
and 78 respectively, followed by quaternization of the azomethine
nitrogen with 2-bromomethylpyrazine [32,33]. Scheme 4 outlines
the synthesis of C2-1, which also applies to C2-2.
3. Results
3.1. Cell-based antiproliferative activities
The synthesized compounds were evaluated for their anti-
proliferative effects on the clear cell renal cell carcinoma (ccRCC)
cell lines (RCC786-0, RCC4/VA), the non-small cell lung carcinoma
(NSCLC) cell lines (H1299, H1666) and the non-malignant human
lung fibroblast IMR-90 cells. RCC786-0 and RCC4/VA are deficient in
the tumor suppressor von Hippel-Lindau (VHL) gene, a genetic
aberration that is commonly encountered in ccRCC. Suppression of
the VHL gene is observed in up to 91% of sporadic ccRCC cases [34].
H1299 cells are p53-null unlike H1666 cells which are p53 positive.
The antiproliferative activities of the test compounds were
expressed in terms of IC50 which is the concentration required to
reduce cell viability to 50% of levels observed in untreated cells
under similar experimental conditions. The median IC50 values
were calculated for each cell line and found to be 113 nM (H1666),
192 nM (H1299), 262 nM (RCC786-0) and 633 nM (RCC4/VA). Dif-
ferences in the sensitivities of the different cell lines were no more
than 6-fold and ranking of compounds in terms of antiproliferative
potency was broadly consistent within each cell line. For example,
Scheme 5 outlines the syntheses of the isomeric indoles C3-
1
and C3-2. 2-Methylindole was reacted with pyrazin-2-
carbaldehyde in the presence of triethylsilane and trifluoroacetic
acid to give 82 [32] which was then N-alkylated in the presence of
KOH in DMSO to afford C3-1 [33]. In the case of C3-2, 2-
methylindole was reacted at the indole nitrogen with methyl
3
magnesium bromide, followed by electrophilic substitution at C
with 2-methoxyethyl bromide [33]. The pyrazinylmethyl side
chain was then introduced at the indole N in the usual way [32].
The synthesis of C4-1 and C4-2 was adapted from Scheme 1.
Nucleophilic displacement of chlorine in 2,3-dichloro-1,4-
naphthoquinone was effected by concentrated ammonia to
give 84, followed by acetylation by acetic anhydride to give
8
5. Nucleophilic displacement of the second chlorine by 2-
methoxyethylamine or pyrazin-2-ylmethylamine gave 86 and 87
respectively. Finally, ring closure using hydrobromic acid yielded
C4-1 and C4-2 respectively (Scheme 6).

46
S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
ꢀ
3
Scheme 4. Reagents and conditions: (a) 2-methoxyethyl bromide, KOH, DMSO, r.t., 1.5 h; (b) 2-bromomethylpyrazine, CH CN, 80 C, 24 h.
ꢀ
Scheme 5. Reagents and conditions: (a) Pyrazin-2-carbaldehyde, Et
0
3
SiH, TFA, DCM, 0 C, 1 h; (b) 2-Methoxyethyl bromide, KOH, DMSO, r.t., 1.5 h; (c) MeMgBr, dry toluene, DCM,
ꢀ
ꢀ
ꢀ
C, 30 min; (d) 2-Methoxyethyl bromide, 0 C / 40 C, 24 h; (d) 2-Bromomethylpyrazine, KOH, DMSO, r.t., 30 min.
ꢀ
Scheme 6. Reagents and conditions: (a) NH
3
in MeOH, EtOH, 35 C, 3 h; (b) Acetic anhydride, conc. H
SO
2 4
, r.t., 1.5 h; (c) 2-Methoxyethylamine/pyrazin-2-ylmethylamine, trie-
thylamine, toluene þ EtOH, 45 ꢀC, 1 h; (d) 48% HBr (aq), 2 drops of 2 M NaOH(aq), EtOH, 50 C, 1 h.
ꢀ
A5-1 had the weakest antiproliferative activity on all cells while the
most potent compounds were either A3-3 or AB7. This trend was
The importance of the quinone moiety was again emphasized by
the negligible activity of the quinoxalinoimidazolium analog C2-3
also reflected in the Spearman correlation coefficients (
r
) derived
(IC50 > 100 mM). Loss of the permanent positive charge on the
from pair-wise comparisons of IC50 values from the four cell lines.
r
scaffold as seen in the dioxonaphthoimidazoles (C4-1, C4-2) and
naphthoquinone which is YM155 without the positively charged
imidazolium ring, likewise reduced IC50 but to a lesser degree (IC50
values exceeded 0.85, indicating a strong monotonic relationship
between the IC50 values (Supporting Information Table S1). Hence
the SAR deduced from one cell line should largely apply to other cell
lines. In the following paragraphs, SAR is discussed with reference
to H1666 cells. Table 1 lists the IC50 values derived from the highly
susceptible H1666 cells. Values for the other cell lines are given in
Supporting Information (Table S2).
We first considered the series C compounds to deduce the
importance of the dioxonaphthoimidazolium scaffold. YM155 had
an IC50 of 13.7 nM on H1666 cells and removal of the benzene ring
to give the dioxodihydrobenzoimidazolium C1-2 (IC50 584 nM)
resulted in a modest loss of activity. On the other hand, omitting the
quinone moiety to give the benzoimidazolium C2-1 completely
10e20 mM) as compared to the omission of the quinone. Broadly,
these results support the view that loss of the quinone from the
tricyclic scaffold was the most detrimental, followed by omission of
the positive charged imidazolium (C4-1, C4-2, naphthoquinone)
while the least disruptive modification was the removal of the
distal benzene ring (C1-2).
Next, we assessed the importance of the three substituents on
the dioxonaphthoimidazolium scaffold of YM155. These were the
0
1
2
0
3
2 -methoxyethyl at N , methyl at C and pyrazin-2 -ylmethyl at N .
2
Only one modification was made to the methyl at C , which was
extension to its ethyl homolog. The resulting compound (C1-1, IC50
96 nM) was less active indicating a preference for a sterically small
group at this position.
abolished activity (IC50 > 100
mM) and the same was true when the
scaffold was trimmed down to a solitary imidazolium ring (C2-2).

S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
47
Table 1
with non-azinylmethyl rings like benzyl (A1-1, IC50 240 nM),
Antiproliferative IC50 values of YM155 and synthesized compounds on non-small
cell lung cancer H1666 and non-malignant IMR-90 cells.
0
cyclohexylmethyl (A1-2, 415 nM), thien-2 -ylmethyl (A1-3, IC50
0
2
07 nM) or imidazol-2 -ylmethyl (A2-1, IC50 279 nM) adversely
Cpd
Antiproliferative IC50
,
m
Ma
IMR-90
affected antiproliferative activity. Third, substituting pyrazine with
other azines (pyridine, pyrimidine and pyridazine) was generally
well tolerated but for selected regioisomers only. There was a
noticeable bias against azines with only ortho-substituted azome-
thine nitrogens. Thus, among the pyridinylmethyl analogs, the
H1666
Selectivity ratiob
YM155
A1-1
A1-2
A1-3
A2-1
A3-1
A3-2
A3-3
A3-4
A3-5
A3-6
A3-7
A3-8
A4-1
A5-1
A6-1
A6-2
A6-3
A6-4
A6-5
A6-6
A6-7
A6-8
B1-1
B1-2
B1-3
B2-1
B2-2
B2-3
B2-4
B2-5
B2-6
B2-7
B2-8
B3-1
B3-2
B4-1
B4-2
C1-1
C1-2
C2-1
C2-2
C2-3
C3-1
C3-2
C4-1
C4-2
AB1
0.0137 ± 0.0009
0.240 ± 0.020
0.415 ± 0.029
0.207 ± 0.017
0.279 ± 0.027
0.231 ± 0.033
0.0404 ± 0.0052
0.0110 ± 0.0007
0.152 ± 0.003
0.00770 ± 0.00070
0.0287 ± 0.0027
0.0199 ± 0.0021
0.157 ± 0.003
0.859 ± 0.090
8.78 ± 0.09
0.0728 ± 0.0056
0.0288 ± 0.0054
0.073 ± 0.0035
0.113 ± 0.018
0.224 ± 0.02
0.0559 ± 0.0111
0.15 ± 0.027
0.278 ± 0.056
0.118 ± 0.017
0.758 ± 0.121
0.0377 ± 0.0045
0.0415 ± 0.0081
0.0349 ± 0.0020
0.0405 ± 0.0033
0.152 ± 0.009
0.0327 ± 0.0012
0.0546 ± 0.0034
0.458 ± 0.025
1.47 ± 0.09
0.247 ± 0.037
1.16 ± 0.07
5.27 ± 0.96
1.45 ± 0.05
2.01 ± 0.11
2.52 ± 0.11
0.611 ± 0.044
0.176 ± 0.009
3.18 ± 0.04
0.205 ± 0.014
0.276 ± 0.027
0.187 ± 0.012
2.00 ± 0.13
2.81 ± 0.52
8.82 ± 1.21
0.743 ± 0.089
0.345 ± 0.038
1.23 ± 0.13
1.72 ± 0.083
3.89 ± 0.29
1.16 ± 0.14
2.64 ± 0.54
6.79 ± 0.29
2.48 ± 0.32
9.89 ± 0.17
1.43 ± 0.26
0.582 ± 0.089
0.480 ± 0.022
0.565 ± 0.027
1.42 ± 0.15
0.213 ± 0.020
1.62 ± 0.14
7.97 ± 0.180
6.58 ± 0.18
5.61 ± 0.38
3.57 ± 0.71
7.42 ± 0.76
3.73 ± 0.11
1.90 ± 0.11
1.36 ± 0.15
>100
18
4.8
13
7.0
7.2
11
15
16
21
27
9.6
9.4
13
3.3
1.0
10
12
17
15
17
21
18
24
21
13
38
14
14
14
9.3
6.5
30
17
4.5
4.1
7.9
5.2
2.1
20
2.3
0
ortho substituted analog (A3-1, pyridin-2 -ylmethyl) fared poorly
compared to its meta (A3-2) and para (A3-3) counterparts. This was
also true for the pyrimidinylmethyl analogs where the exclusively
ortho substituted A3-4 consistently underperformed when
compared to its regioisomers A3-5 (ortho and para Ns) and A3-6
(two meta Ns). The apparent preference for the meta- or para
regioisomer may reflect the optimal location of the azomethine N
for hydrogen bonding. We noted that inserting a methyl at the para
position of the pyrazine ring of YM155 reduced activity by 10-fold
(A3-8, IC50 157 nM) which implied that in the absence of a para-
azomethine N, it is best to keep that position unoccupied. Even
then, the preferred regioisomers (A3-3, A3-5, A3-7) were only
equivalent or modestly exceeded YM155 in terms of anti-
proliferative activity. Lastly, replacing the pyrazinylmethyl side
chain with alkyl side chains (A6-1 to A6-8) was an acceptable
modification. Seven alkyl side chains were investigated and of
these, methyl was the most active on H1666 cells, followed by
cyclopropyl, ethyl, propyl, n-butyl, isopropyl and isobutyl in that
order. The same sequence of methyl > ethyl > propyl was observed
on the other cell lines, corroborating a preference for short alkyl
side chains that were not branched or cyclized. Even so, the highly
ranked methyl analog A6-2 (IC50 29 nM) was at best comparable to
YM155 and the azinylmethyl analogs (A3-3, A3-5, A3-7).
The series B compounds were designed to report on the effects
0
1
of modifying the 2 -methoxyethyl side chain at N while main-
0
2
3
taining methyl and pyrazin-2 -ylmethyl at C and N . Replacing
the methoxy with hydroxy resulted in an 8-fold loss in activity
whereas homologation to ethoxy caused a smaller 3-fold loss.
Embedding the ether functionality in a ring (tetrahydrofurfuryl,
B1-2) or replacing it with an N,N-disubstituted amine (B4-1, B4-2)
pushed IC50 values to the micromolar range. On the other hand,
replacing methoxyethyl with alkyl side chains afforded com-
pounds with surprisingly good activity. The side chains investi-
gated were similar to those in series A (methyl, ethyl, propyl, butyl
isopropyl, isobutyl, cyclopropyl). As before, short unbranched side
chains were best but with a different rank order of activity
(cyclopropyl > ethyl > propyl > methyl > isopropyl > isobutyl)
compared to series A. The preeminent position of the cyclopropyl
side chain (B2-5) was evident in all cells except RCC786-0, where
the methyl and ethyl analogs were more active. Interestingly, we
noted that B2-8 bearing the cyclopentyl side chain fared poorly
compared to its potent cyclopropyl counterpart B2-5. Thus, the
more critical requirement appears to be that of a sterically small
1.37 ± 0.09
0.450 ± 0.052
1.44 ± 0.20
1.81 ± 0.15
0.0960 ± 0.0127
0.584 ± 0.097
>100
>100
>100
>100
>100
22.0 ± 1.5
17.9 ± 2.0
0.0168 ± 0.0017
0.0298 ± 0.0032
0.0862 ± 0.0070
0.0613 ± 0.0054
0.0533 ± 0.0041
0.0355 ± 0.0049
0.00860 ± 0.00030
c
N/A
>100
>100
>100
>100
55.4 ± 4.6
73.0 ± 4.5
0.287 ± 0.059
0.310 ± 0.051
0.649 ± 0.079
0.440 ± 0.007
0.484 ± 0.064
0.475 ± 0.091
0.159 ± 0.016
13.3 ± 2.0
N/Ac
N/Ac
c
N/A
N/Ac
2.5
4.1
17
AB2
AB3
AB4
AB5
AB6
AB7
10
7.5
7.2
9.1
13
18
1.4
9.78 ± 0.56
1
group at N , regardless of whether it is cyclized or not.
We also explored the feasibility of replacing methoxyethyl with
0
aromatic ring structures, namely benzyl and pyrazin-2 -ylmethyl.
a
ꢀ
Evaluated by MTT assay, 72 h incubation, 37 C, 5% CO
2
. Mean ± SD for n ¼ 3 in-
The resulting compounds (B3-1, B3-2) had significantly diminished
dependent determinations.
activities. B3-2 is noteworthy as it is symmetrically substituted
b
Mean IC50 IMR-90/IC50 H1666
.
with pyrazinylmethyl moieties at both N1 and N3 and is thus
c
Not applicable.
Naphthoquinone.
d
comparable to A6-1 which is symmetrically substituted at the same
positions with 2 -methoxyethyl moieties. That A6-1 (IC50 73 nM)
0
was more potent than B3-2 (IC50 450 nM) further underscores the
poor tolerance for bulky substituents on the scaffold. Taken
together, series B was less promising than series A with only one
member, (the cyclopropyl analog B2-5) that was marginally more
potent than YM155, and only on H1299 cells.
The series A compounds were designed to probe the importance
0
3
of the pyrazin-2 -ylmethyl side chain at N . The following obser-
0
vations apply. First, extending the side chain to pyrazin-2 -ylethyl
0
(
8
A4-1, IC50 86 nM) or truncating it to pyrazin-2 -yl (A5-1, IC50
0
3
0
1
.8
m
M) reduced activity. Second, replacing pyrazin-2 -ylmethyl
Having identified N -pyridin-4 -ylmethyl and N -cyclopropyl as

48
S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
the “optimal” substituents at their respective positions, these
functionalities were then introduced to the scaffold with the
expectation that it would yield more potent compounds. Less
optimal substituents (pyridin-3 -ylmethyl, pyrimidin-5 -ylmethyl,
methyl, ethyl) were concurrently investigated to validate this
approach. In all, seven analogs were synthesized, of which six were
diminished affinities (DC50 32e99
pounds bearing ring structures at N and N should be strong
binders was indeed found to apply as seen from B1-2 (13 M), B3-1
(14 M) and B3-2 (12 M). Interestingly, series AB consistently
yielded good DNA intercalators (DC50 12e23 M). It is noted that
mM). The corollary that com-
1
3
m
0
0
m
m
m
the intercalating ability of AB1eAB6 were closely aligned to that of
the corresponding series A analogs. Thus, AB1 and AB2 which have
0
0
derived from permutations of pyridin-4 -ylmethyl, pyridin-3 -
0
3
0
ylmethyl or pyrimidin-5 -ylmethyl at N and ethyl or cyclopropyl at
N . The last compound AB7 was substituted at both N and N with
methyl. Good activity was found in analogs that have pyridin-4 -
ylmethyl at N and cyclopropyl (AB1) or ethyl (AB2) at N . AB1 was
broadly equivalent to YM155 in terms of IC50 while AB2 was
marginally less active than YM155. Clearly, this approach of
combining optimal substituents onto the scaffold did not work as
well as anticipated. Surprisingly, the minimally substituted AB7
displayed good antiproliferative activity (IC50 8.6 nM versus
in common a pyridin-4 -ylmethyl side chain had DC50 values that
1
1
3
3
were within the range of A3-3 which has a similar side chain at N .
0
This was again observed for AB5, AB6 (compared to A3-2) and AB3
(compared to A3-6). AB4 was an outlier, being more potent than
3
1
1
3
A3-6. The strong binding affinity of the N , N -dimethyl substituted
AB7 (DC50 13
(B2-1, N -methyl, DC50 31
m
M) was unusual because the mono-methyl analogs,
1
3
m
M; A6-2, N -methyl, DC50 32
m
M) were
significantly weaker intercalators. Third, even for the strong
binders mentioned earlier, DNA binding affinities could not match
that of doxorubicin. Lastly, when the correlation between anti-
proliferative IC50 and DC50 values were examined in a Spearman
correlation matrix (Supporting Information Table S3), the magni-
13.7 nM for YM155) and is the most promising compound identi-
fied in this investigation.
To determine if YM155 and the test compounds selectively
affected proliferation of malignant cells, we determined their ef-
fects on the viability of non-malignant lung fibroblast IMR-90 cells.
Selectivity was assessed from the ratio of IC50 values on IMR-90
versus the malignant cell line. Values ratios ranging from 1 to 38
were obtained for H1666 (Table 1). Higher selectivities (21e38 fold)
were found for compounds that had relatively weaker growth
inhibitory activities (A6-6, A6-8, B1-3, B2-6) whereas YM155 and
its most potent analogs (A3-3, A3-5, AB1, AB7) had more modest
selectivities ranging from 16 to 27-fold. Selectivity ratios for the
other cell lines are found in Table S1.
tude of the
r values (0.25e0.34) did not point to a convincing
relationship between these two activities.
3.3. Redox cycling and generation of reactive oxygen species
The quinone moiety is known to accept a single electron to yield
the semiquinone radical which can react directly with DNA [37] or
participate in a redox cycle of superoxide radical generation by
transfer of electrons to oxygen [38]. Superoxide and its dismutation
product hydrogen peroxide form hydroxyl radicals, which if
generated in the vicinity of DNA can inflict substantial damage [39].
Thus it was of interest to determine if the quinone-bearing dioxo-
naphthoimidazoliums would predispose the scaffold to free radical
generation and redox cycling. Two assay methods were employed
e an in vitro assay centered on the oxidation of phenol red by
hydrogen peroxide [40] and a cell based method for the detection of
3.2. Intercalation with DNA
Next, we assessed the DNA binding affinity of the test com-
pounds by monitoring the displacement of thiazole orange from
herring sperm DNA [35,36]. Briefly, thiazole orange intercalates
with non-specific DNA sequences in herring sperm DNA with
emission of fluorescence. When it is displaced by a competing
intercalator, fluorescence is diminished. Here, each compound was
tested over a range of concentrations to determine the concentra-
tion (DC50) at which it reduced the basal fluorescence of thiazole
0
0
free radical formation via oxidation of 2 , 7 dichlorodihydro-
fluorescein (H DCF).
2
In the in vitro method, the oxidation of dithiothreitol provides
electrons required to reduce the quinone to semiquinone or hy-
droquinone. This reaction is reversed in the presence of oxygen,
generating superoxide anions which dismutate to hydrogen
peroxide and oxidize phenol red in the presence of horseradish
peroxidase. In the presence of a redox cycler, phenol red is
continually oxidized, causing its absorbance at 610 nm to increase.
The redox cycling propensity of the test compound was determined
by the concentration (EC50) required to increase the absorbance of
oxidized phenol red at 610 nm to half its maximum value (Table 2).
A compound predisposed to facile redox cycling will have a lower
EC50 compared to a weak or non-redox cycler. Representative dose
response plots are shown in Supporting Information (Figure S2).
YM155 and its analogs (except for several members in series C)
were found to be redox cyclers with EC50 values in the range of
orange (10 mM) by 50%. Strong intercalators would have small DC50
values and vice versa. Typical dose response curves depicting the
displacement of thiazole orange by competing test compounds are
provided in Supporting Information (Figure S1).
The DC50 of YM155 was 20
positive control doxorubicin (Table 2). Since thiazole orange was
evaluated at a fixed concentration of 10 M, this would imply that
mM as compared to 2.6 mM for the
m
YM155 had half the binding affinity of thiazole orange for DNA as
compared to doxorubicin which had 4 times the DNA binding af-
finity of thiazole orange. Clearly, the DNA binding affinity of YM155
was significantly weaker than that of doxorubicin.
Some interesting trends in DNA binding affinities were deduced
from the DC50 values in Table 2. First, an intact dioxonaph-
thoimidazolium scaffold was required. This was clearly seen from
0.7
potent than most of its analogs but compared to a known redox
cycler like naphthoquinone (EC50 0.33 M), it was clearly less
mMe13 mM. In this context, YM155 (EC50 1.4 mM) was more
the poor binding affinities (DC50 > 100
m
M) of series C compounds
m
that lack the tricyclic scaffold in its entirety. Omission of the
quinone (series C compounds except C1-1) or the positive charge
effective. In terms of the SAR for redox cycling, the series C com-
pounds provided valuable insight. Only two compounds in series C
were redox cyclers, namely C1-1 and C1-2. None of the non-
quinone bearing series C analogs were redox cyclers but strik-
ingly, some quinone-containing analogs (C4-1, C4-2) also failed to
(
(
C4-1, C4-2) from the scaffold significantly reduced binding affinity
DC50 > 100 M). Naphthoquinone, a bicyclic quinone, was also a
m
weak intercalator. Thus an intact scaffold was critical for DNA
intercalation.
3
redox cycle. Since C4-1 and C4-2 are not positively charged at N ,
Second, for compounds with intact scaffolds, the absence of ring
this would imply that the charged imidazolium and quinone are
both necessary for redox cycling. If so, the outstanding activity of
naphthoquinone (YM155 without imidazolium) is unexpected.
Thus, pending additional data, a reasonable conclusion would be
1
3
structures at N and N adversely affected binding. Thus we noted
that the series A compounds (A6-1 to A6-8) in which the pyrazin-2-
ylmethyl side chain was replaced by alkyl side chains had

S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
49
Table 2
DNA intercalation DC50 and redox cycling EC50 values of test compounds.
Compound
DC50
(
m
M)a for DNA intercalation
EC50 (m
M)b for redox cycling
YM155
A1-1
A1-2
A1-3
A2-1
A3-1
A3-2
A3-3
A3-4
A3-5
A3-6
A3-7
A3-8
A4-1
A5-1
A6-1
A6-2
A6-3
A6-4
A6-5
A6-6
A6-7
A6-8
B1-1
B1-2
B1-3
B2-1
B2-2
B2-3
B2-4
B2-5
B2-6
B2-7
B2-8
B3-1
B3-2
B4-1
B4-2
C1-1
C1-2
C2-1
C2-2
C2-3
C3-1
C3-2
C4-1
C4-2
AB1
20.3 ± 3.1
27.5 ± 0.5
35.4 ± 1.4
28.6 ± 0.4
31.9 ± 1.1
28.6 ± 3.1
16.4 ± 1.3
20.0 ± 0.5
45.0 ± 3.3
25.4 ± 2.4
26.7 ± 3.4
23.0 ± 1.2
18.4 ± 1.2
44.0 ± 3.2
95.4 ± 7.5
32.3 ± 3.8
32.3 ± 1.5
40.1 ± 4.5
48.2 ± 6.0
99.0 ± 0.8
35.4 ± 4.1
44.2 ± 0.2
67.0 ± 5.8
17.7 ± 1.2
13.3 ± 0.9
20.9 ± 1.4
30.6 ± 0.8
24.0 ± 2.4
30.1 ± 1.5
56.9 ± 9.1
25.2 ± 1.6
27.9 ± 3.8
43.3 ± 2.3
78.8 ± 13.2
13.9 ± 0.7
12.0 ± 1.0
19.9 ± 4.0
15.5 ± 0.3
29.5 ± 2.6
>100
1.38 ± 0.05
3.63 ± 0.71
13.0 ± 0.3
3.22 ± 0.14
6.27 ± 0.26
2.10 ± 0.24
1.91 ± 0.24
2.01 ± 0.10
1.36 ± 0.11
4.56 ± 0.13
3.09 ± 0.08
2.37 ± 0.01
2.57 ± 0.16
7.58 ± 0.20
2.96 ± 0.15
9.30 ± 0.43
5.70 ± 0.43
7.08 ± 1.18
8.47 ± 0.15
10.2 ± 0.5
6.37 ± 0.69
6.37 ± 0.69
8.75 ± 0.43
2.36 ± 0.04
3.67 ± 0.20
0.68 ± 0.00
2.94 ± 0.05
3.32 ± 0.13
2.26 ± 0.05
3.47 ± 0.45
3.92 ± 0.10
2.31 ± 0.37
5.82 ± 0.05
7.35 ± 0.06
1.85 ± 0.10
1.34 ± 0.06
1.43 ± 0.06
1.01 ± 0.04
2.00 ± 0.04
1.42 ± 0.01
>400
>100
>100
>100
>100
>100
>100
>100
20.8 ± 1.3
21.3 ± 3.2
22.7 ± 1.6
17.6 ± 1.2
14.7 ± 1.5
12.0 ± 0.3
12.7 ± 0.5
2.64 ± 0.16
>100
>400
>400
>400
>400
>400
>400
6.10 ± 0.38
3.47 ± 0.08
7.61 ± 0.07
3.86 ± 0.11
3.48 ± 0.01
2.75 ± 0.18
7.55 ± 0.71
Not determined
0.33 ± 0.01
AB2
AB3
AB4
AB5
AB6
AB7
Doxorubicin
Naphthoquinone
a
DC50: Concentration required to reduce basal fluorescence of thiazole orange (10
m
M) by 50%.
EC50: Concentration required to reduce oxidized phenol red absorbance to 50% of control values. Mean and SD of n ¼ 3 separate de-
terminations for both experiments. None of the compounds were fluorescent under test conditions.
b
that the quinone moiety per se is not sufficient to ensure redox
cycling.
cyclers. Conversely, potent redox cyclers like B1-3 and naph-
thoquinone had only modest antiproliferative activities. Indeed,
when the correlation between these activities was examined in a
Spearman correlation matrix (Supporting Information Table S3),
Another SAR pertains to the requirement of at least one ring-
1
3
containing substituent at N or N for activity. Analogs with aro-
1
3
matic ring substituents on both imidazole nitrogens (N , N ) were
generally better redox cyclers (B3-1, B3-2) than those without this
motif (AB7, A6-1 to A6-8). Anecdotal comparisons of anti-
proliferative IC50 and redox cycling EC50 values revealed a
disparity between these activities among several analogs. For
example AB1, AB2 and A6-2 which were equipotent to YM155 in
terms of antiproliferative activities were less competent redox
the
relationship.
r
values (ꢁ0.002 to ꢁ0.129) did not support a meaningful
Next, we determined if YM155 and selected analogs generated
free radicals in a cellular environment. To this end, we pretreated
0 0
mM 2 , 7 -dichlorodihydro-
RCC786-0 and H1299 cells with 20
fluorescein (H DCF) diacetate and exposed the loaded cells (2 h) to
various concentrations of test compounds. Briefly, the diacetate is
2

50
S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
hydrolyzed by cellular membrane esterases to give the cell per-
meant, non-fluorescent H DCF which is oxidized intracellularly by
reactive oxygen species (ROS) to the fluorescent product dichlor-
ofluorescein. Hence, an increase in fluorescence would be
observed if the test compound generated ROS by redox cycling or
other means. Representative compounds (YM155, A3-3, B1-3, C1-
fluorescence matching that of naphthoquinone at 5ꢂ IC50
(RCC786-0) and exceeding it at 2ꢂ IC50 (H1299). We noted that
C1-2, like naphthoquinone, is a bicyclic quinone and this struc-
tural feature which is not found in the other YM155 analogs, could
have accounted for its unusual potency. Taken together, both
in vitro and cell based assays confirmed the significantly weaker
redox cycling activities of the dioxonaphthoquinone analogs as
compared to naphthoquinone.
2
2
, AB1, AB7) that have with a range of redox cycling activities were
selected for investigation. B1-3 is a strong redox cycler (EC50
.68 M), YM155, C1-2, A3-3 have modest activities (EC50
.4e2 M) and AB1, AB7 are weaker cyclers (EC50 6.1, 7.6 M).
0
1
m
m
m
3.4. Formation of gH2AX in treated cells
Naphthoquinone was included as positive control. The com-
pounds were tested at concentrations equivalent to 1ꢂ, 2ꢂ, 5ꢂ
and 10ꢂ antiproliferative IC50 (Table S1). As seen from Fig. 4, the
compounds had strikingly similar effects on both cell lines. In each
case, YM155, A3-3, B1-3, AB1 and AB7 did not alter the basal level
of fluorescence over the entire concentration range, indicating
that under these conditions, these compounds did not generate
ROS. In contrast C1-2 demonstrated a dose dependent increase in
YM155 was reported to intercalate with DNA and to induce a
DNA damage response which is purportedly central to its anti-
cancer properties [26,28]. In the presence of YM155, prostate can-
cer cells PC3 showed dose dependent increases in two biomarkers
of DNA damage, phosphorylated histone gH2AX and phosphory-
lated KAP1. These increases were observed at concentrations that
did not reduce survivin, thus implicating DNA damage, rather than
7
RCC786-0
6
5
4
3
2
1
0
NQ
YM155
A3-3
B1-3
C1-2
AB1
AB7
Vehicle ctrl
1x IC50
2x IC50
5x IC50
10x IC50
5
4
3
2
1
0
H1299
NQ
YM155
A3-3
B1-3
C1-2
AB1
AB7
Vehicle ctrl
1x IC50
2x IC50
5x IC50
10x IC50
Fig. 4. Generation of free radicals in RCC786-0 and H1299 cells at different concentrations of test compounds as determined by the fold change in H
change of fluorescence over vehicle control at each concentration is displayed as the mean with SD of n ¼ 3 determinations.
2
DCF fluorescence. The fold

S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
51
suppression of survivin, as the initiating event. We have shown that
YM155 and several analogs had the capacity to intercalate with
DNA but with weaker affinities compared to doxorubicin. To
determine if intercalation is followed by DNA damage, we moni-
observed at the same time or after declines in survivin. In YM155
and AB1-treated cells, increases in gH2AX were also concentra-
tion dependent.
tored levels of
g
H2AX (a marker for double strand breaks in DNA)
3.5. YM155 and selected analogs induced apoptosis in RCC786-
0 cells
[41,42] in RCC786-0 and H1299 cells treated with YM155, AB1 and
AB7. In terms of IC50, AB1 was more potent than YM155 (36 nM
versus 54 nM) but comparable to it on H1299 (36 nM for both). AB7
was more potent than YM155 on both cell lines. gH2AX levels also
increase in response to DNA fragmentation which is a sequel of
apoptosis [43,44]. As such it is necessary to establish if the elevated
A multiplex assay (ApoTox-Glo™) was used to simultaneously
monitor cell viability, cytotoxicity and apoptotic events in RCC786-
0 cells that were exposed (48 h) to varying concentrations of
YM155, A4-1, B2-4 and AB1 (Supporting Information, Figure S4).
The test compounds displayed similar trends in their concentration
dependent effects on viability, cytotoxicity and apoptosis. Cell
viability decreased with increasing concentrations while cytotox-
icity remained at basal levels except at the highest concentrations
of YM155 and AB1. Apoptotic events were monitored by the acti-
vation of caspases 3 and 7 and here, activation was observed to
increase in tandem with the decline in cell viability. These results
showed that loss in cell viability was largely due to apoptosis with
negligible contribution from necrotic cell death.
gH2AX levels in treated cells is a consequence of DNA damage or
DNA fragmentation due to apoptosis. In general, DNA damage is
initiated more rapidly than apoptotic cell death. Thus, time-based
experiments in which levels of an apoptotic marker cleaved cas-
pase 3 and
gH2AX from treated cells were concurrently monitored.
Fig. 5 shows the levels of
gH2AX, cleaved caspase 3 and survivin
in RCC786-0 (panels A, B) and H1299 cells (panels C, D) treated for
various time periods with YM155, AB1 and AB7 at their IC50 con-
centrations. The appearance of
cell lines. In RCC786-0, H2AX was detected at 24 h but only in
YM155-treated cells. In the case of AB1 and AB7, H2AX appeared
later at 48 h. Cleaved caspase 3 levels in treated RCC786-0 cells
were found to increase in tandem with H2AX. In H1299 cells,
H2AX was detected at only 48 h (AB1 and AB7) and 72 h (YM155).
We noted again that increases in cleaved caspase 3 and H2AX
were synchronized. Survivin levels were also tracked and found to
gH2AX was time dependent on both
g
g
3.6. Transcriptome analysis of YM155-treated RCC786-0 cells
g
To further determine if the DNA damage response was critical to
the mode of action of YM155, we compared the transcriptional
profile of RCC768-0 cells that were treated with YM155 at its IC50
(40 nM) for 24 h with control untreated cells. Genes that were
significantly altered in the treated cells by 2-fold or more (adjusted
p-value < 0.05) were identified and subjected to pathway analysis.
Several canonical signaling pathways were significantly affected by
YM155, including the p53 and ErbB signaling pathways (Supporting
Information, Figure S5). We further mined the data for genes that
were involved in DNA repair. Of the 257 DNA repair genes that were
curated, we found significant up-regulation of HES1 (2.24-fold) and
PER (2.15-fold). Notably, these two genes were not involved in the
repair of double-stranded breaks. Genes involved in double-
stranded break repair, such as MRE11 (-1.64-fold) and RAD50
(1.26-fold) were minimally up- or down-regulated. Indeed, an
g
g
decrease before
time (H1299).
gH2AX was detected (RCC786-0) or at the same
Next we investigated the effects of varying the concentrations
0.5ꢂ, 1ꢂ, 2ꢂ IC50) of YM155 and AB1 on the expression of H2AX
in the same cell lines at a fixed time of 48 h. Dose dependent in-
creases in H2AX were observed for both compounds (Supporting
(
g
g
Information, Figure S3). These increases were aligned to increases
in cleaved caspase 3 and decreases in survivin levels. Thus,
YM155, AB1 and AB7 induced time-dependent increases in
g
H2AX levels in both RCC786-0 and H1299 cells. Increases coin-
cided with the elevation in levels of cleaved caspase 3 and were
Fig. 5. YM155, AB1 and AB7 induced time dependent increases in
IC50 concentrations for 6, 24, 48 and 72 h, after which H2AX, survivin and cleaved caspase 3 levels (p17, p19) were probed by immunoblotting. The p19 fragment is the precursor of
p17. GADPH was used as loading control. Molecular weights are shown on the right hand side of the blots.
gH2AX in RCC786-0 (panels A, B) and H1299 cells (panels C, D). Cells were incubated with test compounds at their
g

52
S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
overwhelming majority of the genes involved in DNA repair
were not strongly perturbed by YM155 treatment (Supporting
Information, Table S4).
However, these increases were not entirely compatible with the
DNA damage response for several reasons. First, they were delayed
in terms of detection in treated cells and second, they were closely
aligned to the appearance of the apoptotic marker protein cleaved
4
. Discussion
caspase 3. This raises the possibility that the elevation of gH2AX is
due to DNA fragmentation which is a sequel of apoptotic cell death
[43,47e49]. Further insight was gained from the transcriptional
profile of RCC786-0 cells treated with YM155. Analysis of the DNA
repair genes revealed significant up-regulation of only a limited
number of genes involved in DNA repair (HES1, PER) while the
majority were either down-regulated or unaltered. Hence, DNA
damage is not as pervasive as anticipated. We noted that others
have also found DNA damage in only a fraction of cells treated with
YM155 in comet assays [28]. As anticipated, downregulation of
survivin was observed in cells treated with YM155, AB1 and AB7.
The decreases in survivin were closely aligned to the appearance of
YM155 was identified as a specific repressor of survivin pro-
moter activity in an artificial reporter system containing the sur-
vivin promoter [45]. Ensuing investigations showed that tumor
regressions induced by YM155 in xenograft models were associated
with diminished intratumoral survivin expression levels [46].
Despite its outstanding potency in preclinical studies, YM155 per-
formed disappointingly in clinical trials [21]. Recent findings have
shown that the uptake of YM155 into cells was dependent on the
presence of a specific solute carrier SLC35F2 and it is conceivable
that varying levels of this transporter on the outer membrane
surfaces of tumoral cells could have moderated the cytotoxicity of
YM155 [28]. Various reports have since questioned if the sup-
pression of survivin is the sole and main effect of YM155-induced
cytotoxicity [21,26,28]. The consensus is that YM155 is a DNA
intercalating agent and induces a DNA damage response that leads
to apoptotic cell death. In this report, we adopted an SAR approach
to assess the contribution of DNA intercalation and redox cycling
which are incipient events in the DNA damage response, to the
antiproliferative activities of YM155 and its analogs. The premise
was that overlapping structural requirements would support a
prominent role for DNA damage in the antiproliferative activity of
YM155.
cleaved caspase 3 and either preceded the appearance of gH2AX or
were observed concurrently. That the minimally substituted AB7
evoked down-regulation of survivin and many other properties
associated with YM155 pointed to the sufficiency of the pared
down dioxonaphthoimidazolium scaffold for activity.
In conclusion, we have shown that YM155 and its dioxonaph-
thoimidazolium analogs are modest DNA intercalators with limited
potential as redox cyclers. Intercalation per se does not directly
result in DNA damage but it does promote unwinding of the DNA
helix which would render it susceptible to events (misreading,
mutations, interference with transcription and replication) that
perturb the transcription of gene products in canonical pathways
that are essential for cell survival. Thus, the capacity of function-
alized dioxonaphthoimidazoliums to inflict direct DNA damage
leading to cell death should be reconsidered. While it would be
remiss to disregard DNA as a target of YM155 and its dioxonaph-
thoimidazolium analogs, its role in the antiproliferative activities of
these compounds may not be as significant as originally proposed.
Neither do our results refute the survivin suppressant properties of
YM155 and its analogs. Rather, they point to the possibility that
survivin is not the only gene product whose transcription is dis-
rupted by these compounds.
The structural requirements for growth inhibition were deduced
from the activities of YM155 and its analogs on four malignant cell
lines that belonged to two cancer types. A common finding was that
the intact dioxonaphthoimidazolium scaffold was required for ac-
tivity. Of the various motifs on the tricyclic scaffold, the quinone
was recognized as the most critical, followed by the positively
charged imidazolium and lastly, the benzene ring. Greater diversity
1
3
was tolerated at the N and N side chains of the scaffold. A striking
SAR trend that applied to both positions was the marked preference
for short, unbranched and cyclized alkyl groups. In fact, this was the
0
only acceptable modification of the 2 -methoxyethyl side chain at
1
3
N . Greater leeway was permitted at the N position which accepts
5. Experimental section
short unbranched alkyl groups as well as isosteric modifications of
0
the pyrazine-2 -ylmethyl side chain.
5.1. General details for chemical synthesis
Analysis of the SAR for intercalation and growth inhibition
revealed that aside from a common requirement for an intact
scaffold, there were few overlapping structural features for these
two activities. A case in point is the marked preference for cyclized
Reagents (synthetic grade or better) were obtained from com-
mercial sources and used without further purification. Reactions
were monitored by thin layer chromatography using pre-coated
aluminum plates (silica gel 60, F254, Merck). Column chromatog-
raphy on silica gel 60 (230e400 mesh, Merck) was used for com-
1
2
3
or bulky substituents at N , C and N which was not favored for
growth inhibition. It is noted however that minimally substituted
AB7 was an exception and demonstrated strong intercalation and
growth inhibition. Limited overlaps were observed in the SAR for
redox cycling and growth inhibition. Notably, there is less certainty
as to whether the intact scaffold is required for redox cycling or
even if the presence of quinone will ensure redox cycling.
1
13
pound purification. H NMR spectra (400 MHz) and C spectra
(100 MHz) were recorded on a Bruker Avance 400 Ultrashield in-
strument (Bruker, Billerica, MA, USA), and referenced to residual
solvent peaks. Nominal mass information was captured on an AB
Sciex QTrap 2000 mass spectrometer (AB Sciex, Framingham, MA,
USA) by electrospray ionization (ESI). Accurate mass information
was captured on a Bruker micrOTOFQII mass spectrometer (Bruker,
Billerica, MA, USA) by ESI. Compound purity was ascertained on a
reverse-phase HPLC on a Shimadzu Nexera SR HPLC system (Shi-
madzu Scientific Instruments, Columbia, MD, USA) with a Zorbax
Eclipse XDB-C18 column (Agilent Tech. Inc., Loveland, CO) at two
wavelengths: 254 and 280 nm. Each test compound was evaluated
on two of four isocratic solvent systems: System A: 60%
MeOH þ 40% water, with 15 mM ammonium formate; System B:
1
3
Furthermore, ring bearing groups on N and N were yet again
preferred for redox cycling whereas these same groups adversely
affected growth inhibition. Taken together, the present results do
not convincingly point to overlapping structural requirements for
growth inhibition and DNA intercalation or redox cycling. YM155
and its potent dioxonaphthoimidazolium analogs (A3-3, AB1, AB7)
are modest DNA intercalators with limited capacities to generate
free radicals. The latter was confirmed in cell based assays, lending
further support to the notion that free radical generation does not
contribute significantly to the cellular effects of these compounds.
Notwithstanding these SAR deductions, YM155, AB1 and AB7
30% MeOH þ 30% CH
formate; System C: 60% MeOH þ 40% water; System D: 60%
CH
CN þ 40% water. Compounds with peak areas ꢃ95% on two
3
CN þ 40% water, with 15 mM ammonium
elevated levels of the DNA damage marker
gH2AX in treated cells.
3

S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
53
solvent systems were deemed sufficiently pure for biological in-
vestigations. The purity of C2-3 was determined by elemental
analysis (C, H, N) on the Elementar Vario Micro Cube instrument
under suction and washed with EtOH and distilled water to afford
(64) as a yellow-brown solid, which was unstable and used
immediately for the next step.
(
Elementar Analysensysteme GmbH, Hanau, Germany) and found
to be within ±0.4% of expected values. Spectral and nominal mass
data are provided in the Supporting Information.
5.1.6. N-[1,4-dioxo-3-methylamino-1,4-dihydro-naphthalen-2-yl]-
N-methyl-acetamide (70)
To a stirred suspension of (19) (0.400 g, 1.52 mmol) in 2 mL of
toluene was added dropwise methylamine (0.0564 g, 1.82 mmol)
and triethylamine (0.230 g, 2.28 mmol). The mixture was stirred at
5
.1.1. 2-Chloro-3-(methylamino)-1,4-naphthoquinone (7)
To 2,3-dichloro-1,4-naphthoquinone (0.341 g, 1.50 mmol) in
1
.5 ml of ethanol was added methylamine (0.0465 g, 1.50 mmol)
r.t. for 1 h. The precipitate was filtered under suction, washed with
1
and triethylamine (0.227 g, 2.25 mmol) and the mixture stirred at
r.t. for 18 h. The red precipitate formed was filtered under suction,
washed with distilled water and dried to afford (7) as a red solid,
distilled water and EtOH to afford (70) as red solids, 62.0%.
(CDCl
H
3
)
d
8.15 (dd, 1H, J ¼ 0.83, 7.70 Hz), 8.08 (dd, 1H, J ¼ 0.87,
7.70 Hz), 7.78 (dt, 1H, J ¼ 1.33, 7.59 Hz), 7.66 (dt, 1H, J ¼ 1.30,
1
91.0%. H (CDCl
3
)
d
8.15 (dd, 1H, J ¼ 0.90, 7.69 Hz), 8.03 (dd, 1H,
7.56 Hz), 6.24 (br s, 1H), 3.13 (s, 3H), 3.11 (d, 3H, J ¼ 5.78 Hz), 1.99 (s,
13
J ¼ 0.94, 7.67 Hz), 7.72 (dt, 1H, J ¼ 1.34, 7.60 Hz), 7.62 (dt, 1H, J ¼ 1.31,
3
3H); C (CDCl ) d 182.30, 179.35, 172.36, 144.11, 135.47, 132.77,
13
7
d
.56 Hz), 6.10 (br s, 1H), 3.45 (d, 3H, J ¼ 5.60 Hz); C (CDCl
3
)
132.66, 130.14, 126.86, 126.76, 118.44, 37.50, 30.65, 21.94.
180.53, 144.89, 134.94, 132.78, 132.45, 132.39, 129.74, 126.84,
126.76, 32.55.
5.1.7. 1-Cyclopropyl-2-methyl-4,9-dioxo-3-(pyridin-4-ylmethyl)-
4,9-dihydro-1H-naphtho[2,3-d]imidazol-3-ium hydrogen
5.1.2. 2-Chloro-3-cyclopropylamino-1,4-naphthoquinone (11)
dibromide (AB1)
To 2,3-dichloro-1,4-naphthoquinone (0.341 g, 1.50 mmol) in
.5 ml of ethanol was added cyclopropylamine (0.0856 g,
.50 mmol) and triethylamine (0.227 g, 2.25 mmol) and the mixture
48% Hydrobromic acid (30.0 mmol) was added dropwise to a
solution of crude (64) in a 4 mL mixture of 1:1 EtOH/EtOAc and
stirred at 45 C for 4 h and subsequently r.t. for an additional 12 h.
1
1
ꢀ
stirred at r.t. for 18 h. The red precipitate formed was filtered under
The reaction mixture was concentrated in vacuo and purified by
suction, washed with distilled water and dried to afford (11) as a
column chromatography (8:92 MeOH/CH
2
Cl
2
) to afford (AB1) as a
1
1
red solid, 93.9%. H (CDCl
3
)
d
8.16 (dd, 1H, J ¼ 1.11, 7.69 Hz), 8.02 (dd,
6
yellow solid, 22.0%. H (D -DMSO) d
8.76 (d, 2H, J ¼ 6.36 Hz), 8.23
1
H, J ¼ 1.23, 7.67 Hz), 7.72 (dt, 1H, J ¼ 1.31, 7.61 Hz), 7.62 (dt, 1H,
(dd, 1H, J ¼ 1.14, 7.65 Hz), 8.09 (dd, 1H, J ¼ 1.33, 7.49 Hz), 8.01 (dt,
1H, J ¼ 1.55, 7.55 Hz), 7.96 (dt, 1H, J ¼ 1.44, 7.47 Hz), 7.72 (d, 2H,
J ¼ 6.35 Hz), 6.08 (s, 2H), 3.77e3.71 (m, 1H), 2.90 (s, 3H), 1.43e1.38
J ¼ 1.28, 7.57 Hz), 6.12 (br s, 1H), 3.34e3.28 (m, 1H), 0.97e0.92 (m,
1
3
2
H), 0.78e0.73 (m, 2H); C (CDCl
3
) d 180.27, 177.14, 145.07, 134.91,
13
132.74, 132.46, 129.75, 126.92, 126.72, 111.52, 27.43, 10.40.
6
(m, 2H), 1.30e1.26 (m, 2H); C (D -DMSO) d 174.87, 173.50, 155.98,
147.98, 146.17, 135.58, 135.11, 132.51, 132.09, 131.24, 129.62, 127.16,
5
.1.3. N-(3-chloro-1,4-dioxo-1,4-dihydro-naphthalen-2-yl)-N-
126.76, 123.23, 49.16, 29.84, 11.85, 9.00. ESI-MS: m/z 344.1
þ
þ
methyl-acetamide (19)
Two drops of concentrated sulfuric acid were added to a sus-
pension of (7) (0.443 g, 2.00 mmol) in acetic anhydride (1.84 g,
[MꢁBr] . High resolution MS (ESI) calcd for C21
H18 BrN
3
O
2
[MꢁBr]
344.1394. Found: 344.1405. HPLC purity: system A: 99.85%
(254 nm), 100.00% (280 nm); system B: 99.38% (254 nm), 99.40%
(280 nm).
18.0 mmol) and stirred for 1.5 h at r.t. 10 ml of distilled water was
added slowly to the reaction mixture with stirring to quench excess
anhydride, and extracted with EtOAc. The organic layer was washed
5.1.8. 1,2,3-Trimethyl-4,9-dioxo-4,9-dihydro-1H-naphtho[2,3-d]
imidazol-3-ium bromide (AB7)
with saturated NaHCO
drous Na SO . Purification by column (1:4 EtOAc/Hexanes) afforded
19) as a yellow solid, 92.6%. H (CDCl
.86e7.77 (m, 2H), 3.19 (s, 3H), 1.93 (s, 3H); C (CDCl
79.16, 177.84, 134.89, 134.77, 134.72, 134.69, 134.62, 134.16, 131.31,
27.57, 40.11, 21.76.
3
solution and brine and dried over anhy-
2
4
48% hydrobromic acid (10.0 mmol) was added dropwise to a
solution of (70) (0.263 mg, 1.0 mmol) in a 1.5 ml mixture of 1:1
1
(
7
3
) d 8.28e8.08 (m, 2H),
13
ꢀ
3
)
d
183.10,
EtOH/EtOAc and stirred at 40 C for 4 h followed by r.t. for an
1
1
additional 12 h. The reaction mixture was concentrated in vacuo
and purified by column chromatography (8:92 MeOH/CH
afford AB7 as a beige solid, 21.9%. H (D -DMSO) d
6
2
Cl
8.18 (dd, 1H,
J ¼ 3.31, 5.73 Hz), 7.99 (dd, 1H, J ¼ 3.31, 5.74 Hz), 4.13 (s, 6H), 2.79 (s,
2
) to
1
5.1.4. N-(3-chloro-1,4-dioxo-1,4-dihydro-naphthalen-2-yl)-N-
13
cyclopropyl-acetamide (23)
6
3H); C (D -DMSO) d 175.36, 153.39, 135.60, 131.91, 130.34, 127.18,
Two drops of concentrated sulfuric acid were added to a sus-
pension of (11) (0.495 g, 2.0 mmol) in acetic anhydride (1.84 g,
34.37, 10.17. High resolution MS (ESI) calcd for C14
H
13 BrN
2
O
2
þ
[MꢁBr] 241.0972. Found 241.0967. HPLC purity: system A: 99.83%
(254 nm), 99.89% (280 nm); system B: 99.74% (254 nm), 99.74%
(280 nm).
18.0 mmol) and stirred for 1.5 h at r.t. 10 ml of distilled water was
added slowly to the reaction mixture with stirring to quench excess
anhydride, and extracted with EtOAc. The organic layer was washed
with saturated NaHCO
drous Na SO . Purification by column (1:4 EtOAc/Hexanes) afforded
23) as yellow plates, 89.1%. H (CDCl
3
solution and brine and dried over anhy-
5.2. Biological assays
2
4
1
(
3
)
d
8.20e8.12 (m, 2H),
5.2.1. General biology
7
.84e7.74 (m, 2H), 3.27e3.17 (m, 1H), 2.43 (s, 3H), 0.98e0.68 (m,
Dulbecco's Modified Eagle's Media (DMEM) and Roswell Park
Memorial Institute Media 1640 (RPMI-1640) were purchased from
Caisson Laboratories Inc. (North Logan, UT) and supplemented with
13
4
H); C (CDCl
3
) d 178.01, 134.81, 134.66, 134.56, 134.46, 134.29,
131.40, 131.15, 127.43, 127.40, 31.31, 22.33, 8.86, 8.60.
10% heat inactivated fetal bovine serum (FBS) (Gibco, Life Tech-
5
.1.5. N-cyclopropyl-N-{1,4-dioxo-3-[(pyridin-4-ylmethyl)-amino]-
nologies Corporation, Carlsbad, CA, USA) and 0.01% w/v Penicillin-
G-Streptomycin mixture (GE Healthcare, Buckinghamshire, United
Kingdom). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) was purchased from Alfa Aesar, Inc. (Lancashire,
United Kingdom), reconstituted in phosphate-buffered saline to
2 mg/mL and diluted with appropriate cell culture media before
1,4-dihydro-naphthalen-2-yl}-acetamide (64)
To a suspension of (23) (0.955 g, 3.30 mmol) in 5 ml of toluene
was added pyridin-4-ylmethylamine (0.535 g, 4.95 mmol) and
triethylamine (0.500 mg, 4.95 mmol) and stirred at 45 C for 2 h.
The reaction mixture was cooled, and the precipitate was filtered
ꢀ

54
S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
þ
use. Herring sperm DNA and the Apotox-Glo™ assay kit were ob-
tained from Promega Ptd. Ltd. (Madison, WI, USA). Doxorubicin,
naphthoquinone, thiazole orange, phenol red sodium, dithio-
threitol and horseradish peroxidase (HRP) were purchased from
Na
2
EDTA; total Na concentration 16.0 mM; pH 7.0) containing
2 mg/mL herring sperm DNA and 1 mL of 10 mM thiazole orange
(ThO) in deionized water. The resulting mixture was incubated at
room temperature (25 ) for 5 min. 0.5
concentration in DMSO) was added, kept at room temperature for
ꢀ
mL of compound (200-fold
0
0
SigmaeAldrich (St. Louis, MO, USA). 2 , 7 -Dichlorodihydro-
fluorescein diacetate (H DCF DA) was purchased from Life Tech-
nologies Corporation (Carlsbad, CA, USA). Anti-cleaved caspase-3
and anti- H2AX antibodies were obtained from Cell Signalling
2
8 min, with agitation (500 rpm on a plate shaker) and fluorescence
read at
l
excitation ¼ 503 nm,
lemission ¼ 536 nm on a microplate
g
reader (Tecan Infinite™ M200 Pro). DMSO per well was kept at 0.5%
v/v. Test compound and doxorubicin (positive control) were tested
over a range of concentrations in at least 3 separate experiments.
The percentage of undisplaced thiazole orange at a specific con-
centration of test compound was determined from the expression:
Technology Inc. (Danvers, MA, USA) and anti-GADPH antibody was
from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA).
5.2.2. Cell culture
Human clear cell renal cell carcinoma cell lines RCC786-0, RCC4/
.
%
Undisplacedthiazoleorange¼RFUcompound RFUcontrol ꢂ100%;
VA, non-small cell lung carcinoma cell lines H1299, H1666 and
proliferating normal lung fibroblasts IMR-90 cells were obtained
from American Type Culture Collection (ATCC, Rockville, MD).
RCC786-0, RCC4/VA and IMR-90 cells were cultured in complete
where RFUcompound ¼ fluorescence of test compound and thiazole
orange, and RFUcontrol ¼ fluorescence of thiazole orange without
test compound. The DC50 of test compound (concentration required
DMEM. H1299 and H1666 cells were cultured in complete RPMI-
ꢀ
1
640 and incubated at 37 C, 5% CO
2
. Malignant and IMR-90 cell
to reduce basal fluorescence of thiazole orange (10 mM) by 50%) was
lines were sub-cultured at ratios of 1:6 and 1:3 respectively on
reaching 90% confluence.
determined by plotting % undisplaced thiazole orange against
logarithmic concentration of test compound. (GraphPad Prism 5,
San Diego, CA).
5
.2.3. MTT assay
3
Cells were seeded at a density of 3.0 ꢂ 10 cells/well for ma-
5.2.5. Horseradish peroxidase-phenol red redox cycling assay
3
lignant cell lines and 4.0 ꢂ 10 cells/well for IMR-90 in 96-well
The method described by Johnston et al. [40] and Soares et al.
plates containing 100
m
L media per well and incubated for 24 h
[50] were followed with modifications. 2
mL test compound (200-
ꢀ
(
37 C, 5% CO
2
) for cell adherence. After 24 h, media in each well
L of fresh media was added.
L of test compound (prepared in DMSO stock solutions at 200-
fold concentration in DMSO), 18
mL Hank's Balanced Salt Solution
was removed by aspiration and 199
1
m
(HBSS) and 40 mL dithiothreitol (DTT) solution (2.5 mM in HBSS)
m
were sequentially added to each well in a 96-well clear plate. The
contents were shaken for 45 min at 600 rpm on a plate shaker, after
fold higher concentrations) was added to each well. The final
DMSO concentration in each well was kept at 0.5% v/v. The
which was immediately added 40
phenol red solution (150 g/mL horseradish peroxidase, 1 mM
phenol red sodium in HBSS) and the plate shaken again for another
0 min at 600 rpm. 15 L 1 M NaOH was added to each well,
mL of horseradish peroxidase-
ꢀ
compound-treated cells were incubated for 72 h (37 C, 5% CO
2
),
m
media was subsequently removed by aspiration and an aliquot
200 L) of 0.5 mg/mL MTT in media was added to each well. The
treated plates were incubated for 2 h (malignant cells) or 3 h (IMR-
0 cells) after which the MTT containing media was removed by
aspiration and 100 L of DMSO added to each well to dissolve the
(
m
1
m
agitated (1 min) and absorbance readings were read at 610 nm on a
microplate reader (Tecan Infinite™ M200 Pro). DMSO content per
well was kept at 1% v/v. Test compound and positive control
naphthoquinone were tested over a range of concentrations, with
at least three separate experiments performed for each concen-
tration. The absorbance at 610 nm was plotted against logarithmic
concentration of test compound (GraphPad Prism 5, San Diego, CA)
from which EC50 (concentration required to reduce oxidized phenol
red absorbance to 50% of control value) was determined.
9
m
purple formazan crystals. Plates were agitated at 700 rpm, 5 min on
a plate shaker before absorbance readings were read (570 nm,
Tecan Infinite™ M200 Pro). Viable cells were determined from the
following equation:
hꢀ
ꢁ.
Percentage viability ¼ Abcompound ꢁ Ab
blank
i
ðAbcontrol ꢁ AbblankÞ ꢂ 100%;
5
2
.2.6. H DCF-DA assay
4
RCC786-0 and H1299 cells were seeded at 1.0 ꢂ 10 cells/well
where Abcompound ¼ Absorbance of compound-treated cells,
4
and 1.5 ꢂ 10 cells/well respectively in black-walled, clear-bottom
Abcontrol
¼
Absorbance of untreated/control cells and
ꢀ
9
6-well cell culture plates and incubated for 24 h at 37 C, 5% CO
2
in
Abblank ¼ Absorbance of DMSO.
1
00
well to give a final concentration of 20
for another hour. The media was removed by aspiration and
replaced by 100 L of pre-warmed PBS containing test compound
and 0.5% v/v of DMSO. After incubation for 2 h, plates were read
using microplate reader (Tecan Infinite™ M200 Pro) at
m
L of media. 2
m
L H2DCF-DA in DMSO was then added to each
Each test compound was assessed at 8 concentrations in at least
mM and cells were incubated
3
separate experiments carried out at different times. Cells were
from different passage numbers and each compound was tested
from 2 stock solutions. For each experiment, percentage viability
was calculated using the mean absorbance value from three tech-
nical repeats at a given test concentration. The IC50 (the concen-
tration of test compound required to inhibit cell growth by 50%)
was determined by plotting % viability against logarithmic con-
centration of test compound (GraphPad Prism 5, San Diego, CA).
m
a
l
excitation ¼ 495 nm,
l
emission ¼ 526 nm. Test compounds were
tested at concentrations corresponding to 1ꢂ, 2ꢂ, 5ꢂ and 10ꢂ cell-
based IC50 and fluorescence at each concentration was expressed as
fold-change of control fluorescence recorded in untreated cells.
Each concentration was investigated in three separate experiments.
5
.2.4. Fluorescent intercalator displacement assay to assess binding
to DNA
Previously reported methods were followed with modifications
37,38]. To each well in a black 96-well plate was added 98.5 L of
BPE buffer (comprising 6.0 mM Na HPO , 2.0 mM NaH PO , 1.0 mM
5.2.7. Western blotting
RCC786-0 and H1299 cells were seeded at a cell density of
5
ꢀ
[
m
5.0 ꢂ 10 cells/plate in 100 mm petri dishes (24 h, 37 C, 5% CO
2
)
2
4
2
4
containing 5 mL of media. The media was removed by aspiration

S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
55
and replaced by 10 mL of fresh media containing test compound
[8] S. Gurbuxani, Y. Xu, G. Keerthivasan, A. Wickrema, J.D. Crispino, Differential
requirements for survivin in hematopoietic cell development, Proc. Natl. Acad.
U. S. A. 102 (2005) 11480e11485.
and 0.5% v/v of DMSO. After incubation for the indicated times (6,
®
2
4, 48 or 72 h), cells were harvested and lysed in Cellytic M buffer
[9] D.C. Altieri, Targeting survivin in cancer, Cancer Lett. 332 (2013) 225e228.
(
SigmaeAldrich, St. Louis, MO, USA). Protein content was assessed
by Bradford assay and subjected to SDS-PAGE. Cells were blocked in
% non-fat milk and probed with anti-cleaved caspase-3 or anti-
H2AX antibodies (1:1000 dilution) to determine protein levels.
[10] D.N. Church, D.C. Talbot, Survivin in solid tumors: rationale for development
of inhibitors, Curr. Oncol. Rep. 14 (2012) 120e128.
[11] D.C. Altieri, Survivin, cancer networks and pathway-directed drug discovery,
5
g
Nat. Rev. Cancer 8 (2008) 61e70.
[12] A. Matsuhisa, I. Kinoyama, A. Toyoshima, T. Nakahara, M. Takeuchi, M. Okada,
Fused imidazolium derivatives, US Patent 6,734,203 B2, May 11 2004.
Anti-GADPH antibody was used as a loading control.
[
13] X. Ling, S. Cao, Q. Cheng, J.T. Keefe, Y.M. Rustum, F. Li, A novel small molecule
FL118 that selectively inhibits survivin, Mcl-1, XIAP and cIAP2 in a p53-
independent manner, shows superior antitumor activity, PLoS One 7 (2012)
e45571.
5
.2.8. Transcriptome analysis
Total RNA was isolated using the RNeasy Plus Mini kit (Qiagen
[
14] T. Oikawa, Y. Unno, K. Matsuno, J. Sawada, N. Ogo, K. Tanaka, A. Asai, Iden-
tification of a small-molecule inhibitor of the interaction between Survivin
and Smac/DIABLO, Biochem. Biophys. Res. Commun. 393 (2010) 253e258.
15] M. Tanioka, H. Nokihara, N. Yamamoto, Y. Yamada, K. Yamada, Y. Goto,
T. Fujimoto, R. Sekiguchi, K. Uenaka, S. Callies, T. Tamura, Phase 1 study of
LY2181308, an antisense oligonucleotide against survivin, in patients with
advanced solid tumors, Cancer Chemother. Pharmacol. 68 (2011) 505e511.
GmBH, Hilden, Germany). mRNA was quantified using BioSPEC-
Mini (Shimadzu) and its quality evaluated using the Agilent 2100
Bioanalyzer (Agilent Technologies). The samples were processed
using Affymetrix 3 IVT Express to create biotin-labeled amplified
RNA, which was subsequently fragmented and hybridized to Affy-
[
0
ꢀ
metrix PrimeView Human Gene Expression Arrays (16 h, 45 C,
0 rpm). Arrays were scanned using an Affymetrix 3000 7G scan-
[16] B.D. Cheson, N.L. Bartlett, J.M. Vose, A. Lopez-Hernandez, A.L. Seiz,
A.T. Keating, S. Shamsili, K.P. Papadopoulos, A phase II study of the survivin
suppressant YM155 in patients with refractory diffuse large B-cell lymphoma,
Cancer 118 (2012) 3128e3134.
[17] M.R. Clemens, O.A. Gladkov, E. Gartner, V. Vladimirov, J. Crown, J. Steinberg,
F. Jie, A. Keating, Phase II, multicenter, open-label, randomized study of
YM155 plus docetaxel as first-line treatment in patients with HER2-negative
metastatic breast cancer, Breast Cancer Res. Treat. 149 (2015) 171e179.
[18] N. Nakamura, T. Yamauchi, M. Hiramoto, M. Yuri, M. Naito, M. Takeuchi,
K. Yamanaka, A. Kita, T. Nakahara, I. Kinoyama, A. Matsuhisa, N. Kaneko,
H. Koutoku, M. Sasamata, H. Yokota, S. Kawabata, K. Furiuchi, Interleukin
enhancer-binding factor 3/NF110 is a target of YM155, a suppressant of sur-
vivin, Mol. Cell Proteomics 11 (2012). M111.013243.
6
ner. Data processing was conducted on the Partek Genomics Suite
V6.3 (Partek, St Louis, MI). The raw expression values obtained were
processed using the Robust Multi-chip Average (RMA) method [51].
p Values for individual genes were adjusted using the Benjamini
and Hochberg method. A list of significantly perturbed genes
(
p < 0.05 and fold change >2) was generated and pathway analysis
conducted using the Ingenuity Pathway Analysis (IPA) knowledge
base. The significance of the association between the input data set
and the functions or pathways was determined by two parameters:
[
19] Q. Cheng, X. Ling, A. Haller, T. Nakahara, K. Yamanaka, A. Kita, H. Koutoku,
M. Takeuchi, M.G. Brattain, F. Li, Suppression of survivin promoter activity by
YM155 involves disruption of Sp1-DNA interaction in the survivin core pro-
moter, Int. J. Biochem. Mol. Biol. 3 (2012) 179e197.
(
i) the ratio of the number of genes from the data set that map to
the function/pathway divided by the total number of genes that
map to the function/pathway and (ii) p-value calculated using
Fisher's exact test to determine the probability that the association
between the genes in the data set and the function/pathway arose
by chance alone.
[
20] M. Yamauchi, N. Nakamura, M. Hiramoto, M. Yuri, H. Yokoto, M. Naitou,
M. Takeuchi, K. Yamanaka, A. Kita, T. Nakahara, I. Kinoyama, A. Matsuhisa,
N. Kaneko, H. Koutoku, M. Sasamata, M. Kobori, M. Katou, S. Tawara,
S. Kawabata, K. Furuichi, Sepantronium Bromide (YM155) induces disruption
of the ILF3/p54nrd complex, which is required for survivin expression, Bio-
chem. Biophys. Res. Commun. 425 (2012) 711e716.
[
[
21] A. Rauch, D. Hennig, C. Sch €a fer, M. Wirth, C. Marx, T. Heinzel, G. Schneider,
Acknowledgments
€
O.H. Kramer, Survivin and YM155: how faithful is the liaison? Biochim. Bio-
phys. Acta 1845 (2014) 202e220.
22] H. Tang, H. Shao, C. Yu, J. Hou, Mcl-1 downregulation by YM155 contributes to
its synergistic anti-tumor activities with ABT-263, Biochem. Pharmacol. 82
The authors gratefully acknowledged funding support from
National University of Singapore (ML Go, R148000188112) and
SingHealth Foundation (SPJ Yuen, SHF/FG462P/2011), and Ministry
of Education for provision of the National University of Singapore
President Graduate Fellowship to Sherman Ho.
(2011) 1066e1072.
[23] Y.S. Na, S.J. Yang, S.M. Kim, K.A. Jung, J.H. Moon, J.S. Shin, D.H. Yoon, Y.S. Hong,
M.H. Ryu, J.L. Lee, J.S. Lee, T.W. Kim, YM155 induces EGFR suppression in
pancreatic cancer cells, PLoS One 7 (2012) e38625.
[
[
[
24] P.C. Lai, S.H. Chen, S.H. Yang, C.C. Cheng, T.H. Chiu, Y.T. Huang, Novel survivin
inhibitor YM155 elicits cytotoxicity in glioblastoma cell lines with normal or
deficiency DNA-dependent protein kinase activity, Pediatr. Neonatol. 53
Appendix A. Supplementary data
(2012) 199e204.
25] Y.F. Tao, J. Lu, X.J. Du, L.C. Sun, X. Zhao, L. Peng, L. Cao, P.F. Xiao, L. Pang, D. Wu,
N. Wang, X. Feng, Y.H. Li, J. Ni, J. Wang, J. Pan, Survivin selective inhibitor
YM155 induce apoptosis in SK-NEP-1 Wilms tumor cells, BMC Cancer 12
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2015.09.026.
(2012) 619e632.
26] T.G. Glaros, L.H. Stockwin, M.E. Mullendore, B. Smith, B.L. Morrison,
D.L. Newton, The “survivin suppressants” NSC 80467 and YM155 induce a
DNA damage response, Cancer Chemother. Pharmacol. 70 (2012) 207e212.
References
[
1] C.H.A. Cheung, L.T. Cheng, K.Y. Chang, H.H. Chen, J.Y. Chang, Investigations of
survivin: the past, present and future, Front. Biosci. 16 (2011) 952e961.
2] J.M. Hernandez, J.M. Farma, D. Coppola, A. Hakam, W.J. Fulp, D.T. Chen,
E.M. Siegel, T.J. Yeatman, D. Shibata, Expression of the antiapoptotic protein
survivin in colon cancer, Clin. Colorectal Cancer 10 (2011) 188e193.
3] B.M. Ryan, N. O'Donovan, M.J. Suffy, Survivin: a new target for anti-cancer
therapy, Cancer Treat. Rev. 35 (2009) 553e562.
[27] R.B. Silverman, M.W. Holladay, Chapter 6-DNA-interactive agents, in:
R.B. Silverman, M.W. Holladay (Eds.), The Organic Chemistry of Drug Design
and Drug Action, third ed., Academic Press, Boston, 2014, pp. 275e331.
[28] G.E. Winter, B. Radic, C. Mayor-Ruiz, V.A. Blomen, C. Trefzer, R.K. Kandasamy,
V.K.M. Huber, M. Gridling, D. Chen, T. Klampfl, R. Kralovics, S. Kubicek,
O. Fernandez-Capetillo, T.R. Brummelkamp, G. Superti-Furga, The solute car-
rier SLC35F2 enables YM155-mediated DNA damage toxicity, Nat. Chem. Biol.
10 (2014) 768e773.
[29] A.S. Kalgutkar, I. Gardner, R.S. Obach, C.L. Shaffer, E. Callegari, K.R. Henne,
A.E. Mutlib, D.K. Dalvie, J.S. Lee, Y. Nakai, J.P. O'Donnell, J. Boer, S.P. Harriman,
A comprehensive listing of bioactivation pathways of organic functional
groups, Curr. Drug. Metab. 6 (2005) 161e225.
[30] S.C. Kuo, T. Ibuka, L.J. Huang, J.C. Lien, S.R. Yean, S.C. Huang, D. Lednicer,
S. Morris-Natschke, K.H. Lee, Synthesis and cytotoxicity of 1,2-disubstituted
naphth[2,3-d]imidazole-4,9-diones and related compounds, J. Med. Chem.
39 (1996) 1447e1451.
[31] G. Giaccone, P. Zatloukal, J. Roubec, K. Floor, J. Musil, M. Kuta, R.J. van Kla-
veren, S. Chaudhary, A. Gunther, S. Shamsili, Multicenter phase II trial of
YM155, a small-molecule suppressor of survivin, in patients with advanced,
refractory, non-small-cell lung cancer, J. Clin. Oncol. 27 (2009) 4481e4486.
[32] A. Mahadevan, H. Sard, M. Gonzalez, J.C. McKew, A general method for C3
[
[
[
4] A.C. Mita, M.M. Mita, S.T. Nawrocki, F.J. Giles, Survivin: key regulator of mitosis
and apoptosis and novel target for cancer therapeutics, Clin. Cancer Res. 14
(
2008) 5000e5005.
[
5] S. Weikert, M. Schrader, H. Krause, W. Schulze, M. Muller, K. Miller, The in-
hibitor of apoptosis (IAP) survivin is expressed in human testicular germ cell
tumors and normal testes, Cancer Lett. 223 (2004) 331e337.
[
6] S.P. Tu, X.H. Jiang, M.C.M. Lin, J.T. Cui, Y. Yang, C.T. Lum, B. Zou, Y.B. Zhu,
S.H. Jiang, W.M. Wong, A.O.O. Chan, M.F. Yuen, S.K. Lam, H.F. Kung,
B.C.Y. Wong, Suppression of survivin expression inhibits in vivo tumorige-
nicity and angiogenesis in gastric cancer, Cancer Res. 63 (2003) 7724e7732.
7] I. Tamm, Y. Wang, E. Sausville, D.A. Scudiero, N. Vigna, T. Oltersdorf, J.C. Reed,
IAP-family protein survivin inhibits caspase activity and apoptosis induced by
Fas (CD95), Bax, caspases, and anticancer drugs, Cancer Res. 58 (1998)
[
5315e5320.

56
S.-H.S. Ho et al. / European Journal of Medicinal Chemistry 104 (2015) 42e56
reductive alkylation of indoles, Tetrahedron Lett. 44 (2003) 4589e4591.
33] H. Heaney, S.V. Ley, N-alkylation of indole and pyrroles in dimethyl sulph-
oxide, J. Chem. Soc. Perkin Trans. 1 (1973) 499e500.
34] L. Gossage, T. Eisen, E.R. Maher, VHL, the story of a tumour suppressor gene,
Nat. Rev. Cancer 15 (2015) 55e64.
S. Burma, DNA-PK phosphorylates histone H2AX during apoptotic DNA frag-
mentation in mammalian cells, DNA Repair 5 (2006) 575e590.
[45] T. Nakahara, M. Takeuchi, I. Kinoyama, T. Minematsu, K. Shirasuna,
A. Matsuhisa, A. Kita, F. Tominaga, K. Yamanaka, M. Kudoh, M. Sasamata,
YM155, a novel small-molecule survivin suppressant, induces regression of
established human hormone-refractory prostate tumor xenografts, Cancer
Res. 67 (2007) 8014e8021.
[46] T. Nakahara, A. Kita, K. Yamanaka, M. Mori, N. Amino, M. Takeuchi,
F. Tominaga, I. Kinoyama, A. Matsuhisa, M. Kudoh, M. Sasamata, Broad spec-
trum and potent antitumor activities of YM155, a novel small-molecule sur-
vivin suppressant, in a wide variety of human cancer cell lines and xenograft
models, Cancer Sci. 102 (2011) 614e621.
[47] C. Lu, F. Zhu, Y.Y. Cho, F. Tang, T. Zykova, W.Y. Ma, A.M. Bode, Z. Dong, Cell
apoptosis: requirement of H2AX in DNA ladder formation, but not for the
activation of caspase-3, Mol. Cell. 23 (2006) 121e132.
[48] Y.J. Zhang, C.R. Lu, Y. Cao, Y. Luo, R.F. Bao, S. Yan, M. Xue, F. Zhu, Z. Wang,
L.N. Duan, Imatinib induces H2AX phosphorylation and apoptosis in chronic
myelogenous leukemia cells in vitro via caspase-3/Mst1 pathway, Acta
Pharmacol. Sin. 33 (2012) 551e557.
[
[
[
35] D.L. Boger, W.C. Tse, Thiazole orange as the fluorescent intercalator in a high
resolution fid assay for determining DNA binding affinity and sequence
selectivity of small molecules, Bioorg. Med. Chem. 9 (2001) 2511e2518.
36] W.C. Tse, D.L. Boger, A fluorescent intercalator displacement assay for estab-
lishing DNA binding selectivity and affinity, Acc. Chem. Res. 37 (2004) 61e69.
37] T. Sawahata, R.A. Neal, Biotransformation of phenol to hydroquinone and
catechol by rat liver microsomes, Mol. Pharmacol. 23 (1983) 453e460.
38] T.W. Gant, D.N. Ramakrishna Rao, R.P. Mason, G.M. Cohen, Redox cycling and
sulphydryl arylation; their relative importance in the mechanism of quinone
cytotoxicity to isolated hepatocytes, Chem. Biol. Interact. 65 (1988) 157e173.
39] B. Halliwell, C.E. Cross, Oxygen-derived species: their relation to human dis-
ease and environmental stress, Environ. Health Perspect. 102 (1994) 5e12.
40] P.A. Johnston, K.M. Soares, S.N. Shinde, C.A. Foster, T.Y. Shun, H.K. Takyi,
P. Wipf, J.S. Lazo, Development of a 384-well colorimetric assay to quantify
hydrogen peroxide generated by the redox cycling of compounds in the
presence of reducing agents, Assay Drug Dev. Technol. 6 (2008) 505e518.
41] W.M. Bonner, C.E. Redon, J.S. Dickey, A.J. Nakamura, O.A. Sedelnikova, S. Solier,
[
[
[
[
[
[49] S. Solier, Y. Pommier, The apoptotic ring: a novel entity with phosphorylated
histones H2AX and H2B and activated DNA damage response kinases, Cell
Cycle 8 (2009) 1853e1859.
[
[
Y. Pommier,
g
H2AX and cancer, Nat. Rev. Cancer 8 (2008) 957e967.
[50] K.M. Soares, N. Blackmon, T.Y. Shun, S.N. Shinde, H.K. Takyi, P. Wipf, J.S. Lazo,
P.A. Johnston, Profiling the NIH Small Molecule Repository for compounds
that generate H2O2 by redox cycling in reducing environments, Assay Drug
Dev. Technol. 8 (2010) 152e174.
42] X. Huang, H.D. Halicka, Z. Darzynkiewicz, Detection of histone H2AX phos-
phorylation on Ser-139 as an indicator of DNA damage (DNA double-strand
breaks), Curr. Protoc. Cytom. 30 (2001), 7.27.1e7.27.7.
[
43] E.P. Rogakou, W. Nieves-Neira, C. Boon, Y. Pommier, W.M. Bonner, Initiation of
DNA fragmentation during apoptosis induces phosphorylation of H2AX his-
tone at serine 139, J. Biol. Chem. 275 (2000) 9390e9395.
[51] R.A. Irizarry, B. Hobbs, F. Collin, Y.D. Beazer-Barclay, K.J. Antonellis, U. Scherf,
T.P. Speed, Exploration, normalization, and summaries of high density oligo-
nucleotide array probe level data, Biostatistics 4 (2003) 249e264.
[44] B. Mukherjee, C. Kessinger, J. Kobayashi, B.P.C. Chen, D.J. Chen, A. Chatterjee,