Indole-derived psammaplin a analogues as epigenetic modulators with multiple inhibitory activities

Article abstract of DOI:10.1021/jm300618u

A SAR study has been carried out around a modified scaffold of the natural product psammaplin A obtained by replacing the o-bromophenol unit by an indole ring. A series of indole psammaplin A constructs were generated in a short synthetic sequence that starts with the functionalization of the C3 indole position with in situ generated nitrosoacrylate, and this is followed by protection of the β-indole-α-oximinoesters, saponification, condensation with symmetrical diamines, and deprotection. Biochemical and cellular characterization using U937 and MCF-7 cells confirmed that many of these analogues displayed more potent actitivies than the parent natural product. Moreover, in addition to the reported HDAC and DNMT dual epigenetic inhibitory profile of the parent compound, some analogues, notably 4a (UVI5008), also inhibited the NAD⁺-dependent SIRT deacetylase enzymes. The SAR study provides structural insights into the mechanism of action of these multiple epigenetic ligands and paves the way for additional structural exploration to optimize their pharmacological profiles. Because of their multi(epi)target features and their action in ex vivo samples, the indole-based psammaplin A derivatives are attractive molecules for the modulation of epigenetic disorders.

Full text of DOI:10.1021/jm300618u

Subscriber access provided by UniSA Library
Article
Indole-derived Psammaplin A Analogs as Epigenetic
Modulators With Multiple Inhibitory Activities
Raquel Pereira, Rosaria Benedetti, Santiago Perez-Rodriguez, Angela Nebbioso, José García-
Rodríguez, Vincenzo Carafa, Mariarosaria Conte, Mayra C Stuhldreier, Fatima Rodriguez-
Barrios, Hendrik Stunnenberg, Hinrich Gronemeyer, Lucia Altucci, and Angel R. de Lera
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 03 Oct 2012
Downloaded from http://pubs.acs.org on October 11, 2012
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Indole-derived Psammaplin A Analogs as Epigenetic
Modulators With Multiple Inhibitory Activities
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a,#
b,#
a
b
Raquel Pereira, Rosaria Benedetti, Santiago Pérez-Rodríguez, Angela Nebbioso, José
a
b
a
b
García-Rodríguez, Vincenzo Carafa, Mayra Stuhldreier, Mariarosaria Conte, Fátima
a
d
e
*b,c
Rodríguez-Barrios, Hendrik G. Stunnenberg, Hinrich Gronemeyer, Lucia Altucci, and Ángel
*
a
R. de Lera
a
Departamento de Química Orgánica, Universidade de Vigo, 36310 Vigo, Spain.
b
Dipartimento di Patologia Generale, Seconda Università degli Studi di Napoli, Vico L. de
Crecchio 7 80138 Napoli
c
Institute of Genetics and Biophysics (IGB), CNR, via P. Castellino 111, 80131, Napoli, Italy.
d
NCMLS, Department of Molecular Biology, Radboud University, 6525 GA Nijmegen, NL
#
These authors contributed equally to this work.
e
Department of Cancer Biology - Institut de Génétique et de Biologie Moléculaire et Cellulaire
(
IGBMC) / CNRS / INSERM / ULP, BP 163, 67404 Illkirch Cedex, C. U. de Strasbourg, France.
lucia.altucci@unina2.it, qolera@uvigo.es
List of Abbreviations
AML: Acute Myeloid Leukemia; APL: Acute Promyelocytic Leukemia; CTCL: Cutaneous T Cell Lymphoma;
DNMT: DNA Methyl Transferase; HAT: Histone Acetyl Transferase; HDAC: Histone Deacetylase; HDM: Histone
Demethylase; HMT: Histone Methyl Transferase; MDS: Myelodysplastic Syndromes; NAD: nicotine adenine
dinucleotide; SAH: S-Adenosyl Homocysteine (AdoHcy); SAHA: Suberoylanilide Hydroxamic Acid; SAM: S-
Adenosylmethionine (AdoMet); SAR: Structure-Activity Relationships.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT:
A SAR study has been carried out around a modified scaffold of the natural product psammaplin
A obtained by replacing the o-bromophenol unit by an indole ring. A series of indole-
psammaplin A constructs were generated in a short synthetic sequence that starts with the
functionalization of the C3 indole position with in situ generated nitrosoacrylate and this is
followed by protection of the β-indole-α-oximinoesters, saponification, condensation with
symmetrical diamines and deprotection. Biochemical and cellular characterization using U937
and MCF-7 cells confirmed that many of these analogs displayed more potent actitivies than the
parent natural product. Moreover, in addition to the reported HDAC and DNMT dual epigenetic
inhibitory profile of the parent compound, some analogs, notably 4a (UVI5008) also inhibited
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
the NAD -dependent SIRT deacetylase enzymes. The SAR study provides structural insights
into the mechanism of action of these multiple epigenetic ligands and paves the way for
additional structural exploration to optimize their pharmacological profile. Because of their
multi-(epi)-target features and their action in ex-vivo samples, the indole-based psammaplin A
derivatives are attractive molecules for the modulation of epigenetic disorders.
KEYWORDS
Psammaplin A, epigenetics, enzyme inhibitors, total synthesis, SAR studies
INTRODUCTION
In addition to its structural role in shaping the degree of DNA compaction, chromatin is
1
considered the signal transduction interpreter that regulates all genome functions. The basic unit
of chromatin is the nucleosome, the 146 base-pair stretch of DNA wrapped 1.7 times around an
octameric protein core containing two molecules each of histones H2A, H2B, H3 and H4. The
flexible and highly conserved basic tail regions of the histones can be post-translationally
modified primarily by acetylation, methylation, ADP-ribosylation, phosphorylation,
2
, 3
sumoylation, and ubiquitylation. The covalent modifications of DNA and histone proteins, the
4
,5
majority of which are reversible, control the degree of compaction or relaxation of chromatin
and act as recognition sites for regulatory factors/machineries. Ultimately, these events result in
the regulation of fundamental cellular processes (such as replication, chromatin remodelling,
repair and transcription) through recognition, recruitment and assembly of supramolecular
2
ACS Paragon Plus Environment

Page 3 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
structures (histone-histone and histone-DNA). As the alterations on gene expression are induced
by mechanisms other than changes in DNA sequence, they are termed “epigenetic”. The
enzymes responsible for the most important covalent modifications of DNA and histones with
small chemical marks are DNA methyltransferases (DNMTs), histone acetyltransferases (HATs),
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
histone deacetylases (HDACs), histone methyltransferases (HMTs), and histone demethylases
7
(HDMs).
Emerging evidence suggests that the disruption of the balance of epigenetic networks, and its
8
-14
interaction with the genome and the environment, contribute to several major pathologies.
During onset and/or progression of disease, alterations can occur of both the effectors of
1
5
16-18
chromatin changes (e.g., histone and DNA modifiers) and upstream signalling pathways.
DNA methylation at the cytosine C-5 position within CpG dinucleotide-rich regions (CpG
1
9
20-22
islands) is promoted by the DNMT protein family,
maintenance DNMT1, and the de novo methyl transferases DNMT2 and DNMT3 (a,b and L).
The aberrant expression of DNMT1 induces site-specific hypermethylation and global
which includes in mammals the
2
3
9
, 24-29
hypomethylation of CpG islands, which are biological signatures of cancer.
Methylation of
CpG-islands in the promoter regions of tumour suppressor genes is associated with a silencing
3
0, 31,28
state.
Hypermethylation of DNA promoter and coding sequences is a major epigenetic
mechanism that contributes to cancer progression by causing the inactivation of a number of
32
tumor suppressor genes in a wide range of tumor types, which is the most common molecular
1
2
lesion of cancer cells.
Post-translational reversible acetylation of histone lysine residues regulates gene expression and
3
3
cell cycle progression. Many non-histone proteins also undergo acetylation, and this
modification alters, among other functions, DNA-binding activity, microtubule stabilization,
protein-protein interactions and small-molecule action. At the chromatin level, acetylation is
controlled by the opposite activities of histone acetyltransferases (HATs) and histone
deacetylases (HDACs).
Eighteen mammalian deacetylase enzymes comprise the superfamily of HDACs, which is
conveniently divided in two families that differ in the overall mechanism of lysine deacetylation:
2
+
the Zn -dependent hydrolases (11 members, class I: HDAC1,2,3 and 8; class IIa: HDAC4 and
+
10; class IIb: HDAC5,6,7 and 9; class IV: HDAC11) and the NAD -dependent deacetylases or
3
4-36
sirtuins (7 members: SIRT1-7).
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2+
The Zn -dependent HDAC hydrolases deacetylate the ε-acetylamido lysine by a nucleophilic
2
+
attack of bound water to the Zn -activated carbonyl group, creating a tetrahedral zinc alkoxide
intermediate stabilized by enzyme residues which releases the acetyl group and the lysine
3
7-39
2+
40
product. Class I/II/IV HDAC inhibitors
contain three regions: a Zn -chelating head group
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
hydroxamates, benzamides, short-chain fatty acids, and thiolates, among others, are part of the
38, 39
metal-binding structure
), a linker/spacer that occupies the binding pocket of the native
lysine, and a highly variable cap group that extends along the rim of the pocket. Sirtuins regulate
a variety of cellular functions such as conservation of the genome, longevity and metabolism of
3
6, 41-45
organisms ranging from bacteria to eukaryotes.
An increasing number of reports indicates
13
the implication of multiple epigenetic enzymes in the onset or progression of cancer, among
4
3
them the sirtuins.
The modulation of epigenetic enzymes by small molecules has been correlated with the
4
6, 47
induction of differentiation, cell cycle and growth arrest, and apoptosis in cancer cells.
1
9, 48, 49
8, 13,
Chromatin regulators (“epi-drugs”)
thus represent novel promising anti-cancer drugs.
5
0-52
Indeed, two DNMT inhibitors (DNMTis; 5-aza-cytidine, Vidaza, 5-aza-deoxycytidine,
21, 53
54
Dacogen)
have been approved for the treatment of myelodisplastic syndrome (MDS), and
two HDACis, namely SAHA (suberoylanilide hydroxamic acid, vorinostat) and FK-228 1
(
depsipeptide) have also been marketed for the treatment of cutaneous T cell lymphoma (CTCL).
55
Depsipeptide 1 is a natural product isolated from Chromobacterium violaceum No. 968. Its
approval for therapy demonstrates that natural products continue to be a source of new medicines
5
6-60
or the inspiration for the discovery of new drug leads.
It has been demonstrated that the
disulfide bond of FK228 1 (for other depsipeptides with disulfide bonds in bicyclic structures,
58
such as the spiruchostatins and FR901375, see ) becomes reduced in cells by the activity of
2
+
glutathione to afford a thiolate termed redFK, which then would interact with the Zn ion in the
6
1
active site of HDAC.
A disulfide bond is also present in the dimeric natural product psammaplin A (2, Figure 1), a
representative member of the symmetrical bromotyrosine-cysteine modified tetrapeptides.
6
2-64
Isolated in 1987 from an unidentified sponge
and later, together with 3, from Verongid
6
5, 66
sponges (
and references cited therein), psammaplin A was shown to display antibacterial
67
and antitumor activities. In particular, 2 exhibited significant cytotoxicity against human lung
(
A549), ovarian (SKOV- 3), skin (SK-MEL-2), CNS (XF498), and colon (HCT15) cancer cell
6
6
lines. The dual inhibition of histone deacetylase (HDAC) and DNA methyl transferase
6
5
(
DNMT) enzymes reported for 2 – note that DNMT inhibition by psammaplin A has been
4
ACS Paragon Plus Environment

Page 5 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
68-70
71
72
debated, being reported in
, but not observed in a recent SAR study - captured our interest
given the promising results of ongoing clinical trials with combination therapies using HDACi
2
0, 73
and DNMTi.
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OH
OH
Br
O
N
HN
O
HN
S
H
Br
S
N
O
N
H
S
S
HN
N
O
HO
HO
NH
2
O
O
Psammaplin A
O
Br
1
OH
OH
OH
Br
FK228 (FR901228)
N
OH
O
Br
HN
O
H
N
N
S
NH
S
OH
HO
N
H
S
S
N
Br
O
O
3
N
HO
Bisaprasin
HO
Figure 1. Disulfide-containing natural HDACis.
Inspired by the structure and likely mechanism of action of psammaplin A (2) we recently
7
4
reported the synthesis and characterization of 4a (Figure 2) which, in addition to displaying a
more potent inhibition of HDAC and DNMT than parent 1 it targets also at least one other family
of epigenetic enzymes (SIRT). Here we report on the development of 4a and SAR studies on this
family of epigenetic modulators.
Design and synthesis
The structure of psammaplin A 2 contains a symmetrical tetrapeptide formed by condensation of
two modified 2-bromotyrosine and cystamine units. It was considered that other amide moieties
derived from heteroaryl-containing aminoacids incorporated to the scaffold could improve the
reported HDAC and DNMT inhibitory profile of the parent compound. This assumption was
supported by the discovery of highly potent HDACis with the indole-amide hydroxamic acid
7
5, 76
structure.
Our purpose to undertake SAR studies called for a synthetic approach to these modified
tetrapeptides that could yield analogs (general structure I, Figure 2) at the various positions by
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
slight modifications of the sequence. Initial exploratory studies focused on the variations of the
substituents at the indole ring including the group at N, the oxime functionality, and the size of
the chain connecting the amide to the disulfide. Addressing some of these structural
modifications we expected to additionally get further insights into the mechanism of action of 4a
in its epigenetic inhibitory activities, and for this purpose the terminal group was also chemically
modified.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Br
H
N
X
O
OH
N
O
Z
n
H
N
N
H
S
N
H
S
Y
N
R
O
N
I
N
H
HO
Br
4
a, UVI5008
X = H, halogen, OR'
R = H, Me, 4-Br-Bn
Y = H,H; N-OR'
n = 1-4
Z = S-) , -CH -) , OH, OMe, SMe
2
2 2
Figure 2. General structure of the indole derivatives prepared in the SAR study leading to the
discovery of 4a.
The synthesis of 4a (Scheme 1) is representative of the preparation of the entire collection of
indole-psammaplins. It started with the attachment of the functionalized oxime-ester side chain
7
7
to indole derivative 5a to provide 9a using the method described by Gilchrist. This entails
7
7
reaction of 5a with nitrosoacrylate 6, generated in situ from oxime 7, itself prepared by the
reaction of ethyl bromopyruvate 8 with hydroxylamine hydrochloride. Either an inverse electron-
demand hetero-Diels-Alder cycloaddition or a Friedel-Crafts reaction followed by indole
aromatization are fully compatible with the outcome of the condensation of 5a and 6. Other
analogs 9b-e were similarly obtained (Scheme 1, Table 1) starting from commercial and non-
commercial bromoindoles. The latter were synthesized by application of the Leimgruber-Batcho
7
8
79
methodology or by the Bartoli procedure. N-substituted indoles 5k and 5l were instead
80
synthesized using the protocol described by Stadlwieser.
Minor amounts of the cycloadducts 10 resulting from two-fold condensation at the C3-position
of indole and 7 were isolated and characterized in the case of X = 5-Br (10a, 4%) and X = H
(10b, 7%). X-Ray analysis of compounds 9b and 10b confirmed the E geometry of the oxime
and for 10b the cis-fusion of the indole and oxazine rings (S.I.). The amount of 10 could be
minimized using excess of indole derivative (2 equiv) and base (5.5 equiv) relative to oxime 7.
6
ACS Paragon Plus Environment

Page 7 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CO Et
HO
2
N
H
CO Et
2
X
CO Et
2
N
CO Et
CO Et
2
2
NO
+
Br
Br
X
HO
+
X
6
N
NOH
N
N
7
O
R
N
H
iii
5
a-j
R
ii
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CO Et
10a X = 5-Br
0b X = 5-H
2
9a-l
5
a
1
8
O
i
5
5
k X = 5-Br, R = Me
l X = 5-Br, R = 4-BrBn
Scheme 1. Reagents and conditions: (i) MeI or 4-BrC
6
H
4
CH
2
Br, KOH, DMF, 0 ºC, 30 min (5k, 91%; 5l, 53%) (ii)
CO , CH Cl , 25 ºC, 20 h (9a-l, Table 1).
8, CHCl , NH OH·HCl, MeOH/H O (1:1), 25 ºC, 22 h (86%). (iii) K
3
2
2
2
3
2
2
Initial attempts to couple the oxime indole derivatives 9a with amines using a variety of reaction
8
1, 82
83
conditions (DCC, HOBt,
2
PyBOP, ClCO Et …) were either unsuccessful or produced the
8
4
product in low yields (<50% with DCC, N-hydroxyphtalimide) . Therefore, protection of the
oxime of 9a (Scheme 2, Table 1) as trityl derivative 11a was required before saponification (to
produce acid 12a) and condensation with cystamine 13a via activation as the N-succinimidyl
85
85
ester, a sequence that produced disulfides 14. Cleavage of the trityl protecting group afforded
compound 4a (Table 1).
Following identical sequence, indoles 9b-l were transformed into disulfides 4b-l using either
cystamine 13a or amines 13b-g (Scheme 2, Table 1), some of which were synthesized using the
8
6
general methodology described by Pfammatter. Deprotection of the silyl ether of 14r took
place concomitantly with trityl deprotection. On the other hand, methyl sulfide 4s was obtained
87
4
by reduction of disulfide 4a with NaBH and methylation with MeI (Scheme 2).
8
8
89
The same sequence provided the benzyl and methyl oximes (15b and 15a, respectively,
Scheme 2) following benzylation and methylation of 9a, and along similar steps analogs 17a,b
were obtained from these intermediates.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
CO H
CO Et
2
2
Br
N
N
Br
O
Br
ii
iii
RO
RO
S
N
N
S
N
H
2
H N
2
N
N
H
H
H
2
RO
13a
1
1
5a R = Me
5b R = Bn
16a R = Me
16b R = Bn
17a R = Me
17b R = Bn
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
vi
CO H
CO Et
2
CO Et
2
2
Y
n
H N
N
2
N
N
iii
13
i
ii
TrO
TrO
X
HO
X
X
N
13a Y = -S-) , n = 1
N
R
2
N
R
R
13b Y = -S-) , n = 2
2
1
1
1
3c Y = -S-) , n = 3
2
1
2a-l
9a-l
11a-l
3d Y = -S-) , n = 4
2
3e Y = -CH -) , n = 1
2
2
X
X
13f Y = OMe, n = 1
O
O
iv
13g Y = OTBDMS, n = 1
Y
Y
N
N
H
n
n
H
N
N
N
N
R
TrO
HO
R
14a-r
4a-r
4
4
a X = 5-Br, R = H, Y = S-) , n = 1
2
v
s X = 5-Br, R = H, Y = SMe, n = 1
Scheme 2. Reagents and conditions: (i) K
2
CO
3 2 2 2 2
, CH Cl , TrCl, 25 ºC, 20 h. (ii) LiOH·H O, THF/H O (1:1), 25 ºC,
2
0 h (12a-l, Table 1, combined yield for the two steps); 16a, 84%; 16b, 99%). (iii) i. EDC, NHS, dioxane, 25 ºC, 2
h. ii. Cystamine 13a, Et
3
N, MeOH, dioxane, 25 ºC, 15 h (17a, 80%; 17b, 84%) or amines 13b-f, Et
O, CH Cl , 25 ºC, 2 h (4a-r, Table 1, combined yield for the two steps). (v)
, NaOH, EtOH, 25 ºC, 0.5 h. ii. MeI, 25 ºC, 16 h (4s, 80%). (vi) MeI, Ag O, acetone, 25 ºC, 14 h or BnBr,
3
N, MeOH,
dioxane, 25 ºC, 15 h. (iv) 1M HCl in Et
2
2
2
i. NaBH
4
2
t
KO Bu, DME, 25 ºC, 12 h (15a, 42%; 15b, 72%).
Analogs 20a and 20b, which lack the oxime function, were synthesized by a similar sequence
9
0
using a ZrCl
methyl acrylate followed by hydrolysis and condensation of acids 19a and 19b with cystamine
3a via activation as the N-hydroxysuccinimidyl ester.
4
-catalyzed Friedel–Crafts type alkylation of 5-bromoindoles 5a and 5k with
1
8
ACS Paragon Plus Environment

Page 9 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
Br
OR2
iii
O
Br
Br
i
S
N
H
2
N
N
N
R
R
R1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5a R = H
k R = Me
1
8a R = H, R = Me
20a R = H
20b R = Me
1
2
5
1
8b R = Me, R = Me
1 2
ii
1
1
9a R = H, R = H
1 2
9b R = Me, R = H
1
2
Scheme 3. Reagents and conditions: (i) methyl acrylate, ZrCl
4
, CH
2
Cl
2
, 25 ºC, 5 h. (ii) LiOH·H
2
O, THF/H
2
O (1:1),
N,
2
5 ºC, 20 h (19a, 57%; 19b, 75%; combined yield). (iii) i. EDC, NHS, dioxane, 25 ºC, 2 h. ii. Cystamine 13a, Et
3
MeOH, dioxane, 25 ºC, 15 h (20a, 65%; 20b, 42%).
Table 1. Yields (%) for the compounds of the synthetic sequence of indole-psammaplins.
9
, 12
Yield 9
60
12
73
87
71
73
77
64
59
61
49
37
60
45
4
Yield
70
50
38
30
13
4
a, X = 5-Br, R = H
b, X = H, R = H
a, X = 5-Br, R = H, Y = S-)
2
, n = 1
67
b, X = H, R = H, Y = S-)
2
, n = 1
c, X = 5-F, R = H
87
2
c, X = 5-F, R = H, Y = S-) , n = 1
d, X = 5-Cl, R = H
e, X = 5-I, R = H
42
d, X = 5-Cl, R = H, Y = S-)
2
, n = 1
82
e, X = 5-I, R = H, Y = S-) , n = 1
2
f, X = 5-OMe, R = H
g, X = 5-OBn, R = H
h, X = 4-Br, R = H
i, X = 6-Br, R = H
j, X = 7-Br, R = H
k, X = 5-Br, R = Me
l, X = 5-Br, R = 4-BrBn
51
f, X = 5-OMe, R = H, Y = S-)
g, X = 5-OBn, R = H, Y = S-)
2
2
, n = 1
, n = 1
75
20
54
41
30
71
32
19
65
2
h, X = 4-Br, R = H, Y = S-) , n = 1
43
i, X = 6-Br, R = H, Y = S-)
j, X = 7-Br, R = H, Y = S-)
k, X = 5-Br, R = Me, Y = S-)
2
, n = 1
43
2
, n = 1
68
2
, n = 1
68
2
l, X = 5-Br, R = 4-BrBn, Y = S-) , n = 1
2
m, X = 5-Br, R = H, Y = S-) , n = 2
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
n, X = 5-Br, R = H, Y = S-)
o, X = 5-Br, R = H, Y = S-)
2
, n = 3
42
38
82
37
37
80
2
, n = 4
2 2
p, X = 5-Br, R = H, Y = CH -) , n = 2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
q, X = 5-Br, R = H, Y = OMe, n = 1
r, X = 5-Br, R = H, Y = OH, n = 1
s, X = 5-Br, R = H, Y = SMe, n = 1
Biological effects of psammaplin A derivatives in cellular cancer models
Following the confirmation of the epigenetic activities of psammaplin A 2 using in vitro assays
7
2
on isolated enzymes and cellular experiments, we undertook the search for a more potent
analogs, with a focus on the β-indole-α-oximinoamido scaffold. The series of indole derivatives
synthesized (Schemes 2-3) aimed at understanding the role of structural elements and functional
groups in the modulation of different epigenetic enzymes and at the identification of the
pharmacophore(s) of these inhibitors. We also addressed whether (i) the dimeric structure or just
the monomer are necessary for their biological activity, (ii) if the disulphide bridge is
dispensable, and (iii) if an uncleavable dimeric compound displays similar activities.
Using the U937 acute myeloid leukemia (AML) cell line, the activity readouts systematically
WAF1/CIP1
investigated were cell cycle arrest and p21
induction, induction of differentiation,
acetylation of histone H3 and, for HDAC6 inhibition, the levels of α-tubulin acetylation.
As shown in Figure 3B, after 24 h stimulation with the analogs at 5 µM, derivatives 4a and 4n
showed a greater induction of cell death than psammaplin A 2 (Figure 3D). To confirm these
biological activities, cell cycle analyses were performed after 24 h of induction with psammaplin
A 2 and its derivatives (Figure 3A and C). The pro-apoptotic activities of psammaplin A 2 and
analogs 4a and 4n were also validated in the solid cancer cell lines derived from breast (ZR75.1,
data not shown) and prostate (LnCap; data not shown) tumors. Cell death was observed for
compounds 4i, 4j, 4k, 20a, 4c-4g (see Figure 3 D and 3F), but lower number of cells underwent
apoptosis compared to 4n. Some compounds induced changes in the relative percentage of cells
in the different phases of the cell cycle (Figure 3C and 3E).
1
0
ACS Paragon Plus Environment

Page 11 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
%
&
$'()*#+,
9
'
8
7
"
6
5
4
4
4
4
4
4
4
:;5
7
4
T
:0
:
;&
!
!
!
!
"4
64
54
&4
4
!
T
!
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
!
!
!
!
!
!
T
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
T
T
!
!
!
T
!
!
!
!
!
!
!
!
!
&4
4
!
T
T
T
T
T
T
T T
T
!
!
!
!
23# 1$
"/ "( "+ ") "*
"- ", "% "# "$
".
"/ "( "+ ")
"*
"- ", "%
"# "$
&'(
23# 1$
"
.
0/+/
&'(
0
/+/
!
"
9
4
;5:
0
:
"
6
6
4
'4
84
;
&
T
T
7
4
T
T
T
T
T
7
"
6
5
&
4
4
4
4
4
4
T
T
T
T
T
57
54
&7
&4
7
TT
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T T
"?
T
T
T
T
4
5
"D
"?
"C
23#
1$
5
&
"/
54/
54(
23#
1$
"D
"C
0
/+/
&"/
54/ 54(
0
/+/
#
$
;
5
94
'4
T
0
"
6
6
5
4
;
&
T
7
8
7
"
6
5
4
4
4
4
4
T
T
T
4
7
T
T
T
T
T T
T
T
T
T
54
&7
&4
7
T
T
T
T
T T
T
T
T
T
T
T
T
T
T
T
T
&4
4
T
T
T
4
3
#
"=
2
1$
"2
"A
">
"E
3#
"=
&'/
2
1$
"2
"A
">
"E
&'/
0
/+/
0
/+/
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 3. Analysis of the effects of cell cycle and apoptosis of U937 cells treated with indole-
based psammaplin A analogs. A,C,E. Cytofluorimetric cell cycle analysis after treatment with
the compounds at 5 µM for 24 h. B,D,F. Apoptosis analysis after treatment with the compounds
at 5 µM for 30 h. MS-275 and SAHA (5 µM) were used as controls. Error bars represent
standard deviation (SD) of biological triplicates.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The parent compound formally derived from tryptophan (4b) displayed lower pro-apoptotic
activity than the C5-bromo-indole derivative (4a), but changing the substitution from C5 to C6
(
4i) or C7 (4j) had only a minor effect on cell cycle arrest and apoptosis (Figure 3A and 3B), and
rather increased the induction of differentiation measured by expression of the CD11c marker
Figure 4A).
(
1
2
ACS Paragon Plus Environment

Page 13 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
!
*
"#$%&'()
50
40
T
<:
!"+=//+"
!
"+=//+"
<
9
:
T
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
2
1
0
0
0
0
;
T
T
;9
T
T
/
:
T
T
T
/9
:
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
9
-6 -1 -3 -2 -'
-4 -# -% -. -&
+(. 8&
-5
;
-,
/01
+(. 8&
-?
->
7
636
/-6
;96 ;91
7
636
+
,
@
:
-
9
!
"+=//+"
B+CD1
E.?
9
9
T
%;/
<9
;9
F<6+
F-
T
T
T
(.
+
/
9
T
T
T
-#
-4
-5
-6
-1
8&
T
7636
T
9
(
.
-$
+
8&
-+ -= -*
-A
/
06
7636
"
'
B+CD1
E.?
;/
B+CD1
E.?
%
%;/
F<6+
F-
E.?
F<6+
(.
-'
-2
+
-3
-%
-&
-.
/
01
GH"I*=
7636
;
(.
+
-,
->
-?
8&
;91
;96
/-6
$
B+"CD1
E.?
%;/
E.?
F<6+
GH"I*=
-$
-*
-=
-+
(.
+
/
06
7
636
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 4. A-C. Analysis of cell differentiation (A-C) and Western blot analysis (D-G) of
functional readouts of indole-based psammaplin A analogs. A-C. Effects on granulocytic
differentiation by cytofluorimetric analysis of CD11c expression after treatment of U937 acute
myeloid leukemia cells with indole-based psammaplin A analogs at 5 µM for 30 h; SAHA and
MS-275 (5 µM) were used as positive control. The data represent the average value of
independent triplicates. Error bars represent standard deviation (SD) of biological triplicates. D-
G. Western Blot analysis of p21, histone H3 acetylation and α-tubulin acetylation in U937 cells
after treatment with the indicated compounds at 5 µM for 24 h. Total ERKs were used to
normalise for equal loading of total protein extraction; total H4 or Ponceau red staining was used
to normalise for equal loading of histone extracts. SAHA and MS-275 (5 µM) were used as
controls.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4A-C shows that most compounds (4n, 4i, 4j, 4s, 4k, 4f, 4g and 4m) derived from 4a and
4
b induced the expression of CD11c (pro-differentiation surface marker), which is indicative of
the differentiation of U937 cells to granulocytes.
Cell-based assays confirmed that the presence of one or two sulphur atoms, and/or the disulphide
bridge and/or the in cell cleavage thiols derived from 2 and 4a are required for cell cycle arrest,
induction of apoptosis and differentiation. Interestingly however, the methylthioether 4s is
inactive at inducing apoptosis (Figure 3B) and G2M phase accumulation but it induces cell arrest
WAF1/CIP1
at G1 (Figure 3A) and the corresponding enhanced expression of p21
, as shown in
Figure 4A and 4E. Further studies will be necessary to define the molecular pathways that are
apparently differently affected by 4a and 4s.
Indole psammaplin A derivatives induced α-tubulin acetylation (a marker of HDAC6 inhibition),
in some cases at levels similar to psammaplin A (2) as shown by Western Blot analysis (Figure
4
D-G). Moreover, the majority of compounds increased histone H3 acetylation levels, with 4a
being clearly more efficient (Figure 4D). Note that despite its action on α-tubulin acetylation, 4n
did not show hyperacetylation of histone H3, suggesting a different mechanism of action.
A correlation between hyperacetylation of histone H3 and expression of p21 protein exists.
Indeed, as shown in Figure 4D-G, compounds for which H3ac level is greater (4b, 4a, 4i, 4j, 4s,
4k, 20a, 4c-g), showed also a stronger induction of p21 and the greatest percentage of cell death.
With regard to the effect of other substituents at the C5-indole ring position (OMe, 4f; OBn, 4g),
including other halogens (F, 4c; Cl, 4d; I, 4e), the biological evaluation indicates that the original
scaffold is quite lenient at that position, since most of the C5-substituted indole derivatives
1
4
ACS Paragon Plus Environment

Page 15 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
showed the ability to induce p21, histone H3 and α-tubulin acetylation levels (Figure 4G) and,
with the exception of 4c, also p16 (see Figure 7, below).
Finally, the N-methyl derivative 4k was found to induce apoptosis, change the cell cycle
distribution, induce the acetylation of histone H3 and α-tubulin, and overexpress p21 (at levels
comparable to 4a), but the larger N-p-bromobenzyl analog 4l is quite inactive in these biological
assays, showing a weak induction of p21 after 24 h of treatment of U937 cells (Figure 3C and D;
Figure 4B and F).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Indole psammaplin A derivatives induce apoptosis in AML patient's blasts
The intriguing anticancer potential of these indole-based psammaplin A derivatives prompted us
to test their activity in ex vivo AML patient blasts. As shown in Figure 5 B, 4a induced cell cycle
block and apoptosis measured by FACS in AML #102 ex vivo blasts, similarly to psammaplin A
2. Known HDACis such as MS-275 (Entinostat) and SAHA (Vorinostat) displayed activities
similar to 4a. To verify the action of the selected derivatives on AML patient blasts, the activity
of compounds 4a, 4i, 4j and 4s was tested at 5 µM. As shown in Figure 5A and C, all these
derivatives showed pro-apoptotic activity in two different samples of AML patient's blasts (# 108
and #109). Methylthioether 4s showed a weak induction of apoptosis confirming data obtained
with the U937 cell line.
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
)
*
'
()
*
"+,
-
*
',6781409:47&!%
&
',6781409:47&!"
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
&!!
&
!!
T
T
T
T
T
T
T
T
T
T
T
T
T
%
!
!
T
%!
T
T
T
T
$
$
!
T
T
T
#!
"!
!
#!
"!
!
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
45
T
T
T
T
T
3
,. -121
#1
#0
#/
#.
345
,.
-121
"
#1
#:
+
&
!!
'
,6781409:47&!B
T
T
T
T
T
T
!
T
%
$
#
"
T
T
!
!
!
!
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
3
45
,. -121 #1
#0
#/
#.
Figure 5. A-C. Cell cycle analysis of ex–vivo patient's blasts treated with selected indole-
based psammaplin A analogs. Analysis of cell cycle and apoptosis carried out with selected
compounds at 5 µM in different AML patient blasts (indicated as numbers). SAHA and MS-275
(5 µM) were used as controls. Error bars represent standard deviation (SD) of biological
triplicates.
Inhibition of epigenetic enzymes by indole-psammaplin A derivatives
1
6
ACS Paragon Plus Environment

Page 17 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
Having established the activity of the indole psammaplin A analogs in biological assays, the
active compounds were also tested in vitro (on enzyme-based assays). Because acetylation levels
of many targets were found modified after the treatment with selected compounds (see Figure
4
D-G), those leading to the expression of p21, acetylation of histone H3 and acetylation of α-
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
tubulin were tested (at 5 µM) as inhibitors of HDAC1 and HDAC4 enzymes.
As shown in Figure 6A, compounds 4a, 4i and 4j showed inhibitory activities (around 60-80%)
against HDAC1. The same compounds were also active against HDAC4 (see black and grey bars
in Figure 6A) although with lower potency (40-60% for compounds 4a and 4j). Other analogs
followed the same trend (better inhibitors of HDAC1 than HDAC4) but were less effective
inhibitors of both enzymes, with percentages of inhibition ranging from 20% (HDAC4) to 40%
(HDAC1).
Compounds 4n and 4o were instead rather inactive against HDAC1, but acted as inhibitors of
HDAC4 (in particular compound 4n, for which the percentage of inhibition of HDAC4 is around
6
0%). Compounds 4k and 20a (see cell-based data in Figures 3C, 3D, 4B and 4F), showed
inhibition of HDAC4 (Figure 6B) and to a lesser extent of HDAC1 (4k).
The series of halogen and ether analogs at C5 (4c-4g) reduced the activity of HDAC1 and
HDAC4 in a variable way (different for each derivative) but in general showed a greater activity
against HDAC1 than HDAC4 (Figure 6C). The inhibition induced by the methyloxime 17a
(Figure 6C) was similar to that of 4g.
Likewise, the activity of mechanistically relevant analogs on other epigenetic enzymes was
analyzed, in particular as inhibitors of DNMTs and SIRTs, in order to relate structure and
activities. Using recombinant DNMT1 enzyme and labeled S-adenosymethionine, the cofactor
for methyl transfer (Figure 6D), it was shown that the desoximino compound (20a) was inactive,
but the methylated oxime (17a) was as active as 4a, and the 5-OBn derivative (4g) exhibited
even greater activity than 4a and also than the known DNMT1 inhibitor RG-108.
As shown in Figure 6 E, 4a displayed the greatest potency of the series on SIRT1 inhibition, with
the presence of C5-Br being less important (4b) than other modifications. Whereas the N-Me-
indole derivative 4k proved inactive, the N-p-bromobenzyl derivative 4l and the analog with the
all carbon connector between the amides (4p) showed weak activity. No SIRTi activity was
noted for the corresponding monomers (4s, 4r, 4q).
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
!"#$%&'(
)
*
4
56738
3//
!
!
4567"8
!
!
!
2
1
"
/
/
/
456738
3
//
!
T T
!
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4567"8
2
1
"
0
/
/
/
/
/
!
!
!
!
!
!
T
T T
!
T
!
!
0/
/
!
T
,
-.
")
"(
"'
"&
"%
"$
"#
,-.
"=
0/(
*(+(
*(+(
+
,
456738
5HI!38
3//
!
!
4567"8
3// !
!
!
!
!
!
2
1
"
0
/
/
/
/
/
2/
!
!
!
!
!
!
!
!
!
1/
!
!
"/
!
!
!
!
!
!
!
0/
/
-
.
",
"?
"@
":
"A
,-.
CG3/2
") "( "' "F "# "= 0/( 3>( "A
,
3>(
*(+(
&
!
30/
3//
2/
*
BC!3
!
!
!
!
!
!
!
!
1/
"
/
/
/
!
0
,
-. "( ") "' "F "E ". "# "= "D
Figure 6. Modulation of epigenetic enzymes by indole-based psammaplin A analogs. A-C.
U937 cell-based HDAC1 and HDAC4 assay carried out with the indicated compounds at 5 µM.
1
8
ACS Paragon Plus Environment

Page 19 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
D. In vitro DNMTs assay carried out in U937 extracts treated with the indicated compounds at 5
µM. E. In vitro human recombinant SIRT1 assays with the indicated compounds at 5 µM.
SAHA and RG-108 (5 µM) were used as controls. Error bars represent standard deviation (SD)
of biological triplicates.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Indole psammaplin A analogs induce changes of acetylation and expression levels of
epigenetic targets
Compounds that showed inhibitory activity on DNMT1 and SIRT1 enzymes were also tested in
cell-based assays in order to confirm (by Western Blot) if the in vitro activities correlated with
changes of acetylation and expression states of selected epigenetic targets.
In Figure 7, the reactivation of p16 expression was evaluated after treatment of U937 cells for
24h with selected compounds at two different concentrations (5 and 0.5 µM). Analogs 4a, 17a,
4
n, 4p and 4s proved to be most active since at 0.5 µM they induced the expression of p16
74, 91
(which is causally linked to DNMT1 inhibition
). The level of DNMT1 protein remained
roughly unchanged (Figure 7, panel A). Compounds 4a, 4g, 4r and 4k only induced p16
reactivation at 5 µM.
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
!
"#$%&'(
)
/
%
8-(
,9(
/+
/
(
/0
3,4
#6
3
,4
5
5
#6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
-12
2
-12
2
-12
2
-12
+B
2
-12
2
-12
+B
/+
,9(
8-(
/%
/
(
/0
:;<",
:;<",
5#6
5
#6
-
12
2
-12
2
+B
-12
2
-12
/#
2
-12
2
-12
/?
2
+B
/
;=
/3
/>
/6
3
,4
3
,4
#6
+B
5#6
5
2
-12
2
-12
2
-12
+B
2
-12
2
-12
2
-12
/
;=
/3
/>
/
#
/6
/?
:;<",
:;<",
5
#6
5
#6
2
-12
2
-12
2
-12 +B
2
-12
2
-12
2
-12 +B
*
/(
/0
32@A@.8(!
5
#6
2
-12
2
-12 +B
/%
8-(
,9(
/+
3
2@A@.8(!
5
#6
2
-12
2
-12
2
-12
2
-12 +B
/;=
/3
/>
3
2@A@.8(!
#6
+B
5
2
-12
2
-12
2
-12
/?
/
#
/6
2
3
2@A@.8(!
#6
+B
5
2
-12
-12
2
-12
2
0
ACS Paragon Plus Environment

Page 21 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
Figure 7 Panel A: Western Blot analysis of p16 and DNMT1 levels after treatment of U937
cells with the indicated compounds at 5 and 0.5 µM for 24 h. Total ERKs were used to normalise
for equal loading of total protein extraction; SAHA, RG108 (5 µM) SGI1027 (5 µM) and ALX
(
1 µM) were used as controls. Panel B: Western Blot analysis of p53 acetylation on lysine 382
after treatment of MCF7 cells with the indicated compounds at 5 and 0.5 µM for 24 h. Total
ERKs were used to normalise for equal loading of total protein extraction. SAHA, RG108 (5
µM) SGI1027 (5 µM) and ALX (1 µM) were used as controls.
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
When compared to the data shown in Figure 6D, Figure 7 indicates a strong correlation between
in vivo and cell-based assays for compounds 4a, 4n, 17a and 4g. Surprisingly compounds 4p, 4s
and 4k were inactive on in vitro DNMT1 assay (Figure 6, Panel D) but induced p16 reactivation
(see Figure 7A).
The level of acetylation of SIRT1 epi-targets (as readout for SIRT inhibition) was also
determined by treatment of U937 and MCF7 cells with indole psammaplin A derivatives at two
concentrations. Figure 7B shows increased acetylation of p53 on residue K382 (a well-known
92, 93
SIRT1 target
) after treatment with selected compounds. Whereas some of those showed a
dose-response trend (4k, 4r, 4g and 20a), others (4s, 4a and 4b) were active only at the highest
concentration (5 µM). Note that 4n is the only compound of the series (Figure 6E) that neither
increased p53K382ac levels nor inhibited SIRT in cell-based assays.
9
4, 95
96
The reactivation of other reported SIRT1 targets (acetylation of H3K9
and H3K56 ) was
tested with the indole psammaplin A derivatives. Figure 8A,B shows that in general a correlation
exists between concentration and level of acetylation of H3K9ac (Figure 8A) for the surveyed
compounds. Compounds 4r and 4q, which lack the disulfide bond, did not show SIRT1
inhibition as demonstrated by their effects on H3K9 (Figure 8A) or H3K56 acetylation (Figure
8B), whereas the other analogs showed an effective pro-acetylation action, in particular 4k, 4n,
4
p, 4s, 4l.
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
!
"#$%&'(
)
0
34
05
06
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
01
02
;<=>1!
;<=>1!
;0
;0
+A
7
*87
7
*87 +A
7
*87
7
*87
7
*87
0(
0#
09
0:
0
%
.*1
)/1
;<=>1!
;<=>1!
;0
;0
+A
*87 +A
*87
7
*87
7
*87
7
*
87
7
*87
7
*87
7
7
*
0
1
02
;
<=7@1!
;0
7
*87
7
*87 +A
0#
09
0:
;<=7@1!
;0
7
*87
7
*87
7
*87
+A
034
05
06
;
<=7@1!
;
0
7
*87
7
*87
7
*87
+A
.*1
)/1
0
(
0%
;<=7@1!
;0
7
*87
7
*87 *87
7
*87
7
+A
Figure 8: Western Blot analysis of H3 acetylation on lysine K9 (panel A) and lysine K56 (panel
B) after treatment of U937 cells with the indicated compounds at 5 and 0.5 µM for 24 h. Total
H4 was used to normalise for equal loading of histone extracts. SAHA, RG108 (5 µM) SGI1027
(5 µM) and ALX (1 µM) were used as controls.
2
2
ACS Paragon Plus Environment

Page 23 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
SAR studies on indole psammaplin A derivatives
7
4
The biological results indicate that the indole derivatives of 2 are more potent modulators of
72
epigenetic enzymes than the natural product. . The 5-bromoindole analog 4a has been shown to
highly efficiently induce cancer cell-selective death in a variety of models, and these activities
were also demonstrated in several human tumor xenografts and genetic mouse models of human
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
4
breast cancer in vivo. Mechanistic studies proved that the anticancer activity of 4a involves
activation of death receptors and ROS production in a mutually independent manner.
Importantly, cell death is efficiently induced in cells mutated or deficient for p53, Bcl-2
7
4
modifying factor, BMF, and/or TNF-related apoptosis inducing ligand, TRAIL.
From the biochemical and cellular characterization of the analogs, it was concluded that
positional isomers at the bromoindole ring such as 4i and 4j still showed cell cycle block,
apoptosis induction (although to a lesser extent than 4a), and CD11c differentiation of U937
cells (Figures 3A and B, 4A and E). These derivatives increased α-tubulin acetylation and
induced p21 expression similarly to 4a and psammaplin A (2). On the contrary, protection of the
oxime function either as a trityl (14a) or benzyl derivative (17b) abolished activity (Figure 3A-D
and Figure 4A, B, E, F). Furthermore 4p (Figure 3A and 3B; Figure 4A and 4E) with a
hydrocarbon linker was inactive, suggesting that the activity of these analogs must reside on the
thiolate formed upon disulfide cleavage. Finally, incorporation of an alcohol functional group in
place of the putative thiol as in 4r (Figure 3A and B; Figure 4A and E) afforded a biologically
inactive compound that kept a residual ability to induce α-tubulin acetylation. All together, the
results confirm that the series of tetrapeptide disulfide dimers are pro-drugs that get activated by
reduction to afford the thiols, which is the functional group required for biological activity.
Capping the thiol as the methyl sulfide 4s afforded a derivative with a distinct biological profile,
since it blocked the cell cycle in G1 and induced p21 expression without increasing the
differentiation of U937 cells. We note that its apoptogenic activity was only occurring in primary
leukemic cells from AML patients but not in established U937 cells in culture. It will be
interesting to study 4s and other active compounds of this series in genetic mouse models for
74
different types of (solid) human cancers, as reported for 4a.
HDAC1 and HDAC4 enzymatic inhibition studies revealed that the C5-Br could be replaced by
other halogens (F, Cl, I) or oxygen substituents (OMe, OBn), since these analogs show roughly
equal potency, but the absence of a substituent was detrimental. Positional isomers of the
bromoindole (4i-4j) showed reduced potency as HDAC inhibitors in enzymatic assays, as did the
indole-N-methyl derivative (4k), in particular of HDAC4 (Figure 6A-C). Derivatization of the
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
oxime as an alkyl derivative was tolerated for the small methyl group (17a), but not for the O-
benzyl (17b) or O-trityl (14a) analogs (data not shown). The oxime group is dispensable (20a) if
the indole N-H group is present, as the N-Me derivative (20b) lacks the activity (as confirmed by
in vivo assays). Figure 9 summarizes the SAR study described in this work.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
small N-substituents
alkoxy groups
of indole
at C5 tolerated
disulfide
essential
Br
H
N
OH
N
O
H
N
S
N
H
S
Br at C4 and C7
detrimental
O
N
N
H
HO
Br
UVI5008
2
-carbon chain
halogens at C5
tolerated
optimal
large alkoxy groups
at oxime detrimental
loss of oxime
tolerated
Figure 9. Summary of SAR studies on indole-psammaplin A scaffold for HDAC inhibition.
Regarding DNMT inhibition, Figure 6D shows the beneficial effect of the methylated oxime
(
17a) relative to 4a and, more importantly, the presence of a O-Bn group in place of the bromine
4g). The data also concludes that the dimeric structure is required for efficient inhibition, as
(
monomeric analogs were inactive on DNMT1. We surmise that the disulfide could react with the
enzyme after ocupation of the SAM binding pocket, perhaps undergoing capture by the invariant
9
7
Cys81 active site residue in a disulfide exchange reaction that might be reversible in vivo and
therefore transient. Compounds that inhibited DNMT1 (4a, 4g, 17a) showed the ability to
reactivate p16, an effect that was also shown by other analogs (4n, 4p, 4s; Figure 7A) but less
efficiently (with the exception of 4b).
+
We also studied the structural requirements for the inhibition of the SIRT NAD -dependent
7
4
deacetylases by compound 4a. None of the analogs of 4a assayed with SIRT1 showed
equipotency to the parent compound (except 4b and 4l, Figure 6E), but the negative result
appears also to suggest that the dimeric structure is required for the activity of 4a, since the
corresponding monomers (4s, 4r, 4q) proved inactive in our assay system, and both the N-p-
bromobenzyl derivative 4l and the analog with the all carbon connector between the amides (4p)
showed weak SIRT1 inhibitory activity. The modulation of SIRT1 targets (acetylation of p53
K382, H3K9 and H3K56, the latter also linked to HDAC inhibition) was clearly consistent with
2
4
ACS Paragon Plus Environment

Page 25 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the enzymatic assays for 4a. The same correlation was noted for compounds 4p, 4q and 4r, for
which the very weak inhibitory enzymatic activity (Figure 6E) is consistent with low levels of
acetylation of H3K9 and H3K56 (Figure 8A). Complex readouts on the expression levels of the
different SIRT1 targets hindered further SAR studies although the disulfide bond appears to also
be necessary for SIRT1 inhibition.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The entire data set suggests that the series of tetrapeptide disulfide dimers show greater activity
in cell than in vitro assays (see Figure 6D and 6C, cf. Figures 7 and 8), perhaps due to their pro-
drug nature. The inhibition of SIRT and DNMT enzymes is clearly indicated by the analysis of
the levels of acetylation of sirtuin targets (p53 K382, H3K9 and H3K56) or by the reactivation of
methylated genes (p16), respectively. The discrepancy between data from in vitro and in cell-
based assays could be understood and justified by the role of disulfide group present in the
molecular scaffold, which can be reduced (or undergo disulfide exchange, for example with
glutathione) upon cell uptake. This assumption is consistent with the weak activity displayed in
both settings by alcohol 4r. Other thiol-disguised pro-drugs, such as the disulfide FK228 1 and
octanoyl derivative largazole, have also shown contrasting enzymatic potencies when compared
9
8, 99
to the corresponding reduced forms.
Given their different enzymatic mechanism, sirtuins are not modulated by the inhibitors of Class
I, II and IV HDACs, such as suberoylanilide hydroxamic acid (SAHA) and trichostatin A (TSA)
2
+
which have Zn chelating groups. It is tempting to structurally relate the SIRT inhibition
displayed by 4a to the haloindole motif, which is found in related cyclic indole derivatives
5
9
(EX243, GW5074, kenpaullones) that act as adenosine mimetics.
Mechanistic insights into the multiple epigenetic activities of the indole-based psammaplin
A derivatives
5
9
With the data obtained thus far and the precedents on other sulfur containing inhibitors, we
surmise that the HDAC inhibition exhibited by these compounds is a consequence of disulfide
reduction to the corresponding thiol-capped butene moiety, which likely binds the active site
2
+
Zn . Thiols such as 21 derived from the potent analog 4a (Scheme 4) feature the tripartite
structure characteristic of most effective HDAC inhibitors, as it contains an HDAC rim-
recognition element attached to an active site binding/inactivating group via a linker. The
increased levels of native disulfide reductants (i.e., glutathione, thioredoxin and thioredoxin
reductase) found in many cancer cells likely render such cells particularly susceptible to the
action of 4a and analogs. Their cleavage would occur only following cellular entry after
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
encountering a high glutathione concentration (typically 15 mM intracellular compared to 15 µM
1
00
extracellular).
Molecular Modeling of 21 into the active site of HDAC8 shows the expected pose with the
2
+
chelation of the thiol group to the Zn ion, and the additional stabilization by formation of a
hydrogen bond between the indole N-H and Asp101. This model is also consistent with the
inhibitory activities of the C5-indole analog series and with the lack of activity of monomers
lacking the thiol-chelating unit (Figure 10).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
01
Figure 10. Proposed docking pose for 4a-derived thiol 21 in HDAC8. The Cα trace of the
enzyme is displayed as a ribbon, colored in green. The side chains of His142, His143, Asp178,
Asp267, His180, Met274, Tyr306 are shown as sticks, with carbon atoms colored in green. The
2+
Zn at the catalytic site is shown as a magenta sphere. The inhibitor is displayed also as sticks,
but with carbon atoms colored in gray and the bromine atom colored in brown. The ab initio
1
01
calculated structure of 4a-derived thiol 21 was docked into the structure of human HDAC8
after removal of the trichostatin A (TSA) ligand. The highest scores using an automated docking
2+
method for the interaction with the Zn ion in the active site was validated by the GRID maps.
In these solutions, the thiolate group chelates the metal, the linker domain occupies the channel
and the bromoindole is stabilized through interaction with Tyr100 and Phe152 at the rim of the
active site entrance. This positioning facilitates the formation of a hydrogen bond between the
indole N-H and Asp101. The complex was refined using energy minimization and its dynamic
behaviour was simulated using unrestrained MD (see S.I.).
2
6
ACS Paragon Plus Environment

Page 27 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In support of the mechanism of action of these indole-based psammaplin A analogs,
74
pharmacokinetic studies on lead compound 4a led to the detection by HPLC/MS of the mixed
disulfide 22, a glutathione adduct presumably formed by in-cell modification of the scissile
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
disulphide bond as indicated in Scheme 4.
H
HS
N
O
H
N
Br
H N
CO H
OH
2
2
N
H
N
NH
O
H
Br
OH
CO
2
H
O
HO
O
N
N
N
HS
H
N
HN
O
S
N
H
S
+
NH
N
O
glutathione reductase
S
S
Br
HO
NH
4
a, UVI5008
O
H
N
Br
2
1
H N
CO H
2
2
N
CO H
H
O
2
22
Scheme 4. In vivo conjugation of 4a to glutathione.
5
5
In addition to the disulfide present in other depsipeptides such as FK228 (1), and in
58
102
103
psammaplin A (2) and to the octanoyl thiol ester of depsipeptide largazole, the S-acetyl, S-
1
04
105
isobutyryl and S-2-methyl-3-phenylpropanoyl thiol esters have also been developed as thiol
surrogates in HDACis. Mixed disulfides have been used in lieu of the symmetrical dimeric
1
06
107
structures. All the above ligands are considered as prodrugs that most likely release the
1
08
thiolate group upon reaction with thioesterases or by thiol-disulfide exchange induced by the
addition of sulfhydryl groups of compounds present in the reductive environment of the cell such
as glutathione. In addition, other sulfur-containing functional groups employed as head group of
1
09
110
HDACis have been described, among them trithiocarbonates,
mercaptoacetamides,
1
11
112
mercaptoketones
and methylsulfoxides.
The broad variety of functional groups further
+2
attests to the efficient and energetically favorable chelation of Zn ion by sulfur-containing
4
0
functional groups.
Conclusion
To summarize, the scaffold of the psammaplin A natural product is amenable to synthetic
reengineering without loss of epigenetic modulatory activity. In particular the o-bromophenol
unit of psammaplin A 2 has been replaced by the more chemically robust and drug-friendly
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
indole fragment. A series of indole-psammaplin A constructs have been generated in a short
synthetic sequence and their biological evaluation has led to a better understanding of the
structure-activity relationship of these epigenetic inhibitors (see Table 2 for a summary of
enzymatic and cellular activities). Notably, the incorporation of the 5-bromoindole unit in 4a
replacing the o-bromophenol of 2 not only improved the potency of the parent compound with
regard to the already reported HDAC and DNMT dual epigenetic inhibitory profile, but
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
interestingly added another activity, the inhibition of the NAD -dependent SIRT deacetylase
7
4
enzymes. In vivo pharmacokinetics revealed that this compound is a pro-drug that gets rapidly
74
transformed into the glutathione conjugate. In keeping with the concept of targeting several
pathways, it is encouraging that a large number of phase I/II clinical trials are currently exploring
the combination of an HDAC together with the DNA methyltransferase inhibitor 5-aza-2’-
deoxycytidine (Decitabine) (http://clinicaltrials.gov). Indeed, the simultaneous block of at least
three epigenetic targets offers a valid alternative to combination treatments as well as a unique
59
strength to ensure delivery of an ‘all in one’ drug to the cells.
2
8
ACS Paragon Plus Environment

Page 29 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
!
"#$%&'(&)*++",-&./&0.+1.*23)&"0456545%)
!
"##$%&%#" '"()* !+,,% -%")&#()./0 1234"55./0 6'-!5$ 789:,$ ';<:,
=
,>(
,?(
,?@
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
=A(
A@
=
>(
>@
>%
>
>
+
"
>B
>C
>*
>.
>D
>E
>#
>F
>0
>
>
/
3
>G
>4
>5
Table 2: Summary table of compounds activities. Black boxes show, for each molecule, positive
response to the tests. Column entries from left to right indicate: variations in cell cycle
distribution (Cell cycle); induction of cell death (Death); induction of granulocytic
differentiation (CD11c); variation of the acetylated state of proteins (Acetylation); expression of
key and/or silenced proteins (Expression); in vitro activity against HDAC1/4 (HDACs); in vitro
activity against SIRT1 (SIRT1); and in vitro modulation of DNMT1 activity (DNMT1). The
experimental details are reported in the Materials and Methods section and the numerical values
in Figures 3-8.
A major consequence of 4a-mediated anti-cancer effects involves activation of different
7
4
pathways of apoptosis in the cancer cells due to synergism between these inhibitory activities.
1
13
A balanced modulation of several epigenetic targets makes designed multiple ligands (DMLs)
5
9
a promising class of anticancer drugs with unanticipated therapeutic potential. The fact that 4a
2
9
ACS Paragon Plus Environment