Synthesis and Biological Investigation of Phenothiazine-Based Benzhydroxamic Acids as Selective Histone Deacetylase 6 Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.8b01090

The phenothiazine system was identified as a favorable cap group for potent and selective histone deacetylase 6 (HDAC6) inhibitors. Here, we report the preparation and systematic variation of phenothiazines and their analogues containing a benzhydroxamic acid moiety as the zinc-binding group. We evaluated their ability to selectively inhibit HDAC6 by a recombinant HDAC enzyme assay, by determining the protein acetylation levels in cells by western blotting (tubulin vs histone acetylation), and by assessing their effects on various cancer cell lines. Structure-activity relationship studies revealed that incorporation of a nitrogen atom into the phenothiazine framework results in increased potency and selectivity for HDAC6 (more than 500-fold selectivity relative to the inhibition of HDAC1, HDAC4, and HDAC8), as rationalized by molecular modeling and docking studies. The binding mode was confirmed by co-crystallization of the potent azaphenothiazine inhibitor with catalytic domain 2 from Danio rerio HDAC6.

Full text of DOI:10.1021/acs.jmedchem.8b01090

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE
Article
Synthesis and Biological Investigation of Phenothiazine-Based
Benzhydroxamic Acids as Selective Histone Deacetylase 6 (HDAC6) Inhibitors
Katharina Vögerl, Nghia Ong, Johanna Senger, Daniel Herp, Karin Schmidtkunz, Martin
Marek, Martin Müller, Karin Bartel, Tajith Baba Shaik, Nicholas J. Porter, Dina Robaa, David
W. Christianson, Christophe Romier, Wolfgang Sippl, Manfred Jung, and Franz Bracher
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b01090 • Publication Date (Web): 15 Jan 2019
Downloaded from http://pubs.acs.org on January 16, 2019
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 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 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.
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 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Synthesis and Biological Investigation of Phenothiazine-Based Benzhydroxamic
Acids as Selective Histone Deacetylase 6 (HDAC6) Inhibitors
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
1
1
2
2
2
Katharina Vögerl , Nghia Ong , Johanna Senger , Daniel Herp , Karin Schmidtkunz , Martin
3
1
1
3
4
5
Marek , Martin Müller , Karin Bartel , Tajith B. Shaik , Nicholas J. Porter , Dina Robaa ,
4
3
5
2
David W. Christianson , Christophe Romier , Wolfgang Sippl , Manfred Jung , Franz
Bracher1*
1Department of Pharmacy - Center for Drug Research, Ludwig-Maximilians University
Munich, Butenandtstr. 5-13, 81377 Munich, Germany
2Institute of Pharmaceutical Sciences, University of Freiburg, Albertstraße 25, 79104
Freiburg, Germany
3Département de Biologie Structurale Intégrative, Institut de Génétique et Biologie Moléculaire
et Cellulaire (IGBMC), Université de Strasbourg (UDS), CNRS, INSERM, 67404 Illkirch
Cedex, France
4Department of Chemistry, University of Pennsylvania, 231 South 34th Street,
Philadelphia, PA 19104-6323, USA
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 140
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
5
Institute of Pharmacy, Martin-Luther University of Halle-Wittenberg, 06120
Halle/Saale, Germany
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
*
Corresponding author. Tel.: +49-89-2180-77301, E-Mail: franz.bracher@cup.uni-
muenchen.de
ABSTRACT
The phenothiazine system was identified as a favorable cap group for potent and selective
histone deacetylase 6 (HDAC6) inhibitors. Here we report the preparation and systematic
variation of phenothiazines and analogues containing a benzhydroxamic acid moiety as zinc-
binding group. We evaluated their ability to inhibit selectively HDAC6 by a recombinant
HDAC enzyme assay, by determining the protein acetylation levels in cells by Western blotting
(tubulin vs histone acetylation), and by assessing their effects on various cancer cell lines.
Structure-activity relationship studies revealed that incorporation of a nitrogen atom into the
phenothiazine framework results in increased potency and selectivity for HDAC6 (more than
5
00-fold selectivity relative to the inhibition of HDAC1, HDAC4, and HDAC8), as rationalized
by molecular modeling and docking studies. The binding mode was confirmed by co-
crystallization of the potent azaphenothiazine inhibitor with catalytic domain 2 from Danio
rerio HDAC6.
ACS Paragon Plus Environment
2

Page 3 of 140
Journal of Medicinal Chemistry
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
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
INTRODUCTION
Lysine acetylation is a major, reversible posttranslational modification on proteins that is
regulated by two families of enzymes with opposing activity, the histone acetyltransferases
1
(
HATs) and the histone deacetylases (HDACs) . Whereas the substrates of most zinc-dependent
HDACs are histones, the HDAC6 isozyme is unique in that it is predominantly localized in the
cytoplasm and as a result mainly deacetylates non-histone proteins as α-tubulin in microtubules2,
3
4
the HSP90 chaperone , and cortactin , among others. A structural peculiarity of HDAC6 is its
tandem catalytic domains termed CD1 and CD2, an inter-domain linker, and an ubiquitin-
5
, 6
binding domain . Through this ubiquitin-binding domain HDAC6 recruits misfolded
polyubiquitinated protein aggregates to the dynein motors and promotes subsequent transport to
7
the aggresomes where they are processed . Both catalytic domains are active and structurally
highly conserved, although CD1 has much more stringent substrate selectivity, whereas CD2
5
, 6
exhibits broad substrate specifity . The inter-domain linker has a significant impact on optimal
8
9
activity of HDAC6 and is reportedly involved in binding dynein motor proteins . Besides a
nuclear localization signal (NLS), HDAC6 possesses a nuclear export signal (NES) and serine-
glutamate tetradecapeptide repeat (SE14) ensuring stable anchorage of the enzyme in the
1
0
cytoplasm . By formation of complexes with its partner proteins HDAC6 regulates many
important cellular processes such as cell motility, cell spreading, misfolded protein degradation,
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 140
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
11
transcription, cell proliferation and death, and stress or immune response . In consequence,
misregulation of HDAC6 activity is associated with a variety of human diseases, which
highlights it as a potential therapeutic target. HDAC6 is implicated in various neurodegenerative
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
12
disorders, including Alzheimer´s, Parkinson´s, and Huntington´s diseases as well as
1
3
autoimmune and inflammatory disorders . Moreover high expression levels of HDAC6 seem to
play a crucial role in the pathogenesis of cancer, for instance, as found in acute myeloid
1
4
leukemia, myeoblastic, breast, and ovarian cancer . Until now, all approved HDAC inhibitors
for cancer treatment target multiple HDACs. Their poor selectivity, however, can dose-
dependently cause serious side effects15, 16, which might limit their clinical use in oncology and
beyond. Therefore the development of inhibitors for the particular isoform of interest is of great
relevance, and selective inhibition of HDAC6 may lead to fewer severe side effects17.
Various selective HDAC6 inhibitors have emerged over the years, mainly containing a
1
8
hydroxamic acid as a zinc-binding group (Chart 1). Some of them mimic N-acetyl-lysine
structurally through an alkyl chain. Tubacin was identified by combinatorial chemistry methods
1
9
and was found to decrease cell motility . As a consequence of its nondrug-like properties, high
lipophilicity and tedious synthesis it is more useful as a research tool than a drug candidate20.
Ricolinostat, another potent but less selective inhibitor with at least 10-fold selectivity for
HDAC6 relative to class I HDACs, possesses more favorable characteristics for drug
development. In preclinical studies, Ricolinostat showed synergistic anti-multiple myeloma
2
1
activity in combination with the proteasome inhibitor Bortezomib and Dexamethasone . Other
2
2
clinical trials in combinatory treatments, e.g. with next-generation analogue Citarinostat , are
2
3
still in course . Tubastatin A, a γ-carboline-based hydroxamic acid with a benzylic linker, was
developed utilizing structure-based drug design combined with homology modelling and
ACS Paragon Plus Environment
4

Page 5 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
displays high potency on HDAC6 with an IC of 15 nM and over 1000-fold selectivity over
5
0
2
4
HDAC1 . Tubastatin A showed neuroprotective effects without adverse toxicity and in recent
preclinical testing in animal models it reduced stroke-induced brain infarction and functional
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
2
5
26
deficits as well as reverted cognitive impairment in Alzheimer's disease transgenic mice .
Second-generation Tubastatin A analogues, a series of β- and γ-carboline-based hydroxamic
acids, displayed subnanomolar IC values for inhibition of HDAC6 with an enormous increase
5
0
27
of selectivity against HDAC1 . Tubastatin A and its analogues were found to enhance the
2
7
immunosuppressive effects of Foxp3+ T-regulatory cells . ACY-775, a novel HDAC6 inhibitor
with improved brain bioavailability compared to Tubastatin A, showed low nanomolar potency
and high selectivity over class I HDAC enzymes, linked to an antidepressant activity28.
Marbostat 100, a tetrahydro--carboline derived from Tubastatin A with anti-inflammatory
activity, was published as a very potent and selective HDAC6 inhibitor during the completion of
29
this manuscript . Nexturastat A which inhibits the growth of B16 melanoma cells, bears a
benzylic substituted urea linker. It shows low nanomolar potency against HDAC6 as well and a
selectivity of ~600-fold over HDAC1 making it the first HDAC6-selective inhibitor with
30
antiproliferative effects . A new Nexturastat A analogue demonstrated improved capability to
inhibit tumor growth in melanoma models through the regulation of inflammatory and immune
3
1
responses . Like Nexturastat A, HPOB is also HDAC6 specific, albeit with lower selectivity
(~50-fold over HDAC1) and features similarly a benzylic linker. This inhibitor was shown to
cause growth inhibition without inducing cell death and in combination therapy it may enhance
3
2
the efficacy of other anticancer drugs . Further, inhibitors bearing five-membered
3
3
heteroaromatic linkers such as oxazoles have been reported to exhibit good HDAC6 selectivity .
Diverse novel hydroxamate-based HDAC6 inhibitors have been published very recently34-38
.
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 140
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
A large number of HDAC inhibitors share common structural features, including the zinc-
3
9
binding group (ZBG), the linker, and the cap group . Until recently, the crystal structure of
human HDAC6 was not available and HDAC6-specific inhibitor design approaches were
restricted to homology models along with experimental results. A key point to gain subtype
selectivity is the capping motif, which interacts with the surface of the wide basin surrounding
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
3
9
the binding pocket of the enzyme . The linker segment fits in the hydrophobic channel and
3
9
bridges the zinc-binding hydroxamic acid and the cap group . Given the structural feature of a
benzyl linker in Tubastatin A and related inhibitors, a short bulky linker proved to be beneficial
for potent and selective HDAC6 inhibition.
Based on the knowledge of the structural parameters essential for selective HDAC6 inhibitor
activity, we chose to investigate new hydroxamate-based HDAC6 inhibitors that might gain the
desired isoenzyme selectivity by their capping motif. To guide the design and synthesis of these
inhibitors, the recently reported X-ray structures of HDAC6 in complex with benzhydroxamates
were considered5, 6, 40
. Herein, we report on the identification and development of the
phenothiazine scaffold as a capping motif that confers potent and selective HDAC6 inhibitors in
vitro and in cell culture.
Chart 1. Structures of selective HDAC6 inhibitors. The general structure of these inhibitors
consists of a cap group (in red), a linker (in green), and a zinc-binding group (ZBG; in blue).
ACS Paragon Plus Environment
6

Page 7 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
O
O
H
N
R
N
S
N
HN
OH
Y
O
O
X
N
Cap
N
N
H
N
H
H
O
OH
Linker
ZBG
N
N
OH
OH
OH
O
S
O
O
O
HO
N
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
Tubacin
Tubastatin A (X = CH , Y = NCH , R = H)
Tubastatin A analogues
e.g. X = NCH , Y = CH R = H)
Marbostat-100
Benzothiazine 17c
2
3
(
3 2,
F
O
N
O
N
N
H
H
N
N
N
N
H
H
H
H
H
N
N
N
N
N
N
N
OH
OH
OH
OH
O
O
O
O
O
ACY-775
Ricolinostat
Nexturastat A
HPOB
RESULTS AND DISCUSSION
CHEMISTRY
We began our studies on hydroxamate-based inhibitors with various polycyclic cap groups which
were suitable for connection with the benzyl linker (see Tubastatin A, Chart 1) through
benzylation of a ring nitrogen. We identified 4-[(10H-phenothiazin-10-yl)methyl]-N-
hydroxybenzamide (1a, Chart 2) as a very potent HDAC6 inhibitor with high selectivity over
HDAC1. This hit has structural similarity to a previously published selective benzothiazine-type
4
1
HDAC6 inhibitor (Chart 1) . On the basis of this structure, we set out for structure-activity
relationships to improve selectivity and inhibitory potency.
Chart 2. Lead structure (1a).
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 140
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
S
N
Cap
Linker
ZBG
H
N
OH
1a
O
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
Docking of screening hit 1a to HDAC6 showed that the tricyclic ring system is interacting with
the rim of the binding tunnel but there is also space for further substitutions. Consequently,
initial modifications concerned the phenothiazine ring system of compound 1a. Various
substituted phenothiazines, but also azaphenothiazines and other heterocyclic analogues were
introduced as cap groups according to the synthetic route outlined in Scheme 1 and Table 1. The
cap groups of choice were either commercially available or were prepared following established
procedures. N-alkylation of the cap groups 2a-q with methyl 4-(bromomethyl)benzoate (3)
delivered intermediate esters 4a-q. We first attempted to convert the esters directly into
hydroxamates by treatmernt with hydroxylamine under diverse reaction conditions, but yields
were very poor. Finally, we hydrolysed the esters with sodium hydroxide in aqueous dioxane to
give the corresponding carboxylic acids 5a-q. The hydroxamic acids 1a and 7b-q were obtained
4
2
using COMU as activating agent and either hydroxylamine base (7e-q) or O-(tetrahydro-2H-
pyran-2-yl)hydroxylamine (1a, 7b-d), followed by acidic cleavage of the THP protecting group.
Based on our experience COMU was the most suitable coupling reagent, since only water
soluble co-products arise from this reagent which are conveniently removed during aqueous
4
3
work up . Generally, using hydroxylamine base increased the overall yields of the desired
hydroxamic acids compared to using THP-protected hydroxylamine (as for instance 7i 41 %
versus 21 % over two steps). The same coupling procedure with N,O-dimethylhydroxylamine
provided N,O-dimethyl-substituted hydroxamic acid 23.
ACS Paragon Plus Environment
8

Page 9 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 1 . General synthetic route to the target hydroxamic acids and the N,O-dimethyl-
substituted hydroxamic acid 23.
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
X
N
X
N
A
B
A
B
X
Br
a
b
A
B
N
H
O
O
OH
O
O
O
2a-o
3
4a-o
5a-o
e
c
X
N
X
d
A
B
A
B
N
R1
N
H
N
O
R2
O
O
O
O
1
1
a, 7b-o (R , R2 = H)
6a-d
1
2
3 (R , R2 = CH3)
a
Reagents and conditions: (a) LiHMDS, DMF, 0 °C → rt or 50 °C or 60 °C; (b) NaOH, H O/1,4-
2
dioxane (1:1), rt or 40 °C or 50 °C or 60 °C; (c) DIPEA or Et N, COMU, NH OTHP, DMF, 0 °C
3
2
→
rt; (d) HCl in 1,4-dioxane, CH Cl , rt; (e) DIPEA, COMU, NH OH x HCl, DMF, 0 °C → rt;
2 2 2
for 23: 5a, DIPEA, COMU, CH NHOCH x HCl, DMF, 0 °C → rt.
3
3
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 140
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
Table 1. Line-up of cap groups defining the variations in A, B and X of the phenothiazine
scaffold.
X
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
A
B
N
H
S
O
S
S
S
N
2
a
2e
2f
2i
2m
2n
N
N
N
H
N
H
N
N
H
N
H
O O
S
S
S
N
S
2
b
2j
2k
2l
N
N
H
Cl
N
H
N
H
N
H
S
S
S
N
N
S
S
N
2
c
2g
2h
2o
N
H
N
H
O
N
H
N
H
Cl
S
S
N
S
S
R
2
p/2q
2
d
N
N
H
CF3
N
S
N
H
N
N
N
H
H
2p (R=H)
2q (R=Cl)
Intriguingly, N-alkylation of 1,2-diazaphenothiazine (2p) and its 3-chloro-derivative 2q
delivered the structural isomers 8p and 8q besides the desired esters 4p and 4q (Scheme 2). The
structural peculiarity in constitution of cap group and linker made 8p/8q also interesting building
blocks for the installation of the hydroxamic acid moiety to give compounds 10p/10q.
a
Scheme 2 . Structural isomers from 1,2-diazaphenothiazines.
ACS Paragon Plus Environment
10

Page 11 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
S
N
R
O
S
N
R
O
S
N
R
O
S
R
N
N
N
N
N
N
c
N
b
N
H
N
H
N
O
OH
2
2
p (R=H)
q (R=Cl)
OH
4
4
p (R=H)
q (R=Cl)
5p (R=H)
5q (R=Cl)
7p (R=H)
7q (R=Cl)
a
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
Br
S
N
R
S
N
R
S
N
R
c
N
b
N
N
O
N
N
N
O
H
3
O
OH
N
OH
O
O
O
8
8
p (R=H)
q (R=Cl)
9p (R=H)
9q (R=Cl)
10p (R=H)
10q (R=Cl)
aReagents and conditions: (a) LiHMDS, DMF, 0 °C → rt or 0 °C → rt → 60 °C; (b) NaOH,
H O/1,4-dioxane (1:1), 40 °C or 60 °C; (c) DIPEA, COMU, NH OH x HCl, DMF, 0 °C → rt.
2
2
In order to achieve close structural similarity to the novel, highly active second-generation
2
7
Tubastatin A analogues (Chart 1) we prepared a tetrahydro analogue of azaphenothiazine 7j
Scheme 3). Alkylation of 2-azaphenothiazine 2j as described above yielded ester 4j. This ester
(
was further N-alkylated with methyl iodide to give the quaternary ammonium salt 11j.
Chemoselective reduction with NaBH444 gave the tetrahydro derivative 12j. Finally, ester
hydrolysis and formation of the hydroxamic acid were performed as described above.
a
Scheme 3 . Synthetic route to hydroxamic acid 14j bearing a 2-methyl-1,2,3,4-
tetrahydro-10H-pyrido[4,3-b][1,4]benzothiazine cap.
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 140
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
S
S
N
I
S
Br
N
N
N
N
a
b
N
H
O
O
O
O
O
O
2
j
4j
11j
3
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
S
N
S
S
N
N
N
N
d
N
e
c
H
N
O
OH
OH
O
O
O
12j
13j
14j
a
Reagents and conditions: (a) LiHMDS, DMF, 0 °C → rt; (b) CH I, acetone/EtOH (6.5:1),
3
reflux; (c) NaBH , EtOH, -15 °C → rt; (d) NaOH, H O/1,4-dioxane (1:1), rt → 60 ° C; (e)
4
2
DIPEA, COMU, NH OH x HCl, DMF, 0 °C → rt.
2
Further, we conducted modifications of the linker group of the lead structure 1a. Compound 18
consisting of a shorter and significantly less flexible phenyl linker was synthesised via N-
arylation of parent phenothiazine 2a with 4-fluorobenzonitrile 4 to give nitrile 16 (Scheme 4).
45
Alkaline hydrolysis and acidic work up gave carboxylic acid 17 , which underwent conversion
into the corresponding hydroxamic acid as aforementioned.
ACS Paragon Plus Environment
12

Page 13 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 4 . Synthetic route to the N-phenylphenothiazine hydroxamic acid 18.
S
N
S
N
S
N
F
S
a
b
c
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
N
H
CN
CN
HO
HO
O
N
O
H
2a
15
16
17
18
a
Reagents and conditions: (a) LiHMDS, DMF, 0 °C → rt; (b) KOH, EtOH/H O (7:3), reflux; (c)
2
DIPEA, COMU, NH OH x HCl, DMF, 0 °C → rt
2
To elucidate the relevance of an extended linker compound 22 containing a phenylethyl linker
was prepared (Scheme 5). A novel reductive N-phenylethylation of phenothiazine under acidic
conditions with triethylsilane, trifluoroacetic acid and methyl 4-(2-methoxyvinyl)benzoate gave
4
6
the ester 20 . The required enol ether 19 was obtained by Wittig reaction of methyl 4-
formylbenzoate using (methoxymethyl)triphenylphosphonium chloride and LDA. Subsequent
ester hydrolysis and formation of hydroxamic acid was done as described above.
a
Scheme 5 . Synthetic route to phenylethyl hydroxamic acid 22.
S
N
S
N
S
N
O
S
a
b
c
N
H
O
O
O
O
O
OH
O
NH
OH
2a
19
20
21
22
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 140
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
a
Reagents and conditions: (a) TFA, Et SiH, CH Cl ; (b) NaOH, H O/1,4-dioxane (1:1), 60 °C;
3
2
2
2
(
c) DIPEA, COMU, NH OH x HCl, DMF, 0 °C → rt
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
INHIBITION OF HDAC1, HDAC6, AND HDAC8
All the synthesized compounds were evaluated for their inhibitory activity against human
4
7
HDAC1 and HDAC6 in a biochemical in vitro assay . The data are summarized in Table 2. The
selective inhibitors Tubacin and Tubastatin A were included as reference substances. Already
published values for Tubastatin A are listed additionally for comparison across different assay
procedures. Compared to published data, Tubastatin A turned out to be 9 times more potent on
HDAC1 and 2 times less potent on HDAC6 and therefore significantly less selective for HDAC6
over HDAC1 when tested in our system. These outcomes indicate that our compounds show, at
least in our assay, greater selectivity for HDAC6 over HDAC1 in comparison to Tubastatin A.
Our lead structure 1a already showed very impressive HDAC6 inhibition (IC = 22 nM) and a
5
0
selectivity factor of 231 over HDAC1, thus comparing favourably with the data obtained for
Tubastatin A in our assay (IC = 30 nM, selectivity factor 64). Primarily our studies focused on
5
0
optimizing the capping motif of 1a to even improve potency and selectivity for HDAC6.
However, we also tested alternative linker motifs (18 and 22). The results demonstrated once
3
more that a sp carbon connecting the aromatic part of the linker and the cap is crucial for
maintaining potent inhibition of HDAC6. While the direct connection to a phenyl linker (18)
resulted in almost complete loss of activity and selectivity for HDAC6, the flexible phenylethyl
linker (22) showed merely a reduced ability to inhibit HDAC6 with decreased selectivity. A
similar outcome for HDAC6 inhibition was received for a phenylethyl analogue of Tubastatin
ACS Paragon Plus Environment
14

Page 15 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
24
A . Therefore we did not further explore the linker motif and retained the benzyl linker for the
following compounds.
Regarding the modified capping motif, all congeners of 1a containing additional substituents at
positions 1 or 2 of the phenothiazine scaffold (7c, 7d, 7f, 7g and 7h) also showed inhibitory
activities on HDAC6 in the low nanomolar range (≤ 31 nM) and similar selectivity profiles. The
same applies to the sulfone 7b, whereas the sulfoxide 7e suffered slightly in potency and
selectivity. The incorporation of one or two nitrogen atoms into the phenothiazine scaffold (7i to
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
7
n, 7p) even enhanced the potency for HDAC. In comparison to afore mentioned compounds
these (di)azaphenothiazine-containing compounds exhibited up to fourfold increase in HDAC6
inhibition accompanied by non-uniform changes in selectivity over HDAC1. Additional
substitution at position 3 (chlorine) in compound 7q resulted in a slight loss in selectivity for
HDAC6. Moving the position of the linker from position 10 (thiazine nitrogen; compounds 7p
and 7q) to N-1 of the diazaphenothiazine ring system (compounds 10p and 10q) resulted in the
most detrimental impact on selectivity. Strikingly, compound 7i, bearing an 1-azaphenothiazine
scaffold, attained a very high level of potency (5 nM) and selectivity (538-fold). Substantial
variations of the phenothiazine scaffold as demonstrated with the thiazole congener 7o, resulted
in a slight decline in potency and selectivity compared to compound 1a.
A significant increase (from 1000 to 3700-fold) in selectivity had been reported for the
2
7
tetrahydro-β-carboline regioisomer of Tubastatin A . However, introduction of an analogous
tetrahydropyridine motif into our phenothiazine scaffold (compound 14j) led to a slight decrease
in potency accompanied by a dramatic loss in selectivity.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 140
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
Not surprisingly, the N,O-dimethyl-substituted hydroxamic acid 23 was completely inactive on
both HDACs. Further, the carboxylic acid precursors 5a and 5n did not show noteworthy
HDAC6 inhibition in concentrations up to 25 μM (data not shown here).
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
Table 2. In vitro inhibition of HDAC1 and HDAC6.
X
A
B
N
n
R1
N
O
O
R2
X
N
hHDAC1a
IC [µM]
hHDAC6a
SF
1/6
A
B
1
2
Compound
Tubacin
n =
R , R =
b
IC [µM]
5
0
50
1
.01 ± 0.11
0.007 ± 0.001
0.004 ± 0.00124
0.030 ± 0.002
0.015 ± 0.00124
144
1
.40 ± 0.2424
350
64
1
.91 ± 0.42
N
Tubastatin
A
1
H
N
1
6.4 ± 2.624
1093
Regioisomer
of
Tubastatin
N
1
1
H
H
5.18 ± 0.1227 0.0014 ± 0.000327
3700
231
N
S
N
1
a
5.08 ± 0.51
3.99 ± 1.38
0.022 ± 0.007
0.013 ± 0.001
O
O
S
7
b
1
1
H
H
307
360
N
S
N
7
c
5.76 ± 1.09
0.016 ± 0.002
Cl
ACS Paragon Plus Environment
16

Page 17 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
S
N
7
d
1
H
H
7.49 ± 1.63
0.025 ± 0.002
0.077 ± 0.006
300
88
CF3
O
S
7
e
1
6.81 ± 2.65
N
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
S
7
f
1
1
1
1
1
1
1
1
1
1
1
1
H
H
H
H
H
H
H
H
H
H
H
H
2.95 ± 0.18
2.07 ± 0.19
7.20 ± 0.42
2.69 ± 1.04
1.25 ± 0.01
3.84 ± 0.13
3.39 ± 0.23
0.79 ± 0.14
3.34 ± 0.21
5.05 ± 0.75
1.06 ± 0.08
1.08 ± 0.04
0.012 ± 0.001
0.010 ± 0.001
0.031 ± 0.003
0.005 ± 0.001
0.007 ± 0.001
0.017 ± 0.006
0.014 ± 0.003
0.009 ± 0.001
0.012 ± 0.003
0.037 ± 0.008
0.011 ± 0.001
0.018 ± 0.005
246
207
232
538
179
226
242
88
N
Cl
O
S
7
g
N
S
N
7
h
S
S
N
7
i
N
S
N
7
j
N
S
N
N
N
7
k
N
S
N
7l
N
S
N
N
7
m
N
N
N
S
N
7n
278
136
96
S
S
7
o
N
N
S
7p
N
N
N
S
N
Cl
7
q
N
60
N
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 140
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
S
N
1
0p
N
1
H
H
0.75 ± 0.04
0.019 ± 0.003
0.115 ± 0.026
0.032 ± 0.005
2.67 ± 0.47
0.136 ± 0.023
n.i.
39
40
31
2
N
S
N
Cl
10q
N
1
1
0
2
1
4.58 ± 0.73
0.98 ± 0.20
5.87 ± 0.71
11.68 ± 2.61
n.i.
N
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
S
1
4j
N
H
N
S
1
8
2
3
H
N
S
N
2
H
86
S
N
2
CH3
a
b
IC values are the mean of at least two experiments. SF1/6: selectivity factor for HDAC6 over
5
0
HDAC1 (SF1/6 = IC (HDAC1)/IC (HCAC6)). n.i. was defined as ˂50% inhibition at 100 µM.
5
0
50
Ligand efficiency (LE) was used to estimate the relevance of achieved changes in potency
regarding HDAC6 in terms of the performed modifications (Table 3). LE is a measure of the in
vitro biological activity corrected for the number of heavy atoms (HA) and allows to assess how
4
8
effectively the compound uses its structural features in binding to the target protein . Regarding
the alterations of our lead structure 1a the highest ligand efficiencies were registered for the
(
di)azaphenothiazine-containing compounds (7i to 7n, 7p) with LE values ≥ 0.44 and therefore
surpassing compound 1a and Tubastatin A. Compound 7i with an LE = 0.47 was also the most
4
8
potent one and in good consistence with the mean value of published oral drugs (LE = 0.45) .
ACS Paragon Plus Environment
18

Page 19 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. IC values [µM] and ligand efficiencies (LE) [(kcal/mol)/non-H-Atom] on HDAC6.
5
0
Tub A = Tubastatin A.
Compound
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
Tub A
Tub A
25
1
18
24
22
26
7b
27
7c
26
7d
29
HA
25
25
IC ,
5
0
0
.01524
0.030
0.022
2.67
0.136
0.013
0.016
0.025
HDAC6
LE,
0.44
0.43
0.43
0.33
0.37
0.41
0.43
0.37
HDAC6
Compound
7h
7
f
7e
26
7g
27
7i
25
7j
25
7k
25
7p
25
HA
26
28
IC ,
5
0
0
.012
0.077
0.010
0.031
0.005
0.007
0.017
0.011
HDAC6
LE,
0
.43
0.39
0.42
0.38
0.47
0.46
0.44
0.45
HDAC6
Compound
7q
7
l
7m
25
7n
25
10p
25
10q
26
7o
25
14j
26
HA
25
26
IC ,
5
0
0
.014
0.009
0.012
0.018
0.019
0.115
0.037
0.032
HDAC6
LE,
0
.45
0.46
0.45
0.42
0.44
0.38
0.42
0.41
HDAC6
Some previously reported highly potent HDAC6 inhibitors exhibited moderate selectivity over
HDAC849, 50 or even were HDAC6/8 dual inhibitors . Therefore four representative compounds
were selected to be also tested on HDAC8. Compound 7i showed a selectivity profile
51
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 140
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
considerably better than those of SAHA (Vorinostat) and Tubastatin A, meaning an excellent
selectivity of this compound for HDAC6 (Table 4). Compounds 1a, 7i and 7n were further tested
for HDAC4 inhibition, but showed less than 20% inhibition at 10 μM (data not shown here).
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
Table 4. In vitro inhibition of HDAC1, HDAC6, and HDAC8
X
A
B
N
H
N
OH
O
X
N
hHDAC1a
IC [µM]
hHDAC6a
hHDAC8a
SF
1/6 8/6
SF
A
B
Compound
SAHA
b
c
IC [µM]
IC [µM]
5
0
50
50
0.117 ± 0.006
1.91 ± 0.42
0.104 ± 0.009
0.030 ± 0.002
0.400 ± 0.100
0.695 ± 0.060
1
4
64
23
N
Tubastatin
A
N
1
6.4 ± 2.624
0.015 ± 0.00124 0.854 ± 0.04024 1093 57
S
N
1a
5.08 ± 0.51
2.69 ± 1.04
3.34 ± 0.21
0.98 ± 0.20
0.022 ± 0.007
0.005 ± 0.001
0.012 ± 0.003
0.032 ± 0.005
1.48 ± 0.04
3.10 ± 0.09
231 67
S
N
7
i
538 620
278 167
31 n.d.
N
N
S
N
N
N
7
n
2.00 ± 0.03
S
N
6
6% inhibition
1
4j
@
1 µM
ACS Paragon Plus Environment
20

Page 21 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
b
IC values are the mean of at least two experiments. SF1/6: selectivity factor for HDAC6 over
5
0
c
HDAC1 (SF1/6 = IC (HDAC1)/IC (HCAC6)). SF8/6: selectivity factor for HDAC6 over
5
0
50
HDAC8 (SF8/6 = IC (HDAC8)/IC (HCAC6)). n.d. not determined
5
0
50
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
CELLULAR DATA
To further validate the results obtained in the in vitro assay, Western blotting experiments were
performed with two of our best compounds (7i and 7n) in the leukemic cell line HL 60. The pan-
HDAC inhibitor SAHA was used as a reference compound. HL60 cells were incubated for 4 h
with SAHA, 7i or 7n in three different concentrations (10 µM, 1 µM, and 0.1 µM). Cell lysates
were blotted against Ac-α-tubulin, Ac-histone H3, and GAPDH. The pan-HDAC inhibitor SAHA
induced an increase of tubulin acetylation as well as of histone acetylation in HL60 cells compared
to DMSO control. At least at 1 µM 7i and 7n induced only the acetylation of tubulin but not of
histone H3 (see Figure 1), reflecting the selectivity ascertained in vitro also in the cellular setting
in this range of concentration. At 10 µM some histone hyperacetylation is visible, showing the
concentration dependence of selectivity in the cellular setting.
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 140
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 1. Western blots of acetyl-α-tubulin, acetyl-H3, and GAPDH after treatment of HL60
cells with SAHA, 7i, and 7n. GAPDH was the loading control. DMSO was the negative control.
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 Western blots were quantified simply by determining the ratio of acetylated tubulin versus
GAPDH and of acetylated histone 3 versus GAPDH (Table 5).
As shown in Figure 1 and Table 5, the induction of tubulin acetylation by 7i and 7n was greater
than the induction of histone H3 acetylation at all concentrations tested. Compared to SAHA the
level of acetylated tubulin is higher than the level of acetylated histone H3, except for 7n at the
lowest concentration. Thereby the HDAC6 selectivity of these compounds is also demonstrated
and validated on a cellular level.
Table 5. Quantification of Western blots in HL60 cells
Ac-
selectivity
indexa
Compound
Ac-H3/GAPDH
tubulin/GAPDH
DMSO
1.00b
9.52
9.97
3.79
9.20
9.41
1.00b
2.80
1.91
1.52
1.28
1.07
SAHA, 10 µM
SAHA, 1 µM
SAHA, 0.1 µM
3.40
5.22
2.49
7.19
8.79
7
7
i, 10 µM
i, 1 µM
ACS Paragon Plus Environment
22

Page 23 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
7
i, 0.1 µM
n, 10 µM
5.42
10.93
7.97
1.14
1.47
0.99
1.05
4.75
7.44
8.05
2.06
7n, 1 µM
n, 0.1 µM
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
7
2.16
a
b
Selectivity index defined as quotient of Ac-tubulin/GAPDH divided by Ac-H3/GAPDH. All
values have been normalized to DMSO.
Histone deacetylase 6 (HDAC6) is a crucial regulator of gene expression and transcription in
cancer and several HDAC6 inhibitors are currently investigated in clinics as potential tumor
1
6
therapeutics . Likewise, we aimed at investigating the effects of 7i and 7n on different cancer
cells, to investigate the functional effects of HDAC6 inhibition in different tumor entities.
Additionally, we used non-malignant cells to assess the cancer specificity and to estimate
cytotoxicity (Figure 2). 1a and the pan-HDAC inhibitor SAHA were used as reference compounds.
®
Firstly, CellTiter-Blue cell viability assays indicated that 7i and 7n inhibited proliferation of
HUH7 (hepatocellular carcinoma, Figure 2 A), MDA-MB-231 (breast adenocarcinoma, Figure 2
B) and T24 (bladder carcinoma, Figure 2 C) cells at concentrations of 10 and 100 µM after 72.
SAHA treatment generally resulted in stronger antiproliferative effects, with IC values in the low
5
0
micromolar range (HUH7: 1.2 µM, MDA-MB-231: 2.8 µM, T24: 2.4 µM). Similar findings were
made for other breast cancer (MCF7) and leukemia (HL60, THP-1) cellular models (Figure S7,
Supporting Information), showing an antiproliferative effect on cancer cells of different tissue
types. Secondly, cytotoxic effects on cancer cells were analyzed by measurement of apoptotic cell
death. In accordance with antiproliferative effects, the highest apoptosis rates of all tested HDAC
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 140
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
inhibitors were observed after treatment of MDA-MB-231 breast cancer and T24 bladder cancer
cells (Figure 2 D and E) with SAHA for 48 h, as assessed with propidium iodide staining and flow
cytometry . These findings further emphasized that 1a, 7i and 7n are selective HDAC6 inhibitors
and their application as single treatment only causes minor anticancer activity, which is in line
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
6
with the low clinical activity of selective HDAC6 inhibitors . The HDAC6 inhibitor with reduced
class I HDAC inhibition Ricolinostat, for instance, is well tolerated but it has low clinical activity
5
2
when applied as monotherapy , and Woan et al. showed that both genetic ablation and
pharmacologic inhibition of HDAC6 decreased cancer cell proliferation without inducing
5
3
apoptosis . The limited anticancer activity of specific HDAC6 inhibitors is sought to be overcome
5
2
by combination strategies, e.g. with immunotherapeutic agents like Lenalidomide or anti-PD-L1
antibodies54 or, alternatively, with the proteasome inhibitor Bortezomib55, 56, for which a
5
6
synergistic effect was observed . Based on that, we hypothesized that 1a, 7i and 7n might alter
critical cancer-related pathways, namely immune checkpoints and unfolded protein response
(
UPR). Hence, we checked the influence of specific HDAC6 inhibition on programmed death
receptor ligand-1 (PD-L1, Figure 2 F) and endoplasmatic reticulum chaperone BiP (GRP78, Figure
G). Interestingly, western blot analyses of bladder carcinoma lysates revealed that pharmacologic
2
HDAC inhibition caused a clear upregulation of PD-L1 and a slight increase in GRP78 protein
level. Based on that finding, we tested combinations of 1a, 7i and 7n with Bortezomib and their
influence on cell proliferation by impedance measurements (Figure S9 A and B, Supporting
Information). However, we could not prove a synergism of 1a, 7i and 7n and Bortezomib at the
tested concentrations, which we related to the narrow therapeutic window of Bortezomib in our
setting (Figure S9 C, Supporting Information). Moreover, our data could help to decipher the
5
7
interaction of HDAC6 and PD-L1 in cancer, which is still debated. Woods et al. could show that
ACS Paragon Plus Environment
24

Page 25 of 140
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
class I HDAC inhibition enhances antitumor immunity by upregulating PD-L1 expression in
melanomas, Lienlaf et al. on the other hand found a reduction of PD-L1 protein level after selective
HDAC6 inhibition in melanoma cells. They reported that HDAC6 is essential for cytokine
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
58
mediated expression control of PD-L1 . Yet, Woan et al. could not detect alterations in PD-L1
5
3
protein level after pharmacologic or genetic disruption of HDAC6 activity . It is therefore obvious
that more research has to be performed to clarify the role of HDAC6 in PD-L1 regulation. Adding
still more complexity to the question, the regulation might also be tissue specific as tumors of
5
9
different origin display distinct immunogenicity . Therefore, HDAC6 inhibition can cause an
upregulation of PD-L1 in bladder cancer, as reported in our study, and at the same time a
5
8
downregulation in melanoma as reported by Lienlaf et al. . Hence, our findings confirm the
rationale for combining 1a, 7i and 7n with anti-PD1/anti-PD-L1 antibodies and proteasome
inhibitors.
Furthermore, toxicity of the HDAC inhibitors on non-malignant cells was assessed using
HepaRGTM cells, a non-cancerous hepatic stem cell line that expresses high levels of CYP450
enzymes and thereby allows detection of CYP450-mediated toxicity. Toxicity and unfavourable
1
5
tolerability associated with pan-HDAC inhibitor therapy remain key unmet needs that have
driven the development of specific HDAC6 inhibitors. Viability of HepaRGTM cells after 24 h of
treatment was not affected even at 100 µM concentrations of 1a and 7i, whereas application of
SAHA and 7n slightly decreased cell viability at the highest tested concentrations to around 80%
(Figure 2 H).
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 140
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
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
ACS Paragon Plus Environment
26

Page 27 of 140
Journal of Medicinal Chemistry
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
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
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 140
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
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
ACS Paragon Plus Environment
28

Page 29 of 140
Journal of Medicinal Chemistry
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
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 2. Biological testing of SAHA, 1a, 7i and 7n. Proliferation of (A) HUH7, (B) MDA-
®
MB-231, and (C) T24 cells was analyzed by CellTiter-Blue cell viability assay after
treatment as indicated for 72 h. Apoptosis of (D) MDA-MB-231 and (E) T24 cells was
assessed by propidium iodide staining and flow cytometry after treatment as indicated
for 48 h. Expression of (F) PD-L1 and (G) GRP78 was detected by western blot analysis
of T24 protein lysates upon 10 µM treatment with the indicated compounds (48 h).
Protein bands were normalized to whole lane protein of the DMSO control (loading
control). (H) Viability of HepaRGTM cells was analyzed by CellTiter-Blue cell viability
®
ACS Paragon Plus Environment
29