Marbostat-100 Defines a New Class of Potent and Selective Antiinflammatory and Antirheumatic Histone Deacetylase 6 Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.7b01593

Epigenetic modifiers of the histone deacetylase (HDAC) family contribute to autoimmunity, cancer, HIV infection, inflammation, and neurodegeneration. Hence, histone deacetylase inhibitors (HDACi), which alter protein acetylation, gene expression patterns, and cell fate decisions, represent promising new drugs for the therapy of these diseases. Whereas pan-HDACi inhibit all 11 Zn²⁺-dependent histone deacetylases (HDACs) and cause a broad spectrum of side effects, specific inhibitors of histone deacetylase 6 (HDAC6i) are supposed to have less side effects. We present the synthesis and biological evaluation of Marbostats, novel HDAC6i that contain the hydroxamic acid moiety linked to tetrahydro-β-carboline derivatives. Our lead compound Marbostat-100 is a more potent and more selective HDAC6i than previously established well-characterized compounds in vitro as well as in cells. Moreover, Marbostat-100 is well tolerated by mice and effective against collagen type II induced arthritis. Thus, Marbostat-100 represents a most selective known HDAC6i and the possibility for clinical evaluation of a HDAC isoform-specific drug.

Full text of DOI:10.1021/acs.jmedchem.7b01593

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Marbostat 100 defines a new class of potent and selective
antiinflammatory and antirheumatic histone deacetylase 6 inhibitors
Andreas Sellmer, Hubert Stangl, Mandy Beyer, Elisabeth Grünstein, Michel Leonhardt,
Herwig Pongratz, Emerich Eichhorn, Sigurd Elz, Birgit Striegl, Zsuzsa Jenei-
Lanzl, Stefan Dove, Rainer H Straub, Oliver H Krämer, and Siavosh Mahboobi
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01593 • Publication Date (Web): 28 Mar 2018
Downloaded from http://pubs.acs.org on March 28, 2018
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 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Marbostatꢀ100 defines a new class of potent and
selective antiinflammatory and antirheumatic histone
deacetylase 6 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
2
3
1
1
Andreas Sellmer , Hubert Stangl , Mandy Beyer , Elisabeth Grünstein , Michel Leonhardt ,
1
1
1
4,5
2,6
Herwig Pongratz , Emerich Eichhorn , Sigurd Elz , Birgit Striegl , Zsuzsa JeneiꢀLanzl ,
1
2
3
1
Stefan Dove ‡, Rainer H. Straub ‡, Oliver H. Krämer ‡, Siavosh Mahboobi ‡*
1
Institute of Pharmacy, Faculty of Chemistry and Pharmacy, University of Regensburg,
9
2
3040 Regensburg, Germany
Laboratory of Experimental Rheumatology and Neuroendocrine Immunology, Department
of Internal Medicine, University Hospital 93042 Regensburg, Germany
3
Institute of Toxicology, Johannes Gutenberg University Mainz, Universitätsmedizin 55131
Mainz, Germany
4
Technical University of Applied Sciences (OTH) Regensburg, 93053, Regensburg,
Germany
5
Regensburg Center of Biomedical Engineering (RCBE), OTH and University Regensburg,
9
6
3053 Regensburg, Germany
Dr. Rolf M. Schwiete Research Unit for Osteoarthritis, Orthopedic University Hospital,
Friedrichsheim GmbH, 60528 Frankfurt/Main, Germany
‡
Senior authors who contributed equally
*Corresponding author, Eꢀmail: Siavosh.mahboobi@chemie.uniꢀregensburg.de
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 80
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
KEYWORDS. Acetylated tubulin, HDAC6 inhibitor, Hydroxamic acid, Rheumatoid
arthritis, Tetrahydroꢀßꢀcarboline.
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
Abstract
Epigenetic modifiers of the histone deacetylase (HDAC) family contribute to autoimmunity,
cancer, HIV infection, inflammatory, and neurodegeneration. Hence, histone deacetylase
inhibitors (HDACi), which alter protein acetylation, gene expression patterns, and cell fate
decisions, represent promising new drugs for the therapy of these diseases. Whereas panꢀ
2
+
HDACi inhibit all 11 Zn ꢀdependent histone deacetylases (HDACs) and cause a broad
spectrum of side effects, specific inhibitors of histone deacetylase 6 (HDAC6i) are
supposed to have less side effects. We present the synthesis and biological evaluation of
Marbostats, novel HDAC6i, that contain the hydroxamic acidꢀmoiety linked to tetrahydroꢀ
βꢀcarbolineꢀderivatives. Our lead compound Marbostatꢀ100 is a more potent and more
selective HDAC6i than previously established, wellꢀcharacterized compounds in vitro as
well as in cells. Moreover, Marbostatꢀ100 is well tolerated by mice and effective against
collagen type IIꢀinduced arthritis. Thus, Marbostatꢀ100 represents the most selective known
HDAC6i and the possibility for clinical evaluation of a HDAC isoformꢀspecific drugs.
Introduction
Acetylation and deacetylation of histones significantly contribute to the packing of
chromatin and thereby to the control of gene expression. Acetylation/deacetylation cycles
1
ꢀ3
additionally regulate physiologically important functions of nonꢀhistone proteins.
2
ACS Paragon Plus Environment

Page 3 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Acetylation of proteins is controlled by two types of enzymes: Histone acetyltransferases
(HAT) attach acetyl groups to lysine residues of histone and nonꢀhistone proteins, and
1
, 2, 4
histone deacetylases (HDACs) catalyze the removal of those acetyl groups.
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
HDACs can be divided into four classes (IꢀIV): While class I, II, and IV are zinc
+
dependent, class III deacetylases need NAD as a cofactor. Class II is further subdivided
1
, 2, 4, 5
into two subgroups, IIa (HDAC4, 5, 7, 9) and IIb (HDAC6 and 10).
HDAC6 belongs
5
to class IIb and is mainly distributed in the cytoplasm due to its nuclear export signal.
HDAC6 is the only enzyme of the HDAC family which has two catalytically active
4, 6
domains. It further became clear that HDAC6 has no in vivo activity against histones.
Targets of HDAC6 are HSP90, tubulin, and other proteins that contribute to cancer
4, 6
development and progression.
Hence, agents that selectively target HDAC6 could be
4
valuable antiꢀcancer drugs. Moreover, HDAC6 inhibitors are assumed as a prosing
7
, 8
therapeutic strategy in the treatment of neurodegenerative diseases.
Histone deacetylase inhibitors (HDACi) modulate protein acetylation and thereby gene
expression patterns, protein stability, intracellular targeting of proteins, and other protein
functions.
Due to the target structure, the classical pharmacophore model for HDACi, exemplified
by SAHA (1) (Figure 1) consists of (i) a zincꢀbinding group (ZBG) which chelates the
central zincꢀion and thus inhibits the enzymatic activity of HDACs, (ii) a linker passing
through the hydrophobic ion channel and (iii) a surface recognition region (head group)
4, 9, 10
interacting with amino acid residues at the substrate binding site.
Clinical problems
2
+
that arose with panꢀHDACi, which inhibit all Zn ꢀdependent HDACs, were due to an
1
1
associated broad spectrum of side effects. In contrast, HDAC6ꢀspecific inhibitors
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 80
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
6
(
HDAC6i) are supposed to have less side effects, because HDAC6 knockꢀout mice are
1
2
fully viable and healthy throughout their lives. The first selective HDAC6i reported was
1
3
14
tubacin 2, but it has a nonꢀdrug like structure with a high lipophilicity (CLOGP = 6.36 )
and a high molecular mass. However, as a research tool it contributed to information about
structureꢀactivityꢀrelationships (SAR) for the generation of better HDAC6ꢀselectivec
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
15, 16
ompounds.
Due to the fact thatHDAC6 has the largest active binding site within the
HDAC family, selectivity for HDAC6 can arise from interactions between a bulky head
10, 14
group and specific amino acid residues within the substrate binding site.
Butler, Kalin et
14
al. first synthesized tubastatin A (3),and they characterized further HDAC6ꢀselective
compounds with a tetrahydroꢀγꢀ or tetrahydroꢀβꢀcarboline structure, showing a very good
14, 17
activity and a high selectivity for HDAC6 (over 1000ꢀfold versus HDAC1).
Most of the currently available HDAC6i coꢀinhibit other HDACs and cause lethal effects
instead of a genuine inhibition of HDAC6. The coꢀinhibition of other HDACꢀfamily
members has led to misinterpretations regarding the biological roles of HDAC6 (see e.g.,
18
ref. ) and cytotoxic events prevent the interpretation of HDAC6 functions in vital cells.
While several agents target HDAC6 in the nanomolar range in vitro, biological effects are
often only described for micromolar and even higher doses, which may additionally block
19ꢀ22
other HDACs (e.g.,
). A promising exception are effects against pancreatic cancer cells.
These have been achieved with drugs targeting HDAC6 and HDAC3 in the nanomolar
2
3
range. One also has to consider the caveat that recently developed drugs that may
2+
specifically impair HDAC6 have often not been tested against all other Zn ꢀdependent
HDACs.
4
ACS Paragon Plus Environment

Page 5 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Here we present novel, highly selective and very active inhibitors of HDAC6. These are
based on a modified tetrahydroꢀβꢀcarboline function, incorporating a hydantoin increment
as the elementary part of the head group. We attached this group to a hydroxamic acid by
different benzyl linkers. Our new inhibitors are genuine HDAC6i that reflect the benign
phenotype of HDAC6 knockꢀout mice. Furthermore, they act very fast, in the nMꢀrange,
and more potently and, to the best of our knowledge, more selective than most previously
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
, 2,4
established agents targeting HDAC6.
HDACi are also able to ameliorate inflammatory processes and conditions in vitro and in
vivo. Various studies show the possibility to treat inflammatory conditions, like arthritis,
with HDACi. Often, beneficial effects are based on reduced cytokine responses that derive
24
from human peripheral blood mononuclear cells under inhibition of multiple HDACs,
2
5
26
27
class I HDACs, HDAC3, and HDAC8. Other cell types that are relevant for the
2
8
29
pathogenesis of RA, like fibroblastꢀlike synoviocytes (FLS) and chondrocytes, display
decreased inflammatory responses under unspecific HDAC inhibition by trichostatin A or
givinostat as well. However, panꢀHDACi may also cause undesired effects due to a
feedback loop, in which the proꢀinflammatory cytokines TNF and ILꢀ1β downregulate
3
0
HDAC5 expression in RAꢀFLS.
Of note, emerging evidence suggests beneficial effects of HDAC6i for the treatment of
chronic immune and inflammatory disorders. As demonstrated with the selective HDAC6
inhibitor tubastatin A (3) in a mouseꢀmodel, antiꢀinflammatory and antirheumatic effects
can be obtained by modulation of HDAC6 activity, at least in part, by modulating ILꢀ6
21
18
expression. However, tubastatin A (3) was recently described to inhibit HDAC10,
making an interpretation of HDAC6 as a drug target in arthritis animal models difficult. A
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 80
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
21
caveat of this study is that tubastatin A (3) also inactivates other HDACs at the high doses
that were used to collect the data.
Here we present novel HDAC6i with superior pharmacological and toxicological properties
in vitro, in cells, and in an animal model of rheumatoid arthritis.
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
6
ACS Paragon Plus Environment

Page 7 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A
NH
O
N
H
O
HEAD group
linker
zinc binding
group (ZBG)
Ac-Lysine
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
O
O
O
H
H
N
N
OH
N
HN
OH
H
O
O
SAHA (1)
O
S
O
N
N
HO
Tubacin (2)
N
H
N
OH
O
Tubastatin A (3)
B
HEAD 1
HEAD 2
HEAD 3
O
O
C
N
D
C
N
D
A
A
B
D
O
B
N
N
N
C
N
H
A
B
N
N
OH
H
N
n
O
OH
H
N
n = 0: 5a-f
n = 1: 6;7
n = 0: 8
n = 1: 9
4
O
n
OH
O
HEAD 4
HEAD 5
N
O
H
X
O
N
D
Y
N
O
7
N
D
O
N
8
A
C
B
N
A
B
C
R 9
N
E
N
N
O
H
H
N
NHOH
OH
N
12-15
1
0
11
OH
O
O
Figure 1. Schematic segmentation of the common pharmacophore of HDACi. A) Typical pharmacophore
of HDACi (left) which is designed according to acetylated lysine, i.e. the biological substrate. The cap group
red) binds at the entrance of the HDAC catalytic cleft, a linker interacts with the cleft, and the polar ZBG
group (blue) chelates the catalytically active zinc ion in the HDAC pocket. These structures are exemplified
by the FDA approved HDACi SAHA (1), tubacin(2), or tubastatin (3). B) Structures and general structures of
compounds investigated (4ꢀ15). To obtain isoenzymeꢀselectivity, different tetracyclic bulky head groups were
investigated. Modifications of head 2 were performed in the following.
(
4
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 80
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
Results
Chemistry
The general synthetic strategy to obtain derivatives bearing head group 2 and derivatives
thereof, and head group 4, which contains an additional benzene ring, respectively, is shown
in scheme 1. The methylꢀ2,3,4,9ꢀtetrahydroꢀ1Hꢀpyrido[3,4ꢀb]indoleꢀ4ꢀcarboxylate
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
1
hydrochlorides 16a-d , respectively 17, were accessible according to literature procedures
31ꢀ35
(see also supporting information schemes S4 and S7).
The sixꢀmembered Dꢀring can be obtained by reaction of 16a-d with 2,5ꢀdioxopyrrolidinꢀ1ꢀ
yl methylcarbamate (18) and subsequent cyclization of the methylꢀ2ꢀ(methylcarbamoyl)ꢀ
2
,3,4,9ꢀtetrahydroꢀ1Hꢀpyrido[3,4ꢀb]indoleꢀ4ꢀcarboxylates 19a-d to yield 20a-d.
As shown in scheme 1, introduction of the benzyl spacer bearing a hydroxamic acid as a
2
+
Zn ꢀchelating function was performed in 4 steps: Alkylation of the indole nitrogen using
3
6
tertꢀbutyl 4ꢀ(bromomethyl)benzoate (21) led to 22a-d. The benzyloxyꢀderivative 22c was
further modified to give 22e and 22f. Acidic removal of the tꢀbutyl protecting group to form
24a-f, BOP 25 catalyzed amidation with NH OTHP 26 and acidic cleavage of the resulting
2
acetals 27aꢀf finally led to the target compounds 5a-f.
8
ACS Paragon Plus Environment

Page 9 of 80
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
Scheme 1: Synthesis of 5aꢀf, by modification of head group 2. Reagents and conditions: (a) MeCN,
EtN(isopropyl) , 2,5ꢀDioxopyrrolidinꢀ1ꢀyl methylcarbamate (18), rt, 16 h. (b) Dioxane, Cs CO , 110 °C, 2 h.
2
2
3
(c) i; DMF, NaH, 0 °C, 15 min. ii; rt, 1 h. (d) Pd/C, H . (e) 2ꢀButanone, 4ꢀ(2ꢀChloroethyl)morpholine
2
hydrochloride (23), ꢀ, 4 d. (f) TFA, rt, 30 min. (g) DMF, rt, 2ꢀ3 h. (h) MeOH, HCl , rt. Following in
aq.
principle the same synthetic pathway, compound 10, bearing head group 4, was prepared.
The structure of 5a, hereafter named Marbostatꢀ100, was confirmed by xꢀray structure
1
13
analysis, in addition to characterization by Hꢀ and CꢀNMR spectroscopic data.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 80
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
In principle by the same way the fusion of the 4ꢀmethylenꢀNꢀhydroxybenzamide element
taken from tubastatin A (3) and head groups 1ꢀ4 was performed (Figure 1, n = 0). For head
groups 2 and 3 an elongation of the spacer was investigated in addition (n = 1). Moreover,
we introduced the hydantoin increment at the tetrahydroꢀγꢀcarboline structure (head 5,
γꢀMarbostat, 11) to compare this system with tubastatin A (3) and Marbostatꢀ100 (5a).
Because of tautomerism of the hydantoin ring system followed by alkylation of the oxygen
of the amide group, we modified the synthetic route as shown in scheme 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
.
Scheme 2: Synthesis of γꢀMarbostatꢀ100 11 bearing an angular fused hydantoin increment Reagents and
conditions: (a) DMF, NaH, tertꢀbutyl 4ꢀ(bromomethyl)benzoate (21). (b) NH OAc, CH NO (c) CHCl , iProp,
4
3
2
3
NaBH , SiO . (d) i) HOAc, Zn, ii) NH OH, iii) THF, HCl. (e) MeOH, SiO , ethyl glyoxalate. (f) i) MeCN,
4
2
4
2
i
EtN( Prop) , 2,5ꢀdioxopyrrolidinꢀ1ꢀyl methylcarbamate (18), rt, 16 h. ii) Cs CO , dioxane, ꢀ, (g) CF COOH.
h) DMF, BOP (25), NH OTHP (26), rt, 2 ꢀ 3 h. (i) MeOH, HClaqu., rt.
2
2
3
3
(
2
10
ACS Paragon Plus Environment

Page 11 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In the enzymatic as well as in the cellular assays Marbostatꢀ100 (5a) exhibited the most
potent and selective HDAC6 inhibitory activity. Therefore, to investigate the
pharmacophore requirements for compounds based on this lead structure in more detail,
derivatives by modification of the Dꢀring system were prepared. Hereby we focused on the
influence of the carbonyl functions, structural elements that lack in the original parental
structure of tubastatin A (3). To obtain derivative12, lacking the amide carbonyl group as
well as derivatives 14a and 14b without the urea carbonylꢀgroup (Figure 3) the synthetic
routes were slightly modified as shown in schemes 3 and 4.
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
To obtain 12, in the first step methyl 2,3,4,9ꢀtetrahydroꢀ1Hꢀpyrido[3,4ꢀb]indoleꢀ4ꢀ
carboxylate (16a) was transformed to the amide 37a by treatment with methylamine in
methanolic solution (Scheme 3). In the same way 37b as an intermediate was obtained. 37a
was reduced using LiAlH and the urea structure in 39 was formed by CDI. The benzylic
4
linking element was introduced by alkylation with tertꢀbutyl 4ꢀ(bromomethyl)benzoate
3
6
t
(21) . After acidic cleavage of the butylbenzoate 40 the hydroxamic acid was introduced in
steps as shown.
2
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 80
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
Scheme 3: Synthesis of 12, an inhibitor based on the basic structure the4ꢀmethylꢀ4,5,6,11ꢀtetrahydroꢀ2,6ꢀ
methano[1,3]diazocino[5,6ꢀb]indoleꢀ3(1H)ꢀon.Reagents and conditions: (a) CH NH (40 % in MeOH), rt, 3 d.
3
2
(
b) LiAlH , THF, 16 h, ꢀ. (c) CDI, THF, 16 h, ꢀ. (d) NaH, THF, rt. (e) CF COOH, CH Cl , 3.5 h, rt. (f) THF,
4
3
2
2
BOP, NH OTHP, rt, 2 ꢀ 3 h. (g) MeOH, HCliProp, rt.
2
To obtain derivatives without the urea carbonylꢀgroup (Scheme 4) the amides 37a and 37b
were treated with formaldehyde 43 to form the methylene bridged Dꢀring system (44a,
4
4b). After alkylation with methyl 4ꢀ(bromomethyl)benzoate 45 the hydroxamic acids (14a,
4b) were formed from the methylesters 46a and 46b using hydroxylamine in alkaline
1
solution.
12
ACS Paragon Plus Environment

Page 13 of 80
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
Scheme 4: Synthesis of inhibitors based on the 4ꢀmethylꢀ3,4,6,11ꢀtetrahydroꢀ2,6ꢀmethano[1,3]diazocino[5,6ꢀ
b]indoleꢀ5(1H)ꢀon – basic structure exemplified by compounds 14a and 14b.Reagents and conditions:
(
a) (CH O), CH OH, 70 °C, 2 h. (b) NaH, THF, rt, 30 min. (c) NH OH, CH OH, KOH, 0 °C.
2
3
2
3
The thiourea analoga of Marbostatꢀ100 (5a), compound 15 (Figure 3), can be obtained by
alkaline cleavage of the ester 16a, reaction of the resulting 2,3,4,9ꢀtetrahydroꢀ1Hꢀ
3
7
pyrido[3,4ꢀb]indoleꢀ4ꢀcarboxylic acid (47) with methyl isothiocyanate and ring closure by
3
8
use of CDI , followed by the steps used for the synthesis of Marbostatꢀ100 (5a)
scheme 5).
(
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 80
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
.
Scheme 5: Synthesis of 16, the thiourea derivative of Marbostatꢀ100 (5a) Reagents and conditions: (a) LiOH,
H O, MeOH, THF, rt, 16h, ꢀ, 2h, HOAc. (b) methylisothiocyanate, aceton, DMSO, ꢀ, 2 h. (c) CDI, CH CN,
2
3
rt, 30 min. (d) NaH, DMF, rt. (e) CF COOH, rt, 30 min. (f) NH OTHP, BOP, DMF, NEt , (g) HClaqu., MeOH,
3
2
3
H O.
2
Moreover, thioamide derivative 13 (scheme 6) was prepared by reaction of 20a with
Lawesson`s reagent and the resulting 4ꢀmethylꢀ3ꢀthioxoꢀ3,4,6,11ꢀtetrahydroꢀ2,6ꢀ
methanoindeno[2,1ꢀe][1,3]diazocinꢀ5(1H)ꢀone (53) was transformed to the final test
compound 13 in four steps.
.
Scheme 6: Synthesis of 13, a thioamide derivative of Marbostatꢀ100 (5a) Reagents and conditions:
a) Lawesson's reagent, THF, ꢀ, 48 h. (b) 21, NaH, DMF, rt. (c) CF COOH, rt, 30 min. (d) NH OTHP, BOP,
(
3
2
DMF, NEt , (e) HCl , MeOH, H O.
3
aqu.
2
14
ACS Paragon Plus Environment

Page 15 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
To investigate the influence of the stereochemistry of head group 2 on the biological
3
9
activity, a stereoselective synthesis of Marbostatꢀ100 (5a) by chiral catalysis was
developed. Using 1.5 equiv. of indole (57a) and an enhanced catalyst loading of 25 mol%
of the respective catalyst 59b prepared in situ, the reported enantiomeric excess of 95 % for
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
39
the formation of (S)ꢀethylꢀ2ꢀ(1Hꢀindolꢀ3ꢀyl)ꢀ3ꢀnitropropanoate (60b) by coupling of 57a
with 58 was optimized to an enantioselectivity ≥99 %. Following the same protocol, also
the enantiomeric Rꢀderivative 60a was obtained in a yield of 50 % and an ee of ≥ 99 %.
Based on the synthetic route shown in Scheme 1, the synthesis of (R)ꢀ and (S)ꢀ
Marbostatꢀ100 (R-5a and S-5a) was performed from 60a and 60b in the following as shown
in the supporting information scheme 8. To avoid racemization at least in part at the CH
acidic stereocenter, the reaction conditions were modified.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 80
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
Scheme 7: A: Enantioselective synthesis of (R)ꢀ and (S)ꢀMarbostatꢀ100.The enantiomeric excess was
determined by HPLC analysis.To confirm the stereochemistry, structures of (R)ꢀethyl 2,3,4,9ꢀtetrahydroꢀ1Hꢀ
pyrido[3,4ꢀb]indoleꢀ4ꢀcarboxylate (62a) crystallized with (ꢀ)ꢀdiꢀpꢀtoluolylꢀLꢀtartaric acid from MeOH(B), and
of (S)ꢀethyl 2,3,4,9ꢀtetrahydroꢀ1Hꢀpyrido[3,4ꢀb]indoleꢀ4ꢀcarboxylate (62b) as
a hydrobromide (C),
respectively were determined by xꢀray diffraction analysis.
HDAC6 inhibition and selectivity of Marbostat-100 and derivatives
We synthesized a series of hydroxamic acids bearing tetrahydrocarbolines as parts of
different tetracyclic head groups and investigated whether different linker groups enhance
HDAC6 selectivity which was derived from K values of HDAC6, HDAC2 and HDAC8
i
inhibition measured in assays using recombinant purified human HDAC proteins
16
ACS Paragon Plus Environment

Page 17 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1. K ꢀvalues [nM] for HDACꢀsubtypes HDAC2, HDAC6 and HDAC8 compared to
the panꢀHDACi Trichostatin A (TSA) (Enzymatic Inhibitory Activities)
i
a
Head group 1
Spacer
Cpd. No.
4
HDAC2
HDAC6
HDAC8
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
6 097
6.51
1 540
Head group 2
Marbostat-
774
0.70
173
1
00 (5a)
R-5a
S-5a
3 195
5 205
2.01
0.70
294
999
6
7
66.51
768
4.15
205
72.24
13.09
2
4a
>10 000
>20 000
>40 000
Head group 3
8
9
7 244
234
30.20
14.13
1678
720
Head group 4
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 80
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
1
0
ND
1.78
ND
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
Head group 5
O
N
O
N
11
995
1.86
761
N
TSA
3.55
0.58
7.56
145
Tubastatin A
> 1 130
1 697
a
Compounds were tested in a 10ꢀdose IC mode with 3ꢀfold serial dilution starting from 100 ꢁM solutions
50
for HDAC2 and HDAC8, respectively, starting from 1 ꢁM solutions for HDAC6. IC50 values were extracted
by curveꢀfitting of the dose/response slopes and K values were calculated from the respective IC values,
i
50
40
determined in duplicate, according to the Cheng–Prusoff equation as given in the SI. TSA was used as an
internal standard. ND: Not Determined. Assays were performed by Reaction Biology Corporation, USA.
All compounds inhibit HDAC6 preferentially (Table 1). Compound 4 (Figure 1),
featuring a methyleneꢀbridged tetrahydroꢀßꢀcarboline in conjunction with a benzyl linker,
has an HDAC6 Ki value of 6.51 nM with approximately 940ꢀfold selectivity for HDAC6
over HDAC2. Modification of the Dꢀring system and introduction of a hydantoin increment
(
Figure 3), combined with the addition of a benzyl linker led to Marbostatꢀ100 (5a) with an
Ki of 0.7 nM for HDAC6 and conserved selectivity for HDAC6 (Table 1). S-5a represents
the eutomer (Ki, 0.7 nM), but the eudismic ratio amounts to only 0.35 (Ki of R-5a, 2.01
nM) (Table 1).
18
ACS Paragon Plus Environment

Page 19 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2: Influence the stereochemistry and of modifications of head group 2 in the
enzymatic fluorescence assay on inhibition of HDACꢀsubtypes HDAC2, HDAC6 and
a
HDAC8 (K ꢀvalues [nM])
i
Modifications of head group 2
Cpd. No.
HDAC2
HDAC6
HDAC8
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
R
X
Y
8ꢀOCH3
8ꢀOꢀBz
C=O
C=O
C=O
C=O
C=O
C=O
5b
5c
5e
310
ND
ND
0.15
0.08
0.42
124
ND
ND
8
ꢀOH
8
ꢀ(2ꢀmorpholinoethoxy)
x HCl
C=O
C=O
5f
313
0.48
721
9
ꢀOCH3
C=O
CH2
C=O
C=O
CH2
CH2
C=O
C=S
5d
12
70.8
484
0.32
0.51
0.23
0.16
0.88
0.45
93.3
528
427
299
701
40.4
H
H
C=O
C=O
C=S
C=O
14a
14b
13
1 356
370
8ꢀOCH3
H
H
943
15
469
a
Compounds were tested in a 10ꢀdose IC mode with 3ꢀfold serial dilution starting from 100 ꢁM solutions
50
for HDAC2 and HDAC8, respectively, starting from 1 ꢁM solutions for HDAC6. IC50 values were extracted
by curveꢀfitting of the dose/response slopes and K values were calculated from the respective IC values,
i
50
40
determined in duplicate, according to the Cheng–Prusoff equation as given in the SI. TSA was used as an
internal standard. ND: Not Determined. Assays were performed by Reaction Biology Corporation, USA.
Connecting the tetracylic head group 2 with a paraꢀ or orthoꢀsubstituded cinnamic acidꢀNꢀ
hydroxamide (compounds 6 and 7 Table 1) significantly reduces HDAC6 inhibition and
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 80
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
selectivity. The angular fused Dꢀring system in compounds 8 and 9 lowers activity
compared to the bridged tetracylic system of Marbostatꢀ100 (5a), probably because of steric
hindrance. Reduced activity can also be observed for compound 10, where an additionally
fused Eꢀring might hinder penetration of the drug to the channelꢀrim. The paraꢀsubstituded
cinnamic acidꢀNꢀhydroxamide in case of 9 enhances activity at HDAC2, HDAC6 and
HDAC8 compared to 8, but there is still a preference for HDAC6. Elongation of the linker
possibly compensates steric hindrance due to the head group and allows better penetration
into the catalytic channel rim. As expected, the carboxylic acid 24a (Scheme 1), the
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
+
analogue of Marbostatꢀ100 (5a) lacking the hydroxamic acid asZn ꢀchelating group,
showed no activity on HDAC2 and HDAC6 and a significantly reduced potency on
HDAC8.
Marbostatꢀ100 (5a), the most potent compound in this series, served as lead for further
modifications of head group 2. The introduction of a methoxy group at Cꢀ8 enhances
activity towards HDAC6 in the enzymatic assay (5b, 14b) (Table 2). Substitution at Cꢀ8 by
OH (5e) and by bulkier benzyloxy (5c) or 2ꢀmorpholinoethoxy (5f) groups is also tolerated
as well as a methoxysubstituent at Cꢀ9 (5d). Replacing one of the carbonyl moieties by a
sulfocarbonyl or a CH group does not significantly affect HDAC6 inhibition (13, 14a, 14b,
2
15).
To further validate the selectivity of this series of compounds, Marbostatꢀ100 (5a) was
profiled against a broader panel of isoenzymes. As shown by the data in Table3,
Marbostatꢀ100 (5a
)
exhibited significant selectivity for HDAC6 over all nine
2
+
Zn dependent HDAC isoenzymes investigated, ranging from about 250ꢀfold vs. HDAC8 to
20
ACS Paragon Plus Environment

Page 21 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
600ꢀfold vs. HDAC4. The HDAC6 selectivity of tubastatin A (3), the most selective
inhibitor to date, in contrast is about 150ꢀfold (vs. HDAC8), with respect to the K ꢀvalues
i
determined in our test system.
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
Table 3: Enzyme inhibition data (K ꢀvalues [nM]) for Marbostatꢀ100 HDACꢀsubtypes
HDAC1, HDAC2, HDAC3, HDAC4, HDAC5, HDAC6, HDAC7, HDAC8, HDAC10 and
HDAC11 compared to the panꢀHDACi TSA and the highly potent, selective and cellꢀ
i
4
1
a
permeable class IIa HDAC inhibitor TMP 269, as reference compounds .
HDAC Subtype
Compound
Marbostat-
1
2
3
4
5
6
7
8
10
11
3
58
774
306
1821
701
0.7
442
173
452
495
1
00 (5a)
TSA
0.64
ND
3.55
ND
0.84
ND
ND
ND
0.58
ND
ND
145
4.76
ND
401
ND
TMP 269
90.5
86.9
46.3
2706
a
Compounds were tested in a 10ꢀdose IC mode with 3ꢀfold serial dilution starting from 100 ꢁM solutions
50
for all HDAC subtypes and starting from 1 ꢁM solutions for HDAC6, respectively. IC50 values were extracted
by curveꢀfitting the dose/response slopes and K values were calculated from the respective IC values,
determined in duplicate, according to the Cheng–Prusoff equation as given in the SI. TSA and TMP 269
were used as an internal standard. ND: Not Determined. Assays were performed by Reaction Biology
Corporation USA.
i
50
40
Binding Mode of Marbostat-100 (5a)
4
2,43
Recently, Christianson et al.
released 18 crystal structures of HDAC6 (one human
catalytic domain (CD) II, one zebrafish CD I, 16 zebrafish CD II), mostly in complex with
an inhibitor or a substrate analog, in the Brookhaven Protein Databank (PDB). The
ultrahigh resolution structure of zebrafish HDAC6 CD II (PDB 5WGK) in complex with Nꢀ
hydroxyꢀ4ꢀ[[(2ꢀhydroxyethyl)(phenylacetyl)amino]methyl]benzamide is most suited for
44
docking approaches, since CD II is the major functional catalytic domain of HDAC6 and
4
3
since the binding mode of the phenylhydroxamic acid moiety is provided. Although the
sequence identity of human and zebrafish HDAC6 amounts to only 40%, the latter species
isoform may be used for docking of Marbostatꢀ100 because of 100% identical binding sites
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 80
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
and a root mean square deviation of the backbone of 13 amino acids 3.5Å around
Marbostatꢀ100 of only 0.27Å.
Considering that the K values of (R)ꢀ and (S)ꢀMarbostatꢀ100 differ by a factor not greater
I
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
than 2.9 (Table 1), both enantiomers were manually docked in energy minimum
conformations close to the xꢀray structure (supporting information) retaining the binding
4
3
mode of the phenylhydroxamic acid moiety present in the PDB structure 5WGK. The
detailed binding modes are shown in Figure 2 (amino acid numbers correspond to human
HDAC6). Both isomers do not differ with respect to the docking mode of the hydroxamic
acid moiety and the benzyl linker. The former binds in a rather unusual monodentate
2
+
36 43
2+
hydroxamateꢀZn coordination mode . The Zn ꢀion is coordinated with D649, H651
2
+
and D742 (Zn ꢀO/N distances ca. 2.0 Å), with a catalytic water molecule and with the
hydroxylamine oxygen of Marbostatꢀ100 (5a (distance 2.0 Å). The hydroxylamine oxygen
)
forms a hydrogen bond with the hydroxy group of Y782. Furthermore, the catalytic water
molecule links H610, H611 and the amide oxygen of the hydroxamate group via hydrogen
bonds. The benzyl linker penetrates a partially hydrophobic channel (Figure 2) and contacts
the side chains of F620, F680 and Y782 as well as G619.
The “surface recognizing cap group” of both isomers, represented by the 4ꢀmethylꢀ6,11ꢀ
dihydroꢀ2,6ꢀmethano[1,3]diazocino[5,6ꢀb]indoleꢀ3,5(1H,4H)ꢀdione moiety, is optimally
fitted onto the entrance of the binding pocket. Figure 2C shows an extended hydrophobic
surface region, mainly formed by P501 and L749, and a small hydrophobic cavity with
F679 at the bottom. In case of the (R)ꢀisomer, the 4ꢀmethyl group fits between P501 and
L749, and the diazocino moiety contacts P501 and F620. The phenyl plane covers the
cavity forming van der Waals contacts with H651 and F680. The (S)ꢀisomer binds in a
22
ACS Paragon Plus Environment

Page 23 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
laterally reversed position. The phenyl plane faces P501, L749 and F620. The edge of the
diazocino moiety contacts the phenyl planes of F679 and F680, and the 4ꢀmethyl group fits
between the side chains of F679 and L749. Both carbonyl groups may be involved in waterꢀ
mediated hydrogen bonds with the imidazolyl NH of H651 and the backbone O of L749 on
one hand and with the backbone O of P748 on the other hand. Possibly, these additional
interactions account for the slightly higher affinity of the (S)ꢀisomer (eutomer). Both
binding modes correspond to the structureꢀactivity relationships of Marbostatꢀ100
derivatives substituted at the phenyl moiety (Table 1), since the substituents may project
into the solvent, leading to only small variations of the affinity. Moreover, at least one
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
diazocino carbonyl group may be exchanged by a thiocarbonyl moiety or by CH without
2
significant effects on interactions.
Characterization of the Novel HDACi in Cellular Systems
Having assessed the activities of Marbostatꢀ100 (5a
) and further derivatives in vitro (Tables
1
ꢀ2), we tested these compounds for biological activities. HDAC6 has a number of different
4
5
46
47
cytosolic substrates, such as αꢀtubulin, HSP90, betaꢀcatenin and Akt.. A reliable
marker for the inhibition of HDAC6 is hyperacetylation of tubulin, as tubulin is a bona fide
4
target of the cytoplasmic HDAC6 in cells. We assessed this parameter by Western blot.
4
Due to the high relevance of HDAC6 in hematopoietic cells, we used such cells for our
assays.
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 80
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. Probable binding mode of Marbostat-100 at CD II of human HDAC6. Detailed interactions of
A (R)ꢀMarbostatꢀ100, B (S)ꢀMarbostatꢀ100 with HDAC6, shown are all amino acids within 3.5 Å around the
inhibitor; atom colors N, blue, O, red, Zn , magenta, C and polar H of Marbostatꢀ100, grey, of HDAC6,
cyan; backbone of the binding site, light cyan (αꢀhelices and βꢀstrands, ribbons, loops, tubes), w, water
molecules, hydrogen bonds, red dashed lines. C Binding site represented by the lipophilic potential mapped
onto a MOLCAD Connolly surface (hydrophobic areas, brown, polar areas, green and blue), (R)ꢀ and (S)ꢀ
Marbostatꢀ100, ball and stick models (for clarity, the phenylhydroxamic acid moiety of (S)ꢀMarbostatꢀ100 is
not shown).
2+
24
ACS Paragon Plus Environment

Page 25 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
To define structural requirements for a potent inhibition of HDAC6 in vivo, we analyzed
Marbostatꢀ100 (5a) and some derivatives thereof (5b-5f, 12, 13, 14a, 14b and 15). We
applied these agents to exponentially growing human MV4ꢀ11 cells, which are derived from
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
48
an acute myeloid leukemia case. In contrast to the enzymatic in vitro test, a significant
structural dependency on the modifications of head group 2 can be observed in cellular
assays (Table 1). We found that the urea carbonylꢀfunction in the Dꢀring system (Figure 3)
is essential to block HDAC6, as shown by the induction of tubulin peracetylation mediated
by compounds 5b, 5d, 12 and 13. Compounds 14a, 14b and 15 lack this structural element
(figure 3) and fail to inhibit HDAC6 (Figure S1A). Compounds 12 and 13, which possess
one urea carbonyl function, exhibit slightly reduced activity in comparison to
Marbostatꢀ100 (5a) and its derivatives 5b and 5d with two carbonyl groups. The 8ꢀ
(morpholineethoxy) derivative 5f with enhanced solubility as well exhibited conserved
selectivity and conserved potency.
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 80
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 3. Structureꢀactivity relationship of 5a and derivatives. The urea carbonyl function (magenta;
compounds 12 and 13) is essential for cellular activity. Compounds 14a and 14b exhibit a methyleneꢀgroup,
respectively 15 a sulfur instead this structural element and lack activity on the cellular level.
To ascertain the selectivity of Marbostatꢀ100 (5a
) for HDAC6, we probed for histone
hyperacetylation. HDAC6ꢀspecific drugs should not induce this posttranslational
49
modification, because HDAC6 does not recognize histones as substrates in vivo. Since
4
class I HDACs catalyze the removal of the acetyl group from histones, we included the
benzamide HDACi MSꢀ275 (entinostat) as test compound. MSꢀ275, only inhibits class I
5
0,51
HDACs and can be used as a positive control for histone hyperacetylation.
accumulation of hyperacetylated tubulin and a lack of histone hyperacetylation is observed
for both enantiomers of Marbostatꢀ100 (5a (Figure S1b).
Potent
)
Since the data above illustrate that Marbostatꢀ100 is the most effective new HDAC6i in
enzymatic tests and in cells, we focused our following analyzes on this compound. We
26
ACS Paragon Plus Environment

Page 27 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
treated MV4ꢀ11 cells with 5 nM to 1 µM Marbostatꢀ100 (5a) for 24 h. Western Blot
analyzes revealed that 50 nM Marbostatꢀ100 (5a
hyperacetylated tubulin potently (Figures 4AꢀB). These in vivo findings confirm the in vitro
data from the biochemical fluorescence assays (Table 1), which identify Marbostatꢀ100 (5a
) evoked the accumulation of
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
)
as potent HDAC6i. The roughly tenfold difference between the efficacy of Marbostatꢀ100
in vitro and in cells positions Marbostatꢀ100 as a highly potent new HDAC6i. Such
differences between drug activity in vitro and in cells are typically also seen with highly
potent tyrosine kinase inhibitors.
Next, we determined the effects of Marbostatꢀ100 (5a
MV4ꢀ11 cells with doses from 50 nM to 10 µM Marbostatꢀ100 (5a
were tested for hyperacetylated tubulin (Figures 4C and D). While all doses cause an
accumulation of hyperacetylated tubulin, even high doses of Marbostatꢀ100 (5a do not
)
in higher dose ranges. We treated
)
. Lysates of these cells
)
trigger a hyperacetylation of histones, which verifies Marbostatꢀ100 as a very specific
HDAC6i (Figure 4C). Only a concentration of 10 µM leads to a degradation of HDAC6;
this effect is not linked to a further increase in the acetylation of tubulin (Figure 4C). In
contrast, MSꢀ275 induces hyperacetylated H3 in MV4ꢀ11 cells (Figure 4A).
To exclude cellꢀtype dependent effects, we additionally tested Marbostatꢀ100 (5a
human leukemia cell lines Jurkat, HEL, and BV173, and in the human embryonic kidney
cell line HEK293T. Results collected in these systems confirm that Marbostatꢀ100 (5a is
) in the
)
an inhibitor of HDAC6 at nanomolar concentrations (supporting information Figures S2Aꢀ
F).
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 80
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
Subsequently, we analyzed the kinetics of tubulin hyperacetylation in Marbostatꢀ100 (5a
)
treated MV4ꢀ11 cells. These were incubated with 50 nM Marbostatꢀ100 (5a
to 48 h. Already after 30 minutes, Marbostatꢀ100 (5a causes an accumulation of
) for 10 min up
)
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
hyperacetylated tubulin in MV4ꢀ11 cells (Figure 4E, F). To confirm these results, the same
tests were performed in BV173 cells and yielded similar observations (supporting
information, Figure S2G, H).
To further ascertain the specificity of Marbostatꢀ100 (5a
overexpression model using HDAC6ꢀFLAG and pcDNA3.1 as empty vector control. When
we treated these cells with increasing doses of Marbostatꢀ100 (5a , we noted that no further
increase in tubulin acetylation could be detected with concentrations above 500 nM
), we used a HEK293T cell
)
Marbostatꢀ100 (5a
)
(supporting information, Figure S2I). We therefore treated the
4
transfected HEK293T cells with 500 nM Marbostatꢀ100 (5a
)
. As expected, the
overexpression of HDAC6 reduces the levels of endogenous tubulin acetylation
Figure 4G). Treatment of HEK293T cells expressing pcDNA3.1 or HDAC6ꢀFLAG with
Marbostatꢀ100 (5a augments the acetylation of tubulin (Figure 4G). These data confirm
that Marbostatꢀ100 (5a can antagonize HDAC6ꢀdependent effects in vivo.
(
)
)
Tubastatin A (3) (Figure 1) is currently one of the most frequently used drugs to inhibit
4
HDAC6. Therefore, we compared Marbostatꢀ100 (5a
)
with tubastatin A (3) and with
LBH589 (panobinostat). The hydroxamic acid and panꢀHDACi LBH589 is currently tested
1
4
in several clinical trials and one of the most potent panꢀHDACi. Remarkably, at equimolar
doses, Marbostatꢀ100 (5a is superior to tubastatin A (3) regarding both the strength as well
as the duration of tubulin hyperacetylation in MV4ꢀ11 cells (Figure 4H). Tubastatin A (3)
)
only leads to a similar hyperacetylation of tubulin as Marbostatꢀ100 (5a) when used in100ꢀ
28
ACS Paragon Plus Environment

Page 29 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
fold higher concentrations than Marbostatꢀ100 (5a
tubastatinA (3)) (supporting information, Figure S2J). Thus, Marbostatꢀ100 (5a
over tubastatin A (3). In comparison to LBH589, one of the most potent pan HDACi that is
)
(50 nM Marbostatꢀ100 (5a
)
vs. 5 µM
)
is superior
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
5
0, 51
also highly active against HDAC6,
Marbostatꢀ100 (5a) evokes the hyperacetylation of
tubulin more rapidly and efficiently. In addition, the treatment with LBH589 induces
acetylated histone H3 and a degradation of HDAC6 after 24 h (Figure 4H). An equimolar
dose of Marbostatꢀ100 does not produce this proteolytic effect (Figure 4H).
Based on these findings, we tested if Marbostatꢀ100 (5a
andMSꢀ275 lead to cell cycle alterations and cell death. As presented in Figure 4I,
Marbostatꢀ100 (5a , in contrast to LBH589 and MSꢀ275, did not cause such effects.
), tubastatin A (3), LBH589,
)
LBH589 and MSꢀ275 led to changes in cell cycle distribution and higher amounts of cells
with fragmented DNA (subG1 fraction; Figure 4I).
To corroborate these data, we treated further cell lines and normal peripheral blood
mononuclear cells (PBMCs) with Marbostatꢀ100 (5a
alterations and cell death. Again, we used Marbostatꢀ100 (5a
and compared its effects with those induced by MSꢀ275. While MSꢀ275 dysregulates the
cell cycle and triggered apoptosis in these cells, Marbostatꢀ100 (5a turns out as a nonꢀ
)
and investigated them for cell cycle
)
in a doseꢀescalating manner
)
cytotoxic, genuine inhibitor of HDAC6 (Figure 4I and supporting information Figure S2A,
B).
For further drug assessment, we treated male DBA/1J mice with 0.6 mg Marbostatꢀ100 (5a
)
for 15 days. Livers and kidneys from treated and control mice were examined for
hyperacetylated tubulin by Western Blot. Figure 4J shows that tubulin is hyperacetylated in
mice that received the inhibitor. Hence, Marbostatꢀ100 (5a) also impairs HDAC6 activity in
29
ACS Paragon Plus Environment