Potent and selective HDAC6 inhibitory activity of N-(4- hydroxycarbamoylbenzyl)-1,2,4,9-tetrahydro-3-thia-9-azafluorenes as novel sulfur analogues of Tubastatin A

Article abstract of DOI:10.1039/c3cc41422a

Eight N-(4-hydroxycarbamoylbenzyl)-1,2,4,9-tetrahydro-3-thia-9-azafluorenes were efficiently prepared as sulfur analogues of Tubastatin A and thus evaluated as new HDAC6 inhibitors. All compounds exhibited potency against HDAC6, and four of them were active in the nanomolar range (IC₅₀ = 1.9-22 nM). Further analysis revealed that the sulfone derivatives (designated as Tubathians) are superior to their non-oxidized sulfide analogues, and the two most active sulfones showed good to excellent HDAC6 selectivity compared to all other HDAC isoform classes.

Full text of DOI:10.1039/c3cc41422a

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Potent and selective HDAC6 inhibitory activity of
N-(4-hydroxycarbamoylbenzyl)-1,2,4,9-tetrahydro-
3-thia-9-azafluorenes as novel sulfur analogues
of Tubastatin A†
Cite this: Chem. Commun., 2013,
49, 3775
Received 24th February 2013,
Accepted 7th March 2013
Rob De Vreese,a Tom Verhaeghe,b Tom Desmetb and Matthias D’hooghe*a
DOI: 10.1039/c3cc41422a
www.rsc.org/chemcomm
Eight
N-(4-hydroxycarbamoylbenzyl)-1,2,4,9-tetrahydro-3-thia- isozymes may cause undesirable side effects. Thus, the design
9-azafluorenes were efficiently prepared as sulfur analogues of and development of isozyme-selective inhibitors has emerged
Tubastatin A and thus evaluated as new HDAC6 inhibitors. All as an important challenge within the search for novel HDACIs.5
compounds exhibited potency against HDAC6, and four of them
In recent years, HDAC6 has been acknowledged as an attrac-
were active in the nanomolar range (IC50 = 1.9–22 nM). Further tive target for drug development,6 and an increasing number of
analysis revealed that the sulfone derivatives (designated as research teams are currently involved in the quest for new
Tubathians) are superior to their non-oxidized sulfide analogues, compounds endowed with HDAC6 inhibitory activity.7 In addi-
and the two most active sulfones showed good to excellent tion to the potential of HDAC6-selective inhibitors for applica-
HDAC6 selectivity compared to all other HDAC isoform classes.
tions in the treatment of CNS disorders and neurodegenerative
diseases, these compounds seem to provoke fewer side effects,
The enzymatic addition and removal of acetyl groups at specific hence the growing interest in their preparation.8 An important
lysine residues comprise important biochemical reactions with a milestone in that respect concerns the identification of Tubacin
significant impact on many cellular processes.1 The addition of as a selective HDAC6 inhibitor, although the application of this
acetyl groups within histone proteins, the chief protein compo- compound is hampered by its poor druglikeness and cumbrous
nents of chromatin, is catalyzed by histone acetyltransferases synthesis.9 Since then, considerable advances have been made
(HAT), and histone deacetylases (HDAC) mediate the corre- with regard to the preparation of new HDAC6 inhibitors, leading
sponding deacetylation reactions. The inhibition of the latter to an array of different molecular entities with improved
group of deacetylases has become a hot topic in medicinal chemical and pharmacological properties. From a chemical view-
chemistry, and the use of HDAC inhibitors (HDACIs) has found point, many of these molecules comprise the typical HDACI basic
many applications with regard to cancer and CNS disorder structure accommodating an aromatic cap group (surface recog-
therapies.2 In general, HDACIs act on 11 zinc-dependent HDAC nition domain), a linker and a zinc-binding hydroxamic acid unit.
isozymes, which are divided into four groups: class I (HDACs 1, 2, A major breakthrough was accomplished recently, involving the
3, 8), class IIa (HDACs 4, 5, 7, 9), class IIb (HDACs 6, 10), and class rational design and synthesis of Tubastatin A as a novel and
IV (HDAC11).3 The majority of known HDACIs primarily inhibit selective HDAC6 inhibitor.10 Elaborate studies in this direction
the class I enzymes, making them excellent candidates for cancer showed that the HDAC6 isozyme tolerates modifications of the
therapy applications, but other than class I HDACIs are normally Tubastatin A chemical structure at the level of the cap group and,
required for the pursuit of non-oncological applications.4 Another more specifically, that the introduction of structural diversity at
important issue relates to the potential toxicity of compounds the 2- and 8-position of the tetrahydropyrido[4,3-b]indole scaffold
inhibiting multiple isozymes, as acetylation is involved in can be beneficial with regard to the overall bioactivity.11
the control of many cellular processes and inhibition of some
Inspired by these recent SAR findings, and intrigued by the fact
that several new HDAC6 inhibitors contain a sulfur atom in their
molecular structure,7c–e efforts were made toward the preparation
of a number of sulfur analogues (sulfides and sulfones) of
Tubastatin A in the present study, supported by HDAC6 ligand
docking. Furthermore, the replacement of a methylene group with a
sulfone moiety in medicinally relevant compounds has been shown
to induce a significant beneficial increase in stability,12 suggesting
this modification as a preferred change during compound optimi-
zation and thus providing an additional rationale for the work
a SynBioC Research Group, Department of Sustainable Organic Chemistry and
Technology, Faculty of Bioscience Engineering, Ghent University,
Coupure Links 653, B-9000 Ghent, Belgium. E-mail: matthias.dhooghe@UGent.be;
Fax: +32-9-2646221; Tel: +32-9-2649394
b Centre for Industrial Biotechnology and Biocatalysis, Faculty of Bioscience
Engineering, Ghent University, Coupure Links 653, B-9000 Ghent, Belgium
† Electronic supplementary information (ESI) available: Synthetic procedures and
spectral data of compounds 3, 4, 5, 6, 7 and 8, and bioassay results. See DOI:
10.1039/c3cc41422a
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 3775--3777 3775

View Article Online
Communication
ChemComm
structure of HDAC isozymes as a template. Compounds that do
not carry a methoxy group on their linker (5a, 5c, 8a and 8c) were
found to fit perfectly in the active site of HDAC6 (Fig. 1). In this
case, the linker is positioned in the tubular access channel, with
the carbonyl group of the hydroxamate moiety within chelating
distance from the zinc ion at the bottom of the pocket. As the
linker fills the access channel almost completely, very little space is
left to accommodate a (bulky) substituent such as a methoxy
group, which is in line with previous studies in that respect.
In contrast, modifications of the tricyclic cap group do not seem
to influence the binding mode very much, since the conformation
and orientation of compounds 5a, 5c, 8a and 8c are nearly
identical. However, oxidation of the sulfur atom results in addi-
tional interactions with the enzyme in the form of hydrogen bonds
between the introduced oxygen atoms and the backbone nitrogen
of residues Asp567 and Gly619 (Fig. 1b). The latter observation
provided an interesting motive and an additional reason to
experimentally assess the HDAC6 inhibitory activity of Tubastatin
A analogues in which the NMe moiety is replaced by a sulfone unit.
In vitro pharmacology studies of novel hydroxamic acids 5a–d
and 8a–d with regard to their HDAC1 and HDAC6 inhibitory
activity revealed an interesting potency of these compounds as
HDAC6 inhibitors (see ESI†). In particular, hydroxamic acids 5a,
5c, 8a and 8c showed complete inhibition at a test concentration
of 10 mM, and also compounds 8b and 8d exhibited a good profile
with an inhibition of 73% and 75%, respectively. In addition,
these results pointed to a selectivity of the test compounds
toward HDAC6 inhibition, with HDAC1 inhibition percentages
ranging from 0% to a maximum of 53%. Furthermore, these data
also indicate a detrimental effect of the introduction of a methoxy
group in the linker moiety on the bioactivity (compounds 5b,d
and 8b,d), as indicated by homology modeling. HDAC1 and
HDAC6 were chosen for activity comparison in this preliminary
test, as these two enzymes have a diverse phylogeny and are
members of separate deacetylase classes.
Scheme 1 Synthesis of N-(4-hydroxycarbamoylbenzyl)-1,2,4,9-tetrahydro-3-thia-
9-azafluorenes 5 and their oxidized analogues 8 (R1 = H, F; R2 = H, MeO).
undertaken in this study. The results obtained point to the potential
of sulfur analogues of Tubastatin A as new HDAC6 inhibitors,
especially those containing a sulfone moiety in their structure.
The 1,2,4,9-tetrahydro-3-thia-9-azafluorene scaffold was prepared
via a bismuth nitrate-promoted Fisher indole synthesis employing a
phenylhydrazine hydrochloride 1 and tetrahydrothiopyran-4-one 2,
providing a convenient access to tetrahydro-3-thia-9-azafluorenes 3
in good yields (Scheme 1).13 Subsequently, N-benzylation of com-
pounds 3 was accomplished using a methyl 4-(bromomethyl)-
benzoate in DMF in the presence of sodium hydride and
potassium iodide, furnishing the corresponding N-(4-methoxy-
carbonylbenzyl)-1,2,4,9-tetrahydro-3-thia-9-azafluorenes 4. The
final step of the process comprised an ester to hydroxamic acid
interconversion, which was realized utilizing an excess of
hydroxylamine hydrochloride in the presence of methanolic
sodium methoxide in DMF. In this way, the premised
N-(4-hydroxycarbamoylbenzyl)-1,2,4,9-tetrahydro-3-thia-9-azafluorenes
5 were obtained in an efficient and straightforward approach.
Considering the presence of a (cyclic) sulfone moiety in several
drugs and bioactive compounds,12 the sulfide in systems 3 was
oxidized to the corresponding sulfones 6 by means of meta-
chloroperbenzoic acid treatment in tetrahydrofuran. The thus
obtained sulfones 6 were taken further in the synthesis toward
the contemplated hydroxamic acids 8 via esters 7 applying a
similar strategy to that discussed above for the preparation of
hydroxamic acids 5 (for structural details and reported yields of
compounds 3–8, see ESI†).
The most promising molecules (those showing an inhibition
of >70%) were then selected for determination of their IC50
values with respect to HDAC6 inhibition (Table 1). These assess-
ments confirmed the presumption that molecules bearing a
methoxy-substituted linker exhibit lower – but still moderate –
activities, exemplified by compounds 8b and 8d (with IC50 values
of 2.0 and 1.3 mM, respectively). Furthermore, sulfur oxidation
indeed seems to be beneficial for bioactivity, as sulfones 8a and
8c show even more potent HDAC6 inhibition as compared to
sulfides 5a and 5c. Overall, four compounds (5a, 5c, 8a and 8c)
can be considered to be promising lead templates for further
elaborate studies. Sulfides 5a and 5c (with IC50 values of 15 and
22 nM, respectively) display HDAC6 inhibitory activities similar
to the reference compound Trichostatin A and to Tubastatin A,14
The binding of various ligands in the enzyme’s active site
was evaluated by means of automated docking. Since the crystal
structure of HDAC6 is not available, a homology model was first
generated following the example of Kozikowski10 using the
Table 1 IC50 values for HDAC6 inhibitiona
Compound
IC50 (mM)
Compound
IC50 (mM)
5a
5c
8a
0.015
0.022
0.0019
8b
8c
8d
2.0
0.0037
1.3
Fig. 1 Docking of compound 8a in the active site of HDAC6. (a) View of the tubular
access channel, and (b) additional interactions generated by the oxidation of the sulfur
atom (green: carbon; blue: nitrogen; red: oxygen; yellow: sulfur; magenta: zinc ion).
a
Reference compound: Trichostatin A (IC50 = 0.012 mM).
c
3776 Chem. Commun., 2013, 49, 3775--3777
This journal is The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
Table 2 Comparison of HDAC selectivity
also show an interesting profile for further evaluation based on
their predicted druglikeness (MW, clogP, solubility).
HDAC1
HDAC4
HDAC6
HDAC11
HDAC8
a
a
a
b
The findings described in this communication thus provide a
platform for more elaborate studies with respect to the HDAC6
inhibitory activity of this new class of thiaheterocyclic compounds
which, in combination with further optimization of drug-relevant
molecular properties, might afford promising new lead structures.
In conclusion, N-(4-hydroxycarbamoylbenzyl)-1,2,4,9-tetrahydro-
3-thia-9-azafluorenes were efficiently prepared and shown to be of
interest as novel and selective HDAC6 inhibitors, culminating in
Compoud IC50 (mM) IC50 (mM) IC50 (mM) IC50 (mM) IC50 (mM)
8a
8c
11
12
1.6
1.9
0.0019
0.0037
NC
NC
1.7
0.93
a
b
Reference compound: Trichostatin A. Reference compound: Scriptaid;
NC = Not Calculable (concentration–response curve shows less than 25%
effect at the highest validated testing concentration). 8a: R1 = R2 = H; 8c:
R1 = F, R2 = H.
but sulfones 8a and 8c are even more potent than sulfides 5a and the identification of two sulfone derivatives as interesting lead
5c with IC50 values of 1.9 and 3.7 nM, respectively. structures for further elaboration displaying potent and selective
Finally, the HDAC inhibition selectivity of the two most HDAC6 inhibition in the nanomolar range.
active compounds 8a (R1 = R2 = H) and 8c (R1 = F, R2 = H)
against the other HDAC isoform classes was assessed and, to
this end, a class I (HDAC1), a class IIa (HDAC4), a class IIb
Notes and references
1 M. Dokmanovic, C. Clarke and P. A. Marks, Mol. Cancer Res., 2007,
(HDAC6) and a class IV (HDAC11) isozyme was selected. Con-
sidering the fact that Tubastatin A has over 1000-fold selectivity
against all HDAC isozymes except for HDAC8, where it has only
a 57-fold selectivity, the HDAC8 inhibitory activity of com-
pounds 8a and 8c was also evaluated.
5, 981; P. A. Marks, Biochim. Biophys. Acta, Gene Regul. Mech., 2010,
1799, 717; H. J. Kim and S. C. Bae, Am. J. Transl. Res., 2011, 3, 166.
2 J. E. Bolden, M. J. Peart and R. W. Johnstone, Nat. Rev. Drug
Discovery, 2006, 5, 769; L. D. Marcotullio, G. Canettieri, P. Infante,
A. Greco and A. Gulino, Biochim. Biophys. Acta, 2011, 1815, 241;
N. A. Shein and E. Shohami, Mol. Med., 2011, 17, 448; A. G. Kazantsev
and M. L. Thompson, Nat. Rev. Drug Discovery, 2008, 7, 854.
The data in Table 2 point to a good to excellent HDAC6
selectivity of hydroxamic acids 8a and 8c, with the HDAC6
versus HDAC11 and HDAC1 selectivity being the most pro-
nounced. The HDAC11 inhibitory effect of 8a,c appeared to
be very low and no IC50 values could be obtained. Furthermore,
a 5789-fold and a 3243-fold selectivity against HDAC1 was
determined for compounds 8a and 8c, respectively, which
substantially exceeds the selectivity of Tubastatin A (1093-fold
selectivity).10 In addition, also a high HDAC6 versus HDAC4
selectivity was observed for sulfones 8a and 8c (842- and 513-
fold, respectively). Finally, it is interesting to note that these
compounds show a good HDAC6 versus HDAC8 selectivity, and
both sulfone 8a (895-fold) and sulfone 8c (251-fold) exhibited a
considerably higher selectivity in that respect as compared to
Tubastatin A (57-fold).10
3 A. J. M. De Ruijter, A. H. Van Gennip, H. N. Caron, S. Kemp and
A. B. P. Van Kuilenburg, Biochem. J., 2003, 370, 737.
4 T. C. Karagiannis and A. El-Osta, Leukemia, 2007, 21, 61.
5 F. Thaler and S. Minucci, Expert Opin. Drug Discovery, 2011, 6, 393.
6 C. d’Ydewalle, E. Bogaert and L. Van Den Bosch, Traffic, 2012,
13, 771; G. Li, H. Jiang, M. Chang, H. Xie and L. Hu, J. Neurol.
Sci., 2011, 304, 1; G. I. Aldana-Masangkay and K. M. Sakamoto,
´
J. Biomed. Biotechnol., 2011, 1; A. Valenzuela-Fernandez, J. R.
´
Cabrero, J. M. Serrador and F. Sanchez-Madrid, Trends Cell Biol.,
2008, 18, 291; C. Boyault, K. Sadoul, M. Pabion and S. Khochbin,
Oncogene, 2007, 26, 5468.
7 J. H. Kalin, H. Zhang, S. Gaudrel-Grosay, G. Vistoli and A. P.
Kozikowski, ChemMedChem, 2012, 7, 425; T. Suzuki, A. Kouketsu,
Y. Itoh, S. Hisakawa, S. Maeda, M. Yoshida, H. Nakagawa and
N. Miyata, J. Med. Chem., 2006, 49, 4809; J. M. Ontoria, S. Altamura,
A. Di Marco, F. Ferrigno, R. Laufer, E. Muraglia, M. C. Palumbi,
M. Rowley, R. Scarpelli, C. Schultz-Fademrecht, S. Serafini,
C. Steinku¨hler and P. Jones, J. Med. Chem., 2009, 52, 6782;
D. V. Smil, S. Manku, Y. A. Chanigny, S. Leit, A. Wahhab, T. P. Yan,
M. Fournel, C. Maroun, Z. Li, A. M. Lemieux, A. Nicolescu, J. Rahil,
The experimental results listed in Tables 1 and 2 are in line
with the structure–activity relationship insights provided
by ligand docking. These data show that decoration of the
N-(4-hydroxycarbamoylbenzyl)-1,2,4,9-tetrahydro-3-thia-9-azafluorene
scaffold at the linker unit (in casu by a methoxy group) is
unfavorable for HDAC6 inhibitory activity. On the other hand,
introduction of a substituent (in casu a fluoro atom) at the cap
group did not appear to have a significant effect on the activity
profile. It should also be noted that replacement of the tertiary
amine functionality (NMe moiety) in the tetrahydropyrido-
[4,3-b]indole core structure of Tubastatin A by a sulfide unit
results in compounds with a comparable HDAC6 inhibitory
activity (at least as concerns the IC50 value), whereas replacement
by a sulfone moiety (SO2) affords even more potent HDAC6
´
S. Lefebvre, A. Panetta, J. M. Besterman and R. Deziel, Bioorg. Med.
¨
Chem. Lett., 2009, 19, 688; S. Schafer, L. Saunders, E. Eliseeva,
A. Velena, M. Jung, A. Schwienhorst, A. Strasser, A. Dickmanns,
R. Ficner, S. Schlimme, W. Sippl, E. Verdin and M. Jung, Bioorg.
Med. Chem., 2008, 16, 2011; C. A. Olsen and M. R. Ghadiri, J. Med.
¨
Chem., 2009, 52, 7836; S. Schafer, L. Saunders, S. Schlimme,
V. Valkov, J. M. Wagner, F. Kratz, W. Sippl, E. Verdin and M. Jung,
ChemMedChem, 2009, 4, 283; P. K. Gupta, R. C. Reid, L. G. Liu,
A. J. Lucke, S. A. Broomfield, M. R. Andrews, M. J. Sweet and
D. P. Fairlie, Bioorg. Med. Chem. Lett., 2010, 20, 7067.
8 M. A. Rivieccio, C. Brochier, D. E. Willis, B. A. Walker,
M. A. D’Annibale, K. McLaughlin, A. Siddiq, A. P. Kozikowski,
S. R. Jaffrey, J. L. Twiss, R. R. Ratan and B. Langley, Proc. Natl. Acad.
Sci. U. S. A., 2009, 106, 19599.
9 J. C. Wong, R. Hong and S. L. Schreiber, J. Am. Chem. Soc., 2003,
125, 5586.
10 K. V. Butler, J. Kalin, C. Brochier, G. Vistoli, B. Langley and
A. P. Kozikowski, J. Am. Chem. Soc., 2010, 132, 10842.
inhibitors. The in silico observed occurrence of hydrogen bonds 11 J. H. Kalin, K. V. Butler, T. Akimova, W. W. Hancock and
A. P. Kozikowski, J. Med. Chem., 2012, 55, 639.
12 A. G. Dossetter, Med. Chem. Commun., 2012, 3, 1518.
13 A. Sudhakara, H. Jayadevappa, H. N. H. Kumar and
between the introduced oxygen atoms and the backbone
nitrogen atom of residues Asp567 and Gly619 can account for
the higher in vitro activity of these sulfone derivatives.
In addition to their promising biological potential and their
straightforward and easy synthesis and purification, sulfones 8a
and 8c (designated as Tubathian A and Tubathian B, respectively)
K. M. Mahadevan, Lett. Org. Chem., 2009, 6, 159.
14 Since an IC50 value is not a constant value and the experiments have
not been performed in parallel, a comparison between the IC50 values
for compounds 5 and 8 and the literature value for Tubastatin A
should be made with care.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 3775--3777 3777