Succinimide hydroxamic acids as potent inhibitors of histone deacetylase (HDAC)

Article abstract of DOI:10.1016/S0960-894X(02)00622-4

A series of succinimide hydroxamic acids was prepared and evaluated in vitro for HDAC inhibition and tumor cell antiproliferation. While the original macrocyclic analogue 6 was quite potent in both assays, several appropriately substituted non-macrocyclic succinimides, such as 23, were equipotent.

Full text of DOI:10.1016/S0960-894X(02)00622-4

Bioorganic & Medicinal Chemistry Letters 12 (2002) 2919–2923
Succinimide Hydroxamic Acids as Potent Inhibitors of Histone
Deacetylase (HDAC)
Michael L. Curtin,* Robert B. Garland, H. Robin Heyman, Robin R. Frey,
Michael R. Michaelides, Junling Li, Lori J. Pease, Keith B. Glaser, PatrickA. Marcotte
and Steven K. Davidsen
Cancer Research Area, Abbott Laboratories, Dept. 47J, Bldg. AP10, 100 Abbott Park Road, Abbott Park, IL 60064, USA
Received 20 May 2002; accepted 10 July 2002
Abstract—A series of succinimide hydroxamic acids was prepared and evaluated in vitro for HDAC inhibition and tumor cell
antiproliferation. While the original macrocyclic analogue 6 was quite potent in both assays, several appropriately substituted non-
macrocyclic succinimides, such as 23, were equipotent.
# 2002 Elsevier Science Ltd. All rights reserved.
Histone deacetylases alter the acetylation status of the
amino terminal region of histone proteins complexed
with DNA in the nucleosome; the extent of histone
acetylation determines the accessibility of DNA within
chromatin and thereby influences gene expression.1 The
inappropriate recruitment of HDACs provides at least one
mechanism by which oncogenes can alter gene expression
in favor of excessive proliferation and makes inhibition
of HDACs a potential target for the development of
small molecule anticancer agents.2
Several families of potent (IC50<100 nM), small-mole-
The human histone deacetylases fall into two distinct
classes, the HDACs and the SIRTs.3 The HDACs are
divided into two major sub-classes based on sequence
homology, the class I HDACs (HDAC1, HDAC2,
HDAC3, HDAC8, and HDAC11) which are related to
the yeast histone deacetylase Rpd3 and the class II
HDACs (HDAC4, HDAC5, HDAC6, HDAC7,
HDAC9, HDAC10) which are related to Hda1.4 All of
the HDACs possess a highly conserved zinc-dependent
catalytic domain while the class II HDACs are roughly
twice as large as those of class I. The SIRTs, or surtuins,
are a family of NAD-dependent deacetylases which dis-
play significant sequence divergence from class I and II
HDACs.
cule HDAC inhibitors have been reported in the recent
literature5 and include: SAHA (1), one of several reversible
inhibitors bearing a hydroxamic acid and an aromatic ter-
minus separated by a hydrophobic spacer (lit. IC50 10
nM);6 trapoxin B (2), an irreversible inhibitor from the
cyclic tetrapeptide family which has an epoxyketone at
the terminus of the hydrophobic chain (lit. IC50 0.1
nM);7 and CHAP1 (3), a hybrid analogue of the cyclic
peptide and hydroxamic acid classes (lit. IC50 2 nM).8
The nature of the terminal group of the hydroxamic
acids is an important determinant of inhibitory activity
as exemplified by hydroxamic acid 4, which despite
being structurally similar to 1 is a much weaker HDAC
inhibitor (IC50 1.4 mM).9 Based on a comparison of amino
acid sequences around the active site of the individual
human HDACs, it has been suggested that the terminus
of these inhibitors should interact with a ‘selectivity-
determining region’ and that proper substitution could
lead to subtype selective inhibitors.10
*Corresponding author. Tel.: +1-847-938-0463; fax: +1-847-935-
5165; e-mail: mike.curtin@abbott.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00622-4

2920
M. L. Curtin et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2919–2923
In an effort to prepare novel and less peptidic analogues
of 3, which could be HDAC subtype selective, we
sought to replace the cyclic tetrapeptide of 3 with the
peptidomimetic core of Abbott’s macrocyclic matrix
metalloproteinase inhibitors.11 The synthetic route chosen
for the initial target macrocycle 5 not only gave the
desired product but also the related succinimide hydrox-
amic acid 6 as a significant by-product. While it was
determined that 5 was only a modest inhibitor of
HDAC (IC50 2.1 mM), compound 6 was suprisingly
potent (IC50 38 nM).9a The efficient synthesis and
characterization of 6 as well as the SAR of simplified
analogues is reported here.
Scheme 2. Reagents and conditions: (a) EDC, HOBt, NMM, DMF;
(b) TFA, CH2Cl2, 93% two steps; (c) iPrO2CCl, Et3N, DMF, 0 ꢀC;
MeOH, 93%; (d) KOtBu, THF, 76%; (e) H2, Pd/C, THF, 94%.
The preparation of 5 and 6 (Scheme 1) began with the
coupling of methyl 6-aminocaproate and carboxylic acid
7, which was prepared using published procedures.11
When methyl ester 8 was saponified with LiOH or
NaOH at room temperature, a 1:1 mixture of insepar-
able carboxylic acids were isolated and coupled with O-
benzyl hydroxylamine. The O-benzyl hydroxamates were
partially separated by column chromatography and
deprotected separately to give hydroxamic acids 5 and 6.
The structural assignment of 6 was made based on 1H and
13C NMR and later confirmed by X-ray crystallography
of a related intermediate.12,13 The macrocyclic succin-
imide presumably arises from the intramolecular attack
of the tert-butyl ester by the internal amide nitrogen.14
KOtBu at room temperature for 2 h gave the desired
succinimide 12 in 70–80% isolated yield. O-Benzyl
deprotection then gave hydroxamic acid 6. Treatment of
carboxylic acid 11 under coupling conditions (EDC,
HOBt, NMM or DBU) did not give succinimide for-
mation.
Simplified succinimide hydroxamic acids that lacked the
macrocyclic ring were prepared as shown in Scheme 3.
Amine 13, which was prepared from amine 10 and N-
(tert-butoxycarbonyl)-l-phenylalanine in two steps, and
succinate 1415 were coupled and the adduct was depro-
tected to give carboxylic acid 15. Early experimentation
and literature precedent16 indicated that the acyclic
succinamides should cyclize much more readily than the
macrocyclic amides; however, while treatment of 15
with EDC, HOBt and NMM in DMF did give succinimide
cyclization, the yields were low (30–40%) and the reactions
were capricious. Substituting the stronger base DBU for
NMM and heating (50 ꢀC for 2–3 h) dependably gave
cyclization in good yield (70–75%) regardless of succinate
An improved synthesis of 6 is shown in Scheme 2.
Amine 10, which had been prepared in two steps from
commercially available N-(tert-butoxycarbonyl)-6-amino-
caproic acid, was coupled with macrocycle 7 to give
carbocyclic acid 11 after tert-butyl ester removal. This
acid was converted to the corresponding methyl ester
and, after some experimentation, it was found that
treatment of this intermediate with one equivalent of
Scheme 1. Reagents and conditions: (a) EDC, HOBt, NMM, methyl
6-aminocaproate hydrochloride, DMF, 95%; (b) LiOH, 1:1 THF/
MeOH, 78%; (c) EDC, HOBt, NMM, H2NOBn, DMF; flash
chromatography, 51% total yield; (d) H2, Pd/C, THF, 86% for 5, 91%
for 6.
Scheme 3. Reagents and conditions: (a) EDC, HOBt, NMM, DMF;
(b) TFA, CH2Cl2, 75% two steps; (c) EDC, HOBt, DBU, 50 ꢀC,
DMF, 70%; (d) H2, Pd/C, THF, 96%.

M. L. Curtin et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2919–2923
2921
substitution. Debenzylation gave the desired product,
hydroxamic acid 16.
activity. Replacement of the succinimide with an
unsubstituted five-membered ring lactam (22) gave a
compound with no measureable HDAC activity below 5
mM. Conversely, a second, vicinal substituent on the
succinimide (methyl, n-propyl, iso-butyl) generally led
to an increase in inhibitory activity when compared to
16 (e.g., 23). Additional succinimide substitution did not
give further improvement (24). Replacement of the suc-
cinimide with phthalimide or dimethylhydantoin led to
a 5- to 10-fold loss in activity versus 16.
In order to easily prepare tether and hydroxamate
modified analogues of the monocyclic succinimides, a
synthetic route to succinimide carboxylic acids such as
20 was developed and is shown in Scheme 4. Succinate
1715,17 and l-phenylalanine benzyl ester, prepared from
N-(tert-butoxycarbonyl)-l-phenylalanine in two steps,
were coupled and the product was deprotected to give
carboxylic acid 18. Cyclization as before (EDC, HOBt,
DBU) followed by O-debenzylation gave the desired
intermediate 20, which could be coupled with aliphatic
amines in high yield. The route to the analogous macro-
cyclic carboxylic acid 9 was not successful.
Regarding the amino acid residue, it was found that the
nature of the side chain was very important to enzy-
matic activity. Removal of the benzyl moiety resulted in
a very weakinhibitor ( 25) while its replacement with an
iso-butyl also gave weakenzymatic activity (IC 6 mM,
50
The compounds from this workwere tested in an in
vitro assay against a partially purified nuclear HDAC
preparation from K562 erythroleukemia cells;18 immuno-
characterization indicated that deacetylase activity from
this preparation was predominantly from the class I
HDACs, HDAC1 and HDAC2. Promising compounds
were then assessed for their effect on cell proliferation in
an AlamarBlue assay after 72 h exposure using human
HT1080 fibrosarcoma cells and human MDA435 breast
carcinoma cells.19 In order to verify that the observed
antiproliferative effects were HDAC mediated, selected
inhibitors were evaluated for their ability to cause
hyperacetylation of nuclear histones and induce p21waf1/cip1
expression.20
compound not shown); replacement with an l-cyclo-
hexylalanine residue regained some of the potency (26).
The use of p-methoxy- or p-ethoxy-l-phenylalanine gave
significant potency, both enzymatically and cellularly, as
exemplified by monocyclic succinimide 27 which has
comparable activity to macrocycle 6. Unlike 6, for which
the analogue epimeric at the amino acid center is equipo-
tent, the analogue of 16 prepared from d-phenylalanine
was significantly less potent (IC50 770 nM).
Examination of tether length between the hydroxamic
acid and succinimide of 23 indicated that both the one-
methylene homologated or shortened analogues were
less potent enzymatically (IC50’s of 232 and 1600 nM,
respectively). Replacement of the hydroxamic acid moi-
ety in 16 or 23 with a methyl ester, O-benzyl hydrox-
amate or carboxylic acid gave inactive compounds.
In vitro characterization indicated that macrocyclic
succinimide 6 compared very well to 1, having greater
deacetylase inhibitory activity than the reference inhi-
bitor in the nuclear HDAC assay and superior cellular
activity against the two cell lines tested (Table 1). In
order to determine the pharmacophore of this class of
compounds, an analogue lacking the macrocycle was
prepared; the resulting monocyclic analogue 16 was
quite potent against the nuclear extracted HDACs (99
nM) and still had significant cellular activity (670 nM,
MDA435). The notable potency and the significantly
easier preparation of these analogues led us to conduct
most of the SAR studies on the monocyclic series. Fur-
ther simplification of 16 by removal of the succinimide
iso-butyl moiety (21) gave an even larger decrease in
activity with a corresponding loss in antiproliferative
Cellular assays with T24 bladder carcinoma cells
showed that 6 and 23 increased histone acetylation and
induced expression of p21waf1/cip1, similar to the cellular
effects of 1 and 3.21
The X-ray crystal structure of a bacterial HDAC
homologue (HDLP)22 with bound inhibitor 1 suggested
that this type of compound inhibits the human HDACs
by binding in a tubular hydrophobic pocket and che-
lating the active site zinc. Presumably, other hydrox-
amic acids with lipophilic tethers such as 3 and
succinimides 6 or 23 bind the HDACs in the same
fashion. The data presented in this workis consistent
with this notion as chain length is important and the
presence of a zinc chelator is crucial for activity. How-
ever, because the HDLP has a large deletion at the
entrance to the active site pocket, little is known about
the specific interactions between the HDACs and the
terminus of the hydroxamate inhibitors. While it is clear
that the presence and substitution of the succinate moi-
ety of these inhibitors is very important to HDAC
activity, molecular modeling has not revealed a con-
sistent mode of binding to the human HDACs.
In conclusion, a novel and potent series of hydroxamic
acids has been found that has significant cellular activity
against several tumor cell lines. Consistent with known
HDAC inhibitors, these compounds demonstrate that
subtle adjustments of the succinimide terminus can lead
Scheme 4. Reagents and conditions: (a) EDC, HOBt, NMM, l-phenyl-
alanine benzyl ester, DMF; (b) TFA, CH2Cl2, 88% two steps; (c)
EDC, HOBt, DBU, 50 ꢀC, DMF, 71%; (d) H2, Pd/C, MeOH, 92%.

2922
M. L. Curtin et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2919–2923
Table 1. HDAC inhibition and antiproliferation data for SAHA and selected succinimide hydroxamic acids
Compd
Structure
HDAC IC50
(nM)9a,18
HT1080 proliferation IC50
(mM)19
MDA435 proliferation IC50
(mM)19
1
6
120
38
2.4
0.25
1.9
0.15
16
21
99
2.3
0.67
10.0
660
14.0
22
23
24
25
26
>5000
51
ND
0.40
1.5
ND
0.57
0.7
66
5000
640
ND
ND
ND
ND
27
38
0.53
0.18
5. (a) Jung, M. Curr. Med. Chem. 2001, 8, 1505. (b) Meinke,
P. T.; Liberator, P. Curr. Med. Chem. 2001, 8, 211.
6. (a) The given IC50 was determined using immunoprecipi-
tated HDAC1. Richon, V. M.; Emiliani, S.; Verdin, E.; Webb,
Y.; Breslow, R.; Rifkind, R. A.; Marks, P. A. Proc. Natl.
Acad. Sci. U.S.A. 1998, 95, 3003.
to large changes in HDAC inhibitory activity. The iso-
lation of the different isoforms of HDAC is in progress
and the selectivity profile of these inhibitors will be
reported in due course.
7. The given IC50 was determined in ref 8 against recombinant
HDAC1. Kijima, M.; Yoshida, M.; Sugita, K.; Horinouchi,
S.; Beppu, T. J. Biol. Chem. 1993, 268, 22429.
8. The given IC50 was determined using recombinant
HDAC1. Furumai, R.; Komatsu, Y.; Nishino, N.; Khochbin,
S.; Yoshida, M.; Horinouchi, S. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 87.
Acknowledgements
The authors wish to thankRodger Henry for the X-ray
crystallographic workand Kent Stewart for molecular
modeling studies.
9. (a) The given IC50 was determined using a mixture of
HDAC1 and HDAC2 prepared from the nuclear extraction of
K562 erythroleukemia cells.18 (b) Compound 4 is similar to
inhibitors reported in: Jung, M.; Hoffmann, K.; Brosch, G.;
Loidl, P. Bioorg. Med. Chem. Lett. 1997, 7, 1655.
10. (a) Sternson, S. M.; Wong, J. C.; Grozinger, C. M.;
Schreiber, S. L. Org. Lett. 2001, 3, 4239. (b) Jung, M.; Gerald,
B.; Kolle, D.; Scherf, H.; Gerhauser, C.; Loidl, P. J. Med.
Chem. 1999, 42, 4669.
11. Steinman, D. H.; Curtin, M. L.; Garland, R. B.; David-
sen, S. K.; Heyman, R. H.; Holms, J. H.; Albert, D. H.;
Magoc, T. J.; Nagy, I. B.; Marcotte, P. A.; Li, J.; Morgan,
D. W.; Hutchins, C.; Summers, J. B. Bioorg. Med. Chem. Lett.
1998, 8, 2087.
References and Notes
1. (a) Cheung, W. L.; Briggs, S. D.; Allis, C. D. Curr. Opin.
Cell Biol. 2000, 12, 326. (b) Cress, W. D.; Seto, E. J. Cell.
Physiol. 2000, 184, 1.
2. Coombes, R. C.; Vigushin, D. M. Anti-Cancer Drugs 2002,
13, 1.
3. (a) Gray, S. G.; Ekstrom, T. J. Exp. Cell. Res. 2001, 262,
75. (b) For a discussion of the SIRT active site, see: Dutnall,
R. N.; Pillus, L. Cell 2001, 105, 161.
4. For the most recent published isolation (HDAC 10), see:
Kao, H. Y.; Lee, C. H.; Komarov, A.; Han, C. C.; Evans,
R. M. J. Biol. Chem. 2002, 277, 187.

M. L. Curtin et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2919–2923
2923
12. All compounds were characterized by 1H NMR, mass
spectroscopy and elemental analysis.
18. Yoshida, M.; Kijima, M.; Akita, M.; Beppu, T. J. Biol.
Chem. 1990, 265, 17174.
13. An X-ray crystal structure was obtained of carboxylic acid
9, which was isolated in low yield from a macrocycle synthetic
route similar to that shown in Scheme 4. Carboxylic acid 9 was
coupled with amine 10 and the product was debenzylated to
give authentic hydroxamic acid 6.
19. Procedure for cell growth inhibition assays: HT1080 or
MDA-435 cells were seeded at 5Â103 cells/well in a 96-well
tissue culture 18 h prior to addition of compound or vehicle
(DMSO) in complete growth medium (10% serum). After 72 h
exposure to compound the cells were washed with PBS. Ala-
marBlue (modified MTT) was added in serum free medium or
PBS (10% AlamarBlue in 100 mL) and allowed to incubate for
4 h at 37 ꢀC. Cell number was measured by determining the
fluorescence at 488/520 nm in an fmax fluorescence plate reader
(Molecular Devices, San Diego, CA, USA). Treated cells were
compared to vehicle treated cells to determine growth inhibition.
20. (a) Hyperacetylation of histone H4 was induced in T24
bladder carcinoma cells after 4 h exposure to HDAC inhibi-
tors. Hyperacetylated histone H4 was analyzed by Western
blot using an anti-acetylated H4 antibody (UBI, Lake Placid,
NY) and 5 mg/lane of acid-extracted nuclear histones; (b) The
expression of p21waf1/cip1 was induced in T24 cells by a 24 h
exposure to HDAC inhibitors. The induction of p21waf1/cip1
was determined by Western blot analysis using a p21waf1/cip1
monoclonal antibody (UBI) and 100 mg/lane of cellular lysate.
21. Sambucetti, L. C.; Fischer, D. D.; Zabludoff, S.; Kwon,
P. O.; Chamberlin, H.; Trogani, N.; Xu, H.; Cohen, D. J. Biol.
Chem. 1999, 274, 34940.
14. Cyclizations of succinamide esters under basic conditions
have been reported but only for acyclic systems with unhindered
esters (i.e., methyl, benzyl). For an example, see: Fujii, N.;
Nomizu, M.; Futaki, S.; Otaka, A.; Funakoshi, S.; Akaji, K.;
Watanabe, K.; Yajima, H. Chem. Pharm. Bull. 1986, 34, 864.
15. Beckett, R. P.; Crimmin, M. J.; Davis, M. H.; Spavold, Z.
Synlett 1993, 137.
16. Bodansky, M.; Martinez, J. J. Org. Chem. 1978, 43, 3071.
17. Curtin, M.L; Garland, R. B.; Davidsen, S. K.; Marcotte,
P. A.; Albert, D. H.; Magoc, T. J.; Hutchins, C. Bioorg. Med.
Chem. Lett. 1998, 8, 1443.
22. Finnin, M. S.; Donigian, J. R.; Cohen, A.; Richon, V. M.;
Rifkind, R. A.; Marks, P. A.; Breslow, R.; Pavletich, N. P.
Nature 1999, 401, 188.