Design, synthesis, and activity of HDAC inhibitors with a N-formyl hydroxylamine head group

Article abstract of DOI:10.1016/j.bmcl.2003.10.055

Histone deacetylases (HDAC) are promising targets for cancer chemotherapy. HDAC inhibitors are thought to act in part by disrupting normal cell cycle regulation, resulting in apoptosis and/or differentiation of transformed cells. Several HDAC inhibitors,

Full text of DOI:10.1016/j.bmcl.2003.10.055

Bioorganic & Medicinal Chemistry Letters 14 (2004) 449–453
Design, synthesis, and activity of HDAC inhibitors with a
N-formyl hydroxylamine head group§
Tom Y. H. Wu,a Christian Hassig,b,c Yiqin Wu,d Sheng Dinga and Peter G. Schultza,*
aDepartment of Chemistry, The Scripps Research Institute, La Jolla, CA, USA
bDepartment of Molecular Biology, The Scripps Research Institute, La Jolla, CA, USA
cKalypsys, San Diego, CA, USA
dSyrrx, San Diego, CA, USA
Received 7 October 2003; accepted 23 October 2003
Abstract—Histone deacetylases (HDAC) are promising targets for cancer chemotherapy. HDAC inhibitors are thought to act in
part by disrupting normal cell cycle regulation, resulting in apoptosis and/or differentiation of transformed cells. Several HDAC
inhibitors, which contain hydrophobic tails and the Zn2+ chelator hydroxyamic acid as a head group, are potent inhibitors of
HDACs both in vitro and in vivo. In this study, a related class of compounds with a N-formyl hydroxylamino head group has been
synthesized and their ability to inhibit HDACs have been assayed in biochemical and cellular assays. These compounds were found
to have comparable activities to suberoylanilide hydroxyamic acid (SAHA) in HDAC enzymatic assays and histone hyper-
acetylation cellular assays.
# 2003 Elsevier Ltd. All rights reserved.
Inhibitors of histone deacetylases (HDAC) represent a
new class of potential small molecule cancer ther-
apeutics. Among the hybrid polar class of HDAC inhi-
bitors, suberoylanilide hydroxyamic acid (SAHA) has
demonstrated both in vitro and in vivo efficacy against
several types of cancer and is in clinical trials.1À3 The
molecular basis of HDAC inhibition by SAHA has been
revealed by the X-ray co-crystal structure of SAHA and
a HDAC related polyaminohydrolase (HDLP).4 The
aliphatic chain of SAHA inserts into the narrow hydro-
phobic cleft of the enzyme active site, and the polar
head group bearing the hydroxyamic acid serves as a
bidentate chelator of the catalytic Zn2+ ion bound in
the active site. Other Zn2+ chelators have been used to
inhibit Zn2+ dependent enzymes, including the isomeric
N-formyl hydroxylamino moiety which has been used in
inhibitors of angiotensin converting enzyme (ACE) and
neural endopeptidase (NEP).5 We therefore hypothe-
sized that replacement of the hydroxyamic acid group of
SAHA and other HDAC inhibitors with a N-formyl
hydroxylamine should not adversely affect their ability
to inhibit HDAC, but may lead to a novel HDAC inhi-
bitor with improved biological properties (Fig. 1).
Based on the co-crystal structure of SAHA and HDLP,
several compounds were designed which mimic the
hydrophobic tail group and aliphatic chain of SAHA,
but have a N-formyl hydroxylamino group replacing the
hydroxyamic acid. These compounds should pack into the
hydrophobic cavity in a similar fashion to SAHA, allow-
ing the N-formyl hydroxylamine group to form a biden-
tate chelate to Zn2+. The synthesis of these compounds
is illustrated in Scheme 1.6 Hydroxy-carboxylic acids 1
with varying chain lengths (n=5–7) were first converted
to the corresponding mesylate 2 in approximately 75%
yield. Coupling of 2 with a number of anilines was carried
out using PyBOP/DIEA to yield 3 in 53–65% yield.
Nucleophilic displacement of the mesylate by O-benzyl-
hydroxylamine was carried out by heating the reaction at
80 ꢀC in DMF in 41–50% yield. The resulting secondary
amine 4 was acylated or formylated to give 5 (R2=Me,
OMe, or H). Final deprotection of the benzyl group
under hydrogenation conditions completed the synthesis
of 6. Using this synthetic route, analogues TWZ101 to
TWZ108 were made in overall yield of 10–20%.
To synthesize analogues with varying hydrophobic
groups, another method was devised in which the
hydrophobic moiety was introduced at the last step
of the synthesis (Scheme 2).7 The primary bromide of
N-Boc-1-amino-5-bromopentane 7 was displaced by
O-benzylhydroxylamine to afford 8 in 77% yield. Sub-
sequent formylation of the secondary amine by in situ
§Supplementary data associated with this article can be found at
10.1016/j.bmcl.2003.10.055
* Corresponding author. Tel.: +1-858-784-9300; fax: +1-858-784-
9339; e-mail: schultz@scripps.edu
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.10.055

450
T. Y. H. Wu et al. / Bioorg. Med. Chem. Lett. 14 (2004) 449–453
Figure 1. Structures of SAHA and TWZ analogues.
Scheme 1. (i) MsCl, Et3N; (ii) NaOH, (63–83% over two steps); (iii) R1NH2, PyBOP, DIEA, THF, (53–65%); (iv) BnONH2, DMF, 80 ꢀC, (41–
50%); (v) for R2=Me, OMe: R2COCl, DIEA, DCM, (>95%); for R2=H: HCOOH, CDI, DCM, 0 ꢀC, (>95%); (vi) H2, Pd/C, MeOH/THF (80–
95%).
Scheme 2. (i) BnONH2, K2CO3, DMF, 80 ꢀC, (77%); (ii) HCOOH, CDI, DCM, 0 ꢀC, (>95%); (iii) 10% TFA/DCM, (95%); (iv) R1COCl, DIEA,
DCM; (v) H2, Pd/C, (70% over twp steps).
activation of formic acid using carbonyl diimidazole
(CDI) yielded compound 9 in nearly quantitative yield.
The Boc group was then deprotected by treatment with
10% TFA/DCM to expose the primary amine (10).
Condensation of 10 with various acid chlorides followed
by benzyl deprotection resulted in the synthesis of 11
(TWZ109 to TWZ111) in 49% overall yield. SAHA was
also synthesized as a reference compound for biological
assays.8
modeling experiments, which suggest that the active site
is too small to accommodate a methyl or methoxy
group. However, it is not clear why the N-formyl
hydroxylamino head group caused a dramatic loss in
activity. Further investigations using X-ray crystal-
lography are in progress.
In contrast to the head groups, structural modifications
of the hydrophobic tail had less effect on binding affi-
nity. For example, the length of the aliphatic chain
(n=4–6) does not seem to have a significant effect on
inhibition activity. However, the structure of the
hydrophobic group has a moderate affect on activity.
Increasing steric size from phenyl (TWZ104) to naph-
thyl (TWZ107 and 108) while keeping the alkyl chain
constant improved the inhibitory activity 2- to 10-fold.
Increasing the flexibility of the hydrophobic group led
Enzymatic activities of HDAC1, HDAC2, HDAC6,
HDAC8, as well as total HDAC from nuclear extracts,
were tested in fluorescence-based HDAC enzymatic
assays (Table 1).9,10 SAHA was used as a positive con-
trol in all of the assays. Changes in the Zn2+ chelating
group have significant effects on HDAC inhibition.
Derivatives with the formyl head group (R2=H) inhib-
ited all HDACs with IC50’s in the low micromolar
range: a 15- and 150-fold decrease in activity against
HDAC2 and HDAC8 relative to the corresponding
hydroxyamic acid (SAHA) (Table 1). Compounds with
acetyl (R2=Me) or methyl carbamate (R2=OMe)
groups are inactive against all of the HDACs tested
(IC50>100 mM). This latter result is consistent with
to
a
significant decrease in HDAC inhibition
(TWZ110). Finally, reversal of the amide bond on the
hydrophobic end (TWZ109 and 111) slightly lowered
the activity.
Several SAHA analogues with increased potency have
been reported which retain the hydroxyamic acid group,

T. Y. H. Wu et al. / Bioorg. Med. Chem. Lett. 14 (2004) 449–453
Table 1. Inhibitory activity of TWZ analogues in enzymatic assays
451
Structure
HDAC1 (mM)
HDAC2
(mM)
HDAC6
(mM)
HDAC8
(mM)
HDAC (extract)
(mM)
SAHA
0.048
n.d.a
n.d.
n.d.
n.d.
n.d.
n.d.
0.073
>100
19
0.10
>100
2.8
0.29
>100
6.8
0.10
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
TWZ 101
TWZ 102
TWZ 103
TWZ 104
TWZ 105
TWZ 106
>100
7.8
>100
5.5
>100
2.8
11
n.d.
4.0
>100
>100
>100
TWZ 107
TWZ 108
0.81
5.7
0.89
5.9
0.91
3.1
0.78
0.65
4.0
50
TWZ 109
TWZ 110
TWZ 111
30
49
1.3
26
41
1.4
4.3
22
>100
>100
4.4
30
9.7
7.0
0.60
TWZ 112
0.16
0.20
0.78
0.69
0.97
TWZ 113
0.18
0.27
0.62
0.80
0.75
a n.d., not determined.

452
T. Y. H. Wu et al. / Bioorg. Med. Chem. Lett. 14 (2004) 449–453
Table 2. TWZ inhibitory activities (EC50) in cellular assays
Histone
Hyperacetylation
Proliferation
U-937
Jurkat
Hepatocyte
BMC
SAHA
TWZ112
TWZ113
0.7 mM
5.1 mM
10 mM
33 mM
40 mM
100 mM
2.3 mM
11 mM
31 mM
>500 mM
41 mM
>500 mM
150 mM
78 mM
130 mM
but which have a larger hydrophobic tail.11 Thus, an
inccrease in the size of the hydrophobic tail might com-
pensate for the loss in activity of the TWZ analogues.
The bis-aminoquinoline hydrophobic group was selec-
ted to test this notion. TWZ112 and TWZ113 were
synthesized as described in Scheme 1. The IC50’s of
TWZ112 and TWZ113 were between 150–800 nM
against different HDACs (Table 1). These IC50 values
are approximately 10- to 50-fold lower than previous
hydroxylamine analogues, but 2- to 5-fold higher than
SAHA in the enzymatic assays. These two compounds,
as well as SAHA, were tested in a histone hyper-
acetylation assay in cells (Table 2).12 Consistent with the
in vitro inhibition activity, the EC50’s were 0.7, 5, and
10 mM for SAHA, TWZ112, and TWZ113, respectively.
Cell proliferation assays were also performed on two
human leukemia cell lines (U-937 and Jurkat)13 and
two types of human primary cells (hepatocytes and bone
marrow cells)14 (Table 2). TWZ112 and TWZ113 did
exhibit cytotoxicity effects in the cancer cells tested, but
with a two- to ten-told lower activity than SAHA (Table
2). This result is consistent with the enzymatic and his-
tone hyperacetylation assay results. The toxicity of
TWZ113 against hepatocytes and bone marrow cells is
comparable to that of SAHA, while TWZ112 has
increased cytotoxicity against hepatocytes (EC50=41
mM).
4. 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.
5. Robl, J. A.; Simpkins, L. M.; Asaas, M. M. Bioorg. Med.
Chem. Lett. 2000, 10, 257.
6. A representative synthetic protocol for Scheme 1. To a
stirred solution of 1 (n=6) (250 mg, 1.56 mmol) in THF
(7 mL) at 0 ꢀC was added dropwise MsCl (0.3 mL, 3.91
mmol) followed by Et3N (0.65 mL, 4.68 mmol). The
reaction was warmed to rt over 5 h. A solution of NaOH
(312 mg, 7.8 mmol) in water (7 mL) was added and stirred
for an additional 15min. Ether was added to the reaction
mixture, and the organic layer was extracted and dis-
carded. The aqueous layer was acidified with 3N HCl (to
pH<1) and extracted three times with ethyl acetate. The
combined organic layers were dried with anhydrous
MgSO4 and concentrated under reduced pressure. The
crude product 2 (320 mg, ca. 86%) was used in the next
step without further purification. To a solution of 2 (n=6)
(160 mg, ca. 0.73 mmol) in DCM (7 mL) was added 1-
aminonaphthalene (104 mg, 0.73 mmol), PyBOP (380 mg,
0.73 mmol), and DIEA (0.25 mL, 1.46 mmol). The reac-
tion was stirred at rt for 8 h and poured into water. The
aqueous layer was removed, and the organic layer was
washed with saturated aqueous NH4Cl, saturated aqu-
eous NaHCO3, brine, dried with MgSO4, and con-
centrated under reduced pressure. The resulting crude
product was purified by flash column chromatography on
silica gel using 25% ethyl acetate in hexane to give 172 mg
(65%) of compound 3 (R1=1-naphthyl, n=6). A stirred
solution of 3 (R1=1-naphthyl, n=6) (97 mg, 0.267 mmol)
and O-benzylhydroxyamine (165 mg, 1.34 mmol) in DMF
(5 mL) was heated to 80 ꢀC overnight. The reaction was
cooled to ambient temperature, poured into water, and
the aqueous layer was extracted three times with ethyl
acetate. The combined organic layers were washed with
brine, dried with anhydrous MgSO4, and concentrated
under reduced pressure. The resulting crude was purified
by flash column chromatography on silica gel using 20%
ethyl acetate in hexane to give 52 mg (50%) of 4 (R1=1-
naphthyl, n=6). For the synthesis of 5 where R2=methyl
or methoxy, a solution of 4 (R1=Ph, n=5) (50 mg, 0.144
mmol) in THF (1.4 mL) was added methyl chloroformate
(17 mL, 0.22 mmol) and DIEA (38 mL, 0.22 mmol). The
reaction was stirred at rt for 2 h and the solvent was
removed under reduced pressure. The resulting crude
product 5 (R1=Ph, n=5, R2=OMe) (ca. >95% yield by
LCMS) was used in the next step without further pur-
ification. For the synthesis of 5 where R2=H, a solution
of carbonyl diimidazole (CDI) (117 mg, 0.72 mmol) in
THF (5 mL) at 0 ꢀC was added 95% formic acid (35 mL,
0.72 mmol) and stirred at 0 ꢀC for 30min. Then, a solution
of 4 (R1=Ph, n=5) (50 mg, 0.144 mmol) in THF (2 mL)
was added, and the reaction was stirred at 4 ꢀC overnight.
The solvent was removed under reduced pressure, and the
resulting crude product 5 (R1=Ph, n=5, R2=H) (ca.
>95% yield by LCMS) was used in the next step without
In conclusion, replacing the hydroxyamic acid group on
SAHA with an N-formylhydroxylamino head group led
to inhibitors of HDAC, but with a fifty-fold drop in
potency. This loss in activity could be offset by increas-
ing the size of the hydrophobic region to afford HDAC
inhibitors with low micromolar activity in cellular his-
tone hyperacetylation assays. The anti-proliferation
effects of TWZ112 and TWZ113 in leukemia cells and
additional toxicity tests in hepatocytes and bone mar-
row cells indicated these compounds have a comparable
therapeutic window to SAHA. Further in vitro and in
vivo assays will be required to determine whether these
compounds offer enhanced pharmacological properties
relative to SAHA.
References and notes
1. Richon, V. M.; Webb, Y.; Merger, R.; Sheppard, T.; Jur-
sic, B.; Ngo, L.; Civoli, F.; Breslow, R.; Rifkind, R. A.;
Marks, P. A. PNAS 1996, 11, 5705.
2. Richon, V. M.; Emiliani, S.; Verdin, E.; Webb, Y.; Bre-
slow, R.; Rifkind, R. A.; Marks, P. A. PNAS 1998, 17, 3003.
3. Richon, V. M.; Sandhoff, T. W.; Rifkind, R. A.; Marks,
P. A. PNAS 2000, 947, 10014.
further purification. Compound 5 was dissolved in
MeOH/THF (1:1) and a catalytic amount of 10% Pd/C

T. Y. H. Wu et al. / Bioorg. Med. Chem. Lett. 14 (2004) 449–453
453
was added. The reaction vessel was quickly evacuated and
refilled with hydrogen gas three times. The reaction was
allowed to stir under hydrogen atmosphere for 5–8 h. The
Pd catalyst was filtered through a plug of Celite and
washed with MeOH. The filtrate was concentrated and
the resulting crude product (ca. 80–95% yield) was pur-
ified by HPLC using a C18 reverse phase column with a
15 min gradient of 5–95% acetonitrile/water. The frac-
tions containing the desired product were dried down and
purified by HPLC using a C18 reverse phase column with
a 15 min gradient of 5–95% acetonitrile/water. The frac-
tions containing the desired product were dried down and
1
prepared as 10 mM DMSO stock solutions. TWZ109. H
NMR (500 MHz, CDCl3) d 8.25 (s, 1H), 7.74–7.76 (m,
2H), 7.72 (br, 1H), 7.40–7.50 (m, 3H), 3.42–3.50 (m, 4H),
1.55–1.78 (m, 4H), 1.27–1.39 (m, 2H). Observed [M+H]+
251.13; calcd for 250.13.
8. Stowell, J. C.; Huot, R. I.; Voast, L. V. J. Med. Chem.
1995, 38, 1411.
1
prepared as 10 mM DMSO stock solutions. TWZ102. H
NMR (500 MHz, CDCl3) d 7.83 (s, 1H), 7.49 (d, J=7.7
Hz, 2H), 7.29 (t, J=7.7 Hz, 2H), 7.08 (t, J=7.3 Hz, 1H),
3.49 (t, J=6.6 Hz, 2H), 2.36 (t, J=7.0 Hz, 2H), 1.69–1.77
(m, 4H), 1.34–1.42 (m, 2H). Observed [M+H]+ 251.13;
calcd for 250.13.
9. Protocols for enzymatic assays using purified HDAC1,
HDAC2, HDAC6, and HDAC8 were developed and car-
ried out by Syrrx. Detailed methods will be published
elsewhere.
10. For the preparation of partially purified HeLa nuclear
extract as the total HDAC enzyme source, refer to:
Dignam, J. D.; Lebovitz, R. M.; Roeder, R. G. Nucleic
Acids Res. 1983, 11, 1475. In vitro HDAC-inhibition
assays were performed as follows: HeLa nuclear extract
was incubated with Fluor de Lys acetyllysine substrate
(Biomol) for 30 min at room temperature in the presence
of test compound or DMSO. The reaction was stopped
with excess Trichostatin A and developer solution to
generate a fluorescent product. The fluorophore is detec-
ted on a fluorometric plate reader (Ex 360 nm, Em 460
nm).
7. A representative synthetic protocol for Scheme 2. N-Boc-
1-amino-5-bromopentane 7 (1 g, 3.77 mmol) was stirred
with O-benzylhydroxyamine (1.4 g, 11.3 mmol) and
K2CO3 (1.6 g, 11.3 mmol) in DMF (15 mL). The reaction
was heated to 90C for 3 h and poured into water. The
aqueous layer was extracted three times with ethyl acetate
and the combined organic layers were washed with brine,
dried with anhydrous MgSO4, and concentrated under
reduced pressure. The resulting crude was purified by
flash column chromatography on silica gel using 1%
MeOH in DCM to give 892 mg of 8 (77%). To a solution
of carbonyl diimidazole (CDI) (527 mg, 3.25 mmol) in
THF (30 mL) at 0 ꢀC was added 95% formic acid (158 mL,
3.25 mmol) and stirred at 0 ꢀC for 30 min. Then, a solu-
tion of 8 (200 mg, 0.65 mmol) in THF (10 mL) was added,
and the reaction was stirred at 4 ꢀC overnight. The solvent
was removed under reduced pressure, and the resulting
crude product 9 was treated with 10% TFA/DCM for 1.5
h at rt. The reaction was poured into saturated aqueous
K2CO3 and the aqueous layer was extracted three times
with DCM. The combined organic layers were washed
with brine, dried with anhydrous MgSO4, and con-
centrated under reduced pressure. The resulting product
10 (>95% yield by LCMS) was used in the next step
without further purification. To a solution of 10 (50 mg,
0.22 mmol) in DCM (1 mL) was added benzoyl chloride
(27 mL, 0.233 mmol) and DIEA (58 mL, 0.33 mmol). The
reaction was stirred at rt for 8 h, poured into water, and
the aqueous layer was extracted twice with DCM. The
combined organic layers were washed with brine, dried
with anhydrous MgSO4, and concentrated en vacuo. The
resulting crude product was then dissolved in MeOH/
THF (1:1; 1mL) and a catalytic amount of 10% Pd/C was
added. The reaction vessel was quickly evacuated and
refilled with hydrogen gas three times. The reaction was
allowed to stir under hydrogen atmosphere for 5 h. The
Pd catalyst was filtered through a plug of Celite and
washed with MeOH. The filtrate was concentrated and the
resulting crude product (ca. 70% yield over two steps) was
11. Breslow, R.; Belvedere, S.; Gershell, L.; Miller, T. A.;
Marks, P. A.; Richon, V. M.; Rifkind, R. A. U.S. Patent,
US 6511990 B1, 2003.
12. For detailed procedures used in the histone hyperacetyla-
tion assays in cells, refer to: Stockwell, B. R.; Haggarty,
S. J.; Schreiber, S. L. Chem. Biol. 1999, 6, 71. Adherent
cells (approx. 5000/well) in a 384-well Greiner polystyrene
assay plate were exposed to compound for 0 to 24 h. Cells
are washed once with PBS (80 mL) and then fixed (95%
ethanol, 5% acetic acid) for 1 min at rt (40 mL). Cells are
blocked with 4% BSA for 1 h and washed and stained
with anti-acetylated histone antibody followed by wash-
ing and incubation with an appropriate secondary anti-
body conjugated to HRP or fluorophore.
13. U-937 and Jurkat cell lines were obtained from ATCC.
They were cultured with RPMI1640, 1 mM Na Pyruvate,
MEM essential vitamins, l-glutamine, 10% Fetal Bovine
Serum, and 1% penicillin/streptomycin. Cells were plated
into a white, clear bottom 96-well plate at 8Â104 cell per
well. Compounds were added at 100, 33, 11, 3.7, 1.2, 0.41,
0.14, 0.05 mM (final concentration). The final volume is
100 mL per well. After incubating at 37 ꢀC for 24 h, 10 mL
of Alamar Blue (Trek) was added. After incubation for an
additional 2 h, the fluorescence intensities were read using
an Acquest (Molecular Device) plate reader.
14. We thank Dr. Mingli Chen (GNF, San Diego, CA) for
performing the toxicity assays on human hepatocytes and
bone marrow cells.