Design and campaign synthesis of pyridine-based histone deacetylase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.03.058

A lead benzamide, bearing a cyanopyridyl moiety (3), was identified as a potent and low molecular weight histone deacetylase (HDAC) inhibitor. Various replacements of the cyano group were explored at the C3-position, along with the exploration of solubility-enhancing groups at the C5-position. It was determined that cyano substitution at the C3-position of the pyridyl core, along with a methylazetidinyl substituent at the C5-position yielded optimal HDAC1 inhibition and anti-proliferative activity in HCT-116 cells.

Full text of DOI:10.1016/j.bmcl.2008.03.058

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2525–2529
Design and campaign synthesis of pyridine-based
histone deacetylase inhibitors
David M. Andrews,* Keith M. Gibson, Mark A. Graham, Zbigniew S. Matusiak,
Craig A. Roberts, Elaine S. E. Stokes, Madeleine C. Brady and Christine M. Chresta
Cancer and Infection Research, AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
Received 21 February 2008; revised 18 March 2008; accepted 18 March 2008
Available online 22 March 2008
Abstract—A lead benzamide, bearing a cyanopyridyl moiety (3), was identified as a potent and low molecular weight histone deace-
tylase (HDAC) inhibitor. Various replacements of the cyano group were explored at the C3-position, along with the exploration of
solubility-enhancing groups at the C5-position. It was determined that cyano substitution at the C3-position of the pyridyl core,
along with a methylazetidinyl substituent at the C5-position yielded optimal HDAC1 inhibition and anti-proliferative activity in
HCT-116 cells.
ꢀ 2008 Elsevier Ltd. All rights reserved.
In the eukaryotic cell, DNA is routinely compacted to
prevent transcription factor accessibility. When the cell
is activated this compacted DNA is made available to
DNA-binding proteins, thereby allowing the induction
of gene transcription.1 Nuclear DNA associates with
histones to form chromatin and the N-terminal tails of
the core histones contain lysine residues that are sites
for post-transcriptional acetylation.2,3 This reversible
process is important in transcriptional regulation and
cell-cycle progression.4 Histone deacetylases (HDACs)
are zinc-containing enzymes which catalyze the removal
of acetyl groups from the e-amino termini of lysine res-
idues clustered near the amino terminus of nucleosomal
histones and inhibition of this process is intimately
linked to the induction of gene transcription.5 In
addition, HDAC disregulation has been associated with
several cancers and HDAC inhibitors such as the com-
mercially launched Zolinza (1)2 and MS-275 (2)6 are
undergoing study for the potential treatment of cutane-
ous T-cell lymphoma and various haematological malig-
nancies (Fig. 1).7,8
O
O
O
N
NH2
NH2
H
H
HN
N
N
N
N
2
1
OH
O
O
CN
CN
O
NH2
N
H
N
H
N
4
3
O
O
Figure 1. Zolinza (1), MS-275 (2) and cyanopyridine lead 3.
elaborated into drug-like compounds.9 The cyano com-
pound 3 was shown to be more potent than the des-cy-
ano analogue by approximately 5- to 10-fold in an
HDAC isolated enzyme assay10 and therefore became
the focus of our optimization efforts (data not shown).
We rationalized that the introduction of a hydroxy-
methyl moiety at the 5-position would provide 4 which
could be used in suitably activated form to provide a
large range of base-containing and therefore more solu-
ble analogues. Further expansion of the lead structure,
substituting methyl, chloro and fluoro for cyano at the
3-position provided valuable SAR, which is described
later.
As part of an ongoing effort to identify novel HDAC
inhibitors, compound 3 and its corresponding des-cyano
analogue were identified as potent, low molecular
weight, but moderately soluble leads, capable of being
Our synthetic strategy relied upon the ability to access the
key boronate ester 6 in good yield and to exploit it in a
series of Suzuki coupling reactions with the known or
Keywords: HDAC; Histone; Deacetylase; HCT-116; Benzamide.
*
Corresponding author. Tel.: +44 1625 517126; e-mail: david.andrews
@astrazeneca.com
0960-894X/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.03.058

2526
D. M. Andrews et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2525–2529
R2
O
O
b)
b)
O
7
8
N
N
Br
N
NHBoc
H
N
CN
Cl
11a,b
O
d) f)
R2
R1N
NHBoc
O
B
H2N
O
NHBoc
N
NH2
H
N
H
N
a)
5
6
O
O
13-15 a,b,c,d
R2
Cl
Cl
e) f)
HO
9
b)
HO
N
N
N
NHBoc
H
N
HO2C
F
O
12c,d
10
b) c)
Cl
Cl
Figure 2. Synthesis of aminomethylpyridine derivatives (a) 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzoic acid, DMTMM, acetonitrile, rt,
90%; (b) 7 or 8 or 9 or 10, 6, PdCl2(dppf), NaHCO3, DME, H2O, 60 ꢁC, 5–9 h, 50–76%; (c) LiAlH4/THF, rt 15 min, H2O, 0 ꢁC, NaOH, 2 h; 10% Pd/
C, EtOH, Et3N, H2, rt 18 h 21%; d) R1NH, NaBH(OAc)3, DCM, rt, 3 h, H2O 50–80%; (e) MsCl, DCM, Et3N, 12c or 12d, rt, 2 h, R1NH, rt, 18 h 50–
80%; (f) DCM, TFA, rt 30 min, SCX2 cartridge elution—DCM, MeOH, 2 M NH3/MeOH.
commercially available haloheterocycles 7–10. (Fig. 2) 6
was obtained in 90% yield on a >500 g scale by coupling
amine 5 with 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-
2-yl)benzoic acid using N-(4,6-dimethoxy-1,3,5-triazin-
2-yl)-N-methylmorpholinium chloride (DMTMM) in
acetonitrile at room temperature. The haloheterocycles
7–10 were coupled by heating with 6 in water/DME/NaH-
CO3 at 60 ꢁC and in the presence of tetrakis(triphenyl-
phosphoranyl) palladium as the catalyst.
was more than compensated by the directness of the
approach.
Aldehydes 11a and 11b were then subject to multiple
parallel synthesis, using sodium triacetoxyborohydride
in dichloromethane and a range of secondary amines,
the yields typically being in the range 50–80%. Alco-
hols 12c and 12d were activated using methanesulfo-
nyl chloride and displaced with the same set of
secondary amines, again with overall yield typically
in the range 50–80%. The final test compounds were
liberated in 70–95% yields by Boc deprotection using
TFA/DCM and were characterized by LC–MS and
proton NMR.
The appropriate pyridine precursors for the methyl,11
cyano12 and chloro analogues were either commercially
available, or prepared according to the literature prece-
dent. The precursor to the fluoroanalogue was derived
by coupling commercially available 10 with the boronate
ester 6, as above, followed by reduction using lithium
aluminium hydride to adjust the oxidation level. The
intermediate 6-chloro compound was dechlorinated
using hydrogen gas over 10% palladium on charcoal
in ethanol/triethylamine overnight at room temperature.
Although the yield was only modest (21%), this
The biological data for an illustrative subset of the ter-
tiary amines generated by this approach is summarized
in Table 1. In all, exploitation of the four different 3-
substituted pyridine aldehydes and alcohols led to the
synthesis of approximately 300 secondary and tertiary
amine test compounds.
Table 1. Efficacy and non-efficacy properties of aminomethylpyridine-based HDAC inhibitors
Compound R1
R2
MWt HDAC1 enzyme HCT116
10
ACD
hERG
pH 7.4 mean
pH 7.4
mean pIC50
proliferation
13
mean pIC50
logD (7.4) mean pIC50 solubility (lM) measured logD
3
—
CN
314
7.91
6.54
2.05
<4.5
35
1.58
13a
13b
13c
13d
CH3 372
7.67
8.01
7.73
7.30
6.51
6.86
6.52
6.43
0.87
0.38
1.5
<4.6
<4.6
2240
746
1.18
1.22
NV
CN
Cl
F
383
393
376
4.65
1340
841
1.3
<4.6
1.69
14a
14b
14c
14d
CH3 430
7.73
8.01
7.79
7.42
6.64
6.86
6.47
6.36
1.98
1.09
2.27
2.04
<4.6
<4.6
>2350
>1940
NT
1.20
1.14
1.81
NV
CN
Cl
F
441
450
434
5.08
<4.5
>2420
15a
15b
15c
15d
CH3 444
7.72
7.98
7.88
7.48
6.62
6.85
6.51
6.19
2.31
1.42
2.6
<4.6
>2620
1385
1.44
1.28
1.68
1.60
CN
Cl
F
455
464
448
4.72
5.15
<4.5
>1490
>2160
2.37

D. M. Andrews et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2525–2529
2527
5.2
5.1
7.0
6.8
6.6
6.4
5.0
4.9
4.8
4.7
4.6
4.5
6.2
6.0
4.4
4.3
R =
CH3
CN
Cl
F
R =
CH3
CN
Cl
F
Figure 3. Pyridine core vs cell potency; line connectors show matched
pairs.
Figure 5. Pyridine core vs IonWorks hERG pIC50; line connectors
show matched pairs.
analysis suggested that the lowest hERG liabilities re-
sided within the methyl and cyano subseries, as illus-
trated in Figure 5.
The successful parallel synthesis allowed rapid system-
atic testing and visualization of the results by matched
pair analysis. When the data for a small number of com-
pounds are studied (as in Table 1), the SAR for cell po-
tency appears to be fairly flat, but when the whole
dataset is analyzed, modest but clear trends in potency
for the 3-substituent can be observed. Figure 3 shows
that when matched pairs are compared, the chloro and
fluoro-substituted pyridines are consistently less potent
than their methyl and cyano counterparts. Interestingly,
the enhanced potency of the cyano series does not ap-
pear to arise as a consequence of increased lipophilicity
as the cyano compounds generally display logD 0.4 low-
er than their halo matched pairs (Fig. 4). Extensive use
of the calculated logD was used pre-synthesis to select
compounds with moderate predicted logD, there being
a reasonable correlation between the calculated and
measured values. In each series, compounds were se-
lected from those enumerated with logD generally below
2.5.
Having established that the cyano and methyl pyridines
offered the best balance of efficacy and non-efficacy
Table 2. Rat iv DMPK properties of aminomethylpyridine-based
HDAC inhibitors
Compound Dose iv
CL
t1/2
Vss
AUC(0Àt)
(lmol/kg) (mL/min/kg) iv (h) (l/kg) iv norm
(h kg/L)
3
5.0
5.0
5.0
56.1
30.2
19.6
NV
3.0
3.8
1.0
6.9
5.4
0.30
0.55
0.84
13a
13b
13c
13d
14a
14b
14c
14d
15a
15b
15c
15d
5.0
2.5
34.0
47.5
3.6
1.1
6.3
3.6
0.49
0.35
2.5
44.0
2.0
5.1
0.37
No correlation was observed within the dataset as a
whole for the relationship between hERG IC50 and lipo-
philicity or potency,14 but once again, matched pair
Table 3. Rat oral DMPK properties of aminomethylpyridine-based
HDAC inhibitors
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
Compound
Dose oral
(lmol/kg)
t1/2 oral (h)
AUC(0Àt) oral
norm (h kg/L)
3
10.0
67.1
65.2
63.6
10.0
4.9
NV
5.5
0.19
0.68
0.74
0.51
0.25
0.02
0.32
0.16
0.26
0.18
0.33
0.08
0.21
13a
13b
13c
13d
14a
14b
14c
14d
15a
15b
15c
15d
5.5
5.2
5.4
NVa
3.2
56.7
55.6
4.6
4.4
NVa
3.4
56.4
55.0
3.9
0.8
4.3
R =
CH3
CN
Cl
F
NVa
NVa
4.5
Figure 4. Pyridine core vs measured logD (pH 7.4); line connectors
show matched pairs.
a Compound administered as part of an oral cassette.

2528
D. M. Andrews et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2525–2529
Table 4. DMPK properties of the HDAC inhibitor 13b in a range of pre-clinical species
Species/route
Dose (lmol/kg)
CL (mL/min/kg)
Vss (l/kg)
t1/2 (h)
AUC(0Àt) norm (lM h kg/lmol)
Oral %F
Nude mouse/oral
Beagle dog/iv
Beagle dog/oral
65.2
0.8
—
—
3.0
1.9
2.1
0.51
0.55
0.07
—
—
30.0
—
2.32
—
3.0
13%
properties, we sought to more broadly elucidate the
structure–property relationships by measuring the
DMPK properties of the four subseries. The oral expo-
sure of the lead compound 3 had been limited by high
clearance in the rat and we were also concerned that
its modest solubility would limit its utility in the higher
doses needed for a xenograft efficacy study. In general,
the normalized oral exposures of the halopyridines were
in a similar range to that observed for 3; those with
exposure exceeding the reference generally showed
greater hERG inhibition (Tables 2 and 3).
Smallwood and Rebecca Watson for physical measure-
ments and rat and dog DMPK data; and Howard Beeley
for assistance with NMR interpretation.
References and notes
1. (a) Beato, M. J. Mol. Med. 1996, 74, 711; (b) Wolffe, A. P.
Nature 1997, 387, 16.
2. Grant, S.; Easley, C.; Kirkpatrick, P. Nat. Rev. Drug Disc.
2007, 6, 21.
3. Marchion, D. C.; Bicaku, E.; Daud, A. I.; Sullivan, D. M.;
Munster, P. N. Cancer Res. 2005, 65, 3815.
4. Lindemann, R. K.; Gabrielli, B.; Johnstone, R. W. Cell
Cycle 2004, 3, 779.
5. Brittain, D.; Weinmann, H.; Ottow, E. Annu. Rep. Med.
Chem. 2007, 42, 337.
6. Suzuki, T.; Ando, T.; Tsuchiya, K.; Fukazawa, N.; Saito,
A.; Mariko, Y.; Yamashita, T.; Nakanishi, O. J. Med.
Chem. 1999, 42, 3001.
The methyl pyridines 14a and 15a showed lower oral
exposure than their cyano analogues 14b and 15b, pre-
cluding further study. In the case of the cyclobutyl ana-
logues 13a and 13b,15 the oral AUCs were comparable,
however, clearance of the methylpyridine 13a was higher
than that of cyanopyridine 13b (approximately 30 vs
20 mL/min/kg plasma clearance).
7. (a) Olsen, E. A.; Kim, Y. H.; Kuzel, T. M.; Pacheco, T. R.;
Foss, F. M.; Parker, S.; Frankel, S. R.; Chen, C.; Ricker,
J. L.; Arduino, J. M.; Duvic, M. J. Clin. Oncol. 2007, 25,
3109; (b) O’Connor, O. A.; Heaney, M. L.; Schwartz, L.;
Richardson, S.; Willim, R.; MacGregor-Cortelli, B.;
Curly, T.; Moskowitz, C.; Portlock, C.; Horwitz, S.;
Zelenetz, A. D.; Frankel, S.; Richon, V.; Marks, P.; Kelly,
W. K. J. Clin. Oncol. 2006, 24, 166; (c) Kelly, W. K.;
Richon, V. M.; O’Connor, O. Clin. Cancer Res. 2003, 9,
3578.
8. Kummar, S.; Gutierrez, M.; Gardner, E. R.; Donovan, E.;
Hwang, K.; Chung, E. J.; Lee, M. J.; Maynard, K.;
Kalnitskiy, M.; Chen, A.; Melillo, G.; Ryan, Q. C.;
Conley, B.; Figg, W. D.; Trepel, J. B.; Zwiebel, J.;
Doroshow, J. H.; Murgo, A. J. Clin. Cancer Res. 2007, 13,
5411.
9. Stokes, E. S. E.; Roberts, C. A.; Waring, M. J.
WO2003087057, 2003; Chem. Abstr. 2003, 139, 337995.
10. HDAC inhibitors were screened against recombinant
human HDAC1 produced in Hi5 insect cells. The enzyme
was cloned with a FLAG tag at the C-terminal of the gene
and affinity purified using anti-FLAG M2 agarose. The
deacetylase assays were carried out in a 50 lL reaction.
HDAC1 (75 ng of enzyme) diluted in 15 lL of reaction
buffer (25 mM Tris–HCl (pH 8), 137 mM NaCl, 2.7 mM
KCl, 1 mM MgCl2) was mixed with either buffer alone
(10 lL) or buffer containing compound (10 lL) for 30 min
at ambient temperature. The reaction was started by
addition of an equal volume (25 lL) of acetylated histone
H4 peptide (KI 174 Biomol) (25 lM) and incubated for
one hour at ambient temperature.. The reaction was
stopped by addition of an equal volume (50 lL) Fluor de
Lys developer (Biomol) containing Trichostatin A at
2 lM. The reaction was allowed to develop for 30 min at
ambient temperature and then fluorescence measured at
an excitation wavelength of 360 nm and an emission
wavelength of 465 nm. The IC50 values for HDAC enzyme
inhibitors were determined by performing dose–response
curves with individual compounds and determining the
The solubility of 13b, coupled with its low plasma pro-
tein binding (mouse, 54%; rat, 53%; dog, 48% free),
made it an ideal candidate for further evaluation. Com-
pound 13b showed excellent oral exposure in the rat and
moderate dog oral bioavailability as shown in Table 4.
The development of compounds in the HDAC inhibitor
class has previously been reported as an iterative exer-
cise,16 building upon the structure–activity relationships
derived for the hydroxamate and benzamide families. In
an effort to better understand the utility of the pyridine-
based benzamide inhibitors, we have adopted an
optimization campaign style of approach. Pre-synthesis
calculation of properties was used to define a large com-
pound set capable of being prepared from a small num-
ber of advanced, flexible intermediates. Extensive use of
R-group stripping and matched pair analysis was used
to select compounds for further synthesis and testing,
the focus being upon generating structure–property as
well as structure–activity relationships. Further profiling
of 13b will be reported in due course, along with other
novel HDAC inhibitors.
Acknowledgments
We thank Andrew Mortlock and Mike Waring for sup-
port, encouragement and advice. We thank Graham
Sproat and Heather Haye for cloning of HDAC 1 and
provision of HDAC enzyme; Greg Carr, Graham Dun-
can, Jon Eden, Neil Findlay, Mark Maybury, Steven
Raw, and Andy Turner for provision of intermediates
and final test compounds; Graham Sproat, Helen Cotte-
rill, Natalie Byrne and Neil Hewitt for biochemical test
data; Ross Chawner, Alison Hunter, Clare King, Janet

D. M. Andrews et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2525–2529
2529
concentration of inhibitor producing fifty percent decrease
in the maximal signal (diluent control).
11. Gibson, K. H.; Stokes, E. S. E.; Waring, M. J.; Andrews,
D. M.; Matusiak, Z. S. WO 2006077387, 2006; Chem.
Abstr. 2006, 145, 167099.
and determining the concentration of inhibitor pro-
ducing 50% decrease in maximal signal (diluent
control).
14. Schroeder, K.; Neagle, B.; Trezise, D. J.; Worley, J. J.
Biomol. Screen. 2003, 8, 50.
12. Bayer, A. G. 51373 Leverkusen, DE; Ger. Offen. DE
4429465, 1996; Chem. Abstr. 1996, 124, 343116.
15. 1H NMR spectrum: (DMSO-d6) 2.12 (m, 2H, (–
CH2NCH2CH2–)), 3.45 (m, 4H, (–CH2NCH2CH2–)), 3.92
(s, 2H, NCH2-pyr), 4.92 (br s, 2H, NH2), 6.62 (m, 1H,
CONHCHCHCHCHNH2), 6.80 (m, 1H, CONHCHCH
CHCHNH2), 7.00 (m, 1H, CONHCHCHCHCHNH2),
7.20 (m, 1H, CONHCHCHCHCHNH2), 7.99 (d, 2H,
NCCCH), 8.18 (d, 2H, COCCH), 8.39 (s, 1H,
(NC)CHCH2), 8.92 (s, 1H, NCHCCH2), 9.81 (br s, 1H,
CONH); mass spectrum: M+H+ 384.
16. Siliphaivanh, P.; Harrington, P.; Witter, D. J.; Otte, K.;
Tempest, P.; Kattar, S.; Kral, A. M.; Fleming, J. C.;
Deshmukh, S. V.; Harsch, A.; Secrist, P. J.; Miller, T. A.
Bioorg. Med. Chem. Lett. 2007, 17, 4619.
13. Inhibition of proliferation in whole cells was assayed
using the Promega cell titre 96 aqueous proliferation
assay (Promega #G5421). The HCT116 cells were
seeded in 96-well plates at 1 · 103 cells/well, and
allowed to adhere overnight. They were treated with
inhibitors for 72 h. The 20 lL of the tetrazolium dye
MTS was added to each well and the plates were
reincubated for 3 h. Absorbance was then measured
on a 96-well plate reader at 490 nm. The IC50 values
for HDAC inhibitors were determined by performing
dose–response curves with individual compounds