Design and Synthesis of Janus Kinase 2 (JAK2) and Histone Deacetlyase (HDAC) Bispecific Inhibitors Based on Pacritinib and Evidence of Dual Pathway Inhibition in Hematological Cell Lines

Article abstract of DOI:10.1021/acs.jmedchem.6b00157

Blockage of more than one oncoprotein or pathway is now a standard approach in modern cancer therapy. Multiple inhibition is typically achieved with two or more drugs. Herein, we describe a pharmacophore merging strategy combining the JAK2/FLT3 inhibitor pacritnib with the pan-HDAC inhibitor, vorinostat, to create bispecific single molecules with both JAK and HDAC targeted inhibition. A preferred ether hydroxamate, 51, inhibits JAK2 and HDAC6 with low nanomolar potency, is <100 nM potent against HDACs 2 and 10, submicromolar potent against HDACs 1, 8, and 11, and >50-fold selective for JAK2 in a panel of 97 kinases. Broad cellular antiproliferative potency is supported by demonstration of JAK-STAT and HDAC pathway blockade in several hematological cell lines, inhibition of colony formation in HEL cells, and analysis of apoptosis. This study provides new tool compounds for further exploration of dual JAK-HDAC pathway inhibiton achieved with a single molecule.

Full text of DOI:10.1021/acs.jmedchem.6b00157

Subscriber access provided by UNIV OF OREGON
Article
Design and Synthesis of Janus Kinase 2 (JAK2) and Histone
Deacetlyase (HDAC) Bispecific Inhibitors Based on Pacritinib and
Evidence of Dual Pathway Inhibition in Hematological Cell Lines
Eugene Guorong Yang, Nurulhuda Mustafa, Eng Chong Tan, Anders Poulsen, Pondy
Murugappan Ramanujulu, Wee Joo Chng, Jeffrey J. Y. Yen, and Brian W. Dymock
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00157 • Publication Date (Web): 19 Aug 2016
Downloaded from http://pubs.acs.org on August 20, 2016
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 free 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 accessible to all
readers and 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.
Journal of Medicinal Chemistry 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 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Design and Synthesis of Janus Kinase 2 (JAK2) and Histone Deacetlyase
(HDAC) Bispecific Inhibitors Based on Pacritinib and Evidence of Dual
Pathway Inhibition in Hematological Cell Lines
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Eugene Guorong Yang,1 Nurulhuda Mustafa,2♯ Eng Chong Tan,3♯ Anders
Poulsen,4,5 Pondy Murugappan Ramanujulu,1,6 Wee Joo Chng,2,7,8 Jeffrey J. Y.
Yen,3 Brian W. Dymock1,*
1 Department of Pharmacy, National University of Singapore, 18 Science Drive
4, Singapore 117543.
2
Department of Medicine, Yong Loo Lin School of Medicine, National
University of Singapore, 1E Kent Ridge Road, NUHS Tower Block Level 10,
Singapore 119228.
3 Institute of Biomedical Sciences, Academia Sinica, Taiwan, 11529.
4 Experimental Therapeutics Centre, 31 Biopolis Way, 03ꢀ01 Nanos, Singapore
138669.
5
Department of Chemistry, National University of Singapore, 3 Science Drive
3, Singapore 117543
6 Life Sciences Institute, Centre for Life Sciences Level 5, 28 Medical Drive,
National University of Singapore, Singapore 117456.
7 Cancer Science Institute, Singapore, National University of Singapore
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 146
1
2
3
4
5
8 National University Cancer Institute of Singapore, National University Health
System
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
♯these authors contributed equally
KEYWORDS: Dual inhibitor, Designed Multiple Ligand, Janus Kinase
Inhibitor, Histone Deacetylase Inhibitor.
ABSTRACT: Blockage of more than one oncoprotein or pathway is now a
standard approach in modern cancer therapy. Multiple inhibition is typically
achieved with two or more drugs. Herein we describe a pharmacophore merging
strategy combining the JAK2/FLT3 inhibitor pacritnib with the panꢀHDAC
inhibitor, vorinostat, to create bispecific single molecules with both JAK and
HDAC targeted inhibition. A preferred ether hydroxamate, 51 inhibits JAK2
and HDAC6 with low nanomolar potency, is <100 nM potent against HDACs 2
and 10, subꢀmicromolar potent against HDACs 1, 8 and 11, and >50 fold
selective for JAK2 in a panel of 97 kinases. Broad cellular antiproliferative
potency is supported by demonstration of JAKꢀSTAT and HDAC pathway
blockade in several hematological cell lines, inhibition of colony formation in
HEL cells, and analysis of apoptosis. This study provides new tool compounds
for further exploration of dual JAKꢀHDAC pathway inhibiton achieved with a
ACS Paragon Plus Environment

Page 3 of 146
Journal of Medicinal Chemistry
1
2
3
4
single molecule.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 146
1
2
3
4
5
6
7
INTRODUCTION
8
9
Cancer is a multiꢀfaceted disease often requiring multiple therapeutic
interventions for successful long term outcomes.1 Traditional approaches to
combating cancer, still in widespread use today, are nonꢀselective such as
radiation or chemotherapy.2 At the other extreme, targeted treatment of cancer
typically aims to target cancer cells via blockage of a single biological target.
However, more intensive evaluation of typical ‘targeted’ drugs often reveals
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
multipleꢀtarget mechanisms.3, Realization that simultaneous inhibition of
multiple mechanisms can lead to additive effects, or even synergy, has led to a
rise in combinations of targeted therapies with each other and with nonꢀselective
regimens.2 In today’s therapeutic armoury, an increasing array of complex
combination strategies tailored to individual diseases, and increasingly,
individual patients, is available.5 Despite many significant improvements, it is
clear that less complex approaches are required for multiple pathway
intervention to tackle challenging or resistant tumors.
There are generally three methods to achieve multiꢀtarget knockdown – use of a
cocktail of separate drugs, a multicomponent drug and the multiple ligand
approach.6ꢀ8 Drug cocktails, specifically combinations of single agents each
targeted to specific mechanisms, are associated with poor patient compliance,
complex pharmacokinetics and the challenge of drugꢀdrug interactions.9
ACS Paragon Plus Environment

Page 5 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Similarly, the multicomponent drug regimen, where two or more separate agents
are formulated into a single tablet can suffer from drugꢀdrug interactions and
differing physicochemical properties between components that can affect the
efficacy of the drugs administered. Polypharmacology is recognized as an
approach where more than one therapeutic pathway can be modulated by a
single molecule. Specific strategies to combine targeted mechanisms together
are referred to as the multiple ligand, multiꢀcomponent ligand, or designed
multiple ligand (DML) approaches. In these cases a single drug molecule has
multiple, but targeted, pharmacology.10, 11 With this approach, once the potency
ratio is determined is it fixed within the molecule hence the variable
pharmacokinetics usually seen with combinations of two single agents will not
affect the potency ratio of a DML in the same way. However it should be noted
that DMLs are just as susceptible to metabolism as any other drug, reducing or
eliminating the biological activity or pharmacodynamic profile of the
compound. Being single molecules DMLs will not suffer from drugꢀdrug
interactions, unless combined with another drug. There is also opportunity for
structural novelty – design of a DML is not just a matter of joining two
pharmacophores together, it is a creative process to balance multiple SARs
while maintaining drugꢀlike properties. Development costs for a DML are likely
to be significantly lower than development of a multiple drug combination, but
this must be balanced with other potential downsides such as molecular weight
inflation, the challenge of balancing multiple SARs and potentially more
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 146
1
2
3
4
5
6
7
8
9
challenging synthesis are all present dangers. DMLs are not new, in fact, multiꢀ
GPCR modulators, as well as multiꢀkinase inhibitors, have been known for
many years.12 In this work we focus on DMLs that target two different classes of
enzyme, rather than two enzymes performing the same type of transformation.
Deliberate strategies to adopt such a DML approach at the outset for enzyme
inhibitors in cancer are not commonly reported. Recent progress has been most
successful combining the histone deacetylase (HDAC) pharmacophore with
other enzyme inhibitor activities, such as kinases. Examples of successful
kinaseꢀHDAC DMLs have now reached the clinic in oncology pioneered by
Curis, although none are yet approved by the FDA. CUDCꢀ10113ꢀ17
(HDAC/EGFR2 inhibitor) and CUDCꢀ90718 (PI3K/HDAC inhibitor) have both
entered phase I clinical trials (Figure 1). Inspired by this progress we have been
focused on designed combinations of JAK2 kinase / HDAC DMLs for study in
cancer, using 1 (vorinostat / SAHA) and 2 (pacritinib / SB1518) as initial
templates.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 7 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
HN
O
O
O
N
N
H
HN
MeO
N
OH
9
O
Vorinostat / SAHA (1)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CUDC-101
HN
OH
O
N
O
O
N
O
O
N
S
N
HN
N
N
N
OH
N
N
N
N
H
OMe
CUDC-907
Pacritinib / SB1518 (2)
Figure 1. Structures of dual inhibitors, HDAC inhibitor 1 and JAK2/FLT3
inhibitor 2.
Janus Kinase (JAK) inhibitors19, 20 have been developed rapidly following the
discovery of the JAK2ꢀV617F21ꢀ24 mutation in patients with myeloproliferative
neoplasms (MPNs). This has led to the approval of (3R)ꢀ3ꢀcyclopentylꢀ3ꢀ[4ꢀ
(7Hꢀpyrrolo[2,3ꢀd]pyrimidinꢀ4ꢀyl)pyrazolꢀ1ꢀyl]propanenitrile
(ruxolitinib/INCB018424) and 3ꢀ[(3R,4R)ꢀ4ꢀmethylꢀ3ꢀ[methyl(7Hꢀpyrrolo[2,3ꢀ
d]pyrimidinꢀ4ꢀyl)amino]piperidinꢀ1ꢀyl]ꢀ3ꢀoxopropanenitrile
(tofacitinib/CP690550) as JAK inhibitors for the treatment of myelofibrosis
(MF)25, 26 and rheumatoid arthritis27, respectively. Compound 2 (Figure 1) has
been developed28, 29 as a dual JAK2 and fmsꢀlike tyrosine kinaseꢀ3 (FLT3)
selective inhibitor and has successfully completed its first phase 3 trial,
PERSISTꢀ1, as an orphan drug30 for myelofibrosis (MF) patients with
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 146
1
2
3
4
5
6
7
8
9
thrombocytopenia (low platelet levels).31 On the basis of the phase 3 data, an
NDA filing of 2 was recently initiated at the US FDA.32 Soon after this filing
compound 2 was put on full clinical hold by the FDA while investigations of
patient deaths are investigated. At the time of writing FDA reinstated use of 2
via an individual patient application protocol. Another clinical concern is the
onset of resistance which has been shown in the case of ruxolitinib to be due to
the reactivation of JAKꢀSTAT signaling where heterodimerisation between
activated JAK2 and JAK1 or TYK2 occurs; and not due to mutation or change
in the active site of JAK2 itself.33 JAK2 inhibitor resensitisation occurred upon
withdrawal of treatment. More importantly JAK2 inhibitor resistant cells still
remain dependent on JAK2 protein expression. Alternative therapies that
degrade JAK2 indirectly are also effective and have been postulated to be good
combinations with JAK2 inhibitors. Some of the possible targets for potentially
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
synergistic JAK2 combination therapy include histone deacetylase (HDAC)34ꢀ36
,
Heat Shock Protein 90 (HSP90)34ꢀ36, phosphoinositide 3ꢀkinase (PI3K)37, 38, and
Bꢀcell lymphoma 2 (Bcl2).39 With this in mind we postulated that a drugꢀlike
DML which could inhibit JAK2 and HDAC in a cell in a similar concentration
range could be created and would be effective in JAK2ꢀsensitive or resistant
cells.
HDAC enzymes were initially discovered to deacetylate histones affecting how
tightly the DNA is bound to the histone and subsequently regulation of gene
expression.40 There are four classes of HDAC enzymes based on their function
ACS Paragon Plus Environment

Page 9 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and DNA sequence similarity.41 However HDACs have also been found to
interact with nonꢀhistone proteins,42 thus they are implicated in a wide range of
biological functions. HDAC inhibitors have been shown to be clinically
effective as single agents43 and entered the market earlier than the JAK
inhibitors with the approval of 1 in 2007,44 romidepsin (FK228)45 and the
recently approved belinostat (PXD101, FDA approval July 2014)46 and
panobinostat (LBH589, FDA approval February 2015).47 However HDAC
inhibitors have experienced slow and difficult clinical development as single
agents due to limitations of toxicity.48 Selective inhibition of individual HDAC
isoforms, or smaller groups of isozymes, could prove beneficial in cancer
therapy with some evidence of an improved therapeutic window in clinical
studies, such as with the classꢀI selective benzamides.49 Obtaining selectivity
between individual HDAC isoforms has generally been challenging, but there
have been reports of HDAC3, 6 and 8 selective compounds.49ꢀ53 Selective
HDAC6 inhibitors are achievable by optimisation of the cap region of the
inhibitor taking advantage of the larger lip of the substrate binding pocket.54ꢀ58
Overexpression of HDAC6 correlates with oncogenesis and improved survival.
Some effects of HDAC6 inhibition have been to prevent migration and invasion
of cancer cells by acetylation of αꢀtubulin and cortactin leading to a decrease in
cell motility59, 60, sensitisation of cancer cells to known DNA damaging agents61
and induction of apotosis in cancerous cells.62, 63 Increased HDAC activity has
been observed in patients with MPN.64, 65 Specifically, HDAC inhibitors have
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 146
1
2
3
4
5
6
7
8
9
been shown to be effective against MPN both as a single agent66ꢀ68 and in
combination69ꢀ71 with JAK2 inhibitors. JAK2ꢀselective TG101209 exerts
synergistic toxic effects in MPN cell lines when combined with panobinostat.69
Ruxolitinib and panobinostat in combination lead to improved efficacy in a
mouse model of JAK2V617F driven disease.71 In preclinical mouse models of
AML, a combination of 2 and pracinostat (SB939) were found to be
synergistic.70 Thus, we hypothesised that a single compound that selectively
inhibits both JAK2 as well as HDAC enzymes could have utility, especially in
resistant MPN and AML, but perhaps also more broadly in solid tumors,
compared to a single inhibitor alone. A weaker inhibitor of either enzyme in a
dual acting DML could also potentially lead to lower toxicity levels while
maintaining therapeutic efficacy leading to an overall safer profile.72 In
summary, our goal was to design and synthesise molecules, which inhibit both
JAK2 and HDAC enzymes, seeking to explore selectivity between these two
families of diverse enzymes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Design of Dual Inhibitors
Substrate competitive HDAC inhibitors consist of a zinc binding functional
group, a long chain or extended heterocycle and a cap group (Figure 2). The
zinc binding group can be a bidentate group (eg. hydroxamic acid, benzamide,
ketoꢀthiol) which bind more strongly to the zinc cation or a monodentate group
(eg. carboxylic acid, thiols).73 With monodentate functional groups, the
ACS Paragon Plus Environment

Page 11 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
importance of the “cap” increases as the interaction with the inhibitor increases
relative to the zinc binding group. It was envisioned that we could achieve some
selectivity for HDAC6 which is slightly wider in the “cap” region relative to
HDAC1.54 Some selective HDAC6 inhibitors that have been reported include
tubacin74, tubastatin A54 and a small molecule hydroxamic acid without a cap
group.75 A lead compound was reported by Auzzas et al, where the HDAC
selectivity is reversed to achieve 10 times HDAC6 (IC50 0.4 nM) selectivity
over HDAC1.76 Even more potent compounds, with HDAC6 IC50 down to
2 pM, have been reported based on a phenylisoxazole cap. These compounds
were 10 fold more potent than 1 in pancreatic cells.65 Several natural products
(and analogues) containing macrocycles have been discovered as HDAC
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
inhibitors with most of them having moderate to high selectivity for HDAC1
77ꢀ80
over HDAC6. These are mainly macrocyclic tetrapeptides64,
,
depsipeptides81ꢀ83 (eg. romidepsin), azumamides84 and peptidomimetics.76, 85, 86
key HDAC
Zn binding
head group
O
O
N
H
HN
expected shared
OH
motif required for
binding to both
targets
1
O
linker optimisation
critical for balance of
JAK and HDAC activity
pan-HDAC
merge key
pharmacophores
O
O
H
JAK2
solvent
channel
N
N
OH
O
O
N
N
H
O
JAK2 ATP
domain
JAK2
solvent
channel
N
N
Merged single molecule
binding
macrocycle
N
N
(Designed Multiple Ligand)
H
2
JAK2 selective
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 146
1
2
3
4
5
6
7
8
9
Figure 2. Schematic showing plans for merged JAKꢀHDAC pharmacophore
taking advantage of the solvent binding channel in JAK2 and the probable
interactions to be gained by a merged inhibitor at the entrance to the HDAC
substrate binding pocket.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Aminopyrimidine 2, which contains a macrocycle as its core JAK2ꢀbinding
scaffold provides a suitable moiety to explore our hypothesis as it possesses key
features required in both JAK2 and the cap of a potentially selective HDAC
inhibitor. Furthermore, 2 is selective for JAK2 and FLT3, both enzymes which
have been shown to be beneficial in MPN and AML, and it has been shown to
be synergistic in both cellular and in vivo studies in combination with an HDAC
inhibitor.28, 29, 70 A recent study has shown that JAK2/FLT3 kinaseꢀHDAC dual
inhibitors derived from the potent cyclin dependant kinase (CDK) inhibiting
macrocycle TG02 can be prepared, however HDAC selectivity was not
explored.87 Rather than linking two known inhibitors together with a long chain,
which will typically result in a compound with high molecular weight (addition
of both compound’s molecular weight as well as the linker), a merged
pharmacophore is more desirable.6 For 2, it is known that the pyrolidine ring is
located in a solvent exposed region when binding to the enzyme and its main
purpose is to improve the aqueous solubility of the compound.28, 88 Thus it was
postulated that inserting the HDAC binding group here will retain JAK2 activity
while giving rise to potentially selective HDAC inhibition activity. We then
ACS Paragon Plus Environment

Page 13 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
planned to study the resulting JAKꢀHDAC inhibitors in a range of both solid
tumor and hematological cancer cell lines seeking evidence of target modulation
ideally within the same concentration range.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chemistry
Starting from macrocycle chloride 328 (Scheme 1), substitution with Bocꢀ
protected anilines 4a, 5a and 7a, phenol 6a and piperidineꢀ4ꢀmethyl ester 8a
gave esters 4ꢀ6 and 8 and nitro 7. Boc protection was required to increase the
acidity of the anilines, performing the reaction at 100 °C for 24 h or at 120 °C
for 18 h in DMF. Higher temperature leads to degradation of the product
(Scheme 1). Reaction with 8a was slower requiring heating for 3 days at 80 °C
in acetonitrile. Nitro substituted aniline 7 provided a route to install a longer
linker for the hydroxamic acid, similar to 1. Reduction of nitro 7, followed by
coupling with monoester acids 10aꢀ12a bearing chain lengths of 4ꢀ6 methyenes
gave bocꢀprotected esters 10ꢀ12. A more convergent approach attempted was to
perform the substitution with 4ꢀNHBoc substituted 10a (Supplemental Scheme
S1) but this gave a mixture of the desired product 12 and amide alkylation
product 10b, thus it was abandoned in favour of the longer but more practical
route using 7.
Following a merged pharmacophore strategy, compounds with long chain
hydroxamic acids connected directly to the macrocycle ring, were synthesised
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 146
1
2
3
4
5
6
7
8
9
from commercially available 13 (Scheme 3). The common intermediate phenol
14 was prepared using two strategies. The first was a 3 step protocol, without
purification or protecting groups, entailing the reduction of aldehyde 13 to the
primary benzyl alcohol followed by chlorination using thionyl chloride.
However this chlorination reaction was capricious and does not go to
completion. Furthermore, the benzylic chloride is susceptible to reaction with
water and must be used quickly. Refluxing in neat allyl alcohol gives the desired
allyl ether 14 in 35% yield over 3 steps. A cleaner method was to use
methoxymethyl (MOM) as a temporary phenol protecting group, with
concomitant reduction of the aldehyde, followed by alkylation of the benzyl
alcohol with allyl bromide under phase transfer conditions. Finally, deprotection
of the phenol is achieved by refluxing in aqueous 4 M hydrochloric acid in THF.
This 4ꢀstep protocol, while longer, gives a better overall yield of 72%. An
alternative protecting group, 4ꢀmethoxybenzyl (PMB) gave only a 47% yield
over 4 steps. Aside from the higher yield, the former was preferred due to the
ease of purification of the product from the ether byꢀproduct arising from
cleavage of the protecting group. Alkylation of phenol 14 with the appropriate
bromo esters 15aꢀ19a gave compounds 15ꢀ19. As per the reported28 synthesis of
3, the nitro group was reduced using iron and ammonium chloride. The crude
aniline was stirred with the pyrimidine at 95 °C in dioxane for 18 h to give the
diallyl compounds 21ꢀ25. Ring closing metathesis with the Grubbs II catalyst89
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 15 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
in DCM with aqueous 4 M hydrochloric acid gave macrocycles 26ꢀ30 in good
yields as approximately 85:15 mixtures of trans to cis isomers of the alkene.
7
8
9
Attempts to directly convert esters 26ꢀ30 to the hydroxamic acids using
hydroxylamine hydrochloride and base such as sodium methoxide or sodium
hydroxide at 0 °C or room temperature were not successful, with only the
starting material being recovered. Attempted hydrolysis of the ester with lithium
hydroxide was also not successful. Subsequently, it was found that by treating
the ester with potassium trimethylsilanoate the acid could be successfully
obtained. Attempts to couple the acids with OꢀTHP protected hydroxylamine
were ultimately successful using (1ꢀ[Bis(dimethylamino)methylene]ꢀ1Hꢀ1,2,3ꢀ
triazolo[4,5ꢀb]pyridinium 3ꢀoxidhexafluorophosphate (HATU) but not 1ꢀethylꢀ
3ꢀ(3ꢀdimethylaminopropyl)carbodiimide (EDC). Deprotection of the THP using
anhydrous hydrogen chloride in dioxane released the free hydroxamic acids in
low to moderate yield.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 146
1
2
3
4
5
6
7
Scheme 1. Preparation of intermediates 4-8a
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Y
R Cl
R
4a - 7a
X
3
R =
O
4, X = NBoc, Y = 4-CO2Me
5, X = NBoc, Y = 3-CO2Me
6, X = O, Y = 4-CO2Me
7, X = NBoc, Y = 4-NO2
O
O
N
N
N
H
CO2Me
b
3
N
R
8a
8
CO2Me
CO2Me
BocHN
CO2Me
NO2
HN
CO2Me
BocHN
HO
BocHN
4a
5a
6a
7a
8a
aReagents and Conditions: (a) 4a-7a, Cs2CO3, TBAI, DMF, 120 °C, 18 h, 45 to
82% yield; (b) 8a, TBAI, MeCN, reflux, 72 h, 62% yield.
ACS Paragon Plus Environment

Page 17 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
Scheme 2. Preparation of intermediates 10-12a
8
9
H
N
NO2
NH2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
b
a
R
R
R
N
N
N
CO2Me/Et
10a - 12a
n
Boc
Boc
Boc
7
9
10, n = 2
11, n = 3
12, n = 4
R =
O
CO2Me
CO2Et
HO2C
HO2C
O
10a
O
CO2Me
N
HO2C
N
N
H
11a
12a
aReagents and Conditions: (a) Fe, NH4Cl, EtOH/H2O (2:1), reflux, 3 h; (b) 10a
– 12a, NMM, ClCO2Et, DCM, O °C to RT,18 h, 59 to 79% (over 2 steps).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 146
1
2
3
4
5
6
7
Scheme 3. Preparation of intermediates 26-30a
8
9
O
O
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
j
a, b, c
OH
OH
O
CO2Me
n
OR
d, e, f
OR
15a - 19a
17, n = 3
18, n = 4
19, n = 5
O2N
O2N
O
O2N
15, n = 1
16, n = 2
13
14
g, h, i
O
O
k, l
m
O
O
O
CO2Me
n
N
CO2Me
n
N
N
N
N
N
H
N
N
23, n = 3
24, n = 4
25, n = 5
H
Cl
28, n = 3
29, n = 4
30, n = 5
21, n = 1
22, n = 2
O
20
26, n = 1
27, n = 2
Br
CO2Me
Br
Br
CO2Me
CO2Me
Br
15a
Br
16a
17a
CO2Me
CO2Me
18a
19a
aReagents and Conditions: (a) NaBH4, MeOH, RT, 2 h; (b) SOCl2, DCM, RT,
18 h ; (c) Allyl alcohol, reflux, 3 h, 35% (3 steps); (d) MOMCl, NaBH4, MeOH,
RT, 2 h; (e) c.NaOH, NEt3BnCl, Allyl bromide, RT, 72 h; (f) 4M HCl, THF,
reflux, 3 h, 72% yield (3 steps); (g) NaBH4, MeOH, RT, 2 h; (h) PPh3, CCl3CN,
DCM, RT, 16 h; (i) Allyl alcohol, reflux, 4 h, 35% (3 steps); (j) K2CO3, MeCN,
reflux, 18 h, 84ꢀ100%; (k) 20, Fe, NH4Cl, EtOH/H2O (2:1), reflux, 3 h; (l)
TsOH.H2O, Dioxane, 95 °C, 18 h, 28ꢀ62% (2 steps); (m) Grubbs II, DCM, 4M
HCl, reflux, 3 h, 62 to 83% yield.
ACS Paragon Plus Environment

Page 19 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
Scheme 4. Preparation of test compounds 31ꢀ52a
8
9
O
a, b
c, d
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R CO2Me
R CO2H
OH
N
R
31-41
H
42-52
HN
X
CONHOH 42
HN
X
CO2H
31
32
X =
O
HN
X
HN
X
43
O
CONHOH
CONHOH
CO2H
O
N
44
45
O
X
O
X
CO2H 33
N
N
H
X N
CONHOH
X N
CO2H
34
HN
NH CO2H
X
CO H
X
CONHOH
n
n
HN
X
NH CONHOH
2
X
n
O
n = 1, 37
n = 2, 38
n = 3, 39
n = 4, 40
n = 5, 41
n = 1, 48
n = 2, 49
n = 3, 50
n = 4, 51
n = 5, 52
n
O
n = 5, 35
n = 6, 36
n = 5, 46
n = 6, 47
aReagents and Conditions: (a) KOTMS, THF, RT, 2 h; (b) c.HCl, RT, 18 h, 44 to
83% yield (2 steps); (c) HATU, NH2OTHP, NEt3, DMSO, RT, 3 d; (d) 4M HCl
in dioxane, RT, 18 h, 12ꢀ62% yield.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 146
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Enzyme Inhibitory Activity
Carboxylic acids, such as valproic acid, are known to be weak HDAC inhibitors
due to the weak chelating effect of the carboxylic acid to the zinc cation in the
HDAC active site.90 Thus, the carboxylic and hydroxamic acids were both
tested in enzymatic and cellular assays. Reference compounds include 1 (panꢀ
HDAC), 2 (JAK2 selective) and Tubastatin A (HDAC6 selective). Data for
acids is shown in Table 1 and hydroxamates in Table 2. JAK2, HDAC1 and
HDAC6 were selected for initial SAR studies, the latter due to the large
macrocyclic cap group which we hypothesised would form additional
interactions at the lip of the HDAC pocket potentially conferring HDAC
isoform selectivity.91 Unsurprisingly, given the common macrocyclic ring,
significant inhibition of JAK2 enzymatic activity was observed with acids 31ꢀ34
(IC50 <39 nM). Of these compounds only aminobenzoic acids 31 and 32
afforded weak HDAC6 inhibition and were completely inactive against HDAC1
at the highest concentration tested (10 µM). Likewise ether 33 and piperidine 34
were not HDAC active, probably due to the short distance between the acid and
the bulky macrocycle. JAK2 potency was expected based on our design of the
hybrid molecules whereby the zinc chelating functional group was in the
ACS Paragon Plus Environment

Page 21 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
solvent exposed region of the JAK2 enzyme, while retaining the key binding
features of 2 in the JAK2 enzyme binding pocket. Lengthening the distance
between the Zn chelating group and macrocycle using an aniline linker yielded
progressively more HDAC6 inhibition with a 5 methylene spacer (35, 20%
inhibition at 10 µM) and with 6 methylenes (36, IC50 = 2.6 µM). It is perhaps
not entirely surprising that 36 would have HDAC activity, however, it is
completely inactive against HDAC1 at 10 µM. Compound 36 was potent
against JAK2 (IC50 = 2.3 nM), similar to 2, reassuring us that a wide range of
HDACꢀbinding side chains were fully compatible with our targeted levels of
JAK2 activity. We believed the observed high HDAC6/HDAC1 selectivity is
due to the aforementioned differences in the cap binding region of the HDACs.
Structure activity relationships seen with 35/36 are mirrored in compounds 37ꢀ
41 with aliphatic linkers of increasing length directly attached to the
macrocycle. Again, 40 (HDAC6 IC50 = 2.8 µM) with 6 methylene groups
proved to be the optimal length for HDAC6 activity.17 Following the trend
observed with 35/36, the inhibition of HDAC6 activity increases as the chain
length increases to 6 methylenes before it decreases with an additional
methylene. No HDAC inhibition could be detected with the shortest 3 and 4
carbon linkers in compounds 37 and 38, whereas the 5 carbon linked 39 was
weakly active (HDAC6 IC50 = 45 µM). Increasing the chain length to 7
methylenes decreased the HDAC6 activity to 10 µM. Potency against JAK2
was maintained although with some variations. Shortening the linker as in
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 146
1
2
3
4
5
6
37 brings the negatively charged carboxylate closer to the JAK2 pocket
decreasing JAK2 potency nearly 20 fold from 2.6 nM (40) to 47 nM (37).
7
8
9
This HDAC6 selectivity and SAR profile is mirrored in the enzymatic activity
of the hydroxamic acids but with significantly greater HDAC6 activity (>1000ꢀ
fold in the most active cases). Thus, by changing the zinc binding group it is
possible to ‘tune’ the HDAC activity to explore the effect of different levels of
HDAC activity in these multiꢀligand compounds in biological
systems. Aminobenzhydroxamates 42 and 43 had much improved HDAC6
activity as did piperidine 45, however phenol ether 44 had the most surprising
subꢀmicromolar potency against HDAC6 (IC50 = 66 nM) and also had weak
HDAC1 activity (56% at 10 µM), especially when compared with its acid
derivative 33 (HDAC6 IC50 >100 µM), representing at least a 1,500ꢀfold
improvement. With JAK2 IC50 of 1.3 nM this was the most promising
compound to date. However, we were most interested in achieving compounds
with potency in the same concentration range for both JAK2 and HDAC
isoforms. We were heartened to discover that the longer 1ꢀlike side chains in 46
and 47 conferred high HDAC6 potency, IC50s of 2.5 and 2.3 nM, respectively.
HDAC1 potency was still not detectable at the highest concentration tested
(10 µM). Furthermore, 46, displayed a high level of JAK2 potency with an IC50
of 6.9 nM, while 47 was approximately 4ꢀfold lower at 30 nM. However,
although 46 and 47 had reached our initial potency goals against these two
different enzyme targets, they were not active in cells (with the exception of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 23 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MDAꢀMB231 breast cancer cells for 47 (GI50 = 3.74 µM) (Table 2 – further
discussion below). Although 46 had a desirable LLE92 [LLE =
−LogIC50−cLogP] of 5.2, it had a high molecular weight of 666 and 5 hydrogen
bond donors. Nevertheless we were excited to find that the direct linking of the
hydroxamate to the macrocycle, as in compounds 48ꢀ52, addressed both issues
while maintaining our target enzyme inhibition profile. With the long chain
aliphatic hydroxamic acids 48ꢀ52, JAK2 activity was similar (IC50’s from 1 to
5 nM) but they had striking differences in their HDAC6 IC50s ranging from 2.1
to 510 nM. This data again illustrates the critical nature of the chain length, with
5ꢀfold improvement in HDAC6 potency on going from 3 carbons (48) to 4
carbons (49), a further 4ꢀfold with an additional methylene (50) then a further
10ꢀfold with a sixth carbon in the linker giving 51, which had the best HDAC6
activity of the series with an IC50 of 2.1 nM. Approximately 8 fold HDAC6
potency is lost with a seventh carbon (52). HDAC1 potency roughly tracks that
of HDAC6 with 51 having the highest (IC50 = 221 nM) and, although not as
high as the HDAC6/1 selectivity seen with other compounds, is still significant
at 100ꢀfold. Importantly, these compounds all maintain very strong JAK2
potency with 51 being the most potent with an IC50 of 1.4 nM. Compound 51
had a LLE of 5.2 (cLogP 3.5) matching that of 47 but with a lower MWt of 518
and two fewer hydrogen bond donors. This is probably a contributing reason to
the significant cellular activity seen with 51 and 52 compared with 46 and 47.
We were curious to understand the reasons for the very high HDAC6 potency
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 146
1
2
3
4
and selectivity hence we undertook molecular modelling studies against the 3
5
6
target enzymes focusing on the JAK2/HDAC6 equipotent 51.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 25 of 146
Journal of Medicinal Chemistry
1
2
3
4
5
6
Table 1. Enzyme inhibition data for carboxylic acids
7
8
9
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
N
R
N
N
H
Enzyme IC50 (µM)a
(% Inhibition at 10µM)
Compound R
HDAC1
HDAC6
JAK2
Tubastatin HDAC6 selective
(37%)
0.306b
0.0035
0.020b
ꢀ
ꢀ
1
2
(panꢀHDAC inhibitor)
N
ꢀ
ꢀ
0.001
O
31
32
33
(0%)
64
<0.039c
HN
OH
O
OH
(0%)
(0%)
53
<0.039b
<0.039b
HN
OH
O
>100
O
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 146
1
2
3
4
5
6
7
Enzyme IC50 (µM)a
(% Inhibition at 10µM)
Compound R
8
9
HDAC1
(0%)
HDAC6
>100
JAK2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
N
34
<0.039b
OH
HN
HN
NH
OH
35
36
37
(10.4%)
(0%)
(20.4%)
ꢀ
O
O
OH
0.00231
±0.0001
0.047
NH
O
2.6
ꢀ
O
O
(0%)
OH
O
±0.002
OH
38
39
(0%)
(0%)
ꢀ
<0.039b
<0.039b
O
45
OH
0.00257
±0.0001
OH
40
(2.8%)
(3.8%)
2.8
10
O
O
41
<0.039b
OH
a
all compounds tested in the same assay format. IC50 data are reported as the
average of two independent determinations (see Supporting Information for IC50
ACS Paragon Plus Environment

Page 27 of 146
Journal of Medicinal Chemistry
1
2
3
4
curves for 36 and 40). Note that reported literature values vary depending on
5
6
assay format. bReported values.119 clowest concentration tested 39 nM.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 146
1
2
3
4
5
6
Table 2. Enzyme Inhibition data for hydroxamic acids
7
8
9
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
N
R
N
N
H
Enzyme assay IC50 (µM)a
(% inhibition at 10µM)
Compound R
HDAC1 HDAC6 JAK2
Tubastatin (HDAC6 selective)
(37%)
0.306c
0.0035b
0.020c
ꢀ
ꢀ
1
2
(panꢀHDAC inhibitor)
N
ꢀ
ꢀ
0.001
0.395
±0.013
2.90
0.00171
±0.0003
0.009
O
HN
42
(18%)
HN OH
O
OH
NH
(9%)
43
44
HN
O
±0.31
0.066
±0.0005
0.00127
HN OH
O
(56%)
ACS Paragon Plus Environment

Page 29 of 146
Journal of Medicinal Chemistry
1
2
3
4
Enzyme assay IC50 (µM)a
5
6
7
8
9
(% inhibition at 10µM)
HDAC1 HDAC6 JAK2
±0.024 ±0.00005
Compound R
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0.899
±0.319 ±0.00001
0.0025 0.0069
±0.0014 ±0.00015
0.00097
O
N
(20%)
(16%)
(14%)
45
HN OH
OH
NH
HN
HN
NH
O
46
47
48
49
50
O
0.0023
±0.0009
0.510
0.030
±0.009
0.0047
HN
O
OH
NH
O
2.40
±0.43
3.09
O
OH
N
H
±0.004 ±0.00055
0.107 0.00245
±0.014 ±0.00001
0.028 0.0043
±0.0007 ±0.0005
H
N
OH
O
±0.15
4.07
O
OH
N
H
±0.21
ACS Paragon Plus Environment