Design, Synthesis, and Biological Evaluation of First-in-Class Dual Acting Histone Deacetylases (HDACs) and Phosphodiesterase 5 (PDE5) Inhibitors for the Treatment of Alzheimer's Disease

Article abstract of DOI:10.1021/acs.jmedchem.6b00908

Simultaneous inhibition of phosphodiesterase 5 (PDE5) and histone deacetylases (HDAC) has recently been validated as a potentially novel therapeutic approach for Alzheimer's disease (AD). To further extend this concept, we designed and synthesized the first chemical series of dual acting PDE5 and HDAC inhibitors, and we validated this systems therapeutics approach. Following the implementation of structure- and knowledge-based approaches, initial hits were designed and were shown to validate our hypothesis of dual in vitro inhibition. Then, an optimization strategy was pursued to obtain a proper tool compound for in vivo testing in AD models. Initial hits were translated into molecules with adequate cellular functional responses (histone acetylation and cAMP/cGMP response element-binding (CREB) phosphorylation in the nanomolar range), an acceptable therapeutic window (>1 log unit), and the ability to cross the blood-brain barrier, leading to the identification of 7 as a candidate for in vivo proof-of-concept testing (Cuadrado-Tejedor, M.; Garcia-Barroso, C.; Sánchez-Arias, J. A.; Rabal, O.; Mederos, S.; Ugarte, A.; Franco, R.; Segura, V.; Perea, G.; Oyarzabal, J.; Garcia-Osta, A. Neuropsychopharmacology 2016, in press, doi: 10.1038/npp.2016.163).

Full text of DOI:10.1021/acs.jmedchem.6b00908

Subscriber access provided by Northern Illinois University
Article
Design, synthesis, and biological evaluation of first-in-class dual
acting histone deacetylases (HDACs) and phosphodiesterase
5
(PDE5) inhibitors for the treatment of Alzheimer’s disease
Obdulia Rabal, Juan A Sánchez-Arias, Mar Cuadrado-Tejedor, Irene De Miguel, Marta Pérez-González,
Carolina García-Barroso, Ana Ugarte, Ander Estella-Hermoso de Mendoza, Elena Sáez, Maria
Espelosin, Susana Ursua, Tan Haizhong, Wu Wei, Xu Musheng, Ana Garcia-Osta, and Julen Oyarzabal
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00908 • Publication Date (Web): 08 Sep 2016
Downloaded from http://pubs.acs.org on September 9, 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 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Design, synthesis, and biological evaluation of firstꢀinꢀclass dual acting
histone deacetylases (HDACs) and phosphodiesterase 5
inhibitors for the treatment of Alzheimer’s disease
(PDE5)
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
1
,5
1,5
2,3,5
Obdulia Rabal,
Juan A. SánchezꢀArias,
Mar CuadradoꢀTejedor,
Irene de
1
2
2
1
Miguel, Marta PérezꢀGonzález, Carolina GarcíaꢀBarroso, Ana Ugarte, Ander
1
1
2
2
EstellaꢀHermoso de Mendoza, Elena Sáez, Maria Espelosin, Susana Ursua, Tan
4
4
4
2,*
1,*
Haizhong, Wu Wei, Xu Musheng, Ana GarciaꢀOsta, and Julen Oyarzabal.
1
2
Small Molecule Discovery Platform, Molecular Therapeutics Program, Neurobiology
of Alzheimer´s disease, Neurosciences Division, Center for Applied Medical Research
(
CIMA), University of Navarra, Avenida Pio XII 55, Eꢀ31008 Pamplona, Spain
3
Anatomy Department, School of Medicine, University of Navarra, Irunlarrea 1, Eꢀ
3
1008 Pamplona, Spain
4
WuXi Apptec (Tianjin) Co. Ltd., TEDA, No. 111 HuangHai Road, 4th Avenue, Tianjin
3
00456, PR China
5
These authors contributed equally to this work.
ABSTRACT
Simultaneous inhibition of phosphodiesterase 5 (PDE5) and histone deacetylases
(HDAC) has recently been validated as a potentially novel therapeutic approach for
Alzheimer’s Disease (AD). To further extend this concept, we designed and synthesized
the first chemical series of dual acting PDE5 and HDAC inhibitors, and we validated
this systems therapeutics approach. Following the implementation of structureꢀ and
knowledgeꢀbased approaches, initial hits were designed and were shown to validate our
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
hypothesis of dual in vitro inhibition. Then, an optimization strategy was pursued to
obtain a proper tool compound for in vivo testing in AD models. Initial hits were
translated into molecules with adequate cellular functional responses (histone
acetylation and cAMP/cGMP response elementꢀbinding (CREB) phosphorylation in the
nanomolar range), an acceptable therapeutic window (>1 log unit) and the ability to
cross the bloodꢀbrain barrier, leading to the identification of 7 as a candidate for in vivo
proofꢀofꢀconcept testing (described in ref 23).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION
Multitarget drugs have emerged as an innovative therapeutic approach for Alzheimer’s
disease (AD) due to the complex etiology of this neurodegenerative disease and its
1
multifactorial progression. As observed for other conditions (e.g., complex diseases), it
is becoming clear that AD therapies should focus on the simultaneous modulation of
1
multiple targets implicated in the disease. Among these targets, phosphodiesterases
2
3–6
(
PDEs) and epigenetic targets, primarily histone deacetylases (HDACs) , have
recently attracted much potential therapeutic interest to restore memory function.
PDEs hydrolyze the second messengers cyclic guanosine monophosphate (cGMP) and
7
cyclic adenosine monophosphate (cAMP) and are extensively distributed in the brain.
In fact PDE5, a cGMPꢀspecific phosphodiesterase, is upꢀregulated in the brains of AD
8
patients compared with ageꢀmatched healthy control subjects. Consequently, cGMP
levels, but not cAMP levels, are significantly decreased in the cerebrospinal fluid (CSF)
8
of AD patients when compared with nonꢀdemented controls. Inhibition of
phosphodiesteraseꢀ5 (PDE5), a cGMPꢀspecific phosphodiesterase, elevates cGMP
ACS Paragon Plus Environment

Page 3 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
levels, which may ultimately promote gene transcription by directly and/or indirectly
2
activating CREB. Moreover, by favoring the inactive form of GSK3β (phosphorylated
9
,10
at GSK3βꢀSer9), PDE5 inhibition decreases levels of phosphorylated Tau (pTau).
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
Specific PDE5 inhibitors (sildenafil 1, vardenafil 2 and tadalafil 3, Chart 1) approved
7
for the treatment of erectile dysfunction and pulmonary arterial hypertension have been
shown to improve memory performance and/or enhance synaptic plasticity and
9–11
cognitive function in different animal models of AD.
Histone deacetylases (HDACs) comprise a family of 18 genes in humans and are
divided into four groups: class I (HDACs 1, 2, 3, 8), class IIa (HDACs 4, 5, 7, 9), class
IIb (HDACs 6, 10) and class IV (HDAC11). HDACs are epigenetic modulators that
deacetylate lysine residues in histone and nonꢀhistone substrates. Although already a
12
proven strategy for the treatment of cancers, inhibition of HDACs has attracted much
1
2
interest for the treatment of neurodegenerative disorders in recent years, with class I
HDACs and HDAC6 being implicated in AD memoryꢀrelated dysfunction. Class I
HDACs, particularly HDAC2, predominantly localize in the nucleus and reduce the
13,14
transcription of CREBꢀregulated genes that are important for learning and memory,
15
and HDAC1 activity may be neuroprotective. Notably, HDAC2 and HDAC6 are overꢀ
expressed in the cortex and hippocampus of AD patients, although the cause and effect
14,16
of this upꢀregulation remain unknown.
Chronic treatment with suberoylanilide
hydroxamic acid 4 (SAHA; vorinostat, Chart 1), a clinically approved panꢀHDAC
inhibitor for the treatment of cutaneous T cell lymphoma (CTCL), enhanced memory in
13
animal models. HDAC6, the major cytoplasmatic deacetylase in mammalian cells,
targets αꢀtubulin, among other proteins. Increasing αꢀtubulin acetylation via HDAC6
17,18
19,20
inhibition may facilitate the amelioration of tau
and amyloid pathologies
by
promoting tau clearance and decreasing Aβ levels, respectively.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In this context, we have recently demonstrated the beneficial synergistic effects of
concomitant HDAC and PDE5 inhibition in the Tg2576 murine model of AD using 3
and the panꢀHDAC inhibitor 4, thereby establishing the basis for a potential new
21
symptomatic and diseaseꢀmodifying strategy to treat AD. Based on these results and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
considering that i) toxicity is associated with strong inhibition of HDAC class I
2
2
isoforms, ii) simultaneous inhibition of HDAC and PDE5 exerts a synergistic effect on
21
histone acetylation and thus strong inhibition of HDAC class I is not required, iii)
histone acetylation in conjunction with CREB activation, achieved through PDE5
2
3
inhibition, may facilitate the transcription of specific memoryꢀrelated genes, iv)
22
inhibition of HDAC6 does not affect cell survival and may facilitate the degradation
1
9,20
of misfolded proteins (such as Aβ and pTau),
v) HDAC inhibitors show poor
24,25
26
permeability
and brain availability and vi) a single agent does not lead to the
27
additive toxicity that is often observed with combination therapy, our next step was to
obtain brainꢀpenetrating dual inhibitors with moderate HDAC class I activity as well as
potent HDAC6 and PDE5 inhibition. Thus, we set out with the goal of designing novel
dual PDE5 and HDAC inhibitors with the appropriate profiles for potency, selectivity
and pharmacokinetic properties to consider for in vivo testing in AD mouse models.
Compound 7 (CMꢀ414, Chart 1) fulfilled these requirements in terms of primary
activities (IC50 values of 60 nM, 310 nM, 490 nM, 322 nM and 91 nM against PDE5,
HDAC1, HDAC2, HDAC3 and HDAC6, respectively), crossing the bloodꢀbrain barrier
(BBB), inducing AcH3K9 acetylation and CREB phosphorylation in the hippocampus
and rescuing longꢀterm potentiation (LTP) in APP/PS1 mice. With additional
consideration of its adequate ADME and pharmacokinetic profiles, 7 was selected for in
23
vivo proof of concept (PoC) testing. In this article, we present a detailed account of the
ACS Paragon Plus Environment

Page 5 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
discovery of this novel firstꢀinꢀclass chemical series with dual acting PDE5 and HDAC
inhibitory activities, from initial hits (e.g., 13c) to lead identification (7).
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
N
O
H
N
O
S
HN
N
O
S
HN
N
N
N
N
N
N
O
O
O
O
N
O
O
N
N
N
O
1
2
3
Sildenafil
Vardenafil
Tadalafil
PDE5A IC50 = 8.5 nM
PDE5A IC50 = 0.89 nM
PDE5A IC50 = 9.4 nM
OH
HN
OH
HN
O
O
H
N
OH
N
H
N
N
O
N
O
NH
N
H
N
H
N
4
5
6
SAHA
Panobinostat
Quisinostat
HDAC1 IC50 = 30nM
HDAC2 IC50 = 170nM
HDAC3 IC50 = 100nM
HDAC6 IC50 = 38nM
HDAC1 IC50 = 2.5nM
HDAC2 IC50 = 13.2nM
HDAC3 IC50 = 2.1nM
HDAC6 IC50 = 10.5nM
HDAC1 IC50 = 0.11nM
HDAC2 IC50 = 0.33nM
HDAC3 IC50 = 4.86nM
HDAC6 IC50 = 76.8nM
O
N
O
HN
N
HDAC1 IC50 = 310nM
HDAC2 IC50 = 490nM
HDAC3 IC50 = 322nM
HDAC6 IC50 = 91nM
PDE5A IC50 = 60nM
N
7
CM-414
O
HN
OH
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Chart 1. Known PDE5 inhibitors shown to improve memory (1, 2, 3), HDAC inhibitors
(
4, 5, 6) and the structure of a novel therapeutic tool 7. PDE5 IC50 inhibition values
28
29
taken from Ref. for 1, 2, 3. HDAC inhibition IC50 values for 4 extracted from Ref. ,
30
31
values for 5 extracted from Ref. and values for 6 extracted from Ref. .
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
RESULTS
Rational Design
One interesting approach to generate multivalent ligands is to combine key structural
features facilitating binding to HDAC and PDE into one molecule to obtain a new
chemical entity; however, this is not a straightforward procedure, and we must identify
the appropriate common features, substitution sites and growing vectors to maintain
primary activities without interference. Indeed, in this particular case, given the
structures of both protein families (see below), the design process goes one step further
than simply incorporating the pharmacophoric features of HDAC inhibitors (HDACi)
and PDE inhibitors into one molecule as the sum of two parts to obtain a single agent
with dual activity. This strategy has been particularly useful to derive bifunctional
HDACi anticancer agents due to the presence of large hydrophobic patches at the
32
HDAC surface rim. We envisioned using the sildenafil central core to append HDAC
pharmacophores.
However, before commencing any design and synthetic efforts around the sildenafil
2
1
scaffold, we confirmed that the synergestic effects achieved by 3 and 4
on the
induction of histone 3 acetylation at lysine 9 (AcH3K9) are mechanismꢀofꢀaction
(MoA)ꢀdependent; then, a combination of 4 and 1 was tested and quantified using
ACS Paragon Plus Environment

Page 7 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
AlphaLisa technology in the SHꢀSY5Y neuroblastoma cell line. After treating cells with
5
0 nM of 4, AcH3K9 marks increased by 1.5ꢀfold compared to nonꢀtreated cells (Figure
), and this induction was significantly stronger (P value < 0.01) when combined with
1
concentrations of 1 higher than 400 nM (1.95ꢀfold change over vehicleꢀtreated cells at
00 nM of compound 1, Figure 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
4
Figure 1. Detection of AcH3K9 assayed using SHꢀSY5Y cells and AlphaLisa
technology. SHꢀSY5Y cells were treated with 4 and 1 for 2 hours (** p≤0.01,
***p≤0.001).
Once the synergistic effect between 4 and 1 was confirmed, our design strategy
commenced. Typically, the classical pharmacophore for HDACi consists of a
hydrophobic recognition capping group (also known as a surface recognition motif) that
is able to interact with the rim of the catalytic tunnel, a zincꢀbinding group (ZBG) that is
2
+
able to complex the Zn ion at the bottom of the catalytic cavity and a hydrophobic
1
2,33
linker connecting the two parts along the 11ꢀÅ hydrophobic channel.
There are
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
various ZBGs for HDACi, including hydroxamic acids, aminobenzamides, carboxylates
(shortꢀchain fatty acids), electrophilic ketons, thiols, mercaptoacetamides and 3ꢀ
34
hydroxypyridinꢀ2ꢀthiones.
Initially, and for this proofꢀofꢀconcept series, the
hydroxamic moiety was chosen as a ZBG because it is one of most wellꢀestablished
functionalities for chelating the zinc ion at the catalytic site of HDACs. Examination of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
5
the crystal structure of 1 bound to PDE5 (PDB entry 1TBF , Figure 2a) suggested that
linking the hydroxamic moiety to the methylpiperazine would project this ZBG
substituent into the solvent region and be well tolerated from a potency perspective.
From the viewpoint of HDAC inhibitory activity, the sildenafil core would serve as a
cap group. As a straightforward strategy for HDACi linker design, different linker
moieties contained in reported potent HDAC inhibitors were analyzed in the context of
36
the structure of HDAC2 complexed with 4 (PDB entry 4LXZ , Figure 2b), considering
the 11ꢀÅ cavity length.
3
5
Figure 2. (A) Crystal structure of 1 in the PDE5 cavity (PDB entry 1TBF ). The
pyrazolopyrimidinone group of 1 makes bidentate Hꢀbonds with the conserved Q817.
ACS Paragon Plus Environment

Page 9 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The piperidinylsulfonamide group is solventꢀoriented towards the Hꢀloop region
3
6
(
residues 660ꢀ683). (B) Complex of 4 and HDAC2 (PDB entry 4LXZ ). The NH of the
amide group of 4 makes an Hꢀbond contact with the wellꢀconserved residue Asp104 of
HDAC2 (Asp99 for HDAC1 and Asp567 for HDAC6).
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
This combination of structureꢀbased and knowledgeꢀbased approaches, together with
consideration to synthetic accessibility, enabled the rapid design of potential dual
PDE5/HDAC inhibitors. We envisioned attaching the following to the
piperidinylsulfonamide group of 1: flexible alkyl linkers of varying lengths as in
3
0
compound 4, a cinnamic hydroxamic acid analog as in 5 and pyrimidylhydroxamic
31
acids as in 6 , resulting in novel Nꢀ4ꢀsubstitutedꢀpiperazine derivatives 13aꢀ13d as
potential dual PDE5/HDAC inhibitors. These compounds were synthesized from
commercially
available
5ꢀ(2ꢀethoxyphenyl)ꢀ1ꢀmethylꢀ3ꢀpropylꢀ6Hꢀpyrazolo[4,3ꢀ
d]pyrimidinꢀ7ꢀone (8) (Scheme 1). Selective sulfonylation at the 5’ꢀposition of the
phenyl ring afforded 9, which was converted into esters 10aꢀ10d via reactions with
appropriated amines. Then, the corresponding carboxylic acids were obtained through
hydrolysis and transformed into the THPꢀprotected hydroxamic acids 12aꢀ12d by
reacting with Oꢀ(tetrahydroꢀ2Hꢀpyranꢀ2ꢀyl)hydroxylamine (THPONH
2
)
using
EDC/HOBt as the coupling system. Final deprotection under acidic conditions afforded
us the desired compounds 13aꢀ13d.
Scheme 1.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
i)
ii) or ii) and iii)
iv)
8
9
1
0a-g, R = CH or CH CH
3 2 3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
v)
vi)
1
1a-g
12a-g
1
3a-g
A = a)
b)
c)
d)
e)
f)
g)
3 3
Conditions: i) ClSO H, rt, 2 h; ii) corresponding amine, Et N (optional), EtOH, MW, 100 ºC, 1ꢀ2 h; iii)
ethyl 3ꢀbromopropanoate or ethyl 4ꢀbromobutanoate, K CO , CH CN, MW, 100 ºC, 2 h; iv) LiOH·H O,
2
3
3
2
THF/MeOH/H
2
O (10:1:5), rt, overnight; v) EDC·HCl, HOBt, THPONH
2
, NMM, DMF, rt, overnight; vi)
HCl/1,4ꢀdioxane (2.0 M), 1,4ꢀdioxane or CH Cl (optional) rt, 3 h.
2
2
Compounds 13aꢀ13d were evaluated for their inhibition against PDE5, HDAC1,
HDAC2 and HDAC6 activity (Table 1). Purified fullꢀlength recombinant human HDAC
proteins were used to monitor HDAC activity (HDAC1, 2 and 6 were routinely included
in our screening funnel, and HDAC3 activity was evaluated for selected compounds).
As shown in Table 1, all of these compounds are potent PDE5 inhibitors, with IC50
2
8
values in the low nanomolar range (2ꢀ3 nM), comparable to 1 (IC50 is 8.5 nM or 4 nM
in our assay setꢀup, see Methods). Moreover, these compounds exhibit HDAC
inhibitory activity with different profiles for potency and isoform selectivity, validating
our initial hypothesis to design dual PDE5/HDAC inhibitors. Alkyl (13a, 13b) linkers
resulted in low micromolar or midꢀnanomolar inhibitors (13b against HDAC6). The
ACS Paragon Plus Environment

Page 11 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
comparison of 13a with 13b suggests that increasing the length of the linker may have a
positive effect on the HDAC activities of the three isoforms. However, due to the size of
these derivatives, this strategy was not further contemplated. The cinnamic derivative
13d showed similar potency against HDAC1 and HDAC2 compared to alkyl derivatives
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
but a remarkable potency against HDAC6 (IC50 value of 89 nM), with >1 log units of
selectivity over the class I HDACs (HDAC1 and HDAC2). Conversely, the
pyrimidylhydroxamic 13c showed excellent potency against HDAC1 (IC50 of 8 nM),
comparable to that of the standard compound4 (IC50 of 11 nM, Chart 1), and less
potency against HDAC2 (117 nM) and HDAC6 (268 nM). The suitable potency of this
pyrimidylhydroxamic moiety against class I HDAC isoforms has also been previously
31,37
reported for 6 and related analogues.
Table 1. Initial set of potential PDE5/HDAC inhibitors bearing a sulfonamide moiety.
1
'
Cpd R1
PDE5 HDAC1 HDAC2 HDAC3ꢀ
HDAC6
IC50 nM
A IC50 IC50
IC50
NCOR2
nM
nM
nM
IC50 nM*
13a
3
10500
>20000
2360
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
13b
2
1100
4640
360
13c
13d
3
2
8
117
36
268
89
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1340
6970
13e
0.5
406
1940
87
13f
0.6
1
57
341
54
59
84
13g
356
1310
*
38
HDAC3_NCOR2 values obtained at BPS.
Hit Explosion. Exploring the 5’ꢀposition of the phenyl ring of compound 1 (R1 at
’) and SAR analysis.
1
Encouraged by these early in vitro results for our dual acting compounds that were
initial hits, our strategy focused on identifying molecules with previously defined
primary activities (moderate HDAC class I as well as potent HDAC6 and PDE5
inhibitors) that were CNSꢀpenetrating. The physicochemical properties of compounds
ACS Paragon Plus Environment

Page 13 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
13aꢀ13d are far outside the traditional range for CNS drugs, with high topological
2
39,40
surface areas (TPSA > 90 Å ) and high molecular weights (MW > 450).
Thus we
sought to explore different alternatives for the piperidinylsulfonamide group of
compound 1, not only to optimize potency and examine different in vitro HDAC
inhibitory profiles but also to obtain derivatives that demonstrated a reduced polar
surface area to deꢀrisk poor bloodꢀbrain barrier (BBB) penetration.
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
From the viewpoint of PDE5 activity, based on the crystal structure of compound 1
bound to PDE5 (Figure S1) and the previous analysis of structureꢀactivity relationships
4
1
(SAR) , the piperidinylsulfonamide group of compound 1 is not essential for potent
35,42,43
44
PDE5 inhibition. Moreover, several published complexes of ligands 1
and 2 with
PDE5 have exhibited significantly different orientations of the methylpiperazine portion
of both ligands, stressing the potential of PDE5 to accommodate different substituents at
this region. Thus, our SAR strategy to effectively balance dual PDE5/HDAC potency,
differential selectivity profiles versus HDACs and ADME properties focused on i)
exploring different attachment points (connecting bonds) and enabling different
geometries at the 5’ꢀposition of the phenyl ring of molecule 1 (sulfonamideꢀ, amineꢀ,
etherꢀ and carbonꢀlinked substituents) as well as ii) varying the substituents acting as
HDAC linkers to occupy the catalytic channel between the ZBG and the surface
recognition motif of HDACi (that corresponds to the sildenafil core, acting as a driving
force for PDE binding). Moreover, in this case many analogues can be easily
synthesized due to the selectivity of electrophilic attack on the 5’ꢀposition of the phenyl
ring of intermediate 8.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
As a first initial exploration, the sulfonamide linker was retained with variations in the
piperidinyl ring designed to: i) increase its hydrophobicity by removing the positively
charged nitrogen (13e) and ii) increase the flexibility between the sildenafil core
(capping group) and the hydrophobic linker binding in the HDAC catalytic channel by
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
introducing a secondary sulfonamide as an attachment point (13f), together with
increasing the planarity of the HDACi linker by removing the piperidinyl group (13g).
The synthesis of these compounds was performed as previously described for
hydroxamic acids 13aꢀ13d by coupling benzenesulfonyl chloride (9) with appropriated
amines (Scheme 1), and the in vitro evaluation is listed in Table 1. These three
derivatives (13eꢀ13g) resulted in potent HDAC6 inhibitors with IC50 values < 100 nM.
Concerning class I HDACs, derivatives 13e and 13g exhibited enhanced potency (with
IC50 values of approximately 400 nM against HDAC1) compared to their lessꢀ
hydrophobic parent compounds 13a and 13d (with IC50 >= 1300 nM against HDAC1),
suggesting that a positively charged amine at this position of the class I HDAC channel
is not well tolerated. Conversely, the secondary sulfonamide bearing
a
pyrimidylhydroxamic group 13f was less potent against HDAC1 compared to the
privileged substructure conferring potent class I HDAC activity in 13c (57 nM versus 8
nM; 0.8 log units). As expected, all of these modifications exerted minor influences on
PDE5 activity.
We next examined a variety of heteroatom (nitrogen and oxygen) bonded substituents at
the 5’ꢀposition of the phenyl ring of compound 1 (Table 2): secondary amines (21aꢀ
21e), linear tertiary amines (21f), cyclic tertiary amines (30aꢀ30e) and ethers (30fꢀ30g,
37). Together with the heteroatom connection, a variety of alkylꢀ (21a, 30a, 30b),
cycloalkylꢀ (21b, 30f), phenylꢀ (30g), piperidylphenylꢀ (21c), piperidylpyrimidineꢀ
ACS Paragon Plus Environment

Page 15 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(21d, 21e, 21f) and nitrogenꢀbonded spiro substituents (30c, 30d, 30e) (Table 2) were
selected to cover a large variety of hydrophobic, electronic, and steric properties that
might result in different HDAC profiles, as previously obtained for compounds in Table
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
Amines 21aꢀ21f were synthesized as illustrated in Scheme 2. Intermediate 8 was
converted into amine 15 via nitration and reduction of the nitro group. Then, esters 18aꢀ
1
8f were obtained through reductive amination (and subsequent BOCꢀdeprotection and
coupling with ethyl 2ꢀchloropyrimidineꢀ5ꢀcarboxylate in the case of esters 18d, 18e and
8f). These intermediates were transformed into desired hydroxamic acids via ester
1
hydrolysis, reaction with THPONH
2
and acidic cleavage of the protecting group.
Scheme 2.
iii) or iv)
i)
iii) or iii) and v)
vii)
8
1
4, R = NO2
ii)
1
6a, n = 0, R1 = H, R2 = BOC
18a-f, R = CH3 or CH CH3
2
1
5, R = NH2
16b, n = 1, R1 = H, R2 = BOC
1
6c, n = 0, R1 = CH , R2 = BOC
3
vi)
1
7a, n = 0, R1 = H, R2 = H
17b, n = 1, R1 = H, R2 = H
1
7c, n = 0, R1 = CH , R2 = H
3
viii)
ix)
vi)
19a-f
20a-f
21a-f
A = a)
b)
c)
d)
e)
f)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Conditions: i) H
corresponding carbonyl compound, AcOH, NaBH(OAc)
dimethoxypropanoate, TFA, Et SiH, CH Cl , rt, overnight; v) paraformaldehyde, AcOH, NaBH(OAc)
CH Cl , 60 ºC, overnight; vi) HCl/EtOAc (1.0 or 4.0 M), rt, 1ꢀ4 h; vii) K CO , ethyl 2ꢀchloropyrimidineꢀ
ꢀcarboxylate, CH CN, 40 ºC, overnight; viii) LiOH·H O, THF/MeOH/H O (3:3:2 or 3:1:1), 25ꢀ40 ºC,
overnight; ix) EDC·HCl, HOBt, THPONH , NMM, DMF, rt, overnight.
2
SO
4
, KNO
3
, 0 ºC, 20 minutes; ii) Pd/C, H
2
(1 atm), MeOH, rt, overnight; iii)
3
,
CH Cl rt, overnight; iv) 3,3ꢀ
2
2
,
3
2
2
3
,
2
2
2
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
3
2
2
2
The synthesis of tertiary amines 30aꢀ30e and ethers 30f and 30g was performed as
shown in Scheme 3 from iodure 22. This compound was transformed into boronic acid
26, and then esters 27bꢀ27g were obtained through reactions with different amines or
alcohols. Conversely, ester 27a was synthesized from iodure 22 after Buchwald–
Hartwig amination with 1,4ꢀdioxaꢀ8ꢀazaspiro[4.5]decane, acidic deprotection, Horner–
Wadsworth–Emmons reaction with methyl 2ꢀdiethoxyphosphorylacetate and reduction
2
of the double bond under H . Finally, hydroxamic acids 30aꢀ30g were prepared from
esters 27aꢀ27g employing the strategy previously described.
Scheme 3.
ACS Paragon Plus Environment

Page 17 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
iii)
iv) and v)
23
24
ii)
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
xi)
i)
8
22
2
7a-g, R = CH or CH CH
3
2
3
vi)
viii) or
viii) ix) and x)
vii)
25
26
ix)
xii)
3
0a-g
28a-g
2
9a-g
A = a)
b)
c)
d)
e)
f)
g)
Conditions i) NIS, TFA, 0 ºC, then rt, overnight; ii) tꢀBuOK, Pd
xantphos, toluene, 120 ºC, MW, 1 h; iii) HCl/THF (6.0 M), 70 ºC, overnight; iv) methyl 2ꢀ
diethoxyphosphorylacetate, NaH, THF, 0 ºC, 1 h, then 24, rt, overnight; v) Pd/C, MeOH, H (1 atm), rt, 3
h; vi) 4,4,4′,4′,5,5,5′,5′ꢀoctamethylꢀ2,2′ꢀbiꢀ1,3,2ꢀdioxaborolane, PdCl (dppf), KOAc, 1,4ꢀdioxane, 80ꢀ100
ºC, 48 h; vii) NaIO , NH OAc, acetone, 25 ºC, 16 h; viii) corresponding alcohol or amine, Cu(OAc)
N, DMAP (optional), 4 Å MS, CH Cl , O (1 atm), rt, 2ꢀ12 h; ix) HCl/EtOAc (1.0, 2.0 or 4.0 M), rt, 1ꢀ
h; x) K CO , ethyl 2ꢀchloropyrimidineꢀ5ꢀcarboxylate, CH CN, 60 ºC, overnight; xi) LiOH·H O,
, NMM, DMF, rt,
2 3
(dba) , 1,4ꢀdioxaꢀ8ꢀazaspiro[4.5]decane,
2
2
4
4
2
,
Et
2
3
2
2
2
2
3
3
2
MeOH/THF/H
overnight.
2
O (3:1:3, 1:3:1, 3:3:2), rt, overnight; xii) EDC·HCl, HOBt, THPONH
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The synthetic route for ether 37 with a pyrimidyl group is outlined in Scheme 4.
Starting from boronic ester 25, alcohol 31 was obtained by oxidation. Then, ether 32
was prepared by the Mitsunobu reaction. Acidic removal of the BOC protecting group
led us to amine 33, which was coupled with ethyl 2ꢀchloropyrimidineꢀ5ꢀcarboxylate.
The resulting ester, 34, was converted into the carboxylic acid 35, which was finally
transformed in the desired hydroxamic acid 37 via the THPꢀprotected intermediate 36.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 4.
i
)
ii)
25
31
32
v)
iii)
iv)
33
34
vi)
iii)
35
36
37
Conditions: i) NaOH (aq, 4.0 M), H
2 2 2
O , H O, rt, overnight; ii) tertꢀbutyl 4ꢀhydroxypiperidineꢀ1ꢀ
carboxylate, PPh , DEAD, toluene, 110 ºC, 1 h; iii) HCl/1,4ꢀdioxane (4.0 M), rt, 1ꢀ2 h; iv) ethyl 2ꢀ
3
chloropyrimidineꢀ5ꢀcarboxylate, K
2 3 3 2 2
CO , CH CN, rt, 3 h; v) LiOH·H O, THF/MeOH/H O (10:1:3), rt,
overnight; vi) EDC·HCl, HOBt, THPONH , NMM, DMF, rt, overnight.
2
Table 2. SAR of heteroatomꢀbonded substituents as dual PDE5/HDAC inhibitors.
ACS Paragon Plus Environment

Page 19 of 160
Journal of Medicinal Chemistry
1
2
3
4
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
1'
Cpd R1
PDE5 HDAC1 HDAC2 HDAC3ꢀ
HDAC6
IC50 nM
A IC50 IC50
IC50
NCOR2
nM
nM
nM
IC nM
50
21a
44
7080
>20000
10400
2
1b
1c
22
20
5980
7810
>20000
17500
12700
1600
2
2
1d
1e
5
68
490
712
31
441
709
2
17
118
21f
10
65
25
166
43
584
30a
5860
>20000
1090
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
30b
24
12
603
2030
1060
4340
O
30c
12900
>20000
N
H
O H
N
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
30d
30e
40
74
9270
443
>20000
1850
10100
1900
30f
30g
37
11
23
10
2440
566
66
9830
2250
432
2330
196
373
Regarding PDE5 activity, all compounds in Table 2 retained potent activities in the low
nanomolar range (IC50 < 75 nM); however, the replacement of the sulfonamide group
tended to result in a slight decrease in the potency for PDE5 compared to 1, particularly
for compounds 21a, 30a, 30d and 30e, which exhibited > 1 log unit of decreased
potency.
As anticipated, diverse responses in HDAC activity were observed for the compounds
in Table 2. This differential HDAC inhibitory profile is largely attributable to the nature
of the linker groups bearing the hydroxamic acid, predictably lying deep in the
hydrophobic catalytic channel of the HDACs, rather than due to the influence of the
ACS Paragon Plus Environment

Page 21 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
heteroatom attached to the sildenafil core, closer to the rim surface according to the
proposed binding mode. Thus, differences of low significant in terms of inhibitory
activity against the three isoforms (HDAC1, HDAC2 and HDAC6) are observed when
comparing the secondary amines 21b (5980 nM, >20000 nM, 12700 nM) and 21e (118
nM, 712 nM, 709 nM) with their corresponding etherꢀlinked matched pairs 30f (2440
nM, 9830 nM, 2330 nM) and 37 (66 nM, 432 nM, 373 nM), although a preference for
the more liphophilic ethers can be acknowledged.
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
Concerning the linkers entering deep into the HDAC catalytic channel, the class I
HDAC potency trend towards pyrimidylhydroxamic acids was replicated in the case of
nitrogenꢀ and oxygenꢀbonded variants (21dꢀ21f, 30e, 37). These analogues were the
most potent compounds in Table 2 against HDAC1 and HDAC2 isoforms, with IC50
values close to or below 100 nM for HDAC1 and in the midꢀnanomolar range for
HDAC2 and HDAC6, with the exception of the spiroꢀlinked pyrimidylhydroxamic 30e,
which exhibited reduced potency against the three isoforms, likely due to
conformational constraints to achieve optimal chelation geometry. The impact of this
pyrimidyl group for HDAC activity is clearly recognized when replacing it (21d) with a
phenyl group (21c); derivative 21c demonstrated decreased potency against HDAC1,
HDAC2 and HDAC6 by more than 2, 1.5 and 0.5 log units, respectively. The good
potency of the pyrimidyl group could not be attributed to a plausible explicit hydrogenꢀ
bond contact between any nitrogen of the pyrimidine ring and HDAC residues in the
catalytic pocket. Additionally, the catalytic channel of the three isoforms is highly
conserved such that this class selectivity can be attributed to a particular residue. The
decreased pKa of the hydroxamic group (from 8.73 in 21c to 7.83 in 21d, as calculated
45
with Pipeline Pilot ) might play a role in the good class I potency of the pyrimidyl
4
6
group, although it does not definitively explain the selectivity profile.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Conversely, small linear alkyl (21a, 30a, 30b) and cycloalkyl (21b, 30f) derivatives
were weak micromolar HDAC inhibitors or even inactive against HDAC2 (Table 2),
although increasing the flexible chain length tended to improve HDAC1 and HDAC2
potency compared to shorter linkers (e.g., compare 30b with 30a). Conformationally
constrained spiros (30c, 30d) exhibited no improvement in HDAC potency. Among the
derivatives in Table 2, the optimal linker for HDAC6 was the small phenyl group of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
0g, which had an IC value of 196 nM. Replacement of the planar phenyl ring in 30g
50
by cyclohexyl in derivative 30f was detrimental for HDAC inhibitory activity,
particularly for HDAC6 (IC50 of 30f of 2330 nM). This fully agrees with previous
findings regarding the preference of HDAC6 isoforms for aromatic groups over alkyl
4
7
groups and inspired us to guide the rational design of HDAC6 selective inhibitors
(manuscript in preparation). The hydrophobic nature of the tunnel channel, flanked by
Phe150 in HDAC1 (Phe155 and Phe620 for HDAC2 and HDAC6, respectively) and
Phe205 in HDAC1 (Phe210 and Phe680 for HDAC2 and HDAC6, respectively) as well
as the different conformation of the hydroxamic acid group for optimal bidentate
coordination with Zn metal might explain the preference for phenyl over cyclohexyl
linkers. In summary, despite their reduced polar surface area, compounds in Table 2
showed no clear HDAC improvement in vitro compared to compounds in Table 1.
The next stage of our SAR exploration focused on carbonꢀlinked substituents at the 5’ꢀ
position of the phenyl ring (Table 3) covering both linear alkyl chains (48a, 48b) and
methyleneꢀhomologated rings (7, 48cꢀ48m, 52aꢀ52b) as well as more rigid derivatives
with carbonꢀbonded rings at the 5’ position (48nꢀ48o, 52cꢀ52f). Additionally, the
hydroxamic moiety was directly attached to the phenyl ring of sildenafil (42), resulting
in an inactive derivative against HDAC2 and HDAC6 and low affinity for HDAC1
ACS Paragon Plus Environment

Page 23 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(IC of 6420 nM, Table 3). According to our modelling studies, the ethoxyphenyl ring
50
and sildenafil core causes steric clashes that prevent optimal positioning of the ZBG
within the HDAC cavity. This compound, 42, was prepared from bromide 38 via
reaction with Pd(dppf)
2
Cl
2
3
and Et N in EtOH under a CO atmosphere, ester hydrolysis,
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
treatment with THPONH
2
and acidic removal of the THPꢀprotecting group (Scheme 5).
Scheme 5.
i)
ii)
iii)
8
38
39
iv)
v)
40
41
42
Conditions: i) Br
LiOH·H O, THF/MeOH/H
rt, overnight; v) HCl/EtOAc (4.0 M), rt, 1 h.
2
, AcOH, rt, overnight; ii) Pd(dppf)Cl
2
, Et
3
N, EtOH, CO (atm), 80 ºC, overnight; iii)
2
O (3:1:1), 40 ºC, overnight; iv) EDC·HCl, HOBt, THPONH , NMM, DMF,
2 2
Hydroxamic acids 48aꢀ48o and 7 were synthesized as illustrated in Scheme 6. In this
case, bromine 38 and iodine 22 were employed as starting materials to prepare esters
43aꢀ43m and 43oꢀ43p. Esters 43a and 43dꢀ43g were obtained from bromide 38 through
a Negishi reaction or Suzuki coupling. Intermediates 43b, 43hꢀ43k and 43p were
prepared from iodide 22 by different methods. The key intermediate 22 was also
transformed into aldehyde 45, and then ester 43c could be synthesized. Esters 43l, 43m
and 43o were also prepared from aldehyde 45, in this case via reductive amination,
aldolic condensation and cyclopropanation, respectively. Then, carboxylic acids 46aꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
46m and 46oꢀ46p were isolated as previously described after reaction with LiOH.
Conversely, carboxylic acid 46n was directly obtained from iodide 22 after reaction
with ethyl 2ꢀformylcyclopropanecarboxylate and reduction with Et
desired hydroxamic acids 48aꢀ48o and 7 were achieved through a THPꢀprotected
intermediate or by direct reaction with NH OH hydrochloride.
3
SiH. Finally, the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
Hydroxamic acids 48i1 and 48i2 (trans and cis isomers) were obtained directly after
preparative HPLC purification of the crude reaction mixture. Conversely, hydroxamic
acids 7a and 7b (cis and trans isomers too) could be isolated after supercritical fluid
chromatography (SFC), although the stereochemistry of these two pairs of isomers
could not be confirmed and was randomly assigned.
Scheme 6.
ACS Paragon Plus Environment

Page 25 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
i) or ii), iii) and iv)
or ii), iii) and v)
or ii), iii) and vi)
xix)
iii)
38
4
7a-mand o
xvii)
xx)
or xx) and xxi)
ii) or vii) and viii)
or ix) and x)
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
22
4
6a-p
43a-mand 43o-p
R = CH3 or CH CH3
2
4
8a-o and 7
xiii), iii) and vi) or
xiv) or xv) and xvi)
xi)
4
4, R = CN
5, R = CHO
xviii) and x)
e)
xii)
4
c)
d)
f)
A = a) CH2
g)
h)
i)
i1)
i2)
j)
b) CH CH
2
2
k)
l)
m)
n)
o)
7)
7a)
7b)
Conditions: i) bromoꢀ(2ꢀethoxyꢀ2ꢀoxoꢀethyl)zinc, Pd (dba) , xantphos, THF, 80 ºC, overnight; ii)
2
3
corresponding borane reagent, xantphos, Na
2
CO
3
, Pd
2
(dba)
3
2
, 1,4ꢀdioxane/H O (10:1, 6:1 or 5:1), reflux,
overnight; iii) HCl/EtOAc (0.2, 1.0, 2.0 or 4.0 M), 0ꢀ25 ºC, 1ꢀ3 h; iv) corresponding chloride, K CO ,
2
3
CH
3
CN, 40ꢀ100 ºC, overnight; v) CAN, ethyl propꢀ2ꢀenoate, DIEA, 80 ºC, overnight; vi) (4ꢀ
methoxycarbonylphenyl)boronic acid, Cu(OAc) , Et N, CH Cl , O (1 atm), rt, overnight; vii) ethyl
2
3
2
2
2
acrylate, POT, Et
THF, ꢀ70 ºC, 10 minutes, then ꢀ40 ºC, 1 h, then ethyl 2ꢀformylcyclopropanecarboxylate, rt, 15 h; x) TFA,
Et SiH, CH Cl , 0 ºC, then rt, 10 h; xi) Zn(CN) , Pd(PPh , DMF, 80 ºC, overnight; xii) DIBALꢀH,
3 2
N, DMF, 100 ºC, overnight; vii) Pd/C, H (1 atm), MeOH, rt, overnight; ix) nꢀBuLi,
3
2
2
2
3 4
)
CH Cl , 0 ºC, then rt, overnight; xiii) tertꢀbutyl piperazineꢀ1ꢀcarboxylate, Ti[OCH(CH ) ] , toluene, rt, 90
2
2
3 2 4
minutes, then NaBH(OAc)
3
3 2 2
, rt, overnight; xiv) corresponding amine, AcOH, NaBH(OAc) , CH Cl , rt,
overnight; xv) ethyl 2ꢀdiethoxyphosphorylacetate, NaH, THF, 0 ºC, 1 h, then 45, rt, overnight; xvi)
trimethyloxosulfonium iodide, NaH, DMSO, 40 ºC, 12 h; xvii) LiOH·H O, MeOH/THF/H O (1:3:1 or
:3:2), rt or 40 ºC, overnight; xviii) nꢀBuLi, THF, ꢀ70 ºC, then ꢀ40 ºC, 1 h, then tertꢀbutyl 3ꢀ
2
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
oxocyclobutanecarboxylate, rt, 15 h; xix) EDC·HCl, HOBt, THPONH
2
, NMM, DMF, rt, overnight; xx)
BOP, DIEA, NH OH·HCl, DMF, 80 ºC, overnight; xxi) SFC separation.
2
A similar synthetic route was used to prepare compounds 52aꢀ52f. As shown in Scheme
, ester functionality was conferred via the reaction of boronic ester 25 with the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
appropriated bromide, chloride or triflate and subsequent hydrogenation in the case of
ester 49f. Then, a threeꢀstep protocol (hydrolysis, reaction with THPONH and acidic
2
deprotection) led us to the desired hydroxamic acids 52aꢀ52f.
Scheme 7.
i) or i) and ii)
iii)
25
4
9a-f R = CH or CH CH
3
2
3
50a-f
iv)
v)
52a-f
51a-f
c)
d)
e)
f)
A = a)
b)
Conditions: i) corresponding bromide, chloride or triflate, Pd(PPh
3
)
4
, K
2
3 2
CO , 1,4ꢀdioxane/H O (5:2) or
1,4ꢀdioxane, 85 ºC, MW, 1 h, or conventional heating, 80 ºC, overnight; ii) Pd/C, H (1 atm), MeOH, rt, 1
2
2 2 2
h; iii) LiOH·H O, THF/MeOH/H O (3:3:2 or 3:1:1), rt, overnight; iv) EDC·HCl, HOBt, THPONH ,
NMM, DMF, rt, overnight; v) HCl/EtOAc (2.0 or 4.0 M) or HCl/1,4ꢀdioxane (4.0 M), rt, 1 h.
Table 3. SAR of carbonꢀlinked dual PDE5/HDAC inhibitors.
ACS Paragon Plus Environment

Page 27 of 160
Journal of Medicinal Chemistry
1
2
3
4
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
1
'
Cpd R1
PDE5 HDAC1 HDAC2 HDAC3ꢀ
HDAC6
IC50 nM
A IC50 IC50
IC50
NCOR2
IC50 nM
nM
nM
nM
42
5
6420
>20000
>20000
48a
48b
52a
52b
48c
19
>20000 >20000
>20000
2120
46
4810
2420
1530
>20000
13
>20000 >20000
2350
122
>20000
89
1790
7
14
379
48d
7
63
335
51
1250
4
8e
8f
57
121
57
1880
6430
395
4
>20000 >20000
>20000 >20000
>20000
>20000
48g
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 160
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
48h
48i1
48i2
48j
38
1530
672
>20000
>20000
>20000
13900
344
515
57
13
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
22
346
165
3510
416
7
60
34
310
225
490
729
322
279
239
91
7a
143
7b
45
39
326
1220
126
48k
2610
14100
1920
48l
207
303
7620
>20000
9640
48m
>20000 >20000
>10000
48n
17
20
554
1860
130
48o
8750
>20000
5370
ACS Paragon Plus Environment

Page 29 of 160
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
52c
70
6910
>20000
5130
52d
4
4
354
1870
79
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
52e
2360
>20000
861
52f
5
>20000 >20000
5570
As with previous attachment points and connections at the 5’ position, reported in Table
, short linear alkyl chains (48a, 48b) and homologated arylalkyl chains (52a, 52b)
2
resulted in weak micromolar (HDAC1 and HDAC6) and even inactive HDAC2
compounds. Strikingly, the methyleneꢀlinked cinnamic derivative 52b was far less
potent against HDAC6 (IC50 of 1790 nM) than the corresponding sulfonamideꢀlinked
derivatives 13d and 13g as shown in Table 1 (with HDAC6 IC50 values of 89 and 84
nM, respectively), likely as a result of the different positioning of the sildenafil capping
group on the surface area due to the different geometries of the connecting bonds.
Once again, the influence of the pyrimidylhydroxamic moiety on class I activity was
clearly recognized for compounds 48c and 48d (IC50 values of 14 and 63 nM for
HDAC1 and 89 and 335 nM for HDAC2, respectively). Regarding the nitrogenꢀlinked
pair (21d versus 21c), replacement of the pyrimidine ring (48d) by pyridine (48e) and
phenyl (48f) progressively decreased HDAC activity against the three isoforms, in
agreement with a trend towards increased basicity of the pKa of the hydroxamic acid
ACS Paragon Plus Environment