PTG-0861: A novel HDAC6-selective inhibitor as a therapeutic strategy in acute myeloid leukaemia

Article abstract of DOI:10.1016/j.ejmech.2020.112411

Dysregulated Histone Deacetylase (HDAC) activity across multiple human pathologies have highlighted this family of epigenetic enzymes as critical druggable targets, amenable to small molecule intervention. While efficacious, current approaches using non-selective HDAC inhibitors (HDACi) have been shown to cause a range of undesirable clinical toxicities. To circumvent this, recent efforts have focused on the design of highly selective HDACi as a novel therapeutic strategy. Beyond roles in regulating transcription, the unique HDAC6 (with two catalytic domains) regulates the deacetylation of α-tubulin; promoting growth factor-controlled cell motility, cell division, and metastatic hallmarks. Recent studies have linked aberrant HDAC6 function in various hematological cancers including acute myeloid leukaemia and multiple myeloma. Herein, we report the discovery, in vitro characterization, and biological evaluation of PTG-0861 (JG-265), a novel HDAC6-selective inhibitor with strong isozyme-selectivity (～36× ) and low nanomolar potency (IC₅₀ = 5.92 nM) against HDAC6. This selectivity profile was rationalized via in silico docking studies and also observed in cellulo through cellular target engagement. Moreover, PTG-0861 achieved relevant potency against several blood cancer cell lines (e.g. MV4-11, MM1S), whilst showing limited cytotoxicity against non-malignant cells (e.g. NHF, HUVEC) and CD-1 mice. In examining compound stability and cellular permeability, PTG-0861 revealed a promising in vitro pharmacokinetic (PK) profile. Altogether, in this study we identified a novel and potent HDAC6-selective inhibitor (～4× more selective than current clinical standards – citarinostat, ricolinostat), which achieves cellular target engagement, efficacy in hematological cancer cells with a promising safety profile and in vitro PK.

Full text of DOI:10.1016/j.ejmech.2020.112411

Journal Pre-proof
PTG-0861: A novel HDAC6-selective inhibitor as a therapeutic strategy in acute
myeloid leukaemia
Justyna M. Gawel, Andrew E. Shouksmith, Yasir S. Raouf, Nabanita Nawar, Krimo
Toutah, Shazreh Bukhari, Pimyupa Manaswiyoungkul, Olasunkanmi O. Olaoye,
Johan Israelian, Tudor B. Radu, Aaron D. Cabral, Diana Sina, Abootaleb Sedighi,
Elvin D. de Araujo, Patrick T. Gunning
PII:
S0223-5234(20)30382-2
DOI:
https://doi.org/10.1016/j.ejmech.2020.112411
EJMECH 112411
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 2 March 2020
Revised Date: 27 April 2020
Accepted Date: 27 April 2020
Please cite this article as: J.M. Gawel, A.E. Shouksmith, Y.S. Raouf, N. Nawar, K. Toutah, S. Bukhari,
P. Manaswiyoungkul, O.O. Olaoye, J. Israelian, T.B. Radu, A.D. Cabral, D. Sina, A. Sedighi, E.D.
de Araujo, P.T. Gunning, PTG-0861: A novel HDAC6-selective inhibitor as a therapeutic strategy
in acute myeloid leukaemia, European Journal of Medicinal Chemistry, https://doi.org/10.1016/
j.ejmech.2020.112411.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier Masson SAS.

Title:
PTG-0861: A Novel HDAC6-selective Inhibitor as a Therapeutic Strategy in Acute Myeloid Leukaemia
Author list
Justyna M. Gawel†,#, Andrew E. Shouksmith †,#, Yasir S. Raouf†,‡#, Nabanita Nawar†,‡#, Krimo Toutah†, Shazreh
Bukhari†,‡, Pimyupa Manaswiyoungkul†,‡, Olasunkanmi O. Olaoye†,‡, Johan Israelian†,‡, Tudor B. Radu†,‡, Aaron D.
Cabral†,‡, Diana Sina†,‡, Abootaleb Sedighi†, Elvin D. de Araujo†, Patrick T. Gunning†,‡*
#These authors contributed equally.
Author affiliations
†Department of Chemical and Physical Sciences, University of Toronto Mississauga, 3359 Mississauga Road,
Mississauga, Ontario, L5L 1C6, Canada
‡Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada
Corresponding author
Prof. Patrick T. Gunning (patrick.gunning@utoronto.ca)
Abstract
Dysregulated Histone Deacetylase (HDAC) activity across multiple human pathologies have highlighted this family of
epigenetic enzymes as critical druggable targets, amenable to small molecule intervention. While efficacious, current
approaches using non-selective HDAC inhibitors (HDACi) have been shown to cause a range of undesirable clinical
toxicities. To circumvent this, recent efforts have focused on the design of highly selective HDACi as a novel
therapeutic strategy. Beyond roles in regulating transcription, the unique HDAC6 (with two catalytic domains)
regulates the deacetylation of ꢀ-tubulin; promoting growth factor-controlled cell motility, cell division, and metastatic
hallmarks. Recent studies have linked aberrant HDAC6 function in various hematological cancers including acute
myeloid leukaemia and multiple myeloma. Herein, we report the discovery, in vitro characterization, and biological
evaluation of PTG-0861 (JG-265), a novel HDAC6-selective inhibitor with strong isozyme-selectivity (~36×) and low
nanomolar potency (IC50 = 6 nM) against HDAC6. This selectivity profile was rationalized via in silico docking studies
and also observed in cellulo through cellular target engagement. Moreover, PTG-0861 achieved relevant potency
against several blood cancer cell lines (e.g. MV4-11, MM1S), whilst showing limited cytotoxicity against non-
malignant cells (e.g. NHF, HUVEC) and CD-1 mice. Mechanistic studies into the activity of this HDACi revealed
possible secondary mechanisms behind the observed efficacy, including apoptosis induction and ROS generation. In
examining compound stability and cellular permeability, PTG-0861 revealed a promising in vitro pharmacokinetic (PK)
profile. Altogether, in this study we identified a novel and potent HDAC6-selective inhibitor (~4× more selective than
current clinical standards – citarinostat, ricolinostat), which achieves cellular target engagement, efficacy in
hematological cancer cells with a promising safety profile and in vitro PK.
1

Highlights
•
•
Selective HDAC6 (Class IIb) inhibition offers efficacy against hematological cancers
Novel HDAC6-selective inhibitor PTG-0861 displays single-digit nanomolar potency and high selectivity
(~36×)
•
•
•
Observed in vitro and cellular selectivity superior to phase II HDAC6-selective clinical candidate citarinostat
Relevant potency observed in several blood cancer cell lines with validated cellular target engagement
Promising in vitro pharmacokinetics achieved with good safety profile in cells and in vivo
Keywords
Histone deacetylase 6 (HDAC6), Acute Myeloid Leukaemia, Isoform selectivity, SAR
1. Introduction
Cellular function relies on a diverse set of tightly regulated biological processes, operating under the control of
several genomic mechanisms, be it genetic or epigenetic. Global abnormalities or dysregulated activity in either of
these operations is widely implicated in human pathologies including colorectal, T-cell lymphoma, and other cancers
[1]. Aberrant epigenetic modifications of nucleosomes affect the accessibility of genetic information to the
transcriptional machinery of the cell. These changes, like DNA methylation, play important roles in regulating patterns
of gene expression, which can ultimately drive the incidence and progression of oncogenic neoplasms [2]. Unlike the
genetic framework, epigenetic processes are generally reversible [3], providing the rationale for potential therapeutics
against disease-relevant epigenetic targets. Common epigenetic alterations involve the post-translational modification
(PTM) of proteins, such as histones. These include acetylation, methylation, sumoylation, phosphorylation, and
ubiquitylation [4].
Histone acetylation levels control chromatin structure and are governed by the opposing activity of histone
acetyltransferases (HAT) and histone deacetylases (HDAC). In histones, acetylation of Lys residues unravels DNA,
exposing genes to transcription factors. HDACs reverse this process, creating a condensed chromatin, and thus
regulate gene transcription. The proteome consists of 18 HDACs, with varying size, localization, and expression
levels. Zn2+-dependent HDACs are grouped into four classes; Class I (1, 2, 3, 8), Class IIa (4, 5, 7, 9), Class IIb (6,
10) and Class IV (11). (Fig. 1) [5]. Due to their role in gene expression, aberrant HDAC activity and overexpression
has been implicated in various human cancers, including acute myeloid leukaemia (AML) and multiple myeloma (MM)
[5–9].
Fig. 1. (Left) Superposition of HDAC3, 4, 6, 8 ribbon structures (PDB 4A69, 2VQQ, 5EDU, 5FCW respectively).
Strong alignment between HDACs of varying classes demonstrates the challenge of achieving isozyme selectivity
2

[10]. (Inset) HDAC-active sites (lysine-tunnel) also reveal general overlay of residues neighbouring the enzyme active
site.
With an ever-increasing knowledge of the individual role of HDACs in disease, various drug discovery efforts
were initiated to develop both pan- and selective-HDAC inhibitors (HDACi). As these molecules aim to out compete
N-acetyl-Lys for the enzyme active site, most HDACi are Lys-mimetics, sharing conserved structural characteristics
(Fig. 2). A Zn2+-binding group (ZBG) to coordinate the active site metal ion, an elongated-linker region to traverse the
Lys-tunnel and a surface cap group, anchoring to the exterior surface of the active site. While medicinal chemists
explore many variations of ZBGs, linkers, and surface cap groups, the general described scaffold is typically retained.
To date, four HDACi have received approval for treatment of hematological cancers. These include SAHA
(Vorinostat, 1) Belinostat (PXD-101, 2) Romidepsin (FK-228, 3) and Panobinostat (LBH-589, 4) [11–13]. These highly
potent and clinically efficacious drugs function as pan-HDACi, targeting multiple isozymes with comparable affinities.
However, as non-selective HDAC drugs, undesirable toxicities have since been reported in the clinic. These typically
include fatigue, diarrhea, haematological, gastrointestinal and bone-marrow toxicity [14,15].
As a result, more recent programs have focused on isozyme-specific HDACi in efforts to reduce patient toxicity.
While HDAC active sites share high sequence similarity (Fig. 1), several selective inhibitors have been successfully
reported [16–19]. Notable examples include, HDAC6-selective ricolinostat (ACY-1215, 5) [15] and citarinostat (ACY-
241, 6) [20], (Fig. 2) which are currently in clinical trials for the treatment of MM and metastatic breast cancer, as
single agents or in combination with standard-of-care NAB-paclitaxel [14]. Nonetheless, it is worth-noting that in vitro
selectivity of these drug candidates is reported as ~5-10× [20,21] in biochemical activity assays which may be
insufficient to elicit HDAC6-selective phenotypes in vivo. Other examples of recent HDAC6-selective inhibitors
reported in literature are Tubacin, carbazole-based Tubastatin A, and Bavarostat with an adamantyl cap-group [22].
3

Fig. 2. FDA-approved HDACi (1-4) and Clinical HDAC6-selective inhibitors (ricolinostat, 5 and citarinostat, 6).
The Class IIb HDAC6 (~134 kDa) is a unique isoform, containing two catalytic domains (CD). While both are
structurally accessible, the current understanding suggests HDAC6i achieve their observed potency by binding the
functional and disease-relevant CD2 [8]. Beyond its role in controlling transcription, a primary function of HDAC6 is to
deacetylate ꢀ-tubulin, facilitating microtubule formation, promoting growth factor-controlled cell motility and cell
division [23]. Importantly, over-expression of HDAC6 represses gene transcription for tumor suppressor proteins (e.g.
CYLD), causes global microtubule deacetylation and increased cell motility, leading to metastatic hallmarks [8,23,24].
Herein, we present the discovery and biological evaluation of a novel HDAC6-selective inhibitor, PTG-0861 (54,
JG-265) which exhibited promising biological activity in hematological cancers. This molecule was generated through
a focused SAR study around a recently published N-hydroxamic acid HDACi, AES-350 [25], (Fig. 3) with the key aim
of improving HDAC6 selectivity and drug-like properties. 54 was found to display ~36× selectivity for HDAC6 in a
biochemical activity assay relative to the nearest isozyme; much higher than the available current clinical
counterparts. Biochemical HDAC6 inhibition, cellular activity, and pharmacokinetic studies are presented and
compared to Phase II clinical candidate, citarinostat.
2.
Results and discussion
2.1
Chemical synthesis
Synthetic pathways used to prepare the desired compounds are outlined
in Schemes 1-4. The 4-amino-N-hydroxybenzamide inhibitors 38-47 were
synthesized in 5 steps (Scheme 1). Following carboxyl group protection via
benzylation or Fischer esterification to form the benzyl or ethyl ester,
respectively (7-8); the corresponding amide or sulfonamide was synthesized
using dichlorotriphenylphosphorane (60-95%) or (2-(1H-benzotriazol-1-yl)-
1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU) coupling (90%)
protocols, or via standard acylation/sulfonylation reactions (70-75%) (9-18).
After removal of the protecting group (19-28), the hydroxamate esters were
formed (29-37) before a final deprotection using H2 (g) Pd/C (80-97%) or acidic
conditions (70-80%) to furnish the corresponding final compounds.
Fig. 3 Chemical structure of previously
reported HDAC6i AES-350 along with
summary of advantages and
disadvantages.
4

Scheme 1. Reagents and conditions to synthesize 38-47. a) BnBr, Cs2CO3, DMF, RT, 24 h; b) EtOH/conc. H2SO4
i
(10:1), 90 °C, 16 h; c) R’CO2H, PPh3Cl2, CHCl3, 100 °C, MW, 1–2 h; d) R’CO2H, HBTU, Pr2NEt, DMF, RT, 6 h; e)
R’COCl or R’SO2Cl, CHCl3, 100 °C, microwave, 90 min; f) H2, 10% Pd/C, THF/MeOH (2:1), RT, 2-16 h; g) LiOH.H2O,
THF/MeOH/H2O (3:1:1), RT, 2-3 h; h) i) (COCl)2, THF, DMF, 0 °C, 1 h; ii) H2N-OBn, iPr2NEt, THF, RT, 16 h; i) H2N-
OR (R = THP, tBu), EDCI, HOBt, Et3N, DMF, RT, 16 – 24 h; j) 4M HCl/dioxane, 0 °C-RT, 3 h.
4-(aminomethyl)-N-hydroxybenzamide derivatives 54 and 62-63 were prepared in 5-7 steps as shown in Scheme 2.
Starting with 4-(aminomethyl)-benzoic acid, the synthesis of HDACi 54 (PTG-0861) (Scheme 2A) involved an initial
protection of the 1o amine using di-tert-butyl decarbonate (Boc2O), following by a benzyl group protection of the free
benzoic acid. Subsequent removal of the Boc group furnished the benzyl ester with the 4-(aminomethyl)
hydrochloride (50). This was followed by an acylation using 1,2,3,4,5-pentafluorobenzoyl chloride (60-70%) and
removal of the benzyl group before forming the hydroxamate ester (53). A final deprotection step via catalytic
hydrogenation generated the final compound as required. Structurally similar derivatives (62-63), as shown in
Scheme 2B, were prepared using an initial low-yielding esterification to form the ethyl ester (35-40%), before the
appropriate acylation/sulfonylation reaction (56-57). The route was completed in a similar manner to previously
mentioned inhibitors, where the revealed benzoic acid is converted to the hydroxamate ester and a final deprotection
used to generate the N-hydroxamic acid motif.
5

A.
B.
6

Scheme 2. Reagents and conditions to synthesize 54, 62, 63. (A) k) Boc2O, NaHCO3, THF/H2O (1:1), RT, 19 h; a)
BnBr, Cs2CO3, DMF, RT, 24 h; j) 4M HCl/dioxane, 0 °C-RT, 3 h; l) R’COCl or R’SO2Cl, iPr2NEt, CH2Cl2, 0 °C-RT, 24
h; f) H2, 10% Pd/C, THF/MeOH (2:1), RT, 2-16 h; h) i) (COCl)2, THF, DMF, 0 °C, 1 h; ii) H2N-OBn, iPr2NEt, THF, RT,
16 h. (B) b) EtOH/conc. H2SO4 (10:1), 90 °C, 16 h; g) LiOH.H2O, THF/MeOH/H2O (3:1:1), RT, 2-3 h.
Similarly, 4-(aminomethyl)-N-hydroxybenzamide derivatives with N-alkylations (i.e. cPr, iPr) (76-78) were
prepared in 6 steps (Scheme 3A). Starting with 4-formylbenzoic acid, following a benzyl group protection, the
aliphatic 2o amines were introduced via acid-mediated reductive amination using sodium triacetoxyborohydride
(NaBH(OAc)3) (75-85%) (65-66). As with the routes shown in previous schemes, subsequent acylation, benzyl group
deprotection, addition of the hydroxamate ester and final deprotection via hydrogenation conditions afforded the final
compounds.
Scheme 3B details the seven synthetic steps required to generate compounds (89-90) possessing an elongated
vinylic N-hydroxamic acid structure, similar to the approved HDACi’s, panobinostat and belinostat. First, a Wittig
olefination using tert-butylbromoacetate and 4-methylbenzaldehyde afforded the corresponding ester with the desired
E-alkene spacer (98%) (79). Radical-mediated bromination of the tolyl benzylic -CH3 with N-bromosuccinimide (NBS)
generated the desired benzyl bromide (80). This was subjected to nucleophilic substitution with the appropriate amine
or azide followed by a Staudinger reduction using triphenyl phosphine (PPh3) (81-82). The final steps of the synthesis
A.
were completed in an analogous manner to Scheme 3A to generate the desired final compounds.
7

B.
Scheme 3. Reagents and conditions to synthesize 76-78, 89, 90. (A) a) BnBr, Cs2CO3, DMF, RT, 24 h; m) i) R-NH2,
AcOH, DCE, RT, 2 h; ii) NaBH(OAc)3, RT, 16 h; l) R’COCl, iPr2NEt, CH2Cl2, 0 °C, 24 h; f) H2, 10% Pd/C, THF/MeOH
(2:1), RT, 16 h; h) i) (COCl)2, THF, DMF, 0 °C, 1 h; ii) H2N-OBn, iPr2NEt, THF, RT, 16 h; i) H2N-OR (R = THP, tBu),
t
EDCI, HOBt, Et3N, DMF, RT, 24 h; j) 4M HCl/dioxane, 0 °C-RT, 3 h. (B) n) [Ph3PCH2CO2 Bu]Br, DBU, THF, RT-60
°C, 20 h; o) NBS, AIBN, CCl4, 95 °C, 16 h; p) R-NH2, MeCN, RT, 16 h; q) i) NaN3, DMF, 50 °C, 2 h; ii) PPh3,
THF/H2O (10:1), RT, 20 h; r) TFA/CHCl3 (1:3), RT, 16 h.
The synthesis of 4-amino-N-hydroxybenzamide derivatives with a transposition of the amide or sulfonamide motifs is
listed in Scheme 4A, B (94, 96). In either route, 4-(tert-butyl)aniline is reacted with the required benzoic acid or
sulfonyl chloride starting material (30-85%). Using similar protocols in the schemes above the hydroxamate ester was
introduced before a final deprotection to furnish the N-hydroxamic acids. Inhibitor 101 with an N-ethylbenzamide was
synthesized in five steps (Scheme 4C) via a di-ethyl aniline intermediate (97). Following an amide coupling protocol
using PPh3Cl2 and ester deprotection the hydroxamate was installed and the inhibitor synthesized. Finally, tert-butyl
4-(bromomethyl)benzoate was required to synthesize inhibitor 106 starting via an SN2 type amine substitution
(Scheme 4D). As with the previous compounds, similar subsequent steps generated the desired HDACi.
8

A.
B.
C.
Scheme 4. Reagents and conditions to synthesize 94, 96, 101, 106. (A) c) R-NH2, PPh3Cl2, CHCl3, 100 °C, MW, 1–2
9
D.

i
h; g) LiOH.H2O, THF/MeOH/H2O (3:1:1), RT, 2-3 h; h) i) (COCl)2, THF, DMF, 0 °C, 1 h; ii) H2N-OBn, Pr2NEt, THF,
RT, 16 h; f) H2, 10% Pd/C, THF/MeOH (2:1), RT, 16 h. (B) s) R-NH2, DMAP, Pyridine, Benzene, RT, 24 h; (C) b)
EtOH/conc. H2SO4 (10:1), 90 °C, 16 h t) 10% Pd/C, NH4OAc, MeCN, MeOH, RT, 48 h (D) p) R-NH2, MeCN, RT, 16
h; l) R’COCl, iPr2NEt, CH2Cl2, 0 °C, 24 h; j) 4M HCl/dioxane, 0 °C-RT, 3 h; r) TFA/CHCl3 (1:3), RT, 16 h.
The aforementioned synthetic protocols afforded the following 22 final compounds: 38-47, 54, 62, 63, 76-78, 89, 90,
94, 96, 101, 106. The characterization data for all intermediates are detailed within the Supporting Materials file.
2.2 Structure-activity relationships, biochemical HDAC inhibition & cellular cytotoxicity
Our current investigation involved an SAR of a recently reported HDAC6-selective inhibitor, AES-350. As shown
in Fig. 3, this molecule exhibited strong target potency (IC50 [HDAC6, in vitro] = 24 nM), biological activity
(cytotoxicity) in hematological cell lines (IC50 [MV4-11, in cellulo] = 576 nM), and a promising cytotoxicity window in
healthy fibroblasts (MRC-9) (>50×). However, AES-350 showed only modest in vitro HDAC6 selectivity (8×, EMSA
activity assay), and was not exceptionally potent relative to current clinical leads in vitro against HDAC6 protein.
Fig. 4. Computational docking pose of AES-350 against zfHDAC6 (PDB 6CSR) using Schrödinger’s Maestro v11.9
[26–28]. Portions of the protein have been omitted for clarity. Key intermolecular interactions are also displayed.
The described SAR focused on the value of a direct N-hydroxamic acid aryl group (relative to elongated alkenyl
spacers), the role of the amide linkage relative to sulfonamides, as well as several N-alkylation strategies and the use
of benzylic -CH2- kink atoms for structural flexibility (improving cap group access to the protein surface). Recent
literature has validated the importance of appropriate cap groups in contributing to HDAC potency and selectivity [29].
As shown in Fig. 4, previous in silico docking studies of AES-350 against zfHDAC6 (PDB 6CSR) demonstrated the
importance of the phenylhydroxamate in forming a ꢁ-ꢁ sandwich interaction within the active site for selectivity (as
previously observed in reference [30]) and the free -NH- amide in H-bonding with available surface serine residues
10

(e.g. Ser531). At the same time, the lack of strong interactions by the tert-butyl-substituted phenyl ring B was also
evident. The resulting SAR study generated 22 final compounds, seen in Fig. 5 below.
11

Fig. 5. Chemical structures of HDACi 38-47, 54, 62, 63, 76-78, 89, 90, 94, 96, 101, 106.
12

Inhibitors were first screened against HDAC3, 6, 8, 11 as representative group members of the HDAC family in an
activity assay. Concurrently, due to the validated utility of HDACi in hematological cancers, the inhibitors were tested
for their cellular potency in MV4-11 acute myeloid leukaemia (AML) cell lines and ‘normal’ fibroblasts (MRC-9).
(Table 1) From compounds 38-45, which explore several cap group variations, it was evident that small hydrophobic
groups (-CH3, 38, -cPr 39) improve HDAC6 potency (15×, 9× respectively), while selectivity was attenuated.
Conversely, a bulky cyclohexyl cap group (40) retains a similar biochemical and cellular profile to AES-350.
Interestingly, an increase in HDAC6 selectivity was achieved through installation of a para electron-deficient
heterocycle (i.e. morpholino-, 41) or an electron-deficient (B) ring (-CF3, 42). In both cases, a ~2-fold increase in
selectivity was achieved in vitro for HDAC6 over the nearest isozyme (HDAC8). Also, altering the position of the para-
tBu group (43) of AES-350 to the meta site afforded selectivity for HDAC8 (2-fold). This trend was also observed upon
N-ethylation as seen by compound 101 (3.5-fold). Furthermore, substituting the amide linker with a sulfonamide (46),
improving structural flexibility (sp2 vs. sp3 centers) and introducing an additional H-bond acceptor, yielded ~10×
greater potency for HDAC6 (IC50 = 2.01 nM). However, the selectivity profile of analogs with a benzylic -CH2- kink
atom (47, 62, 63) was either comparable to the parent or reduced. Structural flexibility likely affords greater access to
the active site of multiple HDAC isoforms. Transposition of the amide and sulfonamide linkages in compounds 94 and
96, respectively, resulted in improvements in potency but with reduced selectivity profiles. From these studies, it is
evident that the introduction of structural flexibility (such as a benzylic -CH2- kink atom) improved potency, while
HDAC6 selectivity is significantly improved by introducing electron-deficient cap groups (i.e. pentafluorobenzene).
Table 1. Biochemical HDAC inhibition against recombinant enzymes via EMSA (Nanosyn) and cytotoxicity results in MV4-
11 (Leukaemia) and MRC-9 (Healthy fibroblasts). Note: (a) n=2, (b) n = 3, (c) n > 3, *HDAC6 selectivity not observed.
13

These observations guided the design of analog, 54, (PTG-0861). This compound with a polyfluorinated cap
group exhibited single-digit nanomolar potency against HDAC6 (4× increase as compared to AES-350), with ~36×
selectivity. Impressively, this selectivity for HDAC6 represents a ~6× improvement relative to current HDAC6
selective drug candidates, ricolinostat and citarinostat. In AML cell-based studies, PTG-0861 exhibited analogous
cellular potency to citarinostat, with an IC50 in MV4-11 cells of 1.85 ± 0.72 ꢂM and a favourable cytotoxicity window
(>27×) against non-cancerous fibroblasts. By comparison, citarinostat displayed a cytotoxicity window of (>20×).
To further explore this chemotype, several N-alkylation strategies were utilized including synthesizing -CH3 (106),
i
-iPr (76) and -cPr (78) analogs. In all cases, N-alkylation improved selectivity, with analog 76, possessing an Pr
group, showing 63× selectivity. However, potency and biological activity were negatively affected (MV4-11 IC50 = 16.4
± 1.73 ꢂM, a ~30× loss in cellular activity). It was hypothesized that the aliphatic substitution reduced cellular
permeability. Elongation of the N-hydroxamic acid via alkenyl spacers, akin to belinostat and panobinostat, was
investigated in compounds 89 and 90. Although, potency against MV4-11 was modestly improved (2-5×), HDAC6-
selectivity was attenuated or completely lost. This is consistent with the profile of pan-HDAC inhibitors that possess
an elongated structure which likely accesses the active site Lys tunnel of several HDACs. Conversely, the aryl phenyl
hydroxamate has been cited to afford HDAC6 selectivity due to its tunnel-fitting structure and favourable π-π
sandwich interactions with neighbouring Phe residues. As a result of PTG-0861’s selectivity and potency profile, it
was selected for further biological, computational, and pharmacological studies.
Fig. 6 Dose-response curve detailing the percent inhibition and visible HDAC6-selectivity of 54 (PTG-0861) against
HDAC1-11 via EMSA (Nanosyn)
Table 2. Biochemical HDAC inhibition IC50 values of ricolinostat, citarinostat, quisinostat and 54 (PTG-0861) against
HDAC1-11 using EMSA (Nanosyn).
14

To determine the full HDAC selectivity profile of this scaffold, PTG-0861 was screened against all Zn2+-dependant
HDACs (1-11) using an activity-based EMSA assay (Nanosyn, USA) (Fig. 6 and Table 2). The known pan-HDACi
quisinostat (JNJ-26481585), currently in phase II (for cutaneous T-cell lymphoma, CTCL) was analyzed as a positive
control as well as clinical HDAC6-selective inhibitors, ricolinostat and citarinostat. Quisinostat exhibited similar activity
as reported in literature (low nanomolar potency against several HDACs), while PTG-0861 showed promising HDAC6
selectivity with only low-nM potency shown against HDAC6, with nearest activity against HDAC3 (IC50 = 215 nM)
(Table 2). The in vitro selectivity data for PTG-0861 in the enzymatic activity assay compared favourably to
ricolinostat and citarinostat, which also inhibited HDACs 2, 3, and 8. Recent reports have shown that acetyl-lysine
based enzymatic assays can be unreliable for assessing HDAC10 inhibition.[31] The presence of contaminating
deacetylases, such as Class I HDACs within the multi protein complex, may affect the activity observed. [32,33]
Therefore, PTG-0861’s HDAC10 selectivity will require assessment by an orthogonal assay if progressed further.
2.3 In vitro HDAC Inhibition by Fluorescence Polarization
In conjunction with the enzymatic activity assay, a fluorescence polarization binding assay was performed (Fig. 7)
as an orthogonal in vitro tool to study HDACi binding profiles to recombinant zfHDAC6. In this assay, PTG-861, with
an IC50 = 0.074 ꢂM, was 3.6× more potent a binder than AES-350 (IC50 = 0.269 ꢂM). PTG-0861 was found to have
comparable potency to ricolinostat (IC50 = 0.090 ꢂM), and ~2× more potent than citarinostat (IC50 = 0.154 ꢂM)
Fig. 7 (Above) Dose-response curves of AES-350, PTG-0861, ricolinostat (5) and citarinostat (6) against recombinant
zfHDAC6 using fluorescence polarization. (Below) IC50 data of relevant HDACi against zfHDAC6.
15

2.4 In silico modelling & docking studies
To understand the potency and selectivity profiles experimentally determined, in silico docking experiments using
Schrödinger’s Maestro v11. (Schrödinger, LLC, New York, NY 2019) modelled PTG-0861 with hsHDAC6 (PDB
5EDU, catalytic domain 2) [26–28]. hsHDAC8 (PDB 1T64) was also evaluated as a negative control, given the
relatively lower activity observed against this isozyme (Fig. 8). Against HDAC6, PTG-0861 optimally occupies the
catalytic Lys tunnel, effectively coordinating the active site Zn2+ at the hydroxamate motif. PTG-0861 forms ꢁ-ꢁ
stacking interactions between the central aryl ring (A) and Phe620, Phe680, as well as the H-bond with the Ser568
hydroxyl. As was originally hypothesized, inclusion of the benzylic -CH2- kink atom likely affords flexibility to facilitate
ꢁ-ꢁ stacking interactions between His500 (∆GB = -9.127 kcal mol-1) and the electron-deficient pentafluorobenzene
(PFB) cap group. To the best of our knowledge, the use of electron-deficient cap groups to achieve HDAC6 potency
has not been widely reported. The binding-pose of PTG-0861 in hsHDAC6 was compared tert-butylbenzene
derivative (62), which lacks the PFB cap group within the same active site (Fig. S1B).
From Fig. 8 and Fig. S1B, it is evident the protein surface of HDAC6 neighbouring the active site contains several
aromatic systems (e.g. His500, Tyr570, His611, Phe620, His651, Phe680) with potential to interact with the PFB cap
group. As for 62, the compound appears to maintain the H-bond with the Ser568 hydroxyl. However, this prevents
optimal positioning of the bulky-cap group to achieve favourable interactions without causing undesired steric clashes
with the protein surface. In the next binding pose (Fig. S1B), whilst the benzene cap group can interact with adjacent
pi-systems, the HDACi loses the H-bond with Ser568. Therefore, to conclude, 54 with the PFB cap group appears
electronically and sterically more optimized to achieve both of these interactions on the surface of HDAC6.
In the active site of HDAC8, the surface ꢁ-ꢁ stacking interaction is maintained (Phe208) but both the Ser H-bond with
the free amide -NH- and internal Phe ꢁ-ꢁ stacking sandwich are lost (∆GB = -6.119 kcal mol-1). HDAC8 contains a
Cys, Asp, Tyr in the region where Ser568 is typically location in HDAC6. Also, it is worth noting the varying topology
of both enzymes; while HDAC6 contains a relatively flat terrain outside the active site, HDAC8 is characterized by
large bulky side-chains at the tunnel entrance (e.g. Tyr100), and a more dynamic/irregular region generally (from
surveying available x-ray crystal structures). In Fig. 8B, PTG-0861, appears to avoid a steric clash with Tyr100 by
virtue of the flexibility provided by the -CH2- kink atom. However, while this represents a static image, undesirable
clashes may still occur and potentially contribute to the observed decrease in potency against HDAC8.
16

Fig. 8. (A) In silico docking binding pose of PTG-0861 in human HDAC6 (PDB 5EDU), Zn2+ as a yellow-sphere,
hydrogen-bonds (yellow-dashed line), ꢁ-ꢁ stacking (green-dashed line), (N, blue; O, red; H, white; C, magenta; F,
green). Catalytic triad residues and portions of the protein have been omitted for clarity. (B) In silico docking pose of
PTG-0861 in human HDAC8 (PDB 1T64). (C) Ligand interaction diagram in HDAC6 displaying key intermolecular
interactions within the active site. (D) Ligand interaction diagram in the active site of HDAC8.
2.5 In cellulo target engagement
17

PTG-0861 and citarinostat (6), the orally administered front-running HDAC6-selective clinical candidate, [20] were
assessed for comparative effects on the acetylation of ꢀ-tubulin (a key HDAC6 substrate), histone H3 and histone H4
A.
B.
(HDAC class I substrates) following 6 h incubation in a target indication cell line, MV4-11[29].
Fig. 9. (A) Western blot illustrating ꢀ-tubulin acetylation, histone H3 and histone H4 acetylation levels in MV4-11 AML
cells following 6 h treatment with varying concentrations of PTG-0861 and clinical standard citarinostat (6). Protein
extracts were prepared and subjected to SDS-PAGE and immunoblotting with acetylated ꢀ-tubulin, acetylated histone
H3 and H4, and HSC70 antibodies. A representative western blot of three independent experiments is shown. (B)
Enzyme-linked immunosorbent assay (ELISA) inhibition profile of increasing concentrations of PTG-0861 on HDAC6
(IC50 = 0.59 ꢂM). Error bars represent the standard error from three independent trials.
PTG-0861 showed promising potency and selective stimulation of the accumulation of acetylated ꢀ-tubulin over
acetylation of histone H3 and H4, at sub 500 nM concentrations (Fig. 9 (A)). Citarinostat (6) showed similar levels of
activity and selectivity. However, citarinostat displayed greater off-target effects in cellulo, as can be observed from
the noted upregulation of acetylated histone H3 at the highest concentration of inhibitor (5 μM) relative to PTG-0861.
An enzyme-linked immunosorbent (ELISA)-based HDAC activity assay reliant on mammalian cell-derived HDAC
proteins was used to verify the findings from the activity and binding assays that utilized recombinant proteins.
Literature has highlighted noticeable inconsistencies in the biochemical data obtained from baculovirus versus
mammalian cell–derived HDAC isoforms [34,35]. One notable example is apicidin, which showed remarkable potency
against mammalian cell-derived HDAC1 (IC50 = 23 nM) but poor inhibition in the baculovirus-expressed counterpart
(IC50 > 1000 nM) [32,35]. These conspicuous discrepancies in the inhibitory profiles may have rooted from the source
of the protein. Thus, the ELISA-based activity assay which employs HeLa cell derived HDAC6 was employed to
characterize the lead compound PTG-0861. A range of PTG-0861 concentrations were incubated with active
mammalian HDAC6 coated plates, and the subsequent deacetylation activity was monitored over time (Fig. 9 (B)). In
only 0.25 h of drug dosage, PTG-0861 showed inhibitory activity with an IC50 of 0.59 ꢂM, comparable to the in cellulo
Western-blot findings.
18

A.
B.
Fig. 10. (A) Immunofluorescence analysis of ꢀ-tubulin acetylation and histone H3 acetylation following 6 h treatment
of PTG-0861 (54) in HeLa cells. (acetylated ꢀ-tubulin in red, acetylated histone H3 in green, and nuclear stain DAPI
in blue). This figure is representative of three independent experiments. Data for clinical candidate citarinostat (6) is
available in Supporting Info (Fig. S2A). (B) Quantification of fluorescent signals that correspond to levels of
acetylated α-tubulin and acetylated histone H3 in HeLa cells dosed with PTG-0861 at indicated concentrations.
Western blot data obtained from treated MV4-11 cells was recapitulated via immunofluorescent experiments using
Bright green Alexa Fluor 488 for acetylated histone H3 and Bright far-red Alexa Fluor 647 for acetylated α-tubulin.
PTG-0861 was shown to elevate levels of acetylated ꢀ-tubulin at 0.1 ꢂM, without any observable increase in
acetylated histone H3, confirming the strong in cellulo selectivity profile (Fig. 10).
2.6 Apoptosis induced in a time- and dose-dependent manner
The extent of cell death upon treatment with PTG-0861 was analyzed by flow cytometry through Annexin V/PI
staining. Loss of asymmetry of phosphatidylserine (PS) in the cell membrane leaflets is an early marker of
programmed cell death, which can be readily detected through PS-binding protein Annexin V. This indicator of early
apoptosis as well as the DNA intercalator, propidium iodide (PI), collectively behave as robust markers of the cell
death process [36,37]. At 4.0 ꢂM treatment of PTG-0861, 41% of MV4-11 cells underwent late apoptosis, as can be
seen from their positive staining with PI and Annexin V (Annexin V+/PI+). This is a 5ꢂ increase in late apoptotic cell
population compared to the ~8% observed in the vehicle control. At 4.0 M of citarinostat (6), only 17% of cells
underwent late apoptosis (Fig. S2B, S2C). Similar trends were observed in early apoptotic populations (Annexin
V+/PI-), with a 6-fold increase (3 to 18%) when comparing the DMSO control to PTG-0861 (4.0 ꢂM) treated cells (Fig.
11).
19

A.
Fig. 11. (A) Flow cytometric data of MV4-11 cells treated with increasing concentrations of PTG-0861 for 18 h and
analyzed for apoptosis by Annexin V/PI staining. Representative dot blots from three independent experiments is
presented. (B) Percentage of cell populations in healthy, early apoptosis, and late apoptosis phases following
treatment with PTG-0861 for 18 h at indicated concentrations.
2.7 PTG-0861 observes cytotoxic effect in multiple myeloma cell lines
20

In addition to MV4-11, the biological effects of PTG-0861 were further explored in two MM cell lines, MM1S and
RPMI 8226. Panobinostat (4), Romidepsin (3) and SAHA (1) have all demonstrated clinical efficacy in MM as well as
HDAC6-selective Phase 2 clinical candidate, citarinostat (5) [38],[39]. PTG-0861 showed analogous low μM potency
to citarinostat in both MM1S and RPMI 8226 cell lines.
Table 3. Cytotoxic effect of PTG-0861 in multiple myeloma cell lines obtained using a Cell Titre Blue assay.
2.8 PTG-0861 reveals a promising safety profile in cellulo and in vivo
Selective targeting of disease-associated HDAC isoforms may result in improved safety profiles. For example,
HDAC6 knock-out mice exhibit a normal phenotype, in contrast to genetic ablation of Class I HDACs (HDAC1,
HDAC2 and HDAC3), which impact survival [40]. PTG-0861 was next compared to citarinostat in normal human
fibroblasts, normal human umbilical vein endothelial cells (HUVEC) and primary normal human pooled fibroblasts
from 250 donors. PTG-0861 was found to exhibit a marginally better safety window, with an approximately 2× greater
safety profile. The in cellulo safety profile was upheld in vivo, as PTG-0861 showed no observable weight loss or
visible signs of toxicity in 5 CD1 mice treated daily for 5 days at 20 mg/kg (P.O.) (Fig 12).
Table 4. Cytotoxicity of PTG-0861 and citarinostat (6) in healthy non-cancerous cell lines (Normal Human
Fibroblasts, Normal Human Umbilical Vein Endothelial Cells HUVEC, Primary Normal Human Pooled Fibroblasts
from 250 donors).
Fig. 12. Average mouse weight of CD1 mice (20 mg/kg, P.O.) dosed with PTG-0861 and Vehicle daily for 5 days.
21

2.9 In vitro ADME and in vivo pharmacokinetic profiling of PTG-0861
A range of in vitro PK properties were investigated (Table 5). To assess the stability of the PFB ring to potential
nucleophilic attack, PTG-0861 was evaluated for stability against L-glutathione via an LC-MS/MS analysis that
monitors time-dependent covalent modification of inhibitor (Section 4.14). No discernible reaction was observed after
24 h. In mouse hepatocytes, PTG-0861 was found to have an acceptable half-life of 50.85 ± 3.37 min, comparable to
HDACi in clinic, with an intrinsic clearance rate of 27.32 ± 1.81 ꢂL/min per 106 cells (Fig. S3) [41]. The initial
compound AES-350, was ~2× less stable in the same experiment with faster clearance (T1/2 = 28.3 min, CLint = 49.1
ꢂL/min per 106 cells).
*Mean values from three independent experiments. Data for replicate is available in Supporting information
(Table S1).
Table 5. In vitro pharmacokinetic assessment of PTG-0861.
In a PAMPA permeability assay, PTG-0861 was determined to have a -Log Pe of 5.66 (Pharmaron) (Table S1). In a
Caco-2 assay (Pharmaron) to predict human intestinal permeability, PTG-0861 was calculated to have a Papp (A-B) of
1.33 × 10-6 cm/s and Papp (B-A) of 0.94 × 10-6 cm/s and an efflux ratio of 0.71, suggesting some promise for oral
bioavailability (Table S1). Conversely, clinical standard citarinostat showed a Papp (A-B) of 0.69 × 10-6 cm/s and Papp (B-A)
of 31.51 X 10-6 cm/s with a significantly larger efflux ratio of 45.64 (Table S1). Next, we assessed the in vivo
pharmacokinetics of PTG-0861 compared to citarinostat in CD-1 male mice (single dose of 20 mg/kg I.P.) (Fig. S4).
PTG-0861 was shown to have a half-life of 0.25 h, Cmax of 526 ng/mL with AUClast and AUCinf values of 190 and 219
22

h*ng/mL respectively. Despite a promising in vitro ADME, this in vivo PK profile in a murine model might be the result
of insufficient absorption and/or rapid first-pass metabolism.
3. Conclusions
A series of HDAC6-selective inhibitors were prepared via a focused SAR investigation of a class I/IIb-selective
HDACi, AES-350. Derivatization was chiefly aimed at achieving a balance between HDAC6-targeting potency,
selectivity, and concomitant metabolic stability.
HDACi, PTG-0861 (JG-265), with
a
polyfluorinated
phenylhydroxamate scaffold was identified, possessing single-digit nanomolar potency against HDAC6, and ~36×
selectivity for HDAC6 over the next targeted HDAC isoform. Molecular modelling and in silico docking studies
suggest ꢁ-ꢁ stacking interactions at the phenyl core (Phe620, Phe680) and PFB group (His500) in addition to a H-
bond at the free amide linker (Ser531) contributed to the overall selectivity and potency of PTG-0861. The in vitro
activity was successfully translated to a cellular system, with increases in acylated ꢀ-tubulin as compared to acylated
Histone H3 and H4 being observed in several cell lines via Western blotting, ELISA, and immunofluorescence
experiments.
While biological effects were observed in hematological cell lines (MV4-11, MM.1S, and RPMI 8226), minimal
effects were observed in non-malignant counterparts (MRC9, NHF, HUVEC, pooled primary NHF), as well as no
identifiable toxicities observed in vivo during a 5-day toxicity trial. PTG-0861 treated AML cells revealed dose-
dependent increases in early and late apoptosis populations. While PTG-0861 exhibits low µM IC50’s in transformed
cells, analogous to ricolinostat, it might be concluded that a strong selectivity for one single HDAC, HDAC6, might
actually attenuate cellular cytotoxicity. While the stronger cellular potency exhibited by pan-HDACi is likely due to
their poly-pharmacology, the caveat is that they also exhibit a lower therapeutic window in normal cells. As with most
drug discovery programs, HDACi programs have to achieve an optimal balance between potency and selectivity.
Finally, in vitro evaluation of inhibitor PK revealed promising metabolic stability and permeability. In preliminary PK
studies, PTG-0861 was found to have a lower than predicted profile based on the in vitro PK data. While the low
toxicity profile of PTG-0861 may allow for higher dosing regimes to achieve biologically relevant doses in vivo, the
current effort is to improve the PK before moving into preclinical in vivo models.
23

4. Experimental section
4.1. Chemistry
4.1.1. General
Chemicals and solvents were purchased from Sigma-Aldrich (MilliporeSigma, Canada), VWR International, Alfa
Aesar, Combi-Blocks, Caledon Laboratory Chemicals and Promega. Reactions were performed under nitrogen,
unless indicated otherwise, using anhydrous-grade solvents, and were monitored for completeness by thin-layer
chromatography (TLC) using Merck silica gel 60F254 on aluminium sheets (visualized by 254/365 nm UV light and/or
staining with KMnO4). Preparative HPLC was conducted using a Waters 2487 Dual λ Absorbance Detector, equipped
1
with a Symmetry® C18 4.6 mm × 150 mm cartridge. 400 MHz Bruker NMR was utilized to obtain H, 13C, and 19F
NMR spectra in deuterated solvents (Cambridge Isotope Laboratories Inc. or Sigma-Aldrich). Chemical shifts (δ) are
reported in parts per million (ppm) after calibration to residual isotopic solvent and coupling constants (J) are reported
in Hertz (Hz). Multiplicities are described as singlet (s), doublet (d), triplet (t), quartet (q), pentet (p), sextet (sex),
septet (sep), multiplet (m), broad (br) or a combination of these. Low-resolution mass spectrometry (LRMS) was
carried out using a Waters LC-MS in ESI mode, fitted with a Micromass ZQ MS and an Alliance 2690 LC. High-
resolution mass spectrometry (HRMS) was carried out using an Agilent 6538 UHD Q-TOF MS in ESI mode with a
mass accuracy +/- 1 mDa. Inhibitor purity was evaluated by a Hewlett Packard Series 1100 analytical HPLC system
fitted with a Phenomenex Luna 5.0 µm C18 4.6 mm × 150 mm cartridge and ≥ 95% purity was obtained prior to
biological testing.
4.1.2 Synthetic procedure (a): Carboxyl benzylation
The appropriate benzoic acid (1.0 equiv.) and cesium carbonate (1.1 – 1.2 equiv.) were suspended in DMF (0.33 –
0.5 M) and stirred at RT for 20 min in air, before addition of benzyl bromide (1.0 equiv.) in one go. After 24 h, the
reaction was concentrated in vacuo at 80 °C and partitioned between EtOAc and distilled water. The layers were
separated, and the aqueous layer was extracted with EtOAc. The combined organic layer was dried (MgSO4), filtered
and concentrated in vacuo. Column chromatography isolated the target compound.
4.1.3. Synthetic procedure (b): Carboxyl ethylation
The appropriate benzoic acid (1.0 equiv.) was dissolved in ethanol before mixing with concentrated sulfuric acid
(10:1) (final conc. 0.3 M) dropwise, at RT in air, and stirred at 90 °C. After 16 h, the reaction was cooled to RT and
partitioned with EtOAc and 0.5 M NaOH. The layers were separated, and the organic layer was washed with 0.5 M
NaOH. The aqueous layer was extracted with EtOAc and the combined organic layer was dried (MgSO4), filtered and
concentrated in vacuo to isolate the target compound without further purification.
4.1.4. Synthetic procedure (c): Amide coupling using carboxylic acids I
24

The appropriate carboxylic acid (1.2 equiv.) and dichlorotriphenylphosphorane (1.5 – 2.4 equiv.) were suspended in
chloroform (0.1 – 0.2 M) and stirred for 15 min at RT before addition of the appropriate aniline (1.0 equiv.), and the
mixture was irradiated at 100 °C for 1 – 2 h. Removal of the solvent in vacuo followed by column chromatography
isolated the target compound.
4.1.5. Synthetic procedure (d): Amide coupling using carboxylic acids II
Diisopropylethylamine (6.0 equiv.) was charged in one go to a solution of the appropriate carboxylic acid (2.0 equiv.)
and HBTU (5.0 equiv.) in DMF (0.1 M) at RT. After 15 min, the appropriate aniline (1.0 equiv.) was added in one go,
and the reaction was stirred for 6 h before quenching with saturated aqueous sodium bicarbonate. The layers were
separated, the aqueous layer was extracted with EtOAc, and the organic layer was washed with saturated aqueous
sodium bicarbonate followed by brine. The organic layer was dried (Na2SO4), filtered and concentrated in vacuo, and
column chromatography isolated the target compound.
4.1.6. Synthetic procedure (e): Amide coupling using acid chlorides/sulfonyl chlorides I
The acid chloride or sulfonyl chloride (1.0 – 1.1 equiv.) was added in one go to a solution of the appropriate aniline
(1.0 – 1.05 equiv.) in chloroform (0.1 – 0.5 M) and irradiated in microwave initiator at 100 °C for 90 min. Removal of
the solvent in vacuo followed by column chromatography isolated the target compound.
4.1.7. Synthetic procedure (f): Palladium on Carbon Hydrogenation
To a nitrogen-purged solution of the benzyl or hydroxamate ester (1.0 equiv.) in THF and methanol or ethanol (2:1,
0.05 – 0.1 M) was charged 10% Pd/C (0.04 equiv.). The mixture was purged with hydrogen for 2 h before filtration
through celite, washing with EtOAc, and concentrated in vacuo. Carboxylic acids were isolated without further
purification. Preparative HPLC was used to isolate hydroxamic acids.
4.1.8. Synthetic procedure (g): Ester hydrolysis
Lithium hydroxide monohydrate or lithium hydroxide (3.0 equiv.) was added in one go to a solution of the alkyl ester
(1.0 equiv.) in THF, methanol and distilled water (3:1:1, 0.1 – 0.2 M) and stirred at RT in air. After 2.5 h, the reaction
was diluted with EtOAc and quenched using 1 M HCl. The layers were separated, the organic layer was washed with
1 M HCl and the combined aqueous layer was extracted with EtOAc. The organic layer was dried (MgSO4), filtered
and concentrated in vacuo to isolate the target compound without further purification.
4.1.9. Synthetic procedure (h): Formation of hydroxamate esters I
Oxalyl chloride (5.0 equiv.) was added dropwise to a solution of the appropriate benzoic acid (1.0 equiv.) in THF (0.05
– 0.2 M) and DMF (2 drops) at 0 °C and stirred for 50 min – 3 h. The reaction was concentrated to dryness in vacuo
before re-dissolving in dry THF (0.2 M) and mixing with diisopropylethylamine or triethylamine (3.0 – 4.0 equiv.)
followed by the O-protected hydroxylamine (1.0 – 2.0 equiv.). After 16 h, the reaction was quenched with 1 M HCl
and the layers were separated. The organic layer was washed with 1 M HCl and the combined aqueous layer was
extracted with EtOAc. The organic layer was dried (MgSO4), filtered and concentrated in vacuo, and column
chromatography isolated the target compound.
4.1.10. Synthetic procedure (i): Formation of hydroxamate esters II
25

Triethylamine (5.0 equiv.) was charged in one go to a solution of the appropriate carboxylic acid (1.0 equiv.), 1-(3-
dimethylaminopropyl)-3-ethylcarbodiimide (EDCI, 1.5 equiv.) and 1-hydroxybenzotriazole monohydrate (HOBt, 1.1
equiv.) in DMF (0.1 – 0.2 M) at RT in air. After 10 min, the O-protected hydroxylamine (2.0 equiv.) was added in one
go, and the reaction was stirred for 16 – 24 h before quenching with 0.1 M HCl. The layers were separated, the
organic layer was washed with 0.1 M HCl and the aqueous layer was extracted with EtOAc. The organic layer was
dried (MgSO4), filtered and concentrated in vacuo, and column chromatography isolated the target compound.
4.1.11. Synthetic procedure (j): Acid-mediated hydrolysis of carboxylic or hydroxamate esters I
The carboxylic, hydroxamate ester or carbamate (1.0 equiv.) was charged with ice-cooled 4M HCl in dioxane (0.05
M) and stirred at RT in air. After 18 h, the solvent was removed in vacuo, azeotroping with CH2Cl2. Carboxylic acids
and amines were isolated without further purification. Preparative HPLC was used to isolate hydroxamic acids as final
compounds.
4.1.12. Synthetic procedure (k): Tert-butyloxycarbonyl (Boc) protection
Di-tert-butyl dicarbonate (2.0 equiv.) in THF (0.63 M) was added in one go to a solution of the appropriate amine
hydrochloride (1.0 equiv.) in THF and distilled water (1:1). Sodium bicarbonate (3.0 equiv.) in distilled water (0.95 M)
was added to the reaction mixture in one go and stirred at RT in air. After 19 h, the reaction was quenched by
addition of 1 M HCl until pH 3 and the layers were separated. The aqueous layer was extracted with EtOAc and the
combined organic layer was dried (MgSO4), filtered and concentrated in vacuo to isolate the target compound without
further purification.
4.1.13. Synthetic procedure (l): Amide coupling using acid chlorides/sulfonyl chlorides II
The acid or sulfonyl chloride (1.0 – 1.1 equiv.) was added in one go to an ice-cooled solution of the appropriate amine
or amine hydrochloride (1.0 – 1.1 equiv.) and diisopropylethylamine, triethylamine or pyridine (2.0 – 3.0 equiv.) in
CH2Cl2 or chloroform (0.1 – 0.2 M) and stirred at RT. After 24 h, the reaction was quenched with 1 M HCl. The layers
were separated, and the aqueous layer was extracted with CH2Cl2. The organic layer was dried (MgSO4), filtered and
concentrated in vacuo, and column chromatography isolated the target compound.
4.1.14. Synthetic procedure (m): Reductive amination
The appropriate alkyl amine (1.1 equiv.) was added in one go to a solution of the appropriate aldehyde (1.0 equiv.) in
1,2-DCE (0.1 M) and stirred at RT for 10 min before addition of glacial acetic acid (2.0 equiv.) and further stirring for
30 min. Sodium triacetoxyborohydride (2.0 equiv.) was added in one go and the reaction was stirred for 24 h before
pouring over saturated aqueous sodium bicarbonate. The layers were separated, and the aqueous layer was
extracted with CH2Cl2. The combined organic layer was washed with saturated aqueous sodium bicarbonate,
followed by brine, dried (MgSO4), filtered and concentrated in vacuo. Column chromatography isolated the target
compound.
4.1.15. Synthetic procedure (n): Wittig olefination
Tert-butyl bromoacetate (1.0 equiv.) in toluene (1.0 M) was added in one go to a solution of triphenylphosphine (1.1
equiv.) in toluene (1.1 M), whereupon it was stirred at RT in air. After 24 h, the solid was filtered under atmospheric
pressure, washing with diethyl ether, and dried in vacuo to give the desired Wittig reagent.
26

The Wittig reagent (1.0 equiv.) was suspended in THF (0.5 M) and charged with DBU (2.0 equiv.) at RT for 5 min
before addition of the appropriate aldehyde (1.1 equiv.) whereupon it was stirred at 60 °C in air. After 20 h, the
reaction was cooled to RT, diluted with EtOAc (0.25 M) and quenched with 0.1 M HCl. The layers were separated,
and the organic layer was washed with 0.1 M HCl. The aqueous layer was extracted with EtOAc, and the combined
organic layer was dried (MgSO4), filtered and concentrated in vacuo. Column chromatography isolated the target
compound.
4.1.16. Synthetic procedure (o): Radical bromination
The appropriate tolyl derivative (1.0 equiv.), N-bromosuccinimide (1.1 equiv.) and 2,2’-azobis(2-methylpropionitrile)
(AIBN, 0.02 equiv.) were dissolved at RT in carbon tetrachloride (0.3 M) and stirred at 95 °C. After 19 h, the reaction
was cooled to RT and filtered at atmospheric pressure, washing with carbon tetrachloride. The filtrate was partitioned
with distilled water and the aqueous layer was extracted with CH2Cl2. The organic layer was dried (MgSO4), filtered
and concentrated in vacuo, and column chromatography isolated the target compound.
4.1.17. Synthetic procedure (p): SN2 substitution
The alkyl bromide (1.0 equiv.) was added in one go to a solution of the alkyl amine (4.0 equiv.) in acetonitrile (0.1 M),
and additional alkyl amine (4.0 equiv.) was added after 1 h. After 16 h, the solvent was removed in vacuo and the
product was partitioned with EtOAc or CH2Cl2 and saturated aqueous sodium bicarbonate or 1M NaOH. The organic
layer was washed with saturated aqueous sodium bicarbonate/1M NaOH and the aqueous layer was extracted with
EtOAc/CH2Cl2. The combined organic layer was dried (MgSO4), filtered and concentrated in vacuo to isolate the
target compound without further purification.
4.1.18. Synthetic procedure (q): Azide SN2 and Staudinger reduction
Sodium azide (2.2 equiv.) was added to a solution of the alkyl bromide (1.0 equiv.) in DMF (0.25 M) at RT and stirred
at 50 °C in air. After 2 h, the reaction was cooled to RT, diluted with EtOAc and washed with distilled water, followed
by brine. The organic layer was dried (MgSO4), filtered and concentrated in vacuo to give the desired azide product
without further purification.
The azide (1.0 equiv.) was dissolved in THF and distilled water (10:1) (0.19 M), before addition of triphenylphosphine
(1.3 equiv.), and stirred at RT in air. After 20 h, the solvent was removed in vacuo and the product was dissolved in
diethyl ether and hexanes (1:13) (0.04 M). Any solid was removed by filtration under atmospheric pressure and this
process was repeated until no further solid precipitated. The filtrate was concentrated in vacuo and mixed with
hexanes, followed by 1.0 M HCl in diethyl ether (19:1) (0.02 M). The solid was isolated by filtration under atmospheric
pressure, washing with diethyl ether, and dried in vacuo to isolate the target compound.
4.1.19. Synthetic procedure (r): Acid-mediated hydrolysis of carboxylic or hydroxamate esters II
The carboxylic or hydroxamate ester (1.0 equiv.) was dissolved in chloroform and charged with trifluoroacetic acid
(3:1) (final conc. 0.1 M) at RT in air. After 16 h, the solvent was removed in vacuo, azeotroping with CH2Cl2.
Carboxylic acids were isolated without further purification. Preparative HPLC was used to isolate hydroxamic acids.
4.1.20. Synthetic procedure (s): Amide coupling using acid chlorides/sulfonyl chlorides III
27

The appropriate aniline (1.0 equiv.) was dissolved in pyridine (0.1 M) and anhydrous benzene (0.3 M) before adding
the sulfonyl chloride starting material (1.1 equiv.) and dimethylaminopyridine (DMAP) in catalytic amounts (0.1
equiv.). The resulting solution was left to stir overnight at room temperature. The reaction mixture was extracted with
chloroform and 2.0 M HCl(aq.). The organic layer washed with brine, dried (Na2SO4) before concentrating under
reduced pressure. The resulting crude was either purified via column chromatography or re-crystallized to provide the
desired products.
4.1.21. Synthetic procedure (t): Aniline mono-ethylation
The appropriate aniline (1.0 equiv.), 10% Pd/C (0.02 equiv.) and ammonium acetate (1.0 equiv.) were dissolved in
methanol (0.5 M) under nitrogen. Acetonitrile (5.0 equiv.) was added in one go and the flask was purged with
hydrogen at RT. After 48 h, the reaction was filtered through celite and concentrated in vacuo, before partitioning
between EtOAc and distilled water. The layers were separated, the organic layer was washed with distilled water and
the aqueous layer was extracted with EtOAc. The combined organic layer was dried (MgSO4), filtered and
concentrated in vacuo, and column chromatography isolated the target compound.
4.1.22. Synthetic procedure (u): Aryl carboxylation
The appropriate aryl bromide (1.0 equiv.) was cooled to -78 °C in THF (0.1 M) before addition dropwise of 2.5 M n-
butyllithium in hexanes (1.1 equiv.). After 1 h, dry ice (CO2) was bubbled through the solution via cannula for 3 h
before quenching with 0.1 M HCl. The layers were separated, and the organic layer was washed with 0.1 M HCl and
brine. The combined aqueous layer was extracted with EtOAc and the organic layer was dried (MgSO4), filtered and
concentrated in vacuo.
4.1.23. Synthetic procedure (v): Suzuki-Miyaura coupling
The appropriate aryl bromide (1.0 equiv.), boronic acid (1.5 equiv.), palladium (II) acetate (0.05 equiv.),
tricyclohexylphosphine (0.1 equiv.) and potassium phosphate tribasic (2.0 equiv.) were dissolved in toluene and
distilled water (19:1) (final conc. 0.3 M) and purged with nitrogen for 30 min before stirring at 90 °C. After 16 h, the
reaction was cooled to RT and partitioned with EtOAc and 2:1 solution of distilled water and brine. The organic layer
was washed with 2:1 distilled water and brine, and the aqueous layer was extracted with EtOAc. The combined
organic layer was dried (MgSO4), filtered and concentrated in vacuo, and column chromatography isolated the target
compound.
4.1.24. N-hydroxy-4-(4-methylbenzamido)benzamide (38)
The product was obtained using synthetic procedure (f) as a white-solid, yield 10%. 1H δ/ppm (400 MHz, DMSO-d6)
2.39 (s, 3H, CH3), 7.35 (d, J = 7.9 Hz, 2H, 2 CH), 7.75 (d, J = 8.7 Hz, 2H, 2 CH), 7.84 – 7.89 (m, 4H, 4 CH), 10.37 (br
s, 1H, NH); 13C δ/ppm (100 MHz, DMSO-d6) 21.0, 119.5, 127.4, 127.5, 127.7, 127.7, 128.9, 129.0, 131.8, 141.8,
165.5; LRMS (ESI+) m/z calcd for [C15H14N2O3Na]+: 293.09, found: 293.16; HRMS (ESI+) m/z calcd for
[C15H15N2O3]+: 271.1071, found: 271.1077; HPLC (I) tR = 12.23 min (99.2%); HPLC (II) tR = 18.20 min (98.9%).
4.1.25. 4-cyclopropyl-N-(4-(hydroxycarbamoyl)phenyl)benzamide (39)
1
The product was obtained using synthetic procedure (j) as a white-solid, yield 14%. H δ/ppm (400 MHz, DMSO-d6)
0.75 – 0.79 (m, 2H, 2 CH), 1.01 – 1.06 (m, 2H, 2 CH), 1.98 – 2.04 (m, 1H, CH), 7.22 (d, J = 8.4 Hz, 2H, 2 CH), 7.74
28