Identification and Optimization of Mechanism-Based Fluoroallylamine Inhibitors of Lysyl Oxidase-like 2/3

Article abstract of DOI:10.1021/acs.jmedchem.9b01283

Lysyl oxidase-like 2 (LOXL2) is a secreted enzyme that catalyzes the formation of cross-links in extracellular matrix proteins, namely, collagen and elastin, and is indicated in fibrotic diseases. Herein, we report the identification and subsequent optimi

Full text of DOI:10.1021/acs.jmedchem.9b01283

Subscriber access provided by Caltech Library
Article
Identification and Optimization of Mechanism-Based
Fluoroallylamine Inhibitors of Lysyl Oxidase-Like 2/3.
Alison Findlay, Jonathan S. Foot, Alberto Buson, Mandar Deodhar, Andrew Jarnicki, Philip
Hansbro, Gang Liu, Heidi Schilter, Craig Turner, Wenbin Zhou, and Wolfgang Jarolimek
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01283 • Publication Date (Web): 03 Oct 2019
Downloaded from pubs.acs.org on October 7, 2019
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 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 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.
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 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Identification and Optimization of Mechanism-Based
Fluoroallylamine Inhibitors of Lysyl Oxidase-Like
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
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
/3.
†
*
†*
†
†
‡┴
Alison D. Findlay, Jonathan S. Foot, Alberto Buson, Mandar Deodhar, Andrew G. Jarnicki,
Philip M. Hansbro,‡#^ Gang Liu,#^ Heidi Schilter, Craig I. Turner, Wenbin Zhou and Wolfgang
†
†
†
Jarolimek†
†
Pharmaxis Ltd, 20 Rodborough Road, Frenchs Forest, Sydney, NSW 2086, Australia
‡
Centre for Healthy Lungs, The University of Newcastle and Hunter Medical Research Institute,
Newcastle, NSW 2300, Australia
#
Centre for Inflammation, Centenary Institute, Sydney, NSW 2050, Australia
^
University of Technology Sydney, Faculty of Science, Ultimo, NSW 2007, Australia
KEYWORDS Copper-dependent amine oxidase, lysyl oxidase, lysyl oxidase-like 2, lysyl
oxidase-like 3, mechanism based enzyme inhibitors, collagen, elastin, fibrosis.
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 59
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
ABSTRACT
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
Lysyl oxidase-like 2 (LOXL2) is a secreted enzyme that catalyzes the formation of cross-links in
extracellular matrix proteins, namely collagen and elastin, and is indicated in fibrotic diseases.
Herein we report the identification and subsequent optimization of a series of indole-based
fluoroallylamine inhibitors of LOXL2. The result of this medicinal chemistry campaign is PXS-
5
120A (12k), a potent, irreversible inhibitor that is >300-fold selective for LOXL2 over LOX.
PXS-5120A also shows potent inhibition of LOXL3, an emerging therapeutic target for lung
fibrosis. Key to the development of this compound was the utilization of a compound oxidation
assay. PXS-5120A was optimized to show negligible substrate activity in vitro for related amine
oxidase family members, leading to metabolic stability. PXS-5120A, in a pro-drug form (PXS-
5
129A, 12o), displayed anti-fibrotic activity in models of liver and lung fibrosis, thus confirming
LOXL2 as an important target in diseases where collagen cross-linking is implicated.
INTRODUCTION
In most organisms, including humans, two families of amine oxidases are responsible for the
metabolism of various mono-, di- and polyamines produced endogenously or absorbed from
exogenous sources. The first of these families use covalently bound flavin adenine dinucleotide
(FAD) as the cofactor, and includes the monoamine oxidases (MAOs; MAO-A and MAO-B),
involved in the metabolism of neuroactive amines such as dopamine, serotonin and
1
norepinephrine; polyamine oxidase (PAO), which oxidatively deaminates spermine and
ACS Paragon Plus Environment
2

Page 3 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
spermidine; and lysine-specific histone demethylases, such as LSD1 (or KDM1A), that are
involved in epigenetic function.3
The second family of amines oxidases are dependent on copper and utilize an oxidized tyrosine
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
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
4
residue as the cofactor. Within this family there are two sub-classes. The first one consists of the
Amine Oxidase, Copper-containing (AOC, utilizing the topaquinone (TPQ) cofactor) sub-class
5
encompassing diamine oxidase (DAO; AOC1), involved in the metabolism of histamine; retinal
6
specific amine oxidase (RAO; AOC2) found predominantly in the eye; semicarbazide-sensitive
amine oxidase (SSAO, also known as vascular adhesion protein-1 (VAP-1); AOC3), an
7
ectoenzyme involved in leukocyte recruitment; and plasma amine oxidase (AOC4) which is found
8
in all mammals except rats, mice and humans (who lack a functional AOC4 gene). The second
sub-class consists of the lysyl oxidase family (utilizing the lysine tyrosylquinone (LTQ) cofactor).
There are 5 members of the human lysyl oxidase protein family, designated lysyl oxidase (LOX)
9
and lysyl oxidase-like 1-4 (LOXL1-4). The expression pattern of LOX family members can alter
10
depending on cell type, differentiation state, development state and disease state. The LOX
family act as key regulators in the cross-linking of elastin and collagen fibrils, through the
oxidation and condensation of the -amino group of specific lysine and hydroxylysine residues.11
A fine balance of elasticity and mechanical stiffness is required to maintain appropriate function
in various tissues such as the lung and heart, and this balance is dependent on LOX/LOXL
1
2
activity. Up-regulation of one or more of the LOX family enzymes can lead to aberrant cross-
linking, causing excessive local collagen deposition and propagation of pro-fibrotic signaling
within the extracellular matrix (ECM). If left unchecked, this can induce tissue scarring, fibrosis
and destruction of function, eventually leading to organ failure.13
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 59
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
Under normal physiological conditions, LOX is ubiquitously expressed in most human tissues at
1
0
higher levels than other family members, and is essential in the development of healthy bone,
respiratory system and cardiovascular function.14,15,16 In contrast, LOXL2 mRNA and protein
levels are upregulated in fibrotic tissues and diverse solid tumors when compared with normal
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
1
7
human tissues. LOXL2 is also considered a biomarker for progression in various diseases, such
as liver and lung fibrosis, scleroderma and cancer.18,19,20,21 Given the important role that LOXL2
appears to play in certain fibrotic diseases, selective inhibition of LOXL2 over LOX offers clear
advantages for their long-term treatment.
LOXL2 is significantly increased in human cholangiopathies, and is proposed to contribute to bile
2
2
duct injury through suppression of E-cadherin in biliary epithelial cells. Ikenaga et al. have
shown that LOXL2 mediates collagen crosslinking and fibrotic matrix stabilization during liver
fibrosis, and independently promotes fibrogenic hepatic progenitor cell (HPC) differentiation.23
2
4
Moreover, using the murine LOXL2 antibody AB0023, the authors have shown that inhibition
of LOXL2 can promote a reversal of advanced parenchymal liver fibrosis in mice. How critical
the role of LOXL2 and its enzymatic activity is in hepatic fibrosis compared to the other members
2
5
of the LOX family is yet to be fully proven, with evidence that in some instances LOX can also
act as biomarker for liver fibrosis.26
LOXL2 is also of interest as a therapeutic target in lung fibrosis, including idiopathic pulmonary
2
7
28
fibrosis (IPF). LOXL2 is over-expressed in patients with IPF and is reported to activate lung
fibroblasts via regulation of the TGF‑β/Smad signaling pathway.29
A humanized monoclonal LOXL2 antibody AB0024 (Simtuzumab) had been advanced into
clinical trials, including two fibrosis indications (IPF and NASH).30,31 Whilst Simtuzumab was
ACS Paragon Plus Environment
4

Page 5 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
found to be safe and well tolerated, there was no sign of clinical benefit. Although this result has
thrown doubt on the viability of LOXL2 as a target, it is important to note that the antibody only
partially (~40%) inhibits the catalytic activity of the enzyme,17a,24 and is therefore unlikely to have
a significant effect on collagen cross-linking. Additionally, the absence of a pharmacodynamic
assay for LOXL2 inhibition meant that confirmation of target inhibition with these dosing
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
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
0
regimens was not confirmed, nor was penetration into the organ of interest. To this end, a small
molecule with appropriate drug-like qualities offers an opportunity to fully evaluate the therapeutic
potential of LOXL2 inhibition.
Scarce examples of small molecule lysyl oxidase inhibitors exist, with -aminopropionitrile
32
(
BAPN) (Figure 1) long serving as the archetypal pan-LOX inhibitor. More recently, Springer
3
3
et. al. disclosed an aminomethylthiophene-based inhibitor CCT365626 (1) with sub-micromolar
potency for both the LOX and LOXL2 isoforms, however no selectivity data was provided for the
other family members. PharmAkea have reported the development of a series of aminomethyl
34
pyridine-based, selective LOXL2 inhibitors, of which PAT-1251 (2) completed Phase I trials at
3
5
the end of 2016. Although enzyme inhibition in humans seems to have been achieved with 2, it
appears short lived (complete recovery 24 h after a single dose of 1000 mg per healthy volunteer
36
was reported), and the therapeutic implications of such a profile are at present unclear.
NH HCl
2.
O O
S
O O
S
F3C
NH2
O
S
N
F
N
N
NH2
O
NH2
N
OH
N
N
O
F
S
O
BAPN
CCT365623 (1)
PAT-1251 (2)
PXS-5153A (3)
Figure 1. Chemical structures of BAPN, CCT365623 (1), PAT-1251 (2) and PXS-5153A (3)
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 59
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
We have recently reported the structure of PXS-5153A (3), and outlined its efficacy in two models
of liver fibrosis and a model of myocardial infarction.17b PXS-5153A was developed in tandem to
the chemical series reported herein with the purpose of producing a tool compound to establish
proof-of-concept for in vivo inhibition of LOXL2 in models of fibrosis. PXS-5153A was found to
have good potency for LOXL2 (pIC = 7.7, see also Table 3) and reasonable selectivity over LOX
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
0
(~70-fold, see also Table 3). Importantly PXS-5153A was found to possess a pharmacokinetic
profile that allowed for low dosing of the molecule in in vivo models. Unfortunately PXS-5153A
is recognized as a substrate by both SSAO and plasma amine oxidase (data shown in Table 3), and
that precluded the molecule from being developed further. The present work outlines the
identification and optimization of an indole-based structural class (closely related to PXS-5153A)
and strategies to improve on the potency, selectivity and, most importantly, potential for oxidation
by SSAO and plasma amine oxidase, the latter to enable pre-clinical toxicological evaluation of
this class of LOXL2 inhibitors in higher order mammals (dog, pig, monkey, etc.).
These studies led to the identification of 12k (PXS-5120A), a potent LOXL2/3 inhibitor which
demonstrated efficacy in models of both liver and lung fibrosis, in addition to the previously
reported benefits in a diabetic nephropathy model.37
INHIBITOR DESIGN
Common to the quinone-containing AOC family of enzymes is the ability to recognize and oxidize
short length arylalkyl-based primary amines (histamine, dopamine, tyramine, benzylamine). With
SSAO/VAP-1, this has been exploited using a variety of 3, 4 and 5-atom length arylsubstituted
fluoroallylamine groups (e.g. mofegiline, PXS-4159A, LJP 1586, PXS-4681A)38-41 These
fluoroallylamines are recognized as substrates and are therefore capable of binding to the quinone
ACS Paragon Plus Environment
6

Page 7 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
cofactor directly. In the case of SSAO, it is thought that this leads to the activation of the previously
unreactive vinyl fluoride moiety towards nucleophilic attack, thereby enabling formation of a
further covalent bond within the active site and irreversibly inhibiting enzyme function (Scheme
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
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
).42
1
Scheme 1. Proposed mechanism of inhibition of SSAO/VAP-1 by fluoroallylamine-based
compoundsa
H
H N
R
2
O
R
H
N
H
N
H
N
H
N
O
O
O
O
+
H O
2
O
O
N
H
N
H
N
H
N
H
O
O
HO
HO
-
H O
2
H
R
H H
R
R
O
N
N
NH2
OH
OH
OH
OH
R
F
F
Nu Enzyme
+
++
H O + NH
Cu /Cu / O / H O
2
2
3
2
2
H
H
N
N
O
O
N
H
N
H
HO
HO
H
H
-
F-
R
R
N
N
H
OH
OH
R
Nu Enzyme
R
F
Nu Enzyme
a
Scheme shows the proposed mechanism of inhibition of SSAO by fluoroallylamine-based inhibitors. Black arrows
portray the normal catalytic turnover of amine substrates, the red arrows show how the installation of a vinyl fluoride
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 59
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
into the molecule can redirect the pathway, resulting in the covalent alkylation of the protein active site. “Nu”:
nucleophilic amino acid residue in the active site.
Given the success of this approach in selectively targeting the SSAO TPQ cofactor, we sought to
design a novel series of inhibitors of LOXL2 based on a similar tactic, targeting the LTQ cofactor.
As no detailed structural information regarding the target was available at the outset of our
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
4
3
synthetic campaign we screened several fluoroallylamine templates to identify a tractable starting
point and extend the utility of these scaffolds. We found that a previously reported
4
4
fluoroallylamine 4, displayed modest LOXL2 inhibition, albeit with no selectivity over LOX
(Figure 2) nor SSAO. Further exploration of the fluoroallylamine motif afforded the (Z)-isomer,
5
, a compound with improved LOXL2 potency and reduced SSAO activity. This fluoroallylamine
configuration was deemed to be a suitable starting point for further investigation.
F
(Z)-configuration
preferred over (E)
F
O
O
NH HCl
2
.
H
N
H
N
S
NH HCl
S
2
.
O O
O O
4
5
LOXL2 pIC50 5.6 + 0.3
LOX pIC50 5.7 + 0.1
SSAO pIC50 7.3 + 0.1
LOXL2 pIC50 6.5 + 0.3
LOX pIC50 6.4+ 0.3
SSAO pIC50 <5.5
Figure 2. Identification of a suitable fluoroallylamine configuration.
RESULTS AND DISCUSSION
With a view to identifying the features necessary for boosting LOXL2 potency and opening a
selectivity window over LOX, we opted to replace the central linking phenol moiety appended to
the fluoroallylamine in 5 with a 1,2,3-triazole and expand on the substitution of the terminal phenyl
4
5
ring. This enabled rapid exploration via Cu(I)-catalyzed Huisgen cycloaddition reaction between
ACS Paragon Plus Environment
8

Page 9 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
appropriately functionalized alkynes 6a-h and azide 9, derived from the previously reported
4
4
alcohol 7, affording a suite of 1,2,3-triazole-based inhibitors 10a-h (Scheme 2 and Table 1).
While recombinant human LOXL2 was used for functional assays, bovine aorta LOX was utilized
for selectivity screening the analogues, as it was found that this was pharmacologically similar to
human fibroblast LOX and was more readily available.
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
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 2. Synthesis of 1,2,3-triazole-based inhibitors 10a-h
F
7
HO
Br
NHBoc
b
c
NH HCl
2.
F
8
NHBoc
F
N
N
N
10a-h
HO
O
F
a
d, e
+
O
N3
NHBoc
R
R
Commercially
available phenols, a-h
6a-h
9
R
o
Reagents and conditions: (a) propargyl bromide, K
2
CO
3
, MeCN, rt; (b) MsCl, Et
3
N, acetone, then LiBr, 0 C-rt; (c)
O / EtOAc, rt.
NaN , DMF, rt; (d) CuSO (H O) , sodium ascorbate, t-BuOH, rt; (e) (i) TFA, DCM, rt; (ii) HCl, Et
3
4
2
5
2
Table 1. LOXL2 and LOX inhibition by 1,2,3 triazole analogues 10a-h
LOXL2a,b,c
LOXa,b,c
Compound
BAPN
R group
(
pIC ± SD) (pIC ± SD)
5
0
50
n/a
H
6.4 ± 0.12
6.5 ± 0.08
6.1 ± 0.05
6.7 ± 0.3
5.7 ± 0.13
6.3 ± 0.2
6.3 ± 0.2
6.2 ± 0.1
1
0a
10b
0c
4-tBu
1
4-SO Me
2
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 59
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
10d
10e
2-SO Me
5.7 ± 0.07
6.1 ± 0.05
6.1 ± 0.1
7.3 ± 0.1
7.7 ± 0.09
5.5 ± 0.2
6.0 ± 0.05
5.5 ± 0.4
6.1 ± 0.2
6.1 ± 0.2
2
3-SO Me
2
1
0f
0g
0h
4-CONMe2
4-SO NMe
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
1
2
2
2
-Cl, 4-
1
SO NMe
2
2
a
The measurement of enzymatic activities was based on the detection of hydrogen peroxide with an Amplex-Red
oxidation assay ; Minimum of n=2 for each experiment; “LOXL2” = recombinant human LOXL2 and “LOX” =
bovine aorta LOX; SD = standard deviation
4
6
b
c
Introduction of simple alkyl groups such as a t-butyl group (10b) was found to have a slightly
detrimental effect on LOXL2 potency relative to the unsubstituted example, 10a, while
introduction of the electron withdrawing methyl sulfone marginally improved potency (10c),
importantly favoring LOXL2 over LOX (Table 1). In terms of the optimal position for substitution,
4
-SO Me was preferred over 2-SO Me (10d) and 3-SO Me (10e). While replacement of the para-
2 2 2
methyl sulfone with 4-CONMe (10f) showed no benefit, a significant improvement in LOXL2
2
potency was achieved upon introduction of the dimethylsulfonamide moiety, as in 10g. The
introduction of a chloro substituent in the ortho-position in 10h resulted in a further increase in
LOXL2 potency. Evaluation of 10g and 10h for off-target amine oxidase activity showed that
although neither inhibited SSAO (pIC <4.5) both were inhibitors of DAO (pIC of 6.5 for each),
5
0
50
and 10g (but not 10h) also inhibited MAO-B (pIC 6.9).
5
0
Given the presence of the primary amine, we next evaluated the most potent analogues to establish
if these compounds were recognized as substrates.
ACS Paragon Plus Environment
1
0

Page 11 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Amine oxidase substrate activity may effectively act as a clearance pathway, reducing available
inhibitor levels. Furthermore, it has the potential to produce oxidative stress and undesirable
species in vivo, owing to the fact that hydrogen peroxide and the corresponding aldehyde species
are by-products of the enzyme reaction. In addition, plasma amine oxidase substrate activity can
complicate pre-clinical development, due to its differential presence between species (aside from
rodents and humans, the AOC4 gene product is present in all mammalian species, especially the
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
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
8
dog). Thus, rapid “turnover” by the related amine oxidases was potentially a major limitation and
recognized as an important development consideration. To monitor, and thereby enable
optimization of this, compound oxidation assays to measure the relative substrate activity by
SSAO and plasma amine oxidase were introduced into the screening cascade.
Both compounds were rapidly turned over in the SSAO and plasma amine oxidase oxidation
assays, and when incubated in dog plasma, 10h was found to be completely consumed, with 0%
parent compound remaining after 1 hour. Pre-incubation of the dog plasma with semicarbazide (to
inhibit SSAO and plasma amine oxidase) confirmed that this degradation was dependent on the
activity of these related amine oxidases (data not shown).
In efforts to improve these aspects, we opted to explore the effect of both linker and core
modifications (Figure 3). Due to the observed high turnover by SSAO of the triazole-linked
compounds, we chose to first evaluate this parameter and identify an alternative chemo-type that
would not be recognized as a substrate by SSAO.
ACS Paragon Plus Environment
1
1

Journal of Medicinal Chemistry
Page 12 of 59
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
Core variation:
NH HCl
2.
N
N
N
1,2,3-triazole
pyrrole
N
F
N
X
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
X
N
Linker variation
Y
2
-methylindole
R
Figure 3. Variation of linker and core regions.
Table 2. Structure-activity relationships of linker and core-based modifications
LOXL2a,b
(pIC50 ± SD)
LOX a,b
Linker,
Y
oxidation
by SSAO
Compound
Core
c
(pIC50 ± SD)
R
1
0g
0i
0j
0k
11a
1b
2a
2b
CH O 1,2,3-triazole 4-SO NMe
7.3 ± 0.1
7.7 ± 0.02
7.9 ± 0.01
6.9 ± 0.3
6.1 ± 0.2
6.3 ± 0.2
6.2 ± 0.2
6.1 ± 0.09
6.1 ± 0.1
5.7 ± 0.2
5.5 ± 0.1
5.7 ± 0.1
>4
>4
>4
>4
>4
>4
2.8
1.4
2
2
2
2
2
2
2
2
2
2
1
CH NH 1,2,3-triazole 4-SO NMe
2
2
1
CH2
1,2,3-triazole 4-SO NMe
2
1
-
-
-
-
-
1,2,3-triazole 4-SO NMe
2
pyrrole
pyrrole
4-SO NMe
6.6 ± 0.04
7.0 ± 0.04
6.1 ± 0.04
7.0 ± 0.07
2
1
3-SO NMe
2
1
2-Me indole
2-Me indole
4-SO NMe
2
1
3-SO NMe
2
a
The measurement of enzymatic activities was based on the detection of hydrogen peroxide with an Amplex-Red
oxidation assay ; Minimum of n=2 for each experiment; c Compound oxidation assay measuring the amount of
hydrogen peroxide generated at a single concentration (30 µM) of compound, relative to DMSO; SD = standard
deviation
46 b
ACS Paragon Plus Environment
1
2

Page 13 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Alternative linkers (CH NH, 10i, and CH , 10j, in place of CH O) revealed little difference in
2
2
2
terms of potency relative to 10g (Table 2), nor in turnover by SSAO. Removing rotatable bonds to
the biaryl structure 10k was tolerated, providing a more compact and less flexible structure with
reasonable potency for LOXL2 and good selectivity over LOX, however significant SSAO
turnover was still observed. Replacing the 1,2,3-triazole with a pyrrole (11a and 11b) maintained
potency without improving turnover. Moving to the 2-methylindole scaffold with a direct link to
the aryl sulfonamide (12a and 12b) resulted in a reduction in turnover by SSAO. Presumably this
improvement was due to the increase in steric bulk attached directly to the fluoroallylamine
warhead rather than an electronic effect, given the turnover of the related pyrroles 11a and 11b.
Interestingly for both the pyrroles and indoles, a switch from para- to meta-SO NMe provided an
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
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
2
increase in LOXL2 potency as well as improved selectivity over LOX.
Further profiling against the family of amine oxidases showed 12b to be a highly selective inhibitor
of LOXL2, with no inhibition of MAO-A or B (pIC s <4.5), and minimal activity towards SSAO
5
0
or DAO (pIC of 5.1 and 5.4 respectively). However, despite these improvements, 12b was
5
0
rapidly turned over by plasma amine oxidase, meaning that further modification was required to
enable identification of a compound suitable for preclinical evaluation in higher mammals (e.g.
dog, monkey, pig). Equally, additional improvements in selectivity were desired to enable accurate
evaluation of the effects of LOXL2 inhibition over LOX in preclinical models of disease.
As the 3-substituted dimethylsulfonamide aryl ring appeared necessary for LOXL2 potency and
selectivity over LOX, we chose to “fix” this moiety and investigate substitution of the indole
aromatic ring (Scheme 3 and Table 3). Introduction of an electron-withdrawing dimethyl amide
group (12c, 12d and 12e) at the 5, 6 or 7-position significantly improved SSAO and plasma amine
oxidase substrate activity such that these compounds were not turned over at all by either enzyme.
ACS Paragon Plus Environment
1
3

Journal of Medicinal Chemistry
Page 14 of 59
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
Potency for the 5- and 6-substituted examples also significantly improved, however the 7-
substituted example showed a dramatic drop-off in LOXL2 inhibition. Given the slightly better
potency displayed by 12c, further examples were pursued with substitution in the 5-position. A
chloro group was equally well tolerated (12f), as was a free phenol group (12g), however switching
to the methoxy group (12h) drastically increased turnover by SSAO. A reverse sulfonamide (12i)
also showed an increase in plasma amine oxidase turnover. Introduction of a methyl sulfone (12j)
improved LOXL2 potency further, but this was accompanied by an increase in turnover by SSAO
and plasma amine oxidase. Switching to a carboxylic acid afforded a similarly potent compound
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
1
2k (PXS-5120A) with improved selectivity over LOX (>300-fold) and minimal turnover, an
effect that was not mirrored in the 6-position (12l). Tetrazole (12m) was not a viable bioisosteric
replacement of the carboxylic acid in this case, leading to a drop in potency, whilst the ethyl ester
1
2o (PXS-5129A), which can be regarded as a pro-drug form of 12k (PXS-5120A), had good
potency but moderate SSAO turnover. Removal of the fluorine group from the allylamine warhead
12n, also Figure 4D) led to a 100-fold drop in potency, consistent with its proposed role in the
mechanism of inhibition.
(
ACS Paragon Plus Environment
1
4

Page 15 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 3. Synthesis of 2-methyl, 3-aryl indole-based inhibitors 12b-12oa
NHBoc
H
N
R
F
O
N
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
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
HO
a, b
NHBoc
O
O
d
S
H
N
N
F
R = CO Et
2
NHBoc
c
N
e,f
N
R
R
+
NH HCl
2.
O
F
f
O
Br
O
S
N
O
8a
S
N
F
R
O
O
S
N
a
Reagents and conditions: (a) NBS, then DMAP, Boc
2
O, CH
2
Cl
2
, rt; (b) aryl boronic acid/ester, Pd(PPh
3
)
4
, K
2
CO
3
2
o
(aq), dioxane, 80 C, then TFA, CH
group, then Me NH, HATU, DIPEA, DMF, rt; (f) Either 2 M HCl (in Et
Et O).
2
Cl
2
, rt; (c) Cs
2
CO
3
, DMF, rt; (d) LiOH, MeOH, THF, H
2
O, rt; (e) if R = CONMe
2
2
2
O), MeOH, or TFA, CH Cl2, then HCl (in
2
Table 3. Optimization of the indole core
LOXL2 a,b
LOX a,b
Oxidation by Oxidation by
Compound
12b
R group
c
c
(
pIC ± SD) (pIC ± SD)
SSAO
AOC4
5
0
50
H
7.0 ± 0.07
7.6 ± 0.14
7.2 ± 0.02
5.5 ± 0.05
6.9 ± 0.11
7.3 ± 0.03
5.7 ± 0.1
5.8 ± 0.20
6.1 ± 0.01
4.7 ± 0.11
5.5 ± 0.03
5.9 ± 0.18
1.4
1.2
1
>4
1
1
2c
2d
2e
2f
12g
5-CO NMe
2
2
2
2
1
6-CO NMe
1.1
1.4
1.1
1.1
2
1
7-CO NMe
1.1
1.2
1
2
1
5-Cl
5-OH
ACS Paragon Plus Environment
1
5

Journal of Medicinal Chemistry
Page 16 of 59
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
12h
2i
2j
5-OMe
7.1 ± 0.10
7.7 ± 0.07
8.3 ± 0.17
8.4 ± 0.04
6.3 ± 0.02
6.9 ± 0.03
6.1 ± 0.07
7.1 ± 0.09
7.7 ± 0.23
5.8 ± 0.04
6.0 ± 0.07
6.4 ± 0.06
5.8 ± 0.10
4.8 ± 0.03
4.8 ± 0.21
5.5 ± 0.18
5.2 ± 0.12
5.9 ± 0.09
4.6
1
1.6
1.5
1.3
1
5
-
1
NHSO Me
2
1
5-SO Me
1.3
1
2
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
12k
PXS-5120A
5
-CO H
2
12l
6-CO H
1.1
-d
1
2
12m
5-tetrazole
1
5
-CO H
2
1
2n
1
1.3
1
des-Fluoro
12o
PXS-5129A
5
-CO Et
2.4
1.6
2
PXS-5153A
n/a
6.4
a
The measurement of enzymatic activities was based on the detection of hydrogen peroxide with an Amplex-Red
oxidation assay ; Minimum of n=2 for each experiment; c Compound oxidation assay measuring the amount of
hydrogen peroxide generated at a single concentration (30 µM) of compound, relative to DMSO; Not tested; SD =
standard deviation
46 b
d
PXS-5120A (12k) exhibited good selectivity over the related amine oxidase enzymes and showed
limited auxiliary pharmacology in standard profiling assays (the compound was tested at 10 µM
against 119 different off-targets; see Supporting Information. Only the peripheral benzodiazepine
receptor was inhibited by >50%). Importantly, PXS-5120A exhibited similar potency for LOXL2
across all species tested (Table 4), allowing for easier translation of our in vivo findings when
transitioning into the clinic. In a functional assay measuring the ability of the compound to inhibit
the crosslinking of rat collagen in vitro,17b PXS-5120A displayed dose-dependent inhibition of
crosslinking, with a calculated IC of 13 nM. Comparison with PXS-5153A in the same assay
5
0
showed this to be a ~3-fold improvement, in line with the improved potency in the primary LOXL2
assay (see Supplementary Figure S1 and Table 3). PXS-5120A is metabolically stable in vitro and
has moderate protein plasma binding.
ACS Paragon Plus Environment
1
6

Page 17 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4. In vitro profile of 12k (PXS-5120A)
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
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
Assay- potency and selectivity
IC (nM)
50
Recombinant human LOXL2
Human fibroblast LOXL2
Recombinant mouse LOXL2
Recombinant rat LOXL2
5
9
6
6
Collagen oxidation assay
13
Recombinant human DAO
Recombinant human SSAO, MAO-A/B
Human fibroblast LOX
>10000
>30000
1260
1780
1754
2260
16
Bovine LOX (aorta)
Murine LOX (lung fibroblasts)
Recombinant human LOXL1
Recombinant human LOXL3
Recombinant human LOXL4
280
Assay – ADME parameters
Result
PPB (rat/dog/human), % bound
88/86/92
>120 mins
>90 mins
>120 mins
Plasma stability (rat, dog & human) t1/2
Microsomal stability (rat, dog & human) t1/2
Hepatocyte stability (rat, dog & human) t1/2
ACS Paragon Plus Environment
1
7

Journal of Medicinal Chemistry
Page 18 of 59
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
Profiling of the other LOXL family members (Table 4) shows that PXS-5120A is a potent inhibitor
of the LOXL3 enzyme, and a moderate blocker of LOXL4. The reason for selectivity for the
LOXL2/3 enzymes over LOXL4 is not immediately obvious. All three of these LOXL enzymes
are closely related in terms of protein domains, expressing a signal peptide domain and four
scavenger receptor cysteine-rich (SRCR) domains, in addition to their catalytically active LOX
domain (in contrast to the LOX and LOXL1 isoforms that are secreted in a propeptide form devoid
of SRCR domains). All three possess a high degree of homology at the C-terminal enzymatically
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
1
0
active region of the peptide. Given the lack of structural information concerning the active forms
4
3
of the LOX enzymes, comparisons of the active sites are limited to analysis of residues in the
immediate vicinity of the LTQ cofactors and the copper binding domains. One potentially
significant difference between the LOXL2/3 isoforms and LOXL4 is observed between the four
conserved histidine residues in the copper binding domain, wherein the presence of negatively
charged aspartate (LOXL2) and glutamate (LOXL3) residues is in contrast to a neutral glutamine
residue in LOXL4. It is possible that these negatively charged residues impart a subtle structural
change, leading to the observed preference of PXS-5120A for LOXL2/3 over LOXL4.
2
7
Interestingly, LOXL3 has recently been implicated in the pathology of pulmonary fibrosis, where
it is thought to be involved in both the activation of fibroblasts, and the transition of fibroblasts to
myofibroblasts. Upregulation of both LOXL2 and LOXL3 has also been observed in a carbon
tetrachloride (CCl )-driven rat model of liver fibrosis, suggesting that both enzymes may play a
4
role in the development of acute hepatic injury as well.17b These reports indicate that inhibition of
LOXL2, with additional inhibition of LOXL3, would represent a more effective approach for
disease modulation than targeting a single protein, and therefore have a higher chance in the
successful treatment of fibrotic lung and liver disease.
ACS Paragon Plus Environment
1
8

Page 19 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
PXS-5120A displayed a) time-dependent LOXL2 inhibition (apparent K = 83 nM and k = 0.89
i
inact
-1
min ) (Figure 4A and Supplementary Figure S2); b) substrate competition, in which increasing
concentrations of the reference substrate, putrescine, reduced inhibition by PXS-5120A in a
competitive manner (Figure 4B) and c) irreversible inhibition in jump dilution experiments, in
which, after 100-fold dilution from a 30 x IC concentration of the inhibitor, maximal inhibition
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
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
0
(>80%) of the enzyme activity was maintained when substrate was re-introduced (Figure 4C). To
confirm that the mechanism of inactivation was indeed driven by the presence of the
fluoroallylamine group on the warhead, the des-fluoro analogue 12n was profiled head to head
with PXS-5120A (Figure 4D) in a bound LOXL2 washout assay. Monitoring the enzyme activity
for 30 minutes after washout showed that 12n had been mostly displaced, whereas PXS-5120A
(and BAPN) prevented recovery of enzyme activity, reaffirming that the vinyl fluoride plays a
critical role in the mechanism of inhibition of the enzyme. Furthermore, analysis of compound
oxidation (in the absence of substrate) of 12n vs PXS-5120A by LOXL2, in a similar manner to
what has been described above for SSAO and plasma amine oxidase, shows that whereas PXS-
5
120A has no detectable substrate activity (oxidative turnover after 30 min relative to DMSO =
), 12n is turned over substantially (relative oxidation = 6 at 30 mins). Taken together these
1
observations support the hypothesis that PXS-5120A acts as an irreversible inhibitor of LOXL2
through a mechanism based on the displacement of the fluorine atom from the fluoroallylamine
warhead moiety. Further experiments (e.g. protein digest – LCMS or X-ray crystallography) would
be required to unequivocally prove the proposed mode of action of PXS-5120A and verify if it is
covalently bound to the cofactor.
ACS Paragon Plus Environment
1
9

Journal of Medicinal Chemistry
Page 20 of 59
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
Figure 4. Mechanistic studies on 4 (PXS-5120A)
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
(A) LOXL2 inhibition by PXS-5120A is time dependent.
1
00
1.9 mM
.8 mM
15 mM
3
8
6
4
2
0
0
0
0
0
7
.5 mM
30 mM
-9.0
-8.5
-8.0
-7.5
-7.0
-6.5
-6.0
Log[PXS-5120A] (M)
(B) Substrate competition; Rising concentrations of the substrate putrescine compete with LOXL2 inhibition by PXS-
5
120A, suggesting that PXS-5120A acts directly through interacting with the catalytic cofactor.
ACS Paragon Plus Environment
2
0

Page 21 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
% Recovery
1
100
10
9
0
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
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
80
70
60
50
40
30
20
10
0
DMSO
BAPN
PXS-5120A
(
C) Jump dilution; 1 h soak with inhibitor (at a concentration equivalent to 30 x IC50) followed by a 100-fold dilution,
before addition of putrescine. Assay readout at 30 minutes after substrate addition. No significant return of activity
was observed for either BAPN or PXS-5120A.
ACS Paragon Plus Environment
2
1

Journal of Medicinal Chemistry
Page 22 of 59
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
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
(
D) Washout experiment using bound LOXL2; Time course of substrate turnover post washout showing the return of
enzyme activity over time after exposure to the inhibitor. Neither PXS-5120A nor BAPN (plots overlapping) showed
any return of activity over background after 30 mins. Des-fluoro PXS-5120A, 12n, shows a return of enzyme activity
to a level of ~75% of that observed where no inhibitor was present pre-washout.
PXS-5120A was found to be poorly absorbed and rapidly excreted in both rats and mice, leading
to very low oral bioavailability (<1%, data not shown). PXS-5129A, however, was found to be
readily absorbed following oral gavage, and once in the circulation, rapidly hydrolyzed to release
PXS-5120 (free base) in vivo, affording plasma concentrations well above the LOXL2 IC (6 nM,
5
0
equivalent to ~3 ng/mL) for a prolonged period (>6 hours) in mice, whilst remaining well below
the IC for LOX throughout (Figure 5).
5
0
ACS Paragon Plus Environment
2
2

Page 23 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
000
PXS-5129A
PXS-5120
1
00
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
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
10
1
0
2
4
6
Time (h)
Figure 5. Mouse plasma profile of PXS-5120 (free base of PXS-5120A) after oral (20 mg/kg) administration of the
corresponding ethyl ester, PXS-5129A. Pharmacokinetic analysis of PXS-5129A and PXS-5120 in BALB/c mice,
showing the hydrolysis of PXS-5129A in vivo to afford therapeutically relevent plasma concentrations of PXS-5120.
The plasma concentrations of both PXS-5129A and PXS-5120 were quantified by LCMS/MS. Each data point
represents an average of 3 animals.
With the drug/pro-drug pairing of PXS-5120A/5129A in hand we next sought to test the theory
that selective inhibition of LOXL2 (and LOXL3) over LOX could affect a therapeutic outcome in
relevant models of disease where LOXL2 was either significantly upregulated or has been shown
to play a clear role in the disease pathology through aberrant collagen cross-linking.
To evaluate PXS-5120A/PXS-5129A as a potential therapy for hepatic injury, a 6-week carbon
tetrachloride (CCl )-induced mouse model of liver fibrosis was used, with both prophylactic (15
4
mg/kg, weeks 1-6, n=12) and therapeutic (5 and 15 mg/kg, weeks 4-6, n=15/group) once daily,
oral administration of PXS-5129A employed. BAPN (2 mg/mL in drinking water, provided ad
4
7
libitum and changed every second day for 6 weeks, n=15) and imatinib (orally dosed at 50 mg/kg,
daily for weeks 1-6, n=12) were included as controls. At the end of the study 4% of the total liver
area was fibrotic in the CCl -no treatment group (as quantified by both Picro Sirius Red and α-
4
ACS Paragon Plus Environment
2
3

Journal of Medicinal Chemistry
Page 24 of 59
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
smooth muscle actin -SMA) staining). No beneficial effect was observed in the Picro Sirius Red
readout with the positive controls (BAPN and imatinib) (see Supplementary Figure S3), and only
the lower dose of PXS-5129A showed a significant reduction in fibrotic area. Given that Picro
Sirius Red measures total collagen content (cross-linked and soluble), it is possible that the effects
of the compound are underestimated by the overwhelming quantity of soluble collagen (i.e.
although the total collagen content is unchanged, the proportion of cross-linked to soluble collagen
is different). To account for this possible issue, the area of fibrosis was also quantified by staining
for α-smooth muscle actin (α-SMA), a marker of myofibroblast formation (Figure 6 and
Supplementary Figure S4). In this readout both positive controls showed a strong anti-fibrotic
effect when dosed prophylactically, suggesting that the mechanical stress on the myofibroblasts
had been reduced. When dosed prophylactically PXS-5129A also reduced the fibrotic area, similar
to the positive controls. Additionally, therapeutic dosing effected a significant decrease in fibrosis
at both 5 and 15 mg/kg doses. In this study no significant effect on ALT or AST liver function
markers was observed for PXS-5120A/PXS-5129A, nor imatinib, whereas BAPN increased ALT
and AST levels (Supplementary Figure S5).
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
ACS Paragon Plus Environment
2
4

Page 25 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9

-SMA staining
6
5
4
3
2
1
0
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
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
#
#
##
#
##
#
##
#
##
l
l 4
A
A
s
s
k
s
o
r
C
9
9
k
k
t
2
2
w
w
w
n
C
g
1
1
6
6
6
o
5
5
-
-
-
-
-
C
S
S
A
N
P
b
i
X
X
9
tin
P
P
2
A
1
g
5
B
a
k
k
-
m
i
/
/
l 4
S
g
g
X
C
l 4
m
m
P
C
C
5
5
1
g
C
l 4
k
/
l 4
C
g
C
C
m
C
5
1
l 4
C
C
Figure 6. Inhibition of CCl
4
-induced liver fibrosis by PXS-5120A/PXS-5129A. Quantitative measurement of fibrosis
in a 6 week model of CCl -induced liver fibrosis in BALB/c mice, and the anti-fibrotic effects of PXS‐5129A, BAPN
4
and imatinib. The area of fibrosis was quantified by staining for -SMA and is measured as percentage of total liver
area. PXS-5129A reduces fibrosis when dosed both prophylactically at 15 mg/kg (6 weeks, blue bar) and
therapeutically at 5 and 15 mg/kg (weeks 4-6, green bars). Results are expressed as mean ± SEM. # p<0.05, ## p<0.01
and ### p<0.001, vs. Group 2 (CCl
4
+ vehicle); data analysis was performed by one-way ANOVA followed by
Dunnett’s Multiple Comparison Test.
PXS-5120A/PXS-5129A was also evaluated in a 28 day bleomycin-induced mouse model of lung
fibrosis, using therapeutic (5 and 15 mg/kg, days 7-27, n=12/group) once daily, i.p. administration
of PXS-5129A. Imatinib (50 mg/kg, days 7-27, n=12) was again included as a positive control.48
An increase in fibrosis, as measured by the Ashcroft score,49 was observed at day 28 and
ACS Paragon Plus Environment
2
5

Journal of Medicinal Chemistry
Page 26 of 59
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
therapeutic treatment with imatinib (50 mg/kg i.p.) afforded a small but significant reduction
(Figure 7 and Supplementary Figure S6). PXS-5129A resulted in a dose-dependent reduction of
the Ashcroft score, with the 15 mg/kg dose reaching similar effect size to the positive control
imatinib. The anti-fibrotic reduction seen with imatinib in this instance is consistent with the
reported effects of this drug, however the extent of the anti-fibrotic effect of imatinib is muted
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
4
8
compared to historical data (a reduction of 50-70% is reported). Given the variability of this
model (potency of bleomycin used, route and frequency of administration, strain and age of mice
used, etc.), this disparity is not unexpected, and the model serves to re-inforce the therapeutic
potential of PXS-5129A in models of fibrotic disease.
5
4
3
2
1
0
#
#
###
l
e
b
i
g
g
a
l
k
k
/
c
n
/
m
i
i
t
r
h
g
g
o
e
a
m
m
n
v
m
i
5
5
1
A
9
A
2
9
1
2
5
1
-
5
S
-
X
S
P
X
P
Figure 7. Inhibition of bleomycin-induced lung fibrosis by PXS-5120A/PXS-5129A. Determination of the anti-
fibrotic effects of PXS-5129A and imatinib in a 28 day bleomycin-induced lung fibrosis in C57BL/6 mice. Analysis
4
9
of the amount of fibrosis as measured by the Ashcroft score. PXS-5129A reduces fibrosis when dosed therapeutically
at 15 mg/kg (from day 7 onwards, blue bars). Results are expressed as mean ± SEM. # p<0.05, ## p<0.01 and ###
ACS Paragon Plus Environment
2
6

Page 27 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
p<0.001, vs Group 2 (Bleo + vehicle); data analysis was performed by one-way ANOVA followed by Dunnett’s
Multiple Comparison Test.
CONCLUSION
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
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
In summary, we have identified a fluoroallylamine-based lead (4), with inhibitory activity against
SSAO, LOX and LOXL2, which has been further developed into a potent and selective inhibitor
of LOXL2 (and LOXL3). Rapid exploration for additional pharmacophoric interaction points was
carried out by a “Click-Chemistry” approach, resulting in the identification of a dimethyl
sulfonamide as a key recognition feature required for LOXL2 potency. The potential for turnover
by the related amine oxidases, SSAO and plasma amine oxidase, led us to develop a compound
oxidation assay for these enzymes, which subsequently aided in the identification of the 2-
methylindole core with decreased SSAO substrate activity.
Modification of the 5-position of the indole proved most effective for further optimizing LOXL2
potency and eliminating plasma amine oxidase-mediated turnover of these compounds. This
afforded PXS-5120A, an irreversible inhibitor of LOXL2 with single-digit nanomolar potency for
LOXL2, the most potent LOXL2 inhibitor reported to date. Importantly for both liver and lung
diseases the additional inhibition of LOXL3 activity affords a dual inhibitor with an increased
chance of reducing fibrosis in disease states.
PXS-5120A is selective (~300-fold) over the related LOX enzyme, which is believed to be vital
for maintaining the structural integrity of the skin, heart and bone function. Equally PXS-5120A
shows no off-target inhibition of the related amine oxidases and, importantly, is not recognized as
a substrate by either SSAO or plasma amine oxidase (the AOC4 gene product), affording a
ACS Paragon Plus Environment
2
7

Journal of Medicinal Chemistry
Page 28 of 59
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
metabolically stable compound in vitro, well suited for assessing the in vivo potential of
LOXL2/LOXL3 inhibition.
As the zwitterionic nature of the compound rendered it unsuitable for direct oral administration, a
pro-drug approach was employed, and the administration of PXS-5129A (the corresponding ethyl
ester derivative of PXS-5120A) in two different animal models of disease successfully
demonstrated the promise of selective LOXL2/3 inhibition for the treatment of liver and lung
fibrosis at doses as low as 5 mg/kg.
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
Further progress regarding mechanism-based, orally bioavailable (without the need for a pro-drug)
inhibitors of LOXL2/3 with a fast onset of inhibition and sustained target occupancy will be
reported in due course.
EXPERIMENTAL SECTION
1
Chemistry General Methods. Purity was determined by HPLC and H-NMR. Unless otherwise stated,
the purity of final compounds was ≥95%. HPLC spectra were recorded on a Agilent LCMS 1100 instrument
1
19
using reverse-phase conditions (acetonitrile/water, containing 0.1% formic acid). H-NMR and F-NMR
spectra were recorded using a Bruker 300 MHz NMR spectrometer. Chemical shifts (δ) are reported as
parts per million (ppm) relative to tetramethylsilane (TMS; internal standard). All commercially available
solvents and reagents were used as received. Where appropriate, reactions were carried out under an argon
atmosphere. Reactions were monitored by either analytical thin-layer chromatography (TLC) or by LCMS.
Flash column chromatography was conducted under medium pressure either on silica (Merck silica gel
4
0−63 μm) or on prepacked silica gel cartridges using a flash chromatography system (CombiFlash Rf200,
Teledyne Isco systems, USA), and eluents were monitored by UV light (λ = 254/280 nm). Low resolution
mass spectra (MS) were obtained as electrospray – atmospheric pressure ionization (ES-API) mass spectra,
which were recorded on an Agilent LCMS 1100 instrument using reverse-phase conditions
ACS Paragon Plus Environment
2
8

Page 29 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
acetonitrile/water, containing 0.1% formic acid). Melting points were obtained using a MEL-TEMP®
(
Electrothermal melting point apparatus. All animal experiments performed were conducted in compliance
with institutional guidelines and approval from local ethics committees. The general chemical methods
outlined in the experimental are exemplified for the more important molecules, and may therefore be
referred to before they are described in detail.
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
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
(S,Z)-4-(4-Amino-2-fluorobut-2-enyloxy)-N-(1-phenyl ethyl)benzenesulfonamide (5). Synthesized
using chemistry outlined in reference 44.
(Z)-tert-butyl 3-fluoro-4-hydroxybut-2-enylcarbamate (7).
To a stirring solution of 3-amino-1,2-propanediol (20.0 g, 0.22 mol) in water (200 mL) at 0-5 C was added
di-tert-butyl dicarbonate (55.5 mL, 0.24 mol). After adjusting the alkalinity of the solution to pH~9 by
addition of aq. NaOH (6 N), the mixture was left to stir at rt for 18 h. The reaction mixture was cooled to
0
-5C and then acidified to pH~6 before the addition of sodium metaperiodate (56.3 g, 0.26 mol). The
resulting suspension was stirred at rt for 2 h. The mixture was filtered to remove all solids and the filtrate
was transferred to a separatory funnel and extracted with ethyl acetate (200 mL). Sodium chloride was
added to the aqueous layer until a saturated solution was obtained. The aqueous layer was then extracted
further with ethyl acetate (100 mL). The combined organics were dried over Na SO and then concentrated
2
4
in vacuo to give crude tert-butyl 2-oxoethylcarbamate (45.7 g) as a yellow gum. The crude material was
used in the subsequent step without purification.
To a stirring suspension of crude tert-butyl 2-oxoethylcarbamate (43.7 g, 0.22 mol) and magnesium sulfate
(
32.0 g) in acetonitrile (200 mL) at 0 C under N was added sequentially ethyl 2-fluorophosphonoacetate
2
(55.7 mL, 0.27 mol) and 1,8-diazabicyclo[5.4.0]undec-7-ene (32.8 mL, 0.22 mol). The reaction mixture
was allowed to warm to rt and stirring was continued for 3 h. After removing the solvent under reduced
pressure the residue was taken up in ethyl acetate (200 mL) and then transferred to a separatory funnel. The
organics were washed successively with aq. HCl (2 M; 100 mL x 2), aq. NaOH (2 M; 100 mL x 2) and
brine (100 mL). After drying over MgSO , the organics were concentrated in vacuo to give the crude,
4
ACS Paragon Plus Environment
2
9