Development of Autotaxin Inhibitors: An Overview of the Patent and

Jan 8, 2016 - Diana Castagna obtained her M.Chem. degree at the Heriot-Watt .... A. M. Santini , A. J. Stein , D. Lonergan , C. D. King , I. Calderon ...
0 downloads 0 Views 2MB Size
Subscriber access provided by UNIV OSNABRUECK

Perspective

The Development of Autotaxin Inhibitors: An Overview of the Patent and Primary Literature Diana Castagna, David C. Budd, Simon J. F. Macdonald, Craig Jamieson, and Allan J. B. Watson J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01599 • Publication Date (Web): 08 Jan 2016 Downloaded from http://pubs.acs.org on January 12, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Medicinal Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 57

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

Journal of Medicinal Chemistry

The Development of Autotaxin Inhibitors: An Overview of the Patent and Primary Literature Diana Castagna,a David C. Budd,b Simon J. F. Macdonald,b Craig Jamieson,a* and Allan J. B. Watsona* a

WestCHEM, Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow, G1 1XL, UK. b

GlaxoSmithKline, Medicines Research Centre, Gunnel Wood Road, Stevenage, Hertfordshire, SG1 2NY, UK.

ABSTRACT

The Autotaxin-Lysophophatidic acid (ATX-LPA) signalling pathway is implicated in a variety of human disease states including angiogenesis, autoimmune diseases, cancer, fibrotic diseases, inflammation, neurodegeneration, and neuropathic pain, amongst others. As a result, ATX-LPA has become of significant interest within both the industrial and the academic communities. This review aims to provide a concise overview of the development of novel ATX inhibitors, including the disclosure of the first ATX clinical trial data.

1 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 2 of 57

Introduction Autotaxin (ATX) is an extracellular enzyme, which is part of the ecto-nucleotide pyrophosphatase/phosphodiesterase family (ENPP2), of which there are seven members.1 ATX was first isolated in 1992 from A2058 melanoma cells and categorized as an autocrine motility factor, and was subsequently found to be responsible for cell proliferation, survival, invasion, migration, and neovascularization.2

Following the initial discovery, further analysis by Tokumura et al3 and Umezu-Goto et al4 established that ATX was in fact the glycoprotein lysophospholipase D (lysoPLD),3-4 which is responsible for the hydrolysis of lysophosphatidycholine (LPC) 1 to the bioactive phospholipid lysophosphatidic acid (LPA) 3 with the release of choline 2 (Scheme 1).5

Scheme 1. ATX-mediated hydrolysis of 1 to 3 with concomitant release of 2.

Once generated, LPA can promote a range of physiological events which include cell proliferation, survival, and motility, amongst others, by acting through a set of six G-proteincoupled receptors (GPCRs) known as LPA1-6.2b,6 The LPA receptor family can be further divided into two categories: (1) the EDG sub-family consisting of LPA1-3 and (2) LPA4-6, which contain less than 40% homology to LPA1-3.7 Each receptor couples to specific Gα proteins (Gs, Gi, Gq, and G12/13), as illustrated in Figure 1, which will in turn initiate a range of cellular signaling cascades.2b Some of the main pathways include: the activation of phospholipase C (PLC) leading to the hydrolysis of phosphatidylinositol-bisphosphate (PIP2), 2 ACS Paragon Plus Environment

Page 3 of 57

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

Journal of Medicinal Chemistry

which will, in turn, initiate calcium mobilization and protein kinase C (PKC) activation; activation of the Ras, MERK, ERK pathway leading to cell proliferation; activation of the P13K-AKT survival pathway responsible for the suppression of apoptosis; and, lastly, activation of the Rho pathway leads to cytoskeletal remodeling, shape change, and cell migration.8

Figure 1. Signaling pathways activated by LPA1-6.

Accordingly, intervention into the ATX-LPA signaling cascade has become of increasing interest over the past decade with a substantial rise in activity noted over the past five years, as demonstrated by the number of patents and publications emerging (Figure 2). This heightened level of research has implicated the ATX-LPA signaling pathway in an increased number of physiological and pathological processes that includes angiogenesis,9 autoimmune disease,10 cancer,11 fibrotic diseases,12 inflammation,13 neurodegeneration,2b and pain,14 amongst others. In parallel with this increased level of interest in the therapeutic potential, the 3 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

development of new chemical entities capable of modulating the ATX-LPA pathway has also seen significant investment by both the academic and the industrial communities.

ATX Publications 120 Total No. of Publications

100 No. of Publications

Peer Reviewed Articles 80 No. of Patents 60 40 20

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

0 1992

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

Page 4 of 57

Year

Figure 2. ATX-related publications since 1992. Data collected from SciFinder and Espacenet.15

Structure and Function of the Autotaxin Enzyme Based on its perceived therapeutic potential, a substantial amount of effort has been invested into understanding both the structure and function of ATX. Currently, it has been established that there are five known splice variants of ATX referred to as ATXα (ATXm), ATXβ (ATXt), ATXγ (PD-Iα), ATXδ, and ATXε, all of which are catalytically active.1a,2b,16

In 2008, the cloning and tissue distribution of the three main isoforms (α, β, and γ) using both human and mouse tissue was undertaken in order to characterize and develop an understanding of their respective functions, a summary of which is presented in Table 1.17

4 ACS Paragon Plus Environment

Page 5 of 57

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

Journal of Medicinal Chemistry

Table 1. Splice Variants of ATX Splice Variant

Isoform

Location

Role

Low quantities present in the Original ATXα (ATXm)

Minor

isoform

central and peripheral nervous isolated from melanoma system.17

cells.

Predominately in the brain and Originally cloned from peripheral tissue.17

ATXβ

teratocarcinoma

Major Small

(ATXt)

quantities

present

central nervous system.

in

17

ATXγ Minor

and

found

cells to

be

identical to lysoPLD.2b

Similar

structure

ATXβ

and

to their

Central nervous system.17 biochemical

(PD-Iα)

functions

are indistinguishable.2b

Following the characterization of the different isoforms of ATX, crystal structures were solved in 2011 by Nishimasu and Hausmann, of both rat and mouse ATX, respectively, and which share 93% homology to human ATX.18 The enzyme was found to consist of three main domains; the N-terminal, of which there are two somatomedin B (SMB)-like domains, known as SMB1 and SMB2; a catalytic phosphodiesterase domain (PDE), situated in the centre; and, the C-terminal nuclease-like domain (NUC). In addition to these, there are two linker portions referred to as the L1 linker, which connects the SMB2 domain to the catalytic domain, and the L2 linker (or the lasso loop) connecting the catalytic domain to the nucleaselike domain, as illustrated schematically in Figure 3a.18b, 19 5 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 6 of 57

a)

b)

Figure 3. (a) Schematic representative diagram of ATX domains. (b) 3D structure of ATX representing the four main domains, SMB1 in orange, SMB2 in brown, catalytic domain in cyan, nuclease like domain in purple, L1 in red, and L2 in green.

The modular domain arrangement of ATX is such that the two SMB domains and the nuclease-like domain encapsulate the catalytic domain, with the two linkers (L1 and L2) reinforcing the structure and maintaining the active conformation of the enzyme.2b,18b Further structural characteristics which underpin the association between the PDE and NUC domains include the presence of an interdomain disulfide bond between Cys413-Cys801, the Asn524linked glycan, as well as the presence of a number of hydrophobic and hydrophilic interactions. In addition to these, there are a number of well-ordered water molecules situated between the catalytic and nuclease-like domain, which results in an extensive hydrogenbonding network, ensuring a further degree of thermodynamic stability.

The two SMB domains form a cysteine knot fold18b which consists of four pairs of crossed disulfide bonds; these domains are structurally similar to the SMB domains present in the 6 ACS Paragon Plus Environment

Page 7 of 57

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

Journal of Medicinal Chemistry

integrin ligand vitronectin, which is an extracellular matrix protein that interacts with plasminogen activator inhibitor-1 (PAI-1) and urokinase-type plasminogen activator (uPAR).18b It has also been shown that these domains are important for binding of newly generated LPA to ATX, and is also speculated to be involved in the subsequent release of LPA.18a

The catalytic domain itself can be broken down into two main parts known as the core and the insertion sub-domains. The core adopts an alkaline phosphatase superfamily fold and coordinates to the two essential Zn2+ ions via conserved aspartic acid and histidine residues, Figure 4.2b,18b In 2003, Gijsbers reported that the presence of the threonine (Thr209) residue, located within the phosphodiesterase domain, along with two Zn2+ ions were responsible for the hydrolytic activity of the enzyme (Figure 4).18b,19-20 These structural features are conserved among the ENPP family, which is further exemplified by the similarity in the catalytic domain to that of Xanthomonos axonapodis (XaNPP), a bacterial NPP enzyme.

Figure 4. a) Representation of ATX with the catalytic site highlighted. b) Enlarged image of the catalytic site illustrating the Zn2+ ions, and the essential resides: Thr209 (red), and the catalytic triad made up of His359 (orange), Asp358 (blue), and Asn230 (green).

7 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 8 of 57

Lastly, the C-terminal nuclease-like domain contains a mixed α/β fold and binds to Ca2+, Na+, and K+. These ions are coordinated by residues present in this domain, and are bound close to the interface of the catalytic and the nuclease-like domains, implying an important role in the interaction of the two domains.18b

An interesting structural feature of the enzyme is the presence of a hydrophobic channel connecting the large binding pocket (15 Å), situated within the catalytic domain of the enzyme (Figure 5a), through a T-shaped junction21 (as illustrated in Figure 5b). This can accommodate monoacyl phospholipids such as LPA and LPC. The function of this hydrophobic tunnel remains under discussion as it could play two potential roles: the first to release newly generated LPA to the cellular microenvironment allowing activation of the cognate GPCRs. Conversely, the tunnel entrance may promote LPC uptake or act as an entrance and exit tunnel.21 Additional structural analysis is required in order to validate either of these hypotheses.

Figure 5. (a) Representation of the ATX active site, hydrophobic pocket, and tunnel with LPA. (b) A representation of the T-shaped structure.

8 ACS Paragon Plus Environment

Page 9 of 57

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

Journal of Medicinal Chemistry

Nishimasu and co-workers conducted further structural analysis of the hydrophobic pocket and have disclosed crystal structures of mouse ATX complexed with LPA molecules of varying chain length (14:0, 16:0, 18:1, 18:3, and 22:6).22 The data generated indicated that the phosphate group, glycerol portion, and the acyl-chain of the various forms of LPA are all recognized in a similar way by the enzyme where one of the oxygen atoms from the phosphate group binds to one of the Zn2+ ions while another forms hydrogen bonds to the side chain of Asn230 and the OH and NH of Thr209. This is exemplified in Figure 6, which shows LPA 14:0 4 bound within the active site of ATX.

Figure 6. Structure of 4 complexed within the catalytic site of ATX. a) Ribbon representation of ATX, colored as follows: SMB1 domain in orange, SMB2 domain in brown, catalytic domain in cyan, nuclease-like domain in purple, L1 in red, L2 in green, Zn2+ in grey, and 4 in

9 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 10 of 57

pink. b) Structure of 4. c) and d) Enlargement of the catalytic site indicating the interactions between the phosphate group with the Thr209 residue and one of the Zn2+ ions.

A notable feature of the binding pocket, which differs from other phospholipases, is that no protein plasticity is observed regardless of whether the enzyme is in either a bound state (i.e., with orthosteric ligand) or an unbound state. 21

In addition to the LPA co-crystal structure data with murine ATX, Hausman and co-workers disclosed the crystal structure of the apo form of rat ATX as well as a co-crystal structure of a potent inhibitor HA155 (5, Figure 7b).23 Based on this analysis, it was possible to correlate inhibition with the observed binding mode.18a, 23

From this crystal structure data it was possible to establish key interactions between the enzyme and the boronic acid-derived inhibitor 5 (Figure 7). The data generated illustrated that the boronic acid forms a reversible covalent bond with the nucleophilic hydroxyl group of the Thr209 residue, while one of the hydroxyl groups can be effectively stabilized between the two Zn2+ ions present in the pocket. The hydrophobic ring of the inhibitor forms a network of van der Waals interactions. The fluorobenzene moiety points directly into the hydrophobic pocket and is orientated perpendicular to the protein surface. Therefore, this data implies the need for an acidic warhead to coordinate to the two Zn2+ ions, the presence of a moiety which can interact with the Thr209 residue, and a lipophilic tail which can be accommodated within the hydrophobic pocket. Many of these features are conserved in inhibitors which bind at the catalytic site (vide infra).

10 ACS Paragon Plus Environment

Page 11 of 57

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

Journal of Medicinal Chemistry

Figure 7. Co-crystal of 5 within the catalytic domain of ATX. a) Ribbon representation of ATX, colored as follows: SMB1 domain in orange, SMB2 domain in brown, catalytic domain in cyan, nuclease-like domain in purple, L1 in red, L2 in green, Zn2+ in grey, and 5 in pink. b) Structure of 5. c) and d) Enlargement of the catalytic site indicating the interactions between the boronic acid and both the Thr209 residue and one of the Zn2+ ions.

As stated previously, it is well established that ATX is responsible for the hydrolysis of phospholipids such as LPC,5 yet this is not its only function as it is also known to hydrolyze other phospholipid species. For example, sphingosylphosphorylcholine (SPC, 6) is converted into sphingosine 1-phosphate (S1P, 7) by ATX (Scheme 2).24 S1P is a bioactive lipid displaying similar signaling properties to LPA (as well a range of other signaling consequences via other receptor families);1a,25 however, it should be noted that this is not the

11 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 12 of 57

main route for S1P synthesis – the main source is believed to originate predominately from phosphorylation of sphingosine by sphingosine kinases.1a,26

Scheme 2. Hydrolysis of 6 to 7 catalyzed by ATX.

Disease Implications As described above, the ATX-LPA signaling pathway is implicated in a variety of human pathologies including angiogenesis,9 autoimmune diseases,10 cancer,11 fibrotic diseases,27 inflammation,13 neurodegeneration,2b and neuropathic pain,14 amongst others. However, the most extensively studied disease states include cancer and fibrotic diseases, such as idiopathic pulmonary fibrosis (IPF). IPF is perhaps the most widely studied disease associated with the ATX-LPA pathway, possibly due to the paucity of effective treatment options available for this disease. To date only one ATX inhibitor has been reported to have progressed into clinical trials, GLPG1690 (structure not disclosed).28

ATX has been shown to be linked with a number of cancers due to its ability to stimulate chemokinesis and chemotaxis in melanoma cells.2a This was followed by the discovery that ATX affects motility-dependent processes such as invasion and metastasis.29 Moreover, ATX is highly expressed in many cancer tissues, including non-small-cell lung cancer,30 hepatocellular carcinoma,31 renal cell carcinoma,32 breast cancer,33 thyroid carcinoma,33b,34 ovarian cancer, and Hodgkin’s lymphoma.2b The link between cancer and LPA/ATX has been recently reviewed35 focusing on the roles of ATX and LPA receptors in cancer

12 ACS Paragon Plus Environment

Page 13 of 57

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

Journal of Medicinal Chemistry

progression, tumor cell invasion, and metastasis. Accordingly, the development of an LPA receptor antagonist or ATX inhibitor has been attractive as a therapeutic strategy for the suppression of tumor metastasis.19

The ATX-LPA axis has also been implicated in a number of chronic fibrotic disease conditions including within the liver,36 kidney,37 adipose tissue,38 synovial joints,39 retina,40 and lung.41

In the context of the therapeutically challenging disease of lung fibrosis, locally produced LPA has been suggested to drive several different pathological processes leading to excessive production of extracellular matrix in the interstitial space, permanent scarring of the lung, irreversible loss of tissue architecture, and a reduction in lung function. Expression of ATX has been demonstrated in a variety of tissues; however in the murine lung, ATX expression appears to be highest in bronchial epithelial cells and alveolar macrophages.42 Increased expression of ATX has been reported in the lungs of bleomycin-treated mice and patients with IPF42 in which the pattern of expression suggests that epithelial cells and macrophage exhibit the highest intensity of ATX expression particularly in cells that are juxtaposed to fibroblastic foci.42 In line with the reports of increased expression of ATX in the fibrotic lung, increased levels of LPA isoforms, notably the 22:4 species, have been reported in exhaled breath condensate of IPF patients compared to healthy volunteers.43

The mechanisms underlying the pro-fibrotic properties of the LPA pathway in the lung are complex but have been variously ascribed to accumulation of lung fibroblasts in the lung interstitial space through increased migration and proliferation,27 increased vascular leakage,27 and apoptosis of lung epithelial cells.27 The finding that mice genetically deficient

13 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 14 of 57

in macrophage-specific ATX expression are refractory to bleomycin-induced lung fibrosis42 also suggests an important role for this cell type in driving LPA-mediated pro-fibrotic responses in this organ.

Some dispute exists regarding the identity of the LPA receptors responsible for driving the pro-fibrotic actions of LPA in the lung with data supporting a role for the LPA1 receptor,27,44 LPA2 receptor,45 and LPA3 receptor46 based on a combination of pharmacological and genetic approaches. The use of ATX inhibitors might circumvent the issue of molecular redundancy with regards to which combination of receptor isoform targeting would be most effective with regards to achieving anti-fibrotic efficacy. In this regard, GWJ-A-2347 (8) (Figure 8), delivered intra-peritoneally to C57/BL6 mice reduced LPA levels in the bronchoalveolar lavage fluid and attenuated inflammatory responses and collagen production in the lungs of bleomycin-treated animals42 suggesting this might represent a viable approach.

Figure 8. Compound 8.

Increased expression of ATX has also been established in non-malignant cell tissue, for example in the frontal cortex of Alzheimer-like dementia patients48 and in the cerebrospinal fluid of multiple sclerosis patients.49 ATX has been linked with neuropathic pain, such as that found in patients with osteoarthritis, as the production of LPA influences a range of biological and biochemical processes including up-regulation of pain-related proteins through the LPA1 receptor.14,50

14 ACS Paragon Plus Environment

Page 15 of 57

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

Journal of Medicinal Chemistry

Assays for the identification of ATX inhibition. The availability of a robust and reliable assay to assess ATX inhibition is of great importance in the path towards identification of new inhibitors for this target. Currently, there are a range of different in vitro assays which have been developed to determine ATX inhibition. These can be divided into two main categories: those using the natural enzyme substrate (LPC) and those using unnatural substrates, such as bis-(p-nitrophenyl) phosphate (bis-pNPP). Both of these are discussed in the subsequent section.

Natural Substrates ATX hydrolyzes LPC to LPA and choline, both of which can be monitored in order to measure ATX activity. There are a number of methods that can be used to achieve this including using radiolabelled substrates, liquid chromatography-tandem mass spectrometry, and a choline release method, as illustrated in Table 2.

Table 2. Natural Substrate ATX Assays. Assay

Substrate

Analyte

Comments Source of Enzyme: Recombinant ATX prepared by baculovirus from SF9 cells. Analysis: Autoradiography.25a

Radiolabelling4,21

14

C-1

14

C-3 Comments: Most direct method. Robust and sensitive assay. Less

suitable

for

high-throughput

15 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 16 of 57

screening.1a Source of Enzyme: Recombinant ATX from HEK293 cells.52 Analysis: LC-MS detection of LPA Comments: Very sensitive. 18a,51

LC-MS/MS

1

3 Can

be

used

to

detect

naturally

occurring LPA in biological fluids.1a High-throughput screening possible. Other MS techniques include MALDI.63 Source of Enzyme: Recombinant human ATX from MDA-MB-435 cells.53 Analysis: Absorbance at 405 nm.

Amplex Red 1 Choline release

2 Comments: Routinely used. High-throughput screening possible. Assay interference possible.

Of the assays described in Table 2, the choline release assay is the most popular; however, caution must be taken when using this protocol as there is the potential for small molecules to react with the coloring agent or inhibit the associated enzymes (HRP or choline oxidase) leading to false positives.54 However, incubation of the test compound with choline followed

16 ACS Paragon Plus Environment

Page 17 of 57

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

Journal of Medicinal Chemistry

by addition of the coloring agent has been shown to be an effective way of assessing whether interference has been/will be an issue.1a Unnatural Substrates An alternative means to test for ATX inhibition is through the use of unnatural substrates, which include thymidine 5’-mono-phosphate p-nitrophenol ester (pNP-TMP, 11), bis(pnitrophenyl)phosohate (bis-pNPP, 9), CPF4 (12), and FS-3 (14), with 11 and 9 the simplest methods to measure lysoPLD activity (Table 3).

17 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 18 of 57

Table 3. ATX assays based on unnatural substrates. Substrate

Analyte

Comments

Ref

Source of Enzyme:

25a,55

Recombinant ATX from HEK293 cells. Analysis: Absorbance at 405 nm. Comments: used.

Routinely

Facile,

high

throughput capability, and commercially available. Caution when using: compound binding remote from the catalytic site will give false negative.18c Source

of

Enzyme:

Recombinant ATX from MDA-MB-435 cells. Analysis: Absorbance at 405 nm. Comments: Caution when

18 ACS Paragon Plus Environment

3,20

Page 19 of 57

Journal of Medicinal Chemistry

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

using: compound binding remote from the catalytic site will potentially yield a false negative.56 25a,57

Source of Enzyme: Recombinant ATX using HEK293T (Bac-toBac baculovirus expression system). Analysis: FRET (excitation at 355 nm, emission at 460 and 520 nm). Comments: Use with serum free conditioned culture medium isolated from cells expressing the enzyme. Source

of

Enzyme:

Recombinant ATX from MDA-MB-435 cells. Analysis:

FRET

(excitation at 485 nm and

19 ACS Paragon Plus Environment

59

Journal of Medicinal Chemistry

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

Page 20 of 57

emission at 528 nm). Comments: Caution when using: compound binding remote from the catalytic site will potentially yield a false negative.18c,58

One of the main disadvantages of using synthetic unnatural substrates in place of the natural substrate is the potential lack of specificity for ATX over other phosphatases. These substrates can be readily hydrolyzed by other members of the NPP family, as well as by other less specific phosphodiesterase enzymes. Therefore, it is recommended that these assays should only be used with purified enzyme to ensure accuracy and proper interpretation of the data generated.60

Of the assays described, using either natural and unnatural substrates, the two most commonly used methods to measure ATX activity are the lysoPLD assay using the natural substrate LPC and the unnatural substrate 9. There has been, however, some debate regarding the reliability of these assays. For example, problems have been encountered when using the endogenous substrate LPC due to the presence of choline or the coloring reagents present in the assay, which as discussed above can react with test compounds generating false positives.58 An additional problem has been found when using unnatural substrates, such as 11, as these will interact differently in terms of binding to the enzyme and saturation kinetics compared to the natural LPC substrate. This hypothesis has been validated in molecular modeling studies by Fells et al,54 and Hoeglund et al53 which indicated that compounds that bind in the hydrophobic pocket of the enzyme may not be properly identified as inhibitors as 20 ACS Paragon Plus Environment

Page 21 of 57

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

Journal of Medicinal Chemistry

the substrate occupies a space distinct from the hydrophobic tunnel of the enzyme. This, however, is not the case when using 14 as an assay probe as the long lipophilic portion of the structure is accommodated within the hydrophobic pocket and, as such, may prove to be a more accurate method for identifying inhibitors which bind at a proximal site.54,56,61 Having stated this, while the 14 substrate binds in the hydrophobic pocket, it is remote from the hydrophobic channel and, as such, any compounds which bind at this position will be unable to hydrolyze the 14 substrate and may lead to a false negative.18c

In essence, from the data generated regarding the different ATX assays it seems it may be prudent to evaluate test compounds using a range of assays in order to confirm any activity as ATX inhibitors instead of relying on one assay in isolation.

Inhibitors The design and synthesis of novel compounds with ATX inhibitory activity has been pursued by several academic research groups and industrial entities. The patent data has been previously reviewed in the literature up until 2013.19 The present discussion will take into account the previous reports as well as providing an up to date account on the current competitive landscape, with the aim of providing an authoritative overview of the known inhibitors to date. Some of the first ATX inhibitors were lipid-like structures that mimicked the natural LPC and LPA phospholipids (Table 4). Compound 1651,62 was one of the first ATX inhibitors developed and has been proven to be active (IC50 = 252 nM using 14 as a substrate);51 however, there are several structural features which may inhibit its potential developability, in particular the presence of the phosphothioate, which is a known toxicophore.64 In addition to this, it was found that chirality was an important consideration for enzyme inhibition,51

21 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 22 of 57

which, if taken forward, may lead to issues with cost of goods. Similar problems are associated with BrP-LPA, (17),65 which was again homochiral, as well as incorporating an aliphatic bromide, another known toxicophore.64 In addition to the data presented in Table 4, a range of additional assays were conducted including determining the effect 17 has on migration, proliferation, and invasion on HCT 166 cells, and it was shown to inhibit each of these processes.

Perhaps the most significant discovery with 17 was its ability to

significantly decrease tumor volume in in vivo mouse studies; however, the compromise is poor physicochemical properties, as illustrated in Table 8, which may lead to further development problems.66,67 One of the latest lipid-like structures to be disclosed is 20 (CVS16),68 which has been found to reduce the viability of cancer cells and inhibit angiogenesis.68 20, has been classed as a ‘suicide’ inhibitor of ATX and has been found to inhibit angiogenesis in vivo at a relatively high dose of 50 mg/kg using a xenograft melanoma model.68 Unfortunately, this compound exhibited in vivo toxicity, possibly from unspecified side effects as a result of the alkylating potential.68

A range of very extensive patents, by Merck KGA69 and Novartis70 have been disclosed (Table 5) which exhibit the general chemotype of lipophilic tail, core spacer, and acidic head group. This general motif is exemplified in Scheme 3 using PF-8380 (21)52,71 which contains a benoxazolone as the acidic head group, a functionalized piperazine as the spacer, and the dichlorocarbamate moiety as the lipophilic portion. Due to the similarities between the structures claimed it can be postulated that they interact in similar ways in the active site where it is believed that the acidic head group makes essential interactions with one of the Zn2+ ions and the lipophilic tail is accommodated within the hydrophobic pocket. In general, these compounds look appealing from a developability view point, in terms of there are no obvious issues with their functionality, and the calculated physicochemical properties are in

22 ACS Paragon Plus Environment

Page 23 of 57

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

Journal of Medicinal Chemistry

the acceptable range, with 21 having a slightly high molecular weight as illustrated in Table 8.

Scheme 3. Regions of 21 highlighted with the lipophilic tail in red, the core spacer in blue, and the acidic head group in green.

In addition to compound 5, benzoxazolone derivative 21 has been one of the most extensively studied ATX inhibitors and has been used as a tool compound for elucidating the role of LPA in inflammation.52 In terms of binding to the ATX active site, it is believed that the benzoxazolone makes an essential interaction with one of the Zn2+ ions and the long lipophilic portion resides in a tunnel adjacent to the catalytic site (Figure 9).

23 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 24 of 57

Figure 9. Compound 21 docked within ATX using MOE docking software with PDB structure 3NKR. a) Ribbon representation of ATX, colored as follows: SMB1 domain in orange, SMB2 domain in brown, the catalytic domain in cyan, nuclease-like domain in purple, L1 in red, L2 in green, Zn2+ in grey, and 21 in pink. b) Illustrating close proximity of the acidic head group (benzoxazolidinone) to one of the Zn2+ ions and the Thr209 residue. c) Alternative view of 21 within the active site of the enzyme.

In addition to the biological data generated (Table 5), a range of pharmacokinetic and pharmacodynamic studies have also been conducted.52 The pharmacokinetic profile of 21 was generated using both intravenous doses of 1 mg/kg and oral doses of 1 to 11 mg/kg over a 24 h period, and was found to have a mean clearance of 31 mL/min/kg and an effective t1/2 of 1.2 h. In addition to this, it exhibits good oral bioavailability with values ranging from 43-

24 ACS Paragon Plus Environment

Page 25 of 57

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

Journal of Medicinal Chemistry

83% depending on the oral dose (ranging from 1 mg/kg to 100 mg/kg).52 A PK/PD relationship was also determined by measuring plasma LPA, where it was found that LPA levels were rapidly reduced following oral dosing implying that LPA present at basal levels in the plasma is rapidly degraded.52

Compound 21 was also subjected to an in vivo study using the rat air pouch model of inflammation to determine the concentration of LPA both in the plasma and at the site of inflammation. Oral dosing showed concentration of 21 in the plasma from 0.079 µM at 3 mg/kg to 2.68 µM at 100 mg/kg at 4 h, and it was determined that ATX was inhibited up to 95% in both plasma and air pouch at a dose of 100 mg/kg. To date, no further work has been reported with 21 and no data is available regarding its progression into clinical trials.

In contrast to the analogues of 21, an alternative and novel non-lipid chemotype has been disclosed by Amira72 and PharmAkea,73 which involve the removal of the lipophilic tail portion to produce smaller structures (Table 6).72-74 In addition to the data included in Table 6, Amira and PharmAkea have claimed activity in a range of in vivo and ex vivo assays which include: human serum ATX assay, human whole blood ATX, mouse air pouch assay, collagen induced arthritis, rat model of neuropathic pain, a lung metastasis model, and a mouse carbon tetrachloride-induced fibrosis model, the results of which are not disclosed within the individual patents.

Recent crystallographic data by PharmAkea indicates that compounds such as Pat-078 (25) bind remote from the catalytic site and occupy a site near the hydrophobic channel, as illustrated in Figure 10.18c This alternative binding mode can also be related to the inconsistencies observed when using unnatural substrates such as 9 in biological assays, as

25 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 26 of 57

these assays are unable to detect activity when the inhibitor compound is remote from the catalytic site.18c

Figure 10. Compound 25 crystallized within ATX illustrating a binding site remote from the catalytic site. a) Ribbon representation of ATX, colored as follows: SMB1 domain in orange, SMB2 domain in brown, catalytic domain in cyan, nuclease-like domain in purple, L1 in red, L2 in green, Zn2+ in grey, and 25 in pink. b) Structure of 25. c) and d) Expanded section of the crystal structure to indicate remote binding of 25 from the catalytic site.

From a developability perspective these compounds, while reasonable in molecular weight and polar surface area (Table 8), they do have relatively high lipophilicity, which may result in attrition further in the development process.75

26 ACS Paragon Plus Environment

Page 27 of 57

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

Journal of Medicinal Chemistry

In addition to the templates described previously there are a range of compounds reported that do not fall into any of these categories (Table 7). Some, however, follow a similar chemotype with the three main features remaining; an acidic moiety, core, and lipophilic tail. Hoffman La-Roche has reported a range of compounds, with structures similar to 29, that have ATX IC50 values in the low nanomolar range.76 However, again from considering the physicochemical properties of this series these compounds are on the higher end of molecular weight, which may result in issues downstream.

Eli Lilly77 have also developed a range of nanomolar compounds, with structures such as 27, which appear promising from a developability perspective. Again, no overt Zn2+ binding group is present, potentially indicating that these compounds do not interact with the catalytic site.

As discussed above, the boronic acid 5 has become an important tool compound due to the extensive structural data associated with this molecule. However, the presence of the boronic acid functionality may impact further development due to the potential toxicity issues associated with this motif.78 Compound 31,79 reported from our own laboratories, suggests the ability to inhibit the enzyme without the presence of an acidic moiety, this however did lead to a significant reduction in solubility (7.5 µg/mL).79 Similarly, a range of compounds has been reported that lack the acid moiety, exemplars of which are indicated in Table 7. These include compounds such as 28,56,61 which, in addition to the data provided in Table 7, has been analyzed using a range of ATX assays, including using 12 as a substrate where it displayed an IC50 = 31.4 nM. Interestingly, no inhibition was observed when using the unnatural substrate 11 which can be potentially attributed to 28 binding in a pocket remote from the catalytic site as discussed above.54,56 Additional biological profiling for 28 included

27 ACS Paragon Plus Environment

Journal of Medicinal Chemistry

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

Page 28 of 57

testing against melanoma cell metastasis in B16-F10 cells, where it was found to significantly reduce the number of lung nodules, and analysis for off-target effects using a psychoactive drug screening panel by the National Institution of Mental Health, where no off target effects were observed.54,56

Following this, Galapagos have developed a range of compounds with structures similar to 32,80 which again lack the acidic group. In addition to the data detailed in Table 7, 32 has also been analyzed in an assay using 12 as a substrate (IC50 = 0.01-100 nM), ATX human whole blood assay (IC50 of 100-500 nM), and rat and mouse whole blood assay (IC50 = >500-1000 nM). In vivo studies included tobacco smoke model, lung inflammatory cells recruitment evaluation, a bleomycin-induced model of IPF, pharmacokinetic studies using rodent and dog models, and ADME assays. However, no data was given regarding these assays. Table 8. Calculated physicochemical values for the exemplar compounds. Compound

MW

Log P

Log D

PSA

5

480.3

2.7

2.7

107.3

16

506.9

-4.2

-6.6

71.0

17

487.4

5.3

2.9

106.9

18

648.8

4.0

4.2

141.0

19

368.5

6.5

4.1

60.4

20

494.3

3.9

3.9

34.1

21

478.3

3.6

3.6

88.2

22

533.4

3.3

3.1

91.4

23

440.3

2.5

2.5

100.2

28 ACS Paragon Plus Environment

Page 29 of 57

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

Journal of Medicinal Chemistry

24

398.9

4.8

1.5

75.8

25

431.8

5.8

2.6

81.7

26

432.5

3.7

3.7

59.0

27

438.5

2.1

2.0

112.6

28

433.3

3.3

3.3

75.7

29

531.5

2.6

2.6

119.2

30

521.4

4.0

2.2

104.7

31

342.4

4.0

4.0

40.6

32

588.7

3.6

2.9

105.5

Color coded according to correlation with Gleeson67 (MW