Discovery and optimization of boronic acid based inhibitors of Autotaxin

Article abstract of DOI:10.1021/jm1005012

Autotaxin (ATX) is an extracellular enzyme that hydrolyzes lysophosphatidylcholine (LPC) to produce the lipid mediator lysophosphatidic acid (LPA). The ATX-LPA signaling axis has been implicated in diverse physiological and pathological processes, including vascular development, inflammation, fibrotic disease, and tumor progression. Therefore, targeting ATX with small molecule inhibitors is an attractive therapeutic strategy. We recently reported that 2,4-thiazolidinediones inhibit ATX activity in the micromolar range. Interestingly, inhibitory potency was dramatically increased by introduction of a boronic acid moiety, designed to target the active site threonine in ATX. Here we report on the discovery and further optimization of boronic acid based ATX inhibitors. The most potent of these compounds inhibits ATX-mediated LPC hydrolysis in the nanomolar range (IC₅₀ = 6 nM). The finding that ATX can be targeted by boronic acids may aid the development of ATX inhibitors for therapeutic use.

Full text of DOI:10.1021/jm1005012

4958 J. Med. Chem. 2010, 53, 4958–4967
DOI: 10.1021/jm1005012
Discovery and Optimization of Boronic Acid Based Inhibitors of Autotaxin
Harald M. H. G. Albers,†,‡ Laurens A. van Meeteren,†,^ David A. Egan, ,# Erica W. van Tilburg,†,¥
Wouter H. Moolenaar,†,§ and Huib Ovaa*,†,‡
†Division of Cell Biology, ‡Netherlands Proteomics Centre, §Centre for Biomedical Genetics, and Division of Molecular Carcinogenesis,
The Netherlands Cancer Institute, 1066 CX Amsterdam, The Netherlands. ^Current address: Department of Molecular Cell Biology, Leiden
University Medical Center, Leiden, The Netherlands. #Current address: Department of Cell Biology, Utrecht University Medical Center, Utrecht,
The Netherlands. ¥Current address: Department of Nuclear Medicine, Erasmus Medical Center, Rotterdam, The Netherlands.
Received April 23, 2010
Autotaxin (ATX) is an extracellular enzyme that hydrolyzes lysophosphatidylcholine (LPC) to produce
the lipid mediator lysophosphatidic acid (LPA). The ATX-LPA signaling axis has been implicated in
diverse physiological and pathological processes, including vascular development, inflammation, fibrotic
disease, and tumor progression. Therefore, targeting ATX with small molecule inhibitors is an attractive
therapeutic strategy. We recently reported that 2,4-thiazolidinediones inhibit ATX activity in the
micromolar range. Interestingly, inhibitory potency was dramatically increased by introduction of a
boronic acid moiety, designed to target the active site threonine in ATX. Here we report on the discovery
and further optimization of boronic acid based ATX inhibitors. The most potent of these compounds
inhibits ATX-mediated LPC hydrolysis in the nanomolar range (IC50 = 6 nM). The finding that ATX can
be targeted by boronic acids may aid the development of ATX inhibitors for therapeutic use.
Introduction
certain conditions,15 various synthetic phospholipid analogues
have been explored as ATX inhibitors.16-20 However, lipid-
based inhibitors have the disadvantage that they could act as
agonists or antagonists for any of the LPA/S1P receptors,
thereby resulting in an effect opposite of the one intended. In
addition, non-lipid ATX inhibitors have been identified, but
their inhibitory potential is low (IC50 g 1 μM).21-23
Autotaxin (ATXa or NPP2), originally isolated as an
autocrine motility factor from melanoma cells, belongs to
the ecto-nucleotide pyrophosphatase and phosphodiesterase
(NPP) family.1-3 This extracellular enzyme acts as a lysopho-
pholipase D, hydrolyzinglysophosphatidylcholine (LPC) into
the lipid mediator lysophosphatidic acid (LPA), as depicted in
Scheme 1.4,5 Hydrolytic activity of ATX originates from a
threonine (T210) residue in the active site.1 LPA activates
specific G-protein-coupled receptors and thereby stimulates
the migration, proliferation, and survival of many cell types.6
The ATX-LPA axis has a vital role in vascular develop-
ment.7,8 Furthermore, it has been implicated in various patho-
logies including tumor progression9 and metastasis,10
inflammation,11 and fibrotic disease.12
Given its role in human disease, the ATX-LPA axis is an
obvious target for therapy. The fact that ATX is an extra-
cellular enzyme makes it even more attractive as a drug target.
Since there are at least six distinct LPA receptors, direct
targeting of LPA receptors seems to be a less attractive
strategy.13,14 On the basis of the initial discovery that ATX
is inhibited by LPA and sphingosine 1-phosphate (S1P) under
Recently we reported the non-lipid ATX inhibitor HA130,24
here named compound 72 (Table 3), that rapidly lowers plasma
LPA levels upon intravenous injection in mice. Using inhibitor
72, we found that the turnover of circulating LPA is much faster
than expected, showing the usefulness of ATX inhibitors as
tools to elucidate the role of ATX and LPA in vivo.
Here we describe the discovery and further optimization of
compound 72. We screened ∼40000 small molecules and
identified thiazolidinediones as ATX inhibitors, which proved
to be suitable to further chemical optimization. In particular,
we optimized these molecules by first introducing small sys-
tematic variations in the structure of a screening hit, followed
by the introduction of more random variations. Finally, by
targeting the T210 oxygen nucleophile in ATX through a
boronic acid, the potency of the original thiazolidinedione
screening hit was increased over 400-fold (IC50 = 6 nM).
*To whom correspondence should be addressed. Address: Division of
Cell Biology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066
CX Amsterdam, The Netherlands. Phone: þ31-20-5121979. Fax: þ31-
20-5122029. E-mail: h.ovaa@nki.nl.
Results
ATX Inhibitor Screen. To identify ATX inhibitors, a small
molecule screen was carried out. We first used a collection of
17 500 molecules from the National Cancer Institute (NCI)
to optimize our screening protocol. This NCI library was
initially screened using CPF4, an ATX activity reporter
aAbbreviations: ABTS, 2,20-azino-bis(3-ethylbenzothiazoline-6-sul-
fonic acid); ATX, autotaxin; bis-pNPP, bis-p-nitrophenyl phosphate;
HRP, horseradish peroxidase; HTS, high-throughput screening; FRET,
€
Forster resonance energy transfer; LC-MS, liquid chromatogra-
phy-mass spectrometry; LPA, lysophosphatidic acid; LPC, lysopho-
sphatidylcholine; NCI, National Cancer Institute; NPP, nucleotide
pyrophosphatase and phosphodiesterase; lysoPLD, lysophospholipase
D; PI, percentage inhibition; RA, residual activity; SAR, structure-
activity relationship; S1P, sphingosine 1-phosphate.
€
based on Forster resonance energy transfer (FRET)
(Figure 1A).15,25 The phosphodiester bond which links the
coumarin and fluorescein moieties in CPF4 is hydrolyzed by
r
pubs.acs.org/jmc
Published on Web 06/10/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4959
Scheme 1. Hydrolysis of LPC by ATX into LPA and Cholinea
a The lysoPLD reaction is catalyzed by a threonine oxygen nucleophile (T210). Newly produced LPA triggers subsequent biological events via
activation of specific G-protein-coupled receptors.
ATX, resulting in the loss of FRET. Since we have shown
that LPA can inhibit ATX activity in this assay, it was used as
a positive control for ATX inhibition.15 Approximately 250
active molecules were retested in a second assay using bis-p-
nitrophenyl phosphate (bis-pNPP) as reporter substrate
(Figure 1B).15 In Supporting Information Figure S1 the
results of both assays are compared. Only a few compounds
remained after retesting in the bis-pNPP assay. Since the
CPF4 assay resulted in many false positives, we decided to
reverse the assay order. The NCI library was tested in the
bis-pNPP assay using CPF4 as confirmation tool which
led to reproducible hits corresponding to the findings of
Saunders et al.26 Top three actives of the NCI screen can be
found in Supporting Information Figure S2.
Chemistry. We next explored how to improve the potency
of 2,4-thiazolidinedione 2 as an ATX inhibitor. We designed
the convergent synthetic route toward inhibitor 2 depicted in
Scheme 2. The synthesis requires benzoic acid 7 which can be
obtained via a one-pot-synthesis. First, vanillin is O-alky-
lated with methyl 4-(bromomethyl)benzoate.27 The resulting
benzoate is hydrolyzed to give benzoic acid 7. To afford the
precursor 8, 2,4-thiazolidinedione was dissolved in DMF
and N-alkylated with 4-fluorobenzyl chloride in the presence
of sodium hydride. Finally, monosubstituted thiazolane-2,4-
dione 8 was reacted via a Knoevenagel condensation with
benzoic acid 7.28 Z-Isomer 2 precipitated during reaction,
and washing the precipitate with ethanol resulted in a
homogeneous product.
After validation of our screening protocol, we used a
commercial library (SPECS) consisting of 23 000 small mo-
lecules with predicted druglike properties. For active mole-
cules in this screen the percentage inhibition (PI) at 5 μM and
IC50 values (bis-pNPP assay) are shown in Figure 2. These
actives were confirmed by the CPF4 assay (Figure 2). Among
the confirmed active molecules were several 2,4-thiazolidi-
nediones (1, 2, 4, and 5), a pyrano pyrazole (3), and a
benzothiazole (6). Because the 2,4-thiazolidinediones are
well represented among the positive hits and because of their
amenability to fast chemical diversification, the most potent
2,4-thiazolidinedione 2 (IC50 = 56 nM) was selected for
further optimization.
The synthetic route toward 2 appeared to be applicable for
the fast parallel synthesis and purification of many analogues
without the need for chromatography.
Systematic Optimization. Inhibitor 2 contains two side
chains attached to the thiazolidinedione core, a benzyl and a
benzylidene moiety (Scheme 2). To explore structure-activ-
ity relationship (SAR), we first generated a small library
before embarking on a bigger effort. This resulted in the
molecules listed in Table 1. LPA was used as a control
because it inhibits the ATX-mediated hydrolysis of bis-
pNPP by ATX (IC50 = 422 nM).15
The best series of inhibitors are the benzoic acids (2 and
9-11), where the fluorinated compound 2 has the highest

4960 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
Albers et al.
€
Figure 1. Three ATX assays used for screening and validation of ATX inhibitors. (A) CPF4 assay. This assay is based on Forster resonance
energy transfer (FRET) between coumarin and fluorescein moieties of CPF4. (B) Bis-pNPP assay. In this assay bis-p-nitrophenyl phosphate
(bis-pNPP) is hydrolyzed by ATX into the chromophore p-nitrophenol. (C) Choline release assay. The physiological substrate of ATX, LPC, is
hydrolyzed by ATX to give LPA and choline. The release of choline can be detected in a two-step enzymatic coloring reaction.
activity (IC50 = 56 nM) in the bis-pNPP assay (Table 1).
Replacement of the 4-fluorine atom (2) for hydrogen (9),
4-nitro- (10), or a tert-butyl group (11) reduced the potency
of the inhibitor with increasing size of the functional group.
With replacement of the carboxylic acid in molecules 2 and
9-11 by a methyl ester (12-15), potency is largely lost.
Omitting the carboxylic acid (16-19) has the same effect.
With removal of the methoxy group (R2) in molecules
16-19, activity is regained (compare 16-19 with 20-23).
SAR analysis showed that the optimal combination of
groups would be R1 = -OCH2-Ph-COOH, R2 = H,
and R3 = F.
Encouraged by the SAR results suggesting that the meth-
oxy and carboxylic acid groups are important moieties,
molecules 28-30 were synthesized (Figure 3A). Molecules
2, 9-11, and 28-30 were validated in the physiologically
more relevant LPC hydrolysis (choline release) assay.17 The
ATX-mediated release of choline from LPC is detected by a
two-step enzymatic colorimetric reaction (Figure 1C). In the
first step choline is oxidized by choline oxidase into betaine
and hydrogenperoxide. Horseradish peroxidase (HRP) then
consumes hydrogen peroxide to oxidize 2,20-azino-bis(3-
ethylbenzothiazoline-6-sulfonic acid) (ABTS) to a radical
cation species which absorbs at 405 nm.
The data obtained from the choline release assay are
summarized in Figure 3A. The IC50 values obtained were
some 10-fold higher than observed in with bis-pNPP as a
substrate. It is noteworthy that compound 10 is incapable of
fully inhibiting ATX at high concentrations (Figure 3B).
This residual ATX activity for 10 is about 60%. However, 10
shows inhibition at a lower concentration than 11 which was
also observed in the bis-pNPP assay. LPA did not inhibit
ATX in the choline release assay as observed by Ferry et al.19
The established SAR was confirmed by the more active
molecule 28 (IC50 = 1.63 μM) in which the methoxy group in
2 (IC50 = 2.50 μM) is omitted. Changing the carboxylic acid
of 2 from para to the meta position results in the more potent
inhibitor 29 (IC50 = 2.05 μM) but less potent than 28.
Changing the carboxylic acid in 28 from para to meta (30,
IC50 = 1.07 μM) results in the most active molecule in this
systematic optimization with a 2.5-fold reduction in IC50
value compared to 2 and also to a further reduction in
residual ATX activity (from 35% to 7%) in the choline
release assay.
Variation in the Benzyl Moiety. Since our initial systematic
optimization of 2 provided good structure-activity correla-
tions, we varied either the benzyl or the benzylidene moiety
(Scheme 2). In this approach, either the benzyl or the
benzylidene moiety corresponds to the screening hit 2 while
the other part is alternated. First, the effect of changes in the
benzyl part of the molecule was investigated by introducing
halogens. The percentage inhibition (PI) values at 5 μM
31-40, as measured by the choline release assay, are listed in
Table 2. The most potent inhibitor was bromo compound 39
(PI = 73%). However, its IC50 (2.80 μM) is similar to that of
screening hit 2 (IC50 = 2.50 μM). Other substituents that
were explored did not result in significant improvements.
Variation in the Benzylidene Moiety. Next, variations were
incorporated into the benzylidene moiety, resulting in mole-
cules shown in Supporting Information Table S1. In contrast
to the variation in the benzyl moiety, few molecules were
active; only 30% of the compounds tested showed inhibition
(PImax = 5%). This indicates that variations in the benzyli-
dene moiety of screening hit 2 are not well tolerated.
Boronic Acid Introduction and Optimization. Because the
benzyl and benzylidene moiety variations did not imp-
rove inhibition, we modified inhibitor 30 that resulted from
the systematic optimization. Replacing the carboxylic acid
in molecule 30 by a boronic acid resulted in molecule 72
(Table 3). We recently characterized 72 as a potent ATX inhi-
bitor both in vitro and in vivo.24 We reasoned that the car-
boxylic acid moiety in molecule 30 could act as a phosphate

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4961
Figure 2. Molecules active as ATX inhibitors discovered by high-throughput screening. Percentage inhibition (PI) at 5 μM is based on the
bis-pNPP or CPF4 assay. IC50 values are obtained using the bis-pNPP assay. LPA was used as a control for ATX inhibition.
Scheme 2. Convergent Synthetic Route toward 2,4-Thiazolidinedione 2a
a Reagents and conditions: (i) KOH, DMSO, room temp, 30 min; (ii) NaOH, DMSO/H2O (1:3, v/v), reflux, 4 h, 91%; (iii) NaH, DMF, room temp,
22 h, 74%; (iv) piperidine, EtOH, reflux, 20 h, 63%.

4962 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
Albers et al.
Table 1. Structure-Activity Data for Molecules of the Systematic
Optimizationa
Figure 3. Validation of screening hit 2 and the benzoic acids of the
systematic optimization in the choline release assay. (A) Structure-
activity data of the systematic optimization. Residual ATX activity
is abbreviated as RA (%). (B) Dose response curves for the benzoic
acids.
a IC50 Values and Residual ATX Activity (RA) of the Synthesized
2,4-Thiazolidinedione Derivatives Using the Artificial ATX Substrate
bis-pNPP (na: not active at 5 μM).
and aryl boronic acids have the intrinsic ability to react with
hydrogen peroxide.31-33 We investigated the effect of the
inhibitors 72, 73, 75, and 76 on the coloring reaction, and no
effect was observed (Supporting Information Figure S3).
Therefore, the choline release assay is a valid way to test the
synthesized compounds here on ATX inhibition.
Mode of Inhibition. For inhibitors 2, 30, 72, and 73, the
mode of inhibition was determined from a Lineweaver-
Burk plot (Supporting Information Figure S4). Inhibitors 2
and 30 inhibit in a competitive manner using LPC as sub-
strate. Kinetic analysis for boronic acids 72 and 73 revealed a
mixed type inhibition.
Because these inhibitors contain a Michael acceptor, it
could be that these inhibitors react irreversibly with ATX (or
other enzymes), which is undesirable. When ATX is incu-
bated with 5 μM inhibitor 2, 30, 72, or 73 for 20 min at 37 ꢀC
and these solutions are then washed with ethyl acetate, ATX
activity is restored up to 97% of its original activity in the
choline release assay (Supporting Information Figure S5).
This indicates that these inhibitors reversible bind to ATX.
mimic that binds near or at the active site threonine
(T210). Replacing the carboxylic acid in molecule 30 by a
boronic acid could well be a good ATX targeting strategy.
This strategy was inspired by the proteasome inhibitor
bortezomib, which targets the N-terminal threonine oxygen
nucleophile in the proteasome through a boronic acid.29
Furthermore, boronic acids have been reported to inhibit
β-lactamases through targeting of the active site serine
residue.30
The boronic acid modification resulted in a 100-fold more
potent inhibitor (IC50 = 28 nM) compared to screening hit 2
(IC50 = 2.50 μM). When the position of the boronic acid was
changed from meta (72) to the para (73) position, potency
increased by another 5-fold (IC50 = 5.7 nM) while the potency
of the ortho boronic acid 74 dropped (IC50 > 5.00 μM). Next,
the 3-benzyloxyboronic acids analogues 75-77 were synthe-
sized. The meta boronic acid (75) provedtobemorepotentthan
the para (76), andtheorthoboronicacid(77) has a low potency.
Effect of Inhibitors on Readout Choline Release Assay. An
inherent danger of the choline release assay is that small
molecules can interfere with the readout by inhibiting the
enzymes (horseradish peroxidase or choline oxidase) used in
the coloring reaction, resulting in false positives. Another
way to interfere with the choline release assay is that com-
pounds can react with the ABTS or hydrogen peroxide
generated during the coloring reaction. The latter could be
an issue for the boronic acids reported here because aliphatic
Discussion and Conclusions
This study shows that ATX can be targeted efficiently by
boronic acids. We discovered thiazolidinediones as potent
non-lipid ATX inhibitors. Replacing the carboxylic acid by a
boronic acid in thiazolidinedione screening hit 2 resulted in a
440-fold more active inhibitor (IC50 = 2.50 μM f 6 nM).

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4963
Table 2. Derivatives of Screening Hit 2 with a Structural Variation in the Benzyl Moietya
a Percentage inhibition (PI, %) has been determined in the choline release assay at 5 μM. Molecules that show more than 50% inhibition are in bold
(compound/PI).
Table 3. Potent ATX Inhibitors of the Boronic Acid Optimizationa
a For the boronic acids percentage inhibition (PI, %) at 5 μM and IC50 values have been determined in the choline release assay (PI/IC50).
We handed three different approaches for optimization of
screening hit 2. The first is the systematic optimization of 2,
which led to a 2.5-fold increase in potency in the choline
release assay. The second is a randomized approach, changing
separately the benzyl and the benzylidene parts of the molecule.
This did not result in significantly more potent molecules.
Finally, we replaced the carboxylic acid in screening hit 2 by a
boronic acid. Our rationale was that the carboxylic acid could
function as a phosphate mimic and thereby bind near or at
the T210 oxygen nucleophile. In that case, the T210 oxygen

4964 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
Albers et al.
Figure 4. Our hypothesis on the binding of boronic acid 72 or 73 with the T210 oxygen nucleophile in the ATX active site.
nucleophile in ATX can be targeted via a boronic acid. A
similar strategy has proven successful for the proteasome
inhibitor bortezomib which binds to threonine oxygen nu-
cleophile in the proteasome active site through the boronic
acid moiety.29 Replacing the carboxylic acid moiety in screen-
ing hit 2 (IC50 = 2.50 μM) by a boronic acid resulted in mole-
cule 73 (IC50 = 6 nM), which is over 400-fold more potent
than inhibitor 2. These results demonstrate that ATX activity
can be targeted by boronic acids.
Next to their increased affinity for ATX, the boronic acids
are also expected to improve selectivity over hydrolytic en-
zymes that depend on a sulfur (cysteine) nucleophile, as is
commonly found in phosphate ester hydrolyzing enzymes.
Kinetic analysis showed that inhibitors 2 and 30 inhibit in a
competitive manner indicating that they bind at the active site
of ATX. The acid moiety of these inhibitors likely mimics the
phosphate group, facilitating inhibition. Boronic acids 72 and
73 show a mixed-type inhibition. It is expected that the
boronic acids have the same binding site as their carboxylic
acid equivalents 2 and 30 based on their structural similarity.
Binding of the boronic acid to the T210 oxygen nucleophile
likely results in the Lineweaver-Burk plot observed. The
washout experiment reveals that inhibitors 2, 30, 72, and 73
all bind to ATX in a reversible manner. Figure 4 shows how
the boronic acid inhibitors may bind to the T210 oxygen
nucleophile of ATX.
serum-free medium was added and cells were allowed to secrete
His-tagged ATX into the culture medium for 48 h. Medium
was collected and ATX was purified using TALON-affinity
beads (Clontech) as described. Imidazole was removed by
dialysis against Tris-buffered saline (140 mM NaCl, 5 mM
KCl, 1 mM CaCl2, 1 mM MgCl2, and 50 mM Trizma, pH 8.0).
Purity was verified by SDS-PAGE and Coomassie blue
staining.
Bis-pNPP Assay.15 ATX activity toward bis-pNPP was deter-
mined as follows. In an opaque flat-bottom 384-well plate an
amount of 2 μL of DMSO containing inhibitor was added to
24 μL of recombinant ATX (∼40 nM) in Tris-buffered saline
(140 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and
50 mM Tris-HCl, pH 7.8) which contained albumin from bovine
serum fatty acid free (BSA-FAF) (0.2 mg mL-1). Finally, 24 μL
of bis-pNPP (2 mM) was added to each well and the plate was
incubated for 3 h at room temperature. This mixture with
DMSO alone was used as a control. For each inhibitor 10
concentrations were measured covering a range of 0.025 to
100 μM to determine IC50 values. Percentage inhibition was
determined for a final concentration of 5 μM of inhibitor.
Absorbance was measured in a Perkin-Elmer Envision plate
reader (λ = 405 nm). The absorbance at 0 h was used to correct
for molecules that absorb at 405 nm. Data were analyzed using
Graphpad Prism software. IC50 values and percentage inhibi-
tion were determined in three independent experiments for each
inhibitor. Percentage of residual activity (RA) is given as bottom
value for curve fit.
CPF4 Assay.15 This assay was carried out in the same manner
as the bis-pNPP assay using a substrate concentration of
2 μM. Fluorescence was monitored in a BMG Fluorstar 96-
well plate reader (excitation at 355 nm, emission at 460 and
520 nm).
In summary, we have identified 2,4-thiazolidinediones as
ATX inhibitors and have found that targeting ATX with a
boronic acid moiety resulted in a >400-fold increase in
potency. This strategy to target ATX with a boronic acid
should assist the development of future ATX inhibitors.
Choline Release Assay.17 ATX activity using LPC (18:1) as
substrate was determined as follows. In an opaque flat-bottom
96-wells plate (Greiner) 1 μL of DMSO containing inhibitor was
added to 49 μL of recombinant ATX (∼40 nM) in Tris-HCl
buffer (0.01% Triton X-100 and 50 mM Tris-HCl, pH 7.4).
Finally, 50 μL of 80 μM LPC (18:1) in Tris-HCl buffer (10 mM
MgCl2, 10 mM CaCl2, 0.01% Triton X-100, and 50 mM Tris-
HCl, pH 7.4) was added to each well and the plate was incubated
at 37 ꢀC. The above-described mixture with DMSO alone was
used as a control. LPC without ATX was taken as control for
autohydrolysis of LPC. For each inhibitor 10 concentrations
were measured covering a range of 0.01 to 30 μM to determine
IC50 values. Percentage inhibition was determined for a final
concentration of 5 μM inhibitor. After 3 h of incubation, 50 μL
of ABTS (2 mM) and horseradish peroxidase (5 U mL-1) was
added to 50 μL of the reaction mixture and absorbance was
measured and used to correct for absorbance of the molecules.
Finally, 50 μL of choline oxidase (5 U mL-1) in Tris-HCl (0.01%
Triton X-100 and 50 mM Tris-HCl, pH 7.4) was added for
colorimetric reaction. Absorbance was measured in a Perkin-Elmer
Experimental Section
Chemicals and Enzymes. Small molecule libraries were ob-
tained from the National Cancer Institute (NCI) and purchased
from SPECS, Delft, The Netherlands. NCI compounds NSC-
101794, NSC148368, and NSC48300 are abbreviated in the text as
A, B, and C respectively. SPECS compounds AN-989/40746701,
AN-988/40680277, AJ-292/40674401, AN-989/41697652, AN-989/
41697944, and AQ-088/4201464 are abbreviated in the text as 1, 2, 3,
4, 5, and 6, respectively. 1-Oleoyl-2-hydroxy-sn-glycero-3-phospho-
choline (LPC, 18:1) was purchased from Avanti Polar Lipids.
Oleoyl-L-R-lysophosphatidic acid sodium salt (LPA, 18:1), horse-
radish peroxidase (HRP), choline oxidase, and all other chemicals
were obtained from Sigma-Aldrich and used as received. TALON-
affinity beads were from Clontech, and opaque flat-bottom 96-well
plates were from Greiner.
Recombinant ATX. HEK293 cells were transfected with the
pcDNA3 vector containing a 6xHis-tagged human teratocarci-
noma ATX sequence. After transfection, cells were washed and

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4965
Envision plate reader (λ = 405 nm). Data were analyzed using
Graphpad Prism software.
temperature and was stirred for 4 h. Then the mixture was
poured into ice-water (250 mL), and hexane (100 mL) was
added. After a night at 4 ꢀC the precipitated crystals were filtered
and dried to give a white solid.
In addition, the effect of the inhibitors on the enzymatic
coloring reaction was investigated using 40 μM choline at
30 μM inhibitor using the above-described coloring reagents.
No inhibition of the enzymatic reaction by inhibitors was
observed.
Washout of Inhibitors. ATX (∼40 nM) in the same buffer used
for the choline release assay was incubated with (5 μM) and
without inhibitors at 37 ꢀC. After 20 min the incubation
mixtures were washed out with ethyl acetate with 9 times the
volume of the ATX solution. The ATX activity of these solu-
tions was determined using the choline release assay.
Chemistry. All chemicals were obtained from Sigma-Aldrich
and used without further purification. Dry N,N-dimethylfor-
mamide (DMF) from Biosolve was obtained by treatment with
Yield: 74%. 1HNMR:δ=7.34-7.14 (m, 4H, y), 4.65 (s, 2H, z),
4.26 (s, 2H, b). MS: m/z [M þ H]þ calcd 226.03, obsd 226.04. LC:
tR = 8.22.
General Method for the Synthesis of (2,4-Dioxo-1,3-thiazolan-
5-yliden)benzoic Acids and Benzeneboronic Acids. To a solution
of 1,3-thiazolane-2,4-dione 8 (0.293 mmol) in ethanol (2.5
mL), piperidine (12 μL, 0.207 mmol) and the appropriate
aldehyde (0.352 mmol) were added, and the solution was
refluxed for 22 h. When the mixture was cooled to room
temperature, the product precipitated out of solution. Centri-
fugation and washing with ethanol gave a homogeneous
compound.
˚
molecular sieves (4 A). Analytical thin layer chromatography
was performed on aluminum sheets precoated with silica gel 60
F
254. Column chromatography was carried out on silica gel
˚
(0.035-0.070, 90 A, Acros).
For isolation by centrifugation a Heraeus Multifuge 3S-R
centrifuge was used. Products were spun at 4400g at 4 ꢀC for
5 min. Nuclear magnetic resonance spectra (1H NMR and
COSY) were determined in deuterated dimethyl sulfoxide
(DMSO-d6) using a Bruker ARX 400 spectrometer (1H, 400
MHz) at 298 K, unless indicated otherwise. Peak shapes in the
NMR spectra are indicated with the symbols “d” (doublet), “dd”
(double doublet), “s” (singlet), “bs” (broad singlet), and “m”
(multiplet). Chemical shifts (δ) are given in ppm and coupling
constants J in Hz. DMSO (δ = 2.50 ppm) was used as internal
reference.
4-[(4-{[3-(4-Fluorobenzyl)-2,4-dioxo-1,3-thiazolan-5-yliden]-
methyl}-2-methoxyphenoxy)methyl]benzoic Acid (2).
1
Yield: 63%. COSY was used to assign chemical shifts. H
NMR: δ = 12.99 (bs, 1H, OH), 7.97 (d, J 8.3, 2H, j), 7.92 (s,
1H, c), 7.56 (d, J 8.3, 2H, i), 7.38-7.16 (m, 7H, d þ g þ y),
5.27 (s, 2H, h), 4.82 (s, 2H, z), 3.84 (s, 3H, f). MS: m/z[M- H]-
calcd 492.10, obsd 492.10. LC: tR = 10.54.
3-[(4-{[3-(4-Fluorobenzyl)-2,4-dioxo-1,3-thiazolan-5-yliden]-
methyl}phenoxy)methyl]benzeneboronic Acid (72).
LC-MS measurements were performed on a system equip-
ped with a Waters 2795 separation module (Alliance HT),
Waters 2996 photodiode array detector (190-750 nm), Waters
Alltima C18 column (2.1 mm ꢀ 100 mm), and an LCT ortho-
gonal acceleration time of flight mass spectrometer. Samples
were run at a flow rate of 0.40 mL min-1 using gradient elution
(water/acetonitrile/formic acid) from 950/50/10 (v/v/v) to 50/
950/10 (v/v/v). Retention times (tR) are given in minutes. Purity
of all compounds was verified by LC-MS.
The purity of all tested compounds as determined by LC-MS
analyses was greater than 95%.
Yield: 64%. 1H NMR: δ = 8.08 (s, 2H, OH), 7.92 (s, 1H, c),
7.87 (s, 1H, i2), 7.76 (d, J 7.4, 1H, i1), 7.60 (d, J 7.1, 2H, d),
7.50-7.16 (m, 10H, d þ g þ k þ j þ x þ y), 5.18 (s, 2H, h),
4.82 (s, 2H, z). MS: m/z [M þH]þ calcd 464.11, obsd 464.25.
LC: tR = 10.77.
4-[(4-Formyl-2-methoxyphenoxy)methyl]benzoic Acid (7). To
a solution of vanillin (1.41 g, 9.24 mmol) and KOH (0.640 g,
11.4 mmol) in DMSO (10 mL), methyl 4-(bromomethyl)-
benzoate (2.00 g, 8.73 mmol) was added. The reaction mixture
was stirred at room temperature. After 30 min, water (50 mL)
was added. The solution was heated at 393 K for 2 h.
Subsequently, 1 M NaOH (aq) (10 mL) was added under
reflux until the solution became clear. Finally, the reaction
mixture was poured into water and was acidified with 1 M
HCl to pH 2. The precipitate was isolated by centrifugation,
washed with water, and lyophilized resulting in the title
compound.
4-[(4-{[3-(4-Fluorobenzyl)-2,4-dioxo-1,3-thiazolan-5-yliden]-
methyl}phenoxy)methyl]benzeneboronic Acid (73).
Yield: 81%. 1H NMR: δ = 8.05 (s, 2H, OH), 7.93 (s, 1H, c),
7.81(d, J 8.0, 2H, j), 7.61 (d, J 8.9, 2H, d), 7.42 (d, J 8.3, 2H,
i), 7.41-7.16 (m, 6H, g þ y), 5.21 (s, 2H, h), 4.82 (s, 2H, z).
MS: m/z [M þ H]þ calcd 464.11, obsd 464.21. LC: tR
=
10.45.
2-[(4-{[3-(4-Fluorobenzyl)-2,4-dioxo-1,3-thiazolan-5-yliden]-
methyl}phenoxy)methyl]benzeneboronic Acid (74).
Yield: 91%. 1H NMR: δ = 12.99 (bs, 1H, OH), 9.85 (s, 1H, c),
7.98 (d, J 8.3, 2H, j), 7.58 (d, J 8.3, 2H, i), 7.55 (dd, J 1.9 and 8.3,
1H, d), 7.44 (d, J 1.8, 1H, e), 7.26 (d, J 8.3, 1H, g), 5.32 (s, 2H, h),
3.86 (s, 3H, f). MS: m/z [M þ H]þ calcd 287.09, obsd 287.11. LC:
tR = 7.75.
Synthesis of 3-(4-Fluorobenzyl)-1,3-thiazolane-2,4-dione (8).
To a cooled solution (0 ꢀC) of thiazolidine-2,4-dione (5.87 g,
50 mmol) in DMF (100 mL), sodium hydride (60% in oil, 1.8 g,
45 mmol) was added. A solution of 1-(chloromethyl)-4-fluor-
obenzene (4.3 mL, 36.8 mmol) in DMF (25 mL) was added to
the reaction mixture. The mixture was allowed to warm to room
Yield: 48%. 1H NMR: δ = 8.10 (s, 2H, OH), 7.93 (s, 1H, c),
7.62-7.13 (m, 12H, d þ g þ i þ j1þ2 þ k þ y), 5.32 (s, 2H, h),
4.82 (s, 2H, z). MS: m/z [M þ H]þ calcd 464.11, obsd 464.12.
LC: tR = 10.67.

4966 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
Albers et al.
3-[(3-{[3-(4-Fluorobenzyl)-2,4-dioxo-1,3-thiazolan-5-yliden]-
methyl}phenoxy)methyl]benzeneboronic Acid (75).
(7) van Meeteren, L. A.; Ruurs, P.; Stortelers, C.; Bouwman, P.;
van Rooijen, M. A.; Pradere, J. P.; Pettit, T. R.; Wakelam, M. J.;
Saulnier-Blache, J. S.; Mummery, C. L.; Moolenaar, W. H.;
Jonkers, J. Autotaxin, a secreted lysophospholipase D, is essential
for blood vessel formation during development. Mol. Cell. Biol.
2006, 26, 5015–5022.
(8) Tanaka, M.; Okudaira, S.; Kishi, Y.; Ohkawa, R.; Iseki, S.; Ota,
M.; Noji, S.; Yatomi, Y.; Aoki, J.; Arai, H. Autotaxin stabilizes
blood vessels and is required for embryonic vasculature by produ-
cing lysophosphatidic acid. J. Biol. Chem. 2006, 281, 25822–25830.
(9) van Meeteren, L.; Moolenaar, W. Regulation and biological
activities of the autotaxin-LPA axis. Prog. Lipid Res. 2007, 46,
145–160.
(10) David, M.; Wannecq, E.; Descotes, F.; Jansen, S.; Deux, B.;
Ribeiro, J.; Serre, C. M.; Gres, S.; Bendriss-Vermare, N.; Bollen,
M.; Saez, S.; Aoki, J.; Saulnier-Blache, J. S.; Clezardin, P.;
Peyruchaud, O. Cancer cell expression of autotaxin controls bone
metastasis formation in mouse through lysophosphatidic acid-
dependent activation of osteoclasts. PLoS One 2010, 5, e9741.
(11) Kanda, H.; Newton, R.; Klein, R.; Morita, Y.; Gunn, M.; Rosen,
S. Autotaxin, an ectoenzyme that produces lysophosphatidic acid,
promotes the entry of lymphocytes into secondary lymphoid
organs. Nat. Immunol. 2008, 9, 415–423.
(12) Tager, A.; LaCamera, P.; Shea, B.; Campanella, G.; Selman, M.;
Zhao, Z.; Polosukhin, V.; Wain, J.; Karimi-Shah, B.; Kim, N.;
Hart, W.; Pardo, A.; Blackwell, T.; Xu, Y.; Chun, J.; Luster, A. The
lysophosphatidic acid receptor LPA1 links pulmonary fibrosis to
lung injury by mediating fibroblast recruitment and vascular leak.
Nat. Med. 2008, 14, 45–54.
(13) Meyer zu Heringdorf, D.; Jakobs, K. H. Lysophospholipid recep-
tors: signalling, pharmacology and regulation by lysophospholipid
metabolism. Biochim. Biophys. Acta 2007, 1768, 923–940.
(14) Murakami, M.; Shiraishi, A.; Tabata, K.; Fujita, N. Identification
of the orphan GPCR, P2Y10 receptor as the sphingosine-1-phos-
phate and lysophosphatidic acid receptor. Biochem. Biophys. Res.
Commun. 2008, 371, 707–712.
Yield: 40%. 1H NMR: δ = 8.10 (bs, 2H, OH), 7.94 (s, 1H, c),
7.87 (s, 1H, i2), 7.76 (d, J 7.3, 1H, k), 7.51-7.16 (m, 10H, d þ
i1 þ j þ y), 5.16 (s, 2H, h), 4.82 (s, 2H, z). MS: m/z [M þ H]þ
calcd 464.11, obsd 464.13. LC: tR = 10.72.
4-[(3-{[3-(4-Fluorobenzyl)-2,4-dioxo-1,3-thiazolan-5-yliden]-
methyl}phenoxy)methyl]benzeneboronic Acid (76).
Yield: 26%. 1H NMR: δ = 8.07 (s, 2H, OH), 7.93 (s, 1H, c),
7.81(d, J 8.0, 2H, j), 7.49-7.16 (m, 10H, d þ i þ y), 5.19 (s,
2H, h), 4.82 (s, 2H, z). MS: m/z [M þ H]þ calcd 464.11, obsd
464.12. LC: tR = 10.70.
2-[(3-{[3-(4-Fluorobenzyl)-2,4-dioxo-1,3-thiazolan-5-yliden]-
methyl}phenoxy)methyl]benzeneboronic Acid (77).
(15) van Meeteren, L. A.; Ruurs, P.; Christodoulou, E.; Goding, J. W.;
Takakusa, H.; Kikuchi, K.; Perrakis, A.; Nagano, T.; Moolenaar,
W. H. Inhibition of autotaxin by lysophosphatidic acid and
sphingosine 1-phosphate. J. Biol. Chem. 2005, 280, 21155–21161.
(16) Gajewiak, J.; Tsukahara, R.; Fujiwara, Y.; Tigyi, G.; Prestwich,
G. D. Synthesis, pharmacology, and cell biology of sn-2-aminooxy
analogues of lysophosphatidic acid. Org. Lett. 2008, 10, 1111–1114.
(17) Cui, P.; Tomsig, J.; McCalmont, W.; Lee, S.; Becker, C.; Lynch, K.;
Macdonald, T. Synthesis and biological evaluation of phosphonate
derivatives as autotaxin (ATX) inhibitors. Bioorg. Med. Chem.
Lett. 2007, 17, 1634–1640.
(18) Guowei, J.; Yong, X.; Yuko, F.; Tamotsu, T.; Ryoko, T.; Joanna,
G.; Gabor, T.; Prestwich, G. D. Alpha-substituted phosphonate
analogues of lysophosphatidic acid (LPA) selectively inhibit pro-
duction and action of LPA. ChemMedChem 2007, 2, 679–690.
(19) Ferry, G.; Moulharat, N.; Pradere, J.; Desos, P.; Try, A.; Genton,
A.; Giganti, A.; Beucher-Gaudin, M.; Lonchampt, M.; Bertrand,
M.; Saulnier-Blache, J.; Tucker, G.; Cordi, A.; Boutin, J. S32826, a
nanomolar inhibitor of autotaxin: discovery, synthesis and appli-
cations as a pharmacological tool. J. Pharmacol. Exp. Ther. 2008,
327, 809–819.
Yield: 32%. 1H NMR: δ = 8.09 (s, 2H, OH), 7.94 (s, 1H, c),
7.57-7.12 (m, 12H, d þ i þ y), 5.29 (s, 2H, h), 4.83 (s, 2H, z).
MS: m/z [M þ H]þ calcd 464.11, obsd 464.13. LC: tR
=
10.73.
Acknowledgment. We thank Anastassis Perrakis for help-
ful discussions. This work was supported by grants from the
NetherlandsOrganization for Scientific Research (NWO), the
Dutch Cancer Society, the Netherlands Genomics Initiative,
and the Center for Biomedical Genetics.
Supporting Information Available: Information about the
screening of an NCI collection and the synthetic procedures
and analytical data for compounds 9-30. This material is
available free of charge via the Internet at http://pubs.acs.org.
(20) van Meeteren, L.; Brinkmann, V.; Saulnier-Blache, J.; Lynch, K.;
Moolenaar, W. Anticancer activity of FTY720: phosphorylated
FTY720 inhibits autotaxin, a metastasis-enhancing and angiogenic
lysophospholipase D. Cancer Lett. 2008, 266, 203–208.
(21) Moulharat, N.; Fould, B.; Giganti, A.; Boutin, J.; Ferry, G.
Molecular pharmacology of adipocyte-secreted autotaxin.
Chem.-Biol. Interact. 2008, 172, 115–124.
(22) Parrill, A.; Echols, U.; Nguyen, T.; Pham, T.; Hoeglund, A.; Baker,
D. Virtual screening approaches for the identification of non-lipid
autotaxin inhibitors. Bioorg. Med. Chem. 2008, 16, 1784–1795.
(23) Hoeglund, A.; Bostic, H.; Howard, A.; Wanjala, I.; Best, M.;
Baker, D.; Parrill, A. Optimization of a pipemidic acid autotaxin
inhibitor. J. Med. Chem. 2009, 53, 1056–1066.
(24) Albers, H. M.; Dong, A.; van Meeteren, L. A.; Egan, D. A.;
Sunkara, M.; van Tilburg, E. W.; Schuurman, K.; van Tellingen,
O.; Morris, A. J.; Smyth, S. S.; Moolenaar, W. H.; Ovaa, H.
Boronic acid-based inhibitor of autotaxin reveals rapid turnover
of LPA in the circulation. Proc. Natl. Acad. Sci. U.S.A. 2010, 107,
7257–7262.
References
(1) Gijsbers, R.; Aoki, J.; Arai, H.; Bollen, M. The hydrolysis of
lysophospholipids and nucleotides by autotaxin (NPP2) involves
a single catalytic site. FEBS Lett. 2003, 538, 60–64.
(2) Clair, T.; Lee, H. Y.; Liotta, L. A.; Stracke, M. L. Autotaxin is an
exoenzyme possessing 50-nucleotide phosphodiesterase/ATP pyr-
ophosphatase and ATPase activities. J. Biol. Chem. 1997, 272,
996–1001.
(3) Bollen, M.; Gijsbers, R.; Ceulemans, H.; Stalmans, W.; Stefan, C.
Nucleotide pyrophosphatases/phosphodiesterases on the move.
Crit. Rev. Biochem. Mol. Biol. 2000, 35, 393–432.
(4) Tokumura, A.; Majima, E.; Kariya, Y.; Tominaga, K.; Kogure, K.;
Yasuda, K.; Fukuzawa, K. Identification of human plasma lyso-
phospholipase D, a lysophosphatidic acid-producing enzyme, as
autotaxin, a multifunctional phosphodiesterase. J. Biol. Chem.
2002, 277, 39436–39442.
(5) Umezu-Goto, M.; Kishi, Y.; Taira, A.; Hama, K.; Dohmae, N.;
Takio, K.; Yamori, T.; Mills, G. B.; Inoue, K.; Aoki, J.; Arai, H.
Autotaxin has lysophospholipase D activity leading to tumor cell
growth and motility by lysophosphatidic acid production. J. Cell
Biol. 2002, 158, 227–233.
(6) Moolenaar, W. H.; van Meeteren, L. A.; Giepmans, B. N. The ins
and outs of lysophosphatidic acid signaling. BioEssays 2004, 26,
870–881.
(25) Takakusa, H.; Kikuchi, K.; Urano, Y.; Sakamoto, S.; Yamaguchi,
K.; Nagano, T. Design and synthesis of an enzyme-cleavable sensor
molecule for phosphodiesterase activity based on fluorescence
resonance energy transfer. J. Am. Chem. Soc. 2002, 124, 1653–1657.
(26) Saunders, L. P.; Ouellette, A.; Bandle, R.; Chang, W. C.; Zhou, H.;
Misra, R. N.; De La Cruz, E. M.; Braddock, D. T. Identification of

Article
small-molecule inhibitors of autotaxin that inhibit melanoma cell
migration and invasion. Mol. Cancer Ther. 2008, 7, 3352–3362.
(27) Johnstone, R.; Rose, M. A rapid, simple, and mild procedure for
alkylation of phenols, alcohols, amides and acids. Tetrahedron
1979, 35, 2169–2173.
(28) Bruno, G.; Costantino, L.; Curinga, C.; Maccari, R.; Monforte, F.;
Nicolo, F.; Ottana, R.; Vigorita, M. G. Synthesis and aldose
reductase inhibitory activity of 5-arylidene-2,4-thiazolidinediones.
Bioorg. Med. Chem. 2002, 10, 1077–1084.
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4967
(30) Crompton, I. E.; Cuthbert, B. K.; Lowe, G.; Waley, S. G. Beta-
lactamase inhibitors. The inhibition of serine beta-lactamases by
specific boronic acids. Biochem. J. 1988, 251, 453–459.
(31) Kuivila, H.; Armour, A. Electrophilic displacement reactions. IX.
Effects of substituents on rates of reactions between hydrogen
peroxide and benzeneboronic acid. J. Am. Chem. Soc. 1957, 79,
5659–5662.
(32) Kuivila, H. Electrophilic displacement reactions. VI. Catalysis by
strong acids in the reaction between hydrogen peroxide and
benzeneboronic acid. J. Am. Chem. Soc. 1955, 77, 4014–4016.
(33) Snyder, H. R.; Kuck, J. A.; Johnson, J. Organoboron compounds,
and the study of reaction mechanisms. Primary aliphatic boronic
acids. J. Am. Chem. Soc. 1938, 60, 105–111.
(29) Groll, M.; Berkers, C.; Ploegh, H.; Ovaa, H. Crystal structure
of the boronic acid-based proteasome inhibitor bortezomib in
complex with the yeast 20S proteasome. Structure 2006, 14,
451–456.