Benzyl and naphthalene methylphosphonic acid inhibitors of autotaxin with anti-invasive and anti-metastatic activity

Article abstract of DOI:10.1002/cmdc.201000425

Autotaxin (ATX, NPP2) is a member of the nucleotide pyrophosphate phosphodiesterase enzyme family. ATX catalyzes the hydrolytic cleavage of lysophosphatidylcholine (LPC) by lysophospholipaseD activity, which leads to generation of the growth-factor-like l

Full text of DOI:10.1002/cmdc.201000425

MED
DOI: 10.1002/cmdc.201000425
Benzyl and Naphthalene Methylphosphonic Acid Inhibitors
of Autotaxin with Anti-invasive and Anti-metastatic
Activity
Renuka Gupte,[a] Renukadevi Patil,[a] Jianxiong Liu,[b] Yaohong Wang,[b] Sue C. Lee,[b]
Yuko Fujiwara,[b] James Fells,[b] Alyssa L. Bolen,[b] Karin Emmons-Thompson,[a] C. Ryan Yates,[a]
Anjaih Siddam,[a] Nattapon Panupinthu,[c] Truc-Chi T. Pham,[d] Daniel L. Baker,[d]
Abby L. Parrill,[d, e] Gordon B. Mills,[c] Gabor Tigyi,*[b] and Duane D. Miller*[a]
Autotaxin (ATX, NPP2) is a member of the nucleotide pyro-
phosphate phosphodiesterase enzyme family. ATX catalyzes
the hydrolytic cleavage of lysophosphatidylcholine (LPC) by ly-
sophospholipase D activity, which leads to generation of the
growth-factor-like lipid mediator lysophosphatidic acid (LPA).
ATX is highly upregulated in metastatic and chemotherapy-re-
sistant carcinomas and represents a potential target to medi-
ate cancer invasion and metastasis. Herein we report the syn-
thesis and pharmacological characterization of ATX inhibitors
based on the 4-tetradecanoylaminobenzylphosphonic acid
scaffold, which was previously found to lack sufficient stability
in cellular systems. The new 4-substituted benzylphosphonic
acid and 6-substituted naphthalen-2-ylmethylphosphonic acid
analogues block ATX activity with Ki values in the low micro-
molar to nanomolar range against FS3, LPC, and nucleotide
substrates through a mixed-mode inhibition mechanism. None
of the compounds tested inhibit the activity of related en-
zymes (NPP6 and NPP7). In addition, the compounds were
evaluated as agonists or antagonists of seven LPA receptor
(LPAR) subtypes. Analogues 22 and 30b, the two most potent
ATX inhibitors, inhibit the invasion of MM1 hepatoma cells
across murine mesothelial and human vascular endothelial
monolayers in vitro in a dose-dependent manner. The average
terminal half-life for compound 22 is 10ꢀ5.4 h and it causes a
long-lasting decrease in plasma LPA levels. Compounds 22 and
30b significantly decrease lung metastasis of B16-F10 syngene-
ic mouse melanoma in a post-inoculation treatment paradigm.
The 4-substituted benzylphosphonic acids and 6-substituted
naphthalen-2-ylmethylphosphonic acids described herein rep-
resent new lead compounds that effectively inhibit the ATX–
LPA–LPAR axis both in vitro and in vivo.
Introduction
Autotaxin (ATX, NPP2) is member of the nucleotide pyrophos-
phate phosphodiesterase (NPP) enzyme family. ATX catalyzes
the hydrolysis of lysophosphatidylcholine (LPC) through lyso-
phospholipase D (LPLD) activity, which leads to generation of
the growth-factor-like lipid mediator lysophosphatidic acid
(LPA).[1,2] ATX is a 125 kDa glycoprotein, originally isolated from
human melanoma cells, that stimulates tumor cell motility and
has been implicated in metastatic and invasive properties, as
well as chemotherapeutic resistance of many carcinomas.[3–12]
NPP6 and NPP7 are the only other known NPP isoforms that
use lysophospholipids as substrates. However, in contrast to
ATX, NPP6 and NPP7 act as lysophospholipase C. NPP6 cleaves
phosphocholine from LPC, sphingosylphosphorylcholine, and
glycerophosphorylcholine to generate monoacylglycerol,
sphingosine, and glycerol, respectively.[13] NPP7, also referred
to as alkaline sphingomyelinase, hydrolyzes sphingomyelin to
generate ceramide in the intestinal tract, but also cleaves
phosphocholine from LPC and platelet-activating factor to gen-
erate monoacyl- and alkylacetylglycerols.[14,15]
[a] R. Gupte, Dr. R. Patil, Dr. K. Emmons-Thompson, Prof. C. R. Yates,
Dr. A. Siddam, Prof. D. D. Miller
Department of Pharmaceutical Sciences, College of Pharmacy
University of Tennessee
Health Science Center, Memphis, TN 38163 (USA)
Fax: (+1)901 448-3446
E-mail: dmiller@uthsc.edu
[b] Dr. J. Liu, Dr. Y. Wang, Dr. S. C. Lee, Dr. Y. Fujiwara, J. Fells, Dr. A. L. Bolen,
Prof. G. Tigyi
Department of Physiology, College of Medicine
University of Tennessee
Health Science Center, Memphis, TN 38163 (USA)
Fax: (+1)901 448-7126
E-mail: gtigyi@uthsc.edu
[c] Dr. N. Panupinthu, Prof. G. B. Mills
Department of Systems Biology
The University of Texas M.D. Anderson Cancer Center
Houston, TX 77054 (USA)
[d] Dr. T.-C. T. Pham, Prof. D. L. Baker, Prof. A. L. Parrill
Department of Chemistry, University of Memphis
Memphis, TN 38152 (USA)
[e] Prof. A. L. Parrill
Computational Research on Materials Institute
University of Memphis, Memphis, TN 38152 (USA)
ATX is required for normal development. Homozygous ATX
knockout mice die in utero at day 9.5, coinciding with a period
of vascular stabilization.[16,17] Likewise, ATX plays an important
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000425.
922
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 922 – 935

Autotaxin Inhibitors
role in the development of the nervous system, as ATX knock-
out mice show defects in neural tube development.[18] LPA is a
mitogen, motogen, and anti-apoptotic agent which provides
survival advantages to carcinomas that use LPA in an autocrine
or paracrine fashion. Ovarian cancer cells produce high levels
of LPA.[19–21] Copy numbers increase in ovarian cancers in chro-
mosomal region 8q24, which contains the genes that encode
ATX and the Myc oncogene.[22] Euer et al. found that ATX is
among the 40 most upregulated genes in highly metastatic
cancers.[23] It was recently shown that ectopic expression of
ATX in mice leads to mammary intraepithelial neoplasia, which
develops into invasive and metastatic tumors.[11] ATX inhibits
paclitaxel-induced apoptosis in breast cancer cells,[6] and LPA
renders ovarian cancer cells resistant to cisplatin and adriamy-
cin.[24] ATX is also overexpressed in patients with recurrent dis-
ease after prior chemotherapeutic treatment.[25] In a genome-
wide siRNA screen, we identified ATX as a candidate drug-re-
sistance gene in ovarian cancer.[7] We also showed that a lipid-
like small-molecule inhibitor of ATX, carbacyclic phosphatidic
acid, increases the sensitivity of resistant ovarian cancer cells
to paclitaxel treatment.[7]
model in C57BL/6 mice. Based on these results, compounds 22
and 30b represent promising leads for further synthetic im-
provement and also provide proof of principle that ATX inhibi-
tors offer therapeutic utility in the control of cancer metastasis
in vivo.
Results and Discussion
Synthesis
The synthesis of 4-(hexadecane-1-sulfonylamino)benzylphos-
phonic acid (4) is depicted in Scheme 1. Commercially available
1-hexadecanesulfonyl chloride (2) was added to a mixture of 4-
aminobenzylphosphonic acid diethyl ester (1) and N,N-diiso-
propylethylamine in THF under reflux conditions to give com-
The ATX–LPA–LPAR axis is a promising therapeutic target for
the management of cancer metastasis and therapeutic resist-
ance. ATX shows feedback inhibition by its hydrolysis products
LPA, cyclic phosphatidic acid, and sphingosine-1-phosphate
(S1P).[26,27] Many initially identified ATX inhibitors are lipid-like
substrate or product analogues.[28–35] The characteristics of
these types of compounds limit their utility as potential lead
compounds for drug development. Non-lipid ATX inhibitors
have also been identified, but most of these compounds lack
sufficient stability and characterization in tumor models.[30,36–41]
Ferry and colleagues recently described an ATX inhibitor, 4-tet-
radecanoylaminobenzylphosphonic acid (S32826), that pos-
sesses nanomolar activity in vitro.[42] Unfortunately, S32826
failed to show activity in cellular and in vivo systems. We hy-
pothesized that hydrolysis of the amide bond present in
S32826 could be the reason for its instability and thus lack of
activity in cellular systems.
Scheme 1. Reagents and conditions: a) N,N-diisopropylethylamine, THF,
reflux, 24 h; b) TMSBr, CH3CN, reflux, 1 h; c) MeOH, RT, 30 min.
pound 3, which was deprotected with bromotrimethylsilane
(TMSBr) followed by the addition of methanol under stirring
for 30 min to afford compound 4. The synthesis of compound
10 started with alkylation of 4-hydroxymethylphenol (5)
(Scheme 2) using 1-bromotetradecane (6) in the presence of
potassium carbonate and 18-crown-6 to furnish compound 7
followed by treatment with phosphorus tribromide to obtain 8
in 98% yield. The Michaelis–Arubuzov reaction[43] on com-
pound 8 in trimethyl phosphate provided compound 9, which
was then deprotected using TMSBr in anhydrous acetonitrile
to yield 10.
To overcome the presumed lack of stability of S32826, we
designed analogues that are expected to be more stable than
the parent compound. We generated a panel of analogues
that inhibit ATX with potencies similar to that of S32826. These
stabilized analogues inhibit ATX by a mixed-mode mechanism
in vitro without any effect on the related lysophospholipid
phosphodiesterases NPP6 and NPP7, or on LPAR. Two of these
compounds, 22 and 30b, inhibited ATX-dependent invasion of
rat MM1 hepatoma cells of mesothelial cell and human umbili-
cal cord vascular endothelial cell (HUVEC) monolayers in vitro.
In addition, 22 and 30b showed a profound decrease in lung
foci in vivo using the B16-F10 syngeneic melanoma metastasis
Scheme 2. Reagents and conditions: a) K2CO3, 18-crown-6, acetone, reflux,
16 h; b) PBr3, Et2O, RT, 30 min; c) P(OMe)3, reflux, 18 h; d) TMSBr, CH3CN,
reflux, 1 h; e) MeOH, RT, 30 min.
ChemMedChem 2011, 6, 922 – 935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
923

G. Tigyi, D. D. Miller, et al.
MED
Scheme 3. Reagents and conditions: a) Pd(OAc)2, Et3N, DMF, reflux, 16 h; b) LiAlH4, THF, 08C–RT, 4 h; c) H2, Pd/C, MeOH, RT, 2 h; d) PBr3, Et2O, RT, 30 min;
e) P(OMe)3, reflux, 18 h; f) TMSBr, CH3CN, reflux, 1 h; g) MeOH, RT, 30 min.
Scheme 3 illustrates the synthesis of (E)-4-(pentadec-1-enyl)-
benzylphosphonic acid (18) and 4-pentadecylbenzylphosphon-
ic acid (22). (E)-Methyl-4-(pentadec-1-enyl)benzoate (13) was
synthesized by the Heck coupling of 11 and 12 using palladi-
um(II) acetate and triethylamine in anhydrous DMF. Compound
13 was reduced with lithium aluminum hydride in THF at 08C
to room temperature to produce 14 and was also saturated
under catalytic hydrogenation conditions to generate com-
pound 15. Bromination of compound 14 with phosphorus tri-
bromide in diethyl ether gave 16, which was then subjected to
the Michaelis–Arubuzov reaction[43] using trimethylphosphite
to get 17. Compound 17 was treated with TMSBr in acetoni-
trile followed by the addition of methanol to furnish 18. Com-
pound 22 was synthesized by using the same approach start-
ing from compound 15.
The synthesis of 6-substituted naphthalen-2-ylmethylphos-
phonic acid analogues (30a,b and 34a,b) was performed by
using the same procedure described for the synthesis of com-
pounds 18 and 22 (Scheme 4). We then focused our medicinal
chemistry efforts on phosphomimetic functional groups; the
Scheme 5. Reagents and conditions: a) PDC, CH2Cl2, RT, 16 h; b) HP(O)(OCH3)2,
Et3N, 08C–RT, 4 h; c) TMSBr, CH3CN, reflux, 1 h; d) MeOH, RT, 30 min; e) DAST,
phosphomimetics include compounds 37, 41, 42, and 43
(Scheme 5). We started with the conversion of compound 19
into an aldehyde derivative 35 using pyridinium dichromate
(PDC) in dichloromethane. The resulting aldehyde 35 was then
Et2O, 08C–RT, 1 h; f) SOCl2, CH2Cl2, reflux, 1 h; g) PBr3, Et2O, RT, 30 min.
Scheme 4. Reagents and conditions: a) Pd(OAc)2, Et3N, DMF, reflux, 16 h; b) LiAlH4, THF, 08C–RT, 4 h; c) H2, Pd/C, MeOH, RT, 2 h; d) PBr3, Et2O, RT, 30 min;
e) P(OMe)3, reflux, 18 h; f) TMSBr, CH3CN, reflux, 1 h; g) MeOH, RT, 30 min.
924
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 922 – 935

Autotaxin Inhibitors
Scheme 6. Reagents and conditions: a) 1. bis(2-cyanoethyl)-N,N-diisopropylphosphoramidite, 1H-tetrazole, CH2Cl2, RT, 1 h, 2. 50% H2O2, RT, 1 h; b) 1. bis(2-cya-
noethyl)-N,N-diisopropylphosphoramidite, 1H-tetrazole, CH2Cl2, RT, 1 h, 2. sulfur, reflux, 2 h; c) 1. KOH, MeOH, RT, 3 h, 2. dil HCl, RT, 3 h.
converted into the a-hydroxyphosphonate derivative 36 via
the Pudovik reaction.[33,44] Compound 36 was deprotected with
TMSBr to produce target compound 37. Compound 36 was
treated with diethylaminosulfur trifluoride (DAST) in Et2O,
SOCl2/CH2Cl2, or PBr3/Et2O to obtain the desired compounds
38, 39, and 40, respectively. Finally, compounds 38, 39, and 40
were treated with TMSBr followed by the addition of methanol
to afford compounds 41, 42, and 43 in good yields.
which, in turn, was treated with methanolic potassium hydrox-
ide followed by acidification to yield the thiophosphate 47.
Biological results
Effect of 4-substituted benzylphosphonic acid and 6-substitut-
ed naphthalene-2-ylmethylphosphonic acid derivatives on ATX
and LPAR
The synthesis of target compounds 45 and 47 is outlined in
Scheme 6. Compound 19 was treated with a mixture of 1H-tet-
razole and bis(2-cyanoethyl)-N,N-diisopropylphosphoramidite
in anhydrous dichloromethane followed by the addition of hy-
drogen peroxide to give the bis-cyanoethyl-protected fatty al-
cohol phosphate 44. Removal of the cyanoethyl groups by
treatment with methanolic potassium hydroxide followed by
acidification furnished the phosphate 45. Similarly, 19 was
treated with a mixture of 1H-tetrazole and bis(2-cyanoethyl)-
N,N-diisopropylphosphoramidite in anhydrous dichlorome-
thane followed by reflux in the presence of elemental sulfur to
give bis-cyanoethyl-protected fatty alcohol thiophosphate 46,
The first level of testing was done with compounds at 10 mm
added to recombinant human ATX and the FRET substrate FS3,
each at 2 nm. After incubation for 2 h, the amount of FS3 hy-
drolyzed was measured, and residual ATX activity was ex-
pressed as a percentage of the vehicle-treated sample minus
the autolysis of FS3 in the absence of ATX. In this assay,
S32826 decreased the amount of FS3 hydrolyzed by 97%
(Tables 1 and 2). Among the 4-substituted benzylphosphonic
acid analogues, compound 22 showed 95% inhibition of FS3
hydrolysis (Table 1), whereas compound 30b from the 6-substi-
tuted naphthalene-2-ylmethylphosphonic acid series decreased
Table 1. Characterization of benzylphosphonic acids at ATX, NPP6, and NPP7.[a]
ATX
NPP6
NPP7
Compd
S32826
X
Y
R
Activity [%][b]
2.6
IC50 [mm]
Mechanism[c]
ND
Ki [mm]
Ki’ [mm]
Activity [%][b]
102.1
Activity [%][b]
101.2
O
CH2
ND
NA
NA
4
O
CH2
24.1
ND
ND
NA
NA
102.0
93.5
10
18
22
37
41
42
43
45
47
O
O
O
O
O
O
O
O
S
CH2
CH2
30.4
18.8
5.2
ND
ND
ND
ND
NA
NA
NA
NA
102.0
104.0
96.5
100.6
97.7
96.7
101.0
102.6
99.9
99.5
98.8
92.9
CH2
0.17
0.73
17.9
ND
mixed
mixed
mixed
ND
0.27
0.45
4.97
NA
0.28
0.70
5.54
NA
CHOH
CHF
9.2
99.6
42.3
117.9
60.1
85.2
15.2
101.1
100.6
99.3
CHCl
CHBr
OꢁCH2
OꢁCH2
10.1
ND
mixed
ND
6.10
NA
2.97
NA
100.9
98.8
1.54
mixed
4.45
4.43
[a] ND: not determined; NA: not applicable, as the mechanism of inhibition was not determined. [b] Expressed as a percent of control at 10 mm. [c] Mech-
anism of inhibition.
ChemMedChem 2011, 6, 922 – 935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
925

G. Tigyi, D. D. Miller, et al.
MED
FS3 hydrolysis by 83% (Table 2). Dose–response curves were
generated with compounds S32836, 22, and 30b and com-
pared with the feedback inhibition of the ATX product LPA
(Figure 1). Each of these three compounds inhibited ATX in a
dose-dependent manner.
methyl phosphonate decreased efficacy. However, converting
the phosphate to a thiophosphate (compound 47) in this scaf-
fold increased efficacy 5.6-fold at 10 mm. a-Halogenation of the
linker methyl group generally decreased potency of com-
pounds 41, 42, and 43 relative to compound 22. Insertion of
an a-hydroxy group into the linker methylene increased the
potency of this analogue, but it did not surpass that of com-
pound 22.
Modification of the side chain at position 6 of the naphtha-
lene-2-ylmethylphosphonic acid scaffold showed that the 10-
carbon side chain is more effective than the 11-carbon chain
and that incorporation of an alkene has differential effects de-
pending on the chain length (decreased efficacy of compound
30a relative to 34a, but increased efficacy of 30b over 34b).
Analysis of the seven analogues showed that compound 22 in
the benzylphosphonic acid series and compound 30b in the
naphthalene methylphosphonic acids are the most potent ana-
logues with respect to ATX inhibition.
Because FS3 is not a natural substrate of ATX, we deter-
mined whether the inhibitors can also block LPA production
from a fluorescent analogue of LPC (3-acyl-7-dimethylamino-
naphtyl-1-LPC, ADMAN-LPC) and the hydrolysis of the nucleo-
tide-like substrate p-nitrophenylthymidine 5’-monophosphate
(pNP-TMP). Compound 22 and 30b inhibited LPA production
form ADMAN-LPC and hydrolysis of the pNP-TMP substrate
(Figure 2). To determine the mechanism of ATX inhibition, Ki
(affinity of the compound for free enzyme) and Ki’ (affinity of
the compound for the enzyme–substrate complex) values
were determined by using simultaneous nonlinear fits of the
Michaelis–Menten equations for competitive, uncompetitive,
mixed-mode, and noncompetitive inhibition as we have previ-
ously described.[36,39,45] These experiments uniformly showed
that these compounds display similar Ki and Ki’ values which is
consistent with a mixed-mode type mechanism of inhibition.
ATX inhibitors have been examined against other NPP family
members.[29,36,37] NPP6 and NPP7 are the only NPP isoforms
beside ATX that are known to use lysophospholipid phospho-
diesters as substrates. Thus, specificity was addressed by deter-
mining the activity of NPP6 and NPP7 in the presence and ab-
sence of these analogues. None of the analogues tested
blocked the activity of NPP6 or NPP7 by >10% at a single
10 mm dose (Tables 1 and 2).
Figure 1. Inhibition of ATX-mediated hydrolysis of FS3 by LPA, S32826, 22, or
30b. ATX (2 nm) and FS3 (1 mm) were incubated in the presence of increas-
ing concentrations of the inhibitors for 2 h, and the fluorescent product was
measured and expressed as percent of FS3 hydrolysis by ATX in the pres-
ence of vehicle minus the autolysis of FS3 (n=3ꢀSD).
Next we generated analogues of compounds 22 and 30b
and examined their inhibition of ATX. The modifications were
either in the aliphatic chain, the linker of the phosphate
moiety, or on the phosphate moiety itself (Table 1 and Table 2).
The most effective analogues based on the single 10 mm dose
inhibition assay were extensively characterized to determine
IC50, Ki, and Ki’ values against ATX-mediated hydrolysis of FS3.
Replacement of the amide with a sulfonamide (compound 4)
decreased potency. Introduction of an ether (in 10) or alkene
(in 18) connecting the hydrocarbon chain to the benzyl ring
decreased potency relative to the simple aliphatic chain. Link-
age of the phosphate moiety in compound 45 through a
Table 2. Characterization of naphtylphosphonic acids at ATX, NPP6, and NPP7.[a]
ATX
NPP6
NPP7
Compd
X
Y
R
Activity [%][b]
IC50 [mm]
Mechanism[c]
Ki [mm]
Ki’ [mm]
Activity [%][b]
Activity [%][b]
30a
30b
34a
34b
O
O
O
O
CH2
CH2
CH2
CH2
61.3
17.3
19.9
50.8
ND
1.40
ND
ND
mixed
ND
NA
1.50
NA
NA
1.01
NA
103.1
104.1
101.0
100.0
100.0
96.5
98.8
94.7
ND
ND
NA
NA
[a] ND: not determined; NA: not applicable, as the mechanism of inhibition was not determined. [b] Expressed as a percent of control at 10 mm. [c] Mech-
anism of inhibition.
926
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 922 – 935

Autotaxin Inhibitors
Figure 2. Inhibition of LPA production and pNP-TMP hydrolysis by compounds 22 and 30b. A) Representative TLC image shows that compounds 22 and 30b
applied at 3 mm inhibit the hydrolysis of 0.5 mm ADMAN-LPC by 100 nm ATX. B) Compounds 22 and 30b applied at 10 mm inhibit the hydrolysis of 5 mm pNP-
TMP by 4 nm ATX more than the product LPA 18:1 applied at 10 mm.
Some inhibitors of ATX, including LPA phosphonates[28] and
LPA bromophosphonates,[34] have been shown to inhibit ATX
and block LPAR subtypes. LPA bromophosphonate (LPA-BP)
was also shown to inhibit cancer metastasis and to decrease
tumor size in mice.[34] For this reason, we examined whether
the two leads 22 and 30b can affect the activation of multiple
LPA GPCR targets by using cell lines that overexpress individual
LPA GPCRs coupled to Ca2+ mobilization, as we have previous-
ly described.[34] Analyses were carried out with test compounds
(10 mm) alone or in combination with an EC50 concentration of
LPA for the given receptor subtype to determine agonist and
antagonist activity, respectively (Table 3).
10 mm. Compound 33b elicited a 33% activation of LPA3 and a
17% activation at LPA5 when applied at 10 mm. Likewise, com-
pound 22 (10 mm) was
a weak partial agonist at LPA2
(Emax 10mm ~20%), whereas analogues 37 and 41 inhibited this
receptor subtype by 28 and 26%, respectively. Based on these
findings, we conclude that these analogues are poor ligands of
the LPA GPCR tested, with the exception of compounds 45
and 47, which are full agonists of LPA5 albeit with low potency.
At present, very few LPA5-selective compounds have been de-
scribed,[46] hence the identification of these two hits might
serve as a starting point for a more comprehensive characteri-
zation of the structure–activity relationship of this receptor
subtype.
Some of the compounds that showed >10% attenuation of
the LPA response or activation >10% of LPA at a given recep-
tor subtype at 10 mm were subjected to dose–response/inhibi-
tion experiments. Unexpectedly, compounds 45 and 47 were
weak but full agonists at LPA5 with EC50 values of 7.9 and
2.9 mm, respectively. Compound 47 is also a partial agonist at
LPA3 (Emax 10mm ~44% of maximal LPA response). Compound 45
caused a 50% decrease in the LPA2 response when applied at
ATX inhibitors block hepatocarcinoma invasion of mesothelial
and HUVEC monolayers
Cancer metastasis is a complex process that cannot be accu-
rately modeled in vitro. Models that use cell monolayers to
monitor for the invasion of carcinoma cells come the closest to
Table 3. Activation of multiple LPA GPCR targets by various test compounds.[a]
Compd
RH7777 vector
LPA1
[Emax 10mm
LPA2
[Imax 10mm
LPA3
[Emax 10mm
CHO vector
LPA4
B103 vector
[Emax 10mm]
LPA5
[b]
[b]
[c]
[b]
[b]
[b]
[Emax 10mm
]
]
]
]
[Emax 10mm
]
[b]
[c]
[Emax 10mm
]
[Imax 10mm]
22
30b
37
41
43
42
45
47
>1
>6
>1
>3
>5
>15
>6
>7
NE
NE
NE
NE
NE
ND
NE
10
23
NE
28
22
33
NE
17
NE
ND
13
44
>1
>1
NE
>5
>5
>74
>14
>14
NE
NE
NE
NE
NE
ND
NE
NE
>3
>3
NE
>1
>2
>32
>2
>3
NE
17
NE
NE
NE
NE
NE
51
NE
NE
ND
NE
NE
26
NE
ND
50
ND
100[d]
100[e]
NE
[a] NE: no effect; ND: not definable due to effect on vector-transfected cells. [b] Emax 10mm: maximal response at 10 mm relative to the maximal response to
either ATP in vector-transfected cells or to LPA in LPAR-expressing cells. [c] Imax 10mm: maximal inhibition of the ~EC75 LPA response caused by the compound
applied at 10 mm. [d] EC50 =7.9 mm. [e] EC50 =2.9 mm.
ChemMedChem 2011, 6, 922 – 935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
927

G. Tigyi, D. D. Miller, et al.
MED
the situation in vivo.[47–50] For this we used two different cell
monolayers, mouse mesothelium and human umbilical vein
endothelial cell (HUVEC), to examine the role of ATX in the in-
vasion of MM1 hepatocarcinoma cells. The mesothelial mono-
layer is an accepted model for invasion of body cavities lined
by serous cells, whereas the HUVEC monolayer is considered
to be an in vitro model of hematogenous invasion of carcino-
mas.
exhibiting initial distribution and terminal elimination phases
(Figure 4). Non-compartmental analysis following intravenous
administration showed that the volume of distribution (Vd) for
compound 22 was 3.9 Lkgꢁ1, and the clearance was estimated
to be 8.6ꢀ0.5 mLminꢁ1 kgꢁ1. The average terminal half-life for
compound 22 was 10ꢀ5.4 h, with a mean plasma residence
time of 7.6ꢀ0.5 h. LPA levels were monitored simultaneously
for the 16:0, 18:1, 18:2, and 20:4 acyl species and showed simi-
lar changes regardless of the fatty acyl chain length. Plasma
LPA levels showed a rapid decrease with a nadir at 60 min fol-
lowed by a transient rise at 120 min and second decrease
reaching a minimum in the 480 min sample. LPA levels re-
mained depressed at ~50% of the pretreatment level in the
1440 min sample, the longest sampling time included in the
study. These results indicate a long-lasting decrease in plasma
LPA that appears to be consistent with the 10 h t1/2 value of
this compound. Of note is a confounding factor that at 2 h,
heparin (50 IU) was injected to keep the cannulas open, and
this may have affected plasma LPA levels. Furthermore, these
results may not be reflective of the levels following intraperito-
neal administration used in the murine metastasis model. Fur-
ther experiments are needed to characterize the pharmacoki-
netic properties of compound 22 and other analogues using
different routes of administration. It is important to recognize
that plasma LPA levels are not likely to be indicative of the LPA
level in the tumor microenvironment as supported by the lack
of elevation in LPA or ATX activity in different cancers, with the
exception of follicular lymphoma.[8,19,41,51,52]
First we determined if exposure to compounds 22, 30b, and
the combined ATX inhibitor LPAR pan-antagonist compound
LPA-BP[34] causes toxicity in vitro. We exposed the cells of the
mesothelium monolayer or the MM1 cell line to these com-
pounds (10 mm) dissolved in 1% DMSO for 20 h and examined
cell viability with the trypan blue dye-exclusion assay. Neither
compound caused a decrease in cell viability relative to the ve-
hicle, and all viabilities were >93%.
The mesothelial cells and HUVECs of the monolayer and the
invading MM1 cells express different levels of ATX (Figure 3A–
C). Quantitative real-time PCR performed with mRNA isolated
from MM1, mesothelium, and HUVECs showed that the pri-
mary source of ATX are MM1 cells, as the other two cell types
show very low expression. MM1 cells showed robust expres-
sion of transcripts encoding the P2Y family of LPARs and LPA2
in the EDG receptor subfamily. In mesothelial cells LPA2 is the
predominant receptor, whereas in HUVECs it is the LPA5 recep-
tor.
Addition of LPC to the co-culture increases invasion in a
dose-dependent manner (Figure 3D). We hypothesized that in-
hibition of ATX by our two lead compounds should attenuate
LPA production in situ and decrease invasion of MM1 carcino-
ma cells through the mesothelial and HUVEC monolayers.
Compounds 22 and 30b were applied with or without LPC to
the co-cultures, and the number of invading cells was quanti-
fied after 20–24 h of co-culture (Figure 3E). Both compounds
inhibit MM1 cell invasion, reaching complete inhibition of LPA
and likely ATX-dependent invasion above 3 mm. The two com-
pounds were also tested for their ability to inhibit MM1 cell in-
vasion of HUVEC monolayers (Figure 3F). Similarly to that ob-
served for the invasion of murine mesothelial monolayers,
compounds 22 and 30b dose-dependently inhibited the LPC-
dependent invasion of the HUVEC monolayer, whereas the
compounds alone did not decrease the basal rate of invasion.
We conducted similar experiments with rat microvascular cell
monolayers and found similar inhibition with 22 and 30b, al-
though the number of invading cells per field was considerably
lower than that in HUVEC or mesothelial monolayers (data not
shown). These results provide evidence that the inhibition of
ATX in situ can fully inhibit LPC-dependent invasion of carcino-
ma cells, suggesting the potential applicability of these com-
pounds in animal models of carcinoma metastasis.
ATX inhibitors decrease B16 melanoma metastasis in vivo
We conducted an acute toxicity study with compound 22 by
exposing Swiss mice to 1, 10, and 30 mgkgꢁ1 dayꢁ1 doses for
10 days. We found no noticeable change in animal behavior
within 1 h after injection, or weight loss. No death was detect-
ed in any of the groups. Necropsy conducted on day 14
showed no noticeable organ damage, with the exception of
adhesions and splenomegaly in three mice in the group receiv-
ing the highest dose. We conclude that the non-observable ad-
verse effect level of compound 22 is >10 mgkgꢁ1
.
Inhibitors of ATX have been shown to decrease tumor meta-
stasis in various animal models.[31,34] Here we applied the syn-
geneic B16-F10 mouse melanoma model of hematogenous
lung metastasis in C57BL/6 mice that we previously showed re-
sponds to inhibitors of ATX.[31] C57BL/6 mice were inoculated
with B16-F10 melanoma cells via the tail vein; 30 min later, the
mice were treated with either of the two lead compounds
(0.5 mgkgꢁ1 dayꢁ1) or vehicle (PBS with 1% DMSO) via the in-
traperitoneal route for an additional 10 days. As a positive con-
trol, we also applied LPA-BP, which we showed previously to
inhibit the metastasis of breast and colon cancers in xenograft
models.[34,53] On day 21 the animals were sacrificed and the
lungs were isolated; metastatic nodules in the lungs were then
quantified. In this metastasis model, compounds 22 and 30b
significantly decreased the number of metastatic foci, and their
efficacy at this dose was similar to that of LPA-BP (Figure 5A
and B).
Pharmacokinetic characterization of compound 22
Because of the slightly higher potency of compound 22 as an
ATX inhibitor, we selected this analogue for preliminary phar-
macokinetic characterization. The time-dependent plasma con-
centration profile of compound 22 declined bi-exponentially,
928
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 922 – 935

Autotaxin Inhibitors
Figure 3. Profiling of LPAR and ATX transcripts in A) MM1, B) mouse mesothelial, and C) HUVECs using quantitative real-time PCR. D) LPC dose-dependently in-
creases the invasion of mesothelial monolayers by MM1 cells. E) Compounds 22 and 30b dose-dependently inhibit the invasion of mesothelial monolayers in-
duced by LPC (1.5 mm) that is decreased to the level of control at and above 3 mm. F) Compounds 22 and 30b block LPC-induced but do not alter the basal
rate of invasion of HUVECs.
In this study we provide evidence for the potential thera-
peutic utility of ATX inhibitors derived from the benzylphos-
phonic acid scaffold. We have modified and expanded this
scaffold to 6-substituted naphthalene-methylphosphonic acid
derivatives that are also effective inhibitors of ATX. Our limited
structure–activity analysis showed that some of these com-
pounds are high-potency inhibitors of ATX without significant
effects on LPA GPCR in the nanomolar and low micromolar
range. Unlike the parent compound S32826 that lacks activity
in cellular assays,[42] these novel analogues show strong inhibi-
tion of ATX-mediated tumor cell invasion of mesothelial and
endothelial cell monolayers and are as effective in vivo as the
previously reported dual-action ATX inhibitor and LPA GPCR
antagonist LPA-BP.[31,34,53]
ChemMedChem 2011, 6, 922 – 935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
929

G. Tigyi, D. D. Miller, et al.
MED
Figure 5. Effect of compounds 22, 30b, and LPA-BP on the lung metastasis
of B16-F10 melanoma. A) Representative lungs taken from the treatment
groups. B) Lung nodule counts in the treatment groups; *p<0.05 relative to
vehicle.
ability to control invasion and/or metastasis. Our study pro-
vides initial proof of principle that ATX inhibitors are capable
of controlling cancer metastasis in vivo; it also validates the
monolayer invasion assay as a valuable cell-based screen for
anti-metastatic compounds.
Figure 4. Plasma level of A) compound 22 and B) two LPA species after i.v.
administration of a single dose at 5.8 mgkgꢁ1. Note that compound 22 has
a relatively long half-life calculated at 10.5 h, plasma mean residence time
(456ꢀ33 min), and LPA 18:2 and 20:4 remain depressed over the 24 h dura-
tion of the study.
The preclinical screening template consisting of assays for
the activity of analogues against ATX, followed by some relat-
ed NPP isoforms, LPA GPCR, monolayer invasion assays, and
then followed by the syngeneic animal metastasis model offers
a comprehensive stepwise protocol for the identification of
ATX inhibitors with utility to control tumor progression. Given
the relatively short half-life of ATX[56] and the high rate of LPA
turnover in vivo,[30] we foresee the need of compounds with
long plasma half-lives and broad tissue distribution to control
the constitutive production of ATX and the generation of LPA.
The data presented herein, derived from cellular and in vivo
experiments, suggest that further modification of the benzyl-
and naphthylphosphonic acid scaffolds might yield even more
suitable derivatives for the control of carcinoma metastasis.
Pharmacokinetic profiling of other analogues will be a crucial
step in further evaluation and preclinical development of ATX
inhibitors for the identification of scaffolds that exert long-
term inhibition of ATX in the tumor microenvironment.
There is increasing interest in the development of ATX inhib-
itors to control the tumor-promoting and pro-inflammatory
roles of LPA.[29,30,34,36–39,54] Carbacyclic phosphatidic acid was the
first ATX inhibitor scaffold explored to control carcinoma inva-
sion and metastasis in vivo.[31] Carbacyclic phosphatidic acid in-
hibits ATX without activating LPA1ꢁ4. However, these com-
pounds activate LPA5, and this complicates their utility.[46] Sub-
sequently, the sn-1 or sn-2 hydroxy groups of LPA have been
replaced by fluorine, difluoromethyl, difluoroethyl, O-methyl, or
O-hydroxyethoxy groups to give non-migrating LPA analogues
that resist acyltransferases.[55] Some of these compounds, in-
cluding LPA-BP, have been found to possess dual action by in-
hibiting ATX and LPA GPCR.[34] LPA-BP inhibited breast cancer
metastasis and tumor growth in vivo.[34] Several new ATX inhib-
itor scaffolds have been reported recently;[29,30,36–39,54] however,
these compounds have not been evaluated against other
members of the NPP family and were not explored for their
Conclusions
Two autotaxin inhibitors, 22 and 30b, were identified to be
potent and stable in in vivo models. All of the compounds syn-
930
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 922 – 935

Autotaxin Inhibitors
1-Bromomethyl-4-pentadecylbenzene (20): MS (ESI) m/z 380
[MꢁH]ꢁ.
thesized were specific for NPP2/ATX. The mode of enzyme in-
hibition for the most potent compounds was determined and
was found to be mixed. The fact that the ATX inhibitors block
hepatocarcinoma invasion of mesothelial and HUVEC monolay-
ers lends evidence that the inhibition of ATX in situ can fully in-
hibit LPC-dependent invasion of carcinoma cells. This signifies
the potential applicability of our compounds in animal models
of carcinoma metastasis. Compounds 22 and 30b have shown
outstanding in vivo profiles by decreasing the lung metastases
of B16-F10 syngeneic mouse melanoma in a post-inoculation
treatment model.
General procedure for the synthesis of compounds 9, 17, 21,
29a, 29b, 33a, and 33b (GP 3): Trimethylphosphite (9 equiv) was
added to dry bromide derivative (1 equiv) at RT, and the mixture
was held at reflux for 18 h. The trimethylphosphite was evaporated
under high vacuum overnight, and the crude residue was purified
by column chromatography (CHCl3/MeOH 3:1) to give pure com-
pound in quantitative yield.
Dimethyl-4-pentadecylbenzylphosphonate
(21):
1H NMR
(300 MHz, CDCl3): d=7.40–7.00 overlapping protons (m, 4H),
3.66–3.63 (d, J=15 Hz, 6H), 3.2 (d, J=11 Hz, 2H), 2.6 (m, 2H), 1.6
(m, 2H), 1.2 (s, 24H), 0.90 ppm (t, J=7.5 Hz, 3H); MS (ESI) m/z 433
[M+Na]+.
General procedure for the synthesis of compounds 13, 25a, and
25b (GP 4): To a mixture of aryl bromide derivative (1 equiv), palla-
dium(II) acetate (5 mol%), and Et3N (1 equiv) in anhydrous DMF
was added substituted alkene (1 equiv) successively. The reaction
mixture was held at reflux for 16 h, filtered on a bed of Celite and
extracted with EtOAc and H2O. The compound was purified by
flash column chromatography using EtOAc and hexane (1:2).
Experimental Section
Chemistry
All starting materials, reagents, and solvents were obtained from
commercial suppliers and were used without further purification.
Reactions were performed under an inert atmosphere of N2 unless
otherwise specified. Routine thin-layer chromatography (TLC) was
performed on silica gel plates (Analtech Inc., 250 mm). Flash chro-
matography was conducted on silica gel (Merck, grade 60, 230–
400 mesh). 1H NMR spectra were recorded on a Bruker ARX 300
spectrometer (300 MHz) or Varian spectrometer (500 MHz) using
[D6]DMSO and CDCl3 as solvents, and spectral data were consistent
with assigned structures. Chemical shift values (d) are reported in
ppm, coupling constants (J) are given in Hz, and splitting patterns
are designated as follows: s, singlet; d, doublet; t, triplet; q, quar-
tet; m, multiplet. MS data were collected on a Bruker ESQUIRE elec-
trospray/ion-trap instrument in the positive and negative modes.
High-resolution MS data were obtained using a Micromass Q-TOF2
mass spectrometer. Elemental analyses (C, H, N) were performed
by Atlantic Microlab Inc. (Norcross, GA, USA), and results are within
ꢀ0.4% of the theoretical values for the formula given.
(E)-Methyl-6-(dodec-1-enyl)-2-naphthoate
(25a):
1H NMR
(300 MHz, [D6]DMSO): d=8.58–7.77 overlapping protons (6H),
6.65–6.52 (m, 1H), 5.70–5.20 (m, 1H), 3.91–3.90 (s, 3H), 2.10–1.80
(m, 2H), 1.76–1.55 (m, 2H), 1.21 (s, 14H), 0.86 ppm (t, J=5.1 Hz,
3H); MS (ESI) m/z 375.3 [M+Na]+.
General procedure for the synthesis of compounds 14, 19, 26a,
26b, 31a, and 31b (GP 5): Anhydrous THF was added to LiAlH4
(3 equiv) and stirred for 5 min. To this, a solution of corresponding
methyl ester (1 equiv) in THF was added, and the mixture was
stirred at RT for 4 h. The reaction mixture was cooled to 08C, and
saturated Na2SO4 was added dropwise to the mixture. The product
was then extracted with EtOAc, and the crude product was puri-
fied using flash column chromatography which was eluted with
EtOAc and hexane mixture (1:1).
(E)-6-(Dodec-1-enyl)naphthalen-2-ylmethanol (26a): 1H NMR
(300 MHz, CDCl3): d=7.90–7.30 overlapping protons (6H), 6.60–
6.30 (m, 1H), 5.60–5.40 (m, 1H), 4.81 (s, 2H), 2.35–2.26 (m, 2H),
1.6–1.47 (m, 2H), 1.27 (s, 14H), 0.9 ppm (t, J=4.8 Hz, 3H); MS (ESI)
m/z 347.3 [M+Na]+.
General procedure for the synthesis of phosphonic acid ana-
logues (compounds 4, 10, 18, 22, 30a, 30b, 34a, 34b, 37, 41, 42
and 43) (GP 1): To a suspension of phosphonate derivative
(1 equiv) in anhydrous CH3CN, TMSBr (2.5 equiv) was added, and
the reaction mixture was held at reflux for 1 h. The CH3CN was
then evaporated under reduced pressure, and MeOH was added to
the residue and stirred for 30 min at RT. The desired product was
obtained in good yield after filtration of the MeOH solution as a
white solid.
General procedure for the synthesis of compounds 15, 27a, and
27b (GP 6): To a solution of the corresponding alkene of the
methyl ester derivative (1 equiv) in MeOH was added a catalytic
amount of Pd(OH)2/C, and the mixture was stirred at RT for 2 h
using a catalytic amount of H2 gas. The completion of reaction was
monitored by TLC, and the reaction mixture was filtered on a bed
of Celite. The filtrate was evaporated, product mass was confirmed
by MS, and the crude residue was directly used for next step.
4-Pentadecylbenzylphosphonic acid (22): Yield (0.153 g, 53%);
1H NMR (500 MHz, [D6]DMSO): d=7.14–7.06 overlapping protons
(m, 4H), 2.91 (d, J=20 Hz, 2H), 1.52 (m, 2H), 1.26 (m, 2H), 1.15 (s,
24H), 0.85 ppm (t, J=6.5 Hz, 3H); 13C NMR (500 MHz, [D6]DMSO):
d=140.2, 131.6, 129.8, 128.4, 38.1, 35.1, 31.8, 29.5, 29.3, 29.2, 22.6,
14.2 ppm; MS (ESI) m/z 381.0 [MꢁH]ꢁ; Anal. calcd for C22H39O3P: C
69.08, H 10.28, found: C 69.33, H 10.52.
6-Dodecylnaphthalene-2-carboxylic acid methyl ester (27a): MS
(ESI) m/z 355 [M+H]+.
General procedure for the synthesis of compounds 8, 16, 20,
28a, 28b, 32a, 32b, and 40 (GP 2): PBr3 (0.3 equiv) was added
slowly under stirring to a solution of alcohol derivative (1 equiv) in
anhydrous CH2Cl2 at 08C. The reaction mixture was stirred at RT for
1 h and the formation of the product was monitored by TLC. H2O
was added at 08C to quench the reaction, extracted from CH2Cl2,
and dried over MgSO4. The product mass was confirmed by MS
and the residue was carried on to the next step without further
purification.
Biology
Lysophosphatidic acid (18:1), LPC 18:1, and S1P were purchased
from Avanti Polar Lipids (Alabaster, AL, USA). For calcium mobiliza-
tion assays, LPA, S1P, and the test compounds were prepared as
1 mm stock solutions in phosphate-buffered saline (PBS) in an
equimolar complex with charcoal-stripped, fatty acid free bovine
serum albumin (BSA; Sigma, St. Louis, MO, USA). The fluorescent
ChemMedChem 2011, 6, 922 – 935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
931

G. Tigyi, D. D. Miller, et al.
MED
ATX substrate FS3 was purchased from Echelon Biosciences (Salt
Lake City, UT, USA).
Calcium mobilization assays
Assays for the mobilization of intracellular Ca2+ were performed as
described.[44,57] Test compounds up to a final concentration of
10 mm either alone or mixed with respective ~EC50 concentrations
of LPA 18:1 for each of LPAR subtype were added to cells stably
expressing LPA1.2.3.4.5, GPR87, and P2Y10 LPAR subtypes.[46] The cells
were loaded with Fura-2/AM in Krebs buffer containing 0.001%
pluronic acid for 1 h, and rinsed with Krebs buffer before measur-
ing Ca2+ mobilization. The Ca2+ responses were measured by
using a Flex Station II fluorescent plate reader (Molecular Devices,
Sunnyvale, CA, USA). The ratio of peak emissions at l=510 nm
after 2 min of ligand addition was determined for excitation wave-
lengths of 340/380 nm. All samples were run in triplicate, and
assays were performed at least twice for each receptor.
Autotaxin inhibition assays
As the first level of screen, recombinant-ATX (50 mL, 2 nm final con-
centration) in assay buffer [Tris 50 mm, NaCl 140 mm, KCl 5 mm,
CaCl2 1 mm, MgCl2 1 mm (pH 8.0)] was mixed with 25 mL FS3 (Eche-
lon Biosciences Inc., Salt lake city, UT, USA; final concentration
1 mm) and 25 mL test compound dissolved in assay buffer with
40 mm BSA in 96-well Costar black-well plates. FS3 fluorescence
(lex =485 nm, lem =538 nm) was monitored with a Flex Station II
(Molecular Devices, Sunnyvale, CA, USA) for 2 h incubation at 378C.
The differences between 0 and 120 min were calculated individual-
ly and normalized to vehicle control. The mean ꢀSD of triplicate
samples was expressed as percentage of ATX activity. The ATX ac-
tivity in the presence of the test compounds was compared with
vehicle using Student’s t-test, and p<0.05 was considered signifi-
cant.
NPP6 and NPP7 inhibition assay
Inhibition of NPP6 and NPP7 was evaluated using recombinant pu-
rified proteins, the synthetic substrate para-nitrophenylphospho-
choline (pNPPC), and a Synergy II plate reader (BioTek, Winooski,
VT) as we have previously described.[36,39] The final concentration
of each enzyme was 8.3 nm and pNPPC was 10 mm. All analogues
were tested at single 10 mm concentrations. Absorbance of liberat-
ed para-nitrophenol at l=405 nm was determined up to 1 h
(where responses were linear) and was normalized to vehicle con-
trol.
The second level of screen was aimed at determining the inhibition
of ATX activity against LPC substrate. We used the fluorescent ana-
logue ADMAN-LPC to measure the inhibition of LPA production as
previously reported.[31] ADMAN-LPC at a final concentration of
0.5 mm was mixed with or without the test compound at 3 mm and
100 nm ATX. The reaction was incubated at 378C for 4 h, and lipids
were extracted and separated by silica gel 60 TLC plates (Merck)
using CHCl3/MeOH/NH4OH (60:35:8 v/v/v). Fluorescent LPC and
LPA species were visualized in a Photodyne UV imager using the
TotalLab100 software.
MM1 hepatoma cell invasion of endothelial monolayers
The third level of screen used the nucleotide-like substrate pNP-
TMP (Sigma–Aldrich).[31] The assay contained 4 nm ATX, 5 mm pNP-
TMP, and vehicle with or without the test compound (10 mm), all
dissolved in 60 mL assay buffer. The assay was carried out in 96-
well, half-area plates (Corning Inc., Corning, NY, USA) at 378C in a
BioTek Synergy-2 plate reader (BioTek, Winooski, VT, USA) with ab-
sorbance monitoring at l=405 nm.
The highly invasive MM1 cells originally isolated from the AH130
rat hepatoma cells were a kind gift from Dr. Michiko Mukai (Osaka
University, Japan).[35,49] These cells were grown in suspension in
Dulbecco’s modified Eagle’s medium (DMEM) supplemented with
10% fetal bovine serum (FBS), 2 mm glutamine, 100 UmLꢁ1 penicil-
lin, and 100 mgmLꢁ1 streptomycin. The isolation and culture of
mesothelial cells from C57BL/6 mice was previously described.[35,49]
HUVECs obtained from VEC Technologies Inc. (Rensselaer, NY, USA)
were grown in MCDB-131 complete medium containing 10% FBS,
90 mgmLꢁ1 heparin, 10 ngmLꢁ1 EGF, 1 mgmLꢁ1 hydrocortisone,
0.2 mgmLꢁ1 EndoGrowth supplement, 100 UmLꢁ1 penicillin G,
100 mgmLꢁ1 streptomycin, and 25 mgmLꢁ1 amphotericin B (all from
VEC Technologies). Tumor cell invasion was performed by seeding
1.3ꢁ105 HUVECs (passage 7) into each well of a 12-well plate pre-
coated with 0.2% gelatin (Sigma) and cultured for two days to
form a confluent monolayer. Mesothelial cells harvested from two
to three mice were initially plated in six-well plates and grown to
confluency in DMEM medium supplemented with 10% FBS in the
presence of 100 UmLꢁ1 penicillin G and 100 mgmLꢁ1 streptomycin.
When confluent, the mesothelial cells were split into three equal
aliquots (~5ꢁ104 cells each) and plated to three wells of a 12-well
plate and grown to confluence. For all invasion assays, MM1 cells
were pre-stained with 2 mgmLꢁ1 calcein AM (Invitrogen, USA) for
2 h, rinsed once, and seeded at a density of 5ꢁ104 cells per well
over the monolayers. Tumor cells were left to invade the HUVEC
monolayer for 24 h in MCDB-131 complete media containing 1%
serum with or without addition of 1.5 mm LPC. MM1 cells were
plated on the mesothelial monolayers in 2% FBS-supplemented
DMEM in the presence of 100 UmLꢁ1 penicillin G and 10 mgmLꢁ1
streptomycin with or without 1.5 mm LPC (5 mm for the mesothelial
cells), and invasion was allowed to proceed for 20 h. The day after
MM1 cell seeding, non-invaded tumor cells were removed by five
repeat rinses of the HUVEC monolayer (three rinses for the meso-
Determination of the mechanism of autotaxin inhibition
The mechanism of ATX inhibition was determined with recombi-
nant purified human ATX and FS3, as recently described.[36,37,39]
Final ATX and FS3 concentrations were 8.3 and 1 mm, respectively,
in assay buffer containing 15 mm fatty acid free BSA. To calculate
IC50, full dose responses were determined for the test compounds.
In addition, the mechanism of inhibition of ATX-mediated FS3 hy-
drolysis was determined by varying the concentration of substrate
(0.3–20 mm) in the presence of three concentrations of each inhibi-
tor [0, 0.5ꢁ and 2ꢁ(IC50)]. Kinetic data including Vmax and KM were
determined with Kaleida Graph 4.0 (version 4.03, Synergy Software,
Reading, PA, USA) after the plots of initial velocities versus sub-
strate concentration in the absence or presence of inhibitors were
fit to the following equation: y=m1 ꢁm2 ꢁx/(1+m2ꢁx), in which
KM =1/m2 and Vmax =m1. The average KM value for ATX-mediated
FS3 hydrolysis was determined to be 2.3 mm and was used in the
following calculations. Simultaneous nonlinear regression using
WinNonLin 6.1 (Pharsight, Mountain View, CA, USA) was used to
assign the mechanism of inhibition.[36,37,39] Ki and Ki’ values (affini-
ties for free enzyme and the enzyme–substrate complex, respec-
tively) were determined by calculating the lowest averaged per-
cent residuals for each mechanism derived from curve fitting using
the Michaelis–Menten equations for competitive, uncompetitive,
mixed-mode, and noncompetitive inhibition.
932
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 922 – 935

Autotaxin Inhibitors
thelial monolayers) with PBS (containing Ca2+ and Mg2+) followed
by fixation with 10% buffered formalin. The number of tumor cells
that penetrated the monolayer was photographed under a NIKON
TiU inverted microscope using phase-contrast and fluorescence il-
lumination in a minimum of five non-overlapping fields at 100ꢁ
magnification. The fluorescent images were overlayed on the
phase-contrast images using Elements BR software (version 3.1x),
and the invaded MM1 cells showing the characteristic flattened
morphology in the plane of focus underneath the monolayer were
counted. For the invasion assay, LPC dissolved in chloroform was
dried, re-dissolved in 1 mm charcoal-stripped BSA in PBS and
added immediately to the HUVEC or mesothelium monolayer and
co-cultured with MM1 cells with or without the ATX inhibitors. The
final BSA concentration was 30 mm.
isocratic flow rate of 0.3 mLminꢁ1 [59]
. The spectra were processed
using Analyst software, version 1.5. The molecular species of LPA
and compound 22 were analyzed by multiple reaction monitoring
(MRM) in negative ion mode. The Q3 (product ion) was set at m/z
153.0 for LPA, 79.0 for compound 22, and 186.0 for compound 10.
Q1 was set for the neutral molecular ion. Quantification was done
by calculating the ratio of peak area to that of the appropriate in-
ternal standard.
Pharmacokinetic data analysis
Compound 22 plasma concentration–time data were analyzed by
non-compartmental methods. The area under the plasma concen-
tration–time curve from time 0 to infinity (AUC0ꢁ1) was calculated
by the trapezoidal rule with extrapolation to time infinity. The ter-
minal half-life (t1/2) was calculated as 0.693/lz, for which lz is the
terminal phase elimination constant. Plasma clearance (CL) was cal-
culated as CL=dosei.v./AUCi.v.0ꢁ1, where dosei.v. and AUCi.v.0ꢁ1 are
the i.v. dose and corresponding area under the curve from time 0
to infinity, respectively. The apparent volume of distribution at
equilibrium (Vdss) was calculated as Vdss =CLꢁMRT, where MRT is
the mean residence time after the i.v. bolus dose. MRT was calcu-
lated as MRT=(AUMCi.v.0ꢁ1)/(AUCi.v.0ꢁ1), where AUMCi.v.0ꢁ1 is the
area under the first moment of the plasma concentration–time
curve extrapolated to infinity.
Acute murine toxicity studies
To determine acute toxicity of compound 22 we injected it at 1,
10, and 30 mgkgꢁ1 dayꢁ1 doses intraperitoneally into groups of six
adult Swiss mice for 10 days and monitored behavior and weight
daily. The mice were sacrificed on day 14, and necropsies were
conducted for signs of organ damage.
Pharmacokinetic studies
Studies were approved by the Institute of Animal Care and Use
Committee (IACUC) of the University of Tennessee, Memphis.
Animal treatment was in accordance with regulations outlined in
the United States Department of Agriculture (USDA) Animal Wel-
fare Act (9CFR, Parts 1, 2, and 3) and the conditions specified in
the Guide for the Care and Use of Laboratory Animals. Male Spra-
gue–Dawley rats (250–275 grams) pre-cannulated with jugular and
femoral vein catheters were obtained from Charles River Laborato-
ries (Wilmington, MA, USA). The patency of the jugular and femoral
vein catheters was maintained every other day with heparinized
glycerol solution (500 IU heparin per mL final solution in 50% glyc-
erol) according to the vendor’s instructions. Animals were fed a
normal laboratory diet, allowed access to water ad libitum, and
maintained on a 12 h light–dark cycle (258C).
B16-F10 murine melanoma metastasis model
All animal procedures were approved by the Institutional Animal
Care and Use Committee at the University of Tennessee and the
M.D. Anderson Cancer Center and were consistent with the Guide
for the Care and Use of Laboratory Animals (National Institutes of
Health publication 85–23, revised 1985). Eight-week-old female
C57BL/6 mice were injected with 5ꢁ104 cells per animal via tail
veins and divided randomly into four groups. Each group then re-
ceived selective ATX inhibitor (22 or 30b) or dual ATX and LPAR
antagonist LPA-BP,[34] all at 0.5 mgkgꢁ1 per injection, or vehicle
(PBS with 1% DMSO) via intraperitoneal injection 30 min after the
B16-F10 injection and daily for an additional 10 days. Subsequently,
animals in all groups were monitored for another 10 days without
treatment. At day 21, all mice were sacrificed and lungs were har-
vested, inflated, and fixed with 10% formalin. The number of meta-
static nodules on the lung surface was counted. The number of
lung nodules was compared against vehicle-treated group by one-
way ANOVA followed by Newman–Keuls multiple comparison test
and p<0.05 was considered significant.
Compound 22 was dissolved in propylene glycol and diluted in
normal rat plasma. Rats were given a single intravenous (i.v.) bolus
dose of compound 22 (5.8 mgkgꢁ1) via the femoral vein catheter.
Blood (200 mL) was collected from the jugular vein catheter at
timed intervals (5, 15, 30, 45, 60, 120, 240, 480, and 1440 min) into
Microtainer brand tubes with lithium heparin (BD, Franklin Lakes,
NJ, USA). Samples were centrifuged (10000 g, 258C) for 3 min, and
the plasma was stored at ꢁ808C until analysis. Sample aliquots
(50 mL) were mixed with 100 mL citrate phosphate buffer (30 mm Acknowledgements
citric acid, 40 mm Na2HPO4, pH 4.0), 400 mL water-saturated buta-
nol, 200 ng internal standard (compound 10), and 100 ng LPA 17:0.
This work was supported by NIH-NCI CA 92160 (G.T.), the Van V-
Samples were centrifuged (10000 g) for 1 min, and the butanol
layer was transferred to a new tube and dried under argon gas.
Plasma concentrations were determined similarly to a previously
described method.[58] Compound 22 and LPA concentrations were
determined by LC–MS–MS using an Applied Biosystems Sciex ABI
4000 QTRAP tandem mass spectrometer (Foster City, CA, USA)
equipped with a TurboionsprayTM interface, a Shimadazu LC-
10ADvp HLPC pump (Columbia, MD, USA), and a Leap HTS PAL au-
tosampler (Carrboro, NC, USA). Samples (5 mL) were injected onto a
leet Oncology Research Fund (G.T.), the Breast Cancer Research
Foundation (N.P.), and Lpath Inc. (G.B.M.).
Keywords: phosphonic acids
·
inhibitors
·
receptors
·
structure–activity relationships
[1] M. Umezu-Goto, Y. Kishi, A. Taira, K. Hama, N. Dohmae, K. Takio, T.
Yamori, G. B. Mills, K. Inoue, J. Aoki, H. Arai, J. Cell Biol. 2002, 158, 227–
233.
C8 analytical column (Symmetry, 3.5 mm, 2.1 mm
ꢁ 100 mm;
Waters Corporation, Milford, MA, USA) with a solvent consisting of
CH3CN and H2O containing 10 mm ammonium acetate using an
[2] A. Tokumura, E. Majima, Y. Kariya, K. Tominaga, K. Kogure, K. Yasuda, K.
Fukuzawa, J. Biol. Chem. 2002, 277, 39436–39442.
ChemMedChem 2011, 6, 922 – 935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
933

G. Tigyi, D. D. Miller, et al.
MED
[3] J. M. Wu, Y. Xu, N. J. Skill, H. Sheng, Z. Zhao, M. Yu, R. Saxena, M. A. Ma-
luccio, Mol. Cancer 2010, 9, 71.
[4] M. A. Nouh, X. X. Wu, H. Okazoe, H. Tsunemori, R. Haba, A. M. Abou-
Zeid, M. D. Saleem, M. Inui, M. Sugimoto, J. Aoki, Y. Kakehi, Cancer Sci.
2009, 100, 1631–1638.
[30] H. M. Albers, A. Dong, L. A. van Meeteren, D. A. Egan, M. Sunkara, E. W.
van Tilburg, K. Schuurman, T. O. van, A. J. Morris, S. S. Smyth, W. H. Moo-
lenaar, H. Ovaa, Proc. Natl. Acad. Sci. USA 2010, 107, 7257–7262.
[31] D. L. Baker, Y. Fujiwara, K. R. Pigg, R. Tsukahara, S. Kobayashi, H. Muro-
fushi, A. Uchiyama, K. Murakami-Murofushi, E. Koh, R. W. Bandle, H. S.
Byun, R. Bittman, D. Fan, M. Murph, G. B. Mills, G. Tigyi, J. Biol. Chem.
2006, 281, 22786–22793.
[5] Y. Yang, L. Mou, N. Liu, M. S. Tsao, Am. J. Respir. Cell Mol. Biol. 1999, 21,
216–222.
[6] N. Samadi, C. Gaetano, I. S. Goping, D. N. Brindley, Oncogene 2009, 28,
1028–1039.
[32] J. Gajewiak, R. Tsukahara, Y. Fujiwara, G. Tigyi, G. D. Prestwich, Org. Lett.
2008, 10, 1111–1114.
[7] S. Vidot, J. Witham, R. Agarwal, S. Greenhough, H. S. Bamrah, G. J. Tigyi,
S. B. Kaye, A. Richardson, Cell. Signalling 2010, 22, 926–935.
[8] A. Masuda, K. Nakamura, K. Izutsu, K. Igarashi, R. Ohkawa, M. Jona, K. Hi-
gashi, H. Yokota, S. Okudaira, T. Kishimoto, T. Watanabe, Y. Koike, H.
Ikeda, Y. Kozai, M. Kurokawa, J. Aoki, Y. Yatomi, Br. J. Haematol. 2008,
143, 60–70.
[33] J. Gajewiak, R. Tsukahara, T. Tsukahara, Y. Fujiwara, S. Yu, Y. Lu, M.
Murph, G. B. Mills, G. Tigyi, G. D. Prestwich, ChemMedChem 2007, 2,
1789–1798.
[34] H. Zhang, X. Xu, J. Gajewiak, R. Tsukahara, Y. Fujiwara, J. Liu, J. I. Fells, D.
Perygin, A. L. Parrill, G. Tigyi, G. D. Prestwich, Cancer Res. 2009, 69,
5441–5449.
[9] C. G. Gaetano, N. Samadi, J. L. Tomsig, T. L. Macdonald, K. R. Lynch, D. N.
Brindley, Mol. Carcinog. 2009, 48, 801–809.
[10] A. Kehlen, N. Englert, A. Seifert, T. Klonisch, H. Dralle, J. Langner, C.
Hoang-Vu, Int. J. Cancer 2004, 109, 833–838.
[11] S. Liu, M. Umezu-Goto, M. Murph, Y. Lu, W. Liu, F. Zhang, S. Yu, L. C. Ste-
phens, X. Cui, G. Murrow, K. Coombes, W. Muller, M. C. Hung, C. M.
Perou, A. V. Lee, X. Fang, G. B. Mills, Cancer Cell 2009, 15, 539–550.
[12] M. Gugger, R. White, S. Song, B. Waser, R. Cescato, P. Riviere, J. C. Reubi,
Dis. Markers 2008, 24, 41–50.
[35] A. Uchiyama, M. Mukai, Y. Fujiwara, S. Kobayashi, N. Kawai, H. Murofushi,
M. Inoue, S. Enoki, Y. Tanaka, T. Niki, T. Kobayashi, G. Tigyi, K. Murakami-
Murofushi, Biochim. Biophys. Acta 2007, 1771, 103–112.
[36] A. B. Hoeglund, H. E. Bostic, A. L. Howard, I. W. Wanjala, M. D. Best, D. L.
Baker, A. L. Parrill, J. Med. Chem. 2010, 53, 1056–1066.
[37] A. B. Hoeglund, A. L. Howard, I. W. Wanjala, T. C. Pham, A. L. Parrill, D. L.
Baker, Bioorg. Med. Chem. 2010, 18, 769–776.
[38] E. J. North, D. A. Osborne, P. K. Bridson, D. L. Baker, A. L. Parrill, Bioorg.
Med. Chem. 2009, 17, 3433–3442.
[13] H. Sakagami, J. Aoki, Y. Natori, K. Nishikawa, Y. Kakehi, Y. Natori, H. Arai,
J. Biol. Chem. 2005, 280, 23084–23093.
[39] E. J. North, A. L. Howard, I. W. Wanjala, T. C. Pham, D. L. Baker, A. L. Par-
rill, J. Med. Chem. 2010, 53, 3095–3105.
[14] R. D. Duan, T. Bergman, N. Xu, J. Wu, Y. Cheng, J. Duan, S. Nelander, C.
Palmberg, A. Nilsson, J. Biol. Chem. 2003, 278, 38528–38536.
[15] A. Nilsson, J. Wu, B. A. Jonsson, H. Stenstad, W. Agace, Y. Cheng, R. D.
Duan, Biochem. J. 2006, 394, 299–308.
[16] M. Tanaka, S. Okudaira, Y. Kishi, R. Ohkawa, S. Iseki, M. Ota, S. Noji, Y.
Yatomi, J. Aoki, H. Arai, J. Biol. Chem. 2006, 281, 25822–25830.
[17] L. A. van Meeteren, P. Ruurs, C. Stortelers, P. Bouwman, M. A. van Rooi-
jen, J. P. Pradere, T. R. Pettit, M. J. Wakelam, J. S. Saulnier-Blache, C. L.
Mummery, W. H. Moolenaar, J. Jonkers, Mol. Cell. Biol. 2006, 26, 5015–
5022.
[40] A. L. Parrill, Biochim. Biophys. Acta 2008, 1781, 540–546.
[41] A. L. Parrill, D. L. Baker, Anticancer Agents Med. Chem. 2008, 8, 917–923.
[42] G. Ferry, N. Moulharat, J. P. Pradere, P. Desos, A. Try, A. Genton, A. Gigan-
ti, M. Beucher-Gaudin, M. Lonchampt, M. Bertrand, J. S. Saulnier-Blache,
G. C. Tucker, A. Cordi, J. A. Boutin, J. Pharmacol. Exp. Ther. 2008, 327,
809–819.
[43] Y. C. Kim, S. G. Brown, T. K. Harden, J. L. Boyer, G. Dubyak, B. F. King, G.
Burnstock, K. A. Jacobson, J. Med. Chem. 2001, 44, 340–349.
[44] G. Jiang, Y. Xu, Y. Fujiwara, T. Tsukahara, R. Tsukahara, J. Gajewiak, G.
Tigyi, G. D. Prestwich, ChemMedChem 2007, 2, 679–690.
[45] W. J. Valentine, G. N. Kiss, J. Liu, E S, M. Gotoh, K. Murakami-Murofushi,
T. C. Pham, D. L. Baker, A. L. Parrill, X. Lu, C. Sun, R. Bittman, N. J. Pyne,
G. Tigyi, Cell. Signalling 2010, 22, 1543–1553.
[18] S. Fotopoulou, N. Oikonomou, E. Grigorieva, I. Nikitopoulou, T. Papar-
ountas, A. Thanassopoulou, Z. Zhao, Y. Xu, D. L. Kontoyiannis, E. Re-
mboutsika, V. Aidinis, Dev. Biol. 2010, 339, 451–464.
[19] D. L. Baker, P. Morrison, B. Miller, C. A. Riely, B. Tolley, A. M. Westermann,
J. M. Bonfrer, E. Bais, W. H. Moolenaar, G. Tigyi, JAMA 2002, 287, 3081–
3082.
[46] J. R. Williams, A. L. Khandoga, P. Goyal, J. I. Fells, D. H. Perygin, W. Siess,
A. L. Parrill, G. Tigyi, Y. Fujiwara, J. Biol. Chem. 2009, 284, 17304–17319.
[47] F. Imamura, T. Horai, M. Mukai, K. Shinkai, H. Akedo, Jpn. J. Cancer Res.
1991, 82, 493–496.
[48] F. Imamura, T. Horai, M. Mukai, K. Shinkai, M. Sawada, H. Akedo, Bio-
chem. Biophys. Res. Commun. 1993, 193, 497–503.
[49] M. Mukai, F. Imamura, M. Ayaki, K. Shinkai, T. Iwasaki, K. Murakami-Muro-
fushi, H. Murofushi, S. Kobayashi, T. Yamamoto, H. Nakamura, H. Akedo,
Int. J. Cancer 1999, 81, 918–922.
[50] M. Mukai, H. Nakamura, M. Tatsuta, T. Iwasaki, A. Togawa, F. Imamura, H.
Akedo, FEBS Lett. 2000, 484, 69–73.
[20] R. Sutphen, Y. Xu, G. D. Wilbanks, J. Fiorica, E. C. Grendys, Jr., J. P. LaPol-
la, H. Arango, M. S. Hoffman, M. Martino, K. Wakeley, D. Griffin, R. W.
Blanco, A. B. Cantor, Y. J. Xiao, J. P. Krischer, Cancer Epidemiol. Biomarkers
Prev. 2004, 13, 1185–1191.
[21] Y. Xu, D. C. Gaudette, J. D. Boynton, A. Frankel, X. J. Fang, A. Sharma, J.
Hurteau, G. Casey, A. Goodbody, A. Mellors, Clin. Cancer Res. 1995, 1,
1223–1232.
[22] I. Dimova, S. Raitcheva, R. Dimitrov, N. Doganov, D. Toncheva, Eur. J.
Cancer 2006, 42, 674–679.
[23] N. Euer, M. Schwirzke, V. Evtimova, H. Burtscher, M. Jarsch, D. Tarin, U. H.
Weidle, Anticancer Res. 2002, 22, 733–740.
[51] N. Watanabe, H. Ikeda, K. Nakamura, R. Ohkawa, Y. Kume, T. Tomiya, K.
Tejima, T. Nishikawa, M. Arai, M. Yanase, J. Aoki, H. Arai, M. Omata, K. Fu-
jiwara, Y. Yatomi, Life Sci. 2007, 81, 1009–1015.
[24] E. Shuyu, Y. J. Lai, R. Tsukahara, C. S. Chen, Y. Fujiwara, J. Yue, J. H. Yu, H.
Guo, A. Kihara, G. Tigyi, F. T. Lin, J. Biol. Chem. 2009, 284, 14558–14571.
[25] A. A. Jazaeri, C. S. Awtrey, G. V. Chandramouli, Y. E. Chuang, J. Khan, C.
Sotiriou, O. Aprelikova, C. J. Yee, K. K. Zorn, M. J. Birrer, J. C. Barrett, J.
Boyd, Clin. Cancer Res. 2005, 11, 6300–6310.
[26] S. Tsuda, S. Okudaira, K. Moriya-Ito, C. Shimamoto, M. Tanaka, J. Aoki, H.
Arai, K. Murakami-Murofushi, T. Kobayashi, J. Biol. Chem. 2006, 281,
26081–26088.
[52] N. Watanabe, H. Ikeda, K. Nakamura, R. Ohkawa, Y. Kume, J. Aoki, K.
Hama, S. Okudaira, M. Tanaka, T. Tomiya, M. Yanase, K. Tejima, T. Nishika-
wa, M. Arai, H. Arai, M. Omata, K. Fujiwara, Y. Yatomi, J. Clin. Gastroenter-
ol. 2007, 41, 616–623.
[53] M. K. Altman, V. Gopal, W. Jia, S. Yu, H. Hall, G. B. Mills, A. C. McGinnis,
M. G. Bartlett, G. Jiang, D. Madan, G. D. Prestwich, Y. Xu, M. A. Davies,
M. M. Murph, Mol. Cancer 2010, 9, 140.
[54] J. Gierse, A. Thorarensen, K. Beltey, E. Bradshaw-Pierce, L. Cortes-Burgos,
T. Hall, A. Johnston, M. Murphy, O. Nemirovskiy, S. Ogawa, L. Pegg, M.
Pelc, M. Prinsen, M. Schnute, J. Wendling, S. Wene, R. Weinberg, A.
Wittwer, B. Zweifel, J. Masferrer, J. Pharmacol. Exp. Ther. 2010, 334, 310–
317.
[27] L. A. van Meeteren, P. Ruurs, E. Christodoulou, J. W. Goding, H. Takakusa,
K. Kikuchi, A. Perrakis, T. Nagano, W. H. Moolenaar, J. Biol. Chem. 2005,
280, 21155–21161.
[28] G. G. Durgam, R. Tsukahara, N. Makarova, M. D. Walker, Y. Fujiwara, K. R.
Pigg, D. L. Baker, V. M. Sardar, A. L. Parrill, G. Tigyi, D. D. Miller, Bioorg.
Med. Chem. Lett. 2006, 16, 633–640.
[55] G. D. Prestwich, Y. Xu, L. Qian, J. Gajewiak, G. Jiang, Biochem. Soc. Trans.
2005, 33, 1357–1361.
[29] H. M. Albers, L. A. van Meeteren, D. A. Egan, E. W. van Tilburg, W. H.
Moolenaar, H. Ovaa, J. Med. Chem. 2010, 53, 4958–4967.
[56] S. Jansen, M. Andries, K. Vekemans, H. Vanbilloen, A. Verbruggen, M.
Bollen, Cancer Lett. 2009, 284, 216–221.
934
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 922 – 935

Autotaxin Inhibitors
[57] G. G. Durgam, T. Virag, M. D. Walker, R. Tsukahara, S. Yasuda, K. Liliom,
L. A. van Meeteren, W. H. Moolenaar, N. Wilke, W. Siess, G. Tigyi, D. D.
Miller, J. Med. Chem. 2005, 48, 4919–4930.
[59] A. Tokumura, L. D. Carbone, Y. Yoshioka, J. Morishige, M. Kikuchi, A.
Postlethwaite, M. A. Watsky, Int. J. Med. Sci. 2009, 6, 168–176.
[58] H. Kosanam, F. Ma, H. He, S. Ramagiri, V. Gududuru, G. J. Tigyi, R. K. Van,
D. D. Miller, C. R. Yates, J. Chromatogr. B 2010, 878, 2379–2383.
Received: October 4, 2010
Revised: February 21, 2011
Published online on April 4, 2011
ChemMedChem 2011, 6, 922 – 935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
935