Quantitative Assay for Autotaxin Using Monoclonal

Quantitative Assay for Autotaxin Using Monoclonal Antibodies Specific to Conformational ... antibodies specific to conformational epitope on ... chrom...

0 downloads 54 Views 249KB Size

Quantitative Assay for Autotaxin Using Monoclonal Antibodies Specific to Conformational Epitope Koji IGARASHI, Kazufumi IDE, Yasutami MITOMA Masuo INOUE

Autotaxin (ATX), which is identical to lysophospholipase D (lysoPLD), is a tumor cell motility−stimulating factor originally isolated from melanoma cell supernatant and has been pointed out for its involvement in regulation of invasive and metastatic properties of cancer cells. It is also known as a multifunctional protein that is up-regulated in various malignancies including breast and lung cancer and potently stimulates cell proliferation, cell motility and angiogenesis, which is accounted for by its intrinsic lysoPLD activity that is capable of releasing lysophosphatidic acid (LPA) from lysophosphatidylcholine (LPC) or sphingosine 1−phosphate (S1P) from lysosphingomyelin. LPA is both a mitogen and a motogen that acts through G protein-associated Edg (endothelial differentiation gene) receptors and is well known to be associated with tumor aggressiveness and tumor cell−directed angiogenesis. Considering the importance of ATX as an enzyme that exerts lysoPLD activity and produces LPA, it was deemed that it might be important to measure the serum ATX. Until now ATX has been analyzed by its lysoPLD enzymatic activity or immunological methods, like a Western blotting. In this study we have produced monoclonal antibodies specific to conformational epitope on ATX and developed the ATX immunoenzymetric assay which is simple and can be performed in an automated immunoassay analyzer AIAsystem without cumbersome procedure such a sample pretreatment.

reportedly elevated in plasma of ovarian cancer patients


compared with the healthy control [16,17], while serum

Autotaxin, also known as NPP2 (nucleotide

LPA level is significantly higher in multiple myeloma

pyrophosphatase/phosphodiesterase 2), is a 125kDa

patients [18]. It is now established that these LPA

glycoprotein and has been shown in various malignant

actions are mediated through activation of its G

tumor tissues including non−small cell lung cancer [1],

protein−coupled receptors (GPCR), i.e., LPA1−3 (as the

breast cancer [2], renal cell cancer [3], and

Edg family receptors) [19,20,21] and LPA 4 /p2y9/

hepatocellular carcinoma [4] Indeed, ATX not only

GPR23 [22] and LPA5/GPR92, which are structurally

stimulates the growth of cancer [5], but also acts as a

distant from the Edg family [23]. Although the

tumor motility factor [6,7], augment the tumorigenecity

physiological functions of ATX are still unclear, ATX is

of ras−transformed cells [8] and induces a strong

widely expressed with the highest mRNA levels in

angiogenic response [9]. All biological functions of ATX

brain, placenta, ovary, and small intestine [24] and

can be explained by its ability to act as an extracellular

furthermore its mRNA levels are overexpressed in

lysoPLD [10−13]. A major substrate of ATX is LPC,

various malignancies, such as thyroid, renal cell, breast

which is hydrolyzed into LPA and choline. The

cancer, and glioblastoma [16,25−28].

bioactive lysophospholipid LPA elicits a variety of

The measurement of ATX in serum has currently

biological responses, including cancer initiation,

carried out by lysoPLD enzymatic activity assay or

progression, and metastasis [14,15]. LPA is also

immunological methods, like Western blotting.

( 37 )


TOSOH Research & Technology Review Vol.51(2007)

Unfortunately lysoPLD enzymatic activity assay is lack

Freund’s adjuvant (Difco Laboratories, MI, USA). One

of versatility because of time−consuming and tedious

month after first immunization, the fusion of B cells

method, while it is necessary to consider the

obtained from rat iliac and inguinal lymph nodes with

interference of intrinsic factors, like choline, LPC, and

PAI myeloma cells was performed. Hybridoma cells

other lysoPLDs. Likewise Western blot analysis is also

were cultured in a synthetic medium (GIT medium;

lack of versatility, and furthermore it is just a qualitative


method, not a quantitative method. In this study, we

without serum supplementation. The hybridoma

introduce a quantitative determination of ATX by an

supernatants were screened for the antibodies that bind

immunoenzymometric assay using monoclonal

to the native form ATX in solution by ELISA (Enzyme

antibodies (mAbs) specific to a conformational epitope.

Linked Immunosorbent Assay), not to ATX antigen directly coated on solid phase. Briefly, 50 ng/well of the anti−rat IgG in TBS (10 mmol/L Tris−HCl, 15 mmol/L

2.Materials and methods

NaCl, pH 7.4) was coated on a 96−well Maxisorp

2.1 Preparation of recombinant human ATX

immunoassay plate (NUNC; Nalge Nunc International,

Human cDNA for ATX was amplified by RT−PCR

NY, USA) for 16 h at 4℃. The wells were blocked with

using a human liver cDNA library as template DNA

200 μL of TBS containing 3% bovine serum albumin

base on the sequence information in the data base

(BSA). After washing the wells 3 times with TBS, the

(GenBank accession no. L46720) and was introduced

hybridoma culture supernatant was added to each well

into the baculovirus transfer vector pFASTBac−1

and incubated for 2h at room temperature. After

(Invitrogen, CA, USA). In the same way, cDNA for ATX

washing with TBS containing 0.05% Tween−20 (TBST)

was introduced into pFASTBac HT (Invitrogen) to add

4 times to remove unbound antibody, 50 ng of hisATX

a polyhistidine−tag at the NH2−terminus of ATX. The

in 50 μL of TBST containing 1% BSA (1% BSA−

preparation of recombinant baculovirus and the

TBST) was added to each well. The bound hisATX

expression of recombinant ATXs were performed

was detected by HisProbe−HRP (Pierce Chemical

according to the manufacturer’ s protocol. The

Company, IL, USA). Absorbance at 450nm by the

purification of recombinant polyhistidine−tagged ATX

addition of TMB substrate(3,3’ ,5,5’ −Tetramethylbenzidine,

(hisATX) was performed using metal chelate column

Kirkegaard & Perry Laboratories, Inc., MD, USA) was

chromatography (BD TALON; BD Biosciences, CA,

measured. All procedures were performed without

USA) according to the manufacturer’s protocol. To

hisATX for background of nonspecific bindung. The

analyze the epitope of anti−ATX mAbs, five different

hybridoma cells selected were established by a limiting

regions of ATX were expressed in Escherichia coli

dilution method. In addition, a synthetic peptide

(E.coli) JM109 using the pCold TF vector system

(DSPWTNISGSC, amino acid residue; 49−59) was

(TAKARA, Shiga, Japan) at 15℃. The five fragments,

conjugated with keyhole limpet hemocyanin (KLH),

ATX F1 (amino acid residue; 1−310), ATX F2 (amino

and an anti−peptide mAb, named P101, was established

acid residue; 301−610), ATX F3 (amino acid residue;

as a tracer antibody for detecting ATX by the

601−863), ATX F4 (amino acid residue; 150−459), and

immunization of the peptide−KLH conjugate.

ATX F5 (amino acid residue; 450−726), were expressed as fusion proteins with a trigger factor (48kDa) with


on immunoassay plate

polyhistidine−tag. To purify each ATX fragment, the sonicated supernatant of E.coli was applied onto the metal chelate column.

Reactivity of mAbs with ATX directly coated

To analyze the reactivity of anti−ATX mAbs with ATX antigen directly coated on the immunoassay plate, 50 μL of 1mg/L purified recombinant ATX was coated on

2.2 Preparation of anti−ATX mAbs

a microtiter well for 16h at 4℃. After blocking with 3%

Eight−wk−old Wister Kyoto rat was immunized by

BSA−TBS, the anti−ATX mAbs were added to each well.

the footpad injection with 250μg of purified hisATX

The antibody bound was detected by peroxidase−

antigen which was emulsified with the complete

labeled anti−rat IgG (American Qualex, CA, USA),

( 38 )

東ソー研究・技術報告 第51巻(2007)


followed by addition of TMB substrate as described

Osaka, Japan), horseradish peroxidase (TOYOBO,

above. All procedures were performed without ATX

Osaka, Japan), and TOOS reagent (N−ethyl−N−(2−

coating for background of nonspecific binding.

hydoroxy−3−sulfoproryl)−3−methylaniline; DOJINDO, Kumamoto, Japan) as a hydrogen donor. Absorbance


Reactivity profile of mAbs with ATX by

was read at 550nm and converted to the amount of

Western blotting

choline by comparison with a standard curve by choline

One microgram of the purified recombinant ATX and


the ATX fragments were subjected to SDS−PAGE. After electrophoresis, the proteins on SDS− gel were

2.6 Cross reactivity of anti−ATX mAbs with various animal ATXs

transferred to a PVDF membrane using the Semidry− Blot apparatus (BioRad Laboratories, CA, USA) at 20 2

To analyze the cross−reactivity of anti−ATX mAbs to

A/m −gel for 2h. After transferring the protein, the

ATX of various animal species, an adsorption assay by

PVDF membrane was blocked with TBS containing 1%

anti−ATX mAbs were performed using mouse, rat,

skim milk for 2h at room temperature. The anti−ATX

rabbit, horse and bovine serum. Briefly, various animal

mAb (1mg/L) was added to the membrane and the

sera were incubated with anti−ATX mAbs and the

antibody bound was detected by alkaline phosphatase−

remaining lysoPLD activity of each sample was

labeled anti−rat IgG (American Qualex), followed by

determined describe above.

addition of the CDP−Star chemiluminescense substrate

2.7 Combination of mAbs for two−site immunoenzymetric

(PerkinElmer, MA,USA).


2.5 Adsorption of ATX from human serum by anti− ATX mAbs

Anti−AXT mAbs were purified from the hybridoma culture supernatant using HiTrap Protein G (GE

To analyze an ability of absorption of ATX from

Healthcare, Buckinghamshire, UK). The purified mAbs

human serum by anti−ATX mAbs, human serum was

were then biotinylated with sulfo−NHS−LC−LC−biotin

mixed with the anti−ATX mAbs bound magnetic

reagent (Pierce) according to the manufacturer’ s

particle and the remaining lysoPLD activity in treated

protocol. Microtiter wells were coated with 100 ng of

human serum was determined. Briefly, 10μg of anti−

anti−ATX mAbs and blocked with 3% BSA−TBS. After

ATX mAbs were added to 25μL of the magnetic

blocking, 50 ng of purified recombinant ATX was added

particle suspension immobilized anti−rat IgG (BioMag

to the wells. ATX bound to primary anti−ATX mAbs was

goat anti−rat IgG Fc; QIAGEN, Hilden,


detected by the biotinylated secondary anti−ATX mAbs,

After washing the particles with TBST to remove

followed by peroxidase−labeled streptavidin (Zymed

unbound anti−ATX mAb, 50μL of human serum and

Laboratories, CA, USA) and TMB substrate described

150μL of TBST were added and the mixture was

above. All procedures were performed without ATX for

incubated for 2h at room temperature. The remaining

background of nonspecific binding. Base on the analysis

LysoPLD activity in the supernatant of the mixture was

of this screening for two−site immunoenzymometric


assay, we selected R10.23 as solid phase primary

LysoPLD activity was assessed by measuring choline liberation from the substrate LPC [5]. Briefly, the

antibody and R10.21 as enzyme−labeled secondary antibody.

reactions were performed in 100μL aliquots; the serum samples (20 μL) were incubated with 2 mmol/L 1−


Automated immunoenzymetric assay for quantitative determination of ATX

myristoyl (14:0)−LPC (Avanti Polar Lipids Inc., AL, USA) in the presence of 100 mmol/L Tris−HCl, pH 9.0,

Antigen concentration of ATX in serum was determined

500 mmol/L NaCl, 5 mmol/L MgCl2, 5 mmol/L CaCl2,

using a specific two−site immunoenzymometric assay.

and 0.05% Triton X−100 for 3 h at 37℃. The liberated

To prepare the ATX immunoenzymetric assay reagent,

choline was detected by an enzymatic photometric

R10.23 was digested with pepsin and the purified F(ab)2

method using choline oxidase (Wako Pure Chemical,

form using phenyl−5PW (TOSOH, Tokyo, Japan)

( 39 )


TOSOH Research & Technology Review Vol.51(2007)

hydrophobic column chromatography in order to avoid

recombinant ATX into the ATX depleted serum.

nonspecific binding of human antibodies to various animal IgG in human specimens, like HAMA (human anti−mouse antibodies). R10.23 F(ab)2 was coated by physical absorption on a in house magnetic bead made

3.Results and Discussion 3.1 Recombinant human ATX and ATX fragments

from ethylene−vinyl alcohol and ferrite. Coating

For the production of anti−ATX mAbs, a recombinant

efficiency of antibody on a bead was approximately 100

human hisATX expressed in baculovirus system was

ng. ATX immunoenzymetric assay reagent was

purified using BD TALON metal chelate column.

prepared as follows. Twelve R10.23−coated beads were

Finally, approximately 1 mg of recombinant human

placed in the reaction cup and then 35μg of alkaline

ATX exhibiting lysoPLD activity was obtained. The

phosphatase−labeled R10.21 in assay buffer (5% BSA,

purified ATX stained by CBB (Coomassie brilliant blue)

5% sucrose, 10 mmol/L Tris−HCl, 10 mmol/L MgCl2,

was detected as single band, whereas the doublet

pH 7.4) was added to the cup. After addition of all

bands were detected by Western blot analysis (Fig. 1).

materials, ATX immunoenzymetric assay reagent was

This result could be attributed to a different

prepared by immediate freeze−dry procedure of the

glycosylation of hisATX or a degradation of hisATX in

reaction cup. The ATX immunoassay reagent can be

the course of purification procedure. To determine the

used with an automated immunoassay analyzer AIA−

epitope of anti−ATX mAbs, five ATX fragments were

system (TOSOH).

The AIA− system includes

expressed in E.coli using the pCold TF vector system.

automated specimen dispensation, incubation of the

All fragments expressed at 15℃ were well soluble in

reaction cup, bound/free washing procedure, 4MUP

TBS and it was easy to purify them by metal chelate

(4− methylumbelliferyl


column chromatography (Fig. 2 (a)). Although epitope

dispensation, fluorometric detection, and a result

determination of each mAb was tried by Western

report. The antigen−antibody reaction time is only 10

blotting using these ATX fragments, all 23 mAbs except

min., and the first result is reported within 22 min; the

R10.48 and P101, which is anti−peptide mAb, did not

throughput of the system is 60 and 180 samples per

show any reaction with these ATX fragments (Fig. 2

hour using the AIA−600Ⅱ system and the AIA−1800

(b) , 2(c) and 2(d)). In addition, no inhibition was

system, respectively.

observed by various ATX fragments to the binding of


anti−ATX mAbs to soluble ATX antigen (data not

2.9 Preparation of standard ATX calibrator

shown). These results strongly suggested that the anti−

To purify the non−tagged ATX from the culture supernatant of Sf9 insect cells which expressed was prepared using NHS−activated HiTrap (GE Healthcare), according to the manufacturer’ s protocol.

anti−ATX mAb (P101)

anti−ATX pAb



recombinant human ATX, R10.23−immobilized column


One liter culture supernatant of the Sf9 cells was applied to the 5 mL bed volume of R10.23 column


immobilized 20 mg of antibody. The concentration of


purified ATX was determined by BCA Protein Assay Reagent (Pierce) with BSA as standard. Subsequently the serum was applied to the R10.23−immobilized column to deplete ATX in normal human serum. The pass−through fraction was collected and the remaining lysoPLD activity was measured to confirm the complete depletion of activity. ATX calibration fluids having six different concentrations (0, 0.34, 0.675, 1.35, 2.70, and 5.40mg/L) were prepared by spiking the purified ( 40 )

Fig.1 Purified hisATX and identification as ATX The purified hisATX recombinant protein (1 μg/lane) expressed in baculovirus system was subjected to SDS-PAGE. The proteins were detected by CBB staining (CBB). The purified hisATX was analyzed by Western blotting using antiATX mAb (P101) and anti-ATX polyclonal antibody (pAb) against the synthetic peptide (amino acid residue 652-666; CVRPDVRVSPSFSQN). The results of Western blot analysis with anti-ATX Abs (+) and with only secondary antibodies (-) were shown.

東ソー研究・技術報告 第51巻(2007)









(d)P101 ATX F3


ATX F1 105kDa


105kDa 75kDa





(b)R10.23 ATX F3



(a)CBB Staining

R10.1 R10.6 R10.7 R10.8 R10.9 R10.12 R10.16 R10.17 R10.20 R10.21 R10.23 R10.28 R10.30 R10.31 R10.33 R10.34 R10.36 R10.40 R10.42 R10.47 R10.48 R10.49 R10.50 P101












Fig.2 Epitope analysis of anti-ATX mAbs by Western blotting Five ATX fragments expressed in E.coli were subjected to SDS-PAGE and the proteins were detected by CBB staining (a). Reactivity of R10.23 (b), R10.48 (c) and P101 (d) as a positive control with each ATX fragment was analyzed by Western blotting. Among 23 anti-ATX mAbs, only R10.48 reacted with ATX F1 fragment.









20 30 % adsorption



Fig.3 Binding ability of anti-ATX mAbs to soluble ATX in human serum


conformational epitope of native ATX molecule.

The purified anti-ATX mAbs were incubated with the anti-rat IgG immobilized particle. Normal human serum was added to the particle and the lysoPLD activity in the supernatant of the mixture was determined. The percent adsorption [(initial activity - residual activity in supernatant) / (initial activity) x 100]was calculated. The initial activity was determined by the sample which was treated by the particle without anti-ATX mAb.

3.2 Characterization of anti−ATX mAbs We have established 23 anti−ATX mAbs by a

screening method based on binding ability to ATX

screening method based on an ability of capturing

antigen directly coated on an immunoassay plate. The

soluble ATX antigen. Sixteen of 23 mAbs were

anti−ATX mAbs obtained by this method successfully

efficiently (more than 10% adsorption) able to deplete

work only for immunoassay methods, like Western

ATX from human serum (Fig. 3). Before using the

blotting and cell staining under denatured condition,

screening method by an ability of capturing soluble

not for ELISA. In contrast, almost of all anti−ATX mAbs

ATX, we also tried to obtain anti−ATX mAbs by a

obtained by the screening method based on the ability



R10.1 R10.6 R10.7 R10.8 R10.9 R10.12 R10.16 R10.17 R10.20 R10.21 R10.23 R10.28 R10.30 R10.31 R10.33 R10.34 R10.36 R10.40 R10.42 R10.47 R10.48 R10.49 R10.50 P101 Background

R10.1 R10.6 R10.7 R10.8 R10.9 R10.12 R10.16 R10.17 R10.20 R10.21 R10.23 R10.28 R10.30 R10.31 R10.33 R10.34 R10.36 R10.40 R10.42 R10.47 R10.48 R10.49 R10.50 P101



0.2 0.3 Binding(OD450)





0.4 0.6 Binding(OD450)


Fig.4 Difference of anti-ATX mAbs binding to ATX antigen on the solid phase and soluble ATX antigen (a) The purified anti-ATX mAbs were coated on the microtiter wells. HisATX recombinant protein was added to the wells and hisATX captured by anti-ATX mAb was detected by HisProbe-HRP. (b) The purified ATX recombinant protein was coated on the microtiter wells. The anti-ATX mAbs were added to the wells and the antibody bound was detected by peroxidase-labeled anti-rat IgG.

( 41 )



TOSOH Research & Technology Review Vol.51(2007)

3.3 Preparation of ATX immunoenzymetric assay

of capturing soluble ATX antigen could specifically bind


to only soluble ATX antigen, neither to ATX antigen directly coated on an immunoassay plate (Fig. 4) nor to

To select the pairs of mAbs for two−site immunoenzy

ATX antigen on a Western blot membrane (data not

-mometric assay, all combinations of anti−ATX mAbs

shown). As shown in Fig. 3 and 4, R10.28 and R10.34

were examined. Among examined 529 combinations, 46

could deplete only a recombinant ATX, not an ATX in

pairs of antibodies could form an immunocomplex,

human serum. This result suggested that these

primary antibody − ATX − secondary antibody, and

antibodies recognize the particular epitope on the only

showed an enough binding signal for two− site

recombinant ATX. Anti−ATX mAbs showed the

immunoenzymetric assay to detect serum ATX.

different depletion profile of the lysoPLD activity from

Although R10.16, R10.48 and R10.49 showed the

various animal sera (Fig. 5) and an epitope of anti−ATX

binding to both ATX antigen on solid phase and soluble

mAbs could not be determined using ATX fragments.

ATX antigen as shown in Fig. 4, no applicable partner

These results demonstrated that anti−ATX mAbs

was found for two−site immunoenzymetric assay.



Finally we selected the combination of R10.23 as the

conformational epitope on native ATX molecule and the

primary antibody for solid phase and R10.21 as the

epitope recognized by them was abundantly diverse.

enzyme−labeled secondary antibody for two−site







R10.7 human



























% adsorption

% adsorption
























% adsorption



% adsorption





% adsorption

Fig.5 Cross-reactivity of anti-ATX mAbs with various animal ATXs Anti-ATX mAbs bound on magnetic particle were incubated with various animal sera and the remaining lysoPLD activity in serum was determined. The percent adsorption was calculated as described in Fig. 3. Typical cross-reactivity profiles of anti-ATX mAbs were shown.

( 42 )

50 % adsorption



東ソー研究・技術報告 第51巻(2007)

immunoenzymometric assay because the combination


3.4 Performance of ATX immunoenzymetric assay

of these antibodies showed a good correlation with

We examined the within−run and the between−run

lysoPLD activity in human serum. When we

reproducibility of our new serum ATX immunoenzymetric

constructed two−site immunoenzymetric assay using

assay on AIA−system. Two pooled serum samples from

these antibodies, we initially observed false ATX values

normal healthy subjects were used to examine the

frequently (20−30%) and the relationship between

within−run and the between−run coefficients of

ATX concentration determined by the two−site

variation (CVs). A serum spiked recombinant ATX was

immunoenzymometric assay and lysoPLD activity was

additionally used for the within−run assay. In the

poorly-correlated (data not shown). To solve this

within−run study, the measurement in the three

problem, both rat IgG and HBR−1 (Scantibodies,

samples was replicated 20 times. On the other hand,

CA, USA), which is well known as heterophilic

two samples were measured 20 times, in the between−

antibodies blocker, were added to the buffer of ATX

run study. The mean ± standard deviation (SD) values

immunoenzymetric assay reagent, these false values

for the three different samples in the within−run study

were completely reduced and the correlation was

were 0.59 ± 0.02, 1.25 ± 0.03 and 4.12 ± 0.10 mg/L,

surprisingly better.

while the mean ± SD values for the two samples in the

ATX expressed in baculovirus system for calibrator

between−run study were 0.76 ± 0.03 and 1.48 ± 0.04

standard was effectively purified by a R10.23−

mg/L. The within−run and between−run CVs were 2.5 −

immobilized column chromatography (Fig. 6). ATX

2.9% and 1.6 − 4.6%, respectively.

bound to the R10.23−immobilized column was eluted by

Various serum interferable substances that might

100 mmol/L glycine (pH 3.5). Whereas both the

interfere with the measurement were added to pooled

antigenic activity and the lysoPLD enzymatic activity of

serum samples (1:4 volume), followed by the ATX

ATX significantly decreased in the case of the elution

assay. Addition of up to 174 mg/L of conjugated

with 100 mmol/L glycine (pH 3.0) or more acidic pH.

bilirubin, 174 mg/L of free bilirubin, 4.4 g/L of

To prepare the ATX deficient serum for calibrator base,

hemoglobin, 50 mg/mL of human albumin, 20 g/L of

normal human serum was applied to R10.23−

triglycerides 200 mg/L of ascorbic acid, 100 mg/L of

immobilized column and the lysoPLD activity was

citric acid, 1 g/L of EDTA, 100 U/L of heparin and 5000

completely depleted from 500 mL serum by once path−

IU/L of rheumatoid factor did not affect on the

through procedure. Then, the six points of calibrator

recovery rate less than 10% in the present ATX

fluids were prepared using the purified ATX and the ATX depleted serum. The calibration curve of the AIA− 1000

system was well fitted by four-parameter logistic Binding[nmol/L/sec]

anti−ATX mAb (P101)

anti−ATX pAb



regression (Fig. 7).




105kDa 75kDa 1 0.1




Fig.6 Purified non-tagged ATX and identification as ATX The purified ATX recombinant protein (1 μg/lane) expressed in baculovirus system was subjected to SDS-PAGE. The proteins were detected by CBB staining (CBB). The purified hisATX was analyzed by Western blotting using anti-ATX mAb (P101) and anti-ATX polyclonal antibody (pAb) against the synthetic peptide (amino acid residue 652666; CVRPDVRVSPSFSQN). The result of Western blot analysis with anti-ATX Abs (+) and with only secondary antibodies (-) was shown.

Fig.7 Standard curve for the ATX immunoenzymetric assay on AIA-system The calibration fluids consisting of six different concentrations (0, 0.34, 0.675, 1.35, 2.70, and 5.40 mg/L) were prepared. Binding rate (nmol/L/ sec) of fluorescence intensity changes per second were measured and on AIA-system. Regression curve was drown by the equation; log (y) = a log(x) 3+ b log(x) 2+ c log(x) + d. The constant coefficient was as follows; a

( 43 )


TOSOH Research & Technology Review Vol.51(2007)

immunoenzymetric assay.

hardly bind to ATX directly coated on microtiter wells.

The detection limit of the ATX antigen assay was

This result strongly suggested that we could not

defined as mean + 2 SD for the 5 replicates of zero

efficiently find out the antibody for two− site

calibrator. The minimum detection limit of this assay

immunoenzymometric assay by the ordinary screening

was estimated to be 0.110 mg/L.

method base on antigen directly coated on immunoassay plate. Thus our results also suggested

3.5 Correlation study between ATX concentration and lysoPLD activity

that the screening method based on an ability of capturing native antigen is indispensable for obtaining

It has been reported that ATX accounts for serum

the antibody for immunoassay, like ELISA.

lysoPLD activity [29−30]. In fact, when human serum

Ever since ATX has been reported to be identical to

was applied to the R10.23−immobilized column, both

lysoPLD in 2002[5,31], which is producing enzyme of

ATX antigen and lysoPLD activity were completely

LPA, many scientists has been interested not only in

undetectable. Therefore ATX concentration in human

physiological functions but also as a diagnostic marker.

sera was determined on AIA−system and compared to

Although LPA has been suggested to be a serum

their lysoPLD activity. The correlation between these

biomarker for cancer metastasis by its physiological

two indices was analyzed (Fig. 8). The serum ATX

functions, it is not easy to determine the concentration

antigen was very well−correlated to the serum lysoPLD

of LPA in serum because LPA is not only unstable but

activity (r=0.924). This result demonstrated that ATX

also spontaneously generated in serum. By contrast,

exhibited almost all lysoPLD activity in human serum.

ATX is very stable in serum in comparison to the LPA

Moreover this result indicated that the serum lysoPLD

instability. In fact, ATX remains its antigenic activity

activity can be estimated by the measurement of the

after several freeze−thaw cycles (data not shown). ATX

serum ATX antigen.

concentration was determined by lysoPLD enzymatic activity assay until now but this method is time− consuming and really tedious. Whereas our new


immunoassay is easy, quantitative, and high−

In this study, we found that the antibodies selected

throughput method in comparison with the lysoPLD

based on their capture ability of soluble ATX could

enzymatic activity assay. In addition, our present result showed that the concentration of ATX antigen was well correlated with lysoPLD activity in serum. Therefore the ATX immunoenzymetric assay can be utilized for


the determination of serum ATX concentration instead

Autotaxin on AIA[mg/L]

y = 0.5947 x − 0.036 r = 0.924, n = 46

of the lysoPLD activity assay. Moreover regarding lysoPLD activity human specimens expect serum, like a


seminal fluid, spinal fluid and urine, it is difficult to determine the lysoPLD activity by their intrinsic interfering substances. Especially lysoPLD activity in 10

seminal fluid could not be determined because it contains high concentration, approximately 40 mmol/L, of choline which interferes with the lysoPLD enzymatic choline generation assay using 2 mmol/L LPC as a

0.0 0.0






substrate (data not shown). Our preliminary result

IysoPLD activity [μmol/L/min]

showed the ATX immunoenzymetric assay could determine ATX concentration in these human

Fig.8 Comparison study between the ATX concentration and the lysoPLD activity

specimens (data not shown).

ATX concentration in sera was determined by the AIA immunoenzymetric assay reagent on the AIA-system. The ATX concentration in serum samples was compared its lysoPLD activity by the linear regression method.

in the serum using an automated analyzer and aimed to

We performed the measurement of the ATX antigen

( 44 )

東ソー研究・技術報告 第51巻(2007)

adopt for clinical laboratory testing. Based on the


[9]S. Nam, T. Clair, Y. Kim, A. McMarlin, E.

results of the present study on within−run and

Schiffmann, L. A. Liotta, M. Stracke, Cancer Res.,

between−run precision, interference, detection limit, we

61, 6938−6944 (2001)

speculate that this ATX immunoenzymetric assay can

[10]W. Moolenaar, J. Cell. Biol., 158, 197−199 (2002)

be applied to clinical laboratory testing. The ATX

[11]W. Moolenaar L. van Meeteren, B. Giepmans, BioEssays, 26, 870−881 (2004)

immunoenzymetric assay, which is simple and can be performed in an automated analyzer, will contribute to

[12]A. Tokumura, J. Cell Biochem., 92, 869−881

the investigation of clinical signification of ATX as a

(2004) [13]Y. Xie, K. E. Meier, Cell Signal., 16, 975−981

novel diagnostics marker.

(2004) [14]K. Hama, J. Aoki, M. Fukaya, Y. Kishi, T. Sakai, R.


Suzuki, H. Ohta, T. Yamori, M. Watanabe, J. Chun, H. Arai, J. Biol. Chem., 279, 17634−17639 (2004)

We thank Drs. Yutaka Yatomi (University of Tokyo), Junken Aoki (Tohoku University) and Hiroyuki

[15]W. H. Moolenaar, Trends Cell Biol., 4, 213−219

Arai (University of Tokyo) for helpful discussion. We


would also like to thank Aya Nakagawa, Tomoko

[16]R. Sutphen, Y. Xu, G. D. Wilbanks, J. Fiorica, E. C.

Nishiguchi, Chigusa and Mikihisa Ohmori for their

Grendys, J. P. LaPolla, H. Arango, M. S. Hoffman,

technical support.

M. Martino, K. Wakeley, D. Griffin, R. W. Blanco, A. B. Cantor, Y. J. Xiao, J. P. Krischer, Cancer Epidemiol.Biomarkers Prev., 13, 1185−1191



[1]Y.Yang, L.Mou, N.Liu, M.S.Tsao, Am J Respir Cell

[17]Y. Xu, Z. Shen, D. W. Wiper, M. Wu, R. E. Morton,

Mol.Biol., 21, 216−222 (1999)

P. Elson, A. W. Kennedy, J. Belinson, M. Markman, G. Casey, JAMA, 280, 719−723 (1998)

[2]N.Euer, M.Schwirzke, V.Evtimova, H.Burtscher, M.Jarsch, D.Tarin, U.H.Weidle, Anticancer Res.,

[18]T. Sasagawa, M. Okita, J. Murakami, T. Kato, A.

22, 733−740 (2002)

Watanabe, Lipids, 34, 17−21 (1999)

[3]M.J.Stassar, G.Devitt, M.Brosius, L.Rinnab,

[19]J. H. Hecht, J. A. Weiner, S. R. Post, J. Chun, J. Cell Biol., 135, 1071−1083 (1996)

J.Prang, T.Schradin, J.Simon, S.Petersen, A.Kopp− Schneider, M.Zo ¨ller, Br.J.Cancer, 85, 1372−1382

[20]S. An, T. Bleu, O. G. Hallmark, E. J. Goetzl, J. Biol. Chem., 273, 7906−7910 (1998)

(2001) [4]Z. Zhao, S. Xu, G. Zhang, Zhonghua. Gan. Zang.

[21]K. Bandoh, J. Aoki, H. Hosono, J. Biol.Chem.,

Bing. Za. Zhi., 7, 140−141 (1999)

274, 27776−27785 (1999)

[5]M. Umezu−Goto, Y. Kishi, A. Taira, K. Hama, N.

[22]K. Noguchi, S. Ishii, T. Shimizu, J. Biol.Chem., 278, 25600−25606 (2003)

Dohmae, K. Takio, T. Yamori, G. B. Mills, K. Inoue, J. Aoki, H. Arai, J. Cell Biol., 158, 227−233

[23]C. W. Lee, R. Rivera, S. Gardell, A. E. Dubin, J. Chun, J. Biol. Chem., 281, 23589−23597 (2006)

(2002) [6]E. C. Kohn, G. H. Hollister, J. D. DiPersio, S. Wahi,

[24]H. Y. Lee, J. Murata, T. Clair, M. H.Polymeropoulos,

L. A. Liotta, E. Schiffmann, Int. J. Cancer, 53, 968−

R. Torres, R. E. Manrow, L. A. Liotta, M. L.

972 (1993)

Stracke, Biochem. Biophys. Res. Commun., 218, 714−719 (1996)

[7]H. Y. Lee, J. Murata, T. Clair, M. H. Polymeropolos, R. Torres, R. E. Manrow, L. A. Liotta, M. Stracke,

[25]F. N. Van Leeuwen, C. Olivo, S. Grivell, B. N.

Biochem. Biophys. Res. Commun., 218, 714−719

Giepmans, J. G. Collard, W. H. Moolenaar, J. Biol.


Chem., 278, 400−406 (2003)

[8]S. Nam, T. Clair, C. Campo, H. Lee, L. A. Liotta, M.

[26]A. Kehlen, N. Englert, A. Seifert, Int.J. Cancer,

Stracke, Oncogene, 19, 241−247 (2000)

109, 833−838 (2004)

( 45 )


TOSOH Research & Technology Review Vol.51(2007)

[27]S. Y. Yang, J. Lee, C. G. Park, S. Kim, S. Hong, H. C. Chung, S. K. Min, J. W. Han, H. W. Lee, H. Y. Lee, Clin. Exp.Metastasis, 19, 603−608 (2002) [28]Y. Kishi, S. Okudaira, M. Tanaka, K. Hama, D. Shida, J. Kitayama, T. Yamori, J. Aoki, T. Fujimaki, H. Arai, J. Biol. Chem., 281, 17492−17500 (2006) [29]M. Tanaka, S. Okudaira, Y. Kishi, R. Ohkawa, S. Iseki, M. Ota, S. Noji, Y. Yatomi, J. Aoki, H. Arai, J. Biol. Chem., 281, 25822−25830 (2006) [30]L. A. van Meeteren, P. Ruurs, C. Stortelers, P. Bouwman, M. A. van Rooijen, J. P. Prade `re, T. R. Pettit, M. J. Wakelam, J. S. Saulnier−Blache, C. L. Mummery, W. H. Moolenaar, J. Jonkers, Mol. Cell Biol., 26, 5015−5022 (2006) [31]A. Tokumura, E. Majima, Y. Kariya, K. Tominaga, K. Kogure, T. Yasuda, K. Fukuzawa, J. Biol. Chem., 277, 39436−39442 (2002)

著   者 氏名 五十嵐 浩 二 Koji IGARASHI

著   者 氏名 井 手 和 史

著   者 氏名 三 苫 惠 民

Kazuhumi IDE

Yasutami MITOMA

著   者 氏名 井 上 益 男 Masuoi INOUE

入社 昭和61年 4 月 1 日

入社 平成15年 4 月 1 日

入社 昭和59年 4 月 1 日

入社 昭和59年 9 月16日

所属 バイオサイエンス事業部

所属 バイオサイエンス事業部

所属 バイオサイエンス事業部

所属 バイオサイエンス事業部

開発部 技術開発G

開発部 技術開発G

開発部 技術開発G






( 46 )