[CANCER RESEARCH 58, 26*7-2674, June 15. I998|

Expression and Secretion of Neuroleukin/Phosphohexose Isomerase/MaturationFactor as Autocrine Motility Factor by Tumor Cells1

Yasufumi Niinaka, Sandor Paku, Arayo Haga, Nidi-nini Watanabe, and Avraham Raz2

Metastasis Research Program, Kantianos Cancer Institute, Detroit, Michigan 48201 (Y. N., S. P., A. /?.. A. H.¡:Department of Orthopedic Surgery, Gunma University School ofMedicine, Gunma, 371, Japan ¡H.W.¡;and Department of Pathology and Radiation Oncology. Wayne State University School of Medicine. Detroit. Michigan 48202 ¡A.R.I

ABSTRACT

The results obtained from fragmented protein microsequencing havesuggested that autocrine motility factor (AMF), a tumor-secreted .VI,

55,000 cytokine that regulates cell motility in vitro as well as invasion andmetastasis in vivo, is the neuroleukin (NLK)/phosphohexose isomerase(PHI)/maturation factor (MF) polypeptide. Here, we cloned, sequenced,and studied the expression, secretion, and distribution of AMF/NLK/I'lII/MI in neoplastic and their normal counterpart cells. Although both

normal and neoplastic cells express the gene product, overexpressionassociated with selective secretion of the protein was observed only intumor cells. The cDNA sequences of AMF/NLK/PHI/MF found in bothhuman cancer and normal cells were found to be identical, suggesting thatits secretion by neoplastic cells is independent of mutation or alternativesplicing. Immunohistochemical visualization has depicted AMF/NLK/I'lII/MI to be localized into tubular-like vesicles, diffusely distributed

throughout the cytoplasm and not colocalized with any particular cy-

toskeletal network. Confocal microscopic imaging had shown a partialcolocalization between AMF and its receptor i.U, 78,000 glycoprotein),especially on the malignant cell surface periphery. The results suggest thatextracellular AMF activity may be a result of the product of intracellularcleavage of a precursor polypeptide, which is overexpressed and selectively secreted through a nonclassical secretory mechanism by neoplasticcells.

INTRODUCTION

Malignant tumors are characterized by their unrestrained growth,invasion into the surrounding host tissue, and dissemination to distantorgans ( 1), implying cell migration to be a prerequisite for tumor cellinvasion. Therefore, studies that elucidate mechanisms of motileregulation and signaling should advance our understanding of tumorcell progression and be of clinical significance (2). In vivo cellmotility may be initiated by a group of secreted cytokines havingchemotactic motility induction properties (3). Among them, AMF3 is

a molecule that was originally distinguished by its ability to stimulatedirectional motility (chemotaxis) and random motility (chemokinet-ics) of the AMF-producing cells (4), via binding to its receptor, a cellsurface gp78 (5-7). This motile signal transduction pathway involvesa pertussis toxin-sensitive G-protein activation (8), inositol phosphate

production (9), receptor phosphorylation (7), and protein kinase Cactivation (10, 11).

In our recent study, partial amino acid sequencing of a purifiedmurine AMF showed sequence identity to previously cloned geneproducts: the cytokine NLK and the enzyme PHI (12). NLK is a

Received 1/22/98; accepted 4/20/98.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

'This work was supported in part by NIH Grant R01-CA51714 and by the Paul

Zuckerman Support Foundation for Cancer Research (to A. R.).2 To whom requests for reprints should be addressed, at Metastasis Research Program.

Karmanos Cancer Institute. 110 East Warren. Detroit. MI 48201. Phone: (313) 833-0715;Fax: (313) 831-7518; E-mail: raza@kci.wayne.edu.

3 The abbreviations used are: AMF. autocrine motility factor; gp78. Mr 78,000 glyco

protein; NLK, neuroleukin; PHI, phosphohexose isomerase; MF, maturation factor; ER,endoplasmic reticulum; IL, interleukin; FGF, fibroblast growth factor; pi. ¡soelectricpoint: HRPO. horseradish peroxidase; TXRD. Texas Red: ECL. enhanced chemilumines-

cence.

neutrophic factor that promotes the survival of spinal neurons andsensory neurons (13). NLK is secreted by lectin-stimulated T cells and

induces immunoglobulin secretion (14). PHI, which catalyzes theconversion of glucose-6-phosphate to fructose-6-phosphate (15), was

found to be elevated in the serum of patients with malignant tumors,including colorectal, breast, lung, kidney, and gastrointestinal carcinomas (16-21), and its serum activity was found to be correlated with

the development of métastases(21). Furthermore. AMF motile activity was detected in the urine of patients with bladder carcinoma.Overall, the presence of AMF/NLK/PHI/MF in the extracellular compartments of cancer patients suggests that, in these patients, thedisease may have progressed, and it reflects the motile activity ofinvading and disseminating cells.

Previously, we showed that the secreted murine AMF exhibited aPHI enzymatic activity, whereas the commercial available rabbit PHIinduces cell motility, supporting the initial assignment of AMF asNLK/PHI (12). Recently, the M, 54,300 MF, which mediates thedifferentiation of human myeloid leukemic cells to terminal mono-

cytic cells, was also identified as NLK/PHI (22). Thus, AMF/NLK/PHI/MF exhibits multifunctional activities.

AMF lacks a consensus secretory signal peptide (13), which governs protein secretion via the classical endoplasmic reticulum (ER)-

Golgi route, while functioning, at least in part, as a secreted solublecytokine. It is of interest that AMF is not unique in this respect:several other proteins that lack a signal peptide are secreted from cellsvia an alternative, nonclassical pathway(s), such as IL-Iß(23). Se

cretory mechanism may be controlled by mutation, alternative splicing, posttranslational cleavage, and glycosylation of the proteins.Mutations of lysosomal enzymes result in a change from cytosolicproteins to secretory proteins with modifications of their enzymaticactivity (24). The alternatively spliced variable region of fibronectin isrequired for its secretion and fibronectin dimerization during biosynthesis (25). The posttranslational cleavage and glycosylation are required for secretion of Kaposi's sarcoma FGF, in which inhibition ofglycosylation impairs Kaposi's sarcoma FGF secretion (26). In con

trast, intracellular proteins such as Mr 72,000 ER protein, proteindisulfide isomerase, monomeric sulfatransferase, and rhodanase aresecreted only when they are overexpressed in the cell (27, 28). Thusfar, no evidence of phosphorylation or glycosylation of AMF has beenreported.

There are at least two types of M, 55,000 AMF: acidic AMF andbasic AMF, with isoelectric focusing points of pi 6.5 and pi 8.5,respectively (29). Furthermore, at least three variants of PHI subunit(Mt 60,000, M, 57,000, and M, 56,000, with pi 9.1, pi 8.9, and pi 8.6,respectively) were reported in human gastrointestinal carcinoma cells(16), which may be due to a specific intracellular cleavage (15, 16).

In this report, we investigated the expression, secretion, and distribution of AMF using an AMF-specific antibody, cloned, and se

quenced human tumor and normal cDNAs that encode AMF.

MATERIALS AND METHODS

Cell Culture. The human lung diploid fibroblast cell strain IMR90 (ATCCCCL 186) and the human fibrosarcoma cell line HT1080 (ATCC CCL 121)were obtained from the American Type Culture Collection (Manassas, VA).

2667

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

TUMOR CELL EXPRESSION OF NLK/PHI/MF

The BALB/c 3T3-A31 cell line was obtained originally from Dr. A. Ben-Ze'ev, Weizmann Institute (Rehovot, Israel), and the sublines were derived as

described previously (30). Briefly, the transformed variant A31-TR was sub-cloned from cells that were capable of anchorage-independent growth. Thetumorigenic A31-TU subline arose spontaneously in irradiated mice injecteds.c. with the transformed A31-TR cells. A31-M subline was derived from apulmonary metastasis of a mouse injected i.v. with the A31-TU cells. TheA31-TU and A31-M cells were histologically classified as angiosarcoma (30).

All cells were grown in DMEM containing 10% heat-inactivated fetalbovine serum supplemented with nonessential amino acids and penicillin-

streptomycin antibiotic mixture (Life Technologies, Inc., Grand Island, NY).Cultures were maintained at 37°Cin an air-7% CO2 incubator at constant

humidity. Cells were harvested and passaged for experiments with 0.25%trypsin and 0.025% EDTA, and viability was monitored by trypan blueexclusion. To ensure maximal reproducibility, cultures were grown for nolonger than six passages after recovery from frozen stocks and monitored toprevent mycoplasma contamination.

Antibodies and Chemicals. In advance, we synthesized two independentoligo peptides identical to partial sequences for human AMF (NLK). One ofthem was effective to be an antigen for immunization. Monospecific poly-

clonal antibody directed against AMF was generated by immunization with asynthetic peptide YFQQGDMESNGKYITK, corresponding to 351-366 amino

acid sequence of the human NLK (GenBank accession no. K03515). Monoclonal antibody against gp78 (3F3A) was used either in the form of ascitesfluid or concentrated hybridoma supernatant (5-7). Monoclonal antibodies

AE1 and AE3 to the human cytokeratins (DAKO, Carpintería, CA), monoclonal antibody TUB 2.1 to the human /3-tubulin, monoclonal antibody VIM-13.2 to the human vimentin, rhodamine-labeled antiactin phalloidin, HRPO-conjugated goat antirabbit IgG antibody, HRPO-conjugated goat antiratIgG + M antibody, and -conjugated antiactin phalloidin were all purchasedfrom Sigma Chemical Co. (St. Louis, MO). FITC-conjugated goat antirabbit

IgG (H + L) antibody was purchased from Zymed (South San Francisco, CA).TXRD-conjugated goat antirat immunoglobulin (H + L) antibody was purchased from Southern Biotechnology Associates (Birmingham, AL). TXRD-

conjugated goat antimouse IgG + M+A antibody was purchased from Cappel

Organon Teknika (Durham, NC).SDS-PAGE and Western Blotting. Aliquots of the cell lysates and con

ditioned media were separated by 8% or 10% SDS-PAGE and blotted ontopolyvinylidene fluoride membranes (PVDF-plus; MSI, Westborough, MA)using a Mini-Protean II apparatus (Bio-Rad, Hercules, CA). The blots were

blocked with 5% nonfat dry milk in PBS for overnight, incubated with theanti-AMF polyclonal antibody (1:1000) or with the anti-gp78 monoclonalantibody ( 1:500) for l h at room temperature and then with HRPO-conjugated

antirabbit (1:1000) or with antirat antibody (1:1000) for l h at room temperature. The labeled bands were revealed by chemiluminescence using ECLWestern blotting detection reagents (Amersham, Arlington Heights, IL) andexposed to Kodak X-Omat film.

Cloning of Human AMF cDNA. Total cellular RNA was isolated from70-80% subconfluent IMR90 and HT 1080 cells using RNA isolator (Geno-

sys, The Woodlands, TX) according to the procedure provided by the manufacturer. The cDNA was synthesized by a reverse transcriptase from the totalRNA using random hexamers of first-stranded cDNA synthesis kit (Pharmacia,

Uppsala, Sweden). Oligonucleotide primers for PCR were designed to containthe open reading frame of the AMF (NLK) mRNA sequence reported in 1987(GenBank accession no. K03515). The sequences were as follows: sense(5'-TATTCCGAATTCGCCATGGCCGCTCTCACC) and antisense (5'-

GAGGGACTTAAGCGAGAAGAGAAAGGGGAG), which corresponded tonucleotides —4to +15 and +1708 to +1726, respectively. Each of these

primers was designed to contain an added EcoRl restriction site. The PCRconditions were as follows: 94°Cfor 10 min as an initial denaturation; 30cycles of 94°Cfor 30 s, 55°Cfor 30 s, and 72°Cfor 2 min; and 72°Cfor 10

min as a final extension. The amplified AMF/NLK/PHI/MF cDNA was digested with EcoRI, ligated to pGEM7 (Promega, Madison, WI) or pCR"

(Invitrogen, San Diego, CA), and transformed into JM109 (Promega, Madison,WI). The plasmid DNA was subcloned, amplified, purified, and subjected toautomated sequence analysis using an ABI-373 automated sequencer (Perkin-

Elmer Corp., Norwalk, CT). Alternatively, HT1080 Agtl 1 cDNA library wasscreened with an AMF-specific DNA probe, which was prepared from EcoRI

fragment DNA of the cDNA described above. Positive clones were subcloned

into pGEM7, amplified, purified, and subjected to sequence analysis byGENETYX-MAC software (Software Development Co., Ltd., Tokyo, Japan).Sequence analysis and homology search were done using GENETYX-MAC

and BLAST searches in the National Center for Biotechnology Informationdatabase (http://www.ncbi.nlm.nih.gov).

Northern Blot Analysis. The plasmid DNA, which carries the AMF cDNAsequence, was digested with £coRI.The released DNA fragment was purifiedby Qiaex II Gel Extraction Kit (Qiagen, Chatsworth, CA) and labeled with 32P

using random primed Ready-To-Go DNA labeling kit (Pharmacia, Uppsala,Sweden). A portion (20 /ig) of total RNA was denatured at 65°Cfor 5 min with

formaldehyde and formamide, electrophoresed on a 1% agarose-formaldehyde

gel, and blotted onto a nylon membrane (Magnacharge, MSI, Westborough,MA) with 10X SSC (lx SSC containing 0.15 M NaCl and 0.015 M sodiumcitrate). Prehybridization was performed overnight at 42°Cin 50% formamide,5% dextran sulfate, 5 X Denhardt's solution, 0.05 M Na2HPO4 (pH 7.0),

5 X SSC, and 0.3 mg/ml denatured salmon sperm DNA. Then, the blot washybridized overnight with a 32P-labeled DNA probe at 42°C.The membrane

was washed by vigorous agitation at room temperature for 30 min in each ofthe following solutions: 2x SSC-0.1% SDS, 0.5X SSC-0.1% SDS, and O.lxSSC-0.1% SDS. After washing, the membrane was exposed to film (KodakBiomax MR) with an intensifying screen at -80°C.

FACScan Analysis. All of the following procedures were performed at4°Cor on ice. Cells (1 X IO6 cells/ml) were suspended in PBS containing 0.9mM Ca2+, 0.5 mm Mg2+, 0.2% sodium azide, and 1% BSA and rocked for 30

min. Then, anti-AMF polyclonal antibody (1:50) was added and incubated for

l h with rocking. After the cells were washed twice with PBS containing 0.9mM Ca2"1",0.5 mM Mg2+, 0.2% sodium azide, and 0.1% BSA, the cells were

incubated with FITC-conjugated antirabbit antibody (1:10) as the secondary

antibody, with rocking. After l h of incubation, the cells were washed twice inPBS containing 0.9 mM Ca2+ and 0.5 mM Mg2+, and cell surface fluorescence

was analyzed by flow cytometry using a FACScan (Becton Dickinson, Mountain View, CA). The control cells were incubated identically, except withoutthe first antibody. A scatter window was set to eliminate dead cells and celldebris. The frequency and fluorescence profiles of the stained cells weredetermined using a laser output of 125 mV.

Double Indirect Immunofluorescence and Confocal Imaging. Cells(1 X 104 cells/well) were seeded on 18-mm glass coverslips in six-well platesand cultured overnight and washed twice with PBS containing 0.9 mM Ca2+and 0.5 mM Mg2+, followed by fixation and permeabilization with prechilled

methanol, 3.3% paraformaldehyde, and 0.5% Triton X-100. The cells were

washed briefly with PBS and blocked with 1% PCS in PBS for 30 min,followed by incubation at 4°Cwith rabbit anti-AMF (1:50) and rat anti-gp78

(1:25), mouse anticytokeratin (1:50), mouse anti-ß-tubulin (1:100), mouseantivimentin (1:50), or rhodamine-conjugated antiactin phalloidin (1:100).

After l h of incubation with primary antibodies, the cells were washed withPBS and incubated with FITC-conjugated goat antirabbit antibody (1:75) andeither TXRD-conjugated goat antirat (1:100) or TXRD-conjugated goat anti-mouse antibody (1:100) for 45 min at 4°C,followed by extensive wash with

PBS. The coverslips were then mounted on slides in 13% poly vinyl alcohol,30% glycerol and 0.6 X PBS and visualized using a Nikon Optiphot fluorescence microscope. Alternatively, the cells were processed as described above,except that they were fixed with 3.3% paraformaldehyde in PBS for 30 minafter completion of staining. The latter procedure was used to detect cellsurface antigens.

For the confocal imaging, the cells were processed as described above,except that the coverslips were mounted in SlowFade (Molecular Probes,Eugene, OR) and visualized by Zeiss 310 Laser scanning microscope. Controlsreceiving no primary antibody, a nonspecific rabbit IgG, a nonspecific rat IgG,or a nonspecific mouse IgG exhibited no background labeling. No cross-

reaction of antirabbit antibody with antimouse or antirat antibody was observed. Controls showed significant low signals compared with positive stain

ing.

RESULTS

AMF Cellular Identification and Localization. Previously, itwas suggested that malignant cells might overexpress AMF (2, 6, 7,31). However, direct quantitative evidence was lacking, due to the

2668

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

TUMOR CELL EXPRESSION OF NLK/PHI/MF

unavailability of immunological and molecular probes to detect itsexpression. Here, we synthesized two peptides (see "Materials andMethods") from the sequence of human NLK (13) and produced

monospecific polyclonal antibodies in rabbits. The titers of antiserawere routinely evaluated by ELISA and found to be highly specific tothe synthesized peptide, with high titration (1:5000) and no cross-

reactivity with fetal bovine serum proteins. To quantitate total AMFcellular expression levels, cell lysates of murine [A31 (normal) andA31M (tumorigenic)] and human [IMR90 (normal) and HT1080(tumorigenic)] fibroblasts were subjected to Western blot analysis.AMF expression was observed in all cell lysates, with intracellularmolecular weight form varied. At least five AMF variant forms (MT65,000, M, 57,000, Mr 46,000, Mr 38,000, and Mr 31,000) wereobserved under reduced conditions (Fig. \A). Neither preimmune norcontrol IgG showed these bands. Under nonreduced conditions, threeadditional proteins band were observed, which presumably correspond to the homodimers or heterodimer of either the Mr 65,000 or Mr57,000 forms or both (data not shown). Interestingly, the Mr 46,000variant form of AMF was detected only in the tumor cells (Fig. L4).However, there was no significant difference in total expression levelof AMF between tumor and normal cells, as evaluated by the ELISAdetection system. We compared these intracellular forms with thesecreted ones. Although many AMFs variants were detected in the cellextract, a single secreted major band of Mr 57,000 (reduced conditions) could be detected in the conditioned medium that is concomitant with the molecular mass of AMF (M, 55,000) and NLK (Mr56,000), as shown in Fig. IB. Under nonreducing conditions, an extraband of A/r 110,000 was observed, suggesting a dimer formation of theMT 57,000 AMF. Although several variants of AMF were detected intotal cell extracts, only the Mr 57,000 AMF variant form appears to beselectively secreted. Therefore, we focused our investigation on theexpression of A/r 57,000 AMF secreted variant form.

o_ o coS O) O

B•Ico

co97»-67»-45»-31»-•i--f*j|,_ non-Reduce—^Lft^^<382

!"P'•*

•» 65 67^<¡57-32•

•I"^

45»-<a38

31»-L

M L M

—¿—— 57

Fig. I. A, Western blot of intracellular AMF expression in A31 fibroblasts. A31Mangiosarcoma cells, IMR90 fibroblasts, and HT 1080 fibrosarcoma cells under reducedconditions, ß.Western blot of intracellular (Lanes L) and extracellular (Lanes M) AMF/NLK/PHI/MF expression in A31M cells, under nonreduced and reduced conditions. Thecells reached 70-80% confluency and were washed with PBS three times and then lysedwith cell lysis buffer or incubated with serum-free DMEM for 48 h. Fifty u.g of cell lysateand 25 /¿gof conditioned medium were subjected to 8% SDS-PAGE and Western blot andthen probed with monospecific anti-AMF antibody and HRPO-labeled antirabbit antibody. Labeled bands were revealed by ECL protein detection. No obvious difference isseen in expression of AMF between malignant and normal cells (A}. At least five variantforms (M, 65,000, M, 57,000, M, 45,000, M, 38,000, and M, 31,000) of intracellular AMFare observed (A}. However, a Mr 57,000 extracellular form is identified as a major bandsecreted into medium under either nonreduced or reduced conditions (ß).Neither preimmune nor control IgG showed these bands.

B

AMF

AMF

gp78

Fig. 2. Western blot of AMF (A and B) and gp78 expression in A31 fibroblasts. A31Mangiosarcoma cells. IMR90 fibroblasts. and HT1080 fibrosarcoma cells by reduction. Thecells reached 70-80% confluency and were washed with PBS three times and then lysedwith cell lysis buffer (A and C) or incubated with serum-free DMEM for 48 h (ß).Fiftyfig of cell lysate (A and C) and 25 fig of conditioned medium (B) were subjected to 8%SDS-PAGE and Western blot and then probed with anti-AMF antibody and HRPO-labeled antirabbit antibody (A and B} or anti-gp78 monoclonal antibody and HRPO-labeled anlirat antibody (C). Labeled bands were revealed by ECL protein detectionsystem. Expression of AMF is observed in all cell types, although A31M cells expressmore AMF than their normal counterpart cells (A). However, the expression of AMF inthe conditioned medium is observed only in A31M angiosarcoma and HT 1080 fibrosarcoma cells (ß).The expression of gp78 is enhanced in A31M angiosarcoma and HTI080fibrosarcoma cells (O.

Expression and Secretion of AMF by Normal and NeoplasticCells. To quantitate the relative AMF protein expression levels innormal and transformed cells, cell lysates of A31, A31M, IMR90, andHT 1080 cells were subjected to Western blot analysis. We found thatthe M, 57,000 AMF expression level in the murine A3 IM angiosarcoma cells was 2-fold higher than that in normal counterpart A31

fibroblasts (Fig. 2A), whereas the HT 1080 fibrosarcoma cells expressed approximately the same level of Mt 57,000 AMF as didIMR90 embryonal fibroblasts. However, AMF secretion into theconditioned media could be observed only in the A31M and HT 1080tumorigenic cells. AMF could not be detected in the conditionedmedia of the normal counterpart cells, even after prolonged exposureof the blots (Fig. 2B). The secretion level of AMF seems to beregulated at least in part by cell-cell contacts because we found that

A31M cells at confluency secreted at a significantly higher level ofthe protein than the subconfluent A3 IM cell cultures (data notshown). To investigate whether the expression level of AMF is relatedwith that of its receptor, duplicates of the above cell lysates weresubjected to quantitative Western blot analysis using specific anti-gp78 monoclonal antibody (2, 5-7). The results presented in Fig. 1C

clearly documents that the expression of gp78 was significantly higherin A31M and HT 1080 neoplastic cells, as compared with the A31 andIMR90 normal cell counterparts, respectively. This differential cellular pattern between the tumor and the normal cells of gp78 expressionwas similar to the up-regulated level of secretion of AMF by these

cells (Fig. 25).To test whether the AMF expression and secretion correspond to its

mRNA level, Northern blot analysis was preformed on RNA extractedfrom HT 1080 and IMR90 cells, using the newly cloned human AMFcDNA as a probe (see "Materials and Methods"). Unexpectedly, we

found that the expression of AMF mRNA was markedly higher(4-fold) in the HT 1080 cells than in IMR90 cells (Fig. 3), which did

not directly correspond to the differences in the total cellular levels ofthe protein (Fig. 1A). This result, however, was concomitant with the

2669

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

TUMOR CELL EXPRESSION OF NLK/PHI/MF

Oo coO) O

28S

18S»AMF

ß-actin

Fig. 3. Northern blot of AMF expression in human lung diploid IMR9Üfihroblasls andthe human fibrosarcoma HT108I) cells grown under subconfluenl condition. Total RNAwas isolated from each cell line alter 70-80% confluency was reached. A portion (20 /ig)of denatured tola! RNA was electrophoresed on a l% agarose-formatdehyde gel andblotted onto a nylon membrane After prehybridiiation. the blot was hybridi/ed with12P-labeled AMF cDNA probe. Then, the membrane was washed and hybridi/ed withi2P-labeled ß-actincDNA probe (ß-iiflini. The expression of AMF gene transcripts is

much higher (4-foldl in HT1080 cells than in IMR90 cells.

differential secretion, the result of Mr 57,000 AMF by these cells (Fig.2B). Thus, it may be suggested that AMF protein, when overex-

pressed, is probably secreted and not stored in an intracellular pool.Sequence Analysis of cDNA for AMF mRNA. The above data

indicated that total AMF mRNA expression level was enhanced intumor cells as compared with normal cells (Fig. 3), whereas theintracellular protein levels were similar (Figs. 1 and 2A), and that theoverexpressed protein fraction was secreted (Fig. 2B) and associatedwith the up-regulation of gp78 (Fig. 2C). AMF is a tumor-secreted

motility factor, NLK is a neutrophic factor (13), PHI is a ubiquitousphosphoenzyme that catalyzes the conversion of glucose-6-phosphateto fructose-6-phosphate (15), and MF mediates the differentiation of

human myeloid leukemic cells to terminal monocytic cells (22).Although these molecules have different assigned functions, they arethe product of a single gene (13, 15). The observation that normalcells did not secrete AMF raised the question concerning the reasonfor AMF/NLK/PHI/MF secretion by cancer cells and other activatedcells because secretion efficiency is known to be augmented bymutation, alternative splicing, posttranslational cleavage, and glyco-

sylation (24, 25). Previously, three variants of PHI (M, 60,000, M,57,000, and M, 56,000, with pi 9.1, pi 8.9, and pi 8.6, respectively)have been reported to be present in gastrointestinal carcinoma (16),which may reflect a specific intracellular cleavage of the maturemolecule (15. 16). It is, however, unknown why only the M, 57,000species are selectively secreted, although several variant forms ofAMF molecules were detected in the cell extracts. To address this, weinitially cloned PCR products AMF from HT 1080 and IMR90 cells.Consequently, four independent clones from HT 1080 RNA and twoindependent clones from IMR90 RNA were sequenced. We furthersequenced two independent clones from screening of HT1080 Agtl 1cDNA library. All of the above clones carried the inserts completelyidentical to known sequence of the human NLK (Fig. 4). Thus, we canexclude the possibility that the selective secretion of AMF by cancercells was due to a mutation in the coding sequence.

Cellular Localization of AMF. The binding of AMF to cell surface receptors modulates a signal transduction pathway, resulting incell motility or cell growth (5-11). Similarly, NLK has been shown to

bind to a surface component of the sensory neuron (14). The totalcellular level of AMF in A31M cells was 2-fold higher than that of the

normal A31 counterpart cells (Fig. 2A). We questioned whether thisoverexpression is localized to or associated with a particular cell's

compartment. To address this, a quantitation of AMF expression onthe cell surface in A31 and A31M cells was established by flowcytometry using FACScan analysis. As seen in Fig. 5, A31M cellsexpressed a higher amount of cell surface AMF than A31 cells,indicating the excess of AMF observed in A31M over the parentalA31 is due, in part, to its localization on the cell membrane, probablyin a bound form, to gp78.

Next, to examine in detail the intracellular distribution of AMF inthe cytoplasm, which might hint at the route of its secretion, A31,A31M, IMR90, and HT 1080 cells were sparsely seeded onto cover-slips. After 24 h of culturing at 37°C,they were fixed, permeabilized,

and labeled by immunofluorescence with anti-AMF. We focused our

analysis on single cells, especially semicircular shaped cells, becausethis cell shape was associated with the gliding mode of locomotion(32). AMF was found to be localized predominantly to perinuclearand nuclear region (Fig. 6, A, C, and F), with tubular formationextending to the short cellular axis. A possible colocalization of theseAMF tubular structures and cytoskeleton proteins was investigated bydouble indirect immunofluorescence labeling of AMF and ß-tubulin,vimentin, or actin. ß-Tubulin was diffusely expressed in the cyto

plasm, forming a clear visible meshwork that was well polarized inA31M cells (Fig. 6B). Vimentin was predominantly localized to theperinuclear region, forming fine tubules extending to the cell periphery, some of which overlapped with AMF-containing tubules with no

obvious colocalization (Fig. 6£>).Actin fibers were extending alongthe long cellular axis, which was different from the AMF distributionpattern (Fig. 6F).

Colocalization of AMF with gp78. Flow cytometry analysis haveshown AMF to be present on the cell surface of transformed cells ata higher density, as compared with their normal counterpart cells (Fig.5), which led us to question its possible colocalization with gp78 onthe cell surface. To examine this, A31, A31M, IMR90. and HT 1080cells were sparsely seeded on coverslips and consecutively labeledwith anti-AMF and anti-gp78 antibodies, followed by fixation and

visualization by fluorescence confocal imaging. AMF was found to belocalized to a restricted cell surface area polarized to one cell pole andin some clusters in the periphery on the A31M cell surface. Similarly,the cell surface distribution of gp78 is to a certain peripheral regions,forming a strong signal dense spot. The labeling of these two antigenssignificantly overlapped (Fig. 1A). On the normal A31 cell surface,the same type of colocalization of AMF with gp78 was observed,although the signal was several-fold weaker compared with A31M

cells (Fig. 1A). To examine the colocalization in the cytoplasm,double indirect immunofluorescence labeling of AMF and gp78 wasperformed after fixation and permeabilization. Contrary to the localization on the cell surface, AMF was found diffusely throughout thecytoplasm; similarly, gp78 was observed to be diffused throughout thecytoplasm. However, from the confocal imaging of AMF with gp78(Fig. 1A), it became evident that in both A31 and A31M cells a partialcolocalization (yellow) of AMF with gp78 could be observed, although the signals in the normal A31 cells were much weaker andclosely restricted to one spot, as compared with its transformed cellcounterpart (Fig. 1C). These results suggested that migrating cells,even normal cells as well as malignant cells, indicated colocalizationof AMF with AMF receptor gp78 and the existence of some interaction of AMF with gp78. The colocalization of extracellular AMF with

2670

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

TUMOR CELL EXPRESSION OF NLK/PHI/MF

-4 cgccatggccgctctcacccgggacccccagttccagaagctgcagcaatggtaccgcgagcaccgctccgagctgaacctgcgccgcct 86l MAALTRDPQFQKLQQWYREHRSELNLRRL 29

87 cttcgatgccaacaaggaccgcttcaaccacttcagcttgaccctcaacaccaaccatgggcatatcctggtggatcactccaagaacct 17630 FDANKDRFNHFSLTLNTNHGHILVDYSKNL 59

177 ggtgacggaggacgtgatgcggatgctggtggacttggccaagtccaggggcgtggaggccgcccgggagcggatgttcaatggtgagaa 26660 VTEDVMRMLVDLAKSRGVEAARERMFNGEK 89

267 gatcaactacaccgagggtcgagccgtgctgcacgtggctctgcggaaccggtcaaacacacccatcctggtagacggcaaggatgtgat 35690 I ^^^^^ GRAVLHVALR ^.¿^^^ TPI LVDGKDVM 119

N-Gly N-Gly

357 gccagaggtcaacaaggttctggacaagatgaagtctttctgccagcgtgtccggagcggtgactggaaggggtacacaggcaagaccat 446120 PEVNKVLDKMKSFCQRV R5GDWKGY T G K T. I 149

TK-P PKC-P CK2-P

447 cacggacgtcatcaacattggcattgtcggctccgacctgggacccctcatggtgactgaagcccttaagccatactcttcaggaggtcc 536150 I B VINIGIVGSDLGPLMVTEALKPYSSGGP 179

537 ccgcgtctggtatgtctccaacattgatggaactcacattgccaaaaccctggcccagctgaacccggagtcctccctgttcatcattgc 626180 R V W Y V SNIP GTHIAKTLAQLNPESSLFIIA 209

CK2-P

627 ctccaagacctttactacccaggagaccatcacgaatgcagagacggcgaaggagtggtttctccaggcggccaaggatccttctgcagt 716210 S K T F T T O E T I T N A E TAKE WFLQAAKDPSAV 239

CK2-P CK2-P PKC-P/CK2-P

717 ggcgaagcactttgttgccctgtctactaacacaaccaaagtgaaggagtttggaattgaccctcaaaacatgttcgagttctgggattg 806240 AKHFVALSTN T T K VKEFGIDPQNMFEFWDW 269

tWU^PKC^P

807 ggtgggaggacgctactcgctgtggtcggccatcggactctccattgccctgcacgtgggttttgacaacttcgagcagctgctctcggg 896270 VGGRYSLWSAIGLSIALHVGFDNFEQLLSG 299

897 ggctcactggatggaccagcacttccgcacgacgcccctggagaagaacgcccccgtcttgctggccctgctgggtatctggtacatcaa 986300 AHWMDQHFRT T P L E KNAPVLLALLGIWYIN 329

CK2-P

987 ctgctttgggtgtgagacacacgccatgctgccctatgaccagtacctgcaccgctttgctgcgtacttccagcagggcgacatggagtc 1076330 CFGCETHAMLPYDQYLHRFAAYFQQGDMES 359

1077 caatgggaaatacatcaccaaatctggaacccgtgtggaccaccagacaggccccattgtgtggggggagccagggaccaatggccagca 1166360 NGKYITKSG T R V D HQTGPIVWGEPGTNGQH 389

CK2-P

1167 tgctttttaccagctcatccaccaaggcaccaagatgataccctgtgacttcctcatcccggtccagacccagcaccccatacggaaggg 1256390 AFYQLIHQGTKMIPCDFLIPVQTQHPIRKG 419

1257 tctgcatcacaagatcctcctggccaacttcttggcccagacagaggccctgatgaggggaaaatcgacggaggaggcccgaaaggagct 1346420 LHHKILLANFLAQTEALMRGK S T E E A R K E L 449

CK2-P1347 ccaggctgcgggcaagagtccagaggaccttgagaggctgctgccacataaggtctttgaaggaaatcgcccaaccaactctattgtgtt 1436

450 Q A A G K SPED LERLLPHKVFEGNRPTNSIVF 479CK2-P

1437 caccaagctcacaccattcatgcttggagccttggtcgccatgtatgagcacaagatcttcgttcagggcatcatctgggacatcaacag 1526480 TKLTPFMLGALVAMYEHKIFVQGI IWDINS 509

1527 ctttgaccagtggggagtggagctgggaaagcagctggctaagaaaatagagcctgagcttgatggcagtgctcaagtgacctctcacga 1616510 FDQWGVELGKQLAKKIEPELDGSAQV T S H D 539

CK2-P

1617 cgcttctaccaatgggctcatcaacttcatcaagcagcagcgcgaggccagagtccaataaactcgtgctcatctgcagcctcctctgtg 1706540 ASTNGLINFIKQQRE [A R VI Q * 558

MCT

1707 actcccctttctcttctcgt 1726

Fig. 4. Nucleotide sequence of AMF cDNA and predicted prolein sequence. The coding region is shown from position +1 to + 1674. Underlined ¡filers,potential phosphorylationsites; TK-P. tyrosine kinase phosphorylation site: PKC-P. protein kinase C phosphorylation site; CK2-P. casein kinase II phosphorylation site; double-underlined tellers, potentialA/-glycosylationsites (N-GIyi: boxed tellers, potential microbodies COOH-terminal targeting sequence (MCT}.

gp78 on the cell surface was strong and likely to be restricted or malignant potential (33, 34). Correlation of overexpression of AMFpolarized (Fig. 7/4), whereas the colocalization of intracellular AMF receptor with cell motility and experimental metastasis has beenwith gp78 was a little weak but much more diffusely observed (Fig. reported (2, 5-8), suggesting that gp78 is involved in invasion during1C). These results suggested that AMF interaction with its receptor tumor cell metastasis (8). This thesis has been supported by manymight predominantly occur on the cell surface and be incorporated clinical observations in patients with bladder carcinoma (35), colo-into the cytoplasm. rectal cancer (36), esophageal squamous cell carcinoma (37), cutane

ous malignant melanoma (38), and gastric cancer (39). In thoseDISCUSSION clinical studies, it was found that overexpression of gp78 was corre

lated with poor patient survival. Similarly and independently of theCell motility is an integral aspect of tumor invasion and metastasis, gp78 studies, it has been reported that the expression and activity of

suggesting that differential exploitation of migration-associated pro- AMF may be of prognostic value in cancer patients (2, 6-8, 31). Acesses might endow a tumor cell with a correspondingly greater direct analysis of the relationship between AMF and gp78 was not

2671

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

TUMOR CELL EXPRESSION OF NLK/PHI/MF

A31M

Relative fluorescenceFig. 5. Flow cytometry of AMF expression on the cell surface unaly/ed by FACScan.

A31 and A3IM cells (I X IO6 cells/ml) grown under subconfluent conditions were

incubated with ami-AMF antibody (1:50) for l h and with FITC-conjugated antirabbitantibody (1:10) for l h (shtuled fieuks). Then, cell-surface fluorescence was analyzed by

FACScan. The control cells were incubated identically without the primary antibody (leftsoliti ¡lenk). A scatter window was set to eliminate dead cells and cell debris. Thefrequency and fluorescence profiles of the stained cells were determined using a laseroutput of 125 mV. A31M cells (right shaded peak) express more of AMF on their cellsurface than do A31 cells (left slttided peak).

feasible until now, due to lack of immunological and molecularprobes. Here, for the first time, total cellular AMF protein expressionand secretion were evaluated by quantitative Western blot analyses,and the AMF gene products were identified and quantitated by Northern blot analysis. Northern blot analysis revealed a significant AMFmRNA overexpression by HT 1080 fibrosarcoma, as compared with

the IMR90 normal fibroblasts (Fig. 3). However, the total cellularAMF protein levels did not correspond to the mRNA levels (Figs. 1and 2A). To the contrary, secretion level of AMF/NLK protein into theconditioned medium was up-regulated only by A31M angiosarcoma

cells and HT 1080 fibrosarcoma cells, suggesting that the secretion ofAMF was up-regulated by malignant cells (Fig. IB). Furthermore, thesecretion of AMF was seen in accordance with up-regulation of its

receptor expression (Fig. 2C). The results shown here report that theoverexpression and secretion of AMF were restricted to cancer cells.Serum PHI was shown to serve as a tumor marker for monitoringpatients with malignant progressing tumors (16-21). The results re

ported here give credence to the above PHI studies and togethersuggest that monitoring of AMF in tissue, serum, and urine might bean important tool for clinical prognosis.

In this study, we showed AMF to be colocalized with gp78 on thecell surface (Fig. 7). Similarly, NLK has been shown to bind to a cellsurface component of the sensory neuron (14). NLK, which is partially homologous to the envelope protein gpl20 of HIV, is a neutro-

phic factor, which promotes the survival of spinal neurons and sensoryneurons (13). This activity is blocked by gpl20 (40). Lectin-stimu-lated T cells secret NLK, which induces the secretion of immuno-

globulin by human monolayer cells (14). A maturation inducer hasbeen shown to mediate the differentiation of human myeloid leukemiccells to terminal monocytic cells (22). It is possible that the receptor(s)for NLK or MF of cell lineages may be different from gp78 becausethese factors exert functions different from motile stimulation. Alternatively, these factors may interact with gp78, followed by a differentcascade(s) for individual cell lineage signaling pathway, becauseAMF, NLK, and MF are derived from the same gene, representing

A31 A31M

Fig. 6. Expression and locali/ation of AMF (A, C. and E) compared with ß-tubulin(B), vimentin (D), or actin (Fi. Sparsely seeded A31 and A31M cells were fixed and permeabilizedand then subjected to double indirect immunofluorescence. AMF/NLK/PHI/MF was visualized with FITC (A, C. and £).ß-tubulin(B), and vimentin (D) with TXRD and actin (F)with rhodamine isothiocyanate. as detailed in "Materials and Methods." AMF/NLK/PHI/MF is diffusely expressed throughout the cytoplasm with tubular formation in the periphery,

which partially overlaps with other proteins. However, no colocali/ation with any other cyloskeleton proteins is observed.

2672

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

TUMOR CELL EXPRESSION OF NLK/PHI/MF

A31 A31M

Fig. 7. Confocal imaging of AMF igreenìwith gp78 (red). Sparsely seeded A31 and A31M cells were stained with anli-AMF rabbit antibody and anti-gp78 rat monoclonal antibodyand then with antirabbil-FITC (green} and antirat-TXRD (red) secondary antibodies. In the cell surface staining, fixation was performed after staining (A and B). In the permeabilizedstaining, fixation and permeabilization were performed before staining (C and D). The primary antibodies were omitted in the control cells (B and D). Colocalization (W//ÕMV)of AMFand gp78 is significantly observed on A31M cell surface, which is well polari/ed. On A31 cell surface, the colocalization is much weaker and more restricted to one spot comparedwith A3 IM. On the contrary, more diffuse and weaker overlapping is observed in the cytoplasm.

modifications of the same protein products, as determined by molecular weight on SDS-PAGE (12, 15, 22). Thus far, the receptor(s) for

NLK or MF on their corresponding cells have not been identified.AMF/NLK/PHI/MF protein, as shown here and previously, is

found in a variety of different sizes. Here, at least five variant formswere detected intracellularly, under reduced conditions with molecular weights of Mr 65,000, Mr 57,000, Mr 46,000, Mr 38,000, and Mr31,000. The only difference in expression between tumor and normalcells was the up-regulation of the MT 46,000 intracellular form in

malignant cells. Although its function is unknown, it is probably atruncated nonsecreted form of PHI. The secreted extracellular form ofAMF is detected as a single band, the estimated size of which was M,57,000 (Fig. 1). No other variant form was found in the conditionedmedium at a detectable level, except for a A/r 110,000 dimer form ofthe M, 57,000 subunit under nonreduced conditions, which was dissociated by reducing conditions (Fig. 1). This observation is in accordance with the previously reported molecular weights of AMF [Mr55,000 (4)], NLK [Mr 56,000 (13)], and MF [M, 54,300 (22)]. Although the intracellular expression of MT 57,000 form was not preferentially up-regulated in tumor cells, the secretion of this form was

found to be enhanced only in tumor cells (Fig. 2, A and B).The secretory efficiency of proteins by cells may be altered by

mutation (24), alternative splicing (25), posttranslational cleavage(26), or glycosylation (27). In contrast, intracellular proteins, such asER protein, protein disulfide isomerase, monomeric sulfatransferase,and rhodanase. are secreted only when they are overexpressed (27,28). Thus far, no phosphorylation or glycosylation of AMF has beenreported. We investigated the possible genetic alterations, such asmutation, deletion, insertion, or alternative splicing of AMF, thatmight affect or explain its secretion, but we could not find any geneticalteration in its sequence. These results suggested that the molecularweight variants, which were observed in Fig. 1, were the result of anintracellular cleavage, as predicted previously (15, 16).

AMF peptide lacks a secretory signal peptide (13), which is criticalto the secretion via the classical ER-Golgi route. The proteins devoid

of signal peptide were thought to be released by death and lysis of asmall fraction of the cultures cell population or by transient membranedisruption (41). However, the selective release of Afr57,000 AMF into

the conditioned media cannot be explained by this manner. Therefore,Mr 57,000 AMF is thought to be actively secreted via a novel alternative pathway, which has been indicated in the secretion of theproteins lack of secretory signal sequence, such as IL-1 (23, 42), FGF(43), galectin-1 (44), galectin-3 (45, 46), and others. An enhancement

of secretion by calcium ionophore A23187 or by heat shock iscommonly observed, and the secretion is not inhibited by drugs thatblock ER-Golgi transport or by multidrug-resistant proteins (23, 42-

46). The calcium ionophore A23187 enhances this alternative secretory pathway and exocytosis; however, it does not convert nonsecre-

tory cells into secretory cells (42). Thus far, a consensus sequence (ifany) shared by proteins using the alternative secretory pathway is notknown. Therefore, there must be certain cascades of regulation steps,which may be controlled by various molecules and mechanisms. Forexample, the secretion of FGF-2 is inhibited by serum starvation (43),whereas the secretion of FGF-1 is induced by serum starvation (47),

suggesting the existence of reciprocal regulatory mechanisms. Theresults of deletion and mutation experiments have shown that theprecursor peptide sequence controls the secretion of IL-1/3, which

differs depending on the cell type (48). This regulation is thought tobe controlled by a conformational change, which affects the interaction with the regulatory molecules (48). Together, the current hypothesis is that the secretion of AMF might be initiated by a specificcleavage of the precursor protein, which results in a specific conformational change, leading preferentially to a selective secretion. It ispossible, but not yet tested, that other members of the ectoenzyme/exoenzyme group (49), including other phosphoenzymes like thymi-dine phosphorylase (50) and phosphodiesterase-homologous auto-

taxin (49) shown to induce motility, undergo specific cleavage prior tosecretion. In addition, it was reported that the platelet basic proteins,the cysteine-X-cysteine chemokines connective tissue activating pep-tide-III and neutrophil activating peptide-2, are also heparin/heparansulfate-degrading enzymes (heparanases; Ref. 51) that can exert mi-

togenic activities (52). It was, therefore, suggested that there may bea complex relationship in which each activity could alter the activityor bioavailability of the other function(s) in various pathologicalsituations (51). Thus, the multifunctional activities that AMF/PHI/NLK/MF ascribes to a single polypeptide are not unique. Whether the

2673

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

TUMOR CELL EXPRESSION OF NLK/PHI/MF

PHI activity is critical to AMF, NLK, and MF activities or theyfunction independently of the enzymatic activities under normal andpathological situations is as yet unknown.

In summary, the results suggest that extracellular AMF activity maybe a resulting product of intracellular cleavage of a precursor polypep-tide, which is overexpressed and selectively secreted through a non-

classical secretory mechanism by neoplastic cells and exerts its activity following cell surface binding. Experiments aimed at unveilingthe regulation of AMF expression and mode of secretion are presentlyunderway.

ACKNOWLEDGEMENTS

We thank Victor Hogan and Vivian Powell for typing and editing thismanuscript.

REFERENCES

1. Liolta. L. A.. Rao. C. N.. and Barsky, H. Tumor invasion and the extracellular matrix.Lab. Invest., 49: 636-649. 1983.

2. Nahi, 1. R., Watanabe. H., and Raz. A. Autocrine motility factor and its receptor: rolein cell locomotion and metastasis. Cancer Metastasis Rev., //: 5-20, 1992.

3. Sloker. M., and Gherardi, E. Regulation of cell movement: the motiiity cylokines.Biochim. Biophys. Acta, IÜ72:81-102, 1991.

4. Liotta. L. A., Mandler. R.. Murano. G., Katz, D. A., Gordon, R. K.. Chiang. P. K., andSchiffmann. E. Tumor cell autocrine motility factor. Proc. Nati. Acad. Sci. USA, 83:3302-3306, 1986.

5. Nabi. I. R.. Walanabe. H.. and Raz. A. Identification of B16-F1 melanoma autocrinemotility-like factor receptor. Cancer Res., 50: 409-414, 1990.

6. Sillctti. S., Watanabe, H.. Hogan. V., Nabi. I. R.. and Raz, A. Purification of B16-F1melanoma uutocrine motility factor and its receptor. Cancer Res.. 51: 3507-3511,

1991.7. Watanabe, H., Carmi, P., Hogan, V.. Raz, T., Silletti, S.. Nabi, I. R., and Raz, A.

Purification of human tumor cell autocrine motility factor and molecular cloning ofits receptor. J. Biol. Chem., 266. 13442-13448, 1991.

8. Watanabe. H.. Nabi. I. R., and Raz. A. The relationship between motility factorreceptor internalization and the lung colonization capacity of murine melanoma cells.Cancer Res., 51: 2699-2705, 1991.

9. Kohn, E. C., Liotta, L. A., and Shiffman. E. Autocrine motility factor stimulates athree-fold increase in inositol phosphate in human melanoma cells. Biochem. Biophys. Res. Commun., /66: 757-764. 1990.

10. Silletti. S., Timer, J., Honn, K. V., and Raz, A. Autocrine motility factor inducesdifferential 12-lipoxygenasc expression and activity in high- and low-metastaticK1735 melanoma cell variants. Cancer Res.. 54: 5752-5756. 1994

11. Timar, J., Silletti. S., Bazaz, R., and Honn, K. V. Regulation of melanoma cellmotility by the lipoxygenase metabolite 12(5)-HETE. Int. J. Cancer, 55.- 1003-1010,

1993.12. Watanabe, H., Takehama. K.. Date, M., Shino/aki, T., and Raz. A. Tumor cell

motility factor is the neuroleukin/phosphohcxose ¡somerasepolypeptide. Cancer Res.,56: 2960-2963. 1996.

13. Gurney, M. E., Heinrich, S. P., Lee, M. R., and Yin, H-S. Molecular cloning andexpression of neurolcukin, a neurotrophic factor for spinal and sensory neurons.Science (Washington DC), 234: 566-574, 1986.

14. Gumey, M. E., Apatoff, B. R.. Spear. G. T., Baumel, M. J., Aniel, J. P., Bania. M. B..and Reder. A. T. Neuroleukin: a lymphokine product of lectin-stimulated T cells.Science (Washington DC), 234: 574-581, 1986.

15. Chapul. M., Claes, V., Portetene, D., Cludts. I., Cravador. A., Bumy, A., Gras, H.,and Tartar. A. The neurotrophic factor ncuroleukin is 90% homologous with phos-phohcxose isomerase. Nature (Lond.), 332: 454-455. 1988.

16. Baumann, M., and Brand. K. Purification and characterization of phosphohexoseisomerase from human gastrointestinal carcinoma and its potential relationship toneuroleukin. Cancer Res., 48: 7018-7021, 1988.

17. Baumann. M., Kappel. A.. Brand. K.. Siegfield. W.. and Paterok, E. The diagnosticvalidity of the serum tumor marker phosphohexose isomerase (PHI) in patients withgastrointestinal, kidney, and breast cancer. Cancer Invest.. 8: 351-356, 1990.

18. Fuella. X.. Molina. R., Jo. J., Mas. E.. and Ballesta. A. M. Serum phosphohexoseisomerase activities in patients with colorectal cancer. Tumor Biol.. 12: 360-367.

1991.19. Patel. P. S.. Rawal, G. N. Rawal. R. M., Palei. G. H.. Balar. D. B.. Shah. P. M., and

Patel, D. C. Comparison between serum levels of carcinoembryonic antigen, sialicacid and phosphohexose isomerase in lung cancer. Neoplasia, 42: 271-274, 1995.

20. Bodansky, O. Serum phosphohexose isomerase in cancer. II. As an index of tumorgrowth in metastatic carcinoma of the breast. Cancer (Phila.), 7: 1200-1226, 1954.

21. Shwar/., M. K. Laboratory aids to diagnosis: enzymes. Cancer (Phila.). 37: 542-548.

1976.22. Xu, W., Seiter. K., Feldman. E., Ahmed. T.. and Chiao, J. W. The differentiation and

maturation mediator for human myeloid leukemia cells shares homology with neuroleukin or phosphoglucose isomerase. Blood. 87: 4502-4506, 1996.

23. Rubarteli!, A., Cozzolino, F.. Talio, M., and Sitia, R. A novel secretory pathway forintcrleukin-lp, a protein lacking a signal sequence. EMBO J.. 9: 1503-1510. 1990.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

Ebert, D. L., Jordan. K. B., and Dimond, R. L. Lysosomal enzyme secretory mutantsof Dictyostelium discoideum. J. Cell Sci., 96: 491-500, 1990.Schwarzbauer. J. E.. Spencer, C. S., and Wilson. C. L. Selective secretion ofalternatively spliced fibronectin variants. J. Cell Biol.. 109: 3445-3453, 1989.Delli-Bovi, P., Curatola. A. M., Newman. K. M., Sato, Y., Moscatelli, D.. Hewick,R. M.. Rifkin, D. B., and Basilico. C. Processing, secretion, and biological propertiesof a novel growth factor of the fibroblast growth factor family with oncogenicpotential. Mol. Cell. Biol., 8: 2933-2941. 1988.Dorner. A. J., Wasley. L. C.. Raney. P.. Haugejorden. S.. Green. M., and Kaufman, R. J.The stress response in Chinese hamster ovary cells. Regulation of ERp72 and proteindisulfide isomerase expression and secretion. J. Biol. Chem.. 265: 22029-22034. 1990.

Sloan. I. S., Horowitz, P. M., and Chirgwin. J. M. Rapid secretion by a nonclassicalpathway of overexpressed mammalian mitochondrial rhodanese. J. Biol. Chem., 269:27625-27630. 1994.Watanabe. H.. Kanbe. K.. and Chigira. M. Differential purification of autocrinemotility factor derived from a murine protein-free fibrosarcoma. Clin. Exp. Metastasis, 12: 155-165, 1994.Zvibel. I., and Raz. A. The establishment and characterization of a new Balb/cangiosarcoma tumor system. Int. J. Cancer. 36: 261-272, 1985.Niinaka, Y.. Oida, S.. Ishisaki, A., Takeda, K., limura. T., Maruoka, Y., Momose, F.,Negishi, A., Ichijo, H., Amagasa, T.. Sasaki, S., Watanabe, H., and Raz, A. Autocrinemotility factor and its receptor expressions in oral squamous cell carcinoma (SCC)cells. Int. J. Oncol., 9: 433-438, 19%.Lee, J., Ishihara, A., Theriot, J. A., and Jacobson, K. Principles of locomotion forsimple-shaped cells. Nature (Lond.). 362: 167-171, 1993.

Benlimame, N.. Sim,ml. D.. and Nabi, I. R. Autocrine motility factor receptor is amarker for a distinct membranous organelle. J. Cell Biol.. 729: 459-471. 1995.Doyle, G. E.. Sharie!', Y., and Mohler, J. L. Prediction of metastatic potential by

cancer cell motility in the dunning R-3327 prostatic adenocarcinoma in vitro model.J. Urol., 147: 514-518, 1992.Otto, T., Birchmeier, W.. Schmidt, U., Hinke, A., Schipper, J., Rübben,H., and Raz,A. Inverse relation of E-cadherin and autocrine motility factor receptor expression asprognostic factor in patients with bladder carcinoma. Cancer Res., 54: 3120-3123,

1994.Nakamori, S., Watanabe. H., Kameyama, M.. laoka, S.. Funikawa. H.. Ishikawa, O.,Sasaki, Y., Kabuto, T., and Raz. A. Expression of autocrine motility factor receptorin colorectal cancer as a predictor for disease recurrence. Cancer (Phila.), 74:1855-1862. 1994.Maruyama, K., Watanabe, H.. Hitoshi, S.. Takayama. T., Gofuku, J.. Yano. H., Inoue, M..Tamura, S., Raz, A., and Monden, M. Expression of autocrine motility factor receptor inhuman esophageal squamous cell carcinoma. Int. J. Cancer, 64: 316-321, 1995.

Nagai. Y., Ishikawa, O.. Miyachi. Y., and Watanabe, H. Expression of autocrinemotility factor receptor in cutaneous malignant. Dermatology, 792: 8-11, 1996.

Hirono. Y., Fushida, S., Yonemura, Y.. Yamamoto, H., Watanabe, H.. and Raz, A.Expression of autocrine motility factor correlates with disease progression in humangastric cancer. Br. J. Cancer. 74: 2003-2007, 1996.

Lee. M. R., Ho, D. D., and Gurney. M. E. Functional and partial homology betweenhuman immunodeficiency virus and neuroleukin. Science (Washington DC), 237:1047-1051. 1987.McNeil. R. K.. Muthukrishnan. L.. Warder. E.. and D'Amore, P. A. Growth factors

are released by mechanically wounded cells. J. Cell Biol., 109: 811-822, 1989.Sullies, J., Giri. J. G.. and Mizel. S. B. IL-1 secretion by macrophages. Enhancementof IL-1 secretion and processing by calcium ionophorcs. J. Immunol., 144: 175-182.

1990.Mignatti, P.. Morimoto, T.. and Rifkin. D. B Basic fibroblast growth factor, a proteindevoid of secretory signal sequence, is released by cells via a pathway independent ofthe endoplasmic reticulum-Golgi complex. J. Cell. Physiol., 151: 81-93, 1992.

Cooper, D. N. W., and Barondes. S. H. Evidence for export of a muscle lectin fromcytosol to extracellular matrix and fora novel secretory mechanism. J. Cell Biol., 110:1681-1691, 1990.Sato, S., Burdett, I., Hughes. R. C. Secretion of the baby hamster kidney 30-kDagalactose-binding lectin from polarized and nonpolarized cells: a pathway independent of the endoplasmic reticulum-Golgi complex. Exp. Cell Res., 207: 8-18, 1993.Lindstedl. R., Apodaca. G.. Barondes, S. H.. Mostov, K. E.. and Leffler, H. Apicalsecretion of a cytosolic protein by madin-darby canine kidney cells. J. Biol. Chem..26«:11750-11757, 1993.Shin, J. T., Opalenik, S. R., Wehby, J. N., Mahesh, V. K., Jackson, A., Tarantini, F.,Maciag, T., and Thompson, J. A. Serum-starvation induced the extracellular appearance of FGF-1. Biochim. Biophys. Acta, I3I2: 27-38, 1996.Siders, W. M., and Mizel, S. B. Interleulin-1/3 secretion. J. Biol. Chem., 270:16258-16264, 1996.Murata. J.. Lee, H. Y., Clair, T., Krutzsch, H. C., Arestad. A. A., Sobel, M. E., Liotta,L. A., and Stracke, M. L. cDNA cloning of human tumor motility-stimulating protein,autotaxin, reveals a homology with phosphodiesterase. J. Biol. Chem., 269: 30479-

30484. 1994.Yoshimura. A.. Kuwazuru, Y., Funikawa, T., Yoshida, H., Yamada, K., andAkiyama, S. Purification and tissue distribution of human thymidine phosphorylase;high expression in lymphocytes, reticulocytes and tumors. Biochim. Biophys. Acta,1034: 107-113, 1990.

Hoogewerf, A. J., Leone, J. W., Reardon, I. M., Howe, W. J., Asa, D.. Heinrikson,R. L., and Ledbetter, S. R. CXC chemokines connective tissue activating peptide-IIIand neutrophil activating peptide-2 are heparin/heparan sulfate-degrading enzymes.Am. Soc. Biochem. Mol. Biol.. 270: 3268-3277. 1995.Paul. D., Niewiarowski, S., Varma, K. G.. Rucinski, B., Rucker. S., and Lange, E. Humanplatelet basic protein associated with antiheparin and mitogenic activities: purification andpartial characterization. Proc. Nati. Acad. Sci. USA, 44: 5914-5918. 1980.

2674

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

1998;58:2667-2674. Cancer Res   Yasufumi Niinaka, Sandor Paku, Arayo Haga, et al.   Tumor CellsIsomerase/Maturation Factor as Autocrine Motility Factor by Expression and Secretion of Neuroleukin/Phosphohexose

  Updated version

  http://cancerres.aacrjournals.org/content/58/12/2667

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/58/12/2667To request permission to re-use all or part of this article, use this link

on March 16, 2020. © 1998 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from