[CANCER RESEARCH 56. 2856-2863. June 15.

Neurotrophin Stimulation of Human Melanoma Cell Invasion: SelectedEnhancement of Heparanase Activity and Heparanase Degradationof Specific Heparan Sulfate Subpopulations1

DaríoManchetti,2 David J. McQuillan, William C. Spohn, Dan D. Carson, and Garth L. Nicolson

Departments of Tumor Biology ID. M.. W. C. S., G. L. N.j und Biochemistry anil Molecular Biology [D. D. C./, The University of Texas M. D. Anderem Cancer Center. Houston.Texas 77030. and Institute of Biosciences and Technology, Texas A & M University. Houston. Texas 77030 ¡D.J. M.]

ABSTRACT

Heparanase is an endo-ß-o-glucuronidase, the enzymatic targets of

which are the glycosaminoglycan chains of heparan sulfate proteoglycans.Elevated levels of heparanase are associated with the metastatic potentialof melanoma cells. Treatment of murine and human melanoma cells withthe prototypic neurotrophin nerve growth factor (NGF) increases theproduction of heparanase by melanoma cells. We reported previously thatphysiological concentrations of NGF increased in vitro Matrigel invasionof early-passage human brain-metastatic 70W melanoma cells but not

melanoma cells metastatic to other sites or nonmetastatic melanoma cells.Here we found that treatment of 70W melanoma cells with neurotrophinNT-3 increased Matrigel invasion, whereas treatment with neurotrophinsother than NGF or NT-3 did not influence invasion. Mutants of NGF thatdo not bind to the neurotrophin receptor p75NIK or other nonneuronal

growth factors were not able to enhance the invasion of 70VV melanomacells. When 70W cells were exposed to antisense oligonucleotides directedagainst p75N rR mRNA, there was a reduction in NGF and NT-3 binding,

and the neurotrophins failed to enhance Matrigel invasion. To study theproperties of heparanase in NT-regulated malignant melanoma invasive

processes, we developed a sensitive heparanase assay consisting of purified[35S]heparan sulfate subpopulations separated by agarose gel electro-

phoresis. Incubation of 70W cells with NGF or NT-3, but not brain-derived NT factor, NT-4/5, or mutant NGF, resulted in increased release

of heparanase activity that was capable of degrading a subpopulation ofheparan sulfate molecules.

INTRODUCTION

The vascular endothelium and its underlying BM3 make up an

important barrier that tumor cells must penetrate to colonize distantorgans (1). The major components of BM include type IV collagen,laminin, entactin, and HS proteoglycans (2). During tumor invasion,the BM becomes disorganized due to changes in the expression of BMconstituents, alterations in their assembly, and enzymatic degradation(1,3-5).

Degradation of BM components by invading malignant cells can beaffected by paracrine or trophic factors in the organ microenvironment(6). For example, brain-metastatic murine and human melanoma cellsrespond to certain NTs by increasing their production of ECM-degrading enzymes (5, 7, 8). NTs, including NGF, NT-3, BDNF, andNT-4/5 (9), can bind to the low-affinity NTR (p75NTR; Refs. 9 and 10)

Received 1/22/96; accepted 4/16/96.The costs of publication of this article were defrayed in part by the payment of page

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

1Supported by NIH Grants R29-CA64178 (to D. M.). RO1-AR42826 (to D. J. M.),ROI-HD25235 (to D. D. C), and CA44352 (to G. L. N.) and a grant from the National

Foundation for Cancer Research. Inc. (to G. L. N.).2 To whom requests for reprints should be addressed, at Department of Tumor Biology.

Box 108. The University of Texas M. D. Anderson Cancer Center. 1515 HolcombeBoulevard. Houston, TX 77030.

The abbreviations used are: BM. basement membrane: HS. heparan sulfate; NT.neurotrophin: ECM, extracellular matrix: NGF. nerve growth factor; BDNF. brain-derivedneurotrophic factor; NTR. NT receptor; TRK. tropomyosin-related kinase; FBS, fetalbovine serum: GAG. glycosaminoglycan; CS. chondroitin sulfate: ddH,O. double-distilled water; HSPG. heparan sulfate proteoglycan; CNS. central nervous system; HPLC,high-pressure liquid chromatography.

and also to specific high-affinity NTR represented by the tyrosine

kinase family of TRK receptors (9. 11). There is some specificity ofNT binding to TRK receptors; for example, NGF shows preferentialbinding to TrkA receptors, BDNF to TrkB receptors, and NT-3 to

TrkC receptors (12).Malignant melanoma cells express p75NIK in relation to their

malignancy and ability to metastasize to the brain (7, 8) within regionsthat synthesize and respond to NTs (13). We found that overexpres-sion of p75NFR in brain-metastatic melanoma cells correlates with an

increase in Matrigel invasion and secretion of ECM-degrading enzymes (7, 8). Although the human brain-metastatic melanoma cellsthat we examined did not express fr&A-encoded pl40'rkA, they didexpress i/iC-encoded pl45trkc, the putative NTR for NT-3 (8). These

observations have been confirmed using clinical specimens frompatients with metastatic melanomas at different stages of tumor progression by in situ hybridization and immunohistochemistry. An inverse relationship was found between expression of NT and NTR atthe invasion front of human melanoma brain métastases(14).

Important BM degradation targets of invading melanoma cells arethe HS chains found on HS proteoglycans (5, 7). HS is produced bya wide variety of different cell types and is found at the externalsurfaces of cells and in BMs and other ECMs (15, 16). The major HSproteoglycan of the BM, perlecan (17), is abnormally expressed bymelanoma cells (18, 19). Malignant cells can also modulate theexpression of HS proteoglycans by releasing degradative enzymes,such as heparanase (3), or cytokines can influence the turnover andbiosynthesis of host cell HS proteoglycans (20).

Here we report that invasion of brain-metastatic melanoma cells

through Matrigel is selectively augmented when these cells are exposed to NGF or NT-3 but not BDNF, NT-4/5, or mutated NGF and

other nonneuronal factors. Using antisense oligonucleotides derivedfrom the p75NTR sequence, we demonstrated the biological relevanceof p75NTR in the invasion process. Additionally, by development of asensitive heparanase assay that separates purified [^SjHS species by

agarose gel electrophoresis, we report that expression of heparanase isup-regulated by NGF and NT-3 and that heparanase preferentially

degrades specific HS species.

MATERIALS AND METHODS

Chemicals. Heparin, HS, chondroitin sulfate C, and D-saccharic acid 1,4-lactone were acquired from Sigma Chemical Co. (St. Louis, MO). ['HIHeparin(0.44 mCi/mg) and [35S]sulfate (43 Ci/mg) were purchased from DuPont-New

England Nuclear (Wilmington, DE) and ICN Biochemical (Irvine, CA), respectively. FBS and DMEM were purchased from Gibco. Inc. (Grand Island,NY), guanidine hydrochloride from Life Technologies, Inc. (Gaithersburg,MD), and 3-((3-choamidopropyl)dimethy]ammonio]-2-hydroxy-l-propanesul-

fonic acid from Boehringer Mannheim (Indianapolis, IN). Suramin was agenerous gift from Dr. Motowo Nakajima (Ciba-Geigy Japan Limited. Takara-

zuka, Japan).Tissue Culture. Early-passage human A-875, MeWo, and its wheat germ

agglutinin-resistant 70W melanoma cell lines (21) were maintained as mono-

layer cultures in 1:1 (v/v) DMEM:F12 medium (Life Technologies, Inc.)supplemented with 5% FBS (Gibco. Inc.) at 37°Cin a humidified 5% CGy

2856

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HKPARANASE REGULATION IN HUMAN MALIGNANT MELANOMA

95% air atmosphere. The cells were subcultured every 3-4 days by Irypsin-

EDTA treatment, except tor the 70W cell line, the medium of which waschanged every 24 h. Murine melanoma B16 cells (parental Fl. metastatic-BL6,and brain-metastatic-B 15b: Rets. 3 and 36) of less than eight passages from an

original frozen stock were grown to subconfluence in a 1:1 (v/v) mixture ofDMEM:F12 medium supplemented with 5c/c FBS. Cells were harvested by

rinsing them twice in medium without serum, followed by treatment for 10 minwith 2 HIMEDTA in Ca ' --/Mg+ :-free Dulbecco's PBS. All cell lines were

subcultured when they reached 60-80% confluence. One day before NT (0-4nM) treatment, the medium was replaced with serum-free medium (7). The

human RL95 cell line, an adenocarcinoma of the uterus (22). was chosen as asource of HS proteoglycan because more than 95% of the GAGs made by thesecells is HS (23). These cells were grown in 1:1 (v/v) DME:F12 supplementedwith 10% FBS (v/v). The |15S|GAG labeling required the use of special

medium that was sulfate deplete (RPMI 1640, formula 79-5139 EC: Life

Technologies. Inc.I. All cell lines were periodically checked for Mycoplasmacontamination using a Geneprobe kit (San Diego. C A), and only Myciiplaxma-

free cells were used.Chemoinvasion and Heparanase Assays. Tumor cell invasion following

NT treatment was assayed by use of Transwell (Costar. Cambridge. MA) cellculture chambers with Matrigel-coated filters, and invasion was monitored by

fluorescence using a Cytofluor Model 2300 (Millipore. Bedford. MA: Ref. 7).taking into consideration both background and map reading area of the instrument. Heparanase activity was determined by degradation of |15S]HS using

high-speed gel permeation chromatography (24) with some modifications (7).

Briefly, subconfluent cells, whether human (70W) or murine (BI6BL6 andB16B I5b). were harvested and solubilized in 50 mM Tris-HCI (pH 7.5). 0.05%NaN,. and 0.5% Triton X-100 for 30 min at 4°C.Cell lysates were thencentrifuged at 12.000 X g for 30 min at 4°Cand concentrated by means of

Amicon-30 microconcentration units (Amicon. Beverly. MA), according to themanufacturer's instructions. Cell lysates (50-70 /ng protein) were then incu

bated with radiolabeled HS and 0.2 M sodium acetate (pH 5.0) for 18 h at 37°C

(final reaction volume. 10-100 /¿I).Reaction was terminated by heatingsamples for 15 min at 95°C.Additionally, a delipidation step was applied to

cell lysates following the heparanase assay and before HPLC analysis. A TSKgel G3(XX) PWX2 column (7.8 mm x 30 cm: 6-fim particle size) from TosoHaas. Inc. (Montgomeryville. PA) was used in the high-speed gel permeation

chromatography.NT Inclination and Receptor Binding Assays. Purified recombinant hu

man NGF. BDNF. and NT-3 were iodinated by using an accepted procedure(25) using Na|'"5l| carrier-free (Amersham. Arlington Heights. IL) and treat

ment with matrix-bound lactoperoxidase/glucose oxidase using the Enzy-mobead reagent (Bio-Rad. Richmond. CA). lodination was previously shown

not to interfere with NT bioactivity (25). Briefly, beads were rehydratedovernight at 4°C.centrifuged. and resuspended in 190 fi\ of ddH,O. The

Enzymobcud suspension (50 fj.\) was added to the iodination reaction, asdescribed by the manufacturer, at pH 6.0. An additional 50 /nl of the resuspended beads in ddH,O were added, and incubation occurred for another h.The mixture was incubated for 2 h at room temperature (25°C).and the labeled

NTs were separated from free iodine using disposable desalting columns(Pierce. Rocktord. IL). Usually. NTs (2 /xg) were incubated in the presence of1.5 mCi Na['"5I|. and preparations with specific activities of 3500 cpm/tmolwere typically obtained. The | l21iI]NTs were stored at 4°Cand used within 2

weeks of preparation. NTR were extracted by incubating the cells with continuous gentle mixing in 2% NP40 in PBS for 18 h at 4°C.The extracts were

centrifuged to remove insoluble material, and each pellet was resuspended inNP40 in PBS and centrifuged a second time. The supernatants were combinedand diluted 1:1 with 20 mg/ml cytochrome C in PBS containing 0.5% (v/v)NP40. Receptor binding assays were subsequently performed with appropriatecontrols as described elsewhere (7. 26).

Cell Treatment with Antisense Oligonucleotides. Nuclease-resistantphosphorothioate 21-mer oligonucleotides were synthesized corresponding tothe 5'-translation initiation site of p75NTR (27). Thioate linkages were placed

at each end of the primers with internal thioates to prevent rapid degradationand binding to noncomplementary sequences. The oligonucleotide sequenceswere as follows: p75NTR antisense, 5'-TGGCACCTGCCCCCATCGCCC-3 ' :p75NTR sense. 5'-GGGCGAIGGGGGCAGGTGCCA-3' (Genosys Biotech

nologies. The Woodlands. TX). Oligonucleotides were added to a final concentration of 5 /AM (28) directly to the cells in 24-well plates (1.2 X IO5

cells/well) every 24 h for 4 days in serum-tree medium. Cells were thentransferred onto invasion chambers (2 X IO4 cells/filter), and invasion assays

were performed as described previously (7) in the presence of 2 nM humanNGF or NT-3 and the corresponding oligonucleotide. NTR were measured as

described above.HS and Identification of Cell Surface HS Components. RPMI 1640 was

used as the basal medium for |-'5S|sulfate radiolabeling of human RL95

adenocarcinoma cell HS subpopulations (22). Briefly, near-confluent RL95cells were rinsed several times with serum-free RPMI 1640 (minus sulfate)

supplemented with 3.3 mM MgCU. 1.2 g/l NaHCO,. 15 mM HEPES (pH 7.2),2.5 units/ml penicillin, and 2.5 /¿g/ml streptomycin sulfate. Streptomycinsulfate served as the sole source of nonradioactive sulfate in this medium (finalconcentration, ~2 JUM). The cells were incubated overnight in the samelow-sullate medium described above containing 0.5 mCi/ml |'5S|sultate. The

medium was collected, and the cell monolayers were rinsed several times withice-cold PBS. The cell monolayers were then incubated for 30 min on ice with

PBS containing 50 /ng/ml trypsin to release cell surface proteoglycans. Cellsdid not detach from the tissue culture surfaces under these conditions, nor wascell viability compromised as assessed by trypan blue dye exclusion. Thematerial released into the "trypsinate" was collected and placed in a boiling

water bath for 5 min and immediately cooled on ice to inactivate the trypsin.The released material was extensively dialyzed for 96 h against ddH2(). andaliquots were examined before, during, and after each dialyzing step until nofree sulfate was present. Trypsin-resistant |15S]HS from RL95 cells was

prepared similarly with some modifications that consisted of scraping cells offthe plates in the presence of PBS and adding an equal volume of 20%trichloroacetic acid (w/v):6% phosphotungstic acid (w/v) and placing them onice for 30 min. followed by centritugation at 2(XK)X g for 20 min. Precipitateswere collected by cemrifugalion at K).(KX)x g for IO min, and the pellets were

resuspended and centrifuged two more times in 5 ml 10% trichloroacetic acid.The final pellet was resuspended in 2 ml 0.1 MTris-HCI. 5% (v/v) ethanol. and

2 IHMCaCK (pH 8.0). After boiling for 2 min and cooling to room temperature,the samples were either incubated overnight with promise ( 10 mg/ml; Refs. 48and 49). or they underwent alkaline borohydride treatment (ß-elimination) at45°Cin the presence of 0.05 M NaOH and I M sodium borohydride for 24 h.

followed by neutralization with acetic acid (29). Precipitates were collected,and supernatants were dialyzed extensively with ddH:O. Aliquots were removed, radioactivity was determined, and specific activity was calculated.

Analytical Column Chromatography and HS Chemical Analyses. ASuperóse 6 column (1.0 X 30 cm: Pharmacia LKB. Uppsala. Sweden) waseluted with 4 M guanidine hydrochloride. 0.5% (w/v) 3-[(3-choamidopropyl)-dimethylammonio|-2-hydroxy-l-propanesulfonic acid, and 50 HIMsodium acetate (pH 6.0) at a flow rate of 0.4 ml/min. and 0.4-ml fractions were collected.

Aliquols were taken for determination of radioactivity by scintillation counting. Molecular size estimates for GAG chains were based on the method ofWasteson (30) for Sepharose 6B. The accuracy of molecular weight determination by Superóse 6 gel filtration for CS chains was assessed by directcomparison with a Sepharose 6B column calibrated with CS chains of knownmolecular mass (31 ). For HS chains (derived from alkaline-borohydride cleavage) and HS glycopeptides (derived from pronase digestion of trypsin-released

HS proteoglycan). the calibration held for similarly sulfated GAGs. Analysesinvolving standard nitrous acid hydrolysis (32) were performed to confirmlabeling of hcparan sulfate chains on GAGs. Briefly, samples for digestionwere made up as 0.5:0.5:1.0 of a 20% (v/v) »-butyl-nitrite (in I(X)% ethanol):!

N HC1:H,0 solution. Parallel digestions of samples containing 1 ing heparin inthe same solvent were also done. Samples were digested for 4 h at 25°Cwith

occasional agitation. At the end of incubation. 1/10 volume of 10% (v/v)cetylpyridinium chloride solution was added. Only lubes in which the n-butyl-

nitrite was omitted immediately did a turbid precipitate form. Samples werelyophilized and analyzed by gel-permeation HPLC. Alkaline borohydridetreatment was performed in 0.05 M NaOH at 45°Cfor 24 h in the presence of

l Msodium borohydride. as described previously (29). Excess borohydride wasdestroyed by neutralization with acetic acid.

Agarose (¡el Electrophoresis of HS Subpopulations. A reaction consisted of 10-20 /J.1heparanase enzyme source (7) mixed with electrophoresis

buffer at a final concentration of 0.025% (w/v) bromophenol blue. 0.025%(w/v) xylene cyanol, and 2.5% (w/v) Ficoll (type 4(X)) in ddH,O. Beforeelectrophoresis. 1% (w/v) SDS final concentration was added to each samplefor 15 min at room temperature (25°C) to dissociate <5S-labeled digestion

2857

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HtPARANASE REGULATION IN HUMAN MALIGNANT MELANOMA

products from molecular weight complexes (data not shown). Following thisstep, agarose gel (1.2% w/v) electrophoresis was performed at 75 V for l h at25°Cor until the samples migrated approximately two-thirds of the entire gel

length. Autoradiography was performed on the dried gel by exposure to X-AR5 film (Kodak. Rochester, NY) for 3-7 days. The direction of electrophoretic

mobility shown in the figures is from top to bottom in all cases.In the sequential agarose gel-Superose 6 Chromatographie analysis, gel

pieces from the slower-migrating band of |"S]HS and the intact HS band

following agarose gel electrophoresis were soaked in 2 ml of Tris-borate-

EDTA (TBE) buffer for 2 h. placed in small dialysis bags, and electroeluted for4 h at 50 V using a Minigel apparatus. Liquid was collected from the bag. spunin an Eppendorf centrifuge, and concentrated. Radioactive material then wasloaded onto a Superóse 6 column and eluted under dissociating conditions;

then fractions were collected.

RESULTS

Effects of NTs on in Vitro Invasion of Melanoma Cells. UsingMatrigel-coatecl filters in Transwell invasion units, the effects ofseveral NTs on the invasive capacity of the brain-metastatic melanoma cells were determined. Previously, we found that highly meta-

static human 70W melanoma cells were stimulated to invade theMatrigel barrier by NGF at higher rates than the other lines tested (7,8). This increase in invasion has been shown to be associated withhigh levels of p75NTR transcripts, cell surface NTR molecules, and

NTR sites/cell (7, 8). We have expanded these results to include otherNTs. Melanoma cells were incubated with HPLC-purified prepara

tions of human NT at optimal concentrations to saturate NTRs (NGF,BDNF, NT-3, or NT-4/5). Only NGF and NT-3 stimulated invasion of

70W cells (Fig. 1). In contrast, none of the NTs had dramatic effectson the MeWo parental line. Additionally. NT stimulation of ECMinvasion was specific, since other factors such as keratinocyte growthfactor and Kaposi's sarcoma-derived fibroblast growth factor, which

are not present in the brain microenvironment, did not stimulateinvasion (data not shown). The results were consistent with thepresence of the appropriate NTR (p75NTR and TrkC but not TrkA) in

these cells (8). When we used an NGF mutant in which alanineresidues replaced lysine-32, lysine-34. and glycine-35 (33), we foundsignificantly lower degrees of invasion than with wild-type NGF (Fig.

1). This NGF mutant has less than 1% of the binding capacity of NGF

NT SPECIFICITY

2000 •¿�

IM

1

Control NGF BDNF NT-3 NT-4/5 Mut. NGF

Fig. 1. NT effects on the invasion of human 7ÃœWbrain-metastatic melanoma cells.Human melanoma cells (2.0 x IO4) in O.\c/c (w/v) BSA serum-free medium were seeded

into the upper compartment of Transwell chambers. Filters were precoated with phenolred-free Matrigel (1:30 dilution to final concentrations of 0.46 mg/ml) on their uppersurfaces. After a 72-h incubation at 37°Cwith different concentrations of purified,biologically active human NTs (2 nw) and of a NGF triple mutant (K32A + K34A 4-E35A; Mut. NGF) unable to bind p75NTR. invasion was determined as described in"Materials and Methods." Data are the means of three independent experiments: bars. SD.

OE

ciU

70WNGF70W NT 3Sp75 NGF

ASp75/NGF

Sp75 /NT 3ASp75/NT3

Experiment No.

Fig. 2. Effect of antisense p75NTK on the invasion of human melanoma cells inresponse to NT. Human 70W melanoma cells (1.2 x IOs/24-well plate) were treated for96 h with 5 /IM antisense or sense p75NTR oligonucleotides added every 24 h in serum-freemedium. Cell growth was monitored in parallel. Treated or untreated cells (2 X IO"1)were

transferred into the upper compartment of a Transwell chamber, and chemoinvasionassays were performed in the absence or presence of 2 DMamounts of human NTs, asdescribed in "Materials and Methods." Data are the means of three independent experi

ments: bars. SD.

Table 1 Binding of human ¡'251]NCF.l'"llNT-3. or /'"¡IBDNF la human

melanoma ceils

Cells70W3S5McWoA875NGF63<11198NT-3(femtomoles bound/106cells)51•

1h.6l117BDNF<1<l<11.5

to p75N rR but retains 65% of the specific biological activity of NGF

and, importantly, binds to the TrkA receptor with only a 2-fold

reduction in affinity (33). The result indicates that binding of NTs tothe low-affinity p75NFR receptor is necessary for stimulation of 70W

cell invasion. Incubation of the 70W cells with higher concentrationsof NTs (100 ng/ml) resulted in reduced rates of invasion, as observedpreviously with NGF (7, 8).

We further investigated the role of p75NIR in the invasion of 70W

cells by using specific antisense oligonucleotides designed using the5'-translated region of human p75NTR (27). Analysis of cell growth

indicated that addition of these oligonucleotides to the medium of70W cells did not affect their rate of proliferation (data not shown). Incontrast, incubation of NT-treated 70W cells with p75NTR-specific

antisense oligonucleotides prior to (96 h) and during (72 h) thechemoinvasion assays resulted in inhibition of invasion in the presence of NGF or NT-3 (Fig. 2). In contrast, a p75NTR sense construct

had little effect on the rates of invasion stimulated by NGF or NT-3

(Fig. 2).Inhibition of NT Binding to p75NTR by Antisense Oligonucleo

tides. Previously, we found that p75NIR is overexpressed in 70W

cells (8), resulting in increased NGF binding to these cells (7). We,therefore, examined NT-3 binding to 70W cells and whether theinhibition of invasion observed by the addition of p75NTR antisense

oligonucleotides correlated with reduction of NT-3 receptors. Table 1shows the results of binding assays of human ['25I|NTs to membrane

preparations from 70W, MeWo, and A-875 cells. A875 is a brainmetastasis-derived human melanoma cell line known to express highlevels of p75NTR (0.5 X IO6receptors/cell) but not TRK receptors (34,35). Binding of [l25I]NT-3 to 70W, MeWo, and A875 cells wasobserved and compared with binding of ['-5I]NGF and [12<iI]BDNF,

respectively. The brain-metastatic A875 and 70W cells bound higher

2858

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HEPARANASE REGULATION IN HUMAN MALIGNANT MELANOMA

Table 2 Comparison of antisense or sense p75 treatment on binding of human

Oligonucleotide treatment

Time(days)01234Antisense

p75NTR Sense p75NTR(fmoles [l25I]NT-3 bound/mgprotein)70W10081.367.855.539.1A87510094.773.764.531.670W100ND"94ND77A875100ND107ND81

" ND, not determined.

30000-

§Ietu.

20000-

10000-

20 30 40 SO 60 70

FRACTION NO.Fig. 3. Superóse 6 Chromatographie analysis of [35S]HS subpopulations before and

after nitrous acid degradation. HS radiolabeling, isolation (cell-associated or secreted)from RL95 cells, and subsequent Chromatographie analyses were performed as describedin "Materials and Methods." Column calibration was performed with CS chains of defined

molecular weight. Peak elution position (Kj) was calculated relative lo void (V0, 7.6 ml)and total (V,, 22 ml) column volumes by the formula Kj = (Vc - V,,)/(V, - V0). 0,cell-associated ["SJHS from RL95 eluted in two peaks (Peak 1. Kd = -0.27; Peak 11.Kj = -0.77); A, nitrous acid degradation of cell-associated ["S]HS; D, [%5S]HSsecretedby RL95 into medium; O, ['H]heparin (MT ~11.000) run as standard. Radiolabeled CS

chains were chromatographed in parallel as molecular weight standards and eluted atfractions 35 (M, 60.000 CS chain), 38 (M, 42,000), 42 (M, 20,000). and 47 (M, 17,000).respectively.

amounts of the [l25I]NTs NGF and NT-3 but not BDNF (Table 1).The addition of p75NTR antisense oligonucleotides to the 70W cells

resulted in a decrease in the equilibrium binding values for NGF andNT-3. This reduction was greater as treatment with antisense oligonucleotides continued than in cells treated with p75NTR sense oligo

nucleotides (Table 2). The results indicate that the binding of NT tothe low-affinity receptor p75NTR is probably necessary to induce

increased invasion of the 70W cells.Characterization of HS Components. Since NGF and NT-3 in

teraction with p75NTR appears to be involved in stimulating brain-

metastatic melanoma cell invasion, we next focused on a key BM-

degrading enzyme, heparanase, that is responsible for degradation ofHS molecules in ECM. We found previously that in the absence oftrophic or growth factor stimulation in culture, highly brain-metastatic

murine melanoma cells secrete high levels of degradative enzymes,such as heparanase (3, 5, 7). Similarly, the human brain-metastatic

70W melanoma subline secretes more heparanase and degrades significantly higher amounts of HS than the nonmetastatic 3S5 subline(7). Previous studies also documented the important role heparanaseplays in tumor invasion (24, 36, 37).

We wanted to identify the HS targets of heparanase action andexamine other NTs for their effects on heparanase. Accordingly, we

made use of RL95 cells as a HS source, the rationale being that RL95are able to grow in sulfate-depleted media and most (>95%) of GAG

chains synthesized by these cells is HS, thus providing a convenientsubstrate to investigate heparanase action and heparanase regulationby NTs. RL95 were incubated with [35S]sulfate, and selected HSsubpopulations were obtained as described in "Materials and Methods." Cell lysates were dialyzed to remove unincorporated 35S. Theoriginal [35SJHSPG populations underwent extensive pronase diges

tion and alkaline borohydride treatment (ß-elimination) to digest the

polypeptide backbone while leaving the GAG chain portion intact (38,39). Nitrous acid digestion, followed by Superóse 6 chromatographyof metabolically labeled GAG populations (3-9 X IO9 dpm//ng HS),

confirmed the identity of these GAGs as HS (Fig. 3). The material wasalso sensitive to bacterial heparitinases (23) but resistant to chon-

droitinases AC, C, and ABC digestion (data not shown). Superóse 6chromatography then was performed on the same material usingcolumns calibrated with reference fractions of radiolabeled CS (30).Most of the [35S]HS molecules were of high molecular weight and

eluted as a relatively sharp peak with a median ATdof 0.27 (Fig. 3,Peak I; Mr -70,000). Additional 35S-labeled material was retarded in

the column and eluted at a median K¿of 0.77 (Fig. 3, Peak ¡I).Wenext compared some of the properties of the high molecular weightHS material with those of the low molecular weight HS. PreparativeSuperóse 6 chromatography was performed, and the two HS subpopulations were separated and pooled: peak I (fractions 25-34) represent-

CO

C/)in

PeakI(70kDa)"

PeakII(23kDa)

Fig. 4. Agarose gel electrophoresis of cell-associated [<5S)HS indicating high (M,-70,000: 70 kDa) and low (M, -23.000; 23 kDa) fractions of [MS]HS. Estimation of

molecular sizes is based on elution profiles from Superóse 6 column calibrated withstandard glycans. as mentioned in Fig. 3.

2859

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HtPARANASE REGULATION IN HUMAN MALIGNANT MELANOMA

B

in

Mr by [3H] CS defined size

ing a high molecular weight HS subpopulation: and peak II (fractions35-50) representing low molecular weight HS. Agarose gel electro-

phoresis confirmed the presence of the high and low molecular weightHS components; peak I or the high molecular weight HS migrated onagarose gel electrophoresis as a band of apparent molecular weight ofapproximately —¿�70,000.whereas the low molecular weight HS (peak

II) migrated as a broad band with an approximate median molecularweight of Mr 23,000 (Fig. 4). This reflects a difference at the level ofGAG chains between the two original HSPG populations since beforethe analysis they underwent extensive pronase digestion and alkalineborohydride treatment. Again, nitrous acid deamination was routinelyperformed to confirm the HS nature of our labeled HS chains. Thehigh molecular weight HS (peak I) component was found to beheparanase-sensitive, whereas the low molecular weight HS (peak II)

was not degraded by heparanase (data not shown). Therefore, the roleof high molecular weight HS as a target of heparanase action wasfurther investigated.

The Peak I HS Component Is Degraded by Heparanase. Sincepeak I. the high molecular weight [15S]HS component, was apparently

degraded by heparanase, we performed several controls to insure thatdegradation was indeed due to heparanase enzymatic action. Forexample, we considered that different molecular weight HS mightarise artifactually after binding cellular components in the cell extractsdue to the effects of the electrophoresis (40, 41). To test this, weconcentrated cell extracts from brain-metastatic human melanomacells (VOW) and mixed them with the high molecular weight [15S]HS.

and the mixture was then immediately electrophoresed. In addition,high molecular weight |Õ5S]HSwas incubated at 37°Cfor 18 h under

the conditions used for the heparanase assay (7) but in the absence ofheparanase, and the material migrated similarly to intact HS (Fig. 5A,Lanes I and 2). Additional controls consisted of mixing [35S]HS with

heat-inactivated cell lysates. Excess nonradioactive CS type C (300

fig/ml) was added to the cell lysates and [ S]HS to minimize anyeffects due to nonspecific interactions of the GAG chains with components of the cell extracts (40). None of these procedures changedthe migration pattern of the HS (data not shown). Experiments usingextracts of murine melanoma cell lines known to express differentamounts of heparanase were used (24, 36). The high molecular weight[3SS]HS with various cell lysates resulted in degradation of HS,

especially when murine brain-metastatic melanoma (BI6-B15b)cell lysates were used: RL95 [35S]HS was incubated with B16B15b

cell extracts in the absence or presence of 20 mM D-saccharic acid1,4-lactone, a potent exo-ß-glucuronidase inhibitor (42, 43), withGAG fragments analyzed by both high-speed gel permeation chro-

matography and electrophoretic analysis. We did not observe asignificant inhibition of GAG degradation. Conversely, incubationof the same cellular extracts with heparin or suramin (100 /XM),potent inhibitors of melanoma heparanase and invasion (44), abrogated GAG degradation, confirming a heparanase action (datanot shown). We, therefore, examined whether heparanase activityin 70W cells could be NT modulated, similar to the results seen forNGF (7). Human melanoma cells (70W) were incubated with eachNT member and NGF triple mutant (p75NrR binding-deficient);

then cell lysates were prepared, incubated with high molecularweight |15S]HS. and electrophoresed. In comparison to known

NGF-driven heparanase up-regulation, a marked decrease of intact

2860

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HKPARANASK REGULATION IN HUMAN MALIGNANT MELANOMA

[35S]HS (Fig. 5B, Lane 1) was observed with a concomitantappearance of low molecular weight 35S-labeled material, if the

cells were exposed to NGF or NT-3 (Fig. SB, Lanes 2 and 3).Treatment of 70W cells with other NTs. i.e.. BDNF or NT-4/5, did

not result in increases of heparanase activity, as seen by degradation of the intact high molecular weight [35S]HS material (Fig. 5C).

Electroelution of the heparanase-digested and aggregated material

from the agarose gel, followed by Superóse 6 chromatographyunder dissociating conditions (4 M guanidine-HCl), resulted in a

lower molecular weight peak compared to the starting material(Fig. 6). That this was indeed the result of heparanase action andnot the result of other enzymatic activities, such as sulfatases, wassustained by the fact that we did not observe any detectableradioactivity at the V, region of the column (Fig. 6). Additionalconfirmation came following BioGel P-2 column chromatographyof our cell lysate after reaction with [35S]HS, using in parallel CS

di-, tetra-, and octasaccharides as well as labeled free sulfate as

molecular weight standards. There was no appreciable radioactivity detected in the region corresponding to the migration positionof free sulfate (data not shown).

We next compared heparanase cleavage of our high molecularweight-purified [35S]HS with bovine lung [3H]HS (24, 37); both were

sensitive to heparanase as well as bacterial heparitinases (22, 24, 37).We obtained indications regarding their charge densities by means ofan anion-exchange column (DEAE-Sepharose) chromatography after

equal loading of radiolabeled material and their separation (38). Theirelution profiles were distinct in relation to both elution positions andpeak heterogeneity, as evidences by a different sulfation patternbetween the two substrates (RL95 HS was less sulfated than bovinelung HS; data not shown). GAGs from these two sources were addedto murine metastatic B16BL6 cell lysates because this B16 subline isalready known to contain high amounts of heparanase (3). The heparanase-digested [3H]HS was shifted to a higher elution position than theundigested [3H]HS as expected (24, 37); however, heparanase digestion of purified cell surface [35S]HS resulted in a more dramatic shiftin the size distribution. Indeed, the use of cell surface [35S]HS

strongly enhanced the appearance of intermediate molecular weightfragments (Fig. 7). Similar results were obtained using heparanase

3000

10000

8000 -

o—¿� 6000 -

U

KIL

ìQ.O

4000 -

2000

abc

U I

10 20 30 40 SO 60

FRACTION NO.Fig. 6. Superóse 6 chromatography of agarose gel-purified [ 15S]HS before and after

incubation with brain-metastatic, heparanase-containing 70W cell lysates. A shift in theelution position occurred when digested [15S]HS (^) was compared with starting undigested [35S]HS (Q). V„and V, values were at fractions 16 and 55. respectively. Standard

glycans are: a. M, 60.000; b. M, 42.000: c. M, 20.000; and d. M, 7.000.

Zo\-o<oc

Q.Q

2000 -

1000-

10 20

FRACTION NO.

Fig. 7. Metastatic melanoma heparanase assay using isolated and purified cellsurface [^S]HS subpopulations compared with [1H]HS from bovine lung preparations. The two undigested HS substrates, ['H]HS (H) and [15S]HS •¿�left), had

similar elution profiles in the absence of cell lysates; however, incubation for 18 hwith cell lysates from murine metastatic melanoma cells (B16BL6) containing heparanase gave more pronounced shifts in size distribution when cell surface |"S]HSsubpopulalions were incubated with melanoma cell lysates ([35S]HS HEP.: 0 ) versuscorresponding ['HjHS used as substrate (['H]HS HEP.; «). |"S]HS HEP. (O) and["SJHS HEP.2 (•.right) refer to samples digested with two different heparanasepreparations. High-speed gel permeation chromatography of incubation products wasperformed by using a TSK gel O3000 column (Toso Haas) as described in "Materialsand Methods."

from murine brain-metastatic B16-B15b or human brain-metastatic70W cells (data not shown). Incubation of this cell surface-derived[35S]HS subpopulation with brain-metastatic 70W human melanoma

cell extracts resulted in the generation of characteristic lower molecular weight fragments (Fig. 8/4). Incubation of 70W cells with humanNGF or NT-3, followed by determination of heparanase activity using

HPLC analysis, confirmed a selective NT enhancement of heparanaseactivity in brain-metastatic human melanoma cells, with NT-3 aug

menting heparanase activity at levels higher than NGF or untreatedcells (Fig. SA); heparanase degradation potential remained independent of previous exposure of melanoma cells with members of the NTfamily other than NGF or NT-3, i.e., BDNF, NT-4/5, and the NGF

triple mutant molecule (Fig. 8B).The heparanase degradation potential observed for melanoma cells

"in vitro" was confirmed "in vivo" by immunofluorescence studies

using clinical specimens from patients with brain-metastatic mela

noma by means of purified preparations of monoclonal antibody(10E5 monoclonal antibody) developed against heparanase.4

DISCUSSION

An important clinical end point in patients with melanoma andother cancers is the formation of métastasesin the brain. Tumor cellsthat gain access to the brain must traverse microvessels at variousorgan sites, home to and implant in the brain microcirculation, andeventually invade and survive in the brain (6). The microenvironmentof the brain is unique, and it probably provides malignant cells thatexhibit preferential colonization to different regions of the brain withspecific trophic and other organ factors (6, 45). Since melanomaspossess the highest frequency of CNS colonization (46), it is likely

4 D. Marchetti and G. L. Nicolson. manuscript in preparation.

2861

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HhHAKANASKRtüllLATIONIN HUMAN MALIGNANT MELANOMA

FRACTION NO.

B

zo

gocstCLQ

3000

2000 -

1000 -

[35S)HS

70W + BDNF70W -f NT-4/5

70W + Mut. NGF

FRACTION NO.Fig. 8. NT regulation ut heparanasc activity ill brain-melastatic human 70W melanuma

using isi)lulcd high molecular weight l *\S)HS subpopulations as substrate. A. heparanase

assay ölbrain-mclaslalic human 70W nielanomu cells. Cells were preiiicubated tor 12 hwith MUH or NT-3 before cell lysates »ereobtained and incubated tor IS h al 37°Cwithisolated l'^S)HS subpopulations. HFLC analysis followed, using a TSK gel G3(XX)column as described in "Materials and Methods." Increased heparanase activity was

observed not only for 70W-treatcd NGF (internal controll but also for 70W-ireated NT-3»•eTAU.vcontrols (70W cells without NT or | '^SjHS subjx>pulations alone). H. heparanase

assay of hrain-metastatic human 70W melanoma cells following exposure to BDNF.NT-4/5. and NGF triple mutant. Heparunase activity wras independent of previous cellulare.\[H)sure to NTs other than NGF or NT-3. See "Materials and Methods" for details.

that this type of cancer is quite susceptible to brain trophic and otherfactors.

Invasion into the brain requires penetration of the thick BM thatforms part of the blood brain barrier. Several components of theECM are affected by this process, including the GAG chains of HSproteoglycans (18-20). Using heparanase. metastasizing cells can

degrade the HS proteoglycan chains of the ECM. The suggestionthat NTs might be involved in this process was first advanced whenthe p75 low-affinity neurotrophin receptor (p75NIR) was shown to

be expressed at the highest levels at the later stages of melanomaprogression (47). Using the MeWo human melanoma cell system,we found that exposure of the brain-metastatic 70W cells to

physiological concentrations of NGF resulted in an increased ability to invade Matrigel or purified HS proteoglycan (7). Invasion

was also characterized by secretion of the ECM degradative enzymes, such as heparanase (7). and MT72,000 type IV collagenase(8). Heparanase preferentially cleaves /3-D-glucuronosyl-yV-acetyl-

glucosaminyl linkages of the HS molecule (37). We have foundthat CNS-metastatic melanoma heparanase is capable of cleavingcell surface-derived |<5S]HS subpopulations. This substrate was

found to be sensitive to bacterial heparitinases and to possess adifferent (less sulfated) sulfation pattern when compared to bovinelung HS already used as a heparanase substrate in previous studies(3, 24, 37). Since we previously localized heparanase at the cellsurfaces of metastatic melanoma cells (48). this en/yme maydegrade the cell surface as well as GAG chains of ECM HSproteoglycans. It is likely that metastatic cells require heparanaseas well as other degradative enzymes, such as type IV collag-

enases, for tumor cell invasion to occur through the vascularendothelial ECM (5). The 70W cells expressed high levels ofp75NIK but did not express trkA (8). We have expanded these

studies to include NT-3, BDNF, NT-4/5, and a NGF triple mutant(K32A + K.34A + E35A) unable to bind p75NTR (33). The highlyinvasive cells expressed elevated levels of p75NIR, and invasioncould be inhibited by an antisense oligonucleotide against p75NIR.

NGF and NT-3 also stimulated increased heparanase expression

and concomitant digestion of cell surface and peí¡cellularHS GAGchains of respective HSFGs. By purification and use of a selectedand homogeneous HS population from a cell surface source, weindeed found increased heparanase activity values. Thus, a HS cellsurface origin may provide a more suitable and homogeneoussubstrate to study melanoma heparanase.

The response of brain-metastatic 70W cells to NTs, such as NGFand NT-3, is consistent with their pattern of brain colonization. Wheninjected i.v. into nude mice, brain-metastatic 70W cells colonize thecerebral cortex of the brain (21). This is a major site of NGF and NT-3

secretion and one of the few sites within the brain microenvironmentwhere both NGF and NT-3 are synthesized (49). Preliminary experiments indicate that NGF-treated 70W melanoma cells show increased

metastatic capabilities to the cerebral cortex following i.v. injection innude mice.5 This suggests that melanoma cells may invade, survive,

and grow preferentially in areas within the brain where secretion ofNGF and NT-3 occurs. It is possible that the brain uses NT as

paracrine survival and growth factors during brain injury and wounding (50). Support for this notion is that high concentrations of NGFand NT-3 are found in normal brain tissue at the invasion zones of

human melanomas in the CNS (14, 51). How NTs induce CNSinvasion is unclear, but specific interactions of NT with low-affinityp75NIK receptor and induction of degradative enzymes appear to be

necessary. Studies are presently ongoing to determine exactly thetime-dependent release of heparanase into culture media in relation toHS processing, the presence of HS-binding proteins, and NT regula

tion.

ACKNOWLEDGMENTS

We thank the Neuroscience Group of Genentech Inc. tor their help inproviding NGF. NT-3. and NT-4/5 for this project and the Discovery Group of

Regencron Pharmaceuticals, Inc. tor BDNF. We thank Dr. Carlos Ibanez(Laboratory of Molecular Neurobiology. Karolinska Institute. Stockholm.Sweden) for kindly providing NGF triple mutant and Dr. Magnus Hook(Center tor Extracellular Matrix Biology at Texas A&M University, Houston.TX) for providing radiolabeled CS and oligosacchuride chains as well asaccessibility to chromalugraphic equipment. Special thanks go to ClarenceJohnson and Douglas Chin for their excellent technical assistance.

5 D Marchetti and G. L. Nicolson. unpublished observations.

2862

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HKPARANASi; RI-XìlI.ATION IN Ht MAN MALICÌNANTMHI.ANOMA

REFERENCES

1. Nicolson. G. L. Metustatic tumor cell interactions with endothelium basement membrane and tissue. Curr. Opin. Cell Biol.. /: 1009-1019. 1989.

2. Yurchenco. P. D.. Tsilihary. E. C.. Charonis. A. S.. and Funhmayr. H. Models forself-assembly of basement membrane. J. Histochem. Cytochem.. 34: 93-102, 1986.

3. Nakajima. M. Irimura, T.. and Nieolson. G. L. Heparanase and tumor metastasis. J.Cell. Biochem., 36: 157-167, 1986.

4. Liotta. L. A.. Steeg, P. S., and Stetler-Stevenson, W. G. Cancer metastasis andangiogenesis: an imbalance of positive and negative regulation. Cell, 64: 327-336,1991.

5. Nieolson. G. L.. Nakajima. M., Herrmann. J. L.. Menter, D. G., Cavanaugh. P.. Park,J. S.. and Marchelti. D. Malignant melanoma metastasis to brain: role of degradativienzymes and responses to paracrine growth factors. J. Neuro-Oncol.. IK: 139-149,1994.

6. Nicolson. G. L.. Menler. D. G.. Herrmann. J. L.. Cavanaugh, P., Jia, L., Hamada. J..Yun, Z.. and Marchetti. D. Tumor metastasis to brain: role of endothelial cells,neurotrophin. and paracrine growth factors. Crit. Rev. Oncog.. 5: 451-471. 1994.

7. Marchetti. D.. Menter, D., Jin. L.. Nakajima, M.. and Nicolson. G. L. Nerve growthfactor effects on human and mouse melanoma cell invasion and heparanase production. Int. J. Cancer. 55: 692-699. 1993.

8. Herrmann. J. H., Menter. D. G.. Hamada. J.. Marchetti. D.. Nakajima. M.. andNicolson, G. L. Mediation of NGF-stimulated extracellular matrix invasion by thehuman melanoma low-affinity p75 neurotrophin receptor: melanoma p75 functionsindependently of irkA. Mol. Biol. Cell. 4: 1205-1216. 1993.

9. Chao. M. V. Growth factor signaling: where is the specificity? Cell, (Hi: 995-997.

1992.10. Bothwell. M. Tissue locali/ation of nerve growth factor and nerve growth factor

receptors. Curr. Top. Microbio!. Immunol.. 165: 55-70. 1991.11. Barbacid. M. Nerve growth factor: a tale of two receptors. Oncogene, K: 2033-2044.

1993.12. Lamballe, F., Klein, R., and Barbacid. M. TrkC, a new member of the TRK family of

tyrosine protein kinase. is a receptor for neurotrophin-3. Cell. 66: 967-979. 1991.13. Thoenen. H.. Bandllow. C.. and Heuman. R. The physiological function of nerve

growth factor in the central nervous system: comparison with the periphery. Rev.Physiol. Biochem. Pharmaeol., 109: 145-178. 1987.

14. Marehetti. D.. McCutchcon. !.. Ross. H. I., and Nicolson. G. L. Inverse expression ofneurotrophin receptor and neurolrophins at the invasion front of brain-metastatichuman melanoma tissues. Int. J. Oncol.. 7: 87-94, 1995.

15. Dietrich. C. P.. Sampio. L. ().. and Toledo. O. M. Characteristic distribution ofsulfatcd mucopolysaccharidcs in different tissues and in their respective mitochondria. Biochem. Biophys. Res. Commun.. 71: I-IO, 1976.

16. Hedman. K.. Kiirkinen. M.. Alitalo. K.. Vaheri, A., Johansson. S., and Hook, M.Isolation of the perieellular matrix of human fibroblast cultures. J. Cell Biol.. -V/:83-91. 1979.

17. Murdoch. A. D.. Liu, B., Schwarting, R., Tuan. R. S.. and lozzo. R. V. Widespreadexpression of perlecan proteoglycan in basement membranes and extracellular matrices of human tissues as detected by a novel monoclonal antibody against domain IIIand by in \iiii hybridization. J. Histochem. Cytochem.. 42: 239-249. 1994.

18. Cohen. I. R.. Murdoch. A. D.. Naso. M. F.. Marchetti. D.. Berd. D.. and lozzo. R. V.Abnormal expression of perlecan in metastatie melanoma. Cancer Res.. 54: 5771-

5774. 1994.19. lozzo, R. V., Cohen. I. R.. Grassei, S., and Murdoch. A. 1). The biology of perlecan:

the multifaceted [leparan sulfate proteoglycan of basement membranes and perieellular matrices. Biochem. J.. 302: 625-639. 1994.

20. lozzo. R. V. Proteoglycans: structure, function and role in neoplasia. Lab. Invest.. 53:373-396. 1985.

21. Ishikawa. M., Fernandez, B.. and Kerbel. R. S. Highly pigmented human melanomavariant which mctastasi/.es widely in nude mice, including to skin and brain. CancerRes.. 48: 4897-4903, 1988.

22. Way. D. L.. Grosso, P. S.. Davis. J. R., Surwit, E. A., and Christian. C. D.Characteri/.ation of a new human endometrial carcinoma (RL95-2) established intissue culture. In Vitro. IV: 147-158. 1983.

23. Rohde. L. H.. and C'arson. D. D. Heparin-like glycosaminoglycans participate in

binding of a hum,m trophohlastic cell line (JAR) to a human uterine epithelial cell line(RL95). J. Cell. Physiol.. /55: 185-196. 1993.

24. Nakajima. M.. Irimura. T.. and Nicolson. G. L. A solid phase substrate of heparanase:its application to assay of human melanoma for hcparan sulfate degradative activity.Anal. Biochem.. 157: 162-171. 1986.

25. Escandon. E.. Burton. L. E.. Szonyi. E.. and Nikolics. K. Characterization of neurotrophin receptors by affinity crosslinking. J. Neurosci. Res.. 34: 601-613. 1993.

26. Marchetti. D.. and McManaman. J. L. Characterization of nerve growth factor bindingto embryonic rat spinal cord neurons. J. Neurosci. Res., 27: 211-218. 1990.

27. Johnson. D., Lanahan. A.. Buck, C. R.. Sehgal. A.. Morgan. C.. Mercer. E.. Bothwell.M.. and Chao. M. Expression and structure of the human NGF receptor. Cell. 47545-554, 1986.

28. Sariola. H.. Saarma. M.. Sainio. K.. Arumaae. U.. Palgi. J.. Vaahtokari. A.. Thesleff.I., and Karavanov. A. Dependence of kidney morphogenesis on the expression ofnerve growth factor receptor. Science (Washington DC). 254: 571-573. 1991.

29. Carlson, D. M. Structures and immunochemical properties of oligosaccharides isolated from pig suhmaxillary mucins. J. Biol. Chem.. 243: 616-626. 1968.

30. Wasteson. A. A method for the determination of the molecular weight and molecularweight distribution of chondroitin sulfate. J. Chromatogr., 59: 87-97. 1971.

31. McQuillan. D. J.. Findlay. D. M.. Hocking. A. M.. Yanagishita. M.. Midura. R. J.. andHascall. V. C. Proteoglycans synthesized by an osteoblast-like cell line (UMR106-01). Biochem. J.. 277: 199-206. 1991.

32. Shively. J. E.. and Conrad. H. E. Formation of anhydrous sugars in the chemicaldepolymerization of heparin. Biochemistry. 15: 3932-3942. 1976.

33. Ibanez. C. F., Efhendal. T.. Barhany. G.. Murray-Rust. J., Blundell. T. L.. andPersson. H. Disruption of the low affinity receptor-binding site in NGF allowsneuronal survival and differentiation by binding to the irk gene product. Cell. 69:329-341. 1992.

34. Fabricant. R. N.. De Larco. J. F.. and Todaro. G. L. Nerve growth factor receptors onhuman melanoma cells in culture. Proc. Nati. Acad. Sci. USA. 74: 565-569. 1977.

35. Kaplan, D. R.. Hempstead. B. I... Martin, Z. D., Chao. M. V.. and Parada, L. F. TheIrk proto-oncogene product: a signal transducing receptor lor nerve growth factor.Science (Washington IX'), 252: 554-558, 1991.

36. Nakajima. M.. Irimura. T., Di Ferrante. D.. Di Ferrante. N.. and Nicolson, G. L.Heparan sulfate degradation: relation to tumor invasive and metastatie properties ofmouse BI6 melanoma suhlines. Science (Washington DC). 22ft- 611-613. 1983.

37. Nakajima. M.. Irimura. T.. Di Ferrante. N., and Nicolson. G. L. Metastatie melanomaheparanase. J. Biol. Chem.. 259: 2283-2290. 1984.

38. Sjoberg. !.. and Fransson. L-A. Structural studies on hcparan sulphate from humanlung fibroblasls. Biochem. J.. /9/. 103-110, 1980.

39. Ernst. S.. Langer. R., Cooney. C. L.. and Sasisekharan. R. Hn/ymatic degradation ofglycosaminoglycans. Crii. Rev. Biochem. Mol. Biol.. 30: 387-444. 1995.

40. Gallagher. J. T.. Turnbull. J. E., and Lyon. M. Heparan sulphate proteoglycans.Biochem. Soc. Trans.. IX: 207-209. 1990.

41. Fransson. L. A.. Havsniark. B.. and Sheehan. J. K. Self-association of heparan sulfate.Demonstration of binding by affinity Chromatograph)' of free chains on hcparansulfate-substituted agarose gels. J. Biol. Chem., 256: 13039-13043. 1981.

42. Sloane. B. F., Dunn, J. R., and Honn. K. V. Calcium spike clectrogcncsis and otherelectrical activity in continuously cultured small cell carcinoma of the lung. Science(Washington DC). 2/2: 1151-1153. 1981.

43. Sloane. B. F.. Honn. K. V.. Sadler. J. G., Turner. W. A.. Kimpson. J. J.. and Taylor.J. D. Cathepsin B activity in BI6 melanoma cells: a possible marker lor metastatiepotential. Cancer Res.. 42: 980-986. 1982.

44. Nakajima. M.. De Chavigny. A.. Johnson. C. E.. Hamada. J.. Stein. C. A., andNicolson, G. L. Suramin. J. Biol. Chem.. 266. 9661-9666, 1991.

45. Menter. D. Ci.. Herrmann, J. L.. and Nicolson, G. L. The role of trophic factors andautocrine/paracrine growth factors in brain metastasis. Clin. Exp. Metastasis, 13:67-88. 1995.

46. Palei. J. K.. Didolkar. M. S.. Piekren. J. W.. and Moore. R. H. Metastatic pattern ofmalignant melanoma: a study of 216 autopsy cases. Am. J. Surg.. /.Õ5:807-810,

1978.47. Hcrlyn. M. Human melanoma: development and progression. Cancer Metastasis Rev.,

V: 101-112. 1990.

48. Jin. L.. Nakajima, M.. and Nicolson, G. L. Iminunochemical localization of heparanase in mouse and human melanomas. Int. J. Cancer. 45: 1088-1095. 1990.

49. Whittemore. S. C.. and Seiger. A. The expression, localization anil functional significance of ß-nervegrowth factor in the central nervous system. Neurosci. Rev.. 12:439-464, 1987.

50. Hefli. F. Neurotrophie factor therapy for nervous system degenerative diseases.J. Neurobiol.. 25: 1418-1435. 1994.

51. Menter. D. Ci.. Herrmann. J. L.. Marchetti. D.. and Nicolson. Ci. L. Involvement ofneurotrophins and growth factors in brain metastasis formation. Invasion Metastasis.14: 372-384. 1995.

2863

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1996;56:2856-2863. Cancer Res   Dario Marchetti, David J. McQuillan, William C. Spohn, et al.   Degradation of Specific Heparan Sulfate SubpopulationsSelected Enhancement of Heparanase Activity and Heparanase Neurotrophin Stimulation of Human Melanoma Cell Invasion:

  Updated version

  http://cancerres.aacrjournals.org/content/56/12/2856

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

  .permissions@aacr.orgDepartment at

To request permission to re-use all or part of this article, contact the AACR Publications

on June 24, 2015. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from