functional proteomic analysis of melanoma …...[cancer research 63, 6716–6725, october 15, 2003]...

[CANCER RESEARCH 63, 6716–6725, October 15, 2003]

Functional Proteomic Analysis of Melanoma Progression1

Karine Bernard, Elizabeth Litman, James L. Fitzpatrick, Yiqun G. Shellman, Gretchen Argast, Kirsi Polvinen,Allen D. Everett, Kenji Fukasawa, David A. Norris, Natalie G. Ahn, and Katheryn A. Resing2

Department of Chemistry and Biochemistry [K. B., G. A., K. A. R., N. G. A.] and Howard Hughes Medical Institute [E. L., K. P., N. G. A.], University of Colorado, Boulder,Colorado 80309; Department of Dermatology, School of Medicine, University of Colorado Health Sciences Center, Denver, Colorado 80262 [J. L. F., Y. G. S., D. A. N.];Department of Pediatrics, School of Medicine, University of Virginia, Charlottesville, Virginia 22908 [A. D. E.]; and Department of Cell Biology, University of Cincinnati,Cincinnati, Ohio 45267 [K. F.]

ABSTRACT

Functional proteomics provides a powerful approach to screen foralterations in protein expression and posttranslational modifications un-der conditions of human disease. In this study, we use protein screening toexamine markers of melanoma progression, by profiling melanocyte ver-sus melanoma cell lines using two-dimensional electrophoresis and massspectrometry. Eight candidate markers were identified as differentiallyregulated in transformed cells. In particular, hepatoma-derived growthfactor (HDGF) and nucleophosmin B23 were strongly correlated withmelanoma. Nucleophosmin B23 is a nucleolar and centrosome-associatedprotein, which has been implicated as a target for cyclin E/cyclin-depend-ent kinase 2 (CDK2) in modulating centrosome duplication and cell cyclecontrol. Western blotting of one-dimensional and two-dimensional gelsshowed that the form of nucleophosmin B23 that is up-regulated inmelanoma represents a posttranslationally modified form, most likelyreflecting enhanced phosphorylation in the tumor-derived cells. In con-trast, Western analysis of HDGF demonstrated increased expression of allforms in melanoma cell lines compared with melanocytes. Immunohisto-chemical analysis of human tissue biopsies showed strong expression ofHDGF in early and late stage melanomas and low expression in melano-cytes and nontumorigenic nevi. Interestingly, biopsies of nevi showed agraded effect in which HDGF immunoreactivity was reduced in nevoidnests penetrating deep into the dermis compared with nests at the epider-mal-dermal junction, suggesting that HDGF expression in nevi is depend-ent on epidermal cell interactions. In contrast, biopsies of melanomashowed strong expression of HDGF throughout the tumor, including cellslocated deeply within dermis. Thus, expression of this antigen likelyreports a reduced dependence of protein expression on epidermal inter-actions.

INTRODUCTION

Cutaneous malignant melanoma is among the fastest increasingmalignancies in humans, featuring a capacity to occur in youngpersons and a refractoriness to therapy once metastases have occurred.The paradigm is that melanoma tumors develop from melanocytesafter progressive stages. In this model, sunlight induced mutations ofmelanocytes produce nontumorigenic nevi; these may form atypicalor dysplastic nevi, with relatively high probability for transition toprimary melanoma. Early primary melanomas are characterized byhorizontal growth confined to the epidermis (RGP),3 but can developinto advanced primary melanomas characterized by more aggressivegrowth and dermal invasion (VGP). Late stage VGP tumors arealmost always associated with metastases (1).

The process of melanocyte transformation is also characterized by

reduced anchorage dependence, contact inhibition, and dendricity andincreased surface mobility (2). Important goals are to identify proteinspreferentially expressed during melanoma progression, and to developmarkers for molecular pathology. In particular, markers that distin-guish early melanoma from benign melanocytic nevi as well asmarkers that distinguish RGP melanomas from later malignant stageswould be useful diagnostically. Several molecular markers of mela-noma progression have been identified using different approaches,including antibody screening (3), differential or subtractive cloningprocedures (4), and microarrays (5, 6).

Proteomics is a powerful screening method for alterations in proteinexpression and posttranslational modifications. Recent protein profil-ing in human tumors has been successful in identifying proteinsdifferentially expressed between hepatoma and normal liver cell lines(7), as well as markers that define the degree of differentiation ofbladder squamous cell carcinoma (8). In this study, we assess mela-noma progression by profiling proteins in melanocyte versus mela-noma cell lines using 2-DE and mass spectrometry, from which eightcandidate markers are identified. Two candidate markers were nu-cleophosmin/B23 and HDGF, both which were strongly up-regulatedin melanoma, and cathepsin D which was down-regulated in mela-noma cell lines. Follow-up studies of HDGF expression in humantissue biopsies showed behavior similar to those of cell lines, whereHDGF immunoreactivity discriminates between nontumorigenic nevi/melanocytes versus later stage melanoma tumors. Thus, functionalproteomics successfully reveals potential markers of tumor initiation.

MATERIALS AND METHODS

Cell Culture. All melanoma cell lines originated from the laboratory ofMeenhard Herlyn, Wistar Institute, except for A375 and HS294-T, which wereobtained from American Type Culture Collection. Melanoma cells were cul-tured in 75-cm2 T-flasks containing 25 ml 10% fetal bovine serum, RPMI1640, 5 �g/ml insulin. At 80% confluence, cells were washed once with PBSand incubated for 24 h in 0.05% fetal bovine serum-RPMI 1640. Primarymelanocyte lines FOM 71 and FOM 78 were explanted from patients andcultured as described (9). For harvesting, cells were washed twice with 13 mlof PBS, thoroughly drained, and lysed with 700 �l of lysis buffer containing7 M urea, 2 M thiourea, 4% (w/v) 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid, 1% immobilized pH gradient buffer (pH 4–7; Amer-sham), 1 mM DTT, 1 mM benzamidine, 25 �g/ml leupeptin, 20 �g/ml pepstatinA, 10 �g/ml aprotinin, 1 mM sodium orthovanadate, and 1 �M microcystin.Protein concentrations were determined using the Coomassie Plus reagent(Pierce).

2-DE. Chemicals and reagents used were analytical grade or better. Meth-ods for separation of proteins by 2-DE were performed as detailed in previouspublications (10, 11). Isoelectric focusing was carried out using commercialImmobiline Dry Strips (18 cm; pI 4–7) rehydrated overnight in presence of150 �g of protein lysate and run on a Multiphor II flatbed electrophoresis unitwith an EPS 3500 XL power supply (Amersham). Electrofocusing was per-formed for 28 h at a maximum current of 2 mA and maximum power of 7 Wusing the following gradient: Ramp 0–300 V over 0.01 h; constant 300 V for1 h; ramp 300–3500 V over 9 h; constant 3500 V for 18 h. Before the seconddimension, each immobilized pH gradient strip was reduced in 1% DTT for 15min and alkylated in 4% iodoacetamide for 15 min. Second-dimension SDS-PAGE was performed using 20 � 20-cm slab gels cast and run on a Protean

Received 3/27/03; revised 7/24/03; accepted 7/31/03.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.

1 Supported by NIH Grant R01 CA87648 (K. A. R.) and a postdoctoral fellowshipfrom the Association pour la Recherche contre le Cancer (K. R. B.).

2 To whom requests for reprints should be addressed, at Department of Chemistry andBiochemistry, University of Colorado, Boulder, CO 80309.

3 The abbreviations used are: RGP, radial growth phase; VGP, vertical growth phase;2-DE, two-dimensional electrophoresis; 1-DE, one-dimensional electrophoresis; HDGF,hepatoma-derived growth factor; pI, isoelectric point; ACN, acetonitrile; DAB, 3,3�-diaminobenzidine; TBS, Tris-buffered saline; CDK, cyclin-dependent kinase.

6716

Research. on November 26, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

II xi system (Bio-Rad) for 16 h at 29 mA/gel. Analytical gels were silverstained as described by Blum et al. (12), whereas preparative gels were silverstained using a low fixation protocol (13). Gels were dried between sheets ofcellophane and scanned at 300 dots per inch (UMAX Powerlook II) intransmission mode.

2-DE gels were analyzed using Melanie III software (GeneBio). After spotdetection, percent volume intensities (% vol) were determined by integratingdigitized staining intensities over the areas of individual features, normalizingby total intensity of all features, and multiplying by 100. Comparison ofmelanocyte, RGP, VGP, and metastatic classes of cell lines was assesed usingthe nonparametric Kolmogorov test, the Wilcoxon test, and the parametricStudent t test. Results were considered significant when statistical scores were�0.99 for the Student t test, 2.7 for Wilcoxon, and 0.75 for Kolmogorov-Smirnoff.

Mass Spectrometry. In-gel digestions were performed on silver-stainedproteins as described (14). Proteins were excised from multiple wet gels orfrom dried gels followed by rehydration with deionized water. Gel piecescorresponding to one protein were pooled and washed with 100 mM ammo-nium bicarbonate (NH4HCO3) for 20 min, and 50% ACN-50 mM NH4HCO3

for 20 min. The gel pieces were dehydrated with 50 �l of ACN for 10 min,vacuum dried in a Speedvac (Heto), reswelled in 10–30 �l of sequencing gradeporcine trypsin (Promega) in 25 mM NH4HCO3 to a final concentration of 20ng/�l trypsin, and incubated overnight at 37°C. Supernatants were removedand transferred to separate tubes, and gel pieces were reextracted twice with 50�l of 60% CAN-0.1% trifluoroacetic acid for 20 min in a sonicating waterbath. Supernatants were pooled and dried in a Speedvac, and peptides wereresuspended in 5 �l of 0.1% formic acid. When necessary, C18 silica desalting/concentration microcolumns (ZipTip; Millipore) were used to purify andconcentrate peptides before mass spectrometry.

Peptide digests were analyzed by matrix-assisted desorption ionization-timeof flight mass spectrometry using an ABI Voyager DE-STR instrument (Ap-plied Biosystems) with delayed extraction. Tryptic digests were mixed 1:1(v/v) with �-cyano-4-hydroxy-trans-cinnamic acid (Hewlett Packard) andspotted onto gold-plated matrix-assisted desorption ionization sampleplanchettes. Mass spectra were summed over 150 acquisitions (20 kV accel-erating voltage, 10 V guide wire voltage, 100 ns delay) using Voyager 5.0software, and spectra were processed using Data Explorer. Internal calibrationswere performed using tryptic autolysis fragments at 842.51, 1045.56, 2211.10Da, and masses from contaminating keratin ions were eliminated. Peptidefingerprinting data were matched against the National Center for Biotechnol-ogy Information nonredundant database using ProFound (http://www.

expasy.ch/tools/pi_tool.html), with mass tolerance of 0.1 Da and allowing 1missed cleavage. Artifactual modification of peptides by acrylamide adductswith cysteine and methionine oxidation was considered during databasesearching. Evaluation of results considered the number of peptides matched toidentified proteins, percent coverage of protein sequence, and apparent mass/pIof proteins.

Peptides in tryptic digests were then sequenced using a Q-STAR Pulsarq-time of flight mass spectrometer (Applied Biosystems) equipped with ano-matrix-assisted desorption ionization interface. Accurately measured massesof tryptic peptides and fragment ions were used to search for protein candidatesin the NCBInr database using the program MS-Tag in Protein Prospector(http://prospector.ucsf.edu).

Immunoblotting. Cell lysates were separated by 1-DE or 2-DE, thentransferred to polyvinylidene difluoride (NEN). Primary antibodies used toprobe blots included mouse antinucleophosmin/B23 (1:1000) (15) and rabbitanti-HDGF (1:4000) (16). Blots were probed with primary and secondaryantibodies diluted into Tris-borate-saline, 0.1% Tween 20, and 5% nonfat driedmilk. Blots were then probed with donkey antimouse, goat, or rabbit secondaryantibodies (1:5000) coupled to horseradish peroxidase (The Jackson Labora-tory) and visualized by enhanced chemiluminescence (PerkinElmer).

Immunohistochemistry. Sections from formalin-fixed, paraffin-embeddedpatient biopsies were analyzed. Samples were deparaffinized and rehydrated,and antigen was retrieved using antigen retrieval solution (DAKO Cytomation)in a decloaking chamber (Biocare Medical) first at 120°C for 30 s and then at85°C for 10 s. Slides were then cooled to 25–30°C and rinsed with tap water.Sections were either double stained for HDGF and S100 or S100B (to distin-guish melanocytes) or single-stained for HDGF alone. Primary antibody dilu-tions used were 1:1200 for S100 (polyclonal; DAKO Cytomation), 1:25 forS100B (DAKO Cytomation), and 1:750 for HDGF [affinity purified poly-clonal; Everett et al. (16)]. The order in which sections were double stainedwith S100/S100B � DAB and HDGF � Vector Red were varied to verifyconsistent results. Negative controls (mouse, rabbit; DAKO Cytomation) wereused to verify low background staining. Double staining for S100/S100B andHDGF was performed using the DAKO EnVision Doublestain System (DAKOCytomation) according to the manufacturer’s directions except that beforeincubation with HDGF, sections were treated with Powerblock UniversalBlocking Reagent (BioGenex), HDGF was incubated for 30 min at roomtemperature, and Vector Red Alkaline Phosphatase Substrate Kit 1 (VectorLaboratories) was used in place of Fast Red, which allowed examination byboth bright-field and fluorescence microscopy. For immunohistochemistrywith anti-HDGF antibody alone, samples were blocked with Powerblock,

Table 1 Staging and behavior of cultured melanocytic cells isolated from primary and metastatic melanoma

Cell linesaAJCC

stageb/pathology Clark’s levelc/thickness (mm3)Soft agar growth (CFE)d

(%) Tumorigenicitye

Early primary melanoma cell lines(RGP-VGP)

WM3208V Stage X 1.28 NoSBCL2 0.34 NoWM35 Stage 1/SSM Level II/0.69 1.0 YesWM1789 Stage 1/SSM Level III/0.82 �0.1 YesWM1552C Stage 3/SSM Level IV/5.92 8.32

Primary melanoma (VGP) WM75 Stage 3/SSM Level IV/6.25 10.8 YesWM39 19.1 YesWM278 Stage 2/NM Level IV/3.7 4.2 YesWM115 Stage n.a./SSM Level III/2.24 6 YesWM793B Stage 1/SSM Level II/0.55 3.71–7.2 Yes

Metastasis Corresponding primary Metastatic site

Metastatic melanoma WM164 Lymph node 2 NoLu 451 WM164 Lung (selection in mice) 17 YesWM1232WM852 Skin 22.7 YesWM1617 WM278 Lymph node n.a.f YesWM239A WM115 Lymph node 90 Yes1205Lu WM793 Lung 25 YesHS294-T Lymph node YesA375 Skin Yes

a Names in bold indicate cell lines used in the initial 2-DE screen. Additional cell lines were used in 1-DE immunoblotting.b Disease staging is based on the manual for staging of cancer by the American Joint Committee on Cancer (AJCC). RGP, radial growth phase; VGP, vertical growth phase; SSM,

superficial spreading melanoma; NM, nodular melanoma. Data were compiled from Refs. 9 and 17–20.c Anatomic levels of invasion are based on Clark’s microstaging criteria.d CFE, colony forming efficiency.e Total of 5 � 106 cells/mouse were injected. Tumor growth was monitored after 10 wk.f n.a., not analyzed.

6717

FUNCTIONAL PROTEOMIC ANALYSIS OF MELANOMA PROGRESSION



rinsed twice for 3 min with TBS (DAKO Cytomation) with 0.05% Tween 20(Sigma), and incubated for 30 min at room temperature with anti-HDGFantibody in dilution buffer (TBS with 1% BSA, 0.01% sodium azide). Spec-imens were rinsed 3 � 3 min in TBS with 0.05% Tween 20 and then incubatedwith DAKO Envision Alkaline Phosphatase (DAKO Cytomation) for 30 min.After TBS-Tween 20 rinses, the sections were treated with Vector Red Alka-line Phosphatase Substrate Kit 1 according to the manufacturer’s directions.All sections(single or double stained) were counterstained with hematoxylinand azure B (2.5 mg/ml Azure B; Sigma; 8.5 mM acetic acid, 1.5 mM sodiumacetate, and 12.5% acetone), the latter which stains melanin green, and thenincubated in bluing reagent (Richard Allan Scientific) for 1 min.

RESULTS

Profiling of Melanocyte versus Melanoma Cell Lines by 2-DEand Mass Spectrometry. Whole cell lysates were made from 2primary melanocyte cell strains and 12 melanoma cell lines (Table1). The latter included representative cell lines cultured from RGP,VGP, and metastatic tumors that were catalogued with respect totumor pathology as well as tumorigenicity and invasiveness ofcells in culture and in nude mice (17–20). Extracts from each cellline were analyzed in replicate by 2-DE, using pH 4 –7 focusinggradients in the first dimension and 8 –18% polyacrylamide gradi-ents in the second dimension. Proteins were visualized by silverstaining, which yields highest sensitivity but limited dynamicrange because of saturation. Therefore at least four well-resolvedgels from three different experiments were analyzed for each cellline, resolving �2000 distinct protein spots on each gel. A repre-

sentative gel (Fig. 1) shows resolution of proteins from an earlyprimary melanoma cell line (WM35).

The 14 selected cell lines were assigned into melanocyte, RGP,VGP, and metastatic classes based on tumor pathology and in vitrosoft agar growth. Protein features were clustered into the four classesbased on intensity, normalized to the sum of intensities of all spots onthe gel, and quantified as percentage volume by gel image analysissoftware. Data for individual spots were then averaged over replicategels, and results from automated imaging were confirmed by directvisual inspection of gels. Eight protein spots (Fig. 1) were identifiedin this screen as candidate markers, based on differences in averagednormalized intensity between classes as described in “Materials andMethods.”

Each of the spots was processed by in-gel digestion and analyzedby peptide mass fingerprinting and tandem mass spectrometry. Alleight proteins (M1–M8) were matched to entries in the human non-redundant database. The relative normalized intensities within differ-ent cell lines for these proteins are indicated in Fig. 2, and character-istics of the proteins are summarized in Table 2.

Biochemical Characterization of Candidate Melanoma Mark-ers. Inspection of Fig. 2 illustrates that candidate markers, HDGF(M1) and nucleophosmin/B23 (M2), showed significant and repro-ducible differences between melanocytes and all of the melanoma celllines, whereas cathepsin D (M3, M7) and glutathione S-transferase-�(M6) showed differences between melanocytes and more than one-half of the tumor lines. Other markers, such as quinolinate phospho-

Fig. 1. Representative 2-DE gel, showing resolution of cellular proteins extracted from the RGP melanoma cell line, WM35. As described in “Materials and Methods,” 150 �g ofsolubilized protein (�2 � 106 cells) were separated and visualized by silver staining. Each cell line was examined in triplicate for features that varied with stages in melanomaprogression. Arrows, protein spots identified in the screen that were expressed in WM35.

6718




ribosyltransferase (M4) and 14-3-3� (M8) varied between differentmelanoma stages and are candidates for distinguishing RGP versusVGP cell lines. Further biochemical characterizations of HDGF, ca-thepsin D, and nucleophosmin/B23 were performed on the cell linesused in the screen as well as five additional melanoma cell lines(Table 1).

Nucleophosmin/B23 Is Posttranslationally Modified in Mela-noma Cells. An important advantage of functional proteomic anal-ysis by 2-DE is the ability to reveal changes in protein covalentmodifications. This was illustrated by the behavior of nucleophos-min/B23 during melanoma progression. Silver-stained 2-DE gelsshowed an increase in a single spot (M2) corresponding to nucleo-phosmin/B23, which was located immediately adjacent to twohigher intensity spots that were invariant with cell type (M2b,M2c; Fig. 3A). Mass spectrometry/mass spectrometry analysisidentified the two invariant spots as nucleophosmin/B23, indicat-ing that M2 represents an acidic isoform increased in melanomacells. This was corroborated by immunoblotting 2-DE gels, whichshowed immunoreactivity of all three spots as well as an obviousincrease in immunoreactivity of the form with acidic pI (Fig. 3B).Examination of 19 cell lines by 1-DE immunoblotting showedvariable expression of nucleophosmin/B23 in different mela-

noma cell lines, compared with melanocytes (Fig. 3C). Thus,increased posttranslational modification of nucleophosmin/B23,rather than overall protein expression, correlated with melanomaprogression.

Cathepsin D (Molecular Mass 34 kDa) is Overexpressed inMelanocytes. Cathepsin D is a ubiquitous lysosomal aspartyl en-doproteinase which is processed from an inactive 52-kDa proca-thepsin D to an active 48-kDa single-chain molecule in late endo-somes and early lysosomes (21). The 48-kDa chain is then furtherprocessed in lysosomes to a 14-kDa light chain, derived from theNH2 terminus, and a 34-kDa heavy chain, derived from the COOHterminus. In the 2-DE screen, cathepsin D was observed as twosilver-stained protein spots (Fig. 4A, M3, M7). Both had massescorresponding to the 34-kDa form and showed reduced intensitiesin melanoma cell lines compared with melanocytes. Immunoblotsof 2-DE gels, probed with anticathepsin D antibodies directedagainst the COOH terminus revealed between two and five differ-ent forms in different cell lines, each with various pI values but allwith similar mass. In some cell lines (e.g., SBCL2 and 1205Lu),the reduced total amount of the 34-kDa heavy chain correlated withincreases in acidic isoforms (Fig. 4B). Thus, multiple forms ofcathepsin D were present in various cell lines, but only spots

Fig. 2. Quantitation of variable protein abun-dances in melanocyte and melanoma cell lines. Spotintensities were quantified and normalized as de-scribed in “Materials and Methods” and averagedthree or more gels. Eight features that varied withmelanoma progression are indicated. The cell linesexamined were: A, FOM 71; B, FOM 78; C,WM3208; D, SBCL2; E, WM35; F, WM1789; G,WM278; H, WM793; I, WM115; J, A375; K,HS294-T; L, WM1617; M, WM239A; N, 1205Lu.

6719




corresponding to M3 and M7 were consistently down-regulated inmelanoma compared with melanocytes. Immunoblots of 1-DE gelsconfirmed that the 34-kDa form was present in melanocytes andsignificantly reduced in 15 of 19 melanoma cell lines (Fig. 4C).The 34-kDa heavy chain was the major immunoreactive form inboth melanocytes and melanoma (data not shown).

HDGF Is Overexpressed in Melanoma. HDGF is an intracellularprotein that is mitogenic in several cell types, including endothelialand smooth muscle cell systems. Although this protein was identifiedby silver staining as a single protein spot in the initial screen (Fig. 5A),immunoblotting of 2-DE gels revealed that the antigen resolved intomultiple forms each with the same apparent molecular mass, indicat-ing covalent modification of this marker (Fig. 5B). Nevertheless,immunoblotting after 1-DE revealed that all forms of HDGF showedelevated expression in 19 melanoma cell lines compared with mela-

nocytes (Fig. 5C), indicating strong correlation of expression withearly stage melanoma.

Additional experiments were performed to investigate the ex-pression of HDGF in human melanoma biopsies. Tissue sections ofprimary melanomas at various stages of progression were immu-nostained with anti-HDGF, counterstained with hematoxylin-azureB, and visualized by light microscopy. Nuclear reactivity withanti-HDGF was observed within the epidermal layer in keratino-cytes, which is consistent with the previous characterization of thisantigen as a nuclear protein that is expressed in epithelial cells(Fig. 6A and B). Intense staining was also observed in all stages ofmelanomas, which included in situ melanomas, level II tumorsconfined within the epidermis (in which three of four biopsies werelentigo maligna melanomas), as well as level IV tumors that werestrongly invasive into the dermal layer (Fig. 6, A–F). The speci-

Fig. 3. A modified form of nucleophosmin/B23 is up-regulated inmelanoma cell lines. A, silver-stained 2-DE gels of melanocyte andmelanoma lines illustrate the up-regulation of spot M2 in melanomacells compared with melanocytes. Mass spectrometry/mass spec-trometry analysis of adjacent spots M2b and M2c showed that bothproteins also correspond to nucleophosmin/B23. B, immunoblottingof 2-DE gels show reactivity of M2, M2b, and M2c with antinu-cleophosmin/B23. Thus, the M2 feature most likely represents amore acidic, posttranslationally modified form of this protein. C,immunoblotting of 1-DE gels, probing 2 melanocyte and 19 mela-noma cell lines (10 �g load) show variable expression of nucleo-phosmin/B23 in melanoma compared with melanocytes.

Table 2 Proteins from functional proteomics screening identified by mass spectrometry

FeaturesID Protein name

NCBIsequence ID

Molecular mass(kDa)obs

aMolecular Mass

(kDa)calcc pIobs pIcalc

No. of peptidesmatched by

mass ID

Sequencecoverage of

observedpeptides

(%)No. of peptides

sequencedRepresentative sequenced

peptides

M1 HDGF gi4758516 38.1 26.8 4.6 4.7 10 28 1 DLFPYEESK (1127.5)M2 Nucleophosmin (nucleolar

phosphoprotein B23)gi16307090 38.2 29.6 4.6 4.5 4 14 2 FINYVK (783.4)

VDNDENEHQLSLR (1568.7)M3 Cathepsin D heavy chain P07339 28.5 26.6 5.6 5.5 7 26 3 VGFAEAAR (820.4)

YYTVFDR (963.4)LVDQNIFSFYLSR (1601.8)

M4 Nicotinate-nucleotidepyrophosphorylase

Quinolinatephosphoribosyltransferase

Q15274 32.0 30.9 5.9 5.8 8 24 3 YGLLVGGAASHR (1200.6)GAGWTGHVAGTR (1169.5)VALNTLAR (857.5)

M5 Tryptophanyl-tRNA synthetaseInterferon-induced protein p53

P23381 60.0 53.2 5.8 5.8 8 20 1 RGIFFSH (863.4)

M6 Glutathione S-transferase � P78417 28.5 27.5 6.1 6.2 12 52M7 Cathepsin D heavy chain P07339 28.5 26.6 5.3 5.6 5 16 2 YYTVFDR (963.4)

LVDQNIFSFYLSR (1601.8)M8 14-3-3 protein � Q9UN99 27.4 28 4.7 4.7 7 25 1 DSTLIMQLLR (1189.6)a obs, observed; calc, calculated.

6720




ficity of the antibody is indicated by peptide competition controls(Fig. 6, O and P).

In contrast, melanocytes within the epidermal layer, which wereidentifiable by their hyperchromogenicity and dissociation from basalkeratinocytes, were either negative or weakly stained (Fig. 7). Fur-thermore, nests of nontumorigenic melanocytes forming benign oratypical/dysplastic nevi also showed reduced staining with anti-HDGF, compared with the melanomas (Fig. 6, G–L). Staining inten-sities in nevi revealed a graded effect, in which cells within junctionalnests were more positive for HDGF reactivity than cells deep withinthe dermal layer (Fig. 6, G and I). Interestingly, Spitz nevi werevariably reactive, where many biopsies showed staining intensities

comparable with advanced melanoma (Fig. 6, M and N). Of 24 nevus andtumor samples examined immunohistochemically, 54% of cells withinbenign nevoid nests were positively reactive, whereas 78–90% of cellswithin melanoma tumors were reactive (Table 3). Although the differ-ences between melanocyte/nevi versus melanoma were not absolute,quantitative differences in the number of immunoreactive cells werenevertheless observed. In addition, the level of staining within positivelyreactive cells differed substantially between transformed and nontrans-formed tissue, showing weak reactivity in nevoid cells and strong reac-tivity in tumor cells. The results indicate that HDGF expression in humantissues is correlated with early stages of melanoma, in a manner that isrecapitulated by expression behavior in cell lines.

Fig. 4. Reduced cathepsin D expression in melanoma. A, silver-stained 2-DE gels of extracts, illustrating the up-regulation of spotsM3 and M7 in melanocytes compared with melanoma cells. B,immunoblotting 2-DE gels with anticathepsin D shows reactivity ofseveral features, indicating posttranslationally modified forms ofcathepsin D. M3 and M7 represent the most abundant forms of thisprotein in melanocytes. C, immunoblotting of 1-DE gels, probingmelanocyte and melanoma cell lines (10 �g load) with anticathepsinD, shows high expression in melanocytes and low expression inmost but not all melanoma cell lines.

Fig. 5. Enhanced HDGF expression in melanoma cell lines. A,silver-stained 2-DE gels of extracts, illustrating the up-regulationof spot M1 in melanoma cells compared with melanocytes. Thespots located to the right of M1 are identical with nucleophosmin/B23. B, immunoblotting 2-DE gels with anti-HDGF shows reac-tivity of several features, indicating posttranslationally modifiedforms of HDGF which did not vary with melanoma progression.Spot M1 represents the most abundant form of this protein. C,immunoblotting of 1-DE gels, probing melanocyte and melanomacell lines (10 �g load). Enhanced HDGF expression is highlycorrelated with all melanoma lines, compared with melanocytes.

6721




DISCUSSION

In this study, candidate molecular markers of melanoma pro-gression were identified using 2-DE and mass spectrometry. Theproteomic approach allowed comparative analysis of many geneproducts, using a unique set of melanoma cell lines that werecarefully catalogued with respect to tumor staging and includedcells grown from early primary melanoma (22). Most of theproteins identified showed variations between melanocyte andmelanoma cell lines and may reflect changes occurring on tumorinitiation. In two examples (quinolinate phosphoribosyltransferaseand 14-3-3�), changes in intensity appeared to correlate withprogression from RGP to VGP tumors. Most striking were theabundances of HDGF and an acidic form of nucleophosmin/B23,

which were elevated in all melanoma cell lines compared withmelanocytes, and are thus good candidates for the development ofdiagnostic probes.

Most of the tumor-deregulated proteins identified in the screenwere sample specific, as reported in earlier studies of prostate cancer(23) and breast cancer (24, 25). In a study that systematically inves-tigated the genetic profiles of primary tumors and correspondingmetastases in a number of melanoma patients, the most frequentgenetic alteration detected during metastatic progression was the lossof p16INK4a, which was detected in only 4 of 14 cases (26). Inanother study examining tumoral heterogeneity by 2-DE in ovariancancer, a large degree of intertumoral heterogeneity observed,whereas intratumoral heterogeneity was very low (27). There are

Fig. 6. HDGF overexpression in human melanoma andnevoid biopsies. In A–N, paraffin sections were preparedand stained with anti-HDGF � Vector Red (A, C, E, G, I,K, and M) or anti-HDGF � Vector Red and anti-S100 � DAB (B, D, F, H, J, L, and N), as described in“Materials and Methods.” Representative biopsies showadvanced malignant melanoma (Clark’s IV; A–D), lentigomaligna melanoma (Clark’s II; E and F), atypical com-pound nevus (G and H), benign nevus (I and L), and Spitznevus (M and N). O and P, antigen competition controlshowing advanced malignant melanoma stained with anti-HDGF � Vector Red (O) or with anti-HDGF (P) afterpeptide competition.

6722




various reasons why observed protein changes were not more con-sistent within each class of melanoma cell lines in the study. First, themolecular complexity of melanoma tumor biology may result in fewalterations in common between tumors. In addition, cell lines derivedfrom tumors likely accumulate further mutations and protein changesduring long-term culture. Together, these results support the hypoth-esis that early and fundamental mutations occur in key genes to drivemalignancy, followed by accumulation of further mutations that arenot specifically found in all tumors.

Our study illustrates an advantage of expression profiling at theproteomic level in that it revealed a correlation in melanoma lineswith a posttranslationally modified form of nucleophosmin/B23. Al-though the M2 protein spot showed intensity that increased in amanner strongly correlated with melanoma, analysis of total nucleo-phosmin/B23 protein expression by 1-DE immunoblotting showed

incomplete correlations. Further analysis of neighboring silver-stainedspots by mass spectrometry as well as 2-DE immunoblotting revealedat least three spots that represented differentially modified forms ofnucleophosmin/B23. The spot with most acidic pI was up-regulated inthe initial screen, indicating that covalent modification of nucleophos-min/B23 is elevated in melanoma, an event that would not have beenobserved in experiments profiling mRNA or total protein.

Nucleophosmin/B23 localizes to centrosomes as well as nucleoliand regulates centrosome duplication in a cell cycle-dependent man-ner (15). It is phosphorylated at several residues by various proteinkinases, including CDK1-cyclin B (28), casein kinase II (29), andCDK2-cyclin E (15), the latter event which promotes centrosomeduplication during S-G2. Furthermore, overexpression of nucleophos-min/B23 decreases the susceptibility of human leukemia HL-60 cellsto retinoic acid-induced differentiation and apoptosis (30), and in-

Fig. 6. Continued.

6723




creased stability of nucleophosmin/B23 protein is involved in theantiapoptotic effect of ras during serum deprivation (31). Thus, phos-phorylation and expression of nucleosphosmin/B23 is correlated withcell cycle progression and cell survival, and this likely reflects itscorresponding regulation and function in melanoma. Phosphorylationof nucleophosmin/B23 has not been specifically described in anytumor type, although elevated nucleophosmin protein and mRNAexpression have been associated with breast cancer, colorectal cancer,and leukemia (32–34). It will be important to identify the posttrans-lational modification(s) relevant to melanoma and develop probes tothe modified residues to test their potential relevance to diseaseprogression and to provide useful reagents for the diagnosis andcharacterization of early melanoma.

Cathepsin D has been previously characterized as a marker ofmalignant cancers, including melanoma, breast, colorectal, and squa-mous cell carcinomas (35). This antigen has been shown to be aprognostic factor in breast cancer in a number of studies (36). How-ever, in malignant melanoma, the results in different studies areconflicting, where the protein has been found to be either positively ornegatively correlated with advanced stages of tumorigenesis (37–40).The appearance of acidic isoforms has been described in two breastcancer cell lines in which the amino acid sequence of cathepsin D wasnormal, but its glycosylation state appeared to vary (21). Our resultsindicate that there are several isoforms for cathepsin D with a mass of34 kDa. In two cell lines, SBCL2 and 1205Lu, we also observed theappearance of additional acidic isoforms.

HDGF is an heparin-binding growth factor, originally purifiedfrom conditioned media of the human hepatoma-derived cell line,HuH-7 (41). HDGF has only recently been characterized, andchanges in expression have never been systematically investigatedin any tumor type, although it promotes hepatocyte proliferationand is correlated with reduced radioresistance in esophageal can-cers (42– 44). The translated cDNA sequence of HDGF revealslittle homology with known growth factors but instead showssequence similarity to non-DNA-binding domains of the chroma-

tin-associated protein, HMG1. HDGF is mitogenic and localizes toboth cytosol and nucleus when expressed in cultured cells, andnuclear targeting of HDGF via a consensus nuclear localizationsequence is required for stimulation of DNA replication in vascularsmooth muscle cells (42). When overexpressed, HDGF can bepurified from conditioned media and can act as a mitogen whenadded exogenously to fibroblasts, HuH-7, aortic endothelial cells,and smooth muscle cells. The mechanism for extrusion to condi-tioned media is unknown, and it is curious given the nuclearlocalization of the expressed protein, but the association withgrowth regulation suggests that HDGF may be causal for trans-forming behavior. Nevertheless, suppression of HDGF in mela-noma cell lines using RNA interference had little effect on extra-cellular signal-regulated kinase activation or cell growth rate.4

In our screen, HDGF expression was highly correlated with melanomain all cell lines examined. This was mainly attributable to intracellularexpression, although the protein was also detected in conditioned mediaof at least one cell line tested (WM35). Several forms of HDGF with thesame apparent mass were observed by 2-DE immunoblotting, suggestingthe existence of posttranslationally modified forms. These most likelycorrespond to differentially phosphorylated forms.5

Examination of human biopsies showed a similar expression pattern, inwhich HDGF was absent or weakly present in nontumorigenic melano-cytes and was expressed at a low level in 54% of cells within nevoidtissues. In contrast, HDGF showed high reactivity in both early and latemelanoma, and highest expression in advanced malignant tumors (Table3). The expression of HDGF was graded with progression. Thus, al-though representation of HDGF in benign nevi was significantly lowerthan Clark’s IV melanoma (54% versus 90%, respectively), atypical andSpitz nevi showed higher percentages of positive cells than benign nevi.Furthermore, early in situ melanoma and Clark’s II (mostly lentigo)melanoma showed lower percentages of positive cells than advancedmalignant tumors. This analysis suggests that HDGF is expressed inbenign nevi, but increases its frequency of expression during the transi-tions into atypical nevi, early melanoma, and late melanoma. Such aprotein may be a useful marker for examining properties of nevoid cellsthat eventually undergo selection for conversion into melanoma.

Staining patterns in nevoid tissue showed heterogeneous reactivity, inwhich cells within junctional nests showed higher expression of HDGFcompared with interdermal nests. In contrast, early and advanced mela-noma tissue showed high reactivity toward HDGF that was observedhomogeneously throughout the tumor and was independent of depth ofdermal penetration. This pattern may reflect a selective process, in whicha subset of junctional cells with high HDGF expression preferentiallyprogress into melanoma. Alternatively, the heterogeneity in nevi mayreflect processes in which HDGF expression in nevoid and melanocytecells are subject to control through their interactions with neighboringepidermal cells. For example, melanocyte and nevoid cell interactionswith keratinocytes may lead to elevated HDGF expression within junc-

4 K. Bernard and N. Ahn, unpublished data.5 A. Everett, unpublished results.

Fig. 7. Melanocytes are negatives for HDGF. The paraffin section was prepared andstained with anti-HDGF � Vector Red and anti-S100 � DAB, as described in “Materialsand Methods.” The picture is from the outside edge of a benign nevi biopsy where normalmelanocytes are present (arrows).

Table 3 Expression of HDGF in human melanoma biopsiesa

Stage No. of biopsies % of positive cells for HDGF

Benign nevi 4 53.7 � 12Nevi with atypia 6 65 � 5Spitz nevi 4 72 � 11Melanoma in situ 2 85 � 0Clark’s II 4 78 � 8Clark’s IV 4 90 � 7

a Biopsies examined for immunoreactivity with anti-HDGF � Vector Red. For eachbiospy, 350-450 cells within defined nevoid nests or tumors were scored for reactivity andpercentages of positive cells were reported.

6724




tional layers, whereas loss of epithelial-melanocyte interactions may leadto reduced HDGF expression within interdermal nevoid nests. This isconsistent with the low expression of this marker in melanocytes inculture, where epithelial cell interactions are absent. Likewise, the depthindependence of HDGF reactivity observed in biopsies of early andadvanced melanoma may report independence from transcriptional reg-ulatory processes controlled by tissue interactions in transformed cells.Future studies are needed to investigate the relevance of the melanocyte-melanoma microenvironment in this process and test the hypothesis thatHDGF expression observed in early melanoma cell lines may also reflecta reduced requirement for tissue cell-cell interactions in transformation.

In summary, this study illustrates the use of functional proteomicsto identify relevant molecular markers of cancer progression. Theseresults may provide a better understanding of tumor biology and mayeventually lead to the design and implementation of new cancerdiagnostics and therapies.

ACKNOWLEDGMENTS

We thank Dr. Meenhard Herlyn, Wistar Institute, for providing us withmelanocyte and melanoma cell lines used in this study. We are also grateful toNorma Aumen, University of Colorado Health Sciences Center, Denver, CO,for expert assistance in preparing paraffin slides of human biopsies.

REFERENCES

1. Graeven, U., and Herlyn, M. In vitro growth patterns of normal human melanocytesand melanocytes from different stages of melanoma progression. J. Immunother., 12:199–202, 1992.

2. Soballe, P. W., and Herlyn, M. Cellular pathways leading to melanoma differentia-tion: therapeutic implications. Melanoma Res., 4: 213–223, 1994.

3. Elder, D. E., Rodeck, U., Thurin, J., Cardillo, F., Clark, W. H., Stewart, R., andHerlyn, M. Antigenic profile of tumor progression stages in human melanocytic neviand melanomas. Cancer Res., 49: 5091–5096, 1989.

4. Weterman, M. A., Stoopen, G. M., van Muijen, G. N., Kuznicki, J., Ruiter, D. J., andBloemers, H. P. Expression of calcyclin in human melanoma cell lines correlates withmetastatic behavior in nude mice. Cancer Res., 52: 1291–1296, 1992.

5. DeRisi, J., Penland, L., Brown, P. O., Bittner, M. L., Meltzer, P. S., Ray, M., Chen,Y., Su, Y. A., and Trent, J. M. Use of a cDNA microarray to analyse gene expressionpatterns in human cancer. Nat. Genet., 14: 457–460, 1996.

6. Su, Y. A., Bittner, M. L., Chen, Y., Tao, L., Jiang, Y., Zhang, Y., Stephan, D. A., andTrent, J. M. Identification of tumor-suppressor genes using human melanoma celllines UACC903, UACC903(�6), and SRS3 by comparison of expression profiles.Mol. Carcinog., 28: 119–127, 2000.

7. Yu, L. R., Zeng, R., Shao, X. X., Wang, N., Xu, Y. H., and Xia, Q. C. Identificationof differentially expressed proteins between human hepatoma and normal liver celllines by two-dimensional electrophoresis and liquid chromatography-ion trap massspectrometry. Electrophoresis, 21: 3058–3068, 2000.

8. Ostergaard, M., Rasmussen, H. H., Nielsen, H. V., Vorum, H., Orntoft, T. F., Wolf,H., and Celis, J. E. Proteome profiling of bladder squamous cell carcinomas: iden-tification of markers that define their degree of differentiation. Cancer Res., 57:4111–4117, 1997.

9. Herlyn, M., Thurin, J., Balaban, G., Bennicelli, J. L., Herlyn, D., Elder, D. E., Bondi,E., Guerry, D., Nowell, P., Clark, W. H., et al. Characteristics of cultured humanmelanocytes isolated from different stages of tumor progression. Cancer Res., 45:5670–5676, 1985.

10. Bernard, K., Jonscher, K. R., Resing, K. A., and Ahn, N. G. Methods in functionalproteomics: 2D-PAGE with immobilized pH gradients (IPG), in gel digestion, andidentification of proteins by mass spectrometry. Methods Mol. Biol., 250: 263–282, 2003.

11. Lewis, T. S., Shapiro, P. S., and Ahn, N. G. Signal transduction through MAP kinasecascades. Adv. Cancer Res., 74: 49–139, 1998.

12. Blum, H., Beier, H., and Gross, H. J. Improved silver staining of plant proteins, RNAand DNA in polyacrylamide gels. Electrophoresis, 8: 93–99, 1987.

13. Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. Mass spectrometric sequencingof proteins silver-stained polyacrylamide gels. Anal. Chem., 68: 850–858, 1996.

14. Rosenfeld, J., Capdevielle, J., Guillemot, J. C., and Ferrara, P. In-gel digestion ofproteins for internal sequence analysis after one- or two-dimensional gel electrophore-sis. Anal. Biochem., 203: 173–179, 1992.

15. Okuda, M., Horn, H. F., Tarapore, P., Tokuyama, Y., Smulian, A. G., Chan, P. K.,Knudsen, E. S., Hofmann, I. A., Snyder, J. D., Bove, K. E., and Fukasawa, K.Nucleophosmin/B23 is a target of CDK2/cyclin E in centrosome duplication. Cell,103: 127–140, 2000.

16. Everett, A. D., Lobe, D. R., Matsumura, M. E., Nakamura, H., and McNamara, C. A.Hepatoma-derived growth factor stimulates smooth muscle cell growth and is ex-pressed in vascular development. J. Clin. Invest., 105: 567–575, 2000.

17. Creasey, A. A., Smith, H. S., Hackett, A. J., Fukuyama, K., Epstein, W. L., andMadin, S. H. Biological properties of human melanoma cells in culture. In Vitro, 15:342–350, 1979.

18. Giard, D. J., Aaronson, S. A., Todaro, G. J., Arnstein, P., Kersey, J. H., Dosik, H., andParks, W. P. In vitro cultivation of human tumors: establishment of cell lines derivedfrom a series of solid tumors. J. Natl. Cancer Inst., 51: 1417–1423, 1973.

19. Herlyn, M., Balaban, G., Bennicelli, J., Guerry, D. T., Halaban, R., Herlyn, D., Elder,D. E., Maul, G. G., Steplewski, Z., Nowell, P. C., et al. Primary melanoma cells ofthe vertical growth phase: similarities to metastatic cells. J. Natl. Cancer Inst., 74:283–289, 1985.

20. Satyamoorthy, K., DeJesus, E., Linnenbach, A. J., Kraj, B., Kornreich, D. L., Rendle, S.,Elder, D. E., and Herlyn, M. Melanoma cell lines from different stages of progression andtheir biological and molecular analyses. Melanoma Res., 7 Suppl 2: S35–S42, 1997.

21. Rochefort, H., Liaudet, E., and Garcia, M. Alterations and role of human cathepsin Din cancer metastasis. Enzyme Protein, 49: 106–116, 1996.

22. Hsu, M.-Y., Elder, D. E., and Herlyn, M. Melanoma: the Wistar melanoma (WM) celllines. In: J. R. W. M. a. B. Palsson (ed.), Human Cell Culture, Vol. 1, pp. 259–274.Dordrecht, the Netherlands: Kluwer Academic Publishers, 1999.

23. Ahram, M., Best, C. J., Flaig, M. J., Gillespie, J. W., Leiva, I. M., Chuaqui, R. F.,Zhou, G., Shu, H., Duray, P. H., Linehan, W. M., Raffeld, M., Ornstein, D. K., Zhao,Y., Petricoin, E. F., 3rd, and Emmert-Buck, M. R. Proteomic analysis of humanprostate cancer. Mol. Carcinog., 33: 9–15, 2002.

24. Perou, C. M., Sorlie, T., Eisen, M. B., van de Rijn, M., Jeffrey, S. S., Rees, C. A., Pollack,J. R., Ross, D. T., Johnsen, H., Akslen, L. A., Fluge, O., Pergamenschikov, A., Williams,C., Zhu, S. X., Lonning, P. E., Borresen-Dale, A. L., Brown, P. O., and Botstein, D.Molecular portraits of human breast tumours. Nature (Lond.), 406: 747–752, 2000.

25. Porter, D. A., Krop, I. E., Nasser, S., Sgroi, D., Kaelin, C. M., Marks, J. R., Riggins,G., and Polyak, K. A SAGE (serial analysis of gene expression) view of breast tumorprogression. Cancer Res., 61: 5697–5702, 2001.

26. Morita, R., Fujimoto, A., Hatta, N., Takehara, K., and Takata, M. Comparison of geneticprofiles between primary melanomas and their metastases reveals genetic alterations andclonal evolution during progression. J. Invest. Dermatol., 111: 919–924, 1998.

27. Alaiya, A. A., Franzen, B., Moberger, B., Silfversward, C., Linder, S., and Auer, G.Two-dimensional gel analysis of protein expression in ovarian tumors shows a lowdegree of intratumoral heterogeneity. Electrophoresis, 20: 1039–1046, 1999.

28. Peter, M., Nakagawa, J., Doree, M., Labbe, J. C., and Nigg, E. A. Identification ofmajor nucleolar proteins as candidate mitotic substrates of cdc2 kinase. Cell, 60:791–801, 1990.

29. Jiang, P. S., Chang, J. H., and Yung, B. Y. Different kinases phosphorylate nucleo-phosmin/B23 at different sites during G2 and M phases of the cell cycle. Cancer Lett.,153: 151–160, 2000.

30. Hsu, C. Y., and Yung, B. Y. Over-expression of nucleophosmin/B23 decreases thesusceptibility of human leukemia HL-60 cells to retinoic acid-induced differentiationand apoptosis. Int. J. Cancer, 88: 392–400, 2000.

31. Chou, C. C., and Yung, B. Y. Increased stability of nucleophosmin/B23 in anti-apoptotic effect of ras during serum deprivation. Mol. Pharmacol., 59: 38–45, 2001.

32. Kondo, T., Minamino, N., Nagamura-Inoue, T., Matsumoto, M., Taniguchi, T., andTanaka, N. Identification and characterization of nucleophosmin/B23/numatrin whichbinds the anti-oncogenic transcription factor IRF-1 and manifests oncogenic activity.Oncogene, 15: 1275–81, 1997.

33. Nozawa, Y., Van Belzen, N., Van der Made, A. C., Dinjens, W. N., and Bosman, F. T.Expression of nucleophosmin/B23 in normal and neoplastic colorectal mucosa.J. Pathol., 178: 48–52, 1996.

34. Skaar, T. C., Prasad, S. C., Sharareh, S., Lippman, M. E., Brunner, N., and Clarke, R.Two-dimensional gel electrophoresis analyses identify nucleophosmin as an estrogenregulated protein associated with acquired estrogen-independence in human breastcancer cells. J. Steroid Biochem. Mol. Biol., 67: 391–402, 1998.

35. Duffy, M. J. Proteases as prognostic markers in cancer. Clin. Cancer Res., 2:613–618, 1996.

36. Rochefort, H., Garcia, M., Glondu, M., Laurent, V., Liaudet, E., Rey, J. M., andRoger, P. Cathepsin D in breast cancer: mechanisms and clinical applications, a 1999overview. Clin. Chim. Acta, 291: 157–170, 2000.

37. Goldmann, T., Suter, L., Ribbert, D., and Otto, F. The expression of proteolyticenzymes at the dermal invading front of primary cutaneous melanoma predictsmetastasis. Pathol. Res. Pract., 195: 171–175, 1999.

38. Westhoff, U., Fox, C., and Otto, F. J. Quantification of cathepsin D in plasma ofpatients with malignant melanoma. Anticancer Res., 18: 3785–3788, 1998.

39. Podhajcer, O. L., Bover, L., Bravo, A. I., Ledda, M. F., Kairiyama, C., Calb, I., Guerra,L., Capony, F., and Mordoh, J. Expression of cathepsin D in primary and metastatichuman melanoma and dysplastic nevi. J. Invest. Dermatol., 104: 340–344, 1995.

40. Bartenjev, I., Rudolf, Z., Stabuc, B., Vrhovec, I., Perkovic, T., and Kansky, A.Cathepsin D expression in early cutaneous malignant melanoma. Int. J. Dermatol., 39:599–602, 2000.

41. Nakamura, H., Izumoto, Y., Kambe, H., Kuroda, T., Mori, T., Kawamura, K.,Yamamoto, H., and Kishimoto, T. Molecular cloning of complementary DNA for anovel human hepatoma-derived growth factor. Its homology with high mobilitygroup-1 protein. J. Biol. Chem., 269: 25143–25149, 1994.

42. Everett, A. D., Stoops, T., and McNamara, C. A. Nuclear targeting is required forhepatoma-derived growth factor-stimulated mitogenesis in vascular smooth musclecells. J. Biol. Chem., 276: 37564–37568, 2001.

43. Enomoto, H., Yoshida, K., Kishima, Y., Kinoshita, T., Yamamoto, M., Everett, A. D.,Miyajima, A., and Nakamura, H. Hepatoma-derived growth factor is highly expressedin developing liver and promotes fetal hepatocyte proliferation. Hepatology, 36:1519–1527, 2002.

44. Matsuyama, A., Inoue, H., Shibuta, K., Tanaka, Y., Barnard, G. F., Sugimachi, K.,and Mori, M. Hepatoma-derived growth factor is associated with reduced sensitivityto irradiation in esophageal cancer. Cancer Res., 61: 5714–5717, 2001.

6725




2003;63:6716-6725. Cancer Res Karine Bernard, Elizabeth Litman, James L. Fitzpatrick, et al. Functional Proteomic Analysis of Melanoma Progression

Updated version

http://cancerres.aacrjournals.org/content/63/20/6716

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/63/20/6716.full#ref-list-1

This article cites 40 articles, 10 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/63/20/6716.full#related-urls

This article has been cited by 10 HighWire-hosted articles. Access the articles at:

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

SubscriptionsReprints and

[email protected] at

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

Permissions

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

.http://cancerres.aacrjournals.org/content/63/20/6716To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/content/63/20/6716.full#ref-list-1

http://cancerres.aacrjournals.org/content/63/20/6716.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



functional proteomic analysis of melanoma …...[cancer research 63, 6716–6725, october 15, 2003]...

Documents