Processing of MIA protein during melanoma cell migration

Processing of MIA protein during melanoma cell migration

Jennifer Schmidt and Anja-Katrin Bosserhoff*

Institute of Pathology, Molecular Pathology, University of Regensburg, Regensburg, Germany

MIA (melanoma inhibitory activity) protein, identified as a small 11kDa protein highly expressed and secreted by malignant melanomacells, plays an important functional role in melanoma development,progression and tumor cell invasion. Recent data describe a directinteraction of MIA protein with cell adhesion receptors integrina4b1 and integrin a5b1 and extracellular matrix molecules. By mod-ulating integrin activity MIA protein mediates detachment of mela-noma cells from surrounding structures resulting in enhanced inva-sive and migratory potential. However, until today a detailedunderstanding of the processes of MIA function is missing. In thisstudy, we show that after binding of MIA protein to integrin a5b1,MIA protein is internalized together with this cell adhesion recep-tor at the cell rear. This mechanism enables tumor cells to migratein a defined direction as appropriate for invasion processes. Treat-ment of melanoma cells with PKC-inhibitors strongly reducedinternalization of MIA protein. Endocytosis is followed by dissocia-tion of MIA–integrin complexes. In acidic vesicles MIA protein isdegraded while integrins are recycled. Treatment of melanoma cellswith MIA inhibitory peptides almost completely blocked the MIAprotein uptake into cells. As MIA protein has a major contributionto the aggressive characteristics of malignant melanoma in partic-ular to formation of metastasis, it is important to elucidate theMIA functional mechanism in tumor cells to find novel therapeu-tic strategies in the fight against skin cancer.' 2009 UICC

Key words: melanoma; MIA; endocytosis; migration; integrin

Malignant melanoma is characterized by aggressive localgrowth and early formation of metastasis, and accounts for 75 per-cent of deaths associated with skin cancer. Previously, melanomainhibitory activity (MIA) has been identified as an 11 kDa proteinstrongly expressed and secreted by malignant melanoma cells butnot expressed in melanocytes.1 Subsequent in vitro and in vivoexperiments revealed that MIA protein plays an important func-tional role in melanoma development and cell invasion,2 henceMIA expression levels parallel closely the capability of melanomacells to form metastases in syngeneic animals.3,4 Increased MIAserum concentrations serve as a reliable clinical tumor marker todetect and monitor metastatic diseases in patients with malignantmelanomas.1,5,6

The three-dimensional structure of the protein was solved bymultidimensional nuclear magnetic resonance (NMR)7–9 and X-ray crystallography techniques.10 Corresponding data indicate thatMIA defines a novel type of secreted protein: the MIA proteinfamily, consisting of MIA and the homologous proteins OTOR,MIA-2 and TANGO (MIA-3). The MIA protein family is the firstfamily of secreted proteins comprising an SH3 domain-like fold insolution.11 Furthermore, phage display experiments and NMRspectra revealed that MIA protein interacts with peptides matchingto extracellular matrix proteins including human fibronectin typeIII repeats and laminin structures. In previous studies using farWestern blotting and co-immunoprecipitation MIA protein wasidentified to bind to the cell surface proteins integrin a4b1 andintegrin a5b1.

12 Thus, MIA protein modulates integrin activityand thereby mediates detachment of cells from extracellular ma-trix proteins, resulting in enhanced invasive and migratory poten-tial of melanoma cells.

In cell migration processes integrins, mediating cell–cell andcell–extracellular matrix contacts, undergo endocytic–exocytictransport. Adhesion receptor recycling is described as a processwhere at the cell rear integrins are internalized and subsequentlytransported within recycling vesicles to the leading edge of themigrating cell. Here, they are re-exocytosed to build new adhesion

contacts to extracellular matrix molecules.13,14 Now the questionarises, how MIA protein contributes to migration and invasion af-ter its secretion from tumor cells. This study elucidates the mecha-nism by which MIA protein promotes cell detachment and thusinfluences formation of cancer metastases. We found that extracel-lular MIA protein, directly binding to integrin a5b1, is internalizedtogether with this cell adhesion receptor at the cell rear. Thislocated uptake of MIA protein results in focal cell detachment atthe rear cell pole and allows a directed migration. We also demon-strate that after MIA–integrin endocytosis, these receptor-MIAcomplexes dissociate and MIA protein is degraded in acidicvesicles. Treatment of melanoma cells with MIA-inhibitory pepti-des results in a dramatically decrease of MIA protein internaliza-tion in a dose dependant manner. As MIA protein promotes inva-sive behavior of malignant melanoma cells, it is necessary to finda mechanistic explanation for observed MIA effects to develop anovel therapeutic strategy.

Material and methods

Cell lines and cell culture conditions

The melanoma cell line Mel Im, established from a human met-astatic bioptic sample (generous gift from Dr. Johnson, Universityof Munich, Germany) was used in all experiments. Additionally,main experiments were also conducted using human cell lines MelJu, SK Mel 28 and A375, which were all derived from metastasisof malignant melanoma. Cells were maintained in DMEM (PAALaboratories GmbH, Austria) supplemented with penicillin (400U/ml), streptomycin (50 lg/ml), L-glutamine (300 lg/ml) and10% fetal calf serum (Pan Biotech GmbH, Germany) and split in1:5 ratio every 3 days.

Protein labeling

For the conjugation of the orange fluorescing cyanine dye Cy3,0.11 mg MIA protein or 0.4 mg BSA, respectively, was dissolvedin 1 ml sodium carbonate-sodium bicarbonate buffer (pH 9.3),added to the dye vial (CyTM3 Mono-Reactive Dye Pack, Amer-sham GE Healthcare, UK) and mixed thoroughly. The reactionwas incubated at room temperature for 50 min before separationof protein from free dye using a SephadexTM G-25 M PD-10Desalting column (Amersham Pharmacia Biotech, Sweden). Dur-ing elution 2 pink bands occurred; the faster moving band repre-sents Cy3-labeled MIA protein and Cy3-labeled BSA, respec-

Additional Supporting Information may be found in the online versionof this article.Abbreviations: Arf6, ADP-ribosylation factor 6; BIMI, bisindolylmalei-

mide I; BSA, bovine serum albumin; DAPI, 40,6-diamidino-2-phenylin-dole; DMEM, Dulbecco’s modified essential medium; DIEA, N,N-diiso-propylethylamine; FITC, fluorescence-isothiocyanat; HOBt, hydroxyben-zotriazole; MBHA, 4-methylbenzhydrylamine hydrochloride; MIA,melanoma inhibitory activity; MTOC, microtubule organizing center;NMR, nuclear magnetic resonance; PBS, phosphate buffered saline; PKC,protein kinase C; Rab11, member RAS oncogene family; TBTU, O-(ben-zotriazol-1-yl)-N,N,N0,N0-tetramethyluronium tetrafluoroborate.Grant sponsor: Deutsche Forschungsgemeinschaft (DFG).*Correspondence to: Institute of Pathology, Molecular Pathology, Uni-

versity of Regensburg, Franz-Josef-Strauss-Allee 11, D-93053 Regens-burg, Germany. Fax:149-941-944-6602.E-mail: [email protected] 11 March 2009; Accepted after revision 26 March 2009DOI 10.1002/ijc.24508Published online 14 April 2009 in Wiley InterScience (www.interscience.

wiley.com).

Int. J. Cancer: 125, 1587–1594 (2009)' 2009 UICC

Publication of the International Union Against Cancer

tively. The procedure was designed to label protein to a finalmolar dye/protein ratio between 4 and 12.

The fluorescent Cy3 label does not affect binding properties ofMIA protein, as deduced from Boyden Chamber invasion experi-ments, where Mel Im cells were treated with Cy3-labeled MIAprotein and, in comparison, with unlabeled MIA protein (Sup-porting Information Fig. 1A).

Immunofluorescence assays

Melanoma cells (5 3 105), Mel Im, Mel Ju, SK Mel28 andA375, respectively, were grown in a 4-well chamber slide in 500ll DMEM and incubated with 35 ll of 4.5 lM Cy3-labeled MIAprotein or BSA, respectively, for 90 min at 37�C and 5% CO2.Afterwards, cells were washed and fixed using 4% paraformalde-hyde in 0.1 M phosphate-buffered saline (PBS) for 15 min andpermeabilized.15 After rinsing with PBS for 5 times, coverslipswere mounted on slides using Hard Set Mounting Medium withDAPI (Vectashield, H-1500) and imaged using an Axio ImagerZeiss Z1 fluorescence microscope (Axiovision Rel. 4.6.3)equipped with an Axio Cam MR camera. Images were taken using403 or 633 oil immersion lenses. For a better illustration in allpictures Cy3 staining is depicted in yellow. Conspicuous extracel-lular located yellow dots perceptible in images comprising MIACy3 staining are dye-artefacts.

For Golgi marker experiments cells were seeded and incubatedfor 24 hr before further treatment. During this time span the reorien-tation of the microtubule organizing center (MTOC), a compara-tively slow process that can take several hours after migratory stim-uli, is ensured. After fixation of cells with 4% paraformaldehyde in0.1 M PBS, permeabilizing and blocking of nonspecific bindingsites with blocking solution (1% BSA/PBS) for 1 hr at 4�C and rins-ing was performed. Cells were incubated with primary antibodymouse anti-Golgi protein [58K 9] antibody (Abcam, UK) in concen-trations of 1 lg/ml at 4�C for 2 hr. The amount of migrating cellswas determined by counting 3 times 50 cells. Cells, which show thecharacteristic MTOC staining pattern at the cell front were evaluatedas ‘‘migrating,’’ whereas nonpolarized cells that show homoge-nously distributed staining were counted as ‘‘nonmigrating.’’

To illustrate colocalization of MIA protein with integrin a5b1,cells were incubated with a 1:60 dilution of mouse anti-human integ-rin b1 [CD29] antibody (Chemicon International) or a 1:40 dilutionof mouse anti-human integrin a5 antibody (Chemicon Internatonal),respectively, at 4�C for 2 hr after Cy3-labeled MIA protein treat-ment, fixation with 4% paraformaldehyde in PBS, permeabilizationand blocking of nonspecific binding with 1% BSA/PBS. To excludenonspecific binding of target primary antibody due to Fc-binding or

other protein–protein interactions we also used a mouse IgG isotypecontrol antibody (Chemicon) (data not shown).

To generate MIA/Rab11 costaining, cells were treated withCy3-labeled MIA protein, fixed with 4% paraformaldehyde inPBS, permeabilized and nonspecific binding sites were blockedusing 1% BSA/PBS. Afterwards, cells were incubated with a 1:50dilution of mouse anti-Rab 11 antibody (BD Bioscience Pharmin-gen) at 4�C for 2 hr. After rinsing with PBS for 5 times, cells werecovered with a 1:30 dilution of the secondary antibody (FITC-con-jugated polyclonal rabbit anti-mouse immunoglobulin, DakoCyto-mation) in PBS at 4�C for 1 hr. Afterwards, cells were washedwith PBS and mounted with Hard Set Mounting Medium withDAPI (Vectashield, H-1500) or Hard Set Mounting Medium with-out DAPI (Vectashield, H-1400), respectively.

To selectively stain acidic lysosomes, Mel Im cells, grown on a4-well chamber slide, were incubated with LysoTracker redDND99 (Molecular Probes, Invitrogen) in a concentration of 60nM for 90 min at 37�C, 5% CO2. Afterwards, cells were washed,fixed using 4% paraformaldehyde in 0.1 M PBS for 15 min andpermeabilized. After rinsing with PBS for 5 times, cells were cov-ered with blocking solution (1% BSA/PBS) for 1 hr at 4�C fol-lowed by incubation with an 1:20 dilution of primary antibodyrabbit anti-MIA antibody (Biogenes, Berlin, Germany) for 2 hr at4�C. After washing with PBS cells were incubated with a 1:30dilution of the secondary antibody (FITC-conjugated swine anti-rabbit immunoglobulin, DakoCytomation). In case of simultane-ously staining acidic lysosomes, LysoTracker green DND26 (Mo-lecular Probes, Invitrogen) in a concentration of 600 nM was incu-bated together with Cy3-labeled MIA protein on Mel Im cells for90 min at 37�C and 5% CO2. Without fixation cells were washedwith PBS and mounted using Hard Set Mounting Medium withoutDAPI (Vectashield, H-1400).

MIA inhibitory peptide and PKC inhibitors

For inhibition of MIA protein uptake, Mel Im, Mel Ju, SKMel28 and A375 cells, respectively, together with Cy3-labeledMIA protein and the respective inhibitor were incubated for 90min at 37�C and 5% CO2. Inhibitors were used in several finalconcentrations. AR54 (sequence: NSLLVSFQPPRAR), a MIAbinding peptide deduced from peptide FN14, which was previ-ously identified in a phage display experiment,8 was synthesizedon solid-phase using HOBt/TBTU/DIEA and Rink Amide MBHAresin and was used at concentrations of 0.3 lM, 0.5 lM and 2 lM.Its ability to block MIA function was tested using a BoydenChamber invasion assay.2 AR54 at a final concentration of 1 lMwas able to almost completely inhibit MIA function withoutaffecting integrin activity, indicating that specific binding of

FIGURE 1 – Mel Im cells were treated with Cy3-labeled MIA protein and as a negative control with Cy3-labled BSA protein. (a) Migratingcells internalize MIA protein and show a strong Cy3-fluorescence intensity asymmetrically distributed at 1 cell pole as indicated by the whitearrows. (b) DAPI. (c) Merge. (d) BSA Cy3 negative control. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

1588 SCHMIDT AND BOSSERHOFF

AR54 to MIA protein anticipates MIA interaction to extracellularmatrix molecules and integrins (Supporting Information Fig. 1B).

As a negative control cells were also treated with scrambledpeptide AR5 (sequence: Gly-Gly-Ser-Gly-NH2) in concentrationsof 1 lM and 3 lM. In all cases Cy3-labelled MIA protein uptakewas not affected by AR5 (data not shown).

Both PKCa inhibitors 3-(N-[Dimethylamino]propyl-3-indolyl)-4-(3-indolyl)maleimide3-[1-[3-(Dimethylamino)propyl]1H-indol-3-yl]-4-(1Hindol-3-yl)1H-pyrrole-2,5dione Bisindolyl-maleimideI (BIM I) and 12-(2-Cyanoethyl)-6,7,12,13-tetrahydro-13-methyl-5-oxo-5H-indolo[2,3-a]pyrrolo[3,4-c]carbazole (G€O6976) wereused in a final concentration of 30 lM. As a control, cells weretreated with DMSO.

Results

Integrin heterodimers are known to regulate cell adhesion andmigration. The mechanism of integrin recycling by vesiculartransport contributes to cell migration by internalizing integrins atthe cell rear and thereby facilitates their detachment to surround-ing structures.14

In previous studies, it has been shown that MIA protein directlyinteracts with integrin a4b1 and integrin a5b1. This binding leadsto cell detachment by decreasing interactions between melanomacells and extracellular matrix molecules via inactivation of integ-rins. We, therefore, hypothesized that MIA protein regulatesmigratory behavior of melanoma cells by modulating integrin

FIGURE 2 – Mel Im cells simultaneously were treated with Cy3-labeled MIA protein (a) and AR54, a MIA protein binding peptide, whichwas able to inhibit MIA function. AR54 was added in different concentrations. Final concentrations of 0.3 lM (b) and 0.5 lM (c) moderatelydecrease MIA protein endocytosis whereas concentrations of 2 lM (d) almost completely inhibit MIA protein uptake. [Color figure can beviewed in the online issue, which is available at www.interscience.wiley.com.]

FIGURE 3 – For determination of the direction of migration, Cy3-labeled MIA treated Mel Im cells were stained with a Golgi marker mouse anti-Golgi protein [58K 9] antibody. The uptake of Cy3-labeled MIA containing vesicles takes place at the rear part of the cell (a) since the location ofthe MTOC-Golgi apparatus is toward the leading edge of migrating cells (b). Nonpolarized cells and thus nonmigrating cells, perceptible at the ho-mogeneous green staining, do not show the characteristic MIA fluorescence located at the cell rear. (c) DAPI. (d) Merge. I and II are representativeexamples of 2 independent experiments. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

1589PROCESSING OF MIA PROTEIN

activity. This theory is supported by the fact that MIA expressionlevels directly correlate with the ability of melanoma cells to formskin cancer metastases.3,4

As shown in Figure 1 Mel Im melanoma cells, seeded in me-dium confluence and incubated with Cy3-labeled MIA protein,show a strong Cy3-fluorescence intensity in intracellular vesiclesasymmetrically distributed at 1 cell pole. About 40–50% of seededMel Im cells are migrating; nearly all of them present this charac-teristic unidirectional MIA-Cy3 staining pattern. Cells treatedwith Cy3-labeled BSA protein as a negative control under thesame experimental conditions did not show any fluorescence sig-nal, indicating that the labeled BSA protein is not endocytosed(Fig. 1d). The same characteristic MIA protein uptake was alsofound in all other melanoma cell lines tested: Mel Ju, SK Mel 28and A375 (Supporting Information Fig. 2).

As MIA protein specifically interferes with attachment of mela-noma cells we recently performed a phage display screeningexperiment to identify potential MIA-inhibitory peptides. Thesepeptides were investigated in attachment analysis in a BoydenChamber model on their ability to affect MIA function.2 AR54,one of these peptides deduced from FN14 structure,8 was able toalmost completely inhibit MIA function at concentrations of 1 lMwithout affecting integrin function (Supporting Information Fig.1B). After treatment of Mel Im cells with AR54 the endocytosis ofCy3-labeled MIA protein was reduced in a dose dependent man-ner. AR54 at concentrations of 0.3 and 0.5 lM moderatelydecreases MIA protein endocytosis, whereas a concentration of 2lM almost completely inhibit MIA uptake as shown in Figure 2.For all other melanoma cell lines tested we observed comparableresults (Supporting Information Fig. 3).

To further elucidate the MIA protein function in migratorybehavior of melanoma cells, we determined the direction of cellmigration. Illustrated in Figure 3 are 2 independent examples (Iand II) where the observed MIA fluorescence staining patternargues for a coordinated MIA protein uptake located at 1 cell pole(Fig. 3a, I). As generally accepted, directed migration begins withcell polarization and it has been shown in previous studies that themicrotubule organizing center (MTOC) and the Golgi apparatusare reoriented toward the leading edge of cells in wound-healingmigration assays.16–18 For the determination of the migrationdirection of Mel Im cells, cells were incubated with Cy3-labeledMIA protein (Fig. 3a, I) and afterwards stained with a Golgimarker (mouse anti-Golgi protein [58K 9] antibody) (Fig. 3b, I).Fluorescence analysis shown in Figure 3 revealed the uptake ofMIA protein containing vesicles at the cell rear, confirming ourhypothesis that MIA protein is strongly involved in detachmentprocesses of migrating cells. This mechanism enables tumor cellsto migrate in a defined direction. Nonpolarized cells and thus non-migrating cells, perceptible at the homogeneous green staining(Fig. 3b, II), do not show the characteristically distributed Cy3 flu-orescence (Fig. 3a, II). Using the other melanoma cell lines, inmigrating cells the same fluorescence staining pattern appears.A375 cells, similar to Mel Im cells, show a strong migratory abil-ity compared to the other cell lines Mel Ju and SK Mel 28(Supporting Information Fig. 4).

The observation that the MIA vesicular staining appears at thecell rear and the fact that MIA protein specifically binds to integ-rin structures a4b1 and a5b1 prompted us to investigate whetherMIA protein is internalized together with integrins after binding.Therefore, Mel Im cells were treated with Cy3-labeled MIA pro-tein and stained with anti-integrin b1 [CD29] antibody. As illus-trated in Figure 4, fluorescence-labeled integrins are distributed allover the cell, whereas in close proximity to the cell membraneintegrins appear to accumulate (Fig. 4b). Interestingly, at earlystages of the endocytosis process close to the cell membrane, weobserved that these integrins are colocalized with Cy3-labeledMIA protein, in particular at the cell rear (Fig. 4c). After stainingwith an anti-integrin a5 antibody we observed a similar stainingpattern (data not shown). The same results were found for the

other melanoma cell lines investigated (Supporting InformationFig. 5). These findings are also supported by data presented in pre-vious studies. Based on far Western blotting and co-immunopreci-pitation assays12 we found a direct interaction of MIA proteinwith integrin a4b1 and integrin a5b1. Together these results sug-gest MIA protein to be internalized into the cell together withintegrins a5b1 after binding to these cell adhesion molecules andthereby blocking formation of extracellular matrix contacts.

Nowadays, it is known that integrins are either internalized byclathrin-dependent mechanisms or by nonclathrin-dependentmechanisms.19,20 In terms of the latter, integrins can enter caveo-lae followed by an internalization route that is regulated by proteinkinase Ca and dynamin.21 Since we found that MIA proteinuptake into cells that were treated with G€O6976 (Fig. 5c), block-ing PKCa and b, or BIMI (Fig. 5b), inhibiting PKCa, b and g,

FIGURE 4 – Mel Im cells were treated with Cy3-labeled MIA pro-tein. After fixation cells were stained with anti integrin b1 [CD29]antibody. (a) Cy3-labeled MIA protein is mainly located at 1 cell pole(b) Immunofluorescence-labeled integrins are distributed all over thecell and accumulate at or close to the cell membrane. (c) Merge. Asindicated in the image section by the white arrows, in regions underthe cytoplasmic membrane, we observed that integrin b1 colocalizeswith Cy3-labeled MIA protein, here depicted in red. This phenomenonwas observed especially at the cell rear.


was dramatically decreased (Fig. 5), our theory of integrin-medi-ated MIA protein uptake was further supported.

In the cytoplasm close to the nucleus, Cy3-labeled MIA proteinshows no colocalization with integrins (Fig. 4c). Thus, we concludedthat MIA–integrin complexes were dissociated after endocytosis andthat the 2 proteins now were transported in different ways. As withother cycling receptors, integrin heterodimers internalize to earlyendosomes from which they can be either returned directly to theplasma membrane or further trafficked to the perinuclear recyclingcompartment before recycling through Rab11- and/or Arf6-depend-ent mechanisms.14,22–24 In Figure 6, 2 independent examples (I andII) for Mel Im cells treated with Cy3-labeled MIA protein andstained with anti-Rab11 antibody are illustrated. The MIA proteininternalization takes place at the rear cell pole (Fig. 6a) whereas theRab11 staining, here depicted in red, is homogeneously distributedall over the cell (Fig. 6b). Since Cy3-labeled MIA protein and integ-rin transporter-protein Rab11 do not colocalize (Fig. 6d), our modelof intracellular dissociation of endocytosed MIA-integrin complexesfurther was confirmed. Under the same experimental conditions allother melanoma cell lines show comparable results: Cy3-labeledMIA protein does not colocalize with integrin transporter-proteinRab11 (Supporting Information Fig. 6).

To elucidate the fate of internalized MIA protein we treatedcells with Lysotracker red DND99, a chromophore which specifi-cally stains red acidic vesicles in the cytoplasm of cells (Fig. 7b).Lysosomes are organelles containing digestive enzymes catalyz-ing hydrolysis of macromolecules like proteins, polysaccharides,lipids and nucleic acids. The membrane surrounding lysosomesallows the enzymes to work at a pH value of 4 to 5, where theseenzymes achieve a high activity. Mel Im melanoma cells wereincubated with unlabeled MIA protein. Afterwards, MIA stainingwas performed using a rabbit anti-MIA antibody. As shown in 2independent examples I and II in Figure 7a, MIA protein distribu-tion depicted in green is similar to that of Cy3-labeled MIA pro-tein shown in previous figures: there is a targeted uptake of MIAprotein detectable at the cell rear of migrating cells. As indicatedby the white arrows, exactly in regions comprising assemblies ofacidic compartments colored in red, there was no MIA-staining

detectable, pointing to degradation of MIA protein inside lyso-somes. This phenomenon of disappearance of MIA signalsstrongly contributes to our model of dissociation of MIA proteinfrom integrins after internalization. To further confirm our hypoth-esis of MIA protein degradation, cells were also treated with Lyso-Tracker green DND26 together with Cy3-labeled MIA protein. Asdisplayed in Figure 8c, MIA protein colocalizes with acidic cellcompartments in close proximity to the nucleus. Unlike to detec-tion of MIA protein using an anti-MIA antibody shown in Figure7, this continuous Cy3-fluorescence signal is still detectable insidecytoplasmic acidic vesicles after digestion of the protein at a pHrange of pH 4 to 5. In summary, our results demonstrate that MIAprotein is internalized into the cell together with integrins and thatMIA-integrin binding is dissociated. In the next step, MIA proteinis digested in acidic vesicles while integrins are recycled.

Discussion

In this study, we analyzed the mechanism by which MIA protein,expressed and secreted by malignant melanoma cells, contributes toalteration of migratory and invasive behavior of these tumor cells. Inprevious investigations it was shown that MIA protein binds toextracellular matrix molecules including fibronectin, laminin andtenascin.2 MIA protein was also described to directly interact withthe cell adhesion molecules integrin a4b1 and integrin a5b1.

12 As aresult, matrix structures are masked by MIA protein and moreover,neoplastic melanocytes enhance their metastatic capability by specif-ically changing their attachment to surrounding extracellular matrixmolecules and basement membranes.

Experimental data described in this study demonstrate that secretedMIA protein is internalized together with integrin a5b1 after directlybinding to this adhesion receptor at the cell surface. We also demon-strate that endocytosis is followed by dissociation of MIA–integrincomplexes. Afterwards, MIA protein is degraded in acidic vesicles.

A similar endocytosis mechanism was also described for vitro-nectin, a plasma protein which was also found in the extracellularmatrix. Many functions have been characterized for vitronectin,including regulation of the activity of both thrombin and plasmin-

FIGURE 5 – Mel Im cells were incubated simultaneously with Cy3-labeled MIA and PKC inhibitors. The DMSO control shows the unidirec-tional Cy3-labeled MIA protein staining at the cell rear (a). Treatment with PKCa, b and g inhibitor BIM I (30 lM) (b) and PKCa and b inhibi-tor G€O6976 (30 lM) (c), respectively, leads to a dramatically decrease in endocytosis of Cy3-labeled MIA protein. [Color figure can be viewedin the online issue, which is available at www.interscience.wiley.com.]


FIGURE 7 – Mel Im cells were treated with unlabeled MIA protein and Lysotracker red DND99, a chromophore which specifically stains redacidic vesicles in the cytoplasm of cells. After fixation, MIA protein staining was performed using a rabbit anti-MIA antibody. (a) MIA proteindistribution depicted in green is similar to that of Cy3-labeled MIA protein: there is a targeted MIA protein uptake detectable at the cell rear ofmigrating cells. As indicated by the white arrows, exactly in regions comprising assemblies of acidic compartments colored in red (b), there wasno MIA-staining detectable. (c) DAPI. (d) Merge. I and II represent 2 independent examples.

FIGURE 6 – Mel Im cells were treated with Cy3-labeled MIA protein (a). After fixation, cells were incubated with anti-Rab11 antibody (b).After endocytosis of Cy3-labeled MIA protein it was cleaved from integrins. Thus, the integrin transporter-protein Rab11 depicted in red, media-ting integrin recycling, does not colocalize with MIA protein. (c) DAPI. (d) Merge. I and II are examples of 2 independent experiments.


ogen activator, as well as modulating the membrane attack com-plex of complement. Vitronectin comprises an Arg-Gly-Asp(RGD) sequence25 that can bind to either the avb3 or the avb5

integrin receptor.26,27 Similar to MIA protein, vitronectin alsomediates cell adhesion by this interaction. As a special feature forinternalization of vitronectin together with the cell adhesion recep-tor, the interaction of both the avb5 integrin and a species of hepa-ran sulfate proteoglycan are required. Binding to the extracellularmatrix is a prerequisite for endocytosis of vitronectin deduced bythe observation that multimeric vitronectin does not appear to bedegraded from the fluid phase. Identical to what we observed forMIA protein, receptor-mediated endocytosis is followed by subse-quent degradation of vitronectin in lysosomes. Further, it wasdemonstrated that effectors of protein kinase C, involved in signal-ing pathways between transmembrane signaling receptors, modu-late vitronectin degradation by regulating the internalization.26

The inhibition of receptor mediated endocytosis of MIA protein incells treated with protein kinase C inhibitors BIMI and G€O6976contributes to our hypothesis that MIA protein internalization maybe regulated by a similar mechanism. Unlike vitronectin, MIAprotein is bound to the cell surface receptor integrin a5b1 beforeinternalization. In previous studies, it was reported that remodel-ing of matrix structures occurs via internalization of extracellularmatrix proteins and degradation in lysosomes.28–31 It was shownthat—identical to MIA protein—turnover of extracellular matrixprotein fibronectin is processed via integrin a5b1 internalization.This endocytosis mechanism is constitutively regulated by caveo-lin-1 and can occur in presence or absence of fibronectin and fibro-nectin matrix.32 Not all fibronectin binding integrins can promotefibronectin endocytosis. Of the integrins tested, only a5b1 integrinwas shown to participate in fibronectin endocytosis. Identical toour results for MIA protein it was also demonstrated that matrixturnover of fibronectin is followed by lysosomal degradation.33

Further, the prevention of MIA protein internalization aftertreatment of cells with peptides deduced from extracellular matrixproteins and integrin structures is consistent with our proposed

mechanism. Before initial binding to integrin receptors, MIA pro-tein was captured by these peptides. Next to their canonical role inphysical adhesion of cells, interactions between cell surface mole-cules and matrix components provide pivotal contributions to abroad range of cellular processes in melanocytic cells. Thus,active detachment of melanoma cells induced by MIA proteinmay also be implicated in regulation of migration, apoptosis,secretion of proteases or matrix proteins and cell growth.34–38 It isknown that such interactions between melanocytic cells andextracellular matrix involve foremost binding of integrins to spe-cific epitopes within fibronectin and depend, to a significantextent, on activation of integrin a4b1 and integrin a5b1.

39 Detach-ment from surrounding matrix structures is a basic requirement formelanoma cells to migrate, invade and finally metastasize in a sys-temic disease. To impair formation of metastases and control ma-lignant melanoma metastases at the invasive state it is necessaryto anticipate MIA binding to integrins and extracellular matrixmolecules. Previously published data provide first evidence for areduction of tumor size after application of 2 fibronectin-deducedpeptides in a mouse melanoma model.2

In summary, our results demonstrate that MIA protein, bindingto integrins and thus promoting detachment of cells from extracel-lular matrix structures, is internalized into the cell together withthese cell adhesion receptors at the cell rear. MIA-integrin bindingdissociates and in the next step, MIA protein is digested in acidicvesicles while integrins are recycled. Prevention of MIA proteininternalization by capturing the protein by inhibitors in vivo mayprovide a novel therapeutic strategy for therapy of patients suffer-ing from malignant melanoma.

Acknowledgements

The authors thank Dr. Johnson (University of Munich, Ger-many) for providing melanoma cell lines, Mr. Alexander Riechersfor synthesis of MIA protein inhibitory peptides, Ms. AndreaSassen and Ms. Marietta Bock for technical assistance.

FIGURE 8 – Mel Im cells simultaneously were treated with Cy3-labeled MIA protein (a) and LysoTracker green DND26 staining acidic lyso-somes (b). As illustrated in the merge picture (c), MIA protein colocalizes with acidic cell compartments in the cytoplasm at the centre of thecell close to the nucleus. Colocalization is depicted in red and in the image section it is also indicated by white arrows.


References

1. Bosserhoff AK, Kaufmann M, Kaluza B, Bartke I, Zirngibl H, HeinR, Stolz W, Buettner R. Melanoma-inhibiting activity, a novel serummarker for progression of malignant melanoma. Cancer Res1997;57:3149–53.

2. Bosserhoff AK, Stoll R, Sleeman JP, Bataille F, Buettner R, HolakTA. Active detachment involves inhibition of cell-matrix contacts ofmalignant melanoma cells by secretion of melanoma inhibitory activ-ity. Lab Invest 2003;83:1583–94.

3. Guba M, Bosserhoff AK, Steinbauer M, Abels C, Anthuber M, Buett-ner R, Jauch KW. Overexpression of melanoma inhibitory activity(MIA) enhances extravasation and metastasis of A-mel 3 melanomacells in vivo. Br J Cancer 2000;83:1216–22.

4. Bosserhoff AK, Echtenacher B, Hein R, Buettner R. Functional roleof melanoma inhibitory activity in regulating invasion and metastasisof malignant melanoma cells in vivo. Melanoma Res 2001;11:417–21.

5. Dreau D, Bosserhoff AK, White RL, Buettner R, Holder WD. Mela-noma-inhibitory activity protein concentrations in blood of melanomapatients treated with immunotherapy. Oncol Res 1999;11:55–61.

6. Stahlecker J, Gauger A, Bosserhoff A, Buttner R, Ring J, Hein R.MIA as a reliable tumor marker in the serum of patients with malig-nant melanoma. Anticancer Res 2000;20:5041–4.

7. Stoll R, Renner C, Ambrosius D, Golob M, Voelter W, Buettner R,Bosserhoff AK, Holak TA. Sequence-specific 1H, 13C, and 15Nassignment of the human melanoma inhibitory activity (MIA) protein.J Biomol NMR 2000;17:87–8.

8. Stoll R, Renner C, Zweckstetter M, Bruggert M, Ambrosius D, PalmeS, Engh RA, Golob M, Breibach I, Buettner R, Voelter W, Holak TAet al. The extracellular human melanoma inhibitory activity (MIA)protein adopts an SH3 domain-like fold. EMBO J 2001;20:340–9.

9. Stoll R, Renner C, Buettner R, Voelter W, Bosserhoff AK, Holak TA.Backbone dynamics of the human MIA protein studied by (15)NNMR relaxation: implications for extended interactions of SH3domains. Protein Sci 2003;12:510–19.

10. Lougheed JC, Holton JM, Alber T, Bazan JF, Handel TM. Structureof melanoma inhibitory activity protein, a member of a recently iden-tified family of secreted proteins. Proc Natl Acad Sci USA2001;98:5515–20.

11. Stoll R, Bosserhoff A. Extracellular SH3 domain containing pro-teins—features of a new protein family. Curr Protein Pept Sci2008;9:221–6.

12. Bauer R, Humphries M, Fassler R, Winklmeier A, Craig SE, Bosserh-off AK. Regulation of integrin activity by MIA. J Biol Chem2006;281:11669–77.

13. Bretscher MS. Circulating integrins: alpha 5 beta 1, alpha 6 beta 4and Mac-1, but not alpha 3 beta 1, alpha 4 beta 1 or LFA-1. EMBO J1992;11:405–10.

14. Caswell PT, Norman JC. Integrin trafficking and the control of cellmigration. Traffic 2006;7:14–21.

15. Arndt S, Poser I, Moser M, Bosserhoff AK. Fussel-15, a novel Ski/Sno homolog protein, antagonizes BMP signaling. Mol Cell Neurosci2007;34:603–11.

16. Kupfer A, Louvard D, Singer SJ. Polarization of the Golgi apparatusand the microtubule-organizing center in cultured fibroblasts at theedge of an experimental wound. Proc Natl Acad Sci USA1982;79:2603–7.

17. Magdalena J, Millard TH, Machesky LM. Microtubule involvementin NIH 3T3 Golgi and MTOC polarity establishment. J Cell Sci2003;116:743–56.

18. Prigozhina NL, Waterman-Storer CM. Protein kinase D-mediated an-terograde membrane trafficking is required for fibroblast motility.Curr Biol 2004;14:88–98.

19. Altankov G, Grinnell F. Fibronectin receptor internalization and AP-2complex reorganization in potassium-depleted fibroblasts. Exp CellRes 1995;216:299–309.

20. Liu L, He B, Liu WM, Zhou D, Cox JV, Zhang XA. TetraspaninCD151 promotes cell migration by regulating integrin trafficking.J Biol Chem 2007;282:31631–42.

21. Nishimura T, Kaibuchi K. Numb controls integrin endocytosis fordirectional cell migration with aPKC and PAR-3. Dev Cell 2007;13:15–28.

22. Jones MC, Caswell PT, Norman JC. Endocytic recycling pathways:emerging regulators of cell migration. Curr Opin Cell Biol 2006;18:549–57.

23. Pellinen T, Ivaska J. Integrin traffic. J Cell Sci 2006;119:3723–31.24. Ramsay AG, Marshall JF, Hart IR. Integrin trafficking and its role in

cancer metastasis. Cancer Metastasis Rev 2007;26:567–78.25. Tomasini BR, Mosher DF. Vitronectin. Prog Hemostat Thromb 1991;

10:269–305.26. Panetti TS, Wilcox SA, Horzempa C, McKeown-Longo PJ. Alpha v

beta 5 integrin receptor-mediated endocytosis of vitronectin is proteinkinase C-dependent. J Biol Chem 1995;270:18593–7.

27. Pijuan-Thompson V, Gladson CL. Ligation of integrin alpha5beta1 isrequired for internalization of vitronectin by integrin alphavbeta3.J Biol Chem 1997;272:2736–43.

28. Godyna S, Liau G, Popa I, Stefansson S, Argraves WS. Identification ofthe low density lipoprotein receptor-related protein (LRP) as an endo-cytic receptor for thrombospondin-1. J Cell Biol 1995;129:1403–10.

29. Memmo LM, McKeown-Longo P. The alphavbeta5 integrin functionsas an endocytic receptor for vitronectin. J Cell Sci 1998;111:425–33.

30. Murphy-Ullrich JE, Mosher DF. Interactions of thrombospondin withendothelial cells: receptor-mediated binding and degradation. J CellBiol 1987;105:1603–11.

31. Wienke D, MacFadyen JR, Isacke CM. Identification and characteri-zation of the endocytic transmembrane glycoprotein Endo180 as anovel collagen receptor. Mol Biol Cell 2003;14:3592–604.

32. Sottile J, Chandler J. Fibronectin matrix turnover occurs through acaveolin-1-dependent process. Mol Biol Cell 2005;16:757–68.

33. Shi F, Sottile J. Caveolin-1-dependent {beta}1 integrin endocytosis is acritical regulator of fibronectin turnover. J Cell Sci 2008;121:2360–71.

34. Bates RC, Lincz LF, Burns GF. Involvement of integrins in cellsurvival. Cancer Metastasis Rev 1995;14:191–203.

35. Boudreau N, Sympson CJ, Werb Z, Bissell MJ. Suppression of ICEand apoptosis in mammary epithelial cells by extracellular matrix.Science 1995;267:891–3.

36. Clarke AS, Lotz MM, Chao C, Mercurio AM. Activation of the p21pathway of growth arrest and apoptosis by the beta 4 integrin cyto-plasmic domain. J Biol Chem 1995;270:22673–6.

37. Langholz O, Rockel D, Mauch C, Kozlowska E, Bank I, Krieg T,Eckes B. Collagen and collagenase gene expression in three-dimen-sional collagen lattices are differentially regulated by alpha 1 beta 1and alpha 2 beta 1 integrins. J Cell Biol 1995;131:1903–15.

38. Varner JA, Emerson DA, Juliano RL. Integrin alpha 5 beta 1 expres-sion negatively regulates cell growth: reversal by attachment to fibro-nectin. Mol Biol Cell 1995;6:725–40.

39. Scott G, Ryan DH, McCarthy JB. Molecular mechanisms of humanmelanocyte attachment to fibronectin. J Invest Dermatol 1992;99:787–94.


Documents

Processing of MIA protein during melanoma cell migration