Cystatin E/M suppresses legumain activity and invasion of human melanoma

Cell lines were established from freshly procured samples from patients with primary (MCC5, MCC13, MCC70, MCC80A) or metastatic (MCC11, MCC35, MCC52, MCC57, MCC69B, and MCC72) malignant melanoma at Moffitt Cancer Centre (MCC; Tampa, FL, USA) using standard in vitro cell culturing techniques [31]. The cell line DK928 was established from a patient with primary melanoma at the Mitchell Cancer Institute, University of South Alabama (Mobile, AL, USA). The following cell lines were obtained from patients with malignant melanoma treated at the Surgery Branch, Division of Clinical Sciences, National Cancer Institute (National Institutes of Health, Bethesda, MD, USA); 624.38 [32], 1495 [33], and 2139 (unpublished). The uveal melanoma cell line 92.1 was generously provided by Martine Jager (Leiden University Medical Centre, The Netherlands; [34]). The breast carcinoma cell line MA11 was established from bone marrow aspirate from a patient with lobular carcinoma [35], and the melanoma cell line WM239 was obtained from Meenhard Herlyn (The Wistar Institute, Philadelphia, PA, USA). A375 and SK-MEL-28 (both from malignant melanoma), DU145 (prostate carcinoma), HCT-116 (colorectal carcinoma), MCF7 (breast carcinoma) and MCF10A (non-tumorigenic mammary epithelial) were obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA), and the human dermal fibroblasts (HDF) were purchased from ScienceCell Research Labs (San Diego, CA, USA). Most of the cells were grown in RPMI 1640 (BioWhittaker, Verviers, Belgium) with 10% FBS (Hyclone, Logan, Utah, USA), 1% Glutamax (Invitrogen, Carlsbad, CA, USA), and 1% Penicillin-streptomycin (BioWhittaker). Exceptions were DU145, HCT-116 and HDF which were grown in DMEM (BioWhittaker) with 10% FBS, 1% Glutamax, and 1% Penicillin-streptomycin, and MCF10A which was maintained in DMEM/F12 (BioWhittaker) containing 5% horse serum (Gibco Life Tech., Gaithersburg, MD, USA), 1 μg/ml insulin (Sigma, St Louis, MO, USA), 20 ng/ml EGF (BD Collaborative Research, Bedford, MD, USA), 1% Glutamax, and 1% Penicillin-streptomycin. The cells were passed approximately twice a week using trypsin/EDTA (BioWhittaker). In addition, the human keratinocyte cell line HaCaT [36] was cultured in Keratinocyte-SFM (Invitrogen Life Technologies) supplemented with 0.2 ng/ml human recombinant epidermal growth factor, 30 μg/ml bovine pituitary extract, and Penicillin-streptomycin. Further, whole cell lysates of human melanocytes (Mel) were kindly provided by Léon van Kempen (Dept. of Pathology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands).

Cells (A375 and MCC11) were seeded at 5 × 10⁵ per well in a 6-well plate approximately 24 h prior to transfection. Transfections were performed using Lipofectamine 2000 (Invitrogen) following the manufacturer's protocol. Briefly, the transfection mixture contained a total of 4 μg plasmid (pTracer empty vector or pTracer containing a full-length human cystatin E/M cDNA (pCST6; kindly provided by Daniel Keppler, College of Pharmacy, Touro University-California, Vallejo, CA, USA)) and 10 μl Lipofectamine 2000 in 500 μl DMEM without FBS or antibiotics. This mixture was added to each well containing 2 ml DMEM with 10% FBS and no antibiotics. The following day, the transfection mixture was removed and the appropriate medium for experiments was added.

Cells were washed 3 times with ice cold phosphate-buffered saline (PBS), scraped into PBS and pelleted by centrifugation at 2,500 × g for 5 min at 4°C. Pellets were resuspended in RIPA triple detergent lysis buffer (50 mM Tris (pH 8.0), 150 mM NaCl, 0.02% sodium azide, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, 1 mM PMSF, 2 μg/ml leupeptin, 2 μg/ml aprotinin, and 2 μg/ml pepstatin), and incubated on ice for 30 min. Lysates were centrifuged at 15,000 × g for 30 min. Supernatants were collected and protein concentration was determined using the BCA protein assay kit (Pierce, Rockford, IL, USA).

Cells were seeded at 5 × 10⁵ per well in 6-well plates, and the following day the medium was replaced with 1 ml fresh medium without serum. After 24 h, adherent cells were washed 3 times in PBS before adding lysis buffer containing 100 mM sodium citrate, 1 mM disodium EDTA, 1% n-octyl-β-D-glucopyranoside, pH 5.8. Cell harvesting was assisted by a cell scraper and the collected lysates were frozen at -70°C. Following three cycles of freezing and thawing, the cell lysates were centrifuged at 10,000 × g for 5 min and the supernatants were applied to enzymatic activity analyses. Protein concentrations in the lysates were measured by the procedure described by Bradford [37], and performed according to the manufacturer (Bio-Rad Laboratories, Hercules, CA, USA) in a microplate reader measuring absorbance at 595 nm. Bovine serum albumin (Pierce) was used as a standard and all measurements were performed in triplicate.

The day following transfection or seeding in a 6-well plate, the medium was replaced with 1 ml fresh medium without serum. The conditioned medium was collected 48 h later and passed through a 0.45 μm filter. To precipitate secreted proteins, 0.5 ml 40% trichloroacetic acid (TCA) was added to 0.5 ml supernatant. Samples were vortexed thoroughly and incubated on ice for 1 h. Samples were spun at 14,000 × g for 30 min and supernatant was discarded. Pellets were re-suspended in ice-cold acetone using sonication and samples were incubated overnight at -20°C. Samples were washed twice by spinning at 14,000 × g for 5 min and re-suspension in acetone. Pellets were allowed to air dry for approximately 2 min and re-suspended in RIPA lysis buffer for SDS-PAGE.

Whole cell extract (25 μg total protein), recombinant human legumain (5 ng; R&D Systems, MN, USA; 2199-CY) or pellets from TCA precipitation dissolved in RIPA lysis buffer were supplemented with Laemmli SDS-PAGE sample buffer (Invitrogen) and boiled at 100°C for 5 min. Samples were applied to 4-12% gradient gels in 2-(N-morpholino) ethanesulfonic acid (MES) buffer for SDS-PAGE (Invitrogen). Proteins were transferred to a 0.22 μm PVDF membrane (Millipore Corporation, Bedford, MA, USA) using the Xcell sure-lock mini-cell (Invitrogen) according to the manufacturers instructions. Following transfer, the membrane was blocked by incubation for one hour at room temperature in blocking buffer (TBS containing 5% non-fat dry milk and 0.1% Tween). Membranes were probed overnight at 4°C with primary antibody dissolved in blocking buffer at the following concentrations: cystatin E/M goat polyclonal (1:500; R&D Systems, AF1286), cystatin C goat polyclonal (1:1000; R&D Systems; AF1196), legumain goat polyclonal (1:1000; R&D Systems; AF2199), cathepsin B rabbit polyclonal (1:10,000; Calbiochem, San Diego, CA, USA), cathepsin L goat polyclonal (1:500; R&D Systems; AF952), α-tubulin mouse monoclonal (1:5000; Calbiochem) and β-actin mouse monoclonal (1:5000; Sigma). Membranes were washed twice for 5-10 min with washing buffer (TBS with 0.1% Tween). The membranes were subsequently probed with secondary antibody dissolved in blocking buffer at the following concentrations for one hour: horseradish peroxidase-conjugated goat anti-rabbit (1:5000), rabbit anti-goat (1:5000), or rabbit anti-mouse (1:5000) antibodies (DakoCytomation, Glostrup, Denmark). Subsequently, membranes were washed three times with washing buffer for 20 min and signals were developed with Super Signal West Dura Extended Duration Substrate (Pierce) according to the manufacturer's instructions, and visualized on either Medical X-ray films (Kodak, Rochester, NY) or the digital image station IS2000 (Kodak).

The immunobands visualized using a digital image station were quantified in the related software MI v.4.0.1 (Kodak), and those from X-ray films were scanned in a calibrated densitometer GS-800 (Bio-Rad) and quantified in the related software QuantityOne v.4.6.5 (Bio-Rad). All measured quantities were normalized using the corresponding loading control (α-tubulin) for each lane.

Serum-free conditioned media was collected after 48 h and passed through a 0.45 μm filter. The media was analyzed utilizing published ELISA procedures to determine the amount of cystatin C [38], cystatin E/M [39] and cystatin F [40] secreted by the cells. In addition, a specific ELISA was established for cystatin D using methods and materials described earlier for a mouse cystatin C ELISA [41]. Briefly, the IgG fraction of a high-titer polyclonal rabbit antiserum against recombinant human cystatin D [42] not showing cross-reactivity against cystatins A, B, C, D, E/M, F, S, SN and kininogen was used for construction of a double sandwich ELISA. Wells in polystyrene microplates (MaxiSorp; Nunc, Copenhagen, Denmark) were coated with the IgG fraction as catching antibody. Sample or isolated recombinant cystatin D [42] in dilutions from 0.2 to 100 ng/mL, was added to the wells. A portion of specific antibodies, affinity purified from the same IgG fraction of the anti-serum was biotinylated [43] and used as detecting antibody. Bound cystatin D was quantified using a streptavidin-horseradish peroxidase conjugate (Amersham Biosciences, Buckinghamshire, UK) and the color reaction with 2,2-azino-di-(3-ethyl-benzothiazolin-sulfonate) (Sigma) to detect peroxidase activity as absorbance at 405 nm.

Cells were seeded at 2 × 10⁵ per well in 6-well plates and transfected with pCST6 or empty vector (pTracer) as described above. The following day, the transfection medium was removed and cells were grown in fresh RPMI (with 10% FBS) for approximately 6 hours. This medium was removed and cells were grown in RPMI (10% FBS) supplemented with [methyl-³H] thymidine (Amersham) for 24 h. Cells were detached with trypsin and 5 × 10⁴ [³H]-labeled cells in 100 μl medium were added to each well in a 24-transwell permeable plate (8.0 μm polycarbonate membrane, 6.5 mm insert; Corning Incorporated, Corning, NY, USA) coated with Matrigel basement membrane matrix (1:60 dilution; BD Biosciences, San Jose, CA, USA). For migration studies the wells were left uncoated. The bottom wells received 600 μl medium with the same concentration of serum as the top well, either 1% or 10% FBS. Cells were removed from the upper and lower sides of the filter after 72 h incubation and [³H]-thymidine was counted in a liquid scintillation analyzer (Wallac Oy, Turku, Finland). The ratio of the counts obtained from the bottom versus total counts from both sides was used to determine migration and invasiveness.

Legumain activity was measured by cleavage of the substrate Z-Ala-Ala-Asn-NHMec (Department of Biochemistry, University of Cambridge, UK) as previously described [44, 45]. Cell lysate (20 μl) was added to black 96-well microplates (No. 3915; Costar, Corning, NY, USA). After the addition of 100 μl buffer and 50 μl substrate solution (final concentration 10 μM), a kinetic measurement based on increase in fluorescence over 10 min was performed. Temperature was kept at 30°C and all measurements were performed in triplicate. Cathepsin B activity was measured in a similar manner except for the use of the substrate Z-Arg-Arg-NHMec (20 μM; Bachem AG, Bubendorf, Switzerland) and a buffer system described elsewhere [46, 47]. Cathepsin L activity was measured using the substrate Z-Phe-Arg-NHMec (40 μM; Bachem AG) in the presence of the cathepsin B-specific inhibitor CA074 (0.25 μM; Sigma) as previously reported [46], with and without the addition of the cathepsin L-specific inhibitor Z-Phe-Tyr(t-Bu)-diazomethylketone (2.5 μM; Calbiochem).

Inhibitory activity towards legumain was measured in the culture media from all the cell lines and from the pCST6-transfected cells (A375 and MCC11). Inhibitory activity was evaluated by measuring the residual enzyme activity in a preparation of legumain after the addition of conditioned media. As a source of legumain, homogenized rat kidney was applied to a cation exchange column (Resource S, 1 ml, Amersham Biosciences) and eluted with a gradient of sodium chloride as described previously [48]. Peak legumain activity eluted at about 0.3 M NaCl and this partially purified legumain fraction was used as an enzyme source together with the substrate Z-Ala-Ala-Asn-NHMec. Legumain (20 μl) and media samples (17 μl) were added 83 μl buffer and incubated for 10 min, before 50 μl substrate solution was added (10 μM final concentration). Activity of the legumain preparation alone was used as positive control and the residual legumain activity after addition of conditioned media was measured and degree of inhibition calculated for each cell media.

Although it is known that cystatin E/M is secreted by keratinocytes and is present in the skin [49], little is known concerning its expression in melanocytes or melanoma cells. To study this, we analyzed a panel of low passage cell lines derived from primary and metastatic lesions in addition to well-established melanoma cell lines. Conditioned media from an equal number of cells was collected 48 h after changing to serum-free media. Proteins in the supernatants were TCA-precipitated and examined for cystatin E/M and C expressions using immunoblotting. Several non-melanoma controls secreted cystatin E/M, including DU145, HCT-116 and MA11, and nearly equal amounts of unglycosylated (14 kD) and glycosylated (17 kD) forms were found (Fig. 1A, upper panels). Notably, in control cells derived from the skin, no cystatin E/M was secreted from human dermal fibroblasts (HDF), whereas HaCaT keratinocytes predominately secreted the non-glycosylated 14 kD cystatin E/M form. In contrast, the protease inhibitor was undetectable in media from all established melanoma cell lines examined, whereas strong immunobands were detected in two of the five primary melanoma lines (MCC13 and MCC70), and in two of the six lines derived from patients with metastatic disease (MCC57 and MCC72). Among the four melanoma lines secreting cystatin E/M, the glycosylated form (17 kD) was predominant compared to the non-glycosylated form (14 kD) in media from all but one cell line (MCC70).

Bild vergrößern

Another cysteine protease inhibitor, cystatin C, was detected as one band (15 kD) in media from most of the cell lines examined (Fig. 1A, lower panels). However, in the melanoma lines where cystatin E/M was secreted, cystatin C was generally absent (MCC13, MCC57 and MCC72) or expressed at a very low level (MCC70). None of the cystatins were observed in the primary uveal melanoma cell line 92.1.

ELISA was used to quantify cystatin E/M and C in the conditioned media (Table 1). Among all the melanoma cell lines studied, cystatin E/M was only detected in the primary line MCC70 (0.28 ng/ml), whereas several of the non-melanoma lines displayed detectable levels (0.05-0.55 ng/ml; Table 1). Linear regression analyses showed a positive correlation between cystatin E/M levels measured by ELISA versus both the non-glycosylated (p < 0.0001) and the glycosylated (p = 0.0014) form of cystatin E/M on immunoblots after band intensities were scored 0-3 (exact quantification was difficult due to lack of representative loading controls for media samples). In addition, cystatin C was detected in media from all the cell lines analyzed except MCC13 (Table 1), with a positive correlation (p < 0.0001) between cystatin C as measured by ELISA and intensities of the immunobands (scored 0-3). Of other cystatins, cystatin F was only detected in media from MCC35 cells (5.6 ng/ml), whereas no detectable levels of cystatin D were observed in media from any of the cell lines.

Furthermore, the cellular levels of both cystatin E/M and C have been investigated by immunoblotting of cell lysates from all the cell lines included. No immunobands were obtained indicating that the cellular levels of both cystatins were below the level of detection (data not shown).

The conditioned media from all cell lines was analyzed by fluorometry for their ability to inhibit the activity of a partially purified legumain fraction from rat kidney (Fig. 1B). Legumain inhibition was measured as residual legumain activity after addition of conditioned media to a fixed amount of purified legumain. Legumain activity in the enzyme preparation without addition of media was used as control (100% activity). Of the melanoma lines, MCC13, MCC57 and MCC72 showed the lowest residual legumain activity and thus highest inhibitory capacity. Cystatin E/M levels as detected by immunoblotting were scored (0-3) and a correlation analyzes performed, showing a statistically significant positive correlation in the whole cell panel (p = 0.0012) when comparing the level of cystatin E/M and the ability to inhibit rat legumain by the same cell media sample. In addition, inhibition of papain activity was measured similarly by cleavage of the substrate Z-Phe-Arg-NHMec with or without adding culture media from all cell lines studied (data not shown). Linear regression analyses showed that inhibition of papain correlated statistically significant with the cystatin C levels measured by ELISA (p = 0.0256), but not with the level of cystatin E/M.

Although the activities of cathepsin B and L in melanoma and other tumor types are well documented, the expression and activity of legumain in melanoma is relatively unknown. To determine expression of legumain, cathepsin B and L, whole cell lysates were prepared from the same panel of cell lines as used for cystatin E/M analysis (Fig. 1) following growth in ordinary media for 24 h. Immunoblotting was performed and two molecular bands for legumain were observed in most of the cell lines examined (Fig. 2, upper panels). The highest molecular weight form migrated at approximately 52 kD and represents prolegumain, while the smaller band of approximately 36 kD represents the active form. Three cell lines had only weak expression of the 36 kD band (MCC13, MCC57 and MCC72). Interestingly, these were the same cell lines that predominantly secreted the glycosylated form of cystatin E/M (17 kD) and showed the highest inhibitory activity against exogenously added legumain (Fig. 1).

Bild vergrößern

Both cathepsin B and L were detected in most of the cell lines, except those predominantly secreting the glycosylated form of cystatin E/M (MCC13, MCC57 and MCC72; Fig. 2, middle panels). For cathepsin B, most of the positive melanoma lines expressed the 25 kD molecular weight form, whereas a few cell lines also showed lower levels of the 30 kD form. One cell line (624.38) expressed only low levels of the 30 kD form. Cathepsin L was generally expressed as three forms with molecular weights of 38, 30 and 25 kD, respectively. Interestingly, those melanoma cell lines from patients showing a strong expression of the 36 kD active legumain form, seemed to have a stronger expression of the 25 kD active cathepsin L form compared to the 30 kD form, although not statistically significant. This relationship also seemed to be expressed in HaCaT keratinocytes but not in the melanocyte (Mel) control (Fig. 2).

In addition, legumain, cathepsin B and L activities in the cell lysates were measured by their ability to cleave peptide substrates (Fig. 3). Linear regression analysis of the whole group of cell lines revealed a positive correlation between legumain activity (Fig. 3A) and expression of the putative active 36 kD form (Fig. 2) and between cathepsin B activity (Fig. 3B) and expression of the 25 kD form (Fig. 2) (p = 0.0141 and 0.0351, respectively), as quantified using densitometric scanning of the immunoblots. In contrast, there were no statistical correlation between immunobands (Fig. 2) and activity (Fig. 3C) of cathepsin L. Two cell lines (HDF and 2139) showed particularly high cathepsin L activity (Fig. 3C), also reflected in strong immunobands (Fig. 2). Furthermore, legumain activity was analyzed in conditioned cell media but no active secreted legumain was detected.

Bild vergrößern

To further investigate the relationship between cystatin E/M and legumain, two cell lines with no detectable levels of secreted cystatin E/M (A375 and MCC11; Fig. 1A; Table 1) and low inhibitory activity against legumain in the culture media (shown as high residual legumain activity, Fig. 1B) were utilized. A375 is a commonly used melanoma cell line, with intermediate expression of the 36 kD form of legumain (Fig. 2). MCC11 is a low-passage melanoma line established from a patient's lymph node metastases showing a high level of the 36 kD form of legumain (Fig. 2). These cell lines were transfected with the cystatin E/M expression plasmid pCST6 and the following day the transfection media was changed to serum-free media. Conditioned media and cell lysates were collected 24 h thereafter, and cystatin E/M expression was analyzed by immunoblotting. Whereas no immuno-reactive bands were found in vector-transfected cells, substantial levels of both cystatin E/M forms (14 and 17 kD) were observed both in conditioned media (Fig. 4A, upper panels) and in lysates (Fig. 4A, middle panels) from pCST6-transfected cells. No active legumain was found in conditioned media. The conditioned media was also examined for the ability to inhibit legumain activity using the partially purified legumain fraction from rat kidney as enzyme source (Fig. 4B). Media from both pCST6-transfected cell lines (closed bars) significantly inhibited legumain activity compared to empty vector-transfected controls (open bars). Media from transfected A375 cells almost completely inhibited (97%; p < 0.0001) the activity of the rat legumain fraction, whereas media from transfected MCC11 inhibited the enzyme activity by 22% (p = 0.002).

Bild vergrößern

Legumain activity in the corresponding cell lysates from the pCST6-transfections was also measured, and the activity in both A375 and MCC11 was significantly decreased compared to controls (p < 0.0001 and p = 0.0012, respectively; Fig. 4C). A375 cells had a lower basal level of legumain activity than MCC11, which is reflected in the empty vector-transfected cells (control; Fig. 4C, open bars). Over-expression of cystatin E/M decreased the intracellular legumain activity by 62% and 25% in A375 and MCC11, respectively (closed bars). For comparison, there were no differences between the activities of cathepsin B in the same lysates when comparing control versus pCST6-transfected cells (not shown). Further, immunoblotting and densitometric scanning of the 36 kD legumain band in lysates from pCST6-transfected cells reflected the down-regulation of enzyme activity (not shown).

Although it has been reported that cystatin E/M is capable of suppressing invasion of breast cancer [12, 13], its ability to inhibit melanoma progression has not been investigated. To determine the contribution of cystatin E/M to the suppression of invasion, the pCST6-transfected cell lines A375 and MCC11 were examined for their ability to migrate through a Matrigel-coated filter of a transwell plate. Invasion was significantly decreased by 39% in A375 (p < 0.01) and 41% (p < 0.05) in MCC11 following transfection with pCST6 (Fig. 5). In contrast, there was no change in the rate of growth or motility when transwell plates without Matrigel were utilized (data not shown).

Bild vergrößern