A key requirement for metastatic cancer cell invasion is the penetration of extracellular matrix (ECM) barriers. This process involves the degradation of different ECM proteins and proteoglycans.1, 2 Various proteinases have been implicated in ECM degradation associated with tumour invasion and metastasis, including urokinase-type plasminogen activator, matrix metalloproteinases and cathepsins.3–5 The involvement of the latter enzymes in ECM proteolysis is perplexing, since cathepsins are normally localized in lysosomes. However, tumour cells often secrete significant amounts of these proteinases into the pericellular fluid, as first observed for cathepsin B.6 As typical for lysosomal enzymes, the N-glycan moieties of cathepsins are modified during their biosynthesis with mannose 6-phosphate (M6P) residues which permit interaction with the main lysosomal sorting receptors, the 300-kDa mannose 6-phosphate/insulin-like growth factor II receptor (M6P/IGF2R) and the 46-kDa mannose 6-phosphate receptor (MPR46).7 Evidence has been provided that excessive secretion of cathepsins by tumour cells may be due to transformation-induced changes to the M6P receptor pathway.8–10
Lysosomal sorting via M6P/IGF2R is generally far more efficient than by MPR46, demonstrating that the former is the main lysosomal targeting receptor in mammalian cells.11, 12 However, M6P/IGF2R also binds a variety of other factors that impinge on the proliferation, migration and invasiveness of tumour cells, including insulin-like growth factor II (IGF-II),13 transforming growth factor β,14 urokinase-type plasminogen activator receptor15 and plasminogen.16 Hence, it is of high relevance that the M6P/IGF2R gene is frequently inactivated in human and animal tumours.17 Loss of heterozygosity at the IGF2R locus has been shown to occur in a large number of human cancers, including squamous cell carcinomas,18, 19 often accompanied by missense mutations in the remaining allele.20, 21 In support of the notion that M6P/IGF2R is a tumour suppressor, it has been reported that mammary tumour development is impaired in transgenic mice overexpressing the receptor.22 Furthermore, M6P/IGF2R overexpression has been found to reduce the growth of cancer cells both in vitro and in vivo,23–25 in some cases simultaneously promoting cell death.26 Conversely, M6P/IGF2R downregulation by antisense or ribozyme approaches accelerates tumour cell proliferation and renders the cells more resistant to apoptotic stimuli.27, 28 However, these studies were all performed using receptor-positive cancer cell lines, with only one of them displaying a reduced M6P/IGF2R content.25 Moreover, the impact of M6P/IGF2R on tumour invasion and metastasis is only poorly understood, with controversial results reported for breast cancer cells.24, 25
We have previously shown that M6P/IGF2R-deficient SCC-VII murine squamous cell carcinoma cells secrete large amounts of matrix-degrading cathepsins. Furthermore, SCC-VII cells lack mature lysosomes and store their acid hydrolases instead in compartments reminiscent of late endosomes.10, 29, 30 In this study, we have generated SCC-VII lines stably reconstituted with human M6P/IGF2R to directly assess the impact of this receptor on the tumourigenicity and invasiveness of squamous cell carcinoma cells.
The production of the mouse monoclonal antibody MEM-238 against human M6P/IGF2R was reported previously.31 Rabbit antisera against bovine M6P/IGF2R were provided by Dr. Bernard Hoflack (Technische Universität Dresden, Germany) and Dr. Peter Lobel (Rutgers University, Piscataway, NJ). Dr. Stefan Höning (Universität Bonn, Germany) supplied a rabbit antiserum raised against the cytoplasmic tail of murine MPR46.32 Rabbit antisera against mouse cathepsin D and the cytoplasmic tail of human M6P/IGF2R were kindly provided by Dr. Regina Pohlmann (Westfälische Wilhelms-Universität Münster, Germany).33 Rabbit antiserum against recombinant mouse procathepsin L was generously donated by Dr. Ann H. Erickson (University of North Carolina, Durham, NC). The rat monoclonal antibody to mouse LAMP-1 (clone 1D4B) developed by J. Thomas August (Johns Hopkins University, Baltimore, MD) was obtained from the Developmental Studies Hybridoma Bank maintained by the Department of Biological Sciences (University of Iowa, Iowa City, IA). Monoclonal antibodies against rat GM130 and protein-disulfide isomerase were obtained from BD Biosciences (Erembodegem, Belgium) and Stressgen Bioreagents (Vancouver, Canada), respectively.
SCC-VII murine squamous cell carcinoma cells were propagated in Minimal Essential Medium supplemented with 7% heat-inactivated fetal bovine serum, 1% non-essential amino acids, 2 mM glutamine, 1 mM sodium pyruvate, 100 units/ml penicillin and 100 μg/ml streptomycin.29 HeLa human cervical carcinoma cells and NIH 3T3 murine fibroblasts (both from the American Type Culture Collection, Manassas, VA) were cultured in Dulbecco's Modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum, 2 mM glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin. All tissue culture reagents were purchased from Invitrogen (Carlsbad, CA).
Generation of SCC-VII cells expressing human M6P/IGF2R
SCC-VII cells were transfected with the vector pAHygCMV2/IGF2R using Lipofectin (Invitrogen) according to the manufacturer's instructions.30 Selection of stably transfected cells was achieved by virtue of their ability to grow in the presence of 350 μg/ml hygromycin B (Invitrogen). Drug-resistant clones were isolated after 14 days and tested for M6P/IGF2R production by immunofluorescence analysis. Two of the positive clones, clones 15 (IGF2R-1) and 24 (IGF2R-2), were selected for further studies. Mock-transfected SCC-VII clones were generated by stable transfection with the parental vector pAHygCMV2.
Confocal immunofluorescence microscopy was done as described earlier.30 Briefly, cells grown on glass coverslips were fixed with paraformaldehyde and then incubated with the following antibodies diluted in phosphate-buffered saline containing 0.1% saponin: rabbit anti-bovine M6P/IGF2R (1:100), mouse anti-human M6P/IGF2R (2 μg/ml), rabbit anti-mouse MPR46 (1:100), rabbit anti-mouse cathepsin D (1:100), rat anti-mouse LAMP-1 (1:100), mouse anti-rat GM130 (2 μg/ml) and mouse anti-rat protein-disulfide isomerase (2 μg/ml). Bound primary antibodies were detected by incubation with suitable fluorescein- or Cy3-conjugated affinity-purified secondary antibodies (Jackson ImmunoResearch, West Grove, PA). The immunostained cells were then examined using a Leica TCS SP2 confocal laser-scanning microscope equipped with Ar and He/Ne lasers. Images from the confocal system were imported into Adobe Photoshop 7.0 for colouration.
Preparation of total membrane extracts
Membrane extracts of SCC-VII and SCC-VII/IGF2R cells were prepared as outlined previously.30
Immunoblotting analysis of M6P/IGF2R was performed after separation by 7.5% SDS-PAGE. SDS-PAGE (14%) was used for the analysis of other proteins. Samples were denatured for 5 min at 65°C (M6P/IGF2R) or 95°C (other proteins) prior to electrophoresis and immunoblotting analysis as reported.29, 34 Densitometric analysis of immunoblots was done as described.30
Immobilization and biotinylation of M6P/IGF2R ligands
Hansenula holstii NRRL Y-2448 phosphomannan (a kind gift of Dr. Morey E. Slodki, US Department of Agriculture, Peoria, IL) was rehydrated in 100 ml of 1% KCl at 4°C for 16 hr. The pH of the solution was adjusted to pH 2.5 by addition of Dowex 50W-X8 (H+-form) prior to incubation at 100°C for 1 hr. After cooling to room temperature and removal of the Dowex beads, the pH was readjusted to 7.0 by addition of NH4HCO3. Barium acetate (1 g) was then dissolved in the solution, and 1 N NaOH was added to adjust the pH to 9.5. Addition of ethanol (10 ml) followed by incubation at 0°C for 16 hr resulted in precipitation of the phosphomannan core fragment, which was then dissolved in water and lyophilized. Phosphomannan-Sepharose was then prepared as follows: 3 g CNBr-activated Sepharose beads (GE Healthcare, Little Chalfont, UK) were rehydrated in 1 mM HCl as recommended by the manufacturer. Ten milliliter of a solution of phosphomannan core fragment (2.5 mg/ml) in coupling buffer (0.1 M NaHCO3, 0.5 M NaCl, pH 8.3) were mixed with the rehydrated beads prior to incubation at 4°C for 16 hr. After centrifugation (5 min at 520g), the beads were resuspended in 10 ml of a solution of 1 M ethanolamine in coupling buffer and again incubated at 4°C for 16 hr. The beads were then washed sequentially with washing buffer I (0.1 M sodium acetate, 0.5 M NaCl, pH 4.0) and washing buffer II (0.1 M Tris/HCl, 0.5 M NaCl, pH 8.0). This procedure was repeated twice. Phosphomannan-Sepharose was then stored in washing buffer II containing 0.02% NaN3.
For the preparation of avidin-Sepharose beads, 0.8 g CNBr-activated Sepharose beads were rehydrated in 1 mM HCl and then mixed with a solution of 5 mg chicken egg-white avidin (Sigma, St. Louis, MO) in 4.4 ml coupling buffer prior to incubation at 4°C for 16 hr. Blocking with ethanolamine and subsequent washing steps were performed as outlined above.
For biotinylation of BSA and Glu-plasminogen (Technoclone, Vienna, Austria), 505 μl of protein solution (3.3 mg/ml in phosphate-buffered saline) and 525 μl 0.2 M sodium borate, pH 8.5 were mixed with 20 μl 100 mM Biotin-X-NHS (Sigma) dissolved in DMSO. The mixture was stirred at room temperature for 1 hr prior to addition of 105 μl 1 M Tris/HCl, pH 8.5 to quench any remaining reactive groups. After incubation for another hour, the samples were dialysed twice against phosphate-buffered saline at 4°C.
Phosphomannan binding assays
Total cellular membranes prepared as described30 were resuspended in 100 μl binding buffer (0.15 M NaCl, 50 mM imidazole/HCl pH 7.0, 0.02% NaN3) containing proteinase inhibitors (5 μg/ml E-64, 5 μg/ml leupeptin, 2 mM phenylmethylsulfonyl fluoride) and extracted with 1% (w/v) Triton X-100 for 30 min at 0°C. Aliquots of these membrane extracts corresponding to 150 μg total protein were diluted 10-fold with binding buffer and incubated with 40 μl settled phosphomannan-Sepharose beads on an end-over-end mixer for 16 hr at 4°C. The beads were then washed 5 times with 1 ml binding buffer containing 0.1% Triton X-100. After another wash with 40 μl 5 mM glucose 6-phosphate (Sigma), specifically bound material was then eluted with 40 μl 5 mM M6P (Sigma) in binding buffer.
IGF-II and plasminogen binding assays
Total cellular membrane extracts (100 μg of total protein) were diluted 10-fold with binding buffer (0.4 M KCl, 50 mM imidazole/HCl pH 7.0, 0.02% NaN3 containing 0.1% Triton X-100). The samples were then incubated either with 1 μg biotinylated IGF-II (Gropep, Adelaide, Australia), plasminogen or BSA (control) for 16 hr at 4°C on an end-over-end mixer, prior to capture of the biotinylated proteins with 40 μl avidin-Sepharose beads for another 16 hr at 4°C. The beads were washed 5 times with binding buffer and twice with washing buffer (10 mM Tris/HCl, pH 6.8). Finally, bound proteins were eluted with SDS-PAGE sample buffer by incubation for 5 min at 65°C.
Lysosomal enzyme secretion studies
Intracellular and secreted levels of β-N-acetylhexosaminidase, cathepsin D and cathepsin L were determined as reported earlier.30
Metabolic labelling and immunoprecipitation
Metabolic labelling with [35S]methionine and immunoprecipitation analysis of cathepsins B, D and L was performed as described previously.10
SCC-VII/IGF2R cells and their parental counterparts were incubated in complete culture medium in the absence or presence of 10 mM NH4Cl for 10 hr at 37°C. Post-nuclear supernatants were obtained and fractionated by density-gradient centrifugation in 18% (v/v) Percoll (GE Healthcare) gradients (initial density: 1.055 g/ml) as described.35 The gradients were divided into 10 fractions which were then analyzed for their activity of the lysosomal marker enzyme β-N-acetylhexosaminidase.30 For immunoblot detection of other lysosomal proteins (cathepsins D and L) as well as GM130 (Golgi marker) and protein-disulfide isomerase (endoplasmic reticulum marker), fractions 1–3 (heavy fraction), 4–7 (intermediate fraction) and 8–10 (light fraction) were pooled and extracted as described.30
Cells (6 × 105) were resuspended in 10 ml culture medium and seeded into 58-cm2 dishes. After incubation for 24–72 hr, the cultures were harvested by trypsinization and their cell number determined using a Fuchs-Rosenthal chamber.
Soft agar colony formation assays
Cells grown in a monolayer culture were harvested by trypsinization and resuspended in the respective medium. Cell suspensions (3 × 103 cells) were added to 4 ml medium containing 0.32% agar and seeded in duplicates into 21-cm2 dishes. For this, culture-grade agar (Sigma) was first melted in distilled water, pre-warmed to 50°C and mixed with the respective medium and the cell suspension (both pre-warmed to 37°C). Colonies obtained after culture for 3 weeks were counted, stained with a solution of 0.1% crystal violet (Sigma) and 8.5% formaldehyde in phosphate-buffered saline for 1 hr and then photographed using a Leica MZ FLIII stereomicroscope equipped with a digital camera. The images thus obtained were analysed using the NIH Image J program for determination of colony areas.
In vivo tumour formation assays
In vivo tumour formation assays were performed using 6- to 10-week-old female severe combined immunodeficiency mice (CB-17 SCID; Harlan Winckelmann, Borchen, Germany). SCC-VII and SCC-VII/IGF2R cells were grown to confluency, trypsinized and resuspended each as single cell suspension in sterile Ringer's solution. Each mouse received a single subcutaneous injection of 2.5 × 106 cells in 100 μl Ringer's solution into one of its rear flanks. The mice were sacrificed 20 days after inoculation, and the tumours excised and weighed. All animal experiments were carried out according to the Austrian guidelines for animal care and protection.
Histochemical analysis of tumours
Excised tumours were fixed in 4% phosphate-buffered formaldehyde for 16 hr at 4°C and then embedded in paraffin. Sections (4 μm) were subjected to antigen retrieval by boiling for 20 min in 10 mM sodium citrate buffer (pH 6.0) and then incubated with mouse monoclonal antibodies to either proliferating-cell nuclear antigen (1:500; DAKO, Carpinteria, CA) or cleaved caspase-3 (1:100; BD Biosciences). After incubation with biotinylated anti-mouse antibodies, visualisation was performed using the Vectastain ABC kit employing diaminobenzidine as substrate (Vector Laboratories, Burlingame, CA) prior to counterstaining with hematoxylin and eosin.
In vitro invasion assays
In vitro invasion assays were performed in 24-well plates using Transwell polycarbonate filters (pore diameter: 8 μm; Corning Costar, Cambridge, MA). The upper side of the filters was coated with 5 μg Matrigel (BD Biosciences). Cells were harvested, resuspended in serum-free medium containing 0.1% BSA and seeded (5 × 104 cells in 200 μl) in triplicates on top of the coated filters. The lower compartment was filled with 700 μl conditioned fibroblast medium prepared as described36 or the following alternative chemoattractants diluted in serum-containing medium to a final concentration of 100 ng/ml: hepatocyte growth factor, epidermal growth factor, platelet-derived growth factor, IGF-I, or IGF-II (all from Sigma). After incubation for 24 hr at 37°C, the cells were fixed with methanol and stained in a solution of 0.1% crystal violet and 8.5% formaldehyde in phosphate-buffered saline for 30 min and then rinsed twice with distilled water. Cells remaining on the upper side of the filter were removed with a cotton swab. The filters were then photographed using a Leica MZ FLIII stereomicroscope equipped with a digital camera, and the images thus obtained analysed using Adobe Photoshop 7.0.
For invasion assays in the presence of proteinase inhibitors, stock solutions of aprotinin, pepstatin, E-64 (all from Sigma), GM6001 and GM6001-control (Merck Chemicals, Nottingham, UK) were added to the chemoattractant solution. All stock solutions were prepared in DMSO, except for aprotinin which was dissolved in phosphate-buffered saline. 0.1% DMSO was used as a control where appropriate.
Statistical analyses were performed using Student's t-test, with p < 0.05 being considered significant.
Reconstitution of functional M6P/IGF2R expression in SCC-VII cells
Stable transfection of M6P/IGF2R-deficient SCC-VII cells with human M6P/IGF2R cDNA led to the isolation of several M6P/IGF2R-positive clones. Immunofluorescence detection of M6P/IGF2R in SCC-VII/IGF2R cells revealed a perinuclear distribution reminiscent of the Golgi apparatus and Golgi-associated structures such as the trans-Golgi network. This could be verified by virtually complete colocalisation of M6P/IGF2R with the Golgi marker GM130. Furthermore, M6P/IGF2R colocalised with the other mammalian M6P receptor, MPR46. In contrast, colocalisation with lysosomal proteins (cathepsin D, LAMP-1) or the endoplasmic reticulum marker protein-disulfide isomerase was not observed (Fig. 1a). Similar results were obtained for the subcellular distribution of endogenous M6P/IGF2R in HeLa human cervical carcinoma cells (not shown).
Two SCC-VII/IGF2R clones were selected for further studies, SCC-VII/IGF2R-1 and SCC-VII/IGF2R-2. The M6P/IGF2R content of these clones was estimated by immunoblotting to be 2.1 and 4.1 pmol/mg total cell protein, respectively, using HeLa cells (1.0 pmol/mg total cell protein)37 for comparison (Fig. 1b; data not shown). Thus, the receptor content of the selected SCC-VII/IGF2R clones was within (SCC-VII/IGF2R-1) or slightly above (SCC-VII/IGF2R-2) the physiological range, as the M6P/IGF2R content of different human cell lines including squamous cell carcinoma cells was previously found to be in the range of 1.0–3.4 pmol/mg total protein.37 Consequently, all further experiments were performed either with SCC-VII/IGF2R-1 cells or with both cell lines as indicated.
M6P/IGF2R contains two high-affinity M6P-binding sites and one low-affinity M6P-binding site.7, 38 To test if human M6P/IGF2R expressed in SCC-VII cells is able to bind M6P residues, membrane extracts from SCC-VII/IGF2R cells were incubated with phosphomannan-Sepharose beads. Heterologous M6P/IGF2R bound to the beads, indicating that the recombinant protein is properly folded (Fig. 1c).
M6P/IGF2R has distinct binding sites for M6P and IGF-II. To test the capacity of recombinant human M6P/IGF2R to interact with IGF-II, membrane extracts of SCC-VII/IGF2R cells were incubated with biotinylated IGF-II prior to capture of the latter with avidin-Sepharose beads. Analysis of the captured proteins by immunoblotting demonstrated that recombinant M6P/IGF2R is capable of binding IGF-II (Fig. 1d). A similar assay was used to test recombinant M6P/IGF2R for its ability to bind plasminogen (Plg). We found that only a small fraction of the recombinant receptor associated with biotinylated Plg under our experimental conditions. Because of the long exposure times thus required, non-specific interactions of M6P/IGF2R with the control protein (biotinylated BSA) and/or the avidin-Sepharose beads became apparent. However, this signal was consistently weaker than that obtained after incubation with Plg (Fig. 1e). Similar results were obtained when membrane extracts of HeLa cells were analysed (data not shown), indicating that the affinities of endogenous and heterologous human M6P/IGF2R for Plg are comparable. In conclusion, these data demonstrate functional reconstitution of M6P/IGF2R in SCC-VII cells expressing the human receptor.
Heterologous M6P/IGF2R restores intracellular retention of lysosomal enzymes
SCC-VII/IGF2R-1 cells, which express physiological levels of M6P/IGF2R, secrete far less of the classical lysosomal marker β-N-acetylhexosaminidase (11%) than parental (60%) and mock-transfected SCC-VII cells (62%; Table I). This effect of ectopic M6P/IGF2R is not further enhanced by overexpression of the receptor, with SCC-VII/IGF2R-2 cells secreting 20% of their β-N-acetylhexosaminidase activity. To establish that the improved retention of β-N-acetylhexosaminidase by SCC-VII/IGF2R cells is indeed a direct consequence of M6P/IGF2R expression, we treated the cells with ammonium chloride (NH4Cl), a weak lysosomotropic base interfering with M6P-dependent transport of acid hydrolases to lysosomes.9, 10 NH4Cl treatment of both SCC-VII/IGF2R lines resulted in strongly increased β-N-acetylhexosaminidase secretion (SCC-VII/IGF2R-1: 54%; SCC-VII/IGF2R-2: 58%). Parental and mock-transfected SCC-VII cells showed only a small increase in β-N-acetylhexosaminidase secretion upon addition of NH4Cl. The latter is probably due to the properties of MPR46, which also transports lysosomal enzymes in an NH4Cl-sensitive manner.30
Table I. β-N-Acetylhexosaminidase Secretion of Parental and Transfected SCC-VII Cells Cultured for 24 hr in the Presence or Absence of 10 mM NH4Cl
Extracellular activity (% of total)
Data are derived from 3 to 7 experiments and expressed as mean ± SEM.
p < 0.05 (comparison with mock-transfected SCC-VII cells).
11 ± 2*
54 ± 4*
20 ± 6*
58 ± 8*
62 ± 3*
74 ± 1
60 ± 5*
73 ± 2
Ectopic expression of M6P/IGF2R was also found to improve the intracellular retention and processing of the lysosomal proteinases cathepsin B, cathepsin D and cathepsin L. Immunoblotting experiments revealed that the intracellular fraction of cathepsin D was raised from 53% in the parental cell line to 95% in SCC-VII/IGF2R-1 cells (Fig. 2a, upper panel). Similar results were obtained for SCC-VII/IGF2R-2 cells (data not shown). Furthermore, maturation of single-chain cathepsin D into its two-chain form was only observed in M6P/IGF2R-positive cells. The effect of M6P/IGF2R on cathepsin L retention appears not as strong as that observed for cathepsin D (increase from 15% in the parental line to 28% in SCC-VII/IGF2R-1 cells). This hypersecretion of cathepsin L may be attributable to the intrinsic low affinity of procathepsin L for M6P receptors.39 However, an enhanced processing of single-chain cathepsin L into the two-chain form was observed in SCC-VII/IGF2R cells (Fig. 2a, lower panel).
Biosynthesis and intracellular transport of cathepsins D and L was also assessed by pulse-chase analyses of [35S]methionine-labelled SCC-VII and SCC-VII/IGF2R cells (Fig. 2b). Considering the short incubation times, the results of these experiments are in good agreement with those obtained by immunoblotting. Secretion of newly-synthesised procathepsin D by SCC-VII/IGF2R cells (SCC-VII/IGF2R-1: 15%; SCC-VII/IGF2R-2: 16%) was less efficient than that by parental SCC-VII cells (32%; Fig. 2b, upper panel). Procathepsin L secretion was also clearly reduced by heterologous M6P/IGF2R expression (SCC-VII/IGF2R-1: 38%; SCC-VII/IGF2R-2: 27%; parental SCC-VII: 65%; Fig. 2b, middle panel). Furthermore, it was found that SCC-VII/IGF2R cells secrete far less procathepsin B (SCC-VII/IGF2R-1: 4%; SCC-VII/IGF2R-2: 15%) than their parental counterparts (42%; Fig. 2b, lower panel). Taken together, these results demonstrate that ectopic expression of M6P/IGF2R in SCC-VII cells leads to improved intracellular retention of lysosomal enzymes and their accelerated processing in lysosomal compartments.
Ectopic expression of M6P/IGF2R restores the formation of dense lysosomes
To assess the status of lysosomal biogenesis in SCC-VII/IGF2R cells, we performed Percoll density-gradient centrifugation experiments. For SCC-VII/IGF2R-1 cells, large amounts of β-N-acetylhexosaminidase (47%), cathepsin D (46%) and cathepsin L (27%) were found in the dense (lysosomal) fractions. Similar results were obtained for SCC-VII/IGF2R-2 cells (data not shown). In the case of parental SCC-VII cells, the majority of β-N-acetylhexosaminidase (93%), cathepsin D (80%) and cathepsin L (93%) was detected in the light gradient fractions, which contain endosomes and a range of other compartments such as the endoplasmic reticulum and the Golgi apparatus. However, the Golgi marker GM130 and the endoplasmic reticulum marker protein-disulfide isomerase were located exclusively in the light fractions of all gradients (Fig. 3a). In contrast to parental SCC-VII cells,30 the fraction of SCC-VII/IGF2R lysosomal enzymes residing in dense and intermediate compartments was substantially reduced by treatment with NH4Cl. Furthermore, the proteolytic maturation of cathepsin L was strongly impaired in NH4Cl-treated SCC-VII/IGF2R cells (Fig. 3b). From these results, we draw two conclusions: (i) expression of M6P/IGF2R restores the formation of dense lysosomes in SCC-VII cells; (ii) the steady-state levels of lysosomal enzymes in these newly-formed compartments are sensitive to treatment with lysosomotropic bases.
Anchorage-independent proliferation and tumour growth are reduced by reconstitution of M6P/IGF2R expression in SCC-VII cells
Anchorage-independent growth is a hallmark of cellular transformation.40 Hence, soft-agar assays were performed to test whether the M6P/IGF2R status influences the ability of SCC-VII cells to grow in an anchorage-independent manner. The median diameter of SCC-VII/IGF2R-1 colonies (0.07 ± 0.03 mm) was significantly smaller than that of parental (0.18 ± 0.04 mm) or mock-transfected (0.17 ± 0.03 mm) SCC-VII colonies. Sixty-four percent of the colonies produced by parental and mock-transfected SCC-VII cells exhibited a size larger than 0.02 mm2, whereas 91% of the SCC-VII/IGF2R-1 colonies were smaller than 0.02 mm2 (Fig. 4a). In contrast, the colony-formation efficiencies of SCC-VII/IGF2R-1 (9 ± 4%), parental (11 ± 5%) and mock-transfected (11 ± 5%) SCC-VII cells were similar, as were their proliferation rates up to 48 hr when grown in liquid culture medium (Fig. 4b). However, monolayer cultures of parental and mock-transfected SCC-VII cells reached after 72 hr significantly higher cell densities than those of SCC-VII/IGF2R-1 (Fig. 4b) and SCC-VII/IGF2R-2 cells (data not shown).
We also tested the ability of SCC-VII and SCC-VII/IGF2R cells to form tumours in immunocompromised mice. We observed a significant reduction (p < 0.05) in the mean weight of SCC-VII/IGF2R foci (SCC-VII/IGF2R-1: 0.64 ± 0.47 g; SCC-VII/IGF2R-2: 1.05 ± 0.29 g) as compared to parental SCC-VII tumours (1.95 ± 0.60 g; n = 9 for all cell lines). These data lend support to the notion that reconstitution of functional M6P/IGF2R expression interferes with SCC-VII tumour progression. Interestingly, the proliferation index (82 ± 1%) was the same for SCC-VII/IGF2R-1 and parental SCC-VII tumours, as determined by proliferating-cell nuclear antigen immunohistochemistry. Similarly, histochemical detection of cleaved caspase-3 revealed that the fractions of apoptotic cells present in SCC-VII/IGF2R (2 ± 2%) and M6P/IGF2R-negative SCC-VII tumours (2 ± 1%) were comparable.
Expression of M6P/IGF2R decreases invasion of SCC-VII cells through extracellular-matrix barriers
Parental and mock-transfected SCC-VII cells have the capacity to penetrate ECM barriers when conditioned medium of fibroblasts is used as a chemoattractant.29 Several purified growth factors including IGF-II are also able to induce SCC-VII migration. We have now found that the migratory capacity of SCC-VII/IGF2R-1 and SCC-VII/IGF2R-2 cells is less than 50% of that of parental and mock-transfected SCC-VII cells irrespective of the chemoattractant used. Importantly, the differences in invasiveness in response to IGF-II were not more pronounced than those observed with other chemoattractants (Fig. 5a; data not shown). These results indicate (i) that M6P/IGF2R is a potent anti-invasive factor, and (ii) that its anti-invasive properties are not limited to its capacity to restrict the chemotactic activity of IGF-II.
Because M6P/IGF2R has the capacity to interact with various proteinases including lysosomal cathepsins, we assessed the impact of different proteinase inhibitors on SCC-VII cell migration. Under the conditions used in these experiments, the cathepsin D inhibitor pepstatin A and the serine proteinase inhibitor aprotinin had no pronounced inhibitory effect on SCC-VII invasion. Addition of the cysteine cathepsin inhibitor E-64 caused a striking reduction (48%) of invading cells, its inhibitory effect being even stronger than that of the matrix metalloproteinase inhibitor GM6001 (36%). This points out that lysosomal cysteine proteinases such as cathepsin B and cathepsin L play a prominent role in ECM invasion by SCC-VII cells (Fig. 5b), thus indicating that M6P/IGF2R could restrict the invasiveness of squamous cell carcinoma cells through prevention of the pericellular accumulation of these enzymes.
In this study, we show that expression of human M6P/IGF2R in receptor-negative SCC-VII murine squamous cell carcinoma cells (i) improves the intracellular retention of matrix-degrading cathepsins, (ii) restores the formation of dense lysosomes, (iii) reduces anchorage-independent proliferation and tumour growth and (iv) compromises the invasive potential of the cells.
M6P/IGF2R is considered a tumour suppressor due to frequent loss of heterozygosity at the IGF2R locus. This occurs for instance in 60% of human hepatocellular carcinomas41, 42 and 30% of breast cancers,43 often coupled with missense or nonsense mutations in the remaining allele.20, 21IGF2R loss of heterozygosity has also been detected in lung, head and neck squamous cell carcinomas.18, 19 Consistent with M6P/IGF2R being a tumour suppressor, M6P/IGF2R-expressing SCC-VII cells produced much smaller tumours than their parental counterparts. Furthermore, M6P/IGF2R expression caused a significant reduction in colony size when SCC-VII cells were grown under anchorage-free conditions. These findings are in good agreement with several other studies. Using colorectal carcinoma cells, Souza et al. have been the first to demonstrate that M6P/IGF2R overexpression can impede the proliferation of tumour cells.26 For choriocarcinoma and breast tumour cells, the growth-suppressive capacity of M6P/IGF2R has been also verified in vivo.23–25 It is thought that the anti-proliferative activity of M6P/IGF2R in tumour cells is mainly due to its impact on IGF-II signalling.24, 44 Wise and Pravtcheva have demonstrated that the presence of an Igf2r transgene delays Igf2-induced mammary tumour formation in mice.22 On the other hand, M6P/IGF2R downregulation has been shown to increase the growth rate and tumourigenicity of choriocarcinoma cells,27 further highlighting the relevance of the M6P/IGF2R status for tumour progression.
The information available on the impact of M6P/IGF2R on tumour invasion and metastasis is scarce and controversial. Drastic overexpression of M6P/IGF2R in human breast cancer cells led to reduced invasiveness, which was attributed to decreased IGF-II bioavailability.24 In contrast, subnormal levels of ectopically expressed M6P/IGF2R did not affect the motility of receptor-deficient murine mammary tumour cells.25 In the case of SCC-VII cells, the migratory response of M6P/IGF2R-positive cells was consistently lower than that of their parental counterparts, even though the former displayed near physiological M6P/IGF2R levels. These results indicate that the M6P/IGF2R status affects the ability of squamous cell carcinoma cells to penetrate ECM barriers. Significantly, the anti-invasive potential of the receptor in SCC-VII cells was found to be independent of the chemoattractant used. This suggests that M6P/IGF2R ligands other than IGF-II also contribute to the invasiveness of SCC-VII cells.
M6P/IGF2R is capable of interacting with several matrix-degrading proteinases, including plasminogen and lysosomal cathepsins. Inactivation of cysteine cathepsins profoundly reduced SCC-VII invasion, whereas an inhibitor of plasmin did not impede the invasiveness of this cell line in a substantial manner. Importantly, cysteine cathepsin inhibitors were more effective in preventing SCC-VII migration than inhibitors of matrix metalloproteinases, although the latter enzymes are believed to be heavily involved in squamous cell carcinoma invasion.45, 46 Hence, our results are in good agreement with recent in vivo studies, which provided strong evidence that cysteine cathepsins participate in tumour invasion.47–49 Therefore, it is possible that M6P/IGF2R restricts tumour invasion not only by dampening the biological activity of IGF-II, but also by blocking the pericellular accumulation of matrix-degrading proteinases such as cysteine cathepsins.
Authors express their gratitude to Drs. Ann H. Erickson, Regina Pohlmann, Bernard Hoflack, Stefan Höning, Peter Lobel and Morey Slodki for reagents, and thank Dr. Melinda Abas for critical reading of the manuscript.