RNA Interference-mediated Silencing of the S100A10 Gene Attenuates Plasmin Generation and Invasiveness of Colo 222 Colorectal Cancer Cells*

S100A10 is a key plasminogen receptor of the extracellular cell surface that is overexpressed in many cancer cells. Typically, S100A10 is thought to be anchored to the plasma membrane via the phospholipid-binding sites of its binding partner, annexin A2. Here, using the potent and highly sequence-specific mechanism of RNA interference (RNAi), we have stably silenced the expression of the S100A10 gene in colorectal (CCL-222) cancer cells. We show that siRNA expression mediated by the pSUPER vector causes efficient, stable, and specific down-regulation of S100A10 gene expression. The siRNA-mediated down-regulation of S100A10 gene expression resulted in a major decrease in the appearance of extracellular S100A10 protein and correlated with a 45% loss of plasminogen binding, a 65% loss in cellular plasmin generation and a complete loss in plasminogen-dependent cellular invasiveness. We also observed that the CCL-222 cells do not express annexin A2 on their extracellular surface. Thus, the data show that annexin A2 is not required by S100A10 for its association with the plasma membrane, for its colocalization with uPAR, or for its binding and activation of plasminogen.

Double-stranded RNA (dsRNA) 1 -dependent post-transcriptional gene silencing, or RNA interference (RNAi), refers to the mechanism of sequence-specific, post-transcriptional gene silencing initiated by dsRNA homologous to the gene being suppressed. RNAi was originally described in Caenorhabditis elegans (1,2) and Drosophila melanogaster (3,4) as a mechanism to protect against invasion by foreign genes and has subsequently been demonstrated to be utilized by diverse eukaryotes such as insects, plants, fungi, and vertebrates (reviewed in Refs. [5][6][7][8][9]. However, in most mammalian cells dsRNA provokes a strong cytotoxic effect presumably by the activation of PKR and 2Ј-5Ј oligoadenylate polymerase (10,11). Elbashir et al. (12,13) first reported that in vitro synthesized, small interfering RNA (siRNA) (19 -23 nt), could induce RNA interference in mammalian cells and was as potent and effective as long dsRNA but did not induce global changes in gene expression. It was shown that during RNAi processing, long dsRNA was first degraded into 19 -23 nt siRNA and then recruited into an RNA-induced silencing complex to degrade corresponding mRNA. Mechanistically, dsRNAs are processed to siRNA by Dicer, a cellular ribonuclease III, which generates duplexes of about 21 nt with 3Ј-overhangs (14,15). In mammalian cells siRNA molecules are capable of specifically silencing gene expression without cytotoxic effects. Presumably, these siRNA avoid provoking the PKR response by mimicking the products of the Dicer enzyme. Thus, siRNAs have become a novel and potent alternative to other genetic tools such as antisense oligonucleotides to probe for gene function.
Typically the gene silencing produced by siRNA effects are short-lived, which severely limits its application to cellular systems. An alternative strategy uses the endogenous expression of siRNAs by various polymerase III promoter expression cassettes that allow transcription of functional siRNAs or their precursors. Agami's group reported the use of a vector system, named pSUPER that directs the synthesis of siRNA in mammalian cells (16). They demonstrated that the expression of siRNA by this vector resulted in the efficient and specific downregulation of gene expression. Most importantly, this vector was used to establish the stable repression of gene expression. S100A10, a member of the S100 family of Ca 2ϩ -binding proteins, is a dimeric protein composed of two 11-kDa subunits (reviewed in Ref. 17). The protein is cytosolic when present as a dimer. Typically, S100A10 is found in most cells bound to its annexin A2 ligand as the heterotetrameric (S100A10) 2 -(annexin A2) 2 complex, AIIt. The formation of AIIt results in the translocation of S100A10 to the plasma membrane (reviewed in Refs. 18 -22). S100A10 has been shown to regulate plasma membrane ion channels (23,24) as well as cytosolic phospholipase A2 (25). In addition to an intracellular distribution, it has also been established that the heterotetrameric form of S100A10 is present on the extracellular surface of many cells (26 -32). Extracellularly, the S100A10 subunit functions as a plasminogen receptor (33,34). The penultimate and ultimate carboxyl-terminal lysines of this subunit bind tPA and plasminogen (35) and regulate the stimulation of tPA-dependent plasminogen activation (36).
In the present report we have used a modification of the pSUPER system, the selectable pSUPER.retro.circular.stuffer (OligoEngine) to stably suppress the expression of the S100A10 gene. CCL-22 colorectal cells were transfected with the pSUPER-S100A10 or pSUPER-Control vector, and transfectants were selected with puromycin. Resistant cells were cloned, and several stable lines were established. When analyzed after 4 months, we found that pSUPER-S100A10-transfected clones showed a significant reduction in S100A10 compared with the pSUPER-Control clones. In contrast, a similar strategy utilizing full-length antisense to S100A10 (pLIN-S100A10) failed to significantly lower S100A10 levels in the colorectal cells. The loss of S100A10 from the surface of the pSUPER-S100A10-transfected CCL-222 cells resulted in decreased plasmin production and a loss in cellular invasiveness. Thus we report for the first time the use of a stable siRNA system to study the function of the S100A10 gene.

EXPERIMENTAL PROCEDURES
Materials-Human colorectal adenocarcinoma cell line CCL-222 cells were purchased from the American Type Culture Collection. The carcinoma cells were maintained according to the provider's instruction. The cells were cultured in RPMI-160 supplemented with 10% fetal bovine serum (FBS, Invitrogen). Recombinant S100A10 was expressed in Escherichia coli and purified as described previously (37). Glu-plasminogen and monoclonal anti-plasminogen antibody were purchased from American Diagnostica. Monoclonal anti-S100A10 antibody (BD Transduction Laboratories), polyclonal anti-annexin A2 (Santa Cruz Biotechnology) antibody, and goat polyclonal anti-uPAR antibody (Chemicon) were used for immunofluorescence. Rabbit anti-uPAR antibody and anti-mouse and anti-rabbit horseradish peroxidase-conjugated secondary antibodies was purchased from Santa Cruz Biotechnology. The monoclonal anti-annexin A2 antibody used in Western blots was from BD Transduction Laboratories. Anti-human ␣-tubulin monoclonal antibody was purchased from Oncogene Science.
Small Interference RNA-The mammalian expression vector, pSU-PER.retro.circular.stuffer (OligoEngine) was used for expression of siRNA in CCL-222 cells. The gene-specific insert specifies a 19-nucleotide sequence corresponding to nucleotides 199 -217 downstream of the transcription start site (gtgggcttccagagcttct) of S100A10, which is separated by a 9-nucleotide non-complementary spacer (tctcttgaa) from the reverse complement of the same 19-nucleotide sequence. This vector was referred to as pSUPER-S100A10. A control vector (pSUPER-Control) was constructed using a 19-nucleotide sequence (gcgcgctttgtaggattcg) with no significant homology to any mammalian gene sequence and therefore serves as a non-silencing control (OligoEngine). These sequences were inserted into the pSUPER.retro.circular.stuffer backbone after digestion with BglII and HindIII and transformed into BL21-A1 One Shot TM supercompetent cells (Invitrogen) according to the manufacturer's instructions. Several clones were obtained, and the vectors were amplified. To verify the insertion of the S100A10 sequence into the pSUPER backbone, the purified vector was digested with BstI (New England Biolabs). The S100A10 sequence chosen contains a restriction site for BstI; therefore, successful ligation of the sequence results in a restriction pattern distinct from that of unligated vector or vector ligated with the control sequence (data not shown).
Transfection-CCL-222 cells were plated onto 6-well plates at 200,000 cells per well. After 24 h, cells were transfected with 1 g of RNAi plasmid hybrids using LipofectAMINE 2000 reagent (Invitrogen, Rockville, MD) according to the manufacturer's instructions. 72 h after transfection, cells were processed for immunofluorescence analysis to evaluate S100A10 expression. Stable transfected cell lines were selected with 0.5 g/ml puromycin (Sigma, St. Louis, MO); clonal cell lines were selected by S100A10 protein expression by Western blotting.
Western Blotting-Total cell lysates were prepared in radioimmune precipitation assay lysis buffer (150 mM NaCl, 1% Nonidet P-40, 50 mM Tris-HCl, 0.1% SDS, 5 mM EDTA, and 20 mM NaF) supplemented with 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, and 5 g/ml aprotinin. Lysates were cleared by centrifugation at 14,000 ϫ g for 20 min at 4°C and analyzed by SDS-PAGE and transferred to a nitrocellulose membrane. After blocking with a 5% skim milk solution, the membrane was incubated with 0.25 g/ml of an anti-human S100A10 mAb, 0.25 g/ml of an anti-human Annexin II mAb, or 1 g/ml of an anti-human ␣-tubulin mAb. These mAbs were detected with 0.2 g/ml horseradish peroxidase-conjugated goat anti-mouse IgG and developed using a Super Signal detection kit (Pierce, Rockford, IL).
Immunofluorescence Microscopy-Cells were cultured on coverslips until 80 -90% confluence. Cells were then washed with ice-cold DPBS and fixed with 4% PFA (paraformaldehyde) at 4°C. Alternatively, cells were permeabilized by fixation with ice-cold 100% methanol. After blocking with 1% BSA in PBS at 4°C for 1 h, primary antibodies (monoclonal anti-S100A10 antibody and goat polyclonal anti-annexin A2 antibody) were applied to the cells (1 g/ml in PBS at 4°C) for 1 h. Fluorophore-labeled second antibody (Cy3-conjugated rabbit antimouse and Alexa-conjugated donkey anti-goat) was then applied at 4°C for 1 h. After thoroughly washing with PBS, coverslips were mounted in a solution of Prolong AntiFade (Molecular Probes) and visualized using a Zeiss Axioskop microscope or confocal microscope. To differentiate nonspecific binding of antibodies, isotype-matched, control mouse and goat antibodies were applied to cells, and incubated under the same conditions.
RT-PCR-Total RNA was extracted by using the RNeasy Mini kit (Qiagen). Purified RNA was reverse-transcribed using the One-Step RT-PCR system (Qiagen) according to the manufacturer's protocol in 20 l of final volume. Subsequently, 2 l of cDNA was PCR-amplified using platinum TaqDNA polymerase (Invitrogen) for 22 cycles using S100A10-, annexin A2-, or glyceraldehyde-3-phosphate dehydrogenase cDNA-specific primers.
Plasminogen Binding Assay-Recombinant human Glu-plasminogen (American Diagnostica) was radioiodinated as described previously (34). Plasminogen (10 M) was incubated for 3 min at room temperature with 37 MBq (1 mCi) of Na 125 I and three IODO-BEADs (Pierce) in PBS. Free Na 125 I and protein were separated using a PD-10 column (Sephadex G-25, Amersham Biosciences, Uppsala, Sweden), equilibrated, and eluted with PBS. The specific activity of the protein preparations ranged from 1000 to 2000 cpm/pmol of protein. Confluent cells in 24-well plates were rinsed with ice-cold DPBS and incubated with 50 nM radioiodinated plasminogen and 0.45 M cold plasminogen in the presence or absence of 10 mM ⑀-ACA at 4°C for 1 h. The radioactivity of the cells was determined after washing the cells with ice-cold DPBS three times.
Plasminogen Activation Assay-Both transfected and parental CCL-222 cells were seeded in 24-well culture plates at a density of 5 ϫ 10 6 cells/ml. After incubation in RPMI 1640 with 10% bovine serum albumin for 6 h, cells were rinsed two times with PBS (pH 7.4), and culture media was replaced with fresh phenol-red-and serum-free RPMI-1604 media. Purified Glu-plasminogen (American Diagnostica) was added at a final concentration of 0.5 M. Conditioned media was collected and cleared by centrifugation after 2, 4, and 6 h. The kinetics of cellmediated plasminogen activation was determined by measuring amidolytic activity of the plasmin generated from plasminogen. The reaction was conducted with the substrate H-D-norleucyl-hexahydrotyrosyl-lysinep-nitroanilide (Spectrozyme #251, American Diagnostica) at a final concentration of 100 M. The reaction was initiated by the addition of substrate to 200 l of conditioned media and was monitored at 405 nm in a PerkinElmer Life Sciences HTS 7000 Bioassay reader (Shelton, CT).
Cell Invasion Assay-A cell invasion assay was conducted with QC-M TM 96-well cell invasion assay kit (Chemicon International). This well invasion plate is based on the Boyden chamber principle. This plate contains 96 inserts; each containing an 8-m pore size polycarbonate membrane coated with a layer of ECMatrix TM . Briefly, after detaching with cell dissociation solution (Sigma), CCL-222 cells were cultured in suspension in RPMI 1640 with 10% FBS for 2 h. Then, the cells were resuspended in serum-free RPMI 1640, and 5 ϫ 10 4 cells were seeded into the extracellular matrix layer, which had been previously rehydrated at room temperature for 1-2 h. For the plasmin-dependent invasion assay, 0.2 M plasminogen was added to the cell suspensions. 150 l of RPMI 1640 media containing 10% fetal bovine serum was added to the lower chamber as chemoattractant. Cells were incubated for 24 h at 37°C in a CO 2 incubator (5% CO 2 ). Invaded cells on the bottom of the insert membrane were dissociated from the membrane by incubation with cell detachment buffer and subsequently lysed and detected by CyQuant GR dye. The fluorescence was quantified with a fluorescence plate reader using a 485/535-nm filter set.
Chamber Migration Assay-Migration was evaluated using a modified Boyden chamber assay. 8-m cell culture inserts containing polyethylene tetrephthalate (BD Biosciences) were placed within a 24-well chamber containing 0.8 ml of RPMI 1640 medium with 10% FBS. 1.5 ϫ 10 5 cells were seeded into the inserts suspended in 0.3 ml of serum-free RPMI 1640 media with 0.2 M plasminogen. After incubation for 24 h at 37°C in a CO 2 incubator (5% CO 2 ), the upper surface of the filter was scraped to remove non-migratory cells. Migrated cells were fixed with 4% PFA and stained with crystal violet. For quantification, the average number of migrating cells per field was assessed by counting 10 random fields under a light microscope (400ϫ). Data indicate the mean obtained from three separate chambers.
Miscellaneous Techniques-Protein concentrations were determined using Coomassie Brilliant Blue and BSA standards as described by Bradford (38). All reagents used were of analytical grade or better. Data were analyzed using Sigma Plot (Jandel Scientific).

RESULTS
Colo 222 Colorectal Cells Express Extracellular S100A10 but Not Annexin A2-The CCL-222 cell line is a human intestinal tumor cell line that forms liver micrometastases in nude mice. To investigate the intracellular distribution of S100A10 and annexin A2, the colorectal cells were permeabilized and simul-FIG. 1. Immunofluorescence localization of S100A10 in CCL-222 colorectal cells. Permeabilized (A) or nonpermeabilized (B) CCL-222 colorectal cells were treated with anti-S100A10 monoclonal and anti-annexin A2 polyclonal antibodies followed by Cy3-labeled (red) and Alexa-labeled (green) secondary antibodies. The nucleus was also stained with 4Ј,6-diamidino-2-phenylindole (blue). Yellow fluorescence indicates colocalization of these proteins. Staining of permeabilized cells revealed an intense colocalization of S100A10 and annexin A2 in the submembranous region (A). As shown in B, immunofluorescence staining of non-permeabilized cells reveals the presence of S100A10 in discrete patches on the cell surface, whereas annexin A2 was not detectable on the cell surface. Immunofluorescence staining was also performed at higher magnification by confocal microscopy (C and D). At this higher magnification, the intense colocalization of S100A10 and annexin A2 in the submembranous region is readily observed (C). In the non-permeabilized cells, the patch of S100A10 at the cell surface appears diffuse and non-uniform (D). Normal mouse and goat IgG (isotypematched) were used as controls to represent background fluorescence (not shown). Scale bar represents 10 m. E, detection of extracellular S100A10 by cell surface biotinylation. CCL-222 cell surface proteins were biotinylated using membrane-impermeable sulfo-NHS-biotin. Biotinylated proteins were pelleted with streptavidin-conjugated Dynabeads, loaded onto a 12% SDS-PAGE, and analyzed by Western blotting with both monoclonal S100A10 and monoclonal annexin A2 antibody.
FIG. 2. Colocalization of S100A10, uPAR, and plasminogen on the CCL-222 cell surface. Cells grown on glass coverslips were fixed with 4% Paraformaldehyde (PFA) and stained for both S100A10 and uPAR using anti-S100A10 monoclonal antibody and anti-uPAR polyclonal antibody, as described in the Fig. 1 legend. Immunofluorescence photomicrographs detailing the extracellular expression of S100A10 and uPAR are compared at low (A and D) and higher (B and C) magnification using confocal microscopy. C, a patch of p11/uPAR immunofluorescence taneously stained for both S100A10 and annexin A2. As shown in Fig. 1A, immunofluorescence microscopic analysis established the presence of both S100A10 and its ligand annexin A2 within these cells. The distribution of S100A10 and annexin A2 immunofluorescence was consistent with the majority of these proteins colocalizing at the submembranous region of the cell. Confocal microscopy also confirmed that the majority of intracellular S100A10 and annexin A2 colocalized at the plasma membrane (Fig. 1C).
Next, the immunofluorescence distribution of S100A10 and annexin A2 was examined in non-permeabilized colorectal cells (Fig. 1, B and D). We observed that S100A10 was present in discrete patches on the cell surface similar to structures observed for this protein on the extracellular surface of breast carcinoma, HT1080 fibrosarcoma, glioma, and human umbilical vein endothelial cells (32,34,39). Surprisingly, we did not observe immunofluorescence staining for annexin A2, suggesting that annexin A2 was not present on the extracellular surface. Because the anti-annexin A2 antibody easily detected intracellular annexin A2, our inability to detect extracellular annexin A2 was not due to a problem with antibody reactivity.
To confirm the absence of annexin A2 from the cell surface, cell surface proteins were biotinylated and isolated by avidin-Sepharose pull-down. The biotinylated protein fraction was analyzed by Western blotting. As shown in Fig. 1E, only S100A10 and not annexin A2 was detected in the biotinylated protein fraction. Thus, both immunofluorescence microscopy and surface biotinylation suggested that annexin A2 was not present on the surface of the colorectal cells. The absence of annexin A2 from the extracellular surface was unexpected, because typically both S100A10 and its binding partner, annexin A2, are found on the extracellular surface as the heterotetrameric complex, AIIt (26,31,32,34). The absence of annexin A2 from the extracellular surface presented the opportunity to examine the role of S100A10 in plasminogen regulation in the absence of its annexin A2 binding partner. S100A10 Colocalizes with uPAR-The conversion of the inactive zymogen, plasminogen, to the broad specificity protease, plasmin, is mediated at the cell surface by the action of the urokinase plasminogen activator, uPA (reviewed in Refs. 40 -45). Both plasminogen and uPA interact with specific cellsurface receptors: plasminogen binds to receptors such as S100A10 (31,33,35,46), cytokeratin-8 (47)(48)(49), TIP49a (50), and ␣-enolase (51-53), whereas uPA binds to its receptor, uPAR. As shown in Fig. 2A, S100A10 and uPAR share a similar distribution on the cell surface and appear to localize to one or two distinct regions of the cell surface. Confocal microscopy also confirmed the similar distribution of these proteins, although the distribution of S100A10 appeared to be more restricted than that of uPAR (Fig. 2B). At high magnification S100A10 and uPAR appear to colocalize in some regions of the cell surface (Fig. 2C). Interestingly, we also observed that the majority of plasminogen binding sites at the cell surface colocalized with uPAR (Fig. 2D). Collectively, these results suggest the presence of S100A10, uPAR, and plasminogen at a common locus on the extracellular surface.
Silencing of the S100A10 Gene by RNA Interference-To study the role of S100A10 in plasminogen regulation we used the pSUPER system to stably suppress the expression of the S100A10 gene. The pSUPER construct consists of a H1-RNA promoter cloned next to the 19-nucleotide S100A10 sequence (nucleotides 199 -217) separated by a short 9-nucleotide spacer that forms the hairpin, followed by the reverse compliment of the same nucleotide sequence. The pSUPER-Con vector was identical to the pSUPER-S100A10 vector except the 19-nucleotide sequence was derived from an irrelevant nucleotide sequence. CCL-22 colorectal cells were transfected with the pSUPER-S100A10 or pSUPER-Con, and cells were selected with puromycin and cultured under these conditions for about 1 month. As shown in Fig. 3, both immunofluorescence microscopic analysis (Fig. 3, A-D) and Western blotting (Fig. 3, E and  F) showed that the S100A10 levels are reduced in the pSUPER-S100A10 cells compared with the pSUPER-Con cells. Furthermore, the annexin A2 levels were not affected by the S100A10 knockdown, establishing the specificity of the action of the S100A10 siRNA (Fig. 3F). In previous studies we used the pLIN vector to stably transfect HT1080 fibrosarcoma cells with a full-length antisense cDNA to S100A10 (34). It was interesting to note that this antisense strategy was unsuccessful in the CCL-222 cells. Although transfection of CCL-222 cells with the pLIN-S100A10 vector resulted in stable G418-resistant cells, as shown in Fig. 3E, the S100A10 protein levels were unchanged.
Next, the pSUPER-S100A10 cells were cloned and analyzed for S100A10 levels. As shown in Fig. 4A, when analyzed after 2 months of culture we found that several pSUPER-S100A10transfected clonal cell lines maintained a significant reduction in S100A10 protein levels compared with the pSUPER-Contransfected cells. Furthermore, the protein levels of uPAR and annexin A2 were unaffected by the transfections of pSUPER-S100A10 or pSUPER-Con.
Of the three established cell lines, clone SIR6 showed the lowest total cellular levels of S100A10. Therefore, clone SIR6 was chosen for further analysis. We examined the affect of the siRNA on the S100A10 mRNA levels. As shown in Fig. 4B, S100A10 mRNA levels of SIR6 were reduced as compared with the pSUPER-Con cells or parental cells. Furthermore, the annexin A2 mRNA levels were unchanged. This indicated that the siRNA approach was specific and loss of S100A10 mRNA or protein levels did not affect the mRNA or protein levels of its intracellular binding partner, annexin A2. Second, we performed an analysis of the distribution of S100A10 by immunofluorescence microscopy. We observed that the cellular levels of S100A10 were diminished but not totally eliminated. Furthermore, when purified recombinant S100A10 was incubated with the SIR6 clone, it bound to the cell surface and demonstrated a distribution that was similar to the staining pattern observed for endogenous extracellular S100A10 (Fig. 3C).
Modulation of Plasmin Formation by S100A10 -We examined the effect of altered extracellular levels on the ability of the cells to convert plasminogen to plasmin. CCL-222 cells, like many cancer cells, constitutively secrete uPA, which functions as a plasminogen activator. Cellular plasmin formation was monitored by washing the cells with serum-free media and measuring the initial rates of plasmin formation after addition of plasminogen and a colorimetric plasmin substrate. We initially compared the activity of the CCL-222 cells that had been transfected with either pSUPER-S100A10 or pSUPER-Con and selected with puromycin but had not been subjected to clonal selection. As shown in Fig. 5A, the pSUPER-S100A10-transfected cell population that had not been cloned showed a loss of about 35% of their plasmin-generating capability. By comparison, about 65% of the plasmin-generating activity of the SIR6 cell line was lost compared with the pSUPER-Con cell line.
consisting of a 0.3-m-thick optical section, was analyzed by confocal microscopy and illustrates the colocalization of S100A10 and uPAR. D, non-permeabilized CCL-222 colorectal cells were treated with anti-plasminogen monoclonal and anti-uPAR polyclonal antibodies followed by Alexa-labeled (green) and Cy3-labeled (red) secondary antibodies. uPAR and plasminogen colocalize on the CCL-222 cell surface. Scale bar, 10 m.
Furthermore, addition of purified recombinant S100A10 to the SIR6 cell line restored plasmin production. The 65% loss in plasmin generation by the SIR6 cells was maintained irrespective of the time of incubation of cells with plasminogen (Fig. 5B).
The binding of plasminogen to cells involves the interaction of the lysine-binding kringle domains of plasminogen with the carboxyl-terminal lysines of plasminogen receptors. This binding interaction can be inhibited by lysine analogues such as ⑀-aminocaproic acid (⑀-ACA), which competes for the lysinebinding kringle domains of plasminogen. To examine if changes in the extracellular levels of S100A10 corresponded to changes FIG. 3. S100A10 gene silencing by siRNA. CCL-222 cells were transfected with the pSUPER-S100A10 or pSUPER-Con. After transfection, cells were selected with puromycin. For immunofluorescence staining, cells were permeabilized and stained for S100A10 and annexin A2 (A and B). C and D, S100A10 was incubated with the SIR6 clone overnight. After washing three times with DPBS, cells were fixed and stained for S100A10. Staining was performed for both non-permeabilized (C) and permeabilized cells (D). Relative S100A10 and annexin A2 expression levels were determined by Western blot analysis (E and F). Cells were lysed with radioimmune precipitation assay lysis buffer, separated on SDS-PAGE, and blotted with a monoclonal S100A10 and polyclonal annexin A2 antibody. Detection of ␣-tubulin was used as a loading control. R, pSUPER-S100A10 stable transfectants selected with puromycin; V R , cells transfected with pSUPER-Con and selected with puromycin; C, parental CCl-222 cells; AS, cells transfected with pLIN vector containing full-length antisense to S100A10 and selected with G418; V AS , cell transfected with pLIN vector containing no insert and selected with G418. The AIIt standard is also shown. in the plasminogen binding capacity of the CCL-222 cells, we incubated the SIR6 cell line with 125 I-labeled plasminogen and, after washing the cells with buffer, determined the levels of bound plasminogen. As shown in Fig. 5C, the SIR6 clone showed a 45% loss of specific plasminogen binding compared with the vector control cells. Plasminogen binding to all cell lines was abrogated by ⑀-ACA treatment.
Role of S100A10 in the Invasiveness of CCL-222 Cells-The invasion of cancer cells through the extracellular matrix involves both the activation of cell surface proteases as well as changes in cell motility. We used a Matrigel assay system to study the invasiveness of the colorectal cells. The cells were seeded in the upper chamber of an ECM TM invasion assay chamber, and the number of cells that traversed the Matrigelcoated membrane and appeared in the lower chamber was determined using the fluorescent CyQuant dye. When the experiment was performed in the absence of plasminogen, the invasiveness of the SIR6, vector control, and untransfected cells were similar (Fig. 6A). In contrast, we found that, in the presence of plasminogen, the invasiveness of the pSUPER-p11 cells was decreased by 38% compared with the vector control cells. Interestingly, the SIR6 cells showed a complete loss of plasminogen-dependent invasiveness.
Changes in the invasiveness of the CCL-222 cells could be due to changes in the cell-surface proteolytic activity or motility of the cells or both. The motility of the cell lines was examined by determination of their migration through the polyethylene filter in the absence of Matrigel. As shown in Fig. 6B, all the clones had similar migration rates indicating that observed changes in cellular invasiveness was not due to differences in cellular motility. DISCUSSION A variety of experimental approaches have been developed to evaluate gene function. The most effective of these approaches include DNAzymes, ribozymes, antisense, and RNA interference (RNAi) (54). Fire and co-workers (1) originally reported that injection of double-stranded RNA (dsRNA) into Caenorhabditis elegans blocked the expression of genes highly homologous in nucleotide sequence to the delivered dsRNA. However, the introduction of dsRNA into mammalian cells can invoke a strong apoptotic response due to activation of dsRNA-dependent protein kinase and RNase L by the dsRNA (55,56). This apoptotic response can be circumvented by the use of short, 21to 22-nucleotide interfering RNAs (siRNA) (12,13). Although introduction into cells of 22-nucleotide dsRNA by transfection methods leads to specific gene silencing, this knock-down is transient and, considering the half-lives of most proteins, usually is not sustained for the full period required for complete degradation of the gene product. Recently, with the development of a new vector system, pSUPER, for the stable expression of siRNA, it has become possible to generate permanent cell lines that constitutively express siRNA (16).
In the present report we have used the pSUPER system to stably silence the S100A10 gene. CCL-222 colorectal cells were transfected with the pSUPER-S100A10 or pSUPER-Convector, selected with puromycin and resistant clones cultured. We observed that clones that exhibited stable suppression of S100A10 protein expression show a loss of plasminogen binding and plasmin generation and show a decrease in plasminogendependent cellular invasiveness. However, siRNA-transfected cells demonstrated similar protein levels of annexin A2 and uPAR compared with control-transfected or untransfected cells, establishing the specificity of the siRNA approach.
Typically S100A10 and annexin A2 colocalize intracellularly. In the case of HepG2 cells and F9 teratocarcinoma cells, which do not normally express annexin A2, S100A10 is cytosolic. However, the expression of annexin A2 by these cells results in the formation of the S100A10-annexin A2 complex and the submembranous localization of the complex (57,58). Therefore, it has been proposed that annexin A2 is primarily responsible for anchoring the complex to the cytosolic face of the plasma FIG. 3-continued membrane. We have also observed that, in colorectal cells, S100A10 and annexin A2 were present intracellularly at the submembranous region of the cell. However, only S100A10 but not annexin A2 was detected on the extracellular surface of the cells (Fig. 1). This was unexpected, because S100A10 has not yet been reported to exist on the extracellular surface of the plasma membrane in the absence of annexin A2. Typically, both S100A10 and annexin A2 are found on the extracellular surface (26,31,32,34). The reported colocalization of the two proteins on the extracellular surface has supported the concept that, analogous to the intracellular distribution of these proteins, the majority of extracellular S100A10 is present in the heterotetramer form, AIIt. The phospholipid-binding sites of annexin A2 are therefore thought to anchor the S100A10-annexin A2 complex to the extracellular surface of the plasma membrane. The absence of annexin A2 from the extracellular surface of CCL-222 colorectal cells suggests that S100A10 may bind to a distinct cell surface receptor. Furthermore, the absence of annexin A2 extracellularly presented the opportunity to examine the role of S100A10 in plasminogen regulation in the absence of its common binding partner, annexin A2.
The plasminogen activators, tissue plasminogen activator (tPA) and urokinase-type plasminogen activator (uPA), activate plasminogen by cleaving the Arg 561 -Val 562 site, resulting in a conformation change and formation of the substrate-binding site. Plasminogen activation by cancer cells is usually initiated by the release of uPA. The presence of specific receptors for both uPA and plasminogen at the cell surface is responsible FIG. 4. Specificity of S100A10 gene knock-down. A, S100A10, annexin A2, and uPAR expression in several pSUPER-S100A10-transfected clonal cell lines was compared. Of the three established cell lines, clone SIR6 showed the lowest total cellular levels of S100A10. S100A10 mRNA levels were also examined by RT-PCR in clone SIR6. As shown in B, S100A10 mRNA levels of SIR6 cells were reduced as compared with the pSUPER-Con cells or parental cells. Furthermore, annexin A2 mRNA levels were unchanged in all cell lines.
FIG. 5. Regulation of plasminogen binding and activation by S100A10. Plasminogen activation was studied in parental CCL-222 cells and two transfectants with pSUPER-S100A10 or pSUPER-Con. Plasmin generation was also quantitated in the presence or absence of 0.5 M Glu-plasminogen. Conditioned media was collected from different cell lines at different times, and cell-generated plasmin was measured at 405 nm with the amidolytic plasmin substrate H-D-norleucylhexahydrotyrosyl-lysine-p-nitroanilide after 2 h. A, plasmin generation was examined for the pSUPER-S100A10 transfectants selected with puromycin but not cloned (RNAi pre-cloning) cells transfected with the pSUPER-Con vector and selected with puromycin (Vector) and the SIR6 clonal cell line (SIR6). Alternatively, SIR6 cells were preincubated with S100A10 (0.5 M), washed, and incubated with plasminogen, and plasmin activity of the conditioned media was examined (siR6ϩS100A10). Parental cells (Non-transfected) were also examined. B, time course of plasmin generation by SIR6 clonal cell line in the presence or absence of 0.5 M plasminogen. The vector control cells and parental cells were also included in the analysis. C, plasminogen binding to SIR6 cells and vector control cells was examined with 125 I-labeled plasminogen. Plasminogen binding in the presence of ⑀-ACA is included as a specificity control.
for the spatial and temporal regulation of the conversion of plasminogen to plasmin (59,60). The cell-surface receptor for uPA, uPAR, acts as a scaffold for the conversion of the zymogen, pro-uPA to catalytically active form, uPA. Subsequently, cell-bound uPA converts cell-bound plasminogen to plasmin. Binding of plasminogen to one or more of its cell surface receptors is thought to be rate-limiting for efficient activation of plasminogen by uPA (60,61). In fact, we observed that, on the surface of HT1080 fibrosarcoma cells, S100A10 regulates the ability of the cells to convert pro-uPA to uPA, most likely secondary to the loss in the plasmin-generating activity in these cells (34). Our observation that uPAR and S100A10 colocalize on the surface of the colorectal cells (Fig. 2) suggests that S100A10 does not require annexin A2 to localize to regions of the cell in which active plasmin production is occurring. Although S100A10 does not directly bind uPAR, it does bind to the cysteine protease cathepsin B (32). Therefore, it is possible that S100A10 participates in, or even initiates, a proteolytic cascade on the tumor cell surface. In support of this hypothesis, our data establish the colocalization of plasminogen, S100A10, and uPAR to discrete regions of the cellular surface (Fig. 2). It is therefore reasonable to speculate that these regions of the colorectal cells are regions actively converting plasminogen to plasmin and participating in extracellular proteolysis.
Plasminogen binds to cells with low affinity (K d ϭ 0.3-2 M) and high capacity (10 4 -10 7 binding sites per cell). The plasminogen-binding sites on cells are heterogeneous in nature, and both proteins and non-proteins such as glycosaminoglycans and gangliosides participate in plasminogen binding. However, a series of studies established the paradigm that only a small subset of cellular plasminogen receptors, those that possess a carboxyl-terminal lysine residue, participates in cell surface plasminogen activation (reviewed in Ref. 62). Candidate plasminogen receptors possessing carboxyl-terminal lysines include S100A10 (31,33,46), cytokeratin-8 (47)(48)(49), TIP49a (50), and ␣-enolase (51)(52)(53). Although the identity of the one or more plasminogen receptors that participate in plasminogen activation is at present controversial, our observation that plasminogen binding was reduced by 45% in S100A10 knock-down cells establishes S100A10 as a key plasminogen receptor of colorectal cells.
Hajjar's group (63) has suggested that annexin A2 serves as a profibrinolytic coreceptor for both plasminogen and tPA on a variety of cells (reviewed in Ref. 63). These investigators reported that tPA and plasminogen bound to annexin A2 at distinct amino-terminal ( 7 LCKLSL 12 ) and carboxyl-terminal (Lys 307 ) sites, respectively. It was also postulated that annexin A2 played a key role in cellular plasmin production by bringing tPA and plasminogen into close proximity. Central to the Hajjar hypothesis is the suggestion that the proteolytic cleavage of annexin A2 at Lys 307 -Arg 308 could generate the carboxylterminal lysine that participates in plasminogen binding. Because we observed that phospholipid-bound annexin A2 bound plasmin but did not bind tPA or plasminogen, we have postulated that annexin A2 might act as a plasmin receptor (35).
We originally reported that recombinant S100A10 stimulated plasmin formation about 46-fold (31,33,39). Because S100A10 possesses two lysines at its carboxyl terminus, we postulated that the carboxyl-terminal lysines of S100A10 played a key role in the stimulatory activity of AIIt (33). To support this hypothesis we have shown that the loss of the two carboxyl-terminal lysines of S100A10 blocks the ability of AIIt to stimulate tPA-dependent plasminogen activation (36). We have also used surface plasmon resonance to show that S100A10 binds tPA (K d ϭ 0.45 M), plasminogen (K d ϭ 1.81 M), and plasmin (K d ϭ 0.36 M). Removal of the carboxylterminal lysines from S100A10 attenuated tPA and plasminogen binding to the protein (35). More recently, we transfected HT1080 fibrosarcoma cells with the S100A10 gene in the sense or antisense orientation and established several stable cell lines with altered extracellular levels of S100A10. We demonstrated with the antisense-S100A10 cell lines that a decrease in the extracellular levels of S100A10 resulted in a significant loss (about 80%) of cellular plasmin production. Moreover, the sense-S100A10-transfected cells showed enhanced extracellular S100A10 levels and enhanced plasmin production. We also observed that the extracellular levels of annexin II remained unchanged among all the S100A10-sense or antisense transfected cell lines (34).
Using an siRNA approach we have now shown that decreased levels of S100A10 correspond to a loss of about 65% of the plasmin-generating capability of the CCL-222 colorectal cells. Furthermore, we observed a complete loss in the plasm inogen-dependent invasiveness of the siRNA-transfected colo-FIG. 6. Role of S100A10 in cellular invasiveness. A, CCL-222 parental cells (C), SIR6 (R), and pSUPER-Con transfectant (V) were incubated in the absence or presence of 0.2 M plasminogen. Cell invasiveness was examined with Matrigel-coated invasion chambers. Invasive cells were lysed and stained with CyQuant GR dye. B, cell migration was assessed with Boyden chamber inserts. Migratory cells were fixed with 4% PFA and stained with crystal violet. For quantification, the cells were counted in 10 random fields under a light microscope (400ϫ). rectal cells. Therefore, these data provide an independent verification of the importance of S100A10 in cellular plasmin production and cellular invasiveness. In addition, the observations made in the present study show for the first time that annexin A2 is not required for the association of S100A10 with the plasma membrane, for S100A10 colocalization with uPAR, or for S100A10 to bind plasminogen and stimulate its conversion to plasmin.