SHPS-1: A Budding Molecule against Cancer Dissemination¹

The cancer dissemination, which is a versatility of tumor cells, remains the most important cause of death for cancer patients. Despite several trials attempted to control the dissemination at a final stage of cancer by chemotherapy and hyperthermia, no significant prolongation of survival has been found. The progression seems to be a complex process involving cell attachment, invasion, and cell growth. However, the molecular mechanisms of the dissemination of cancer cells remain to be elucidated, and the genetic events that induce or suppress the dissemination process are incompletely understood. Accordingly, identification of the regulatory molecules in the process of dissemination is one of the major goals of cancer research.

SHPS-1 (also known as signal-regulatory protein α) is a transmembrane glycoprotein possessing three immunoglobulin-like domains in the EC⁴ as well as four potential tyrosine phosphorylation sites and Src homology 2 domain-binding sites in the cytoplasmic region (1, 2, 3). It is indicated that SHPS-1 may be a direct substrate for activated protein tyrosine kinases and for nontransmembrane protein tyrosine phosphatases to regulate some physiological and pathological signals potentially involved in cell-cell interactions (4, 5, 6). In the present report, we will show that the SHPS-1 is also involved in the process of anchorage-independent cell growth and cancer dissemination.

COS7 cells, BALB/c3T3 cells, v-Src transformed BALB/c3T3 (B-SR) cells, and the derived cells were cultured in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum and antibiotics.

Full-length wild-type human SHPS-1 cDNA was inserted into the EcoRI and XhoI sites of pcDNA3 or pHis-cDNA3 (Invitrogen). For introduction of SHPS-1, COS7 cells were transiently transfected, and BALB/c3T3 cells and B-SR cells were stably transfected with the SHPS-1 cDNAs.

Peroxidase-labeled anti-phosphotyrosine (PY20H; Transduction Laboratories), anti-SHP-2 (Transduction Laboratories), anti-Src (Santa Cruz Biotechnology), anti-EAK2 (Santa Cruz Biotechnology), anti-phospho-MAPK (New England Biolabs), anti-Myc (Bahco Berlzley), anti-FN (Santa Cruz Biotechnology), and anticaveolin (Transduction Laboratories) were purchased. Anti-SHPS-1 polyclonal antibody was generated as described previously. The other antibodies used in this experiment were purchased from Transduction Laboratories.

To generate the GST fusion protein, the EC of SHPS-1 was prepared by reverse transcription-PCR. The primers used for the PCR were 5′-gcgggatccaactgaaggtgactcaggctg-3′ and 5′-cgaattcccgcacggtatggttttcggt-3′. The resulting fragments were ligated into pGEX-5 × 1 using BamH1 and EcoR1 site. The GST fusion protein was purified from Escherichia coli lysates by affinity chromatography using glutathione beads and used for the pull-down assay. Briefly, the GST fusion protein and control GST were incubated with lysates in TNE buffer conditioned medium (50 mm Tris [pH 7.4], 1% Triton X-100, 1 mm EDTA). After incubation for 2 h, the beads were then washed five times with lysis buffer (PBS, 1% Triton X-100, and 100 mm EDTA) and resuspended in SDS containing sample buffer. Gel electrophoresis and immunoblot analysis with indicated antibodies, and an enhanced chemiluminescence detection were performed.

Trypsinized cells (1 × 10⁶) were plated onto 100-mm culture dishes, which had been coated with 8 ml of 0.8% hard agarose onto the dish. As culture controls, cells were plated onto dishes and cultured under adhesion condition. After incubation, cells were collected from hard agar coated dishes with pipetting and used for Western blotting or flow cytometric analysis.

Cells were incubated on tissue culture or on hard agar-coated dishes for 24 h, collected, and fixed with ice-cold 70% ethanol. The cells were then stained in PBS containing 50 μg/ml of propidium iodide and 100 units/ml RNase. Flow cytometry was performed using an EPIC/XL cell analyzer (Coulter).

Cells were assayed for their motility by a computer-assisted modification of the phagokinetic assays with gold colloid-coated glass plates described previously. Briefly, cells (2 × 10³ cells/3.5-cm plate) were seeded on colloidal gold particle-coated glass coverslips and incubated for 24 h. After fixation with 4% paraformaldehyde, the coverslips were mounted onto the glass microscope slides, and photographs were then taken by a computer-assisted digital camera (model HS-300; Olympus, Tokyo, Japan) connected to a microscope. The areas of particle swept where cells moved around during incubation were measured by NIH image (version 1.62) and statistically analyzed by Stat View (version 4.51). For the assay, 12 fields/sample were randomly selected and 5–6 cells/field were examined.

Cells (1 × 10⁶ cells/6-cm plate) were incubated in serum-free DMEM at 37°C for 18 h, and the conditioned medium was collected. After clarification by centrifugation (10 min at 1000 rpm), the medium was diluted in 4 × nonreducing sample buffer [0.27 m Tris (pH 6.8), 8% SDS, 40% glycerol, and 0.04 mg/ml bromphenol blue] and electrophoresed in 10% SDS-PAGE containing 0.1% (w/v) gelatin for MMP detection. Gels were washed repeatedly with 2.5% Triton X-100 for 30 min at room temperature and then incubated with substrate buffer [50 mm Tris (pH 7.4), 0.02% NaN₃, and 10 mm CaCl₂ (MMP detection)] for 16 h at 37°C. The gels were then stained with Coomassie Brilliant Blue and destained until white zones on a dark background appeared.

Male BALB/c nude mice were obtained at 5 weeks of age. The indicated cells were collected by trypsinization and were washed with HBSS. To produce experimental tumor dissemination, 0.5 × 10⁶ cells were injected into the abdominal cavity of the athymic nude mice. After ∼2 weeks, the mice were killed and analyzed.

Standard methods of cell culture, plasmids construction, Western blotting, immunoprecipitation, immunofluorescence study, TUNEL staining, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay, and soft agar colony formation assay were used as described elsewhere (7).

We found that expression of SHPS-1 was markedly decreased in the cells transformed with oncogene products such as v-Src, k-Ras, v-Fps, and papillomavirus large T compared with a parental F2408 rat fibroblast cell line, suggesting that a functional oncogenic cascade in malignant tumors could have overridden the biochemical consequences of expressed SHPS-1 (Fig. 1,a). In the transfectants that caused reduced expression of SHPS-1, anchorage-independent cell growth was obvious (Fig. 1,b), although the morphological phenotypes of cells were all transformed by the respective oncogenes. We next evaluated whether SHPS-1 expression was similarly changed in cancerous tissues to search for the involvement of SHPS-1 in tumorigenesis. Examining the tumor tissues compared with the paired normal tissues obtained from human breast cancer (7), Western blot analysis revealed immunoreactive SHPS-1 bands of M_r ∼120,000 in the normal tissues that were infrequently detected in the cancerous tissues (Fig. 1 c). More than 40% of breast cancer tissues appeared to have a decreased expression of SHPS-1 (data not shown). Another group also has reported a lack of SHPS-1 expression in leukemic myeloid cells but not in normal cells (8). This data suggested that SHPS-1 expression might be involved in some step of tumor development. However, it was possible that the reduced expression in the transformed cells or cancerous tissues was either a cause or a consequence of aberrant proliferation of those cells.

On the other hand, forced expression of full-length SHPS-1 into v-src-transformed BALB/c3T3 cells (B-SR) showed abrogation of its anchorage-independent cell growth in soft agar (Fig. 2,c). Consistent with this observation, it has been reported that overexpression of SHPS-1 in NIH3T3 cells causes poor responsiveness to growth factors and subsequent reduced proliferation (2). SHPS-1 is a transmembrane protein that consists of a highly glycosylated extracelluar component, which is responsible for ligand binding and intracellular component, which provides four tyrosine residues for Src family kinases as well as SHP-1 or SHP-2. To elucidate which domain was responsible for the inhibitory activity, two simple truncation mutants of SHPS-1 were produced (Fig. 2, a and b) and introduced into the transformed cells (B-SR). To our surprise, the EC was fairly effective (Fig. 2,c). However, we could not rule out the possibility that the intracellular region (TMIC) functioned in cooperation with EC causing the suppression. We hypothesized that SHPS-1 had a great influence on signaling cascades leading to the suppression of anchorage-independent growth, as B-SR cells expressing wild-type SHPS-1 (FL) showed decreased potential of cell growth in soft agar. It has been suggested that SHPS-1 may play a crucial role in the recruitment of SHP-2 from the cytosol to a site near the plasma membrane and in increasing catalytic activity, thereby regulating the RAS-MAPK signaling cascade (9). To find clues of the mechanism for the anchorage-independent growth inhibition, we checked a series of signaling molecules relevant to cell growth or cell cycle (10, 11). As shown in Fig. 2,d, phosphorylation states of MAPK, JNK, STAT, and AKT in B-SR or FL cells were almost similar when cells were cultured either in the adherent or in the suspension condition. Likewise, the amount of cell cycle regulators such as cyclin A, cyclin B, cyclin D1, CDK4, p27, and RB present in the two cell lines showed no significant differences by Western blotting (Fig. 2,e). However, the expression of caveolin-1, which was down-regulated by v-Src, was restored by the SHPS-1 or the EC but not by the TMIC (Fig. 2,f). FL cells, which expressed almost equal amounts of caveolin-1 compared with that of BALB/c3T3 cells (B), had a slightly slower growth rate than B-SR cells in adherent cell culture (Fig. 2 g), which is consistent with the observation about caveolin-1 function (12). Although the EC fragment seems to be a poor inducer for restoration of caveolin-1 expression, it is possible for the EC to need to act synergistically with TMIC for full restoration activity. Taken together, raising caveolin-1 expression by SHPS-1 seems to reversely correlate with anchorage-independent cell growth.

These findings suggested that key regulators of growth control might themselves be regulated by SHPS-1, causing biological responses such as G₁ arrest, S phase retardation, or induction of uspension-dependent apoptotic cell death, called “anoikis.” We additionally investigated the property of the cells expressing SHPS-1 (FL) to know the mechanism by which anchorage-independent growth was suppressed. Fig. 3,a by FACS analysis showed that SHPS-1 raised the incidence of apoptosis in a suspension-dependent manner. In addition, the results of TUNEL assay supported the idea (Fig. 3,b). In confluent culture conditions, the rate of S phase in FL cells was similar to the BALB/c3T3 cells (B) but less than that of B-SR, suggesting that the SHPS-1 might induce a state of “contact inhibition” (Fig. 3,c). In a sense, that is consistent with “anchorage-dependent cell growth.” The effects of SHPS-1 on cell motility were next measured using a phagokinetic track assay (13). As shown in Fig. 3,d, significant decrease in motility was observed in FL cells compared with B-SR cells. In addition, both FL and EC cells were well spread and less rounded than B-SR cells on the FN, and attachment of cells was not impaired (data not shown). Because cell migration and cell spreading could be implicated with the formation of focal adhesion (14), we checked the cytoskeleton by immunostaining. Rhodamine-phalloidin staining indicated that there were stress fibers in the FL and EC cells (Fig. 3,e). B-SR cells had disrupted actin microfilaments characteristic of SR3Y1 but also showed enhancement of peripheral stress fiber formation. Focal adhesions were then examined by staining with antibodies against paxillin (Fig. 3 e) and vinculin (data not shown), which are major components of focal adhesions (15). Both FL and EC cells manifested focal adhesions stained as patches although the number was less than BALB/c3T3 cells (B).

It is indicated that intracellular domain of SHPS-1, which possesses four potential tyrosine phosphorylation sites, Src homology 2 domain-binding sites, and serves as a substrate for activated receptor-tyrosine kinases (1, 2), may be a direct substrate for activated protein tyrosine kinases as well as nontransmembrane protein tyrosine phosphatases to regulate some physiological and pathological signals (16, 17). However, our data presented here clearly indicated that one of the biological effects of SHPS-1 was at least mediated via the extracellular domain. As CD47 has been shown to be a receptor for the SHPS-1 (18, 19), it seemed most likely that EC exerted its biological effects via the cross-linking with the CD47. Yet, this may not be an exclusive mechanism for suppressing anchorage-independent growth by the EC, because it was effective on all of the cell lines examined including B-SR and a breast cancer cell line in which CD47 was scarcely detected (data not shown). Besides, the expression of SHPS-1 including the deletion mutants could not influence CD47 expression (data not shown). Moreover, GST fused to the EC also could not mimic the inhibitory effect (Fig. 4,e), and addition of a reducing agent such as DTT to the culture medium completely negated the effect of the EC (Fig. 4,f), which we will discuss later. Taken together with the demonstration in Fig. 1, this evidence at least implies that SHPS-1 specifically the EC domain may be directly or indirectly involved in the regulation of cell-growth and tumor development in vivo. We then hypothesized that the EC might exert super-FN (20) activity to exhibit the inhibitory effect on cells because of the following reasons. First, as mentioned before, the EC of SHPS-1 is composed of three immunoglobulin-like domains. Second, interaction of FN with immunoglobulin has been reported previously (21). Third, it has been demonstrated that a polymeric form of FN (super-FN), which is caused by disulfide binding, has an antitumor effect on various tumors grown in immunodeficient mice (20, 22).

To examine the interaction between EC and FN in vitro, we used GST-EC pull-down experiments. As shown in Fig. 4,a, GST-EC but not control GST, specifically precipitated FN from both Balb-3T3 and B-SR cells. When cells that expressed EC were subjected to coimmunoprecipitation experiments, FN was detected in anti-Myc EC immunoprecipitates (Fig. 4,b), demonstrating the interaction of these two proteins in vivo. Similarly, EC was coimmunoprecipitated by an anti-FN antibody (Fig. 4,c). These results demonstrate that EC binds to FN both in vitro and in vivo. We next examined if the GST-EC, which contained undetectable FN (data not shown), had an inhibitory effect on anchorage-independent growth. As shown in Fig. 4,e, GST-EC, as well as control GST, had no significant effects on the growth of B-SR cells in soft agar. In addition, low dose of DTT to the soft agar culture completely neutralized the inhibitory effect of conditioned medium of EC cells. Neither fresh medium used in cell cultures nor the conditioned medium of B-SR cells had the effects regardless of DTT. These results suggest that continuous stimulation of EC is effective for the inhibition of tumor cell invasion by the interaction with FN in vivo. We could not determine if the multimeric form of FN was responsible for its interaction with EC, as it is difficult to detect the multimeric complex by Western blotting because the molecular weight of a monomeric FN is > M_r 240,000. It is plausible that EC can stabilize and/or protect FN from degradation by binding to it. Before the conventional in vivo experiments, we examined MMP2 activity, one of the well-known MMPs related to invasion (23), with zymographic analysis. Although clear proteolytic activity of MMP2 was observed in B-SR cells, there were no differences between control B-SR cells and the cells treated with conditioned medium of EC or GST-EC (Fig. 4,d). We next evaluated the effects of ectopic expression of EC onto B-SR cells in vivo. Injection of B-SR cells caused massive tumor dissemination in 100% of the injected nude mice. However, not only EC cells (data not shown) but also EC plus B-SR cells (1:1 mixture in cell numbers) caused decreased nodule formation in mice (Fig. 4, g and h). We also injected B-SR cells treated with EC-conditioned medium before injection. The result showed slight inhibition of the nodule formation (Fig. 4, g and h). These results suggest that EC is trans-effective to the dissemination of B-SR cells. Hence, transfer of EC might be an appropriate modality for therapeutic intervention, which was consistent with our in vitro observation shown above and with the results of super-FN reported previously (22). In the future, this study may provide an experimental basis for further analysis of SHPS-1 in human cancers in addition to providing further insights into molecular therapy against cancer dissemination. Investigation of the mechanisms by which SHPS-1 and FN promote apoptosis in suspension culture and/or precise evaluation of the effectiveness in vivo should open a door to discovery of new therapeutic strategies.

We thank Sachi Kozawa for excellent technical assistance and Ryan Blackman for the paper correction. We also thank all of the members of our laboratory who participated in this work.