The essential roles of Syk in lymphocyte development and activation of hematopoietic cells have been extensively studied.1, 2, 3 Among its functions in the immune system, Syk plays essential roles in signaling pathways that regulate proliferation, differentiation and apoptosis. Syk is also functionally important in nonhematopoietic cells such as fibroblasts and epithelial cells.4 In these cell types, Syk relays signals from other receptors, including integrins,5, 6, 7 cytokine receptors,8 and G protein–coupled receptors.9 Coopman et al.10 first reported that Syk is commonly expressed in normal human breast tissue, is low or undetectable in invasive breast carcinoma and can function as a tumor suppressor in breast cancer through reduction of cell motility and invasion. Transfection of wild-type Syk into a Syk-negative breast cancer cell line markedly inhibited its growth as well as the formation of metastatic lesions in athymic mice. Support for the antitumor effect of Syk in human breast cancer has come from Toyama et al.,11 who showed that reduced expression of the Syk gene is correlated with distant metastasis and poor prognosis. To help explain the reduced Syk activity, Wang et al.12 presented evidence of an alternatively spliced nonfunctioning Syk variant which is frequently expressed in breast cancer cells but not in normal tissue. Another mechanism thought to lead to the loss of Syk activity is hypermethylation of critical CpG islands, which results in reduced transcriptional activation.13
Very little is known about the mechanism(s) by which Syk modulates the malignant phenotype of breast cancer. Previously, it was demonstrated that Syk can block NFκB activity by inhibiting tyrosine phosphorylation of IκBα, resulting in reduction of cell motility and uPA secretion.14 The cellular events and soluble mediators linking Syk-mediated modulation of NFκB activity with effects on cell motility have yet to be elucidated.
Several cytokines, growth factors and other agents are known to regulate cell motility. In breast cancer, a number of proteins belonging to the chemokine family of inflammatory cytokines have been shown to affect tumor cell invasion and metastasis.15, 16, 17 Among the compounds studied, members of the CXC subgroup of chemokines designated as GRO are important regulators of cell motility and growth. Three isoforms of GRO, α, β and γ, have been characterized and show similar functional activity.18 In addition to their role as neutrophil chemoattractants and mediators of angiogenesis, GRO proteins have been implicated in various stages of melanocyte progression to malignant melanoma19, 20 and in the promotion of squamous cell carcinoma through regulation of cellular proliferation, metastasis, leukocyte infiltration and angiogenesis.21 GRO has also been shown to stimulate the migration and invasive potential of human breast cancer cells.22
In the present study, we utilized human cytokine protein array technology to investigate possible cytokine mediators of Syk activity in human breast cancer. Our results demonstrate endogenous production of GRO family members in this type of cancer and implicate these chemokines as essential mediators of the antiinvasive activity of Syk.
The following reagents were utilized: mouse anti-Syk MAb (Santa Cruz Biotechnology, Santa Cruz, CA), mouse antiphosphotyrosine antibody (4G10; Upstate Biotechnology, Lake Placid, NY), mouse anti-GRO MAb (10G4) and Matrigel (Pharmingen, San Diego, CA), GRO-neutralizing antibody (R&D, Minneapolis, MN), recombinant human GRO and biotinylated anti-GRO (Peprotech, Rocky Hill, NJ), piceatannol (Sigma, St. Louis, MO), polyfect transfection reagent (Qiagen, Alameda, CA) and Boyden-type cell migration chambers (Corning Costar, Corning, NY).
MDA-MB-231 and MCF-7 cells were originally obtained from the ATCC (Manassas, VA). Both MDA-MB-231 and MCF-7 cells were cultured in RPMI-1640 supplemented with 10% FCS, 100 units/ml penicillin, 100 μg/ml streptomycin and 2 mM glutamine in a humidified atmosphere of 5% CO2 and 95% air at 37°C.
Wild-type Syk cDNA in an expression vector (pCAF1) and a mutant Syk lacking kinase activity (K402R) were a generous gift from Drs. P. Coopman and S.C. Mueller (Department of Oncology, Georgetown University Medical School, Washington DC).10 MDA-MB-231 cells were split 12 hr prior to transfection and then transiently transfected with Syk cDNA using polyfect transfection reagent according to the manufacturer's instructions. Briefly, wild-type, mutant or vector (4 μg) was mixed with serum-free medium and then incubated with transfection reagent for 15 min. The complex was added to the cells and incubated further at 37°C for 24 or 48 hr. Cell viability was detected by a trypan blue dye exclusion test. Transfected cells were assessed for expression of Syk by Western blot analysis and used in other experiments. MCF-7 cells were transfected with Syk-specific S-oligonucleotides (4 μg) according to the methods described above. Human ASSyk (5′-TGC CGC TGC TGG CCA TGC TT-3′) and SSyk (5′-AAG CAT GGC CAG CAG CGG CA-3′) with phosphorothioate linkages were synthesized (Qiagen). These oligonucleotide-transfected cells were used for assaying changes in Syk expression by PCR and Western blotting, and the supernatant was used for detecting GRO protein levels by ELISA.
Western blot analysis
To assess the levels of Syk protein expression in MCF-7 and MDA-MB-231 cells, the cells were lysed in lysis buffer (1% Triton X-100 solution containing 1 mM PMSF, 20 μg/ml leupeptin, 2 mM EDTA, 200 mM Na3VO4, 200 mM Na4P2O7, 500 mM NaF) and the cleared lysates collected by centrifugation at 12,000g for 15 min at 4°C. Protein concentrations in the lysates were measured by BCA protein assay (Sigma). Lysates containing equal amounts of total proteins were resolved by SDS-PAGE and the proteins electrotransferred from gel to nitrocellulose membranes. Membranes were incubated with mouse anti-Syk MAb (1:1,000) and incubated further with antimouse HRP-conjugated IgG (1:1,000). Membranes were then washed and developed by the ECL detection system (Amersham Biosciences, Piscataway, NJ) according to the manufacturer's instructions. To check the tyrosine phosphorylation status of Syk, cell lysates were immunoprecipitated with anti-Syk antibody according to the manufacturer's instructions (Roche, Indianapolis, IN), after which equal amounts of the immunoprecipitated samples were processed for Western blotting as described above using mouse antiphosphotyrosine antibody (Zymed, San Francisco, CA). These blots were reprobed with anti-Syk antibody to confirm equal loading of total Syk protein.
Cytokine protein arrays
Assays for cytokine protein arrays were performed as described previously.23, 24, 25 Briefly, protein array membranes were blocked with 5% BSA/TBS [0.01 M TRIS HCl (pH 7.6)/0.15 M NaCl] for 1 hr, then incubated with 1 ml of cell line conditioned medium. After extensive washing with TBS/0.1% Tween-20 (3 times, 5 min each) and TBS (2 times, 5 min each) to remove unbounded material, membranes were incubated with a cocktail of biotin-labeled antibodies against different individual cytokines. Membranes were then washed and incubated with HRP-conjugated streptavidin (2.5 pg/ml) for 1 hr at room temperature. Unbound HRP-streptavidin was washed out with TBS/0.1% Tween-20 and TBS. Finally, signals were detected by ECL (Amersham Pharmacia Biotech, Aylesbury, UK).
Total RNA was prepared using Trizol reagent (Sigma). Total RNA (2 μg) was reverse-transcribed using 18-mer oligodeoxythymidylate and Superscript II (Life Technologies, Rockville, MD) in a volume of 20 μl. Reactions lacking RT were used to verify the absence of amplification from genomic DNA contamination. cDNA templates were subjected to PCR amplification. As a control for cDNA integrity, GAPDH expression was analyzed as well. StarBlast software was applied (DNASTAR, Inc. Madison, WI). Primer sets were 5′-CCATGGCCCGCGCTGCTCTC-3′ (forward) and 5′-TTCCCCTGCCTTCACAATG-3′ (reverse) for GROα, 5′-CGGGTGGCGCTGCTGCTC-3′ (forward) and 5′-CACATACATTTCCCTGCCGTCACA-3′ (reverse) for GROβ, 5′-TTCCCCACCCTGTCATTTATCAAG-3′ (forward) and 5′-GCCAGCTCTCCCGCATTCT-3′ (reverse) for GROγ, 5′-GAGTCAACGGATTTGGTCGT-3′ (forward) and 5′-GACAAGCTTCCCGTTCTCAG-3′ (reverse) for GADPH. PCR conditions were 35 cycles of 94°C for 30 sec, 56°C for 30 sec and 72°C for 40 sec for GROα, β and γ; and 25 cycles of 94°C for 30 sec, 58°C for 30 sec and 72°C for 30 sec for GADPH. All RT-PCR experiments were repeated at least 3 times.
ELISA for GRO
Conditioned medium was collected from cells seeded at 104/well in 6-well plates (VWR, Suwanee, GA) after culture times as indicated in the figure legends. Samples and standard cytokines were then added to 96-well ELISA plates, which had been coated using 100 μl of 2 μg/ml capture antibody (anti-GRO, Pharmingen); and plates were then incubated overnight at 4°C. Unbound material was washed out with PBS/0.05% Tween-20. Biotinylated anti-GRO detection antibody (100 μl of 1 μg/ml, Pharmingen) was added to each well. After incubation for 1 hr at room temperature, plates were washed, 100 μl of streptavidin-HRP conjugated antibodies (1:1,000) were added to wells and incubation was continued for an additional 30 min at room temperature. After extensive washing, color was developed by incubation with substrate solution containing ethylbenzthiazoline sulfonate (Sigma), and plates were read at OD 405 nm.
Cell migration assay
The migration assay was performed using 12-well polycarbonate filter (8 μm pore size) Transwells (Corning Costar) coated with 20 μl Matrigel (Pharmingen) diluted with 2 × volume of serum-free RPMI-1640. The lower chamber contained RPMI supplemented with 10% fibroblast conditioned medium in the absence or presence of serially diluted recombinant GRO. Pooled, transfected tsyk or MDA-MB-231 vector control cells (1 to 5 × 104) resuspended in 300 μl of serum-free RPMI medium containing 1% BSA were plated in the upper compartment of transwells. In some experiments, anti-GRO antibody was also added to the upper chamber, as indicated. After 24–36 hr of incubation, cells that migrated to the lower side of the filter were fixed in 4% formaldehyde and stained with Giemsa solution (Sigma). Cells remaining in the Matrigel side were removed by cotton swabs. Cells that had invaded through the Matrigel to the bottom side of the transwells were counted in 3 random fields under a microscope.
The software SPSS (Chicago, IL) 10.0 was used for statistical analysis. Statistical significance was determined with Student's t-test (2-tailed) comparison between 2 data sets. Asterisks shown in the figures indicate significant differences of experimental groups in comparison to the corresponding control condition. p < 0.05 was considered a significant difference.
Modulation of Syk expression in breast cancer cells
Endogenous expression of Syk is not detectable in the human breast cancer cell line MDA-MB-231. As a model for investigating effector mechanisms of Syk activity, we transfected these cells with a wild-type Syk cDNA expression plasmid. To determine whether our transfectional methodology produced the desired effects, expression of Syk in parental and transiently transfected cells was analyzed by Western blotting. Figure 1a shows that, as expected, Syk expression was not detectable in control (empty vector–transfected) cells. In contrast, cells transfected with 4 μg of Syk cDNA (tsyk) showed ample levels of Syk protein, which persisted for at least 2 days. To assess the functional competence of the overexpressed Syk protein, its tyrosine phosphorylation status was evaluated. In these experiments, the cell lysate from tsyk MDA-MB-231 cells was immunoprecipitated with anti-Syk MAb, then analyzed by Western blotting using antibody against phosphotyrosine to detect the phosphorylated form of the protein. As seen in Figure 1b, tsyk cells showed tyrosine-phosphorylated Syk expression. To assess whether this phosphorylation was due to autophosphorylation by Syk activation, we also transfected MDA-MB-231 cells with a kinase-deficient mutant of Syk (SykK−).10 Figure 1b shows that these cells did not express tyrosine phosphorylation. Reprobing the same filter with an anti-Syk antibody revealed that the amount of Syk immunoprecipitated from all samples was comparable, indicating that the tyrosine phosphorylation of Syk resulted mainly from autophosphorylation. This finding confirms the functional competence of Syk kinase activity in transfected MDA-MB-231 cells. As expected, no Syk protein was detected in immunoprecipitant from parental MDA-MB-231 cells (not shown).
As another model for investigating downstream events involved in the Syk signaling pathway, we utilized the MCF-7 breast cancer cell line, which expresses readily detectable levels of the kinase.10 In these cells, we specifically downregulated constitutive Syk expression by transfection with antisense oligonucleotides; sense oligonucleotides served as control vectors. The data showed that 4 μg of antisense, but not sense, oligos effectively decreased Syk expression at both the protein and mRNA levels (Fig. 1c and data not shown).
Identification of Syk-regulated cytokines by protein array
As an initial screen to investigate the possible involvement of cytokines, growth factors or other agents that may be involved in the cellular action of Syk, we utilized an antibody-based protein array system as previously described.23, 24, 25 This methodology can simultaneously screen for the presence and relative production of more than 70 cytokines/chemokines with high specificity and sensitivity. Figure 2 illustrates the characteristics of the ECL array membranes that were developed from the supernatant of 24 hr cultured control and tsyk MDA-MB-231 cells. As shown, this methodology detected the presence of GM-CSF, Ang, IGFBP-1, TIMP-2, GRO, IL-6 and IL-8 in supernatant from control cells. As seen in Figure 2b, the same cytokines were detected and at similar relative concentrations in the conditioned medium from tsyk cells, with the exception of GRO. This growth factor was markedly reduced in supernatant. To further investigate and quantitate this finding, we standardized a GRO ELISA that showed a limit of detection at 0.1 pg/ml GRO protein. As seen in Figure 3, supernatant from 48 hr MDA-MB-231 and MCF-7 cultures contained approximately 3,500 and 260 pg/ml GRO protein, respectively. Overexpression of Syk in tsyk cells resulted in a 73% reduction in secreted GRO protein, while treatment of these cells with piceatannol, a Syk inhibitor,14 enhanced GRO secretion to levels similar to those secreted from control MDA-MB-231 cells. The decreased Syk expression in antisense transfected MCF-7 cells caused an approximate 50% increase in GRO (Fig. 3).
Syk differentially modulates GRO family members
Studies have identified 3 isoforms of GRO, α, β and γ, all of which have shown biologic activity as neutrophil chemoattractants18, 26 and migration inducers of certain epithelial cells, including breast cancer.22 Production of GROα and β in human endometrial cells, GROα in squamous cell carcinoma and GROα, β and γ in colon carcinoma cells has been demonstrated.27, 28 However, production or secretion of these chemokines by mammary tissue has heretofore not been reported. As seen in Figure 4a, MDA-MB-231 cells express all 3 GRO isoforms, with α showing the greatest abundance, followed by β and then γ. These transcripts were differentially regulated by overexpression of Syk in the cells; mRNA levels of GROα and γ decreased significantly in tsyk cells compared to parental and vector control cells. In contrast, GROβ showed no significant difference in transcript levels resulting from Syk overexpression in tsyk (Fig. 4b).
Downregulation of GRO is involved in Syk-mediated suppression of cell migration
To delineate the effect of Syk and GRO on cell migration, we utilized a Matrigel migration assay. In these experiments, the ability of exogenous GRO or neutralizing anti-GRO antibody to alter the invasive activity of cells was quantitated. As seen in Figure 5, cell invasion was markedly reduced when MDA-MB-231 cells were treated with anti-GRO antibody or when transfected with the Syk expression plasmid (tsyk cells). Addition of exogenous recombinant GRO to tsyk cultures caused these cells to show similar invasive properties as parental and control vector-transfected MDA-MB-231 cells. Dose–response experiments indicated that 100 ng/ml GRO completely reversed the lowered invasive properties of tsyk compared to control cells, while GRO had no apparent effect on control cells (Fig. 6a). Conversely, addition of anti-GRO reduced the invasive activity of control cells starting at 20 μg/ml while showing about a 35% reduction at 40 μg/ml (Fig. 6b). For tsyk cells, the effect of anti-GRO antibody was less apparent, demonstrating about a 20% further reduction in invasive activity at the highest concentration of antibody tested compared to untreated transfectants.
Syk is commonly expressed in normal human breast tissue, benign breast lesions and low-tumorigenic breast cancer cell lines. In contrast, there is low or undetectable level of Syk expression in invasive breast carcinoma tissue and cell lines.10 Previous reports have shown that Syk is not detectable in highly invasive MDA-MB-231 cells but is expressed in the much less invasive MCF-7 line.10 In the present study, we manipulated the expression of Syk in these 2 cell lines to determine whether some of the functional effects of this gene could be mediated by cytokines known to play a role in cell migration, invasion and/or metastasis.
Forced expression of Syk in MDA-MB-231 cells was achieved by transient transfection of a Syk cDNA expression plasmid, while reduction of endogenous Syk protein levels in MCF-7 cells was accomplished by transfection with phosphorothioate-linked antisense Syk oligonucleotides. Western blotting of lysate from both cell line models confirmed that the desired modulation of Syk protein levels was achieved. Using protein array technology developed by Huang et al.,23 we demonstrated that MDA-MB-231 cells secrete detectable levels of a number of cytokines that directly act on breast cancer cells to regulate growth, invasion and motility. Of the regulatory cytokines detected, IL-6, IL-8, IGFBP-1 and TIMP-2 have previously been shown to be secreted by MDA-MB-231 and other breast cancer cells and tissue.29, 30, 31, 32 Interestingly, the protein array technique indicated the presence of GRO protein using reagents that were not isotype-specific (i.e., detected total GRO protein), while specific detection of the α isotype was not apparent (Fig. 2b, control cells). This finding was surprising since GROα has been reported to be the most prevalent isotype in most tissue types studied.33, 34 However, using total GRO vs. GROα-specific ELISAs, we determined that >75% of the GRO protein secreted by MDA-MB-231 cells is, in fact, GROα (data not shown). This result is consistent with our data showing that, while GROα, β and γ are all expressed in MDA-MB-231 cells, mRNA levels of GROα are substantially greater than those of the other 2 isotypes. We have concluded that the protein array technology shows markedly greater sensitivity at detecting total GRO vs. GROα using commercial antibody reagents.
In the MDA-MB-231 cell line, the concentration of GRO in the culture supernatant was 3–4 ng/ml, while less than one-tenth of these levels were secreted by MCF-7 cells. These levels are difficult to compare to other cell types due to variation in culture conditions. However, high GRO secretion such as that from MDA-MB-231 cells has been demonstrated from bronchial epithelial cells, while normal human eosinophils secrete lower levels, in the range produced by MCF-7.34, 35 Since GRO is known to promote cellular migration and invasion,28 this difference in breast cancer cell lines is consistent with the high vs. low invasive properties of MDA-MB-231 and MCF-7 cells, respectively. That endogenous GRO production actively contributes to the malignant phenotype of MDA-MB-231 was indicated by the ability of anti-GRO antibody to markedly reduce the in vitro migration of these cells through Matrigel. Conversely, addition of recombinant GRO to low GRO-secreting MCF-7 cells was previously shown to stimulate cellular migration and invasion.22
Of the cytokines detected in the supernatant of MDA-MB-231 cultures, GRO was the only one that was significantly altered by overexpression of Syk; secreted levels of GRO from tsyk cells were reduced by over 65% compared to controls. Conversely, downregulation of relatively high levels of constitutively expressed Syk in MCF-7 cells with antisense oligos upregulated GRO secretion, consistent with reports that such treatment stimulates their migration in Matrigel.14 These data provide compelling evidence that Syk activity can directly modulate the production and secretion of GRO in human breast cancer cells. Our data show that increasing Syk differentially regulated the GRO isotypes; mRNA expression of GRO-α and γ were downregulated while GROβ was not affected. This finding was surprising since overexpression of Syk in MDA-MB-231 has been shown to block “global” NFκB activity14 and all 3 GRO isoforms have NFκB regulatory elements in the promoter.36, 37 Thus, it is unlikely that modulation of NFκB activity can, by itself, account for all the regulatory effects of Syk on GRO expression. In this sense, cooperation between NFκB and an additional sequence motif (GGGCGTAGC) located 10 bp downstream of NFκB has been implicated in regulating the IL-2Rα gene in human T cells.38 Interestingly, GROα and γ, but not GROβ, have a sequence (GCTCCGCCC) 12 bp downstream of the NFκB site, whose reverse complement has only one mismatch with the above sequence.39 The possible role of this element in regulating GROα and γ expression has not been determined. Other regulatory sequences that segregate with GROα and γ vs. β are a CRE site40 and 2 SP1 sites41 that are located in the GROβ promoter but not in that of the other 2 isoforms. Further support for the differential regulation of the GRO isoforms by Syk comes from studies showing that different inflammatory inducers can stimulate different patterns of GRO expression in the same cell type.27 The contention that Syk regulation of GRO and other chemokines may be highly complex is further supported by the fact that secretion of IL-8, another NFκB-responsive gene,37 was also not altered by overexpression of Syk in MDA-MB-231 transfectants (i.e., tsyk cells).
Mahabeleshwar and Kundu14 suggested that at least part of the ability of Syk to suppress the motility of MDA-MB-231 cells was through inhibition of uPA secretion. Although this work convincingly demonstrated that NFκB plays a crucial role in uPA secretion and migration of the cells, the exact cellular events linking Syk with uPA modulation have not been elucidated. In lieu of our present findings, a role for Syk regulation of GRO in mediating its effects on uPA must be considered. Thus, cytokines such as tumor necrosis factor, lymphotoxin, IL-1 and IL-8 have been shown, by themselves, to induce production and secretion of uPA from a variety of cell types.42, 43, 44 Although GRO-dependent regulation of uPA has not been reported, the ability of GRO to promote tumor cell invasion through extracellular matrix degradation suggests a direct link between GRO, uPA and other fibrinolytic pathway(s). Alternatively, GRO and uPA may play necessary, but independent, roles that together contribute to the highly invasive phenotype of MDA-MB-231 cells.
In summary, our findings provide evidence that human breast cancer cells express and secrete GRO and that production of this protein is regulated by Syk tyrosine kinase activity. As such, the data demonstrate that GRO may play a critical role in mediating the antimetastatic effects of Syk and support the conclusion of a direct relationship between Syk activity and breast cancer prognosis.