Protein tyrosine phosphatase Meg2 dephosphorylates signal transducer and activator of transcription 3 and suppresses tumor growth in breast cancer

MCF7, MDA-MB-231, and HEK293T cells were obtained and characterized by a cytogenetic analysis by American Type Culture Collection (ATCC) and maintained in this lab according to the recommendation of ATCC. The cell lines were characterized in this lab by morphological analysis before using for experiments. The v-Src/NIH3T3 cell line was a gift from Dr. H Yu at City of Hope Comprehensive Cancer Center, California, USA and was characterized by morphological analysis in this lab according to her recommendation. The stable cell line for depletion of PTPMeg2 by shRNA was generated in this lab based on MCF7 and characterized by morphological analysis and the expression of targeted gene was characterized by a Western blot. The cells were cultured in DMEM medium supplemented with 10% fetal bovine serum in 5% CO₂ astrosphere in 37°C. The mouse hepatic cell lines STAT3^-/- (KO) and STAT3^+/+ (WT) derived from STAT3 conditional knockout and wild-type mice were also cultured in DMEM medium.

Anti-sera against PTPMeg2 were generated by immunizing rabbits with purified GST-PTPMeg2 (1-1779) proteins in ZJ Zhao's lab. Anti-Myc (9E10), anti-HA (F-7), anti-GFP (FL), anti-pSTAT3(Tyr705, B-7), anti-pSTAT3(Ser727), anti-STAT3 (F-2) and anti-STAT3 (C-20) antibodies, and protein G/A plus agarose beads were purchased from Santa Cruz Biotechnology, and anti-β-Actin (AC-15) antibody from Sigma. MG132 and leupepstin were purchased from Amresco (AMRESCO Inc. OH), and human recombinant IL-6 and IL-6 receptor from B&D (B&D SYSTEM, USA). Plasmids including GST-STAT3, pXJ40-STAT3, and its deletions were kindly provided by Dr. Xinmin Cao [25]. Expression vectors for human PTPMeg2 and PTPMeg2C515S or deletion mutants were constructed based on pcDNA3.1-Myc or pEFneo-Myc. Other plasmids involved in this study were stored in the lab.

MCF7 cells were plated in 24-well plates the day before transfection. An amount of 0.1 μg of reporter plasmid pAPRE-luc or M67-luc together with 5 ng of an internal control plasmid pRL-TK was transfected. Constructs expressing STAT3, PTPMeg2 and its mutants were co-transfected at an amount of 0.4 μg per well. To deplete endogenous PTPMeg2, 0.8 μg of vectors with shRNA targeting PTPMeg2 (shRNA1: 5'-GATCCGGAAAGGCATTGTAAATTCAAGAGATTTACAATGCCTTCCTTCCTTA-3', shRNA2: 5'-GATCCGCAAGGAATCTATGAGGAATTCAAGAGATT CCTCATAGATTCCTTGCTTA-3', shRNA3: 5'-GATCCCTAGAGTGAAGCT AACAATTCAAGAGATTGTTAGCTTCACTCTAGTTTA-3') in pSilencer-4.1 was transfected. Twenty-four hours after transfection, luciferase assays were performed with a dual-luciferase reporter assay system (Vigofect Inc. Beijing, China) and the luciferase activity was normalized by firefly against the renilla luciferase activity.

Proteins were analyzed by SDS-PAGE and Western blot [26]. For immunoprecipitation experiments, HEK293T cells grown in 60 mm dishes were transfected with indicated expression plasmids and were lysed in cell lysis buffer (20 mM Tris-Cl, pH 7.6, 150 mM NaCl, 5 mM EDTA, 10 mM MgCl₂, 0.5% NP40, 10% glycerol, 1 mM DTT, 0.1 mM Na₃VO₄, 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 μg/ml pepstatin) for 30 min on ice, and then the lysates were centrifuged at a maximum speed for 15 min. Supernatants of cell lysates were incubated with 2 μg of indicated antibodies overnight at 4°C, and 30 μl of protein IgG/A agarose plus beads were added for binding for 4 h at 4°C. Beads were washed with cell lysis buffer (containing 0.1 mM vanadium) 4 times and bound proteins were eluted with 2 × loading sample buffer and analyzed by Western blot with indicated antibodies. For GST-pull down assays, the procedure was similar to that in immunoprecipitation experiments except that GST beads were used and washed by PBST buffer.

GST-PTPMeg2 WT and GST-PTPMeg2CS proteins were expressed and purified as described previously [27]. Phosphorylated Flag-STAT3 was prepared from HEK 293T cells transfected with Flag-STAT3 for 48 h and then stimulated with IL-6 for 30 min. Purified phosphorylated Flag-STAT3 was incubated with GST and GST-PTPMeg2 WT or GST-PTPMeg2CS proteins at 37°C for 30 min with PTPase reaction buffer (25 mM Tris-Cl, pH 7.5, 2.5 mM EDTA, 1 mM EGTA, 1 mg/ml BSA, 10 μM TPCK, 1 mM Benzamidine, 5 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 μg/ml pepstatin). The dephosphorylation reaction was terminated by directly boiling. Proteins were separated with SDS-PAGE and analyzed with an anti-pSTAT3 and anti-Flag antibodies by a Western blot.

Cells were seeded on glass coverslips for 24 h and subjected to serum-starvation for 18 h before treatments with IL-6 for 30 min. Cell was fixed in 4% paraformaldehyde for 20 min at 4°C and permeabilized in 0.3% Triton X-100 for 15 min. Cells were blocked with 10% normal rabbit serum at room temperature and incubated in primary antibody overnight at 4°C. Primary antibodies used were an anti-PTPMeg2 or an anti-Flag antibody. Secondary antibodies used were FITC-conjugated and TRITC-conjugated IgG. Nucleus was stained with DAPI.

MCF7 cells stably transfected with the shPTPMeg2 plasmid, or MDA-MB-231 or mouse hepatic STAT3 KO cells infected with an adenovirus for over-expression of PTPMeg2, were seeded on 96-well plates at a density of 1000 cells/well in triplicate. Cells were cultured for different times and 5 g/L of MTT was added 4 h before termination of cell growth. The purple blue sediment was dissolved in 150 μl of DMSO before harvest. The relative optical density (OD)/well was determined at wavelength of 570 nm in a WELLSCAN MK3 ELIASA (Lab systems, Dragon, Finland) using a 630 nm reference filter. Cell growth curve was drawn according to the average of OD570-OD630.

Exponentially growing MCF7 cells were stably transfected with the PTPMeg2 shRNA vector and MDA-MB-231 were infected with the adenovirus or retrovirus and v-Src/NIH3T3 cells with the adenovirus for over-expression of PTPMeg2. Cells were suspended in 1 ml physiological saline for preparation of injection into mice. BALB/c-nu/nu mice at 5 weeks of age were subjected to bilateral subcutaneous injections with 1.0 × 10⁷ (MCF-7) or 5.0 × 10⁶ (MDA-MB-231) or 5.0 × 10⁵ (v-Src/NIH3T3) cells in a volume of 0.1 ml saline. Tumor volume defined as (length × width²)/2 was measured every two days with a caliper up to study termination. Tumor growth curves were drawn according to the average of tumor volumes (mm³). All animal experiments were performed in accordance with the institutional animal experiment guidelines.

The formalin-embedded tissue samples from 73 patients with breast carcinomas diagnosed between 2008 and 2010 were obtained from the Surgical Pathology in the TangShan People's Hospital. All breast cancer specimens from female patients were obtained from clinical surgery. Information of age, histological type, differentiation grade, and lymph node metastasis of breast carcinomas were obtained from the Surgical Pathology Files in the hospital. The clinicopathological diagnosis on the tumor status was evaluated by the clinical pathologists in the hospital. All samples used in the study were approved by the TangShan People's Hospital Ethical Committee under the guidance of tissue collection procedure with informed consent.

Sections of formalin-fixed breast carcinoma tissues were treated with 0.3% hydrogen peroxidase/methanol and incubated with primary antibodies followed by incubation with secondary antibodies (biotin-labeled; Santa Cruz) and third antibodies (peroxidase-labeled; Santa Cruz). Samples were developed using DAB as substrates (Santa Cruz). Scoring criteria for tumor degrees as reported previously were used [28]. Briefly, the grade was classified as 0 for negative, 1 for weak (< 10%), 2 for moderate (10-25%), 3 for strong (26-75%) and 4 for very strong (> 75%) staining according to percentage of positively staining cells. The staining index was subsequently obtained by multiplication of the proportion and intensity and calculated index was finally assessed by a simplified score (score 0, index 0-1; score 1, index 2-4; score 2, index 6-8; score 3, index 9-12). Samples with staining score of at least 1 were classified as positive staining, score 2 and 3 were strong positive staining. The percentage of pSTAT3 nuclear positive cells were used to classify the grade of its expression as negative (< 5%), weak (6-50%), and strong (51-100%).

All experiments were repeated at least three times. The Student's t test was used to evaluate the significance of differences between experimental and control groups. Data were analyzed by one-way ANOVA with SPSS13.0 (SPSS, Int, Chicago, IL). Frequencies of PTPMeg2, STAT3 and pSTAT3 expressions among cancer samples were analyzed by the x ² test with a modification by the Fisher's exact test to account for frequency values < 5. The correlation between protein levels was evaluated by the pair-wise Pearson correlation coefficient and by bi-dimensional hierarchical clustering. All Ps reported were two-sided. Significance was defined at the level of P ≤ 0.05.

To search for negative regulators of STAT3, we examined the possibility of its interaction with different phosphatases, and PTPMeg2 was identified as a potential interacting protein. To confirm the interaction, Myc-PTPMeg2 and Flag-STAT3 were co-expressed in HEK 293T cells and co-immunoprecipitation and GST pull-down experiments were performed. The results showed that PTPMeg2 interacts with STAT3 in vitro (Figure 1A and 1B). Interestingly, we observed that PTPMeg2 preferentially interacted with STAT3 as it had either a weak or no interaction with STAT5 or STAT1 (Figure 1C). An in vivo interaction of endogenous PTPMeg2 and STAT3 proteins was observed in the mouse brain tissue (Figure 1D, left panel) and breast cancer MCF7 cells (Figure 1D, right panel). All these results suggested that PTPMeg2 interacts with STAT3 under physiological and pathological conditions.

To determine if the interaction of PTPMeg2 with STAT3 is regulated by cytokines, HEK 293T cells transfected with Flag-STAT3 and Myc-PTPMeg2 were stimulated by IL-6 for 30 min. A reciprocal immunoprecipitation experiment indicated that the interaction of PTPMeg2 and STAT3 was increased dramatically under stimulation of IL-6 (Figure 1E). Interestingly, we observed a strong band of phosphorylated STAT3 in a complex precipitated with an anti-Myc antibody (for Myc-PTPMEG2) (Figure 1E). Consistently we observed that IL-6 induced the interaction of endogenous STAT3 and PTPMeg2 in MCF7 cells (Figure 1F). These results suggest that PTPMeg2 interacts with the phosphorylated form of STAT3 (pSTAT3). Based on the observation that PTPMeg2 interacts with STAT3 in the absence of IL-6, we concluded that PTPMeg2 interacts with both the phosphorylated and unphosphorylated STAT3.

To reveal the cellular location of the PTPMeg2/STAT3 complex, we performed an immunofluorescence staining assay in MCF7 cells transfected with STAT3 and PTPMeg2. The results showed that STAT3 was located in the cytoplasm under a quiet condition, but translocated into the nucleus after IL-6 stimulation (Figure 1G, b2-b3). When STAT3 was co-expressed together with PTPMeg2, a notable co-localization of the two proteins in the cytoplasm was observed (Figure 1G, c3). Interestingly, we observed that STAT3 remained in the cytoplasm under the stimulation of IL-6 when PTPMeg2 was co-expressed (Figure 1, 2, 3). This result suggests that PTPMeg2 blocks the translocation of STAT3 from the cytoplasm into the nucleus upon IL-6 stimulation. To support this notation, a mutant PTPMeg2CS, which lost the ability to dephosphorylate STAT3, failed to block STAT3 localization into the nucleus in response to IL-6 stimulation (Figure 1, 2, 3). These results suggest that STAT3 colocalizes with PTPMeg2 in the cytoplasm and overexpression of PTPMeg2 inhibits the translocation of STAT3 upon cytokine stimulation.

Our observation that over-expression of PTPMeg2 blocks STAT3 translocation implied that PTPMeg2 may regulate STAT3 phosphorylation. Since PTPMeg2 is a phosphatase, we determined to examine whether PTPMeg2 dephosphorylates STAT3. To this end, HEK293T cells were transfected with Flag-STAT3 and Myc-PTPMeg2 plasmids under IL-6 treatment for 30 min. The results showed that the level of pSTAT3 was decreased when PTPMeg2 was co-expressed with STAT3 (Figure 2A). In contrast, transfection of mutant PTPMeg2CS failed to decrease the level of pSTAT3 (Figure 2A). To examine whether the decreased level of pSTAT3 is induced by a dephosphorylation or protein degradation process, the level of pSTAT3 was examined after withdrawal of IL-6 and in the presence of MG132, an inhibitor of proteosome. Results showed that the level of pSTAT3 was decreased much more quickly when PTPMeg2 was over-expressed than that without PTPMeg2 (Figure 2B). At the same time, the level of pSTAT3 remained unchanged in the presence or absence of MG132 (Figure 2B). These dada indicated that PTPMeg2 induces dephosphorylation of pSTAT3 rather than its degradation. Furthermore, we showed that over-expression of PTPMeg2 promoted the dephosphorylation of STAT3 at the residue Tyr 705 but had no effect on the phosphorylation level of pSTAT3 at the residue Ser727 (Figure 2C). The role of PTPMeg2 on the dephosphorylation of pSTAT3 (Tyr705) was further confirmed in a dosage dependent experiment (Figure 2D). These results suggested that ectopic expression of PTPMeg2 regulates the tyrosine phosphorylation of STAT3.

To further confirm the role of PTPMeg2 on dephosphorylation of STAT3, purified GST-PTPMeg2 and GST-PTPMeg2CS fusion proteins were used to incubate with pSTAT3 prepared from mammalian cells for an in vitro phosphatase activity experiment. The results showed that the tyrosine phosphorylation level of STAT3 was dramatically reduced when GST-Meg2 protein was added in a dose dependent manner (Figure 2E). As controls, addition of GST or GST-PTPMeg2CS had no effect on the level of pSTAT3 (Figure 2E, lane 1 to 2). This result indicated that STAT3 is a substrate of PTPMeg2. To address whether the PTP domain of PTPMeg2 has the phosphatase activity, the SEC domain, PTP domain and mutations of different deletions (ΔSEC, ΔPTP) were generated to examine the effect on the level of pSTAT3. A Western blot result showed that both PTP domain and ΔSEC domain had the ability to dephosphorylate pSTAT3 (Figure 2F, lanes 4 and 5). These data indicated that the PTP domain is responsible for the phosphatase activity of PTPMeg2, which is in consistency with the role of the PTP domain in other phosphatases.

We questioned whether PTPMeg2 regulates the transcriptional activity of STAT3 based on its interaction with STAT3. To this end, we used an APRE luciferase reporter, which responds to STAT3 activation, to examine the effect of PTPMeg2 on STAT3 mediated transcriptional activity. The results showed that over-expression of PTPMeg2 in MCF7 cells resulted in a decrease of the luciferase activity in response to over-expressed STAT3 and stimulation of IL-6 (Figure 3A). The inhibitory role of PTPMeg2 on the STAT3-mediated luciferase activity was dose dependent (Figure 3B, left columns). Interestingly, when the mutant PTPMeg2CS was increasingly expressed the STAT3-mediated luciferase activity was increased (Figure 3B, right columns). These results suggest that the mutant PTPMeg2CS acts as a dominant negative antagonist of endogenous PTPMeg2 in regulating STAT3 phosphorylation. In consistence, depletion of the PTP domain impaired the activity of the phosphatase (Figure 3C). Finally, we showed that depletion of PTPMeg2 by three shRNAs increased the luciferase activity mediated by STAT3 (Figure 3D) while these shRNAs dramatically recovered the phosphorylation of the endogenous STAT3 protein (Figure 3E). In contrast, over-expression of PTPMeg2 had no effect on the transcriptional activity of STAT1 in response to INF-gamma stimulation (Figure 3F). These results indicate that PTPMeg2 inhibits STAT3 activation with certain specificity.

Since STAT3 phosphorylation is highly related to tumorigenesis, we attempted to examine whether PTPMeg2 could affect tumor progression. For this purpose, we used two human breast cancer cell lines MCF7 and MDA-MB-231. We observed that the endogenous PTPMeg2 protein level was low in MDA-MB-231 cells but high in MCF7 cells while the level of endogenous pSTAT3 displayed a reversed trend (Figure 4A). Therefore we determined to establish a gain-of-function model in MDA-MB-231 cells and a loss-of-function model in MCF7 cells.

To address the increased pSTAT3 was the cause of decreased PTPMeg2, we stably depleted PTPMeg2 by using an shRNA targeting PTPMeg2 in MCF7 cells. A Western blot analysis showed that pSTAT3 was dramatically increased when PTPMeg2 was depleted (Figure 4B, right lane). Intriguingly, a cell proliferation experiment result showed that the growth of MCF7 cells was increased when PTPMeg2 was depleted (Figure 4C). An in vivo tumor growth experiment in a xenograft tumor model in mice showed that MCF7 cells with stable depletion of PTPMeg2 formed larger tumors than mock transfected cells (Figure 4D-E) and grew more rapidly (Figure 4F). On the other hand, we over-expressed PTPMeg2 in MDA-MB-231 cells using an adenovirus expression system. The results showed that MDA-MB-231 cells infected with the adenovirus expressing PTPMeg2 had a lower level of endogenous pSTAT3 than the cells infected with a control adenovirus (GFP) (Figure 4G, right lanes). And these cells grew much more slowly (Figure 4H) and the cells formed smaller sized tumors (Figure 4I-J) and had slower tumor growth rate (Figure 4K), lower tumor weight (Figure 4J) and slower tumor growth (Figure 4K). To further confirm the inhibitory role of PTPMeg2 on tumor growth in a moderate expression system, we used the retroviral system to ectopically express PTPMeg2 (pMSCV/PTPMeg2) in MDA-MB-231 cells. The results were similar to that using the adenovirus expression system (Figure 4L-N). All these results indicated that PTPMeg2 inhibits STAT3 phosphorylation directly and PTPMeg2 is a tumor suppressor.

To confirm the inhibitory role of PTPMeg2 on tumor growth is depended on regulation of STAT3 phosphorylation, we used v-Src transformed NIH3T3 fibroblasts in a xenograft tumor model. The result showed that v-Src transformed cells had a much higher STAT3 phosphorylation level than non-transformed cells and over-expression of PTPMeg2 significantly decreased the level of pSTAT3 (Figure 5A). Consistent with the decreased level of pSTAT3, the tumor size (Figure 5B), weight (Figure 5C) and tumor growth (Figure 5D) from v-Src transformed cells were decreased when PTPMeg2 was forcedly expressed. These data implied a correlation of PTPMeg2-reduced tumor growth and the decreased level of pSTAT3. To address whether pSTAT3 is a key target by PTPMeg2, we examined the cell proliferation ability in the STAT3 KO cells. A MTT experiment indicated that overexpression of PTPMeg2 inhibited the cell growth dramatically in wild-type cell but had no effect in the STAT3 KO cells (Figure 5E-F), suggesting that the inhibitory role of PTPMeg2 on the cell proliferation is depended on STAT3. Together with the biochemical data, these results suggested that the ability of PTPMeg2 to inhibit the tumor growth and cell proliferation is depending on its role of regulation of phosphorylated STAT3.

In consistence with the tumor growth experiment, we examined the expression of STAT3 downstream genes. The result showed that both Bcl-xL and cyclin D1 were dramatically increased when PTPMeg2 was depleted in MCF7 cells (Figure 4B). In contact, Bcl-xL and cyclin D1 were decreased when PTPMeg2 was over-expressed in MDA-MB-231 cells (Figure 4G). We observed no alteration of STAT3 expression but the phosphorylated STAT3 level was changed with either over-expression or depletion of PTPMeg2 (Figure 4B, G). These results suggest that PTPMeg2 regulates STAT3 phosphorylation and thereafter the downstream gene expression.

To address whether PTPMeg2 regulates STAT3 dephosphorylation in human tumors, we examined the correlation of pSTAT3 level and expression of PTPMeg2 in human breast cancers. The result showed that expression of PTPMeg2 was in a strong positive status in peritumoral tissues (90% = 37/41) and in a negative status in paired tumor tissues (83% = 59/71) (Figure 5E, Table 1). In contrast, pSTAT3 remained at a low (or negative) level in the peritumoral tissues but at a high level (71% = 52/73) in the paired tumor tissues (Figure 5E). We observed a negative correlation between PTPMeg2 expression and the pSTAT3 level (spearman's correlation coefficient was -3.33, p = 0.004, Table 2) from a Spearman's correlation test. The analysis also revealed that the increased STAT3 level was correlated with reduced PTPMeg2 expression in the breast carcinoma (correlation coefficient is -2.65, p = 0.023, Table 2). These data indicated that PTPMeg2 might be an important regulator of STAT3 dephosphorylation in tumors.