ANTICANCER RESEARCH 30: 1937-1946 (2010)

Function and Regulatory Mechanisms of the Candidate Tumor Suppressor Receptor Protein Tyrosine Phosphatase Gamma (PTPRG) in Breast Cancer Cells

SHERRY T. SHU1*, YASURO SUGIMOTO1,2*, SULING LIU1, HSIANG-LIN CHANG1, WEIPING YE1, LI-SHU WANG1, YI-WEN HUANG1, PEARLLY YAN2,3 and YOUNG C. LIN1,2

1Laboratory of Reproductive and Molecular Endocrinology, Department of Veterinary Biosciences, College of Veterinary Medicine, 2The Ohio State University Comprehensive Cancer Center, and 3Department of Molecular Virology, Immunology and Medical Genetics, The Ohio State University, Columbus, OH 43210-1093, U.S.A.

Abstract. Background: Protein phosphorylation is one of the transfected with vector alone. Similarly, overexpression of essential steps in , and aberrant phosphorylation PTPRG also delayed the re-entry of breast cancer cells into is a common event in human cancer. The expression of the cell cycle after serum starvation, and reduced the receptor type protein tyrosine phosphatase gamma (PTPRG) phosphorylation levels ERK1/2 in MCF-7 cells. In addition, in normal breast is found to be ~50-60% higher than that of methylation assays in PTPRG promoter in breast cancer cell breast tumor tissue. Overexpression of PTPRG inhibits lines (including SK-Br-3) revealed an aberrant methylation anchorage-independent growth and proliferation of breast pattern. When SK-Br-3 and MCF-7 cells were treated with cancer cells. To understand the tumor suppression deoxy-5-azacytidine (DAC) and trichostatin A (TSA), these characteristics of PTPRG, we studied its tumor suppressive compounds reactivated the expression of PTPRG, suggesting function in an athymic mouse model and evaluated factors that an epigenetic control on its expression. Conclusion: Our can potentially regulate its expression in breast cancer cells. results indicated that PTPRG inhibited breast tumor formation Materials and Methods: To investigate the function of PTPRG in vivo; PTPRG may up-regulate p21cip and p27kip proteins in vivo, athymic nude mice were implanted with MCF-7 cells through the ERK1/2 pathway. This study also showed overexpressing PTPRG. For in vitro study, protein levels of cell methylation-mediated silencing of PTPRG in breast cancer cycle regulators, cell cycle re-entry, and the phosphorylation cell lines. These data indicate that PTPRG exhibits the levels of extracellular signal-regulated protein kinases 1/2 characteristics of a breast tumor suppressor. (ERK1/2) were examined. In addition, methylation assays were conducted to investigate the epigenetic modification on the Tyrosine phosphorylation is an important cellular process in promoter of PTPRG. Results: Athymic nude mice bearing multicellular eukaryotes (1). Two major groups of enzymes are MCF-7 cells overexpressing PTPRG showed a reduction in involved in regulating tyrosine phosphorylation: protein tumor burden in comparison to animals implanted with MCF- tyrosine phophatases (PTPs) and protein tyrosine kinases 7 cells transfected with vector alone. When these two cell lines (PTKs). A number of PTKs have been functionally were studied in an in vitro system, elevated mRNA and protein characterized and several PTK inhibitors have been studied in levels of cell cycle regulators, p21cip and p27kip were detected clinical trials for their therapeutic values (2, 3). PTPs, however, in MCF-7 cells overexpressing PTPRG compared to cells are less well-studied, although their involvement in cell proliferation and cell transformation has been reported (4, 5). PTPs can be classified into four distinct families based on the amino acid sequences of their catalytic domains (1, 6). *Both authors contributed equally to this work. The class I, cysteine-based PTP family can be further classified into two subfamilies depending on their cellular Correspondence to: Young C. Lin, DVM, Ph.D., 1925 Coffey Road, localization, namely non-transmembrane PTP subtype Columbus, OH 43210-1093, U.S.A. Tel: +1 6142929706, Fax: +1 6142926473, e-mail: [email protected] (NTPTPs) and receptor–like PTP subtype (RPTPs). The NTPTPs contain only intracellular motif with one catalytic Key Words: PTPG, breast cancer, tumor suppressor, cell cycle domain (PTP domain), whereas RPTPs contain extracellular, regulators. transmembrane and intracellular motifs with one or more PTP

0250-7005/2010 $2.00+.40 1937 ANTICANCER RESEARCH 30: 1937-1946 (2010) domains. Reports have shown that PTPs play important roles Materials and Methods in transformation and tumorigenesis. For examples, SHP2, PTP1B, and low molecular weight PTP (LMW-PTP) Mouse xenograft model. Female NCr-nu/nu athymic mice at 5-7 overexpressed in cancer cells promote oncogenesis and tumor weeks of age were purchased from the National Cancer Institute (Frederic, MD, USA). The animals received food and water (ad progression (7, 8). However, reports have also demonstrated libitum) and were housed in a pathogen-free environment. All animal the possible mechanisms of how PTPs repress cell growth studies were performed according to guidelines approved by The through cell cycle regulation. PTPH1/PTPN3 suppresses cell Ohio State University ILACUC. Sub-confluent M7-PCR and M7- cycle progression through dephosphorylation of valosin- PTPRG cells were harvested, and cell viability was determined by containing protein (VCP/p97), an ATPase that mediates trypan blue exclusion assay. Single cell suspension (5×106 cells/ml) proteasomal degradation (9). Overexpression of protein of >90% viability was washed using PBS, mixed in Matrigel (10 tyrosine phosphatase receptor-type O (PTPRO) can delay the mg/ml), and injected into mice subcutaneously. In addition, a 100 day-release pellet containing 10 mg of estradiol-17β (E , 6.0 re-entry of lung cancer cells into the cell cycle (10). Rat PTP 2 μg/day/kg body weight) (Innovative Research of America, FL, USA) η (PTPRJ) exerts growth inhibition activity through was implanted subcutaneously to facilitate tumor formation (18). kip stabilization of -dependent kinase inhibitor p27 (11). Mice were weighed and tumor formation was monitored by Receptor type protein tyrosine phosphatase gamma (PTPRG) measuring tumors with calipers every 3 days. Tumor volumes were is an RPTP and is broadly expressed in many different organs calculated using the formula, L × W2/2 = mm3, where L is the length (12). PTPRG located at 3p14.2 has been and W is the width of the tumor (18). The mice were sacrificed on implicated as a candidate tumor suppressor gene based on day 55 and a whole-body tumor scan was performed using a 4.7 Tesla/40cm magnetic resonance imaging (MRI) system controlled by loss of heterozygosity studies in renal cancer (13). Reduced a Bruker Advance Console, at the OSU Animal MRI Facility. Tumors PTPRG expression in lung, ovarian and breast cancer cells were processed for histologic analysis and evaluated by the OSU has been reported (14-16). However, little is known about the Histology/Immunohistochemistry Core service. functions and regulation of PTPRG and how it regulates cell cycle events in breast cancer cells. Cell lines and cell treatment. MCF-7, SK-Br-3, MDA-MB-453 and During cancer development and progression, tumor MCF-10A were purchased from American Type Culture Collection, VA. suppressor are increasingly silenced, leading to growth MCF-10A was maintained in phenol red-free low calcium (0.04 mM) advantage in the cancerous cells. One avenue of gene silencing Dulbecco’s modified Eagle’s medium and Ham’s F12 Medium (1:1) (DMEM/F12) supplemented with 10% (v/v) Chelex-100 (Bio-Rad is by promoter epigenetic modulations. Promoters of PTPRO Laboratory, CA, USA) –treated fetal calf serum (FCS), 10 μg/ml human and SHP1 were methylated in lung cancer and lymphoma, recombinant epidermal growth factor (EGF) and 100 μg/ml cholera respectively, and the methylation levels were shown to be toxin (both from Invitrogen, CA, USA). MCF-7 cells stably transfected correlated with the loss of gene transcripts in these cancer with full length PTPRG cDNA (M7-PTPRG cells) or stably transfected samples (10, 17). In this study, we also investigated the with pCR3.1 vector alone (M7-PCR cells) were established and selected methylation patterns in the PTPRG promoter region in breast by use of G-418 as previously described (5). M7-PTPRG, M7-PCR, cancer cell lines and examined how the methylation and gene MCF-7, SK-Br-3, and MDA-MB-453 cells were maintained in DMEM/F-12 (1:1) supplemented with 5% (v/v) FBS (final expression patterns were altered by deoxy-5-azacytidine (DAC) concentration). For methylation assay, cells were treated with 10 μM and trichostatin A (TSA) treatments. DAC (Sigma, MO, USA), 1 μM TSA (Sigma), the combined treatment Previously, we have established and characterized MCF-7 of DAC and TSA (10 μM and 1 μM, respectively), or vehicle for 4 days. cells stably transfected with full length PTPRG cDNA (M7- PTPRG cells) (5). PTPRG levels for both mRNA and protein Western blot assays. Whole cell lysate from M7-PTPRG and M7- expression were showed to be elevated in the PTPRG- PCR cells were separated by SDS-PAGE, and proteins were transfected cells when compared to cells transfected with transferred onto polyvinylidene fluoride (PVDF) membranes. vector alone (M7-PCR cells). Overexpression of PTPRG Enhanced chemiluminescence (ECL) plus™ reagents (Amersham reduces anchorage-independent colonization and proliferation Biosciences, NJ, USA) were used to detect the signals of the following proteins: anti-PTPRG (5), anti-p27kip (BD Biosciences, of breast cancer cell lines (5), indicating that PTPRG may CA, USA), anti-actin (Santa Cruz Biotechnology, CA, USA), anti- function as an important modulator of tumorigenesis in p21cip, anti-ERK1/2, and anti-phosphorylated-ERK1/2 (Cell human breast. In this report, we provide additional in vivo Signaling Technology, MA, USA). The chemiluminescent signals evidence to support the tumor suppressive function of PTPRG were acquired by an LAS-3000 imaging system (Fujifilm, CT, and how it may play a role in the cell cycle re-entry USA). The splicing of images is for illustration purposes. The most pathways. The stated hypotheses for this study are as follows: representative images are shown. i) PTPRG inhibits tumor formation in vivo; ii) overexpression GEArray assays. The cDNA GEArray (SuperArray, MD, USA) Q of PTPRG delays cell cycle re-entry by increasing the levels series Human Breast Cancer and Estrogen Receptor Signaling Gene cip kip of cell cycle regulators p21 and p27 ; iii) the loss of Array (HS-020) and Human Signal Transduction in Cancer Gene PTPRG in human breast cancer is due to epigenetic Array (HS-044) were used according to the manufacturer’s modifications in the PTPRG promoter. instruction manual. In brief, total RNA was isolated from M7-PCR

1938 Shu et al: PTPRG in Breast Carcinogenesis and M7-PTPRG cells using TRIZOL® Reagent (Invitrogen). One recognition site in the franked region is methylated (Figure 6A). The μg of total RNA was biotin-16-dUTP-labeled by the AmpoLabeling- PCR reaction was performed in a 10 μl volume containing 0.1 μM LPR (SuperArray). The membranes described above were primer, 0.8 mM dNTP, 2 μM MgCl2, 1X PCR buffer II, 0.25 U Taq hybridized with the biotin-labeled probe overnight. The hybridized polymerase (Applied Biosystems, CA, USA) and 10–3 mCi 32P- array was washed according to manufacture’s directions and the dCTP (Amersham Biosciences) at 94˚C for 1 min, 53˚C for 1 min, chemiluminescent signals were acquired by the LAS-3000 imaging and 72˚C for 1 min for 45 cycles. The radiolabeled PCR products system (Fujifilm). The images were analyzed using the GEArray were digested with TaqI at 65˚C for 4 h. The mock-digested control Analyzer software. The thresholds for significant up- or down- and the enzyme-digested products were run in parallel on regulation were set at >1.5 or <0.66, respectively, when the polyacrylamide gels and subjected to autoradiography. hybridization intensity of the treated group was compared to the control group. Selected gene probes with signal intensities less than Statistical analysis. The error bars in each figure mark±one standard 15% of the internal control gene (cyclophilin A) were removed from deviation from the mean values unless otherwise noted. An asterisk further consideration. indicates significant changes at the p<0.05 level using Student’s t-test.

Bromodeoxyuridine (BrdUrd) incorporation assay and cell cycle Results analyses. M7-PCR and M7-PTPRG cells were seeded in chamber slides and synchronized by culturing in medium containing 0.2% (v/v) FCS for 30 h prior to releasing them into complete growth medium for PTPRG inhibited tumor formation in vivo. To characterize 4 h. Cells were incubated with 10 μM BrdUrd (Sigma) for 2 h, fixed PTPRG anti-tumor effects, MCF-7 cells were stably with 70% (v/v) ethanol, denatured with 2 M HCl and incubated with transfected with full length PTPRG cDNA (M7-PTPRG) and anti-BrdUrd antibody (Santa Cruz Biotechnology) for 1 h followed by the mRNA and protein expression were evaluated (5) (Figure staining with horseradish peroxidase-conjugated secondary antibody 1A and Figure 1B). M7-PTPRG cell had significantly higher (Vector Laboratories, CA, USA). Diaminobenzidine (Vector Laboratories) was added to the stained cells followed by eosin Y PTPRG mRNA and protein levels (3.9- and 3.0-fold of (Sigma) counterstaining to permit visualization of BrdUrd increase, respectively) in comparison to MCF-7 transfected incorporation. Cells were photographed with a digital camera under a with the empty vector (M7-PCR cells). Both cell types were phase-contrast microscope. A hundred cells were counted within a injected subcutaneously in athymic nude mice (Figure 2A). randomly selected area and only the notably stained cells were counted Tumor formation was found in all M7-PCR inoculated sites, as positively staining cells. with some appearing as early as 2 weeks after inoculation (8/8). Once detected, the size of these tumors increased Real-time RT-PCR. Reverse transcription of 1 μg RNA was rapidly thereafter (Figure 2A and Table I). In contrast, there performed with random hexamers and Moloney murine leukemia virus reverse transcriptase (Invitrogen) in a 50 μl reaction volume. was only one tumor formed at one of the M7-PTPRG For real-time PCR, the 25 μl reaction mix consisted of 5 μl RT- inoculation sites (1/12). Histopathology examination reactant, 12.5 μl 2X SYBR® Green PCR Master Mix (Applied confirmed that these tumors were adenocarcinomas (data not biosciences, CA, USA) and 1.4 μM primer for PTPRG: F5’-GCA shown). MRI imaging show no sign of tumor metastasis in TCC TCT GCC ACA TAC TAC G-3’and R5’-TCA TCT TCT GCC the inoculated mice (Figure 2B). These highly repeatable and cip AAG CTC TGG T-3’; for p21 : F5’-GGA AGA CCA TGT GGA significant results demonstrate that overexpression of PTPRG CCT GT-3’ and R5’-GGA TTA GGG CTT CCT CTT GG-3’, and inhibited tumor formation in the athymic mouse model. for p27kip: F5’- CCG GCT AAC TCT GAG GAC AC-3’ and R5’-AGA AGA ATC GTC GGT TGC AG-3’. For 36B4 real-time PCR, 25 μl reaction mix consisted of 5 μl RT-reactant, 12.5 μl 2X Overexpression of PTPRG delayed cell cycle re-entry and TaqMan® Universal PCR Master Mix (Applied Biosciences), 1.125 increased p21cip and p27kip protein levels. To understand μM forward primer (5’-CTG GAG ACA AAG TGG GAG CC-3’), how PTPRG may reduce tumor formation observed in our in 0.75 μM of reverse primer (5’-TCG AAC ACC TGC TGG ATG AC- vivo system, we performed two types of gene expression 3’), and 0.625 μM TaqMan probe (FAM-ACG CTG CTG AAC ATG microarrays (Human Breast Cancer and Estrogen Receptor CTC AAC ATC TCC-TAMRA). RT-reactant was amplified 45 cycles Signaling Gene Array and Human Signal Transduction in after 95˚C 10 min incubation followed by 95˚C for 30 s and 60˚C for 1 min. The relative mRNA expression levels were calculated using Cancer Gene Array) to identify differentially expressed genes 2–ΔΔCT method described by Livak and Schmittgen (19). that may be involved in this tumor suppressive function. As depicted in Table II, our analyses revealed 54 genes, Combined bisulfite restriction analysis (COBRA). Genomic DNA including p27kip, p21cip, cyclin A1, p57kip, TRAILR2, and was extracted using Wizard® SV Genomic DNA Purification wig-1, which showed at least a two-fold difference in System (Promega, WI, USA) according to the manufacturer’s expression between M7-PTPRG cells and M7-PCR cells. guidelines. One μg genomic DNA was bisulfite-converted using the This result indicates that PTPRG may influence a wide EZ DNA Methylation Kit (Zymo Research Corporation, CA, USA). spectrum of cellular functions, including cell cycle Modified DNA (~10 ng) was used as templates for PCR with a primer set (F5’-GGA GGT AAG AAT TTA TTT AAT AAG TTT-3’ regulation, as implied by the impressive impact it has on and R5’-CAC AAT TCC AAT AAC CTC CA-3’) which flanks the lowering the tumor burden in our athymic mouse study. PTPRG promoter region containing one TaqI restriction site if the Other evidence that PTPRG may be involved in the

1939 ANTICANCER RESEARCH 30: 1937-1946 (2010)

Figure 2. PTPRG inhibits tumor formation in a mouse xenograft model. A: The inoculate sites of M7-PTPRG and M7-PCR cells; 5×105 cells were injected subcutaneously. E2 pellets were implanted subcutaneously in the back of the mouse. B: The MRI images from day 55. The left side of the mouse is at the bottom of the image.

of p21cip and p27kip in M7-PTPRG cells were 1.8- and 2.3- fold higher than in M7-PCR cells, respectively. Next we Figure 1. Over-expressing PTPRG in MCF-7 cells. A: PTPRG mRNA tracked the increase in p21cip and p27kip protein expressions expression in M7-PTPRG and M7-PCR cells were determined using using Western blotting analysis. Figure 3B shows a 3-fold real-time RT-PCR. Target gene expression levels were normalized to cip kip 36B4. The normalized PTPRG was shown to be significantly different increase in p21 and a 2-fold increase in p27 protein in between M7-PTPRG and M7-PCR cells at *p<0.05. The error bars M7-PTPRG cells when compared to M7-PCR cells. These correspond to the standard deviation. B: PTPRG protein in M7-PTPRG data indicate that the levels of p21cip and p27kip (both mRNA and M7-PCR cells. Western blot images for PTPRG and actin are and protein) were elevated in the human breast cancer cells presented in the insert. Histograms depict the quantitative normalized over-expressing PTPRG. level of PTPRG in the two cell types. The error bars correspond to the cip kip standard error. The elevated levels of p21 and p27 in M7-PTPRG cells along with our previously data demonstrating that PTPRG prolonged breast cancer cell doubling time (5) prompted us to investigate if PTPRG plays a role in cell cycle important process of cell cycle regulation comes from reports progression. BrdUrd incorporation assays were conducted to demonstrating that PTPRO, a related PTP member, can delay evaluate if PTPRG can indeed delay the process of cell cycle cell cycle re-entry in lung cancer cells (10), and that PTPη is re-entry. After M7-PTPRG and M7-PCR cells were capable of suppressing the neoplastic phenotype of synchronized by serum starvation, complete growth medium retrovirally transformed thyroid cells through the was introduced into the cell culture, and the numbers of cells stabilization of p27kip1 (11). With this supporting evidence, containing newly synthesized DNA were counted. As shown we decided to study cell cycle regulation genes, p21cip and in Figure 4A and Figure 4B, the number of cells containing p27kip, shown to be expressed at 3.7- and 2.6-fold higher in newly synthesized DNA was significantly lower (~50%) in M7-PTPRG than in M7-PCR cells, respectively (Table II). M7-PTPRG cells than M7-PCR cells. This indicates that the These mRNA expression data were further validated with the presence of PTPRG delays the re-entry of arrested cells back real-time RT-PCR analyses (Figure 3A). In three independent into the cell cycling process, a potentially important real-time RT-PCR experiments, the mRNA expression levels mechanism via which PTPRG can inhibit tumor growth.

1940 Shu et al: PTPRG in Breast Carcinogenesis

Table I. Tumor formation on the sites inoculated with M7-PTPRG or Table II. Overexpression of PTPRG results in alteration of gene M7-PCR cells. expression.

Cell Inoculation Day 0 Day 15 Day 19 Day 22 Day 28 Gene name M7-PCR M7-PTPRG M7-PTPRG Group U/D site /M7-PCR

M7-PTPRG 1 – – – – – p21/Cip1 1.445E-1 5.346E-1 3.700E+0 (A) U 2––– – –p27/Kip1 1.564E-1 4.133E-1 2.642E+0 (B) U 3––– – –p57/Kip2 4.760E-1 9.840E-1 2.067E+0 (B) U 4––– – –WIG1 4.215E-1 1.219E+0 2.891E+0 (C) U 5––– – –Cyclin A1 9.932E-1 3.726E-1 3.751E-1 (D) D 6––– – –GABRP 1.389E-2 8.009E-2 5.767E+0 (D) U 7––– – –C3/Lipophilin 6.013E-1 3.858E-1 6.415E-1 (D) D 8 – – – +++ ++++ Catenin 4.903E-1 3.004E-1 6.127E-1 (D) D 9––– – –TRAILR2 4.172E-1 1.067E+0 2.556E+0 (E) U 10 – – – – – Cox-2 1.984E-1 3.771E-2 1.900E-1 (E) D 11 – – – – – Fas/Apo-1 1.067E-1 3.725E-2 3.491E-1 (F) D 12 – – – – – WISP1 9.959E-2 3.405E-2 3.419E-1 (G) D PTCH 6.056E-2 1.117E-1 1.844E+0 (H) U M7-PCR 1 – + ++ +++ SAC BMP2 1.171E-1 2.601E-2 2.220E-1 (H) D 2 – ++ +++ +++ SAC 3 – +++ ++++ ++++ SAC This table provides a list of selected genes (14/109 genes) identified by 4 – +++ ++++ ++++ SAC GEArray to be differentially expressed between M7-PTPRG and M7- 5––++ ++PCR cells. Normalized microarray signal intensities (the ratios of gene 6––– – +of interest to internal controls) are given for each cell type and for their 7 – + ++ +++ ++++ comparison. The pathway information listed under the column “Group” 8 – + + ++ ++++ is the most relevant pathway for the listed gene as described by the array manufacturer (SuperArray). The group names are as follows: (A) –: No tumor formation, +: tumor was 0-100 mm3, ++: tumor was 101- Androgen pathway; (B) TGF beta pathway; (C) DNA damage/p53 200 mm3, +++: tumor was 201-300 mm3, ++++: tumor was > 300 mm3, pathway; (D) breast cancer-associated gene; (E) Cox-2 pathway; (F) SAC: mouse was sacrificed because tumor was >800 mm3 or mouse lost DNA damage/p53 pathway; (G) Wnt pathway; (H) Hedgehog pathway. ≥20% of body weight. U/D indicates induction (U) or reduction (D) when compared to the control gene (>1.5 or <0.66, respectively).

PTPRG reduced the phosphorylation levels of ERK1/2. Trapasso et al. have shown that PTPRJ/PTPη suppresses cancerous cell line MCF-10A as previously described (14). the neoplastic phenotype of retrovirally transformed thyroid The ratio of observed to expected CpG in PTPRG is 0.81, cells through the stabilization of p27kip (11). ERK1/2 are suggesting that CpG island exists in the PTPRG gene (21). involved in the p27kip degradation pathways and have been To examine the methylation pattern in the PTPRG promoter proposed to be either direct or indirect substrates of PTPH1 region, COBRA assays using TaqI restriction enzyme were (9). In addition, the MAP kinase pathway has been shown performed on DNA isolated from these breast cancer cell to accelerate p27kip degradation (20). To test if the MEK lines. We observed differential methylation in one of the and ERK1/2 pathway is involved in the up-regulation of enzyme recognition sites which is 18 bp upstream of the p21cip and p27kip observed in M7-PTPRG cells, the transcriptional start site of PTPRG in all three breast cancer phosphorylation levels of ERK1/2 were examined using cell lines (Figure 6A and Figure 6B), but not in MCF-10A. Western blot analyses. Importantly, the phosphorylation levels of ERK1/2 were less in the M7-PTPRG cells when PTPRG expression is reduced in breast cancer cell lines and compared to the M7-PCR cells (Figure 5), indicating that can be re-activated with TSA and DAC treatments. To affirm ERK1/2 may be involved in PTPRG-related signaling that the differential methylation observed in the PTPRG pathways. promoter may indeed lead to decrease expression of PTPRG in human breast cancer cell lines, we treated SK-Br-3 and Aberrant methylation pattern in PTPRG CpG island in breast MCF-7 cells with DAC alone, TSA alone or combined cancer cells. To evaluate PTPRG expression in a quantitative treatment. Subsequently we performed real-time RT-PCR manner, we designed and performed real-time RT-PCR on analysis to evaluate how epigenetic treatments altered several breast cancer cell lines. PTPRG expression is PTPRG expression in the treated samples. As shown in significantly lower in breast cancer cell lines MCF-7, MDA- Figure 6C, PTPRG expression levels were significant higher MB-453 and SK-Br-3 cells when compared to the non in cells treated with DAC alone, TSA alone and their

1941 ANTICANCER RESEARCH 30: 1937-1946 (2010)

Figure 3. Overexpression of PTPRG increased p21cip and p27kip levels. A: Real-time RT-PCR for p21cip and p27kip in M7-PTPRG and M7-PCR cells. Target gene expression levels were normalized to 36B4 in M7-PTPRG and M7-PCR cells. M7-PTPRG had significantly higher levels of p21cip and p27kip expression at p<0.05. B: Western blot assays for p21cip and p27kip. Actin served as a loading and normalization control. C: Densitometric analyses of Western blot images revealed higher levels of p21cip and p27kip in M7-PTPRG cells in comparison to M7-PCR cells. The difference for p21cip was significant (p<0.05) but not that for p27kip.

combination when compared to the controls in SK-Br-3 Discussion cells. Similar results were observed in MCF-7 cells (Figure 6C). These results indicate that both DNA methylation and Protein phosphatases can affect cell signaling pathways histone modification play a role in the silencing of PTPRG in including cellular events important in tumorigenesis. However, human breast cancer. very little is known about the functions and the regulation of

1942 Shu et al: PTPRG in Breast Carcinogenesis

Figure 5. Overexpressing PTPRG decreased ERK 1/2 phosphorylation levels. Western blot images show the levels of phosphorylated ERK1/2 (p-ERK1/2) and total ERK1/2 in M7-PTPRG (lane 1 and 2) and M7- PCR (lane 3 and 4) cells.

this class of enzyme, and even less is known about potential ligands for RPTPs. Some of the known functions for PTPRG include its roles in hematopoietic differentiation (22, 23), sensory neuron development (24), and retardation of NGF- induced neurite outgrowth (25). We have previously reported the tumor suppressive properties of PTPRG (5). Evidence supporting this property include the observation of reduced PTPRG expression in primary breast cancer samples and breast cancer cell lines, as well as the inhibition of cell proliferation and anchorage-independent growth in vitro in cells overexpressing PTPRG. Here we report that over-expression of PTPRG can reduce tumor formation in vivo and delay cell cycle re-entry by increasing the level of cell cycle regulators p21cip and p27kip. We also provide data in support of promoter methylation in PTPRG in human breast cancer cell lines and that the silenced gene can be reactivated by treatments targeting both DNA methylation and histone modifications. The in vivo results reported in this paper further strengthen our hypothesis that PTPRG may function as a tumor suppressor in human breast cancer development. Over-expression of PTPRG dramatically reduced the ability of MCF-7 cells to form tumor under the stimulation of estradiol in athymic mice. In addition to PTPRG, other PTP members have been shown to possess anti-tumorigenic properties in vivo. Rat PTPRJ was able to suppress the growth of human thyroid carcinoma in a mouse xenograft model (26). Lamprianou and co-workers have Figure 4. Overexpression of PTPRG delayed cell cycle re-entry. A: Immunohistochemical detection of incorporated BrdUrd in M7-PTPRG generated PTPRG null mice to examine its role in sensory and M7-PCR cells after cell synchronization by serum starvation and neuron development/differentiation (24). PTPRG null mice released into growth medium for 4 hours. The BrdUrd-negative cells are were found to develop normally but exhibited a low level of identified by dotted arrows and the BrdUrd-positive cells by the solid behavioral changes. Thus far, no neoplastic phenotype has been arrows. B: Percentage of BrdUrd-positive M7-PCR and M7-PTPRG cells. reported. It would indeed be interesting to challenge these One hundred cells were counted within a randomly selected area and only the notably stained cells were counted as positively staining cells. Cells animals with estradiol or other tumor-inducing agents to overexpressing PTPRG had significantly fewer BrdUrd-positive cells than evaluate their tumor susceptibility in comparison to their the mock-transfected cells. The asterisk (*) indicates p<0.05. normal counterparts.

1943 ANTICANCER RESEARCH 30: 1937-1946 (2010)

Figure 6. Methylation patterns at the promoter region of PTPRG in human breast cancerous and non-cancerous cells. A: Genomic plot of PTPRG promoter, CpG islands and TaqI restriction sites. Thick arrow indicates transcription start site. Thin arrows indicate primer locations. B: COBRA results showing the TaqI-digested PCR products. Broad arrow indicates undigested PCR products (89 bps) due to unmethylated DNA; short arrows indicate digested products (50 and 41 bps) due to methylated DNA; long arrows indicate non-specific products that may be due to degraded DNA. (–): undigested lane; (+): digested lane; SSSI: SssI methylase-treated DNA was used as a positive control; Blood; DNA extract from blood was used as a negative control. C: PTPRG expression determined by real-time RT-PCR after SK-Br-3 and MCF-7 cells was treated with DAC, TSA or the combined treatment. All three treated cells showed significant PTPRG expression in comparison to untreated control cells.

Cell cycle regulation by PTPs has been demonstrated MEK in M7-PTPRG cells (data not shown), suggesting that previously (10, 27, 28). PTPRO was shown to delay cell the effects of PTPRG on ERK1/2 are not due to a decreased cycle re-entry after lung cancer cells were released from cell- activation of MEK. The activation of the ERK1/2 kinase kip cycle arrest (10). Rat PTPRJ can cause G1 growth arrest and signaling pathway can increase p27 phosphorylation and this is shown to be mediated through the increase of p27kip its subsequent clearance by proteasome degradation (20). It by reducing its proteasome-dependent degradation in human is possible that PTPRG may de-phosphorylate ERK1/2 thyroid cancer (11). In this study, we also found that the through a MEK-independent pathway in human breast cancer mRNA and protein levels of p21cip and p27kip are elevated cells thereby exerting its effect on cell cycle re-entry. There in M7-PTPRG cells (real-time RT-PCR and Western blotting may well be other pathways at work by which PTPRG exerts data as shown in Figure 3A and Figure 3B). We also its tumor suppressive functions. Other forms of global observed that PTPRG can delay cell cycle re-entry after analyses such as phosphorylation array or pathway-specific breast cancer cells were synchronized by media containing array may produce illuminating results on identifying low fetal bovine serum. ERK1/2 has been proposed to be PTPRG functions in health and disease. either a direct or an indirect substrate for rat PTPRJ (27), and There are many tumor suppressor genes shown to be the phosphorylation levels of ERK1/2 have been associated silenced by promoter methylation leading to tumorigenesis. with PTP-SL activity (29). Our results showed that DNA methylation, as well as other epigenetic alterations, phosphorylation levels of ERK1/2 were lower in M7-PTPRG can be reversed by epigenetic treatments, thus opening new cells when compared to M7-PCR cells. Interestingly, we did approaches to treat cancer (30, 31). Here, we present not observe any changes in the levels of phosphorylated- evidence that aberrant DNA methylation occurs at the

1944 Shu et al: PTPRG in Breast Carcinogenesis promoter region of PTPRG. We found that DAC alone, 3 Kantarjian HM, Giles F, Gattermann N, Bhalla K, Alimena G, TSA alone, and their combined treatment were all able to Palandri F, Ossenkoppele GJ, Nicolini FE, O’Brien SG, Litzow reactivate PTPRG expression. These results indicate that M, Bhatia R, Cervantes F, Haque A, Shou Y, Resta DJ, Weitzman A, Hochhaus A and le Coutre P: Nilotinib (formerly both DNA methylation (as evident from COBRA data) and AMN107), a highly selective bcr-abl tyrosine kinase inhibitor, histone modification (as evident indicate from treatment is effective in patients with philadelphia chromosome-positive outcomes of TSA alone and the combined treatment) are chronic myelogenous leukemia in chronic phase following important in the silencing of PTPRG in human breast imatinib resistance and intolerance. Blood 110(10): 3540-3546, cancer. At present, the methylation status in the region 2007. explored by the COBRA assay did not reflect the level of 4 Mustelin T and Tasken K: Positive and negative regulation of T- PTPRG expression in MCF-7 and SK-Br-3 (data not cell activation through kinases and phosphatases. Biochem J shown). This is indicative that other methylation events 371(Pt 1): 15-27, 2003 5 Liu S, Sugimoto Y, Sorio C, Tecchio C and Lin YC: Function PTPRG may be present in the promoter and detailed analysis of estrogenically regulated protein tyrosine phosphatase methylation analysis needs to be performed to identify gamma (PTPgamma) in human breast cancer cell line MCF-7. these locations. Oncogene 23(6): 1256-1262, 2004. In summary, we found that the presence of PTPRG 6 Andersen JN, Mortensen OH, Peters GH, Drake PG, Iversen significantly reduces tumor formation in athymic mice. Our LF, Olsen OH, Jansen PG, Andersen HS, Tonks NK and data revealed elevated levels of p21cip and p27kip (mRNA Moller NP: Structural and evolutionary relationships among and protein) in PTPRG overexpressing human breast cancer protein tyrosine phosphatase domains. Mol Cell Biol 21(21): 7117-7136, 2001. cells, and the presence of PTPRG delays cell cycle re-entry. 7 Jiang ZX and Zhang ZY: Targeting PTPs with small molecule Lastly, we identified aberrant methylation in the PTPRG inhibitors in cancer treatment. Cancer Metastasis Rev 27(2): promoter region in breast cancer cell lines and the expression 263-272, 2008. of PTPRG can be reactivated by treating the cells with DAC 8 Marzocchini R, Malentacchi F, Biagini M, Cirelli D, Luceri C, and TSA. Our data provide evidence for the epigenetic Caderni G and Raugei G: The expression of low molecular silencing of PTPRG in human breast cancer. Therefore, it weight protein tyrosine phosphatase is up-regulated in 1,2- falls into a unique group of tumor suppressor genes that can dimethylhydrazine-induced colon tumours in rats. Int J Cancer 122(7): 1675-1678, 2008. potentially be targeted by epigenetic therapies to reinstate 9 Zhang SH, Liu J, Kobayashi R and Tonks NK: Identification of their functions. More needs to be done to firmly establish the the cell cycle regulator VCP (p97/CDC48) as a substrate of the tumor suppressive functions of PTPRG in breast cancer and band 4.1-related protein-tyrosine phosphatase PTPH1. J Biol some of the studies proposed in the discussion are already Chem 274(25): 17806-17812, 1999. under way. 10 Motiwala T, Kutay H, Ghoshal K, Bai S, Seimiya H, Tsuruo T, Suster S, Morrison C and Jacob ST: Protein tyrosine phosphatase Acknowledgements receptor-type O (PTPRO) exhibits characteristics of a candidate tumor suppressor in human lung cancer. Proc Natl Acad Sci USA 101(38): 13844-13849, 2004. We thank Dr. Tim Huang (The Ohio State University) for providing 11 Trapasso F, Iuliano R, Boccia A, Stella A, Visconti R, Bruni P, assistance in the COBRA assays. We also thank Dr. Petra Baldassarre G, Santoro M, Viglietto G and Fusco A: Rat protein Schomalbrock (The Ohio State University) for performing MRI and tyrosine phosphatase eta suppresses the neoplastic phenotype of data analysis, Chieh-Ti Kuo for figure preparation, and Dr. Donna retrovirally transformed thyroid cells through the stabilization of Kusewitt (The Ohio State University) for histopathological p27(Kip1). Mol Cell Biol 20(24): 9236-9246, 2000. examination. This project was supported by the Cooperative Human Tissue Network of The Ohio State University Hospital, Department 12 Tsukamoto T, Takahashi T, Ueda R, Hibi K, Saito H and of Defense Breast Cancer Research Program Grant (DAMD 8140), Takahashi T: Molecular analysis of the protein tyrosine and National Institutes of Health Grants (CA94718 and R01 phosphatase gamma gene in human lung cancer cell lines. ES015212). Cancer Res 52(12): 3506-3509, 1992. 13 Lubinski J, Hadaczek P, Podolski J, Toloczko A, Sikorski A, McCue P, Druck T and Huebner K: Common regions of References deletion in chromosome regions 3p12 and 3p14.2 in primary clear cell renal carcinomas. Cancer Res 54(14): 3710-3713, 1 Alonso A, Sasin J, Bottini N, Friedberg I, Friedberg I, Osterman 1994. A, Godzik A, Hunter T, Dixon J and Mustelin TO: Protein 14 Zheng J, Kulp SK, Zhang Y, Sugimoto Y, Dayton MA, Govindan tyrosine phosphatases in the . Cell 117(6): 699- MV, Brueggemeier RW and Lin YC: 17 Beta-estradiol-regulated 711, 2004. expression of protein tyrosine phosphatase gamma gene in 2 Modi S, Seidman AD, Dickler M, Moasser M, D’Andrea G, cultured human normal breast and breast cancer cells. Anticancer Moynahan ME, Menell J, Panageas KS, Tan LK, Norton L and Res 20(1A): 11-19, 2000. Hudis CA: A phase II trial of imatinib mesylate monotherapy in 15 van Niekerk CC and Poels LG: Reduced expression of protein patients with metastatic breast cancer. Breast Cancer Res Treat tyrosine phosphatase gamma in lung and ovarian tumors. Cancer 90(2): 157-163, 2005 Lett 137(1): 61-73, 1999.

1945 ANTICANCER RESEARCH 30: 1937-1946 (2010)

16 Vezzalini M, Mombello A, Menestrina F, Mafficini A, Della 25 Shintani T, Maeda N and Noda M: Receptor-like protein tyrosine Peruta M, van Niekerk C, Barbareschi M, Scarpa A and Sorio phosphatase gamma (RPTPgamma), but not PTPzeta/RPTPbeta, C: Expression of transmembrane protein tyrosine phosphatase inhibits nerve-growth-factor-induced neurite outgrowth in gamma (PTPgamma) in normal and neoplastic human tissues. PC12D cells. Dev Neurosci 23(1): 55-69, 2001. Histopathology 50(5): 615-628, 2007. 26 Iuliano R, Trapasso F, Le Pera I, Schepis F, Sama I, Clodomiro 17 Khoury JD, Rassidakis GZ, Medeiros LJ, Amin HM and Lai R: A, Dumon KR, Santoro M, Chiariotti L, Viglietto G and Fusco Methylation of SHP1 gene and loss of SHP1 protein expression A: An adenovirus carrying the rat protein tyrosine phosphatase are frequent in systemic anaplastic large cell lymphoma. Blood eta suppresses the growth of human thyroid carcinoma cell lines 104(5): 1580-1581, 2004. in vitro and in vivo. Cancer Res 63(4): 882-886, 2003. 18 Brady H, Desai S, Gayo-Fung LM, Khammungkhune S, McKie 27 Florio T, Arena S, Thellung S, Iuliano R, Corsaro A, Massa A, JA, O’Leary E, Pascasio L, Sutherland MK, Anderson DW, Pattarozzi A, Bajetto A, Trapasso F, Fusco A and Schettini G: The Bhagwat SS and Stein B: Effects of SP500263, a novel, potent activation of the phosphotyrosine phosphatase eta (r-PTP eta) is antiestrogen, on breast cancer cells and in xenograft models. responsible for the somatostatin inhibition of PC Cl3 thyroid cell Cancer Res 62(5): 1439-1442, 2002. proliferation. Mol Endocrinol 15(10): 1838-1852, 2001. 19 Livak KJ and Schmittgen TD: Analysis of relative gene 28 Harris DL and Joyce NC: Protein tyrosine phosphatase, PTP1B, expression data using real-time quantitative PCR and the 2(-delta expression and activity in rat corneal endothelial cells. Mol Vis delta C(T)) method. Methods 25(4): 402-408, 2001. 13: 785-796, 2007 20 Kawada M, Yamagoe S, Murakami Y, Suzuki K, Mizuno S and 29 Pulido R, Zuniga A and Ullrich AP: PTP-SL and STEP protein Uehara Y: Induction of p27Kip1 degradation and anchorage tyrosine phosphatases regulate the activation of the extracellular independence by ras through the MAP kinase signaling pathway. signal-regulated kinases ERK1 and ERK2 by association through Oncogene 15(6): 629-637, 1997. a kinase interaction motif. EMBO J 17(24): 7337-7350, 1998. 21 Takai D and Jones PA: Comprehensive analysis of CpG islands 30 Ghoshal K, Datta J, Majumder S, Bai S, Kutay H, Motiwala T and in human 21 and 22. Proc Natl Acad Sci USA Jacob ST: 5-Aza-deoxycytidine induces selective degradation of 99(6): 3740-3745, 2002. DNA methyltransferase 1 by a proteasomal pathway that requires 22 Lissandrini D, Vermi W, Vezzalini M, Sozzani S, Facchetti F, the KEN box, bromo-adjacent homology domain, and nuclear Bellone G, Mafficini A, Gentili F, Ennas MG, Tecchio C and localization signal. Mol Cell Biol 25(11): 4727-4741, 2005. sorio C: Receptor-type protein tyrosine phosphatase gamma 31 Mei S, Ho AD and Mahlknecht U: Role of histone deacetylase (PTPgamma), a new identifier for myeloid dendritic cells and inhibitors in the treatment of cancer (review). Int J Oncol 25(6): specialized macrophages. Blood 108(13): 4223-4231, 2006. 1509-1519, 2004. 23 Sorio C, Melotti P, D’Arcangelo D, Mendrola J, Calabretta B, Croce CM and Huebner K: Receptor protein tyrosine phosphatase gamma, ptp gamma, regulates hematopoietic differentiation. Blood 90(1): 49-57, 1997. 24 Lamprianou S, Vacaresse N, Suzuki Y et al: Receptor protein tyrosine phosphatase gamma is a marker for pyramidal cells and Received February 26, 2010 sensory neurons in the nervous system and is not necessary for Revised April 8, 2010 normal development. Mol Cell Biol 26(13): 5106-5119, 2006. Accepted April 15, 2010

1946