PTN Signaling: Components and Mechanistic Insights in Human Ovarian Cancer

Cell Culture

EOC cell lines (A1847, A2780, OVCAR3, OVCAR5 and SKOV3) [25–28] were grown in RPMI 1640 (Invitrogen), [supplemented with 10% FBS (Hyclone, Gibco), 2 mM L-glutamine (Invitrogen), 100 IU/mL penicillin (Invitrogen), and 100 µg/mL streptomycin (Invitrogen)]. The cell lines were maintained at 37 °C in a humidified atmosphere with 5% CO₂ [29].

Reagents and antibodies

Goat anti-PTN was obtained from Abcam. Mouse anti-phospho ERK1/2 and anti-ERK1/2 were obtained from Cell Signaling Technology. Mouse anti-PTPRZ1 was obtained from BD. HRP-conjugated anti-goat antibodies were obtained from Santa-Cruz Biotechnology. HRP-conjugated anti-mouse antibodies were obtained from Bio-Rad. Recombinant human PTN was obtained from R&D Systems. Mouse anti-β-actin antibody was obtained from Sigma.

Analysis of secreted PTN levels by direct ELISA

A1847, A2780, OVCAR3, OVCAR5 and SKOV3 cells were seeded at a cell density of 30,000 cells/100 µL per well in a 96-well plate in complete medium. After 24 h, the media was changed to serum free and the cells were starved for additional 48 h. The conditioned media was collected and spun down at 3,000×g for 5 min. The cell free supernatant was analyzed by direct ELISA as described previously.[9] Briefly conditioned media were incubated in Immunolon flat-bottomed 96-well plates overnight at 4 °C. Plates were then washed 3 times with phosphate buffered saline (PBS) and blocked with PBST (PBS with 0.1% Tween-20, supplemented with 1% bovine serum albumin) at room temperature for 2 h. Next, the plates were incubated with 100 µL/well of primary antibody (0.5 µg/mL biotinylated anti-PTN, R&D Systems) at 4 °C overnight, washed 3 times with PBST incubated with 100 µL secondary antibody (alkaline-phosphatase conjugated streptavidin, 200 mg/mL, 1:200, KPL) for 1 h at room temperature and washed 4 times with PBST. Bound proteins were detected using Blue-Phos microwell phosphatase substrate (KPL) following manufacturer’s directions and plates were read using an Infinite M200Pro plate reader (Tecan) at 620nm. Experiments were performed in triplicates. A standard curve was generated using 0 to 2 ng/mL rhPTN (recombinant human PTN, R&D Systems).

qRT-PCR and PCR array

To measure PTN and PTPRZ1 mRNA levels across cell lines, total RNA was isolated using TRIzol reagent as described previously [6]. Reverse transcription was performed to prepare cDNA [30] followed by specific target amplification (14 cycle PCR reaction) for each sample using 100 ng cDNA combined with the pooled TaqMan primers in Pre-Amp Mastermix (Life Technologies) using manufacturer’s protocol. Quantitative RT-PCR (qRT-PCR) was performed using the Fluidigm BioMark microfluidic real-time PCR platform. To quantify the mRNA levels of a specific gene, the cycle threshold (C_t) values were determined from the amplification curves for each sample and then normalized to the level of PPIA (internal control reference gene) in that sample. Gene expression profiles in cells treated with siRNA targeting GL2 or PTN mRNA were characterized using a RT² Profiler™ PCR Array (Qiagen PAHS027Z, commercially targeted array). Since we observed an apoptotic phenotype on PTN knockdown, therefore a panel of 36 survival-related genes (shortlisted by enrichment in biological functions of cell death and-survival using IPA as described previously [31, 32]) were analyzed (Supplementary Table S5) for differential expression. qRT-PCR was performed on a Bio-Rad CFX96 C1000 Touch real-time detection apparatus (Bio-Rad Laboratories) using a RT SYBR Green/Fluorescein qPCR Master Mix (SABiosciences). The conditions used for PCR amplification were: 15 s at 95 °C, 30 s at 60 °C, and 30 s at 72 °C for 45 cycles, and fluorescence was measured after each amplification cycle. To quantify the mRNA levels of a specific gene, the C_t values were determined from the amplification curves for each sample and then normalized to the mean C_t values of the five housekeeping genes (ACTB, GAPDH, HPRT1, B2M, and RPL13A) in each sample. Fold changes in gene expression were calculated using the ΔΔCT method using the web-based software (SuperArray Biosciences Corp) in PTN siRNA treated as compared to GL2 siRNA (negative control) treated A2780 cells.

Western blotting

Cell lysates were prepared in RIPA buffer (10 mM NaPi, pH 7.8, 60 mM NaCl, 1% Triton X100, 0.5% deoxycholic acid, 0.1% SDS, 10% glycerol, 25 mM β-glycerol phosphate, 50 mM sodium fluoride, 2 mM sodium pyrophosphate, 1 mM sodium orthovanadate) and analyzed by immunoblotting using primary antibodies specific for PTN (1:500), PTPRZ1 (B D, 1:250), ERK (1:250), pERK1/2 (1:2000), and β-actin (1:10,000). The blots were quantified by densitometry using Alpha View software, version 3.3 (Cell Biosciences).

siRNA reagents

PTN siRNA (Qiagen), PTPRZ1 siRNA (Qiagen), reagents were purchased from Qiagen. Each siRNA was targeted at a final concentration of 50nm that comprised of equimolar concentration of two independent siRNA sequences (PTN: product ID SI03093083 & SI00043750, PTPRZ1: product ID SI00044457 & SI00044471).

siRNA transfections in 96 well plates and cell viability

All siRNA transfections were done using the reverse transfection method as described previously [6]. Briefly, DharmaFECT-1 was diluted in reduced-serum media (OptiMEM, Invitrogen) and added to the siRNAs arrayed in v-bottom 96-well dilution plates using a bulk reagent microplate dispenser. The concentration of siRNA pools was 50 nM. The siRNA-lipid complexes were allowed to form for 30 min at room temperature followed by addition of cells in an antibiotic-free medium using a bulk reagent microplate dispenser (100 µL final volume per well). Following 72 h of incubation at 37 °C, cell viability was determined by using Cell Titer-Blue (CTB, Promega). The CTB reagent was diluted 3-fold in PBS prior to its addition to the assay plates (20 µL per well added using a bulk reagent microplate dispenser). Fluorescence intensity was measured by using an Infinite M200Pro plate reader (Tecan) 3 h following addition of the CTB reagent. Data were analyzed using the viability score method as described previously [6] where viability in test well was normalized to the viability in the negative control wells. Parameters which affect efficiency of transfection such as the dilution factor of the DharmaFECT-1 lipid reagent, the concentration of the siRNA, and the cell seeding density were optimized for each of the five cell lines as described previously [6].

siRNA treatment in 10 cm plates

Cells were seeded (4 × 10⁵ – 6 × 10⁵) in 10 cm plates overnight. Next day, media was changed to OptiMEM (Invitrogen) for 4 h. Transfection complexes were then added to the plate with final siRNAs (50 nM). After 48–72 h, cells were harvested and lysates were prepared for immunoblotting.

Apoptosis and Cell Cycle analysis

A2780 and OVCAR5 cells were transfected with GL2 (negative control) and PTPRZ1 siRNA in 96-well plates using the transfection method as described above for the 96-well format. After 72 h, cells were trypsinized and processed for apoptosis and cell cycle analysis using the Guava Nexin and Guava Cell Cycle assays, respectively, following the manufacturer’s instructions (Guava Technologies, Millipore).

Analysis of genomic data sets

The log₂ ratios for copy number and β-values for DNA methylation were downloaded from TCGA portal (http://tcga-data.nci.nih.gov/tcga/tcgaHome2.jsp) for a total of 494 cases for the PTPRZ1 gene. A mean β-value was calculated since the results from 2 probes were provided

Serum and Tissue samples from ovarian cancer patients

All tumor (high-grade serous ovarian cancer) and non-tumor serum and tissue samples (that were homogenized in lysis buffer and used for immunoblotting experiments) were obtained from the KU Cancer Center’s Biospecimen Repository Core Facility (BRCF). Tumor grades and histological subtypes were obtained from the pathology reports without reference to the patient’s personal information. Written informed consent was obtained by the members of the BRCF from each patient and extensive precautions were taken to preserve the privacy of the participants. The University of Kansas Medical Center’s Human Subjects Committee approved the use of these tissues. All safety and ethical guidelines were followed.

Tumor samples for Immunostaining

Ovarian tissue microarrays (TMAs) were constructed by the KUCC BRCF and contained duplicate cores of both normal ovarian tissue samples (n = 26) and serous ovarian tumor cases (n = 56).

Immunohistochemistry

The method for IHC has been described previously.[33] Briefly, rehydrated TMA sections were treated with target retrieval solution (pH 6.0, Dako). Sections were then incubated with primary antibody (PTPRZ1 at 1:100 dilution) for overnight at 4 °C and successively conjugated with peroxidase-conjugated polymer directed against mouse (Dako) and diaminobenidine (DAB) substrates. Stained sections were imaged using Nikon 80 microscope (Nikon Inc.). Both the epithelial ovarian cancer (EOC, serous subtype) tissue cores and normal ovarian surface epithelia (OSE) were assessed for intensity of staining by two independent pathologists (visualization experiment 1 (VE1) and visualization experiment 2 (VE2)) and scored as 0 (for undetectable staining), 1 (for weak intensity), 2 (for moderate intensity) or 3 (for strong intensity) as well as for the percentage of positively staining cells. In addition, staining was separately assessed for subcellular location. Spearman's correlation test was performed to analyze correlations between VE1 and VE2.

Bioinformatics and enrichment analysis

Functional analysis and network generation was done using Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems, www.ingenuity.com)

Statistical analysis

Descriptive statistics including mean and standard deviation, student’s two-tailed t-tests and Fisher’s exact test were computed using Graph Pad Prism 5.0 software. P values < 0.05 were considered significant.

PTN expression in EOC cell lines and serous ovarian carcinoma tissue

In order to dissect the potential role of PTN in ovarian cancer, we determined the PTN expression levels across a panel of five EOC cell lines (A1847, A2780, OVCAR3, OVCAR5 and SKOV3). We evaluated expression of PTN mRNA using qRT-PCR which showed differential levels of expression in the EOC cell lines (normalized C_t values, Figure 1A). PTN was detected as a band with a molecular weight (MW) of 18 kDa, corresponding to the migration of recombinant human PTN (rhPTN), across these EOC cell lines by immunoblotting (Figure 1B). Primary human fibroblasts derived from normal ovaries (HFNO-402) or from ovarian tumors (HFOT-208) [34] have undetectable levels of PTN (Supplementary Figure S1A). We next measured secreted levels of PTN across the EOC cells and a positive control glioblastoma cell line, LN229, by ELISA. PTN was detected in all the EOC cell lines evaluated, showing an average level of 0.49 ng/mL (range ~0.30 – 0.65 ng/mL, n = 4, Figure 1C) and was 2- to 3-fold higher than levels detected in the positive control LN229 cell line (0.19 ng/mL). Based on these results, it was apparent that PTN is expressed, produced and secreted by the ovarian cancer cells evaluated. We next evaluated PTN levels in tumor tissue and serum samples from patients with ovarian carcinoma. Using a small sample set that included normal ovarian tissue (n = 2) and high-grade serous ovarian tumor tissue (n = 5), protein extracts were prepared and analyzed for PTN expression using immunoblotting (Figure 1D). PTN was detected in 4 out of the 5 tumor lysates with 3 of the samples exhibiting 2-fold or higher expression levels than the normal tissues (Figure 1D). The mean level of PTN across the five tumor tissues was ~3.5-fold higher than the mean level in normal ovarian tissue (Supplementary Figure S1B). However, when we measured circulating levels of PTN in serum samples from patients with high grade serous ovarian cancer (grade III, n = 10) and normal human age- and gender-matched controls (n = 10), no statistically significant difference was observed (p = 0.63) (Supplementary Figure S1C), suggesting that the circulating levels of PTN as measured in this study are not a strong discriminator of disease state.

Expression of PTPRZ1 in EOC cell lines and serous ovarian carcinoma tissue

We next evaluated the expression levels of the PTN receptor, PTPRZ1, in EOC cells. As with PTN transcript levels, PTPRZ1 mRNA could be detected across all of the EOC cell lines (Figure 2A). Immunoblotting revealed the presence of a transmembrane protein with an approximate MW of 300 kDa in all EOC cells, similar to one that was observed in the positive control cell line LN229 [19] (Figure 2B). In comparison, the primary human ovarian fibroblast lines, HFNO-402 and HFOT-208, appear to have very low levels of PTPRZ1 (Supplementary Figure S2). Furthermore, TCGA ovarian dataset [7] was analyzed for PTPRZ1 DNA methylation and copy number across 494 high grade serous adenocarcinoma specimens (Figure 2C and 2D). β-values for DNA methylation status of PTPRZ1 indicate that the gene was hypomethylated (β-values < 0.25 [35]) across 96% of the specimens. These results imply that the gene is actively transcribed across the majority of the tumor specimens. Copy number data showed that PTPRZ1 exhibits low level copy number gains (log₂ tumor/normal ratio > 0.3) across 30% of the samples.

We further evaluated PTPRZ1 expression in clinical samples by immunostaining. The ovarian TMAs examined contained epithelial components from a total of 26 normal ovarian tissue samples and 54 ovarian tumor (serous histotype) cases (two cores per case) (Table 1, patient characteristics). Each core was independently scored by two pathologists (VE1 and VE2) as described in the Methods section. PTPRZ1 could be detected in 100% of the tumor samples (Supplementary Table S2) with PTPRZ1 staining noted in 60%-100% of the epithelial cells (Supplementary Table S2). In contrast, PTPRZ1 staining was undetectable in the majority of normal ovarian surface epithelial (OSE) cells with measurable staining being typically weak (Supplementary Table S1). The sub-cellular localization of PTPRZ1 was not uniform across the serous ovarian cancers or the normal ovarian tissues studied. Fifteen percent of the tumor samples showed strong membranous staining at the tumor-stromal junction (Table 2 Fisher’s exact test p-value < 0.05; Supplementary Tables S1 and S2 and Figure 3A (left) for a representative image). No membranous staining was observed in normal OSE cells. PTPRZ1 expression was elevated in the cytoplasm of all the transformed tissues with weak, moderate and strong staining noted in 15%, 59% and 26% of the samples respectively. The tumor epithelial components scored higher when compared to normal OSE (Table 3 Fisher’s exact test p-value < 0.0001; Supplementary Tables S3; Figure 3A for a representative image; and Figure 3B)). Nuclear PTPRZ1 expression was elevated in the tumor epithelium as compared to normal OSE (Table 4 Fisher’s exact test p-value = 0.0018; Supplementary Table S4; Figure 3A for a representative image; and Figure 3C). Nuclear staining for PTPRZ1 was undetected in 42% of normal OSE samples and was weakly positive for 54% of normal OSE samples with only one sample staining moderately. Nuclei of 89% of tumor samples stained positive with 39% staining weakly, 45% staining moderately and about only 5% staining strongly. A statistically significant positive correlation between VE1 and VE2 scores for the cytoplasmic (r = 1) and nuclear (r = 0.93, p < 0.001) PTPRZ1 intensity across 26 OSE cores was observed (Supplementary Table S1). Similarly, a statistically significant positive correlation between VE1 and VE2 scores for the cytoplasmic (r = 0.86, p < 0.001) and nuclear (r = 0.61, p < 0.001) PTPRZ1 intensity across 54 EOC cores was observed (Supplementary Table S2). Although a weakly positive correlation (r = 0.3) between VE1 and VE2 scores for the membranous PTPRZ1 intensity across the fifteen percent of the tumor samples was observed (Supplementary Table S2), but the VE1 and VE2 scores did not significantly overlapped due to non-uniform staining patterns. Furthermore, Kaplan-Meier survival analysis of the IHC staining intensities (undetected/weak vs. moderate/strong) with the corresponding clinical survival data from each patient was performed to determine the prognostic value of PTPRZ1 expression at all three cellular sites (cytoplasm, membrane and nucleus). We did not observe statistical significance (p > 0.05) in the survival curves for any of the sub-cellular stains, with nuclear staining, although not statistically significant, showing the greatest difference in survival between the two staining intensities (Supplementary Figure S3).

Knockdown of PTPRZ1 effects viability of EOC cells

In our previous study [6] we showed that gene silencing of PTN by RNAi approaches affected EOC cell viability and decreased survival. In this study we evaluated whether knockdown of its receptor, PTPRZ1 affected EOC cell viability as well. We transfected A1847, A2780, OVCAR3, OVCAR5 and SKOV3 with PTPRZ1 siRNA and a negative control siRNA (GL2 siRNA targeting firefly luciferase) and measured the effect on cell viability after 72 h. All EOC cell lines exhibited variable sensitivity to knockdown of PTPRZ1 with normalized viability scores ranging from 0.40–0.89 (decrease in viability ranged from 11–60%) across the five EOCs as compared to respective controls (Figure 4A). More than 40% reduction in cell viability was observed in 3 out of the 5 cell lines. Among these cell lines, A2780 and OVCAR5 were the most sensitive to PTPRZ1 knockdown. Phase contrast images of GL2 siRNA treated and PTPRZ1 siRNA treated A2780 cells (Figure 4B) and OVCAR5 cells (Supplementary Figure S4) are shown. As an additional check on the ability of PTPRZ1 siRNA to correctly target PTPRZ1, we performed immunoblotting to determine the level of PTPRZ1 protein 72 h after siRNA treatments in A2780 cells (Figure 4C). An approximate 80% decrease in PTPRZ1 level was observed in PTPRZ1 siRNA treated as compared to GL2 siRNA treated A2780 cells.

Since cell viability measurements provide limited information about the mechanism of decreased cell viability following gene silencing, we investigated the functional aspects of targeting PTPRZ1 on survival (apoptosis and necrosis) and cell cycle progression in EOC cells. A2780 and OVCAR5 cells were transfected with PTPRZ1 siRNA and GL2 siRNA (50 nM) in 96- well plates. Cells were harvested 72 h post transfection. Thereafter, effect on survival (using the Guava Nexin assay which combines annexin V staining to measure apoptosis and 7-Amino-actinomycin D (7-AAD) staining to measure necrosis) and on cell cycle progression (using the Guava Cell Cycle assay) were analyzed (Figures 4D and 4E, respectively). A significant 1.4-fold increase in annexin V positive and a significant 1.3-fold increase in 7-AAD positive cells was observed in PTPRZ1 siRNA treated as compared to GL2 siRNA treated A2780 cells (Figure 4D, left panel). In OVCAR5 cells, there was a significant 1.2-fold increase in annexin V positive and a significant 1.4-fold increase (n = 3, p < 0.05) in 7ADD positive cells in PTPRZ1 siRNA treated as compared to GL2 siRNA treated cells (Figure 4D, right panel). However, no significant effect on the cell cycle phase distribution was observed following knockdown of PTPRZ1 gene in either of the two cell lines (Figure 4E).

PTN mediates its pro-survival effect through ERK1/2 (extracellular signal-regulated kinase) in EOC cells

We have previously shown that knockdown of PTN induces significant apoptosis in A2780 cells [6]. To investigate the potential signaling pathways induced by PTN in EOC cells, we adopted an integrated experimental and computational approach. First, we evaluated the differences in expression across a panel of 36 survival related genes in PTN siRNA treated A2780 cells as compared to negative control siRNA treated A2780 cells. Fold-change in gene expression was calculated (Supplementary Table S5) and it was found that 24 genes exhibited a 1.5-fold or greater change in expression level following PTN silencing relative to GL2 siRNA treated cells (Figure 5A). Of these, 15 genes were down-regulated by 1.5-fold or more following PTN depletion (highlighted in green in Supplementary Table S5), and were subsequently analyzed for signaling effectors responsible for the observed changes in their gene expression using a computational method (IPA network analysis, Figure 5B). Functional mapping and network characterization of these 15 genes was performed (shaded nodes, Figure 5B). The results showed that the network with the highest significance [36, 37] to which the genes mapped exhibited downstream effects through multiple signaling effectors (non-shaded nodes, Figure 5B) namely ERK1/2, RB, and cyclin A. Of these, PTN has been previously shown to exert its pro-tumorigenic effects through ERK1/2.[38, 39] Moreover, ERK1/2 is a well-known effector protein that has been implicated as an important mediator of growth factor induced oncogenic pathways in ovarian cancer [40–42].

Based on these observations, we evaluated whether ERK1/2 is involved in PTN mediated signaling in EOCs. First, we investigated whether PTN knockdown affects ERK1/2 signaling in A2780 cells. PTN siRNA treatment down-regulated the cellular PTN levels by more than 50% at 48 h in A2780 cells (Figure 5C) and secreted PTN levels as measured by ELISA were decreased by more than 75% (Supplementary Figure S5). Furthermore, PTN siRNA treatment in A2780 cells reduced phosphophorylated ERK1/2 (pERK1/2) levels by 86% and 100% at 24 h and 48 h after siRNA transfection, respectively (Figure 5C). Similar to PTN siRNA treatment, PTPRZ1 siRNA treatment in A2780 cells also reduced phosphophorylated ERK1/2 (pERK1/2) by 70% at 48 h after siRNA transfection (Supplementary Figure 6A and 6B).

Next, we examined the effect of PTN stimulation on ERK1/2 phosphorylation. The three cell lines A2780, OVCAR3 and OVCAR5 that showed maximum sensitivity to PTN siRNA treatment (Supplementary Figure S7A) were tested. Cells were seeded overnight and serum starved for 48 h followed by treatment with PTN (100 ng/mL). In A2780, peak induction of pERK1/2 levels was observed at 5 min (2.4-fold increase in pERK1/2 levels as compared to unstimulated cells) following PTN treatment (Supplementary Figure S7B). OVCAR3 and OVCAR5 also showed pERK1/2 induction following PTN treatment (Supplementary Figures S7C and S7D). Finally, this induction of pERK1/2 by PTN treatment was attenuated by treatment with a MEK inhibitor U0126 (30 µM [30], 2 h) (Figure 5D) suggesting that induction was directly related to PTN stimulation.