Therapeutic potential of PRL-3 targeting and clinical significance of PRL-3 genomic amplification in gastric cancer

The GC cell line MKN7 was kindly provided from the Cell Resource Center for Biomedical Research Institute of Development, Aging and Cancer, Tohoku University (Sendai, Japan). Seven other GC cell lines (GCIY, AZ521, KatoIII, SH10, H111, MKN74, and NUGC4) were purchased from RIKEN BioResource Center (Ibaraki, Japan). These cell lines cover the two main types of GC [12], intestinal type (MKN7, MKN74, AZ521, and H111 cells) and diffuse type (GCIY, KatoIII, SH10, and NUGC cells) [13‐15]. MKN7, NUGC4, and AZ521 cells were established from lymph node metastasis (LNM), and MKN74 cells were from liver metastasis. KATOIII and GCIY cells were established from metastatic pleural effusion and ascites, respectively. H111 and SH10 cells were established from the xenotransplantation. Normal skeletal muscle C2C12 cells were purchased from DS Pharma Biomedical Co., Ltd (Osaka, Japan). AZ521 and C2C12 cells were grown in DMEM medium (GIBCO, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS). The other cells were grown in RPMI1640 medium (GIBCO) supplemented with 10% FBS. 1-4-bromo-2-benzylidene rhodanine was purchased from Calbiochem Corp (San Diego, CA), which was identified as a PRL-3 inhibitor through high throughput screening using chemical library of Korea Chemical Bank, and inhibited PRL-3 phosphatase activity [16]. Indeed, phosphorylation of KRT8, PRL-3-interacting protein, induced by catalytically inactive mutant of PRL-3, but not by wild type, was confirmed by PRL-3 inhibitor treatment in a dose-dependent manner [17]. Moreover, anticancer efficacy of PRL-3 inhibitor treatment also showed to be similar to that of siRNA treatment in esophageal cancer or colorectal cancer [11, 17].

Out of 173 formalin-fixed, paraffin-embedded, tissue samples series where we previously assessed PRL-3 expression status using immunohistochemical staining (IHC) in GC [6], 77 matched pairs of primary tumor tissues and the corresponding normal mucosa tissues were randomly selected from patients with differential stages according to the 13^th edition of the Japanese Classification of Gastric Carcinoma (JCGC) [18]; 40 pairs with positive PRL-3 expression (10 patients in Stage I, 10 in II, 10 in III, and 10 in IV) and 37 pairs with negative expression (10 patients in stage I, 10 in II, 9 in III, and 8 in IV). All patients underwent gastrectomy according to the gastric cancer treatment guidelines in Japan [19], and histopathologic examinations were done according to the JCGC. The 6^th edition of the International Union Against Cancer (UICC)/TNM classification was also used [20]. Table 1 depicts the detailed information on 77 patients. All tissue samples were collected at the Kitasato University Hospital, and informed consent was obtained from all patients. The present study was approved by the Ethics Committee of the Kitasato University.

Fluorescence in situ hybridization (FISH) analysis was performed, as described previously [11]. PRL-3 is located on chromosome 8q24.3 (GenBank accession number NT 000008.9), and the chromosome 8 centromeric probe was used to estimate the copy number. Because PRL-3 FISH scoring algorithms had not been standardized, the assessment was based on the criteria of HER2 [21]. For each sample, at least 60 cancer cells were scored. Positive PRL-3 genomic amplification was defined as a ratio of PRL-3 to chromosome 8 centromere more than 2.2, and negative was the ratio of less than 1.8. If the ratio of PRL-3 to chromosome 8 centromere was 1.8 to 2.2, additional cells were counted, and the ratio of more than 2.0 was finally considered as positive [21]. Polysomy was defined as the mean chromosome 8 centromeric signals more than 3.0 per nucleus [22].

Tissue sections from tumor and the corresponding normal mucosa, obtained at least 5 cm from the tumor edge, were sharply dissected on hematoxylin and eosin-stained slides, and genomic DNA was subsequently extracted using of a QIAamp DNA FFPE Kit (QIAGEN Sciences, Hilden). Quantitative-genomic polymerase chain reaction (Q-PCR) was performed to quantify PRL-3 gene copy numbers using iQ™ Supermix (Bio-Rad Laboratories, Hercules, CA) in triplicate on the iCycler iQ™ Real-Time PCR Detection system (Bio-Rad). To normalize PRL-3 gene copy number per cell, ADAM metallopeptidase domain 2 (ADAM2, NT 923907.1), located on chromosome 8p11.2, was used as an endogenous reference because that gene amplification is defined as a copy number increase of a restricted region of a chromosome arm [10]. ∆Ct values were calculated as Ct (PRL-3)-Ct (ADAM2) for each sample. Relative copy number was determined as 2^{- ∆∆ Ct}, where ∆∆C_t = ∆C_t (tumor)-∆C_t (corresponding normal) [23]. The increases of more than 2-fold relative to the corresponding normal were considered as genomic amplification. Additional file 1 depicts detailed PCR condition and sequences of primer and probe used in the present study.

Whole cell lysates were extracted in RIPA buffer (Pierce, Rockford, IL) supplemented with 10 μL/mL Halt™ Protease Inhibitor Cocktail Kit (Pierce) and Halt™ Phosphatase Inhibitor Cocktail Kit (Pierce), and the protein were separated on NuPAGE^® 4-12% Bis-Tris Gel (Invitrogen) according to the manufacturer's protocol. Both detection and quantification of the specific proteins were performed using ATTO Light Capture (ATTO Corporation, Tokyo, Japan). Two colorectal cancer cell lines DLD-1 and SW480 cells (RIKEN BioResource) were used as the low and high expression controls, respectively, as described previously [11].

PRL-3 mouse monoclonal antibody (R&D Systems, Minneapolis, MN) and β-actin mouse monoclonal antibody (Sigma, St. Louis, MO) were used as described previously [11].

Cells were transfected with 1 μmol/L Accell SMARTpool, siRNA-PRL-3 (Thermo Fisher Scientific, Lafayette, CO) mixed with Accell siRNA Delivery Media (Thermo Fisher Scientific) according to the Thermo Scientific Dharmacon^® Accell™ siRNA Delivery Protocol [24]. The Accell Non-targeting Pool (siRNA-ctr) and Accell siRNA Delivery Media alone were used as a control for non-sequence-specific effects and as a mock-treatment, respectively.

Anchorage-independent cell growth was analyzed by plating 0.36% top agarose (Bacto™ Agar, Becton, Dickinson and Company, Franklin Lakes, NJ) containing 1 × 10⁵ cells on a surface of 0.72% bottom agarose in 6-well plates [11]. Cells were fed weekly by overlying fresh soft-agar solution, and colonies were photographed after 2 weeks of incubation. The 50% effective concentration (EC₅₀) value of PRL-3 inhibitor treatment was calculated based on the measurement of colony count.

The proliferation assay was performed using Premix WST-1 Cell Proliferation Assay System (Takara Bio, Tokyo). Cells (2 × 10³) were seeded in 96-well, and the proliferative activity was measured by absorbance at 450 nm on designated sampling days. The sensitivity to PRL-3 inhibitor on antiproliferation was determined using the 50% inhibitory concentration (IC₅₀) value after treatment for 72 hours.

The invasion assay was performed in the 24-well BD BioCoat™ Matrigel™ Invasion Chamber (BD Biosciences Discovery Labware, Bedford, MA). Cells that had invaded through the membrane were counted in four separated fields per well. Both experiments were done in triplicate.

Apoptosis assays were performed using Guava PCA System (Guava Technologies, Inc., Hayward, CA). Cells (2 × 10⁵) were treated with the PRL-3 inhibitor at the indicated concentration in medium supplemental with 1.0% FBS for 72 hours, then stained with Annexin V and 7-AAD (Guava Nexin Reagent). The experiment was done in triplicate and analyzed using CytoSoft 2.1.5 software (Guava Technologies).

Fisher's exact test, or the Mann-Whitney U-test was used to statistically analyze the relationship between PRL-3 gene amplification and clinicopathological variables. One-way analysis of variance (ANOVA) with post-hoc test was used to compare between three groups for siRNA treatment (siRNA-PRL-3, siRNA-ctr, and mock). Student t test was used to evaluate therapeutic effect for the individual concentrations of PRL-3 inhibitor, compared with 0 μmol/L of PRL-3 inhibitor. The Kaplan-Meier method was used to estimate cumulative survival rates, and differences in survival rates were assessed with the use of the log-rank test. All deaths of patients were cancer-related, and disease specific survival (DSS) was measured from the date of surgery to the date of death or the last follow-up. P < 0.05 was considered to indicate statistical significance. All statistical analyses were conducted with JMP 7.0 software (SAS Institute, Cary, NC).

Initially, PRL-3 expression status was evaluated using western blotting in 8 GC cell lines (Figure 1A). PRL-3 expression was observed at a detectable level in all the cell lines, among which 5 cell lines (KatoIII, H111, MKN7, MKN74, and NUGC4 cells) and 3 cell lines (GCIY, AZ521, and SH10 cells) exhibited high and relatively low expression, respectively. Subsequently, FISH analysis was performed to examine whether PRL-3 expression was caused through its genomic amplification (Figure 1B). Genomic amplification was obviously positive in 2 cell lines (MKN7 and MKN74 cells) and negative in 6 cell lines. 3 of the six were dysomic (AZ521, GCIY, and NUGC4 cells), and three were polysomic (KatoIII, SH10, and H111 cells). PRL-3 genomic amplification frequently occurred in the different regions from chromosome 8, so-called distributed insertions, on metaphase [10], and was concordant with its high expression.

In our previous study, PRL-3 expression was detected in 95 (55%) out of 173 primary GCs by IHC [6]. To explore the link between PRL-3 expression and its genomic amplification, Q-PCR was performed for both the 40 tumors with positive PRL-3 expression and 37 tumors with negative expression, which were randomly selected from differential stages in the 173 primary tumors. All the primary tumors without PRL-3 expression were not amplified, whereas 8 (20%) out of the 40 primary tumors with PRL-3 expression were amplified (Figure 2A). FISH analyses also confirmed obvious genomic amplification as the cancer-specific alteration (Figure 2B), and exhibited at nearly homogenous pattern in both the central area and invasive area within tumor. Subsequently, the relationship with clinicopathological factors was assessed for PRL-3 genomic amplification (Table 1), where it was significantly associated not only with its expression (P = 0.006), but also with depth of tumor invasion (P = 0.006), presence of LNM (P = 0.022), LNM status (P = 0.004 in JCGC, P = 0.002 in UICC), and stage (P = 0.005 in JCGC, P = 0.003 in UICC). Additionally, all the primary tumors with genomic amplification were stage III or IV disease (40%, 8/20). Moreover, the genomic amplification negatively affected the outcomes of the histologically node-positive patients (P = 0.021, Figure 2C), although PRL-3 expression did not in our and other previous reports [6, 25].

In GC, the functional roles of PRL-3, including invasion and proliferation abilities, have been documented only in SGC7901 cells [25]. To confirm these metastatic properties using 3 cell lines with different PRL-3 expression and genetic status, knock-down of endogenous PRL-3 expression was performed using siRNA transfection; AZ521 cells (low expression and disomy), H111 cells (high expression and polysomy), MKN74 cells (high expression and genomic amplification). These cell lines were transfected with siRNA-PRL-3 or siRNA-ctr, and western blotting showed the decreased level of PRL-3 protein in siRNA-PRL-3 cells, but not siRNA-ctr cells, compared with mock-treatment cells (Figure 3A). One of the important characteristic of the metastatic phenotype is supposed as the ability for cancer cells to grow under anchorage-independent conditions [26], but the involvement in PRL-3 remains unknown in GC. All siRNA-PRL-3 cells showed the significantly decreased size and number of colonies, compared to siRNA-ctr cells or mock-treatment cells (Figure 3B). Moreover, in line with previous reports for other GC cell lines [25, 27], we also confirmed that siRNA-PRL-3 cells showed the significantly less proliferative activity (Figure 3C) and invasive ability (Figure 3D).

To assess the therapeutic potential and examine a landmark guiding the response to PRL-3-targeted therapy, we evaluated the anticancer activity of PRL-3 inhibitor, cell-permeable benzylidene rhodanine compound [16], against 6 cell lines with different PRL-3 expression and genetic status; GCIY and AZ521 cells (low expression and disomy), KatoIII cells (high expression and polysomy), SH10 cells (low expression and polysomy), MKN7 and MKN74 cells (high expression and genomic amplification). Cells were treated with PRL-3 inhibitor at concentrations ranging from 0 to 50 μmol/L. PRL-3 inhibitor showed dose- and time-dependent antiproliferative efficacy on all the tested cell lines, irrespective of different PRL-3 expression level and genetic status, and the IC₅₀ values of GCIY, AZ521, KatoIII, SH10, MKN7, and MKN74 cells were 26.77, 9.98, 24.26, 23.95, 22.29, and 9.45 μmol/L, respectively (Figure 4A). AZ521 and MKN74 cells were more sensitive to PRL-3 inhibitor treatment than GCIY and MKN7 cells that were categorized as the identical groups in terms of expression and genetic status, respectively. Namely, genetic or expression status was not associated with sensitivity of GC cells against the PRL-3 inhibitor. Similar efficacy was shown in anchorage-independent colony formation, and the EC₅₀ values of GCIY, AZ521, SH10, and MKN74 cells were 6.99, 9.52, 13.05, 9.09 μmol/L, respectively (Figure 4B). GCIY cells exhibited more sensitive inhibition in contrast with the anti-proliferation. Additionally, this inhibitor also robustly abrogated the invasive ability of GC cells (Figure 4C). To further characterize the anticancer efficacy of PRL-3 inhibitor treatment, apoptosis assay was performed (Figure 5A). Although 1 μmol/L of the inhibitor was insufficient to induce apoptosis beyond the baseline (0 μmol/L), 10 μmol/L of the inhibitor robustly caused the drastic apoptosis on all the tested cell lines, where there were the 3-fold and 11-fold increases beyond the baseline in GCIY and MKN74 cells, respectively. Thus, PRL-3 inhibitor repressed these metastatic properties on all the tested cell lines in dose-dependent manner, and neither expression level nor genetic status showed clear correlation with the sensitivity.

Finally, we assessed whether PRL-3 inhibitor induced cytotoxicity in normal skeletal muscle, where PRL-3 is predominantly expressed [28]. Both proliferation and apoptosis assays were performed using normal skeletal muscle C2C12 cells treated with the inhibitor, and showed that 10 μmol/L of the inhibitor failed to cause antiproliferative and apoptotic response on C2C12 in contrast with the efficacies on all the tested GC cell lines (Figure 5A and 5B).