The depletion of PinX1 involved in the tumorigenesis of non-small cell lung cancer promotes cell proliferation via p15/cyclin D1 pathway

The cBioPortal for Cancer Genomics is an open-access downloaded bio-database, providing visualization and analyzing tool for large-scale cancer genomics data sets (http://​cbioportal.​org). This portal collected records that were derived from 147 individual cancer studies, in which 31 types of cancer were analyzed, which included over 21000 samples [23, 24]. Analysis of the 1788 NSCLC samples (1098 lung adenocarcinoma cases and 682 lung squamous cell carcinoma cases) from this database was performed in silico.

Ninety-three NSCLC patients from the Department of Radiotherapy, Cancer Center, Sun Yat-Sen University between January 2008 and December 2014 were assigned as learning cohort. Another 51 NSCLC cases in the same period from the first Affiliated Hospital, Sun Yat-Sen University were assigned as validation cohort. These cases were selected based on the availability of resection tissues and follow-up data. All selected cases had not received preoperative radiation or chemotherapy before diagnosis biopsies. All samples were classified according to the TNM (UICC, 2014) staging and evaluated for histological type by the same pathologist. The study was approved by the Medical Ethics Committee of the two institutions and involved only informed consent patients. The clinical-pathological characteristics of the two cohorts are summarized in Table 1.

Slides were dried overnight at 37 °C, dewaxed in xylene, rehydrated with graded alcohol, and immersed in 3% hydrogen peroxide for 20 min to block endogenous peroxidase activity. For antigen retrieval, tissue slides were boiled in tris (hydroxymethyl) aminomethane-EDTA buffer (pH 8.0) in a pressure cooker for 10 min. The slides were incubated with 10% normal rabbit serum at room temperature for 20 min to reduce nonspecific interactions. Subsequently, tissue slides were incubated with anti-PinX1 antibody (1:200, ProteinTech Group, Inc.) for 60 min at 37 °C in a moist chamber. After five rinses with 0.01 mol/L phosphate-buffered saline (PBS, pH = 7.4) for 10 min, the slides were incubated with a secondary antibody (Envision, Dako, Glostrup, Denmark) at a concentration of 1:100 for 30 min at 37 °C, followed by PBS washes and finally stained with DAB (3,3-diaminobenzidine). The nucleus was counterstained with Meyer’s hematoxylin. PBS alone was used as a negative control.

PinX1 immunoreactivity was classified by receiver-operator curve (ROC) analysis: (1) low expression defined as less than 65% PinX1 positive cells and (2) high expression defined as greater than 65% PinX1 positive cells. BMP5 positive staining was also divided into low expression cases (cases with score 0–6) and high expression (cases with scores 8–12). (See Additional file 1: Supplementary Materials and Methods).

The construction of PinX1 and GAPDH sense/antisense primers has been previously described [10]. RNA was reverse-transcribed using SuperScript First Strand cDNA System (Invitrogen, USA) according to the manufacturer’s instructions. qRT-PCR was performed using Real-time PCR system (Applied Biosystems, USA) as follows: 50 °C for 2 min, 95 °C for 10 min, 40 cycles of 95 °C for 15 s, and 60 °C for 60 s. The relative levels of gene expression were represented as ΔCt = Ct_gene- Ct_reference, and the fold change of gene expression was calculated by the 2^-ΔΔCt Method.

H125, A549, SK-MES-1 and H1299 cells were maintained in DMEM and/or RPMI 1640 supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin at 37 °C in 5% CO₂. The PinX1 cDNA was cloned into the pCDH-CMV-EF1-copGFP lentivector (System Biosciences, Mountain View, CA, USA). The PinX1-shRNA lentivirus vector has been previously described [10, 11, 16, 20]. The PinX1siRNA transient transfection (GGAGCTACCATCAATAATG) was designed to decrease PinX1 expression temporary.

Cellular viability was measured using the MTT proliferation assay (Sigma) according to the manufacture’s protocol. In brief, 1000 cells were seeded in 96-well plates and cultured/treated for 24 h. Viability was measured at different time points from 12 h to 72 h after post-treatment on the basis of experimental requirement.

Approximately 500 cells were seeded in each well of a six plate for 24 h. The media was then replaced with fresh RPMI1640 or DMEM containing 10% FBS and the cells were maintained for two weeks. Colonies were fixed with methanol and stained with 0.1% crystal violet in 20% methanol for 15 min. Western blot methods were performed with standard procedure [18]. Details may be found in the Additional file 1: Supplementary materials and methods.

EdU is a thymidine analog whose incorporation can be used like BrdU to label cells undergoing DNA replication. Cells were performed using Cell-Light™ EdU Apollo®488 In Vitro Imaging Kit (Ribobio, Guangzhou China) according to the manufacture’s protocol. The EdU positive cells (red cells) were counted using Image-Pro Plus (IPP) 6.0 software (Media Cybernetics, Bethesda, MD, USA). The EdU incorporation rate was expressed as the ratio of EdU positive cells to total DAPI positive cells (blue cells).

The cells were collected, washed with PBS three times and fixed overnight with cold 70% clod ethanol at 4 °C. The fixed cells were analyzed with flow cytometry (Becton Dickinson, San Jose, CA, USA) using the Alexa Fluor 488 Annexin V/Propidium iodide Apoptosis Assay kit (Invitrogen) following the manufacturer’s instructions.

Total lysates obtained from A549 cells and with or without PinX1 siRNA were used to perform a Cell Cycle Antibody Array according to the manufacture’s protocol (Full Moon BioSystems, Inc., Sunnyvale, CA, USA).

According to institutional guidelines, female 5-week-old athymic nude mice were used for in vivo experiments. Cells (5 × 10⁶ A549-vector, 5 × 10⁶ A549-PinX1 shRNA, 5 × 10⁶ SK-MES-1, 5 × 10⁶ SK-MES-1 PinX1) were injected into the lateral aspect of the rear leg. The tumor diameters were measured with calipers every 4 days, and tumor volumes were calculated (width² × length/2).

The bioinformatics analysis was first performed to detect PinX1 gene alterations in NSCLC, using cBioportal Web resource online (cBioportal for Cancer Genomic). The deletion of PinX1 gene accounted for the most alterations and was visualized in six NSCLC database (Lung Adenocarcinoma – Broad, Cell 2012; TCGA, Provisional; TCGA, Nature 2014; TSP, Nature 2008; Lung Squamous Carcinoma – TCGA, Provisional; TCGA, Nature 2012) (Fig. 1a). There were more PinX1 homozygous deletions and PinX1 heterozygous deficiencies and less PinX1 amplifications in two specific NSCLC datasets (namely Lung Squamous Carcinoma – TCGA, Provisional and Lung Adenocarcinoma –TCGA, Provisional) (Fig. 1b).

To verify the results observed from historical data on cBioportal for Cancer Genomic, we selected primary NSCLC tissue samples from our affiliated hospitals. In these selected primary NSCLC tissues, PinX1 mRNA expression was observed in 8 out of 12 (66.7%) NSCLC samples as compared to their normal counterparts by qRT-PCR (Fig. 1c). Western blot analyses revealed that 7 out of 12 (58.3%) NSCLC samples had decreased PinX1 expression, when compared with adjacent normal lung tissues (Fig. 1d).

Immunohistochemistry was utilized to examine the expression of PinX1 in NSCLC patients (Fig. 2b). According to the ROC curve (Fig. 2a, Additional file 2: Table S1 and Additional file 3: Figure S1), low expression of PinX1 was discovered in 56/93 (60.2%) in the learning cohort and 27/51 (52.9%) in the validation cohort. As shown in Table 1, the association between PinX1 expression and the clinical-pathological parameters in learning cohort and validation cohort suggested that low expression of PinX1 is significantly associated with NSCLC patients’ smoking condition, histological type, T stage, N stage, M stage and TNM stage. However, no significant associated between PinX1 expression and other clinical-pathological features, such as age, gender, WHO grade (Table 1).

In univariate analysis, Kaplan-Meier analysis also demonstrated a significant impact of well-known clinicopathological prognostic parameters, such as Smoking Condition (P = 0.000 and 0.042, respectively), T stage (P = 0.000 and 0.009, respectively), N stage (P = 0.000 and 0.011, respectively), M stage (p = 0.000 and 0.004, respectively), TNM stage (P = 0.003 and 0.008, respectively) on NSCLC patient’s survival in both cohorts (Table 2). However, these parameters, such as Age (P = 0.199 and 0.899, respectively), Gender (P = 0.996 and 0.385, respectively), Histological Type (P = 0.293 and 0.252, respectively) had no correlation with NSCLC patient’s survival in both cohorts (Table 2). Furthermore, low PinX1 staining was correlated with poor survival of NSCLC patient in the learning cohort (media 21 months versus 34 months, P = 0.008) and in the validation cohort (media 18 months versus 34 months, P = 0.005) (Fig. 2c). In addition, PinX1 gene alteration also correlated with overall survival (media 38.47 months versus 52.56 months, P = 0.0353) and Disease-free survival (media 17.0 months versus 51.51 months, P = 0.000), which is explored in cBioportal Web resource online (Fig. 2d). In multivariate analysis, the expression of PinX1 expression was found to be an independent prognostic factor for NSCLC patients’ survival in the learning cohort (P = 0.046) and in validation cohort (P = 0.023) (Table 3).

We continued our study in vitro by investigating how PinX1 levels correlate with NSCLC cell survival. Of the four NSCLC cell lines (SK-MES-1, H125, A549, H1299) analyzed by Western blot, and all four lines showed lower levels of endogenous PinX1 than those in normal control lung tissues (Fig. 3a). Next, we choose to decrease the PinX1 expression in A549 and H1299, which have higher PinX1 expression among the four cell lines. Meanwhile, we further increased PinX1 expression in SK-MES-1 which has the lowest PinX1 expression. We examined Bcl-2 and Bax protein levels, and discovered that the ratio of Bcl-2/Bax expression was substantially enhanced in PinX1-silenced SK-MES-1 cells and reduced in PinX1-expression A549 and H1299 cells (Fig. 3b). A subsequent MTT analysis and colony-forming assay showed that both A549 and H1299 cells transfected with PinX1 siRNA and/or PinX1 displayed a substantial increase in cell proliferation and/or clone capacity compared with control cells after PinX1 knockdown, while the reintroduction of PinX1 in SK-MES-1 cell line suppressed the cell viability and/or clone ability (Fig. 3c and d). In addition, we also observed that knocked-down PinX1 expression in BEAS-2B cells (Normal lung epithelial cells), the survival capacity of cells was substantially enhanced, and BEAS-2B transfected with PinX1 displayed a substantial drop compared with that of control cells (Additional file 4: Figure S2).

According to our previous study, we asked whether down-regulation of PinX1 influenced the cell cycle in G1/S phase. An EdU cooperation assay and flow cytometry were used to analyze the potential mechanisms of PinX1 suppresses NSCLC cells proliferation. As shown in Fig. 4a and b, the EdU positive cells were dramatically boosted in the both PinX1-silenced cells and significant reduced in PinX1-overexpression cells compare with that control cells. Knockdown of PinX1 resulted in a decrease in the percentage of cells in the G0/G1 phase from 63.4%, 57.7% in control cells to 50.5%, 49.6% in H1299 and A549 cells lines transfected with PinX1 siRNA respectively. Additionally, the population of cells in S phase also displayed the same trend from 17.2%, 20.1% in control cells to 29.3%, 31.7% in PinX1-slienced cells. Simultaneously, an increase in the percentage of cells in G0/G1 phase from 56.8% in control cells to 66.9% and S phase from 25.4% in control cells to 15.9% in PinX1-overexpression cells (Fig. 4c and d). These findings indicated that PinX1 in blocking G1 to S phase transition might contribute to NSCLC cells survival ability.

To identify the molecular basis of PinX1 regulation of target proteins in NSCLC cells transition from G0/G1 phase to S phase, the expression of several relevant activators or inhibitors of the cell cycle were detected by antibody microarrays and western blot assay in PinX1-silenced A549 cells. In this antibody arrays, we discovered the expression of P15^INK4b and Rb decreased 0.4-, 0.18- fold and the expression of CDK4, CyclinD1, CyclinD2, p-Rb(p-Ser608) were increased 0.25-, 0.31-, 0.2-, 0.5- fold in PinX1-slienced A549 cells (Fig. 5a). Furthermore, western blot also confirmed that decreased P15^INK4b and increased CDK4, CyclinD1, p-Rb(P-Ser608) expression were observed in PinX1-slienced A549 cells (Fig. 5c). These results indicated that P15/cyclinD1 pathway might contribute to the PinX1-arrested G0/G1 to S phase transition.

Xenograft assay was used to investigate if PinX1 knockdown could stimulate NSCLC cells proliferation in vivo. As shown in Fig. 5b, tumors from nude mice injected with A549-PinX1shRNA grew much faster and weighed significantly more at day 44 (mean tumor volume of 200 mm³ to 1820 ± 155 mm³), compared to control A459 tumors (mean tumor volume of 200 mm³ to 1420 ± 138 mm³, P < 0.05). Furthermore, the growth of the SK-MES-1-PinX1 tumors was substantially suppressed (a mean tumor volume of 1220 ± 116 mm³), compared with that in control SK-MES-1 tumors (a mean tumor volume of 896 ± 95 mm³, P < 0.05).

According to another study, the differentially expressed mRNAs were identified by mRNA microarray analysis in PinX1 overexpressed group. The fold change (14.54) of BMP5 was significantly increased in microarray analysis and was expressed at the highest levels [25]. Therefore, BMP5 was involved in PinX1-slienced A549 cells and PinX1-overexpressed SK-MES-1 cells to understand the possible molecules of PinX1 act on P15/cyclinD1 pathway. As it shown in Fig. 5d, BMP5 expression was significantly decreased in PinX1-slienced A549 cells, and increased in PinX1-overexpressed SK-MES-1 cells. As it shown in Fig. 6a, Bcl-2/P-Rb expression was up-regulated in PinX1-slienced A549 cells and was down-regulated in PinX1-overexpressed SK-MES-1 cells. Bax/P15^INK expression was decreased in PinX1-slienced A549 cells and increased in PinX1-overexpressed SK-MES-1 cells. After introducing BMP5-inhibitor noggin, Bcl-2/P-Rb expression was increased and Bax/P15^INK expression was decreased in PinX1-overexpressed SK-MES-1 cells. Moreover, Bcl-2/P-Rb expression was decreased and Bax/P15^INK expression was increased in PinX1-slienced A549 cells by BMP (10 μg/ml) treatment (Fig. 6a). Next, MTT analysis, flow cytometry and EdU cooperation assay further demonstrated that the function of PinX1 in cell proliferation was inhibited by noggin and restored by BMP-5 (Fig. 6b, c and d). Together, these results support the notion the BMP5 may be important for PinX1 mediated cell cycle and its expression may be related to NSCLC proliferation.

Utilizing the previously criterions of a semi-quantitative scale evaluation, low expression of BMP5 was observed in 41/93 (44.1%) in the learning cohort and 23/51 (45.1%) in the validation cohort. Further correlation analysis demonstrated that a prominently consistency correlation between PinX1 expression and BMP5 in both cohorts (P = 0.015 in learning cohort and P = 0.006 in validation cohort, Fisher’s exact test, Table 4).