SEARCH

SEARCH BY CITATION

Keywords:

  • Biomarker;
  • Non-small cell lung cancer;
  • KPNA2;
  • Secretome;
  • Transcriptome

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

The cancer cell secretome may contain potentially useful biomarkers. In this study, we integrated the profiles of secreted proteins in lung cancer cell lines with mRNA expression levels from pulmonary adenocarcinoma tissue, with a view to identify effective biomarkers for non-small cell lung cancer (NSCLC). Among the novel candidates isolated, importin subunit alpha-2 (also known as karyopherin subunit alpha [KPNA]-2), was selected for further validation. Immunohistochemical staining revealed overexpression of KPNA2 in the nuclei of tumor cells, compared with adjacent normal cells. A sandwich ELISA assay developed to detect KPNA2 levels in serum samples showed significantly higher serum KPNA2 in NSCLC patients than in healthy controls. A combination of serum KPNA2 and carcinoembryonic antigen displayed higher diagnostic capacity than either marker alone. Importantly, protein levels of KPNA2 in pleural effusion from NSCLC patients were also significantly higher than those from non-lung cancer. Moreover, knockdown of KPNA2 inhibited the migration ability and viability of lung cancer cells. Our results collectively suggest that integration of the cancer cell secretome and transcriptome datasets provides an efficient means of identifying novel biomarkers for NSCLC, such as KPNA2.

Lung cancer is the most common type of cancer worldwide1 accounting for 12.4% of all newly diagnosed cases,2 and it represents the leading cause of cancer-related deaths in Taiwan. Based on biology, therapy and prognosis, lung cancer is broadly divided into two classes, specifically, small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC).3 NSCLC is the most common lung cancer type, comprising ∼80% of all lung cancers.4 Despite major advances in cancer therapy over the past two decades, the prognosis of patients with NSCLC has improved only minimally, and the 5-year survival rate remains less than 15%.5 One reason for this poor outcome is that most patients are diagnosed at the late stages as they initially present at outpatient clinics.6

Patient history, physical examination and radiological imaging studies, including chest X-ray and computerized axial tomography scans, are currently the gold standard for lung cancer screening. However, these tests are not sufficiently accurate for effective cancer diagnosis and disease staging.7, 8 Serum proteins, such as carcinoembryonic antigen (CEA), CYFRA 21-1 (cytokeratin 19 fragment), CA125 (cancer antigen 125), squamous cell carcinoma antigen, neuron-specific enolase, progastrin-releasing peptide, tumor M2-pyruvate kinase and C-reactive protein are potential markers for lung cancer, and their levels may signify the presence of tumors, facilitate histological analysis, and allow prediction of cancer progression.9, 10 Because of limited sensitivity and specificity, these potential markers are not currently recommended or encouraged in routine clinical practice.11

Proteomic approaches have been widely applied to investigate malignant diseases, particularly in the field of plasma/serum tumor marker identification.12, 13 For practical usage in tumor screening, biomarkers should be measurable in body fluid samples. Thus, proteins secreted by or shed from tumor cells are of particular interest.14, 15 Proteins present in the conditioned media of cultured cell lines derived from specific cancer types present attractive potential tumor biomarker candidates, as they are more likely to be detected in body fluids, such as serum or plasma. Previously, we analyzed the secretomes from two adenocarcinoma cell lines, CL1-0 and CL1-5, by one-dimensional SDS-PAGE in conjunction with the nano LC tandem mass spectrometry (GeLC-MS/MS) approach. Our analysis led to the identification of 1096 and 1830 proteins from serum-free conditioned media of CL1-0 and CL1-5 cells, respectively.16 Although cancer cell secretome profiling is a promising strategy used to identify potential body fluid-accessible cancer biomarkers, questions remain regarding the depth to which the cancer cell secretome can be mined and the efficiency with which researchers can select useful candidates from the growing list of identified proteins. Taking advantage of the available cancer cell secretome datasets and public-accessible gene expression microarray databases of several human cancer types, we have proposed a simple strategy to facilitate the discovery of potential body fluid-accessible cancer biomarkers overexpressed in tumor tissue and secreted into body fluids. This theory is supported by our recent finding that Mac-2 binding protein is an effective colorectal carcinoma (CRC) plasma biomarker, based on its apparent secretion by CRC cell lines and elevated transcriptional level in public array-based analysis of CRC tissues.17

Here, we use the above strategy to detect potential lung cancer biomarkers. We integrated two lung adenocarcinoma cell line secretome datasets (CL1-0 and CL1-5) with one pulmonary adenocarcinoma microarray dataset to identify targets that are significantly up-regulated in lung cancer tissues and secreted/released from lung cancer cells.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Cell culture

Human lung adenocarcinoma cancer cell lines, CL1-0 and CL1-5, were kindly provided by Professor P.C. Yang (Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan, Republic of China). CL1-0 and CL1-5 cells were maintained in RPMI 1640 with 10% fetal bovine serum plus antibiotics at 37°C at a humidified atmosphere of 95% air/5% CO2.18

Transcriptome dataset of pulmonary adenocarcinoma tissue

The NCBI GEO database (http://www.ncbi.nlm.nih.gov/projects/geo/) was searched for mRNA expression profiles of pulmonary adenocarcinoma tissues. Currently, three human lung cancer microarray datasets are deposited in the GEO database, including two squamous cell carcinoma datasets (GDS2373 and GDS1312) and one human pulmonary adenocarcinoma dataset (GDS1650). GDS1650 contains differential gene profiles generated from 20 pulmonary adenocarcinoma tissues and 19 paired adjacent normal tissues, and this was selected as our target transcriptome.

Patient populations and clinical specimens

Written informed consent was received from all patients involved in this study before collection, and the study was approved by the Institutional Review Board, Chang Gung Memorial Hospital. The 66 paired surgically resected lung cancer and adjacent normal tissue samples (42 males and 24 females; 46 adenocarcinoma and 20 squamous cell carcinoma; stages I to IV) were obtained from patients subjected to surgery at Chang Gung Memorial Hospital. Patients' medical records were reviewed, and we ensured that identities were protected. All tissue specimens were formalin-fixed and paraffin-embedded. Before treatment, serum samples were obtained from 126 patients (83 men and 43 women; range 27-79 years) with lung cancer and 64 healthy control volunteers (42 men and 22 women; range 25-81 years). Control subjects did not display any type of cancer at least 6 months after serum sample collection. Pleural effusion samples were obtained from 82 patients (51 men and 31 women; range 38-85 years) with lung cancer, 23 with breast cancer (23 women; range 34-83 years), 12 with gastric cancer (6 men and 6 women; range 40-72 years) and 13 with colon cancer (6 men and 7 women; range 53-83 years). All the pleural effusion samples obtained from lung cancer and other cancer patients are malignant pleural effusions, which were proved by cancer cells deposited in pleural spaces. Venous blood (10 ml) was collected and centrifuged at 4°C for serum collection. The serum and pleural effusion samples collected were stored at −80°C until further analysis.

Immunohistochemistry and scoring

Immunohistochemistry was performed as described previously.16 Consecutive sections (5 μm thickness) of formalin-fixed paraffin-embedded specimens from NSCLC patients were subjected to H&E staining and immunohistochemical (IHC) analysis with the anti-karyopherin subunit alpha (KPNA)-2 antibody (Proteintech, Chicago, IL) and anti-NBS1 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), respectively. Tissue sections were deparaffinized, treated with 3% hydrogen peroxide for 10 min at room temperature and microwaved in 0.01 M citrate buffer (pH 6.0) for retrieval of antigenicity. Sections were incubated with blocking solution (1% bovine serum albumin in phosphate buffered saline [PBS]) for 20 min at room temperature. Samples were incubated with anti-KPNA2 (1:40) or anti-NBS1 (1:50) overnight at 4°C. Secondary anti-rabbit antibody-coated polymer peroxidase complexes (DAKO Corp., Carpinteria, CA) were applied for 30 min at room temperature, followed by treatment with substrate/chromogen (DAKO Corp., Carpinteria, CA) and further incubation for 5 to 10 min at room temperature. Slides were counterstained with hematoxylin. Stained sections were scored by two independent pathologists (C.-W.W. and Y.L.) according to the intensity and percentage of stain from whole tissue section per specimen at 200× magnification. Score was calculated by multiplying the score of overall percentage of stained cells and the score of median intensity staining, and finally scored as 0, 1+, 2+ and 3+, respectively. Scores of 1+/2+/3+ were defined as positive stain. Scores of 2+/3+ were defined as moderate/strong satin, and −/1+ as negative/weak satin.

Fluorometric sandwich ELISA

KPNA2 protein levels in human serum and pleural effusion were determined using a sandwich ELISA assay developed in-house. Briefly, white polystyrene microtiter plates (Corning, Canton, NY) were coated with an anti-KPNA2 mouse monoclonal antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and blocked with blocking buffer (1% bovine serum albumin/PBS). Serum or pleural effusion samples diluted 1:10 in blocking buffer and various amounts of KPNA2 recombinant protein (Abnova, Taipei, Taiwan) were added to the wells. Each well was incubated with anti-KPNA2 rabbit polyclonal antibody (Proteintech Inc, Chicago, IL). After washing, alkaline phosphatase-conjugated rabbit anti-rabbit IgG (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was added to the individual wells and plates. After six further washes, 4-methylumbelliferyl phosphate (Molecular Probes, Eugene, OR) was added as the substrate, and the fluorescence intensity (excitation: 355 nm, emission: 460 nm) was measured with a SpectraMax M5 microplate reader (Molecular Devices, Sunnyvale, CA).

Gene knockdown of KPNA2 with small interfering RNA

For gene knockdown of KPNA2, 19-nucleotide RNA duplexes for targeting human KPNA2 were synthesized and annealed by Dharmacon (Thermo Fisher Scientific, Lafayette, CO). In brief, CL1-0 and CL1-5 cells were transfected with control siRNA or SMART pool siRNA for KPNA2 (GAAAUGAGGCGUC GCAGAA, GAAGCUACGUGGACAAUGU, AAUCUUACC UGGACACUUU, GUAAAUUGGUCUGUUGAUG) using Lipofectamine RNAiMAX reagents (Invitrogen, Grand Island, NY), according to the protocol provided by the manufacturer. At 48 hr after transfection, cell lysates were prepared for Western blotting to determine gene knockdown efficacy.

Cloning and expression of KPNA2 in lung cancer cell line

The open reading frame of KPNA2 was obtained from CL1-0 cells by polymerase chain reaction using a sense primer, 5′-GTT ATGTCCACCAACGAGAATG-3′, and an antisense primer, 5′-CTAAAAGTTAAAGGTCCCAGGA-3′. The full-length KPNA2 gene fragment was ligated into the pGEM-T easy vector (Promega Corporation, Madison, WI) and designed to introduce an EcoRI site at the initiating methionine and an EcoRI site six base pairs downstream of the stop codon, then subcloned into the eukaryotic expression vector, pcDNA 3.1/myc-His (Invitrogen, Grand Island, NY). The CL1-0 lung cancer cells were transfected with ICA Fectin 441RL reagent (IN CELL ART Biopharmaceutical Company, Nantes, France), according to the manufacturer's procedures. At 48 hr after transfection, cell lysates were prepared for Western blotting to determine gene overexpression efficacy.

Cell viability assay

Cell viability was evaluated with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenytetrazolium bromide (MTT) colorimetric growth assay. Briefly, cells (1 × 103 cells/well) were plated in 96-well plates and cultured for the indicated time intervals. After culture, MTT solution (5 mg/ml) was added and the cells were incubated at 37°C for 1 hr. The supernatant was aspirated, cells treated with dimethyl sulfoxide (100 μl), and absorbance measured at 540 nm using an ELISA reader (Molecular Devices, SpectraMax M2).

Transwell migration assay

Cells transfected with control siRNA or KPNA2 siRNA for 48 hr were harvested by trypsinization, and suspended in serum-free culture medium. Cell migration was assayed using a 24-well format transwell chamber (8.0-μm pore size filter; Corning, Canton, NY). The cell suspension (300 μl; 3 × 104 cells) was added to each insert of the upper chamber, and each lower chamber was filled with 600 μl culture medium containing 10 μl/ml fibronectin. After 6 hr incubation at 37°C, chambers were gently washed twice with PBS and fixed with methanol, followed by Giemsa staining. Cells that had traversed the filter to the lower chamber were counted microscopically (100×) in six different fields per filter.

Statistical analysis

The statistical package, SPSS 13.0 (SPSS Inc., Chicago, IL), was used for all analyses. All continuous variables were expressed as means ± SD. The nonparametric Mann-Whitney U test was used to analyze variations in ELISA results for different clinical characteristics. The chi-square test was used to determine the proportional differences in IHC intensity between clinicopathologic factors. Two-tailed p values of 0.05 or less were considered significant. The receiver operating characteristic (ROC) curve was constructed by plotting sensitivity versus (1 − specificity), considering each observed value as a possible cutoff value. The optimal cutoff point for establishing an accuracy score in biomarker was determined using Youden's index (J) calculated as J = 1 − (false positive rate + false negative rate) = 1 − ((1 − sensitivity) + (1 − specificity)) = sensitivity + specificity − 1.19 The area under the ROC curve (AUC) was calculated as a single measurement to establish the discriminative efficacy of each marker.20

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Generation of a potential adenocarcinoma biomarker dataset by combined analysis of lung cancer cell secretome and transcriptome of pulmonary adenocarcinoma tissues.

To identify potential serum biomarkers for NSCLC, we integrate two secretome datasets obtained from lung cancer cell line (CL1-0 and CL1-5) with the GDS1650 pulmonary adenocarcinoma tissue transcriptome. A schematic representation of the strategy is shown in Supplementary Figure S1. We detected 285 genes displaying more than fourfold upregulation in pulmonary adenocarcinoma tissues, compared with adjacent normal tissues. On combination of these 285 upregulated genes with CL1-0 and CL1-5 cancer cell secretome datasets, 35 potential biomarkers present in single or both secretome datasets were isolated (Supplementary Fig. S1 and Table 1). Among these potential biomarkers, 17 were previously reported as dysregulated proteins/genes in lung cancer and 18 were novel candidates (Table 1). Five novel candidates were identified in both cancer cell secretome datasets, including GINS1, KPNA2, PAFAH1B3, PFKP and SORD. To validate the potential biomarkers for NSCLC, we selected the good candidates according to prediction of secretion pathway, functional classification, commercial availability of antibody, and IHC validation. First, we got commercially available antibodies against seven candidates, including ECM1, KIF23, KPNA2, PAEP, PAFAH1B3, MAD2L1 and SORD for immunodetection. We confirmed the quality of these antibodies by Western blotting and also IHC analysis and found that four antibodies including antibodies against KPNA2, PAFAH1B3, MAD2L1 and SORD could be applied in IHC successfully. Based on preliminary data obtained from IHC (n = 11–33), the frequency of overexpressed KPNA2 in tumor tissue was highest among the four candidates examined (79%, 26/33). According to these results, we choose KPNA2 for further validation using a large sample size of clinical specimen.

Table 1. List of 35 potential biomarkers for NSCLC based on combined analysis of lung cancer cell secretomes and adenocarcinoma microarray dataset
inline image

Overexpression of KPNA2 in NSCLC tissues

To ascertain the clinical significance of KPNA2 in NSCLC, we examined protein expression in 66 NSCLC tissue specimens containing tumors and their adjacent normal tissues by IHC. The results showed that 87.9% (58/66) of lung tumor tissues and only 4.5% (3/66) of the adjacent normal tissues were stained positively for KPNA2 in the nuclei (Table 2). In addition, 28.8% (19/66) of lung tumor tissues and 4.5% (3/66) of adjacent normal tissues displayed KPNA2-positive staining in the cytoplasm. The results clearly indicated that KPNA2 is mainly overexpressed in the nuclei of tumor cells. Furthermore, Spearman correlation analysis demonstrated that the cytoplasmic KPNA2 expression is positively correlated to nuclear KPNA2 expression in lung tumor tissues (Supplementary Fig. S2). One representative case of positive staining of KPNA2 in tumor cell nuclei is shown in Figure 1a. KPNA2 belongs to the karyopherin family and delivers numerous cargo proteins to the nucleus, followed by translocation from nuclear back to cytoplasmic compartments in a Ran-GTP-dependent manner.21 In view of the finding that KPNA2 is predominantly overexpressed in the nucleus in lung tumor cells, we speculate that the function of KPNA2 in cargo protein transport between nucleus and cytoplasm is altered. This speculation is supported by increasing cytosolic fraction of NBS1, one of the cargo protein of KPNA2,22–24 in NSCLC. Analysis of the intracellular distributions of NBS1 and KPNA2 in the same NSCLC tissue specimens revealed that the cytoplasmic NBS1 signal was prevalent in 68.4% (26/38) of KPNA2-positive tumor cells, compared with adjacent normal cells (Fig. 1b, Supplementary Table S1).

Table 2. Immunohistochemical analysis of KPNA2 expression in 66 NSCLC tissue specimens containing both tumor and adjacent normal cells
inline image
thumbnail image

Figure 1. Detection of KPNA2 and NBS1 in NSCLC tissues. (a) IHC analysis of KPNA2 in paired tumor and adjacent normal tissues from one representative NSCLC case. Brown signals indicate the distribution of KPNA2-positive staining in nuclei of tumor cells. (b) Altered distribution of NBS1 in NSCLC tissues with nuclear KPNA2 positive satin. IHC analysis of NBS1 in paired tumor and adjacent normal tissues from one representative NSCLC case with nuclear KPNA2 positive stain. (c) IHC analysis of KPNA2 from two representative cases with well- and poorly differentiated NSCLC (upper panel) and two representative cases of NSCLC with low and high mitosis (low panel) are shown.

Download figure to PowerPoint

No statistical association was observed between KPNA2 expression levels in NSCLC tissues and age, smoking, angiolymphatic invasion and tumor stage of NSCLC (Table 3). Next, multivariate analysis using logistic regression model was performed to evaluate the independence of KPNA2 expression to clinicopatholgic significances. We found that cancer cells with poor differentiation (OR=4.8; 95% CI, 1.03–22.43; p = 0.046) and with high mitosis (OR=5.19; 95% CI, 1.55–17.34; p = 0.007) were recognized as independent determinants for nuclear KPNA2 expression in NSCLC, whereas gender and histological type did not play independent role significantly. Four representative adenocarcinoma cases with different differentiation and mitotic activity were presented in Figure 1c. It is notable that 80% (16/20) of squamous cell carcinoma tissues exhibited moderate/strong nuclear KPNA2 stain (Table 3), indicating that dysregulation of KPNA2 is not limited to adenocarcinomas, but is a common phenomenon in NSCLC.

Table 3. Relationship between tissue KPNA2 expression and clinicopathologic characteristics of 66 NSCLC patients
inline image

Elevation of KPNA2 serum levels in NSCLC patients

As KPNA2 is overexpressed in NSCLC tissues and present in the secretome of lung cancer cell lines, we speculated that it may be detectable in serum samples of NSCLC patients. To determine the serum levels of KPNA2, we developed an in-house fluorometric sandwich ELISA technique, as described in Material and Methods section. To demonstrate accuracy and specificity of in-house KPNA2 ELISA assay, the spike-and-recovery and linearity-of-dilution experiments in both serum and pleural effusion samples were done and the results were shown in Supplementary Tables S2 and S3. The linear dynamic range for detection of soluble KPNA2 ranged from 7.8 to 500 ng/ml. We examined the KPNA2 levels in serum samples collected from NSCLC patients (n = 126; n = 37 for squamous cell carcinoma and n = 89 for adenocarcinoma) and healthy controls (n = 64). As shown in Figure 2a, the serum levels of KPNA2 were significantly higher in NSCLC patients versus those in healthy controls (mean ± SD, 646.8 ± 392.3 ng/ml vs. 485.1 ± 168.5 ng/ml, p < 0.001). We additionally determined the CEA levels in the same serum samples. Serum levels of CEA were higher in NSCLC patients versus healthy controls (10 ± 17.0 ng/ml vs. 0.1 ± 0.6 ng/ml, p < 0.001). Serum levels of KPNA2 and CEA in NSCLC patients (Supplementary Table S4) and in healthy controls (Supplementary Table S5) with different clinical characteristics were shown in additional supporting information. At 5.0 ng/ml CEA, a cutoff value currently used for detection of lung cancer, sensitivity and specificity values were 34.9% and 100%, respectively. Notably, on selection of a cutoff value of 548.89 ng/ml for KPNA2, 38 of 82 lung cancer patients with CEA levels lower than 5.0 ng/ml were further distinguished from healthy individuals. Additionally, ROC curves were constructed for KPNA2, CEA and both proteins together (Fig. 2b). The AUC was determined as 0.63 (95% CI, 0.55–0.71) for KPNA2 and 0.78 (95% CI, 0.72–0.85) for CEA. Importantly, a combination of KPNA2 and CEA displayed higher diagnostic capacity than either marker alone (AUC = 0.89; 95% CI, 0.85–0.94; Fig. 2b). These results collectively indicate that KPNA2 is a potentially useful serum biomarker for NSCLC, particularly in conjunction with CEA.

thumbnail image

Figure 2. KPNA2 levels in serum and pleural effusion from NSCLC patients. (a) Serum levels of KPNA2 in healthy controls (control) and NSCLC patients were determined using ELISA, as described in Material and Methods section. (b) ROC curve analysis of the diagnostic efficacy of serum KPNA2, CEA, and a combination of both proteins. (c) The protein levels of KPNA2 in pleural effusion from patients with NSCLC and non-lung cancer were determined by ELISA, as described in Material and Methods section. (d) ROC curve analysis of the diagnostic efficacy of pleural effusion KPNA2 and CEA. Data are presented as upper and lower quartile and range (box), median value (horizontal line), and the middle 90% distribution (dashed line). A p value less than 0.05 indicates statistical significance using the nonparametric Mann-Whitney U test.

Download figure to PowerPoint

Measurement of KPNA2 levels in pleural effusion from cancer patients

Here, we determine the KPNA2 protein levels in human serum specimens for the first time. KPNA2 levels in pleural effusion, a type of easily accessible body fluid frequently observed in patients suffering from specific cancer types, including lung cancer, breast cancer, gastric cancer and colon cancer, were evaluated. We assessed KPNA2 levels in pleural effusion from 82 NSCLC, 23 breast cancer, 12 gastric cancer and 13 colon cancer patients using ELISA, as described earlier. In NSCLC patients, the KPNA2 levels in pleural effusion were similar to those in serum samples (699.5 ± 684.6 ng/ml vs. 646.8 ± 392.3 ng/ml). Moreover, KPNA2 levels in pleural effusion from NSCLC patients were significantly higher than those from non-lung cancer (699.5 ± 684.6 ng/ml vs. 361± 240.7 ng/ml) (Fig. 2c). The AUC was determined as 0.78 (95% CI, 0.70–0.87) for KPNA2 (Fig. 2d). We also determined the CEA levels in pleural effusion, which were higher than those in serum samples (mean ± SD, 106.0 ± 252.4 ng/ml vs. 10 ± 17.0 ng/ml). However, CEA levels in pleural effusion from NSCLC patients were not significantly higher than those from non-lung cancer (106.0 ± 252.4 ng/ml vs. 29.8± 58.7 ng/ml, p = 0.325). The AUC was determined as 0.55 (95% CI, 0.45–0.66) for CEA (Fig. 2d). Taken together, our data show for the first time that the KPNA2 level in pleural effusion of NSCLC patients is similar to that in NSCLC serum sample and suggest that KPNA2 is an ideal pleural effusion marker to differentiate NSCLC from non-lung cancer.

KPNA2 involves in the migration ability and viability of lung cancer cells

To examine the possible roles of KPNA2 in tumor progression of lung cancer, we applied the siRNA approach to suppress the expression of endogenous KPNA2 in CL1-0 and CL1-5 and assessed the effects on cell migration and survival. A Western blotting showed that KPNA2 protein levels were reduced significantly in cells transfected with KPNA2 siRNA, compared with control siRNA (Fig. 3a). Data obtained from the transwell migration assay (6 hr incubation) indicate that the migration ability of KPNA2 knock-downed cells is severely impaired, compared with that of control cells (Figs. 3b and 3c). The migration capacity was reduced to 24.7% and 47.1% in KPNA2 knockdown CL1-0 and CL1-5, respectively (Fig. 3c). The MTT assay revealed different effects on cell survival as cancer cells transfected with KPNA2 siRNA, compared with those with control siRNA. Cell viability was significantly reduced in KPNA2-knockdown CL1-5 cells, whereas displayed no changes on CL1-0 cells (Fig. 3d). To confirm whether overexpression of KPNA2 increases the number of migrating cells or not, an expression plasmid with full-length KPNA2 was constructed and expressed in CL1-0 cells. Figure 3e showed that overexpression of KPNA2 (2.2-fold change) increased approximately 23.6% of the migration ability in CL1-0 cells when compared with the cells transfected with empty vector. Collectively, our results indicate that KPNA2 is involved in cell migration and survival during tumorigenesis of lung cancer.

thumbnail image

Figure 3. KPNA2 involves in the migration ability and viability of lung cancer cells. (a) Knockdown of KPNA2 inhibits the migration ability of lung cancer cells. Lung cancer cell lines were transfected with control siRNA and KPNA2-specific siRNA, respectively. After transfection for 2 days, cell lysates were prepared and the extracted proteins (20 μg) were analyzed by Western blotting. (b) Simultaneously, cells were subjected to the migration and MTT assays, as described in Material and Methods section. Representative microphotographs of filters obtained from the migration assay. Original magnification: 100×. (c) Quantitative analysis of the migration assay. Data are presented as mean values obtained from three independent experiments. Error bars indicate standard error. A p value less than 0.05 indicates statistical significance using the Mann-Whitney U test. (d) Quantitative analysis of the MTT assay. Data are presented as mean values obtained from three independent experiments. Error bars indicate standard error. A p value less than 0.05 indicates statistical significance using the two-way ANOVA method. (e) Overexpression of KPNA2 enhances the migration ability of lung cancer cells. CL1-0 cells were transfected with pcDNA 3.1/myc-His empty vector or KPNA2/pcDNA 3.1/myc-His vectors. After transfection for 2 days, cell lysates were prepared and proteins were detected by Western blotting (left). Simultaneously, cells were subjected to the migration assay, as described in Material and Methods section. Representative microphotographs of filters obtained from the migration assay (middle). Original magnification: 100×. Data are presented as mean values of cell counts obtained from three independent experiments of migration assay (right). Error bars indicate standard error. A p value less than 0.05 indicates statistical significance using the Mann-Whitney U test.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

NSCLC accounts for almost 80% of lung cancers, of which 50% are adenocarcinoma. To facilitate the discovery of novel biomarkers for NSCLC, we previously selected two adenocarcinoma cell lines with different invasion ability to generate a lung cancer cell line secretome dataset.16 In this study, we tried to narrow down the list of candidate biomarkers by combined analysis with adenocarcinoma trascriptome dataset. Recently, Planque et al.25 identified five candidate lung cancer biomarkers by proteomics analysis of secretomes of four lung cancer cell lines, including H23 (adenocarcinoma), H460 (large cell carcinoma), H1688 (small cell lung cancer) and H520 (squamous cell carcinoma). It is possible that the use of multiple cell lines could be more effective to capture the disease heterogeneity and consequently to identify more useful diagnostic biomarkers. We have compared the two lung cancer lines secretomes (CL1-0 and CL1-5 adenocarcinoma) used in this study with those four lung cancer cell lines secretomes generated by Planque et al. and found that there were 509, 401, 518 and 486 proteins overlapped in our datasets and H23, H460, H1688 and H520, respectively (Supplementary Fig. S3). Among 35 potential biomarkers discovered in this study, there were eight candidates (EEF1A2, HMGB3, KPNA2, MDK, NQO1, PAICS, TOP2A and UCHL1) also identified from H23 adenocarcinoma secretome. Importantly, KPNA2 could be identified from all six lung cancer cell lines, indicating that release of KPNA2 into the conditioned medium is a common phenomenon in lung cancer cell lines derived from different histological types. This observation also supports the notion that analysis of secretomes from multiple cell lines is one of the efficient methods to discover potentially useful biomarker for lung cancer.

CEA is a common tumor markers used clinically for lung cancer detection. The sensitivity and specificity reported ranged from 17% to 78%.26–28 In this study, a combination of serum KPNA2 and CEA (AUC = 0.89) displayed higher diagnostic capacity than CEA alone (AUC = 0.78). Both the mean value and dynamic range of CEA levels in pleural effusion were much higher than those in serum sample. However, unlike the KPNA2 levels in pleural effusion that might be an ideal candidate marker to differentiate NSCLC from non-lung cancer, the CEA levels in pleural effusions from NSCLC patients were not significantly higher than those from non-lung cancer patients. At a cutoff value of 441.2 ng/ml for KPNA2 in pleural effusion, the sensitivity, specificity, positive predict value and negative predict value were 74.4%, 77.1%, 84.5% and 62.7%, respectively. Collectively, we expect the clinical significance of KPNA2 in lung cancer is that serum level of KPNA2 can improve the sensitivity and specificity of CEA and the pleural effusion level of KPNA2 may assist clinician to define malignant pleural effusion with unknown origin.

To our knowledge, KPNA2 is a reported potential tissue biomarker for breast cancer, but has not been investigated in other types of human cancer to date. Dahl et al.29 initially identified KPNA2 as a potential novel prognostic marker in breast cancer after comprehensive transcriptome analysis. Subsequently, Dankof et al.30 examined KPNA2 protein expression in invasive breast carcinoma, which matched peritumoral ductal carcinoma in situ. Recently, Gluz et al. reported that nuclear KPNA2 expression is a predictor of poor survival in patients with advanced breast cancer, irrespective of treatment intensity.31 In this study, we selected KPNA2 as a potential novel biomarker for NSCLC not only due to its unregulated mRNA level in lung cancer tissue but also its secreted characteristic in conditioned medium and unidentified biological function in lung cancer. KPNA2 has never been reported as a secreted protein or body fluid accessible molecule. Based on SecretomeP software program analysis,32 the score of KPNA2 is 0.561 that is higher than the threshold of 0.5, and suggests that KPNA2 is secreted through nonclassical protein secretion pathway. The exact pathway responsible for KPNA2 secretion is currently unknown and remains to be established. It is interesting to note that KPNA2 has been reported to be one of the interacting partners of importin beta 1 (KPNB1),33, 34 which is a nucleocytoplasmic shuttle protein that could be detected in the urinary exosomes and exosomes of colorectal cancer cells.35–37 Thus, interaction between KPNA2 and KPNB1 in cancer cells and secretion of the KPNA2–KPNB1 complex via exosome might represent a plausible explanation why the nuclear transporter protein KPNA2 is secreted or detectable in the cancer cell secretome.

KPNA2 is possibly involved in the regulation of cell proliferation, differentiation, DNA repair and tumorigenesis.24, 29–31, 38 Microarray and quantitative real-time polymerase chain reaction analyses of the mRNA expression profiles in normal human epidermal keratinocytes with either overexpression or knockdown of KPNA2 clearly indicate that KPNA2 is involved in various signal transduction pathways that regulate epidermal proliferation and differentiation.38 In this study, our findings showed that cancer cells with poor differentiation and with high mitosis were recognized as independent determinants for nuclear KPNA2 expression in NSCLC. Gene knockdown of KPNA2 inhibited the migration ability and also viability of lung cancer cell (Fig. 3), supporting the multiple roles of KPNA2 in tumorigenesis. It is interesting that a nucleocytoplasmic shuttle protein, such as KPNA2, is overexpressed in NSCLC tissue and also detectable in human body fluid. As mentioned earlier, KPNB1 has been recently detected in exosomes derived from different origins35–37; however, the extracellular function of importin family, if any, has not been elucidated so far. Whether secreted KPNA2 plays a role in autocrine or paracrine that contributes to tumorigenesis needs further investigation in the future.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
IJC_25568_sm_suppinfofigures.pdf582KSupplementary Figure S1. Schematic illustration of the strategy used to identify potential serum biomarkers for NSCLC. The diagram represents integration of CL1-0 and CL1-5 secretome datasets with the GDS1650 pulmonary adenocarcinoma tissue transcriptome dataset. The CL1-0 and CL1-5 secretome datasets contain 1096 and 1830 proteins, while the GDS1650 transcriptome dataset contains 285 unique genes with tumor/normal ratio > 4-fold. In total, 35 potential biomarkers are presented in single or both secretome datasets. Supplementary Figure S2. The relationship between cytoplasmic and nuclear expression of KPNA2. The cytoplasmic KPNA2 expression was positively correlated to nuclear KPNA2 expression in 66 NSCLC tumor tissues. Spearman correlation coefficient= 0.518 , p<0.001. IHC score, (score of overall percentage of stained cells × score of median intensity staining) ×100. Supplementary Figure S3. Comparison of lung cancer cell lines secretomes used in the current study (Wang et al.) with four other lung cancer cell lines secretomes (Planque et al.) (a) The overlap of proteins identified from CL1-0, CL1-5 (Wang et al.) and H23 (Planque et al.) adenocarcinoma cell lines. (b) The overlap of proteins identified from CL1-0 and CL1-5 cell lines (Wang et al.), and four lung cancer cell lines with different histological backgrounds (Planque et al.).
IJC_25568_sm_suppinfotable.doc95KSupplementary Table S1. Immunohistochemical analysis of NBS1 expression in 38 NSCLC tumor tissues with positive nuclear KPNA2 stain.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.