Using tissue microarray, we investigated MAGEA4 expression in 240 patients who received surgery for non-small cell lung cancer between 1996 and 2004 at the Department of Surgical Oncology, Hokkaido University Hospital, Sapporo, Japan. Patients were classified according to the TNM Classification of Malignant Tumours 7th edition (17). This study was approved by the Ethics Committee of Hokkaido University. Informed consent for the use of tissue samples which were stored prior to the establishment of ethics approval, were originally obtained from the patients at the time of surgery. The Hokkaido University Institutional Review Board confirmed that this study was fully ethically compliant and the second informed consent was waived by the ethics committee of our institution.

The 293F and 293FT cells were obtained from Invitrogen (Carlsbad, CA, USA), while the human lung squamous cell carcinoma cell lines, PC10, H226, LK2 and LC-1, were obtained from the Japanese Cancer Research Resources Bank (Ibaraki, Japan), along with lung adenocarcinoma cell lines, A549, REPF-LC-MS, VMRC-LCD and ABC-1. The cells were grown at 37°C in a humidified incubator with 5% CO₂, and in culture medium with 10% fetal bovine serum and 1% penicillin/streptomycin.

MAGEA genes were amplified from human testis cDNA (Takara Bio, Inc., Tokyo, Japan) using suitable PCR primers and high-fidelity KOD-Plus-polymerase (Toyobo, Tokyo, Japan). Amplicons were then inserted into pcDNA-IRES-GFP vector by restriction enzyme cloning using T4 ligase (Promega, Madison, WI, USA). The vector is a pcDNA3.1(+) plasmid (Invitrogen) that contains an internal ribosomal entry site and green fluorescent protein (GFP) originally synthesized in our previous study (18). Following plasmid amplification in JM-109 competent cells (Takara Bio, Inc.), inserts were confirmed by full sequencing (Hokkaido System Science Co., Ltd., Sapporo, Japan).

The 293F or 293FT cells were transfected with the MAGEA expression plasmids using Lipofectamine LTX (Invitrogen), following the manufacturer' s instructions. Transfection was confirmed by GFP fluorescence. Transfectants were then used to optimize MAGEA detection by RT-qPCR, western blot analysis and immunohistochemistry. Untransformed cells were used as negative controls, along with cells transfected with the empty vector. Transfected cells were exposed for 24–48 h to 1 μM cisplatin (Bristol-Myers Squibb, New York, NY, USA) when investigating subcellular localization under cytotoxic conditions.

Total RNA was extracted from the transfected cells or non-small cell lung cancer cells using TRI reagent (Sigma-Aldrich, St. Louis, MO, USA) according to the manufacturer's instructions, and digested with RQ1 RNase-Free DNase (Promega). cDNA was then reverse transcribed from 2 μg RNA using the Superscript II or Superscript VILO cDNA Synthesis kit (Invitrogen), and digested with Ribonuclease H (Takara Bio, Inc.). MAGEA4 was then amplified with GoTaq^® DNA Polymerase (Promega), using the primers MAGEA4_forward, 5′-ATGTCTTCTGAGCAGAAG AGTCAGC-3′ and MAGEA4_reverse, 5′-TCAGACTCCCTCT TCCTCCTCT-3′. The reaction consisted of pre-heating at 94°C for 5 min and 30 cycles of denaturation at 94°C for 30 sec, annealing at 60.4°C for 30 sec, extension at 72°C for 1 min, followed by final extension for 5 min at 72°C. Subsequently, MAGEA4 expression was assessed by quantitative RT-PCR on an ABI PRISM 7000, using Power SYBR^®-Green PCR Master Mix (both from Life Technologies/Applied Biosystems, Carlsbad, CA, USA), the internal probe 5′-CTGGAGCATGT GGTCAGGGTCAAT-3′ and the reverse primer used in the initial PCR reaction. Expression was normalized to β-actin, which was amplified with the internal probe, 5′-CCAGGC TGTGCTATCCCTGTACGC-3′ and reverse primer, 5′-ACC GGAGTCCATCACGATGC-3′.

CB6F1 mice were immunized with human recombinant MAGEA4, and clone designated as MCV-1 was affinity-purified with protein G. MCV-1 was selected based on epitope specificities of hybridoma supernatants, as assessed by ELISA. This process was conducted at Mie University, Tsu, Japan, and purified antibody to MCV-1 was supplied by the University.

CB17/SCID mice were purchased from Charles River Laboratories Japan (Yokohama, Japan). All mice were female, 4–6 weeks of age, and were maintained under specific pathogen-free conditions and were treated under the guidelines of the Hokkaido University Institutional Animal Care and Use Committee. PC10, H226, LK2, LC-1, A549, REPF-LC-MS, VMRC-LCD and ABC-1 cells (5×10⁶) were subcutaneously injected in a volume of 100 μl of phosphate-buffered saline into the left flank region of each mouse. The mice were monitored once every 2 or 3 days after the injection. When the tumor diameter exceeded 10 mm, the mice were sacrificed and the tumors were separated into 2 blocks: one block was frozen using liquid nitrogen to extract proteins for western blot analysis, and the other was immersed in formalin for immunohistochemical analysis. If multiple tumors were developed in a mouse, the largest one was used as a sample. All animal experiments were conducted with the Institutional Animal Care and Use Committee approval at that time.

Lysates from cell lines and from SCID mouse xenografts were prepared in SDS buffer containing 62.5 mm Tris-HCL (pH 6.8), 2% w/v SDS, 10% glycerol, 50 mm DTT, 0.1% w/v bromphenol blue and 1 mm PMSF. Protein concentration was measured by the Bradford method using a commercial protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA). Heat-denatured cell lysate (10 μg) were electrophorased in 15% SDS-polyacrylamyde gels and were blotted on nitrocellulose membranes. They were probed for 1 h at room temperature with 1:1,000 dilutions of monoclonal antibodies to MAGEA4 (MCV-1), β-actin (#MAB1501, clone C4; Millipore, Temecula, CA, USA), a control for protein loading, and GFP (#632375, Living Colors GFP Monoclonal Antibody; Clontech, Mountain View, CA, USA), a marker of transfection. The blots were then labeled for 1 h at room temperature with a 1:10,000 dilution of peroxidase-conjugated goat anti-mouse IgG (#115-035-003; Jackson ImmunoResearch Laboratories, West Grove, PA, USA) and visualized with ECL Plus Western Blotting Detection Reagents (Amersham Biosciences, Buckinghamshire, UK).

Formalin-fixed, paraffin-embedded cell lines, non-small cell lung cancer xenografts and clinical specimens were evaluated by immunohistochemistry based on streptavidin, biotin and peroxidase. The samples were labeled overnight at 4°C with monoclonal antibodies to MAGEA4 (MCV-1, 2.8 mg/ml) at a dilution of 1:2,000. To stain p53, the samples were labeled for 1 h at room temperature with mouse monoclonal antibodies to human p53 (#M7001, DO-7; Dako Japan, Tokyo, Japan). After washing, the sections were incubated for 30 min at room temperature with peroxidase-labeled goat anti-mouse and anti-rabbit polyclonal IgG (Fab'), and then with Histofine Simple Stain MAXPO (MULTI) (Nichirei, Tokyo, Japan). After further washing, specimens were visualized with freshly prepared 3,3′-diaminobenzidine tetrahydrochloride supplied with the Histofine SAB-PO (M) kit (Nichirei), counter-stained with hematoxylin and mounted. Secimens probed with a mix of mouse isotype IgG1 and IgG2a (Dako, Glostrup, Denmark) at a dilution of 1:20 were used as negative controls, whereas normal human testis (#T2234260, Paraffin Tissue Section; BioChain, Newark, CA, USA) was used as a positive control. Apoptosis was assessed by immunostaining for cleaved caspase-3, using Autostainer Plus (Dako, Carpinteria, CA, USA) and a rabbit monoclonal antibody to cleaved caspase-3 (Asp175) (5A1E) (#9664; Cell Signaling Technology, Danvers, MA, USA), which was diluted 1:100 in S2022 Antibody Diluent (Dako). Specimens were then reacted for 60 min at room temperature with the EnVision^™ + Dual Link Detection kit (K4063; Dako), a polymer-based detection system with 3′,3′-diaminobenzidine as the chromogen. Staining was quantified in five fields per specimen using Aperio Image Scope software (Leica Biosystems, Nussloch, Germany) and averaged.

We selected three cancer spots and two non-cancerous spots from paraffin-embedded resected non-small cell lung cancer tissues in each of the 240 patients described above. These specimens were arrayed in a second paraffin block using Tissue Arrayer (Beecher Instruments, Alphelys, Plaisir, France) with a diameter of 0.6 mm. MAGEA4 and p53 staining in 1,200 samples from 240 different tissues was independently assessed by A.F.-K. and the late Dr Masaki Miyamoto. The staining intensity was evaluated by scoring ('−' for 0%; '+' for 1–100%) in both the nucleus and cytoplasm, and we considered staining intensity as positive if any of three cancerous spots indicated positive (+) expression for MAGEA4 or p53. p53 staining was evaluated only with respect to nuclear expression as all specimens expressing cytoplasmic p53 also expressed nuclear p53. A specimen was considered positive if judged by both investigators as positive.

The association between MAGEA4 and clinicopathological variables was evaluated using the χ² test with StatView J version 5.0 software (SAS Institute Inc., Cary, NC, USA). Apoptotic levels based on cleaved caspase-3 were compared using the Student's t-test. The overall survival was assessed by Kaplan-Meier analysis and the log-rank test. Uni- and multivariate regression was performed using Cox's proportional hazards model. P-values <0.05 were considered to indicate statistically significant differences.

MAGEA4 was detected by RT-qPCR in 293FT cells transfected with MAGEA4 expression plasmid, but not in cells transfected with other MAGEA genes (Fig. 1A). Endogenous MAGEA4 expression was also detected in four (PC10, LC-1, REPF-LCMS and VMRC-LCD) of eight non-small cell lung cancer cell lines (50.0%, Fig. 1B). The endogenous expression of other MAGEA genes was also quantified (Table III).

In western blots of 293FT cells transfected with various MAGEA genes, the anti-MAGEA4 antibody, MCV-1, crossreacted with MAGEA2B and MAGEA12 (Fig. 1C, upper panel). However, MCV-1 detected only the 37 kDa MAGEA4 protein in the cell lines, as determined by RT-qPCR. The cells endogenously expressed the MAGEA4 gene, but not MAGEA2B (at 35 kDa) (Fig. 1C, lower panel). Although MCV-1 also crossreacted with MAGEA2B and MAGEA12 when used for immunohistochemistry of the transfected cells (Fig. 2A), the antibody detected only MAGEA4 in MAGEA4-positive cell lines (Fig. 2B). For example, MCV-1 did not stain xenografted LK2 tissue, which was shown to express MAGEA2B, but not MAGEA4 by RT-qPCR (Fig. 1B and Table III). Collectively, these results indicate that MCV-1 specifically detects endogenous MAGEA4 (Table IV).

MAGEA4 was shown to be accumulated in the cytoplasm of transfected 293FT cells, as assessed by fluorescence from the genetically fused GFP (Fig. 2C and D). However, endogenous MAGEA4 was detected not only in the cytoplasm, but also in the nuclei of xenografts derived from MAGEA4-positive cell lines (Fig. 2B). Notably, exposure for 24–48 h to 1 μM cisplatin, a cytotoxic anticancer agent, did not alter the intracellular localization of MAGEA4 in transfected 293F cells (Fig. 3).

Using immunohistochemistry, we examined caspase-3 activation in stably transfected 293F cells to investigate whether MAGEA4 expression induces or inhibits apoptosis in response to genotoxic stress. We found that cells expressing cleaved caspase-3 were significantly fewer in number in the cultures transfected with MAGEA4 than in the cultures transfected with the empty vector, with (P=0.0078) or without (P=0.043) exposure to cisplatin for 24 h (Fig. 4). These results suggest that MAGEA4 inhibits apoptosis via caspase.

MAGEA4 was detected in 61 of 240 patients (25.4%); 33 cases in the cytoplasm only, 12 cases in the nucleus only and 16 cases in both the cytoplasm and nucleus (Table V). MAGEA4 expression was significantly associated with male patients, of whom 31% were positive for MAGEA4, although only 18% of female patients exhibited positivity (P=0.0251). MAGEA4 expression was also associated with squamous cell carcinomas, with MAGEA4 detected in 47% of cases, but in only 16% of adenocarcinomas (P<0.0001). Finally, the positive expression rate was significantly higher in tissues with an advanced pathological stage, being detected in 21, 32 and 34% of patients in stage I, II and III/IV disease (P=0.0309), respectively (Table VI). Of note, cytoplasmic MAGEA4 was also associated with the male sex (P=0.0049) and squamous cell carcinomas (P<0.0001). However, no significant association was found between nuclear MAGEA4 expression and clinicopathological variables (Table VI). A representative immunohistochemical analysis of cytoplasmic and nuclear MAGEA4 expression is shown in Fig. 5A. Of the 240 tumor specimens, 12 (5.0%) exhibited nuclear, but not cytoplasmic MAGEA4 expression, whereas 33 specimens (13.8%) exhibited cytoplasmic, but not nuclear MAGEA4 expression. Both the nuclear and cytoplasmic forms were detected in 16 cases (6.7%, Table V). Patients expressing only nuclear or only cytoplasmic MAGEA4 exhibited a significantly poorer overall survival (P=0.0424 and P=0.0340, respectively) (Fig. 5B), than those with neither. Intriguingly, the accumulation of both the nuclear and cytoplasmic forms was not prognostic (P=0.9101, Fig. 5B). Therefore, we hypothesized that the intracellular localization of MAGEA4 was functionally significant, and was thus associated with prognosis.

In light of the link between the MAGEA4 subcellular localization and prognosis, we also examined the association between p53 and MAGEA4, as MAGEA proteins have been reported to be in complex with p53 at p53 cognate sites in chromatin (5). The p53 status was immunohistochemically surveyed in clinical specimens (Fig. 6A). p53 expression itself did not have significant prognostic value (Fig. 6B). In addition, there was no difference in survival among patients without nuclear MAGEA4, regardless of p53 expression. However, patients with nuclear MAGEA4 expression, but not p53 expression exhibited a significantly poorer survival than others (P=0.0017; Fig. 6C, left panel). Conversely, the survival rate was 100% in patients with nuclear MAGEA4 and p53 expression (Fig. 6C, right panel). Moreover, the patients with lung adenocarcinoma with nuclear MAGEA4 expression, but not p53 expression exhibited a shorter overall five-year survival than patients with other forms of adenocarcinoma (P=0.0012, Fig. 7). A similar trend was observed in patients with squamous cell carcinomas, although the difference was not statistically significant (P=0.1113).

Uni- and multivariate analysis implied that an advanced pT status (P= 0.0001), advanced pN status (P<0.0001), non-adenocarcinoma histology (P=0.0436) and the accumulation of nuclear MAGEA4, but not p53 expression (P=0.0022) were significantly associated with a poor prognosis (Table VII). Multivariate analysis also showed that expression of nuclear MAGEA4 but not p53 was an independent prognostic factor in patients with non-small cell lung cancers (P=0.0116), as were pT (P=0.0066) and pN status (P=0.0035, Table VII).