We obtained five human mesothelioma cell lines (211H, H28, H226, H2052, and H2452) and the human mesothelial cell line MeT-5A from American Type Culture Collections (Manassas, VA, USA). Two human mesothelioma cell lines ACC-MESO-1 (MESO1) and ACC-MESO-4 (MESO4) were obtained from RIKEN Cell Bank (Tsukuba, Japan). The cells were maintained in RPMI-1640 medium supplemented with 5% FBS (JRH Biosciences, Lenexa, KS, USA) in a humidified incubator maintained at 37°C with 5% CO₂.

We obtained an anti-COPA antibody by immunizing rabbits with the synthetic peptide CAQFHPTEDL VVSASLDQTV that is a component part of human COPA. The membrane and cytoplasmic protein fractions were separated using a ProteoExtract Transmembrane Protein Extraction kit (Merck Millipore, Billerica, MA, USA). Western blotting was conducted using anti-COPA or anti-β-actin antibody (Sigma).

The unknown protein band was detected by western blotting using the anti-COPA antibody, and gel pieces of the corresponding size were excised. Proteins in the gel pieces were analyzed by a liquid chromatography-tandem mass spectrometry (LC-MS/MS) system and mass spectrometry data were processed to provide protein identifications using the non-redundant NCBI protein database with MASCOT software (Matrix Science, Boston, MA, USA). LC-MS/MS analysis and database search were conducted by APRO Science (Tokushima, Japan).

We synthesized first-strand cDNAs from the cells using a FastLane Cell cDNA kit (Qiagen, Hilden, Germany). Predesigned and preoptimized TaqMan probes to detect AHNAK, non-erythroid α-spectrin, and 18S rRNA were purchased from Applied Biosystems (Foster City, CA, USA). Real-time RT-PCR was conducted in triplicate using a Premix Ex Taq reagent (Takara Bio, Otsu, Japan). Gene expression levels were normalized to 18S rRNA expression in each sample. Three independent experiments were conducted.

Cells were grown on glass coverslips and fixed with 4% paraformaldehyde. Immunofluorescence staining was conducted using anti-AHNAK (Abcom, Cambridge, UK) and Alexa Fluor 594 goat anti-mouse antibodies (Molecular Probes, Eugene, OR, USA). The coverslips were mounted in mounting medium with DAPI (Vector Laboratories, Burlingame, CA, USA). Fluorescence images were captured with an exposure time of 1/1.5 sec for AHNAK.

The animal experiment was approved by the Institutional Animal Care and Use Committee of the National Institute of Radiological Sciences. Mesothelioma xenografts were established by injecting mesothelioma cells subcutaneously into the flanks of BALB/c-nu/nu male mice (CLEA Japan, Tokyo, Japan).

Mesothelioma tissue specimens were resected from 10 patients (six sarcomatoid, two epithelioid, and two biphasic subtypes) in Juntendo University Hospital. This study was approved by the Institutional Review Boards of Juntendo University School of Medicine and the National Institute of Radiological Sciences.

Xenografted tumors and human specimens were fixed in 10% neutral buffered formalin and embedded in paraffin for sectioning. The tissue sections were stained with the anti-AHNAK or anti-ERC/mesothelin (22A31) (10) antibody and counterstained with hematoxylin. The adjacent sections were stained with hematoxylin-eosin (H&E) staining.

Stealth RNAi siRNAs for AHNAK and negative control were purchased from Applied Biosystems. The cells were transfected with 50 nM siRNA using Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA, USA). Forty-eight hours after transfection, the knockdown efficiency was confirmed by quantitative real-time PCR as mentioned above.

Migration and invasion assays were conducted in 24-well plates using 8-μm pore size inserts (Becton-Dickinson Bioscience, Bedford, MA, USA) without Matrigel coating (for the migration assay) and with Matrigel coating (for the invasion assay). The inserts were placed into the wells of the 24-well plates containing RPMI-1640 with 10% FBS. Cells (5×10⁴ cells) in RPMI-1640 with 1% FBS were seeded to the inserts. After 22 h, non-migrating and non-invasive cells were removed from the top of the filter, and migrating cells and invasive cells on the bottom of the filter were fixed with 100% methanol and stained with 0.125% crystal violet (Wako Pure Chemical Industries). The cells were counted in five randomly selected fields at ×200 magnification.

The data were statistically analyzed by ANOVA with the Dunnett’s multiple comparison test, or by Student’s t-test.

Western blotting with the anti-COPA antibody for membrane and cytosolic protein fractions of 211H and MeT-5A cells detected a protein band with the same molecular weight as COPA (140 kDa) in the cytosolic fraction of 211H and MeT-5A cells (Fig. 1). In addition, another protein band with a smaller molecular weight (120 kDa) compared to COPA was detected only in the membrane fraction of 211H cells (Fig. 1). To explore whether this protein band is a fragment of COPA or a different protein, the piece corresponding to the protein band of 120 kDa was excised from the gel and subjected to LC-MS/MS analysis. According to the results of a protein database search of mass spectrometry data, the peptides in the gel piece of 211H and MeT-5A cells were matched to nine and 12 proteins, respectively. The identified proteins are listed in Table I with their accession number (GI no.), name, total score and matched peptide number. Four proteins (talin, fatty acid synthase, filamin A and nuclear pore complex protein Nup214) were detected in both 211H and MeT-5A cells. Although filamin A α (actin binding protein 280), isoform CRA_a in 211H cells, and filamin-A isoform 1 in MeT-5A cells, have a distinct accession number (119593150 and 116063573, respectively), these two proteins are isoforms of filamin A and these amino acid sequences are 99.5% identical. Five other proteins (neuroblast differentiation-associated protein AHNAK, non-erythroid α-spectrin, Treacher Collins syndrome, pro-ubiquitin and protein BAT2-like 2) were identified only in 211H cells. The remaining eight proteins (β-spectrin, spectrin-α, β-filamin, non-muscle myosin heavy chain, retinoid-acid induced protein 1, plectin, NUP210 protein, and eukaryotic protein synthesis initiation factor) were identified only in MeT-5A cells. There was no peptide sequence matching COPA. Between the five proteins identified only in 211H cells, the total score and matched peptide number of AHNAK (total score, 1,186; matched peptide number, 24) were markedly higher than those of the others. Although AHNAK and non-erythroid α-spectrin were matched by multiple peptides (24 and 5 peptides, respectively), three other proteins (Treacher Collins syndrome, pro-ubiquitin, and protein BAT2-like 2) were matched by only one peptide.

The mRNA expression of AHNAK and non-erythroid α-spectrin in mesothelioma and mesothelial cell lines was measured by real-time RT-PCR. AHNAK mRNA expression in all seven mesothelioma cell lines (211H, H28, H226, H2052, H2452, MESO1 and MESO4) was significantly higher compared with that in the mesothelial cell line MeT-5A (P<0.05) (Fig. 2A). AHNAK mRNA levels of mesothelioma cell lines increased by 1.5- to 5.1-fold in comparison with that of MeT-5A (Fig. 2A), whereas the non-erythroid α-spectrin mRNA level of 211H cells was less than half of MeT-5A (data not shown). By immunofluorescence staining, a high protein expression of AHNAK was observed in all mesothelioma cell lines, and particularly in H2052 cells (Fig. 2B), whereas the AHNAK expression in MeT-5A cells was faint (Fig. 2B). The protein expression levels seemed to be correlated with the mRNA levels.

To examine the protein expression of AHNAK in xenograft tumors, the mesothelioma cell lines were injected subcutaneously into the flank of nude mice. Four cell lines (211H, H226, MESO1 and MESO4) formed tumors at the site of inoculation, but the remaining three cell lines (H28, H2052, and H2452) failed to induce tumor formation. AHNAK protein was expressed in all four xenograft tumors as determined by immunohistochemical staining using the anti-AHNAK antibody (Fig. 3). The cytoplasm and plasma membrane of 211H tumors was intensely stained with the anti-AHNAK antibody and the plasma membrane of the other xenografts was intensely stained. The epithelioid marker ERC/mesothelin was detected in H226, MESO1 and MESO4 xenografts, but not in 211H. The 211H xenograft had spindle elongated cell morphology and was negative for ERC/mesothelin, indicating that it formed a tumor with sarcomatoid features. The other xenografts showed glandular and solid epithelial patterns and were positive for ERC/mesothelin, indicating that they formed tumors with epithelioid features.

We assessed the AHNAK and ERC/mesothelin expression in resected specimens from patients with sarcomatoid, epithelioid, or biphasic mesothelioma by immunohistochemical staining (Fig. 4). In the sarcomatoid specimens, the cytoplasm and plasma membrane of the tumor cells were intensely stained with the anti-AHNAK antibody, but the intensity of the plasma membrane tended to be stronger compared with that of the cytoplasm. ERC/mesothelin expression was not observed. In the epithelioid specimens, strong expression of both AHNAK and ERC/mesothelin was detected on the plasma membrane, but not in the cytoplasm. In one of two biphasic specimens, strong AHNAK expression was observed on the plasma membrane, but not in the other biphasic specimen. ERC/mesothelin expression was not observed in either of the biphasic cases. The summary of immunohistochemical staining in 10 specimens is shown in Table II. The expression of AHNAK was detected in 90% (9/10) of all mesothelioma specimens: 100% (6/6) of sarcomatoid, 100% (2/2) of epithelioid and 50% (1/2) of biphasic. The ERC/mesothelin expression was detected in 0% (0/6) of sarcomatoid, 100% (2/2) of epithelioid and 0% (0/2) of biphasic mesothelioma specimens.

We assessed the cell migration and invasion ability of the mesothelioma cell lines. The number of migrating cells in six of the mesothelioma cell lines (211H, H28, H226, H2452, MESO1 and MESO4) was 4.3- to 12.5-fold higher than that in MeT-5A (P<0.01) (Fig. 5A). No significant difference was observed between H2052 and MeT-5A (Fig. 5A). On the other hand, the number of invading cells in all seven mesothelioma cell lines was 33.5-to 143.9-fold higher than that in MeT-5A (P<0.01) (Fig. 5B).

To examine the association of AHNAK and the cell migration or invasion in mesothelioma cells, we conducted migration and invasion assays combined with AHNAK knockdown using the siRNA technique. The knockdown efficiency of the siRNA on AHNAK mRNA expression was measured by real-time RT-PCR at 48 h after transfection. The AHNAK mRNA expression in AHNAK siRNA-transfected mesothelioma cells was reduced to <25 or 23% compared with that in the mock-transfected cells (without siRNA) or negative control siRNA-transfected cells, respectively (Fig. 6). In six mesothelioma cells (211H, H28, H226, H2052, H2452 and MESO4), the number of migrating cells in AHNAK siRNA-transfected cells was decreased to <21% compared with that in cells transfected with the negative control siRNA at 48 h after transfection (Fig. 7A). Although the number of migrating cells in MESO1 transfected with the AHNAK siRNA was decreased to 62% compared with that in cells transfected with the negative control siRNA, there was a significant difference (P<0.01) (Fig. 7A). In five mesothelioma cells (211H, H28, H226, H2052 and H2452), the number of invading cells in AHNAK siRNA-transfected cells was reduced to <26% compared with that in cells transfected with the negative control siRNA (Fig. 7B). The number of invading cells in MESO1 and MESO4 cells transfected with the AHNAK siRNA was reduced to <63 and 46% compared with that in each negative control siRNA-transfected cell, respectively (Fig. 7B). However, in all cell lines, significant differences were observed between the AHNAK siRNA-transfected and negative control siRNA-transfected cells (P<0.05 in 211H, H2452 and MESO1; P<0.01 in H28, H226, H2052 and MESO4) (Fig. 7B).