A comprehensive expression analysis of the MIA gene family in malignancies: MIA gene family members are novel, useful markers of esophageal, lung, and cervical squamous cell carcinoma

Melanoma inhibitory activity (MIA) gene family members include MIA, MIA2, and Transport and Golgi organization protein 1 (TANGO). Although MIA gene family members have several tumor-related functions, their detailed roles in malignancies remain poorly elucidated. In this study, 477 tumor specimens were subjected to immunohistochemical screening to evaluate MIA gene family expression. For a validation analysis, we also examined the association between MIA gene family expression and clinicopathological factors in 66 cases of esophageal cancer, 145 cases of lung cancer, and 126 cases of cervical cancer. The frequency of MIA gene family expression was higher among squamous cell carcinomas than among other tumor types subjected to screening. In the validation analysis, MIA gene family staining was observed frequently in esophageal and lung cancers associated with nodal and/or distant metastasis. In cervical cancers, MIA and TANGO immunostaining also correlated with tumor progression and metastasis. Furthermore, MIA2 expression levels in invasive cervical cancer were upregulated relative to those in cervical intraepithelial neoplasia 3. A disease-free survival analysis revealed that MIA-, MIA2, or TANGO-positive patients had a significantly shorter disease-free survival than did those patients who were negative. Our results suggest that MIA, MIA2, and TANGO may be useful diagnostic and therapeutic molecular targets in human malignancies.


INTRODUCTION
An estimated 455 800 and 527 600 new cases and 400 200 and 265 700 deaths related to esophageal and cervical cancer, respectively, have been reported worldwide [1]. Notably, postoperative recurrences occur in approximately half of all patients with esophageal squamous cell carcinoma (ESCC) [2], and it is the third leading cause of cancer deaths among women in developing countries [1]. Additionally, an estimated 1.8 million new lung cancer cases have occurred worldwide, accounting for approximately 13% of all cancers [1]. Therefore, the early detection of such malignancies is urgently necessary.
The melanoma inhibitory activity (MIA) gene family includes MIA, MIA2, Transport and Golgi organization protein 1 (TANGO), and otoraplin (OTOR). Members of this family share 34%-45% amino acid homology and 47%-59% cDNA sequence homology and feature a highly conserved SH3-like domain and hydrophobic N-terminal secretory signal sequences [3][4][5][6]. Although OTOR expression is highly restricted to healthy eyes, cochlea, and cartilage [7], other members of the MIA gene family have several tumor-related functions. MIA expression correlates with cancer cell detachment, migration, invasion, and apoptotic repression and is accordingly related to malignant tumor progression, metastasis, and poor prognosis [8][9][10][11][12]. MIA2 is induced in liver fibrosis or cirrhosis by activating transforming growth factor-beta (TGF-β) signaling [13,14] and serves as a tumor suppressor in liver cancers following the loss of hepatocyte nuclear factor-1 (HNF-1) expression [15]. However, wild-type MIA2 promotes the loss of chemosensitivity in pancreatic cancers, thus worsening an already poor prognosis [16]. Regarding other MIA gene family members, TANGO has been suggested as a tumor suppressor in malignant melanoma, colorectal cancer, and hepatoma [4,6]. In summary, the functions of MIA gene family members in malignancies have not been well documented.
Tumor biomarkers have been classified as screening (used diagnostically to identify patients), staging (used to stage disease), prognostic (used to predict outcome), and predictive and monitoring markers (used to speculate and observe clinical responses to any treatment) [21,22]. Cancer biomarkers must also satisfy the following conditions: (1) the transition can be objectively determined the quality; (2) must be measureable in small sample amounts; (3) must be altered in tumors but not in normal tissues; and (4) must be altered at an early phase of cancer development [22,23]. However, the role of MIA gene family as tumor markers in various human malignancies remains controversial. The purpose of this study was to investigate the usefulness of MIA gene family as novel tumor markers in various human neoplastic specimens, including ESCC, lung cancer, and cervical cancer.

Screening for MIA gene family expression in human tumors
We initially used immunohistochemistry to examine the expression of MIA gene family members in 477 cases of different tumors. The specificity of the antibodies for MIA gene family was confirmed by Western blotting with recombinant proteins (data not shown). These results are summarized in Table 1. Briefly, higher MIA, MIA2, and TANGO expression levels were observed in 80 (16.8%), 67 (14.1%), and 76 (15.9%) of these cases, respectively. All immunopositive cases exhibited cytoplasmic MIA gene family staining. Several representative images of MIA gene family immunostaining in tumors are shown in Figure 1A to 1I.

Association between MIA gene family expression and clinicopathological characteristics in lung cancers
A summary of the results pertaining to lung cancer is shown in Table 3. Non-cancerous lungs did not overexpress MIA gene family members; in contrast, cytoplasmic MIA, MIA2, and TANGO expression was found in 49 (33.8%), 45 (31%) and 47 (32.4%) of the 145 cases, respectively ( Figure

Gene expression of MIA gene family and secretion of MIA in esophageal, lung, and cervical cancers
Next, we verified the expression of MIA family genes in cases with esophageal, lung, and cervical cancers. In malignancies, expression levels of MIA, MIA2, and TANGO were significantly higher than in non-tumorous specimens ( Figure 5A). Moreover, the expression of MIA family genes was significantly associated with immunohistochemical grade in esophageal, lung, and cervical cancers ( Figure 5B). Expression levels of MIA gene family in primary tumor and metastatic sites remained unchanged (data not shown).
Next, MIA gene family expression levels were compared between serum samples and tumor specimens. Serum secretion levels of MIA measured by enzymelinked immunosorbent assay (ELISA) were well correlated with those of tumor expression levels quantified by quantitative (qRT-PCR) ( Figure 5C).

Disease free survival analysis of esophageal, lung, and cervical cancers
Finally, we conducted a Kaplan-Meier survival analysis. We found that patients with ESCC whose samples exhibited positive MIA, MIA2, and TANGO immunostaining had significantly shorter disease-free survival intervals, compared to patients with negative expression (P < 0.0001, P = 0.0135, and P = 0.0131, respectively; Figure 6A-6C). Among patients with lung cancer, those with MIA, MIA2, and TANGO-positive samples had a significantly worse disease free survival than did those with negative samples (P < 0.0001, P < 0.0001, and P = 0.0006, respectively; Figure 6D-6F). Furthermore, MIA (P < 0.0001), MIA2 (P = 0.0144), and TANGO expression (P = 0.0151) were associated with a poor prognosis among patients with cervical cancer (Figure 6G-6I).

DISCUSSION
Although MIA gene family members serve several tumor-related functions, to our knowledge, this is the first report to subject a variety of human malignancies to semicomprehensive immunohistochemical MIA gene family expression profiling. In this investigation, we found that MIA gene family members are frequently expressed in several types of human tumors, including SCCs. We also confirmed the significance of MIA gene family expression in ESCC, lung cancer, and cervical cancer. In particular, cases of ESCC and lung cancer with nodal and/or distant metastases were frequently positive for MIA gene family expression; similarly, lung SCCs were frequently positive for these proteins. In addition, MIA gene family expression was also associated with a poor prognosis among cancer patients. However, further research is needed to determine the association between MIA gene family expression and clinicopathological significance in tumors. TMA has recently become as powerful tool for large-scale expression analysis. TMA immunohistochemistry is a valuable high-throughput analysis technique because it eliminates technical variations among cases by subjecting all tissue cores to equal immunostaining conditions [24]. Additional immunohistochemical analyses of MIA gene family expression using large numbers of TMA slides will likely be effective.
In the present study, we have demonstrated the expression of MIA in previously uninvestigated tumors.
Notably, MIA promotes cell detachment, migration, and invasion and suppresses cancer cell apoptosis and lymphokine activated killer cell (LAK) infiltration. In addition, MIA is a ligand for the cell surface receptors integrin α 4 β 1 /α 5 β 1 and binds to fibronectin via SH3 domain-like structures to inhibit cell-to-stromal adhesion [9,25,26]. In a previous report, we described the activation of MIA via interactions of intracellular HMGB1 with NFkBp65 and observed the strong implication of MIA in tumor progression and nodal metastasis through the induction of angiogenesis and lymphangiogenesis in OSCC [17,18]. MIA expression has also been observed in malignant melanoma, gastric cancer, pancreatic cancer, breast cancer, chondrosarcoma, glioma, and OSCC [7-12, 17, 18, 27-30].
Several reports have revealed that MIA2 and TANGO can act as tumor suppressors [4][5][6]15]; it is therefore interesting that according to our current results and previous findings, MIA2 and TANGO might behave as proto-oncogenes in SCCs of the esophagus, lung, and cervix [19,20]. These potential oncogenic roles might depend on receptor-related signaling differences. Indeed, we revealed that signaling through the MIA2-integrin α 5 β 1 -JNK pathway promotes apoptosis, whereas signaling through the MIA2-integrin α 4 β 1 -MAPK p38 pathway suppresses apoptosis [19]. Furthermore, MIA2 inhibits lymphocyte infiltration into tumors by binding integrin α 4 , thus dysregulating the host immune system [19]. Similar to MIA, MIA2 might also interact with fibronectin, which induces T lymphocyte chemotaxis when combined with stromal cell-derived factor 1α [31]; therefore, MIA or MIA2 might suppress T lymphocyte chemotaxis by masking fibronectin. TANGO expression is observed in many adult tissues [3]; we also confirmed weakly expression of TANGO in cancer-adjacent tissues (data not shown). In addition, we previously found that TANGO promotes tumor angiogenesis and lymphangiogenesis by activating PDGFB and neuropilin 2 [20]. Although TANGO is a ligand for CD11c/CD18 [5], we did not observe a direct interaction between TANGO and this receptor in OSCC cells [20]. More detailed studies will be needed to identify alternate receptors for TANGO in tumor cells; these might include other integrins or adhesion molecules.
In conclusion, we have demonstrated the utility of MIA gene family members as tumor markers, using a wide range of esophageal, lung, and cervical cancer samples. Although innumerable studies have investigated tumor biomarkers, the usefulness of molecular biomarkers for malignant tumors remains controversial. As MIA gene family members are secretory proteins [3], they might be detectable in serum, saliva, urine, ascites, pleural fluid, and other samples that can be collected more easily than tumors. Although additional detailed and large-scale examinations will be fundamental to determining the importance of MIA gene family members in cancers, our findings indicate that these proteins are alternative and efficacious diagnostic and treatment targets in human cancers. Our results therefore provide new knowledge about molecular tumor markers that could potentially improve the clinical outcomes and quality of life of affected patients.

Tissue specimens
Randomly selected formalin-fixed, paraffinembedded (FFPE) specimens were used for the present analysis. All FFPE samples were diagnosed at the Department of Molecular Pathology, Nara Medical University. To screen the expression of MIA family genes, No patient received preoperative therapy. Tumor staging was performed according to the Union for International Cancer Control TNM classification system (seventh edition), and tumor histology was classified according to the World Health Organization criteria. Because written informed consent was not obtained for the immunohistochemical analysis, any identifying information was removed from the samples before the analysis to ensure the strict protection of patient privacy (unlinkable anonymizing). Written informed consent was obtained from individual patients for the use of their samples in the gene expression analysis and ELISA. The study plan was performed according to the ethical standards proposed in the Declaration of Helsinki and was approved by the Medical Ethics Committee of Nara Medical University, Kashihara, Japan (approval number. 719).

Immunohistochemistry
Consecutive 3-μm sections were cut from each block and subjected to immunohistochemical staining with the EnVision+ DualLink system (DAKO, Carpinteria, CA, USA). After a 20-min antigen retrieval treatment with pepsin (DAKO), the sections underwent staining using an immunoperoxidase technique. Briefly, after a 15-min endogenous peroxidase block with 3% H 2 O 2 -methanol, specimens were incubated in a 10% skim milk solution (Morinaga Milk, Tokyo, Japan) for 20 min to block non-specific antibody reactions and rinsed 3 times with phosphate-buffered saline (PBS) (Sigma-Aldrich, St. Louis, MO, USA). Anti-MIA (Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-MIA2 (Abcam, Tokyo, Japan), and anti-TANGO/MIA3 antibodies (LifeSpan, Seattle, WA, USA) were diluted to 1 μg/ml and used as primary antibodies; after a two hour primary antibody incubation, the sections were incubated with a secondary antibody for 30 minutes at room temperature. The specimens were subsequently rinsed three times with PBS and color-developed using a diaminobenzidine (DAB) solution (DAKO). After washing to remove excess DAB solution, the specimens were counterstained with Meyer's hematoxylin (Sakura Finetek Japan, Tokyo, Japan). All samples were immunostained under the same antibody reaction and DAB exposure conditions. Appropriate negative and positive control slides were used.

RNA extraction and qRT-PCR
Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and 1 mg of total RNA was converted to cDNA with a ReverTra Ace qPCR RT Kit (Toyobo, Osaka, Japan). Real-time RT-PCR was performed on a StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) using TaqMan Fast Universal PCR Master Mix (Applied Biosystems), and analyzed using the relative standard curve quantification method. The PCR conditions used were selected according to the manufacturer's manual and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA was amplified as an internal control. TaqMan Gene Expression Assays of MIA (Hs00197954_m1), MIA2 (Hs00365015_m1), MIA3 (TANGO) (Hs00412706_ m1), and GAPDH (ID: Hs03929097_g1) were purchased from Applied Biosystems. All PCRs were performed in triplicate.

ELISA for MIA
The serum samples were obtained before treatment and stored at -80°C. Serum levels of MIA were measured by ELISA system for MIA (Roche Diagnostics, www.impactjournals.com/oncotarget Mannheim, Germany) according to the manufacturer's instructions. All samples were tested in triplicate.

Statistical analysis
Relationships between MIA gene family expression and clinicopathological parameters were calculated using the chi-square test or Fisher's exact test. Disease free survival was analyzed according to the Kaplan-Meier method, and differences between groups were calculated using a log-rank test. JMP8 software (SAS Institute, Cary, NC, USA) was used for all statistical analyses. P values < 0.05 were considered statistically significant.

ACKNOWLEDGMENTS AND FUNDING
This work was supported in part by Grant-in-Aid for Scientific Research from Japan Society for the Promotion of Science, Japan.