Purpose: We previously identified mina53, a novel Myc target gene. Here we investigated whether mina53 is related to esophageal squamous cell carcinoma (ESCC), a disease with poor prognosis.
Experimental Design: Mina53 expression was suppressed in ESCC cell lines by a RNA interference method to investigate whether Mina53 is involved in cell proliferation. Expression of Mina53 was investigated by Western blotting in tissue sections from patients with ESCC. Immunohistochemical analysis of Mina53 was carried out and compared with that using anti–Ki-67 antibody. Finally, the level of Mina53 expression was compared with the length of survival of patients with ESCC.
Results: Reduction of mina53 expression by RNA interference suppressed cell proliferation in ESCC cell lines. Western blot analysis of surgically resected ESCC specimens indicated that the expression of Mina53 in tumors was increased compared with that in adjacent nonneoplastic tissues in all four specimens examined. When formalin-fixed specimens from 52 patients with ESCC were stained immunohistochemically, it was found that Mina53 was highly expressed in 83% of specimens. Anti-Mina53 antibody stained tumors more efficiently than antibody against Ki-67, a cell proliferation biomarker, in some cancer specimens. Patients with high expression of Mina53 had shorter survival periods, whereas the expression level of Ki-67 in ESCC showed no relationship to patient outcome.
Conclusions: Taken together, our results indicate that expression of Mina53 is a characteristic feature of ESCC and suggest that immunostaining by anti-Mina53 antibody may be useful as a potential prognostic indicator.
Esophageal squamous cell carcinoma (ESCC) is a highly aggressive disease with a poor prognosis (1) . Genetic changes associated with the development of ESCC involve activation of proto-oncogenes including cyclin D1, c-erbB2, and c-myc and inactivation of several tumor suppressor genes including p53, Rb, and p16 (2) . However, the precise mechanisms in the development of ESCC still remain unclear. Several studies suggested that the survival rate correlates with biological factors such as the expression of p53 (3) , cyclin D1 (4) , and Ki-67 (5 , 6) or with clinicopathologic factors such as tumor size and lymph node status (6 , 7) . However, whether these factors can be used as prognostic indicators is still controversial, although they appear to be involved in the carcinogenesis of ESCC (6 , 8 , 9) . With the development of cDNA microarray technology, it has recently become possible to screen for alterations in the expression of many genes simultaneously (10) . This new technique has been used to investigate mechanisms in the development of ESCC, to search for prognostic factors, and to overcome therapeutic difficulties (11, 12, 13) . However, no single gene or set of genes that is commonly up- or down-regulated in all malignant tumors has been generally recognized yet (14) . Because tumor tissues contain various types of cells, such as mesenchymal cells and blood cells, the gene expression profiles observed inevitably reflect the profiles of these cells in addition to tumor cells. Thus, expression analysis of only resected tumor cell tissues or immunohistochemical studies for each candidate gene may be necessary to depict more precisely the gene expression of in vivo tumor cells.
The myc family of proto-oncogenes consists of three main genes: c-myc, N-myc, and L-myc (15, 16, 17, 18, 19) . Deregulated expression of myc family genes has long been known to be associated with neoplastic diseases in a wide range of vertebrates, including humans (18, 19, 20) . The genes of the myc family also relate to many biological phenomena besides tumorigenesis, and they control apoptosis and cell differentiation in addition to cell proliferation (21 , 22) . The proteins encoded by the myc family genes are members of the basic helix-loop-helix leucine zipper transcription factors (19 , 20 , 23 , 24) and appear to control the expression of several genes that mediate each of the myc functions. However, there appear to be Myc target genes that remain to be identified and characterized, and to date, the mechanisms by which myc contributes to tumorigenesis are still not fully resolved.
Recently, we identified a novel gene, mina53, whose expression was demonstrated through experimental evidence to be directly induced by c-myc (25) . C-myc is one of the most widely studied proto-oncogenes. In general, its expression is associated with cell proliferation and down-regulated in quiescent and differentiated cells. The mina53 gene encodes a protein with a molecular weight of 53,000 that is localized in the nucleus, with part of the protein concentrated in the nucleolus. Expression of mina53 is increased during cell proliferation, and specific inhibition of mina53 expression by an RNA interference method suppressed cell proliferation in some cultured cell lines (25 , 26) . Recently, we generated a specific monoclonal antibody against human Mina53 protein and found that expression of Mina53 is frequently increased in human colon cancer (26) . Here we show that Mina53 is highly expressed in most ESCC cells and that its expression level in tumors relates to the outcome of patients with ESCC.
MATERIALS AND METHODS
We established hybridoma clones secreting anti-human Mina53 antibodies (26) . One hybridoma secreting an IgG2a antibody, clone M532, was used in this study. The specificity of M532 to human Mina53 was demonstrated in our previous study (26) . The antibody was produced as ascitic fluid in prestane-preinjected mice and purified. Rabbit anti–c-Myc polyclonal antibody (sc-764; Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-nucleolin polyclonal antibody (sc-13057; Santa Cruz Biotechnology), Alexa 488-conjugated goat antimouse IgG (Molecular Probes, Eugene, OR), goat anti-mouse IgG-horseradish peroxidase and Cy3-conjugated anti-rabbit IgG (Zymed Laboratories, South San Francisco, CA), anti–β-actin monoclonal antibody (AC-15; Sigma-Aldrich Japan, Tokyo, Japan), mouse anti–Ki-67 antibody (clone MIB-1; DAKO A/S, Glostrup, Denmark), biotinylated rabbit anti-mouse IgG (Nichirei, Tokyo, Japan), biotinylated goat anti-rabbit IgG (Nichirei), peroxidase-conjugated streptavidin (Nichirei), peroxidase-labeled goat anti-mouse IgG Fab′ (Nichirei), and peroxidase-labeled goat anti-rabbit IgG Fab′ (Nichirei) were purchased.
Immunoblotting and Indirect Immunofluorescence Staining.
The moderately differentiated human ESCC cell line TE-11 and the poorly differentiated human ESCC cell line TE-9 were graciously provided by Dr. Tetsuro Nishihira (Seta Clinic Group, Tokyo, Japan) and the Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer, Tohoku University (Senda, Japan; ref. 27 ). TE-11 and TE-9 cells were maintained in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum). The human promyelocytic leukemia cell line HL60 was cultured in RPMI 1640 supplemented with 10% fetal calf serum. For Western blotting of cell lysates, cells were collected by treatment with trypsin and EDTA in PBS [PBS = 137 mmol/L NaCl, 2.7 mmol/L KCl, 10 mmol/L Na2HPO4, 2 mmol/L KH2PO4 (pH 7.4)], washed with PBS, extracted with a solubilization buffer [3% SDS, 0.125 mmol/L Tris-HCl (pH 6.8), 50 mmol/L dithiothreitol], boiled for 10 minutes, and centrifuged at 14,000 rpm for 20 minutes. The total protein concentration of the supernatant was determined using Bio-Rad protein assay reagent (Bio-Rad Laboratories, Hercules, CA). Cell extracts with equal amounts of protein were subjected to Western blotting as described previously (25) . Membranes in which Mina53 was detected by anti-Mina53 antibody were reprobed with a monoclonal anti–β-actin antibody after treatment with a stripping buffer (Pierce, Rockford, IL). For Western blotting of tissue lysates, four surgically resected tissues were cut into small pieces in ice-cold PBS, added to an equal volume of the solubilization buffer, boiled for 10 minutes, and centrifuged at 14,000 rpm for 20 minutes. Tissue extracts with equal amounts of protein were subjected to Western blotting, except that peroxidase-labeled goat anti-mouse IgG Fab′ or peroxidase-labeled goat anti-rabbit IgG Fab′ in PBS containing 0.5% skim milk and 0.5% nonimmune goat serum was used for the second antibody.
For indirect immunofluorescence staining, TE-11 cells grown on glass coverslips were fixed in methanol for 10 minutes at −20°C and processed as described previously (25) . In brief, after incubation in 1% skim milk to block nonspecific binding sites, cells were reacted with mouse anti-Mina53 monoclonal antibody (M532) and rabbit anti-nucleolin antibody and then further incubated with Alexa 488-conjugated antimouse IgG and Cy3-conjugated anti-rabbit IgG. Cells were observed by fluorescence microscopy.
Introduction of Small Interfering RNA into Cells.
Small interfering RNA (siRNA) was introduced into cells basically as described previously (25 , 28) . Small interfering RNA with 21 nucleotides targeting human mina53 and a control siRNA sequence (a duplex with the inverted sequence of human mina53 siRNA) were chemically synthesized (Hokkaido System Science, Sapporo, Japan; ref. 25 ). The siRNA targeting human c-myc and the control RNA molecule were purchased from Ambion (Austin, TX). Twenty-four hours before transfection, cells in an exponentially growing phase were trypsinized and transferred to a 12-well plate. Transfection was carried out with 200 pmol of siRNA per well using Oligofectamine (Invitrogen Japan, Tokyo, Japan) according to the manufacturer’s instructions, except that TE-11 cells and TE-9 cells were cultured without serum for 24 hours for transfection.
Tissues for Immunostaining.
Routinely processed formalin-fixed and paraffin-embedded specimens from 52 consecutive patients with ESCC operated on from 1996 to 1998 at the Department of Surgery of Kurume University Hospital were used. All of the patients decided for themselves whether to receive chemotherapy after being informed of their pathological stagings. Postoperative chemotherapy was elected by 25 of the patients studied. The chemotherapy regimen consisted of two courses of 70 mg/m2 cisplatin as an intravenous drip infusion for 2 hours on day 1, and 700 mg/m2 5-fluorouracil as a continuous intravenous infusion for 24 hours on days 1 to 5. The specimens were assigned a tumor-node-metastasis (TNM) stage based on the International Union Against Cancer (UICC) criteria for tumor classification (29) . Hematoxylin and eosin (H&E)-stained sections were also classified according to the histopathological grading system as well (G1), moderately (G2), or poorly (G3) differentiated. The histopathological grade and stage of cancer were determined by two pathologists. The characteristics of the tissues are outlined in Table 1⇓ . One tissue block containing a marginal region of tumor was selected from each patient. Immunostaining of tissues was performed essentially as described previously (26) . In brief, thin sections of formalin-fixed paraffin-embedded tissue specimens were mounted on glass slides, deparaffinized, and autoclaved for 20 minutes in 10 mmol/L sodium citrate buffer (pH 6.0) for antigen retrieval. After pretreatment with 3% H2O2 in PBS and then with 1% skim milk and 5% rabbit serum or 1% skim milk and 5% goat serum in PBS, sections were incubated with the primary antibodies overnight at 4°C and then incubated with biotinylated rabbit anti-mouse IgG or biotinylated goat anti-rabbit IgG, followed by incubation with peroxidase-conjugated streptavidin. Color was developed with 3,3-diaminobenzidine and H2O2. After light counterstaining with hematoxylin, the slides were dehydrated, coverslipped, and observed with an Olympus AX80 microscope (Olympus Optical, Tokyo, Japan). For evaluation of the level of staining by antibodies, the image of the most highly stained area in each slide was captured at ×400 with a digital camera and printed. Each section was scored on a scale from 0 to 4 by visual observation. The highest staining intensity was scored as 4, the lowest was scored as 0.5, and no staining at all was scored as 0. To estimate the percentage of stained cells, the numbers of positive and negative tumor cells in the field were counted, and the ratio of positive cells to the total number of cells was expressed as a percentage. Between 47 and 582 cells were counted. The difference was due to the population of tumor cells in the field and the size of each tumor cell. A staining index was calculated as staining intensity × percentage of cells stained. Without prior knowledge of the patients’ other data, sections were scored by two independent observers (M. T. and K. T.), with very high correlation between scores (P < 0.0001).
The correlations between the Mina53 staining index and other factors were examined using the Spearman rank correlation coefficient and statistical test of independence. The crude survival rates were calculated using the Kaplan-Meier method (30) . The period from the date of surgical excision of the primary tumor to the date of death or last follow-up was used as the survival interval. For disease-specific survival, data for patients who died from causes other than ESCC were censored at the time of death. Patients with staining index values higher than the mean value of staining indexes from all 52 patients were classified into the high staining index group. The mean staining index for Mina53 was 1.19. The number of patients with high and low staining index for Mina53 was 28 and 24, respectively. The mean staining index for Ki-67 was 0.949. The number of patients with high and low staining index for Ki-67 were 21 and 31, respectively. Whether or not the Mina53 staining index, histopathological grade, stage of cancer, and chemotherapy after the operation had an effect on survival rates was examined using the Cox’s proportional hazards linear model (31) . All statistical analyses were performed using the statistical analysis program StatView version 5 (SAS Institute, Cary, NC).
Expression of Mina53 in Esophageal Squamous Cell Carcinoma Cells.
The expression of Mina53 was examined in a human ESCC cell line, TE-11. A cell lysate was prepared from proliferating cells and analyzed by immunoblotting with anti-Mina53 mouse monoclonal antibody. The antibody recognized a single band of Mr 53,000 (Fig. 1A⇓ , Lane 3). The result indicates that TE-11 cells express Mina53 and that anti-Mina53 antibody specifically recognizes Mina53 protein in ESCC cells with no cross-reactivity to other proteins. HL60 cells are terminally differentiated by phorbol 2-myristate 13-acetate (PMA), which reduces the c-myc expression level (32 , 33) , and mina53 expression is subsequently reduced (ref. 25 ; Fig. 1A⇓ , Lanes 1 and 2 in the top panel). The specificity of the reduction was confirmed because PMA treatment did not reduce β-actin expression (Fig. 1A⇓ , Lanes 1 and 2 in the bottom panel). The mobility of the band was the same as that in TE-11 cells, confirming that the antibody specifically recognized Mina53 in TE-11 cells. The expression level of Mina53 in TE-11 cells was similar to that of the proliferating human promyelocytic leukemia cell line HL60 (Fig. 1A⇓ , Lane 1), in which c-myc and mina53 are well expressed (25) .
To investigate the localization of Mina53 in ESCC cells, TE-11 cells were stained by a double immunofluorescence technique using anti-Mina53 antibody and anti-nucleolin antibody. As shown in Fig. 2A⇓ , anti-Mina53 antibody specifically stained nuclei of TE-11 cells with a strong, dotted pattern. The signals overlapped with those for nucleolin. These results indicate that Mina53 is located in the nucleus with concentrated amounts in the nucleolus in ESCC cells, as we have shown previously in HeLa cells and colon cancer cells (25 , 26) .
Involvement of Mina53 in Proliferation, and Regulation of Mina53 Expression by c-Myc in Cultured Esophageal Squamous Cell Carcinoma Cells.
Recently it was reported that a 21-nucleotide siRNA duplex specifically suppresses gene expression in mammalian cell lines (28) . We used this technique to investigate the function of Mina53 and the regulation of mina53 expression by c-myc in human ESCC TE-11 cells. When TE-11 cells were transfected with a siRNA molecule specific for mina53, the expression of mina53 was reduced (Fig. 1B⇓ , compare Lane 3 with Lanes 2 and 4). The expression of β-actin was not affected by the siRNA molecules, indicating that the effect of the siRNA was specific. The reduction of Mina53 expression resulted in suppression of cell proliferation (Fig. 1C)⇓ . To confirm the suppression of cell proliferation by the siRNA for mina53, another ESCC cell line, TE-9, was tested, and it was found that the reduction of Mina53 expression (Fig. 1B⇓ , Lanes 5 and 6) resulted in suppression of cell proliferation in TE-9 cells (Fig. 1C)⇓ . These results suggest that mina53 is involved in proliferation of ESCC cells.
When TE-11 cells were transfected with a specific siRNA molecule for c-myc, the expression of c-myc was reduced, and substantial reduction of Mina53 expression was also observed (Fig. 1B⇓ , compare Lane 1 with Lanes 2 and 4). The expression of β-actin was not affected by the siRNA for c-myc, indicating that the effect of the siRNA was specific. These results suggest that c-myc positively regulates the expression of mina53 in ESCC cells.
Western Blotting Analysis of Mina53 Expression in Esophageal Squamous Cell Carcinoma Tissues.
Four sets of tumors and adjacent nonneoplastic tissues were resected from surgical ESCC specimens and analyzed by Western blotting. In all four cases (Fig. 1D⇓ , top panel), the expression of Mina53 was high in tumor tissues compared with that in their nonneoplastic counterparts. In all cases, the level of β-actin was similar (Fig. 1D⇓ , middle panel). These results suggest that Mina53 expression is elevated in at least some ESCC tissues.
Next, the expression of c-Myc was examined. In the first case, the amount of c-Myc did not differ much between nonneoplastic and tumor areas; in the second case, it was lower in the tumor area; and in the third and fourth cases, it was higher in the tumor areas. Whereas it is not clear at present why all tumor tissues expressing elevated levels of Mina53 did not contain the elevated level of c-Myc, these results may suggest that the expression of Mina53 is not determined only by c-Myc in ESCC tissues.
Immunohistochemical Analysis of Mina53 Expression in Esophageal Squamous Cell Carcinoma Tissues.
Anti-Mina53 antibody was used to detect Mina53 protein immunohistochemically in formalin-fixed ESCC tissues. H&E staining was used to demarcate tumor areas. The section shown in Fig. 2B⇓ contains a well-differentiated squamous cell carcinoma in which keratinization was often observed. Fig. 2D⇓ shows marked staining for Mina53 in tumor areas. Staining was weak in highly keratinized cells, and most nonneoplastic stromal cells around tumors showed little staining. Staining for Mina53 was found in nuclei (Fig. 2E)⇓ , in which dotted staining was observed, showing a pattern similar to that obtained in the ESCC cell line (Fig. 1B)⇓ . Specific nuclear staining was not observed when the first antibody was omitted (Fig. 2C)⇓ or when the antibody was displaced by an excess amount of recombinant Mina53 protein (data not shown).
In some sections, tumor cells were found in blood vessels, and/or deeply invading the muscular layer of the esophageal wall (Fig. 3, A and C⇓ , respectively). Anti-Mina53 antibody stained tumor cells in these areas (Fig. 3, B and D⇓ for blood vessels and muscular layer, respectively), whereas nonneoplastic cells surrounding tumors were not stained. Tumor cells in lymphoid vessels were also markedly stained by anti-Mina53 antibody (data not shown). These results suggest that the expression of Mina53 is frequent in tumors that remain localized as well as in those with invasive and metastatic potential.
Comparison of Mina53 and Ki-67 Expressions in Esophageal Squamous Cell Carcinoma Tissues.
Ki-67 is expressed in proliferating cells (34 , 35) , and anti–Ki-67 antibody is widely used to detect neoplastic cells, including ESCC (5 , 36) . Expressions of Mina53 and Ki-67 were examined in serial sections of specimens from 52 patients with ESCC. A summary of the immunohistochemical staining is presented in Table 1⇓ . Fig. 2F⇓ shows that anti–Ki-67 antibody strongly stained ESCC, which was also stained by anti-Mina53 antibody. In some specimens, however, anti–Ki-67 antibody rarely stained tumor cells that were strongly stained by anti-Mina53 antibody (Fig. 3E⇓ for Mina53 and Fig. 3F⇓ for Ki-67).
Staining indexes of Mina53 and Ki-67 for areas of nonneoplastic epithelium from 38 patients were determined. The rest of the specimens did not have a typical nonneoplastic epithelium. The mean indexes for Mina53 and Ki-67 in nonneoplastic epithelium were 0.38 and 0.37, respectively. Eighty-three percent (43 of 52) of the tissue samples had higher Mina53 staining values than the mean index of Mina53 for nonneoplastic epithelium (Table 1)⇓ . Next, the staining indexes of Mina53 and Ki-67 in tumor areas were divided by their respective mean index of nonneoplastic areas and plotted (Fig. 5A)⇓ . The results suggest that the expression of Mina53 was roughly proportional to that of Ki-67; however, there were several tissue samples in which expression of Mina53 was high, but that of Ki-67 was very low.
Expression of Mina53 in a Nonneoplastic Esophageal Tissue.
The esophagus is lined by a stratified squamous epithelium that is composed of basal, epibasal, suprabasal, intermediate, and superficial squamous cell layers. The basal and epibasal layers contain proliferating cells. In the suprabasal layer, cells cease to proliferate and start to differentiate, and in the intermediate and superficial squamous layers, cells differentiate further (37 , 38) . The expression patterns of Mina53, c-Myc, and Ki-67 were compared in nonneoplastic epithelium. As shown in Fig. 4A⇓ , Mina53 staining was present in cells from the basal to the suprabasal layers of the epithelium (Fig. 4C)⇓ . Anti–c-Myc antibody also stained cells from the basal layer to the suprabasal layers of the epithelium (Fig. 4B)⇓ . As shown in Fig. 4D⇓ , Ki-67 staining was observed in a few cells in the basal and epibasal regions but was hardly observed in other layers of the epithelium, as reported previously (38 , 39) . These results suggest that both Mina53 and Ki-67 are expressed in proliferating cells in epithelium but that Mina53 is expressed in more types of cells than Ki-67.
Lymphoid germinal centers, which contain nonneoplastic cells with high mitotic activity, were intensely stained by anti–Ki-67 antibody (Fig. 4F)⇓ but only weakly stained by anti-Mina53 (Fig. 4E)⇓ and anti–c-Myc (data not shown) antibodies. These results are consistent with previous observations that Mina53 and c-Myc are rarely expressed in germinal centers of lymphoid follicles in colon tissues (26) and that c-Myc is rarely expressed in germinal centers of the tonsil (40) .
The Level of Mina53 Expression in Relation to Survival Period.
Because the expression level of Mina53 differed between specimens (Table 1⇓ and Fig. 5A⇓ ), we investigated whether the level of Mina53 expression is associated with any biological event in ESCC. The Mina53 staining index was compared with the histopathological grade, stage of cancer, and Ki-67 staining index to calculate correlation coefficients and to conduct independent tests. We did not observe any correlation between the staining index of Mina53 and histopathological grade (the correlation coefficient r was −0.008; P = 0.953) or stage of cancer (r = 0.147; P = 0.299) in this study. The Mina53 staining index was correlated with the Ki-67 staining index (r = 0.374; P = 0.006), suggesting that Mina53 is related to cell proliferation. However, as seen in Fig. 5A⇓ , there are several points that have a high Mina53 staining index but a very low staining index for Ki-67. In nonneoplastic epithelium, more types of cells were stained by anti-Mina53 antibody than by anti-Ki-67 antibody (Fig. 4)⇓ . These results may reflect some biological characteristics specifically associated with Mina53.
Next, the patients were divided into two classes with a low or high Mina53 staining index. The crude survival curves using the Kaplan-Meier method were estimated for each category. Fig. 5B⇓ shows that patients with a low Mina53 staining index had longer survival periods than those with a high Mina53 staining index (P = 0.0085). When the patients were divided into two classes with a low or high Ki-67 staining index, no difference between the two classes was demonstrated (Fig. 5C⇓ , P > 0.5).
Possible prognostic factors, including stage of cancer, histopathological grade, chemotherapy after operation, and the staining index for Mina53, were analyzed by Cox’s proportional hazards method. As shown in Table 2⇓ , among these four factors, only the staining index of Mina53 (P = 0.0122) and stage of cancer (P = 0.0280) were found to reflect the survival rate.
Western blotting analyses and immunohistological studies showed that Mina53 is highly expressed in most ESCC tissues (Figs. 1⇓ and 2⇓ ; Table 1⇓ ). The expression level of Mina53 was not related to histopathological grade. Deeply invading tumors, as well as tumors that had penetrated blood and lymphatic vessels, also showed strong immunoreactivity to Mina53 (Fig. 3)⇓ . Anti-Mina53 antibody stained tumor cells in some ESCC tissues more efficiently than anti–Ki-67 antibody (Fig. 5A)⇓ , which is widely used as a cell proliferation marker. On the other hand, germinal centers of lymphoid follicles, which contain nonneoplastic cells with high mitotic activity, were stained strongly by anti–Ki-67 antibody but were rarely stained by anti-Mina53 antibody (Fig. 4)⇓ . We performed immunohistochemical studies in lymphomas using anti-Mina53 antibody and found that lymphoma was rarely stained by anti-Mina53 antibody.4 Therefore, the elevated expression of Mina53 is not associated with tumor cells in general but rather with certain types of tumor cells, including ESCC and colon cancer. Our results indicate that anti-Mina53 antibody efficiently and preferentially stains ESCC and that Mina53 can be used as a histochemical marker for ESCC.
The levels of staining by anti-Mina53 antibody varied between specimens (Fig. 5A)⇓ . A relationship between the staining index of Mina53 and the prognostic outcome of patients was demonstrated by a Kaplan-Meier plot: patients with high expression of Mina53 had a poor outcome (Fig. 5B)⇓ . Analysis by the Cox’s proportional hazards method showed that the staining index of Mina53 correlated well with patient outcome (Table 2)⇓ . The stage of cancer was also demonstrated to correlate well with patient outcome by Cox’s proportional hazards method, although the staining index of Mina53 was not associated with the stage of cancer. Patients with ESCC take various clinical courses: even though the stage of cancer is the same at the time of surgery, some patients succumb to the disease soon after surgical treatment, whereas others experience long-term survival (14) . If the staining index of Mina53 can be used as a prognostic indicator for patients with ESCC independent of the stage of cancer, the stage of cancer plus the Mina53 staining index may provide a more accurate prognostic tool for patients with ESCC.
It was observed that anti-Mina53 antibody detected tumor cells more efficiently than anti–Ki-67 antibody in several ESCC tissues (Fig. 5A)⇓ . We have also observed that the percentage of tumor cells stained by anti-Mina53 antibody was higher than that stained by anti–Ki-67 in colon cancer tissues (26) . When the expression patterns of Mina53 and Ki-67 were compared in nonneoplastic epithelium, expression of Mina53 was observed in the basal, epibasal, and suprabasal layers, but expression of Ki-67 was restricted to a smaller population of cells in the basal and epibasal layers (Fig. 4)⇓ . Ki-67 is preferentially expressed in proliferating cells in late G1, S, G2, and M phases (34 , 41, 42, 43) but disappears in cells with a prolonged G1 phase; Mina53 appears to remain expressed even when cells are in a prolonged G1 phase. The expression of Mina53 in these cells allows the efficient detection of tumor cells by anti-Mina53 antibody.
When expression of c-myc was specifically suppressed by the siRNA molecule for c-myc in cultured ESCC cells, expression of Mina53 was reduced (Fig. 1)⇓ . The expression pattern of c-Myc was similar to that of Mina53 in nonneoplastic esophageal epithelium (Fig. 4)⇓ . These results are consistent with our previous finding that expression of mina53 is directly induced by c-myc (25) . However, Western blotting analysis of ESCC tissues showed that the expression level of Mina53 was not proportional to that of c-Myc in some cases (Fig. 2)⇓ . Therefore, whereas c-Myc appears to be involved in control of Mina53 expression, there are likely to be other factors that control Mina53 expression in ESCC. These factors could be proteins in the Myc/Mad/Max network, which can control expression of some Myc target genes through the same DNA elements as those for Myc (19 , 44) . There is a previous report that the cdk4 gene in proliferating cells appears to be directly regulated by both c-Myc and Mnt, the latter is a protein in the Myc/Mad/Max network that has an antagonistic activity to Myc (45) . It is also possible that some other factor or factors control Mina53 expression through other elements in the mina53 promoter or posttranscriptional mechanisms.
The staining index of Mina53 was inversely correlated with patient survival (Fig. 5⇓ ; Table 2⇓ ), suggesting that Mina53 is involved in tumor progression. In ESCC tissues, Mina53 expression correlated with Ki-67 expression (P = 0.006). Specific inhibition of Mina53 expression suppressed cell proliferation of cultured ESCC cells (Fig. 1)⇓ . In nonneoplastic esophageal epithelium and ESCC, some cells express Mina53 but not Ki-67. These results suggest that Mina53 is involved in proliferation of ESCC cells, although expression of Mina53 itself is not sufficient for cell proliferation. Cell proliferation itself can be a causal factor in malignant transformation because an increased number of cell divisions increases chances that mutations will be introduced into the genomic DNA (46) . Thus, Mina53 may be involved in tumor progression by stimulating proliferation.
Mina53 was preferentially expressed in ESCC with a poor prognosis (Fig. 5⇓ ; Table 2⇓ ) and involved in proliferation of ESCC cells (Fig. 1)⇓ . Therefore, although Mina53 is expressed not only in tumor cells but also in nonneoplastic tissues, it is worthwhile to consider whether Mina53 could be a target for ESCC therapy. Whereas deregulated expression of myc family genes has long been known to be associated with neoplastic diseases (18, 19, 20) , myc is also involved in signaling pathways that are important for nontumorigenic biological events (47, 48, 49) . Thus, long-term use of medicines that are designed to inactivate myc would be expected to have serious toxicity (50) . There is, however, a report suggesting that the toxicity might be avoided by making the period of inhibition of Myc activity short (50) . Thus, inactivation of mina53, which is a Myc target gene involved in cell proliferation, might not produce serious toxicity if the inhibition of Mina53 activity is restricted to a short period. Additional studies are necessary to address the specific role of Mina53 in cell proliferation, which may lead to development of a new cancer therapy.
We thank Dr. Tetsuro Nishihira and the Cell Resource Center of the Biomedical Research, Institute of Development, Aging and Cancer, Tohoku University, for providing the human squamous carcinoma cell lines TE-11 and TE-9. We thank Shigeo Kamimura, Kurume University, Kurume, Japan, for preparing serial sections of tissues and Yasuko Noguchi, Kurume University, Kurume, Japan, for technical assistance.
Grant support: Grants-in-Aid from the Ministry of Education, Science, Sports and Culture of Japan and the Shimabara Science Foundation (Japan).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Makoto Tsuneoka, Division of Human Genetics, Department of Forensic Medicine, Kurume University School of Medicine, Kurume 830-0011, Japan. Phone: 81-942-31-7554; Fax: 81-942-31-7700; E-mail:
↵4 K. Teye, M. Tsuneoka, N. Arima, H. Kimura and Y. Koda, unpublished results.
- Received November 6, 2003.
- Revision received June 10, 2004.
- Accepted August 13, 2004.