I. Introduction

The KAI1 gene was originally identified as a putative metastasis suppressor gene for prostate cancer ^[1] and its protein is a member of the transmembrane-4 superfamily (TM4SF) ^[2,3]. Additional studies have shown that decreased KAI1 expression could be a useful marker for the metastatic/invasive potential in a series of human tumor types, including cancers of the prostate ^[4,5], bladder ^[6], breast ^[7], colon ^[8,9], and pancreas ^[10,11]. In addition, other studies have indicated that the KAI1 mRNA expression level is associated with prognosis in patients with lung ^[12-14], breast ^[15], and pancreatic cancers ^[16]. These observations indicate that down-regulation of KAI1 expression may be an important step in the progression of many types of human malignancies. With regard to SCC, a higher incidence of decreased expression of KAI1 protein was recently reported in esophageal SCCs ^[17,18]. In contrast, a steady level of KAI1 mRNA expression has been found to be similar in both primary and metastatic esophageal SCCs ^[19]. Thus, whether loss of KAI1 function could contribute to the pathogenesis of human SCC including oral SCC is inconclusive.

In oral SCC, our previous study revealed that neither methylation nor mutation contributes to carcinogenesis, whereas loss of KAI1 expression is a very common event in metastatic lymph node tumor ^[20]. However, prevalent primary sites of oral SCC are so many that the mechanisms and pathways of progression and metastasis are very confusing for precise understanding. In the present study, we focused on only one primary site, tongue that is the most prevalent site of SCC in oral cavity, and examined the mutation and the state of KAI1 expression through multi-steps of progression and metastasis of tongue SCC.

II. Materials and Methods

Tissue Samples

Forty-one pairs of tumor and corresponding normal tongue mucosa specimens and 15 metastatic lesions were obtained from 41 unrelated Japanese patients with tongue SCC at the time of surgical resection at the Department of Oral Surgery, Chiba University Hospital, between 2000 and 2003. Informed consent was obtained from all patients and the patients' families. In addition, 25 pre-cancerous tissues, which were pathologically diagnosed as leukoplakia, were obtained in the same manner as mentioned above. The resected tissues were divided into two parts, one of which was frozen immediately after careful removal of the surrounding normal tissues and stored at -80℃ until use; the other was fixed in 10% buffered formaldehyde solution for pathologic diagnosis and for immunohistochemical staining. Histopathologic diagnosis for each cancerous tissue was performed according to the International Histological Classification of Tumours ^[21] by the Department of Pathology, Chiba University Hospital. The clinicopathologic staging was determined by the TNM classification of the International Union against Cancer ^[22]. The tumor samples were checked to ensure that tumor tissue was present (>80% of specimens).

Immunohistochemistry

Paraffin-embedded tissue sections (4μm) from 41 paired primary tongue SCCs, adjacent normal tongue tissues, 15 lymph node metastases and 25 pre-cancerous lesions (leukoplakias) were used for immunohistochemical detection of KAI1. After deparaffinizing and hydrating, the slides were heated in a microwave for 15 minutes in 0.01 M citrate buffer (pH 6), and rinsed three times in phosphate-buffered saline (PBS) solution. After quenching the endogenous peroxidase activity in 0.3% H2O2 for 30 minutes, the sections were blocked for 2 hours at room temperature with 5% bovine serum albumin before reacting with mouse anti-human KAI1 monoclonal antibody, C33 (kindly provided by Dr Osamu Yoshie, Shionogi Institute for Medical Science, Osaka, Japan) at a dilution of 1:50. The sections then were incubated with the primary antibody for 1 hour at room temperature in a moist chamber. Upon incubation with the primary antibody of KAI1, they were washed three times in PBS buffer and treated with ENVISION reagent (DAKO JAPAN Inc., Kyoto, Japan) followed by color development in 3, 3'-diaminobenzidine tetrahydrochloride (DAKO). The slides were then lightly counterstained with hematoxylin, dehydrated in graded ethanol, cleaned in xylene, and mounted. To quantitate the state of KAI1 protein expression, a scoring method was applied ^[9]. A mean percentage of positive tumor cells was determined on at least ten fields at 400X magnification in each section. The intensity of KAI1 immunoreaction was scored as follows: (a) weak, 1+; (b) moderate, 2+; and (c) intense, 3+. The percentage of positive cells and the intensity of the stain were multiplied to produce a KAI1 immunostaining score for each case. Cases with the KAI1 score <30 were defined as negative. Statistical significance was evaluated byχ2 analysis.

PCR-SSCP Analysis for the KAI1 Gene Mutation

To screen the sequence variations of the KAI1 gene, PCR-SSCP analysis was performed as described previously ^[23] for all tongue SCC cases examined in the immunohistochemical analysis. Genomic DNA was isolated as described previously ^[24]. Ten sets of oligonucleotide primers were used to amplify exons from 1 to 10 of the KAI1 gene. The sequence of each primer and each PCR condition was based on those reported previously ^[2]. After the amplification, the PCR products were then electrophoresed under several different conditions; at 4℃, 15℃, and at room temperature.

RT-PCR Assay of the KAI1 Gene in Tongue SCCs

Total RNA was isolated from randomly selected primary tongue SCC specimens (12 KAI1-positive and 4 KAI1-negative cases) and from paired specimens of non-cancerous oral tissue using an SV Total RNA Isolation System (Promega, Madison, WI, USA) according to the manufacturer's protocol. To create first-strand cDNA, 1.5 μg of total RNA was used for RT reaction. The reaction was performed using a Ready-To-Go T-Primed First-Strand Kit (Amersham Pharmacia Biotech, Uppsala, Sweden). To obtain reproducible quantitative performance of the RT-PCR assay, we titrated the amount of starting cDNA and the number of amplification cycles. All subsequent assays were carried out using the parameters that yielded amplification of both KAI1 and GAPDH genes within a linear range. cDNA was amplified by PCR using primers specific for cDNA of the KAI1 gene (5'-AGTCCTCCCTGCTGCTGTGTG-3', sense; 5'-TCAGTCAGGGTGGGCAAGAGG-3', antisense). The GAPDH gene (5'-CATCTCTGCCCCCTCTGCTGA-3', sense; 5'-GGATGACCTTGCCACAGCCT-3', antisense) was also amplified as an interior control resulting in 305-bp product. cDNA preparations were done in the presence and absence of reverse transcriptase, the latter acting as a control for contaminating genomic DNA from which fragments of the pseudogene can be amplified with these primers. PCR reactions were performed in a 9700 Perkin-Elmer Thermal Cycler at 94℃ for 1 min, 29 cycles at 94℃ for 40 s, 60℃ for 90 s, 72℃ for 90 s, followed by an extension step at 72℃ for 5 min. After amplification, an aliquot of the PCR product was separated on a 1.5% TAE-agarose gel and stained with ethidium bromide. The density of the ethidium bromide-stained bands was analyzed using the NIH image software. The results were normalized as a ratio of each specific mRNA signal to the GAPDH gene signal within the same RNA sample. cDNA obtained from normal oral epithelium was used as a positive control. Reproducibility was confirmed by processing all samples at least twice.

III. Results

Immunohistochemistry

A series of 41 tongue SCC patients (29 males and 12 females) were examined. There was no significant relationship between loss of KAI1 gene expression and age and gender of the patients with tongue SCC ^{(Table 1)}. The correlation between the clinicopathologic characteristics of the cases with tongue SCC and KAI1 expression is summarized in Table 2. A strong KAI1 immunoreaction was successfully detected in the cellular membrane of the normal tongue epithelial cells on paraffin-embedded normal tongue tissues ^{(Fig. 1A)}. The KAI1 immunostaining scores for leukoplakias, primary tongue SCCs, and metastatic tongue SCCs are from 0 to 220, from 0 to 280, and from 0 to 75, respectively. Among the primary tumors examined, 35 cases (85%) showed significant down-regulation of KAI1 protein ^{(Table 3; Fig. 1B, E, F)}. All of metastatic tongue SCCs within the lymph nodes showed absent or reduced KAI1 immunostaining ^{(Table 3)}. Moreover, almost all of the primary tongue SCC with regional lymph node metastasis showed down-regulation of KAI1 protein expression and a statistical significance was observed between the KAI1-reduced primary tongue SCCs with and without lymph node metastasis (P=0.044; ^{Table 2}). On the other hand, there was no statistically significant difference between KAI1 expression and other clinicopathologic features ^{(Table 2)}. Twenty-five tongue pre-cancerous lesions histologically diagnosed as leukoplakia were also examined for the state of KAI1 protein expression using the same immunohistochemical method. Eleven cases (44%) of the leukoplakias were defined as KAI1-negative ^{(Table 3)}. There was a significant down-regulation of KAI1 protein expression associated with tumor progression; from normal tissues to metastatic tumors (P<0.01). Representative results are shown in Fig. 1C, and 1D.

Mutation Analyses of the KAI1 Gene
PCR-SSCP analysis to screen for the KAI1 gene mutation was done for all subjects examined in immunohistochemical analysis. No mobility shift in SSCP patterns, however, was detected in either KAI1-negative or KAI1-overexpressed cases (data not shown).

KAI1 Gene Expression in Tongue SCC Tissues Analyzed by RT-PCR Assay
All primary tumors showing down-regulation of KAI1 protein expression (n=35) by immunohistochemistry revealed significantly reduced mRNA expression, whereas a steady state level of the gene expression was observed in KAI1-positive cases (n=6). Typical examples of the RT-PCR analysis are shown in Fig. 2.

IV. Discussion

KAI1, a novel member of the transmembrane-4 superfamily (TM4SF), is located at chromosome 11p11.2 and encodes a transmembrane protein of 267 amino acids ^[1]. KAI1 has been shown to suppress metastasis in a Dunning rat model of prostatic adenocarcinoma ^[1], and the down-regulation of KAI1 is associated with tumor metastasis in multiple human tumors of lung ^[13], breast ^[15], pancreas ^[16], bladder ^[6], colon ^[8,9], and esophagus ^[17,18]. Recent studies have reported that this gene inhibits cellular invasion and mortality in colon cancer cell lines in vitro ^[25] and overexpression of KAI1 in breast cancer cells resulted in suppressed experimental invasion in vitro and metastasis in vivo ^[26].

In the current study, we found that no genetic mutation of the KAI1 gene occurred in either primary or metastatic tongue SCCs, agreeing with other previous reports ^[5,18,27]. Thus, it seems that down-regulation of the KAI1 gene was not commonly caused by gene mutation.

To date, most studies of the KAI1 gene have suggested that this gene may influence the metastatic ability of human cancers. We observed down-regulation of the KAI1 protein in all (100%) of metastatic oral tumors. Therefore, our observations for tongue SCCs agree with the evidence that KAI1 expression is reduced in the majority of metastatic tumors. On the other hand, we observed its down-regulation in almost half (44%) of pre-cancerous lesions investigated ^{(Table 3)}. Therefore, we suggest that the marked down-regulation of KAI1 can be seen as a precipitous event in tongue carcinogenesis. All of the primary tongue SCCs that had metastatic lesions was found to have extensive KAI1 down-regulation. Finally, as shown in Table 3, the frequency of KAI1 expression loss was differentially increased according to the steps of tumor progression (precancerous lesion, primary carcinoma, and metastatic tumor). Our data indicate that loss of KAI1 expression is involved in tumor progression as well as metastasis.

Acknowledgments

We would like to thank Dr. Osamu Yoshie for the generous and much appreciated gift of C33 antibody, which was essential to this research.