The KAI1 gene was first isolated from the human chromosome 11p11.2 in prostate carcinoma cells.1 CD82 was primarily identified by cDNA cloning as the R2 antigen in mitogen-activated human T cells and was subsequently cloned as IA4 and C33 antigens.2-5 The KAI1 gene product is identical to CD82, designated as KAI1/CD82, which is a representative member of the transmembrane-4 superfamily (TM4SF) of cell membrane glycoproteins containing a type III integral membrane structure with four transmembrane domains.6 This novel family has been shown to be widely distributed in some human nonhematopoietic cells, leukocytes, and neoplastic cells.7
It is noteworthy that, although the biophysiologic function of this superfamily remains unclear, it has been strongly suggested that alterations in expression of these cell membrane glycoproteins may be associated with tumor progression in some tumor cells. Indeed, prostate carcinomas with low levels of KAI1/CD82 gene expression have had more aggressive biologic characteristics than those with high levels,1, 8, 9 and similar observations have been noted for pancreatic,10 bladder,11 and breast carcinomas.12
Recently, we have examined KAI1/CD82 expression in nonsmall cell lung carcinoma (NSCLC) by the reverse transcriptase-polymerase chain reaction (RT-PCR) method, using KAI1/CD82 specific oligonucleotides, and we found that patients with a high level of KAI1 gene expression had better prognoses than those with a low level.13 However, in this preliminary study, these findings may reflect the total amounts of KAI1/CD82 within the tumor tissue, not KAI1/CD82 expression status in cancer cells, because stromal cells expressing KAI1/CD82 infiltrated the tissue. Furthermore, the value of KAI1/CD82 expression in predicting disease free survival (DFS), considered to be more important for determining its association with tumor malignancy, remains undetermined. Moreover, KAI1/CD82 expression status in metastatic lesions from NSCLC is unknown in comparison with that in primary lesions.
Therefore, to determine the possible inverse association of KAI1/CD82 expression and tumor progression purely by estimating its expression in lung carcinoma cells, we performed an immunohistochemical study using monoclonal antibody against the KAI1/CD82 gene product in tumor tissues from patients with this disease. In addition, its expression status in matched pairs of primary and metastatic lesions was comparatively analyzed.
- Top of page
- PATIENTS AND METHODS
Using the RT-PCR method, Adachi et al.13 demonstrated that a high level of KAI1/CD82 gene expression in NSCLC, especially lung adenocarcinoma, is a favorable prognostic factor. In the current immunohistochemical analysis of KAI1/CD82 expression, even solely in cancer cells (although there was no significant association between KAI1/CD82 expression status and clinicopathologic factors), patients with KAI1/CD82 positive expression in cancer cells had significantly good prognoses; these observations were particularly significant among those with adenocarcinoma. Thus, it appears that the findings in the current immunohistochemical study were almost in agreement with those shown by gene analysis using the RT-PCR method.13
Samples were grouped as positive, reduced, or negative, according to the percentage of immunohistochemically KAI1/CD82 positive cancer cells. However, there was no clinicopathologic or prognostic difference between samples from the reduced and negative groups. Therefore, based on an immunostaining cutoff point of 50%, it seemed appropriate to combine these two groups. With this cutoff point, the ratio of KAI1/CD82 positive expression to tested patients was close to that in the previous study, which used the RT-PCR method:13 32.5% in this study vs. 23.2% in RT-PCR analysis.
Preliminarily, in order to determine whether there was an association of KAI1/CD82 expression between immunohistochemistry and RT-PCR method, its expression status was comparatively analyzed in some patients who had been included in our previously reported series13 (data not shown). According to the McNemar test, there was no statistically significant difference between the current immunohistochemical data and the prior RT-PCR data, suggesting that there was a good correlation of these methods in the estimation of KAI1/CD82 expression in each NSCLC patient. However, a few patients with adenocarcinoma, regarded as KAI1/CD82 positive by RT-PCR, were included in the KAI1/CD82 reduced group by immunohistochemistry. Considering that the results obtained in RT-PCR analysis were based on total amounts of the KAI1/CD82 transcript level in the tissue, KAI1/CD82-expressing stromal cells and/or noncancerous epithelia components might strongly influence its expression status in the tissue in such patients. In contrast, there were also a few patients, mainly with nonadenocarcinoma type, who belonged to the reduced group according to RT-PCR but to the positive group according to immunohistochemistry. Intratumoral heterogeneity of KAI1/CD82 expression in the samples may have been strongly associated, but the reason for this is not clear. At any rate, by the current study, it was again demonstrated that these exceptions occurred only in a minority of patients.
In the current study, in univariate analysis using the log rank test, patients with positive expression of KAI1/CD82 had favorable prognoses for overall survival (P = 0.0026), and prognostic value was also observed in multivariate analysis (P = 0.035). Especially in patients with adenocarcinoma, these observations were more markedly observed (univariate analysis using the log rank test, P = 0.0010; multivariate analysis, P = 0.041). Moreover, in the current study, it was noted that immunohistochemical KAI1/CD82 expression status in NSCLCs, especially in adenocarcinomas, is also an independent prognostic marker for DFS (NSCLCs: univariate analysis using the log rank test, P = 0.0007; multivariate analysis, P = 0.035; adenocarcinomas: univariate analysis using the log rank test, P = 0.0005; multivariate analysis, P = 0.013). In determining the biologic characteristics of tumor malignancy accurately, DFS analysis is considered to be more important and preferable, because postoperative overall survival periods may be influenced by postrecurrent therapy. Therefore, immunohistochemical KAI1/CD82 expression status in NSCLC, particularly in lung adenocarcinoma, may be clinically useful for detecting a subgroup with a high risk of recurrence among postoperative patients. Our finding that the prognostic value of KAI1/CD82 expression status was second to that of lymph node involvement suggests that it is an indicator of strategy for the postoperative treatment of patients without lymph node involvement. Moreover, considering that immunohistochemistry is routinely and easily performed in comparison with the RT-PCR method, it appears that immunohistochemical KAI1/CD82 expression status in NSCLC is more available in clinical practice.
Besides KAI1/CD82, there are many representative members of the TM4SF, including motility-related protein (MRP-1)/CD9,15-19 CD37,20 CD53,21 ME491/CD63,22 TAPA-1/CD81,23 PETA3/CD151,24 CO-029,25 and Sm23,26 some of which have been shown to be inversely associated with tumor progression in several malignant tumors.17-19, 22 Indeed, we previously reported that reduced MRP-1/CD9 expression in NSCLCs, especially in lung adenocarcinoma,19 as well as in breast carcinoma is a useful unfavorable prognostic factor associated with tumor aggressiveness.18 Considering a possible biologic role of MRP1/CD9 as a regulator of cell motility,15, 16 these findings may be reasonable.
In contrast, the C33 antigen, i.e., KAI1/CD82 protein, was identified by monoclonal antibody, the C33 antibody used in the current study, was originally shown as inhibitory to syncytium formation induced by human T-cell leukemia virus type I,3-5 and this specific inhibition to syncytium formation induced by some human T-cell line by this antibody was strongly associated with altered glycosylation of cell surface antigen,3-5 suggesting that the C33 antigen, i.e., KAI1/CD82 protein, might play a possible role in cell-to-cell adhesion mechanism. Recently, it was also reported that some breast carcinoma cell lines with highly metastatic propensity and invasive ability showed lower levels of KAI1 mRNA.12 These results suggest that this gene product might be a metastasis-suppressor or a less aggressive associated glycoprotein.
In this respect, in addition to the favorable prognostic significance of KAI1/CD82 protein expression, it was also noted that in a few patients, especially those with lung adenocarcinoma, the degree of its expression in metastatic lesions was decreased in comparison with that in the matched primary lesions. Regarding MRP1/CD9 expression, this similar observation was also previously observed in breast carcinoma.18 Therefore, in NSCLCs, especially adenocarcinomas, the possibility exists that decreased KAI1/CD82 expression may be associated with enhanced metastatic potential. Conversely, it is speculated that KAI1/CD82 protein may be, in part, functionally related to the metastasis-inhibiting activity.
However, in the current study, there was no association between KAI1/CD82 expression and prognostic value in patients with histologically nonadenocarcinoma type, mainly squamous cell carcinoma of the lung. Interestingly, we previously noted that the prognostic value of MRP-1/CD9 expression in tumor cells may vary depending on tissue and/or histologic type.19 Therefore, the role of KAI1/CD82 protein in inhibiting metastasis of NSCLCs, if any, may be limited and complicated, and may vary depending on the specificity of tissue and/or the histologic type.1 In fact, it was experimentally shown that the introduction of chromosome 11, containing KAI1/CD82 gene, into highly metastatic rat mammary tumor cells had no effect on metastatic potential.27 Thus, the precise mechanism of the inverse effect of KAI1/CD82 expression in NSCLCs (especially adenocarcinoma cells) on tumor progression, namely, the reason why its expression status in these diseases is a favorable prognostic factor, is not yet known. Further attempts to determine the role of TM4SF, including KAI1/CD82 and MRP-1/CD9, in patients with NSCLC are needed.