LOCALIZATION AND EXPRESSION OF P27(KIP1) IN MULTISTAGE COLORECTAL CARCINOGENESIS

The cyclin-dependent kinase inhibitor p27(Kip1) can inhibit the G(1) t o S transition of the cell cycle and is a putative tumor suppressor. H owever, our laboratory found that a variety of human cancer cell lines express relatively high levels of this protein and that this is often associated with increased expression of cyclin D1 or cyclin E. Theref ore, in the present study we analyzed by immunohistochemistry the expr ession of p27(Kip1) in a series of human tissue samples representing v arious stages of colon carcinogenesis, using 20 samples of normal colo n mucosa, 20 hyperplastic polyps, 19 samples of adenomatous polyps, an d 40 samples of various types of colorectal carcinomas. Parallel immun ostaining was done for cyclin D1 and also for Ki67 to evaluate cell pr oliferation, An additional 17 human colon carcinoma samples, together with paired adjacent normal mucosa samples, were analyzed for levels o f expression of the p27(Kip1) protein by Western blot analysis, and 7 of these pairs of samples were examined by Northern blot analysis for levels of p27(Kip1) mRNA. We did not find a positive or negative corre lation between p27(Kip1) expression and cell proliferation in the norm al mucosa and tumor samples. There was, however, an inverse correlatio n between p27(Kip1) and Ki67 expression in the Lymphoid follicles pres ent in the colonic mucosa. There was no evidence for a consistent incr ease or decrease in p27(Kip1) expression in the mucosal cells during c olon carcinogenesis, because the mean values for percentage p27(Kip1)- positive cells were similar in the normal mucosa, adenomatous polyps, and carcinoma samples. This is in contrast to Ki67 and cyclin D1 expre ssion, which did show significant increases in mean values with tumor development. A subset (35%) of the carcinomas displayed diffuse cytopl asmic staining, in addition to nuclear staining, for p27(Kip1), and in these eases the percentage of cells that were positive for p27(Kip1) was higher than in cases that had only nuclear staining. There was a s ignificant correlation between p27(Kip1) expression and tumor grade; i .e., well and moderately differentiated carcinomas had high p27(Kip1) expression, whereas poorly differentiated carcinomas had lower express ion. The Western blot analysis data on p27(Kip1) expression confirmed this correlation. Comparisons of Northern and Western blots did not sh ow a correlation between the level of p27(Kip1) mRNA and the correspon ding protein, a finding consistent with evidence that the p27(Kip1) pr otein is regulated mainly via a posttranscriptional mechanism. The imm unostaining studies revealed a significant correlation between high p2 7(Kip1) protein expression and high cyclin D1 expression in the adenom atous polyps and in the subset of carcinomas that had only nuclear p27 (Kip1) expression. This may reflect the existence of a homeostatic fee dback mechanism that is lost in the high-grade carcinomas that express low levels of p27(Kip1).