Presence and instability of repetitive elements in sequences the altered expression of which characterizes risk for colonic cancer.

50C10 and 52H10 are two DNA clones previously reported by us to be overexpressed in human colonic mucosa at high risk for development of colonic tumors. This report presents sequencing data that reveal that these clones contain repetitive Alu elements. Each Alu sequence is associated with a 3'-oligoadenylate [oligo(A)] sequence, which is demonstrated to exhibit instability in human colonic tumors. The oligo(A) sequences only decrease in length, unlike microsatellites, which can either increase or decrease. Rigorous quantitative analysis of the length of the oligo(A) sequence in colonic tumors demonstrates that the standard deviation of the length of the sequence in tumors is a function of the mean length; i.e., as the oligo(A) sequence becomes progressively shorter, the variance increases. Both measurements, therefore, provide a quantitative index of the extent of instability in a tissue. Comparison of instability at the oligo(A) loci defined by 50C10 and 52H10 to instability at a CA microsatellite upstream of the apoD gene, and comparison of an oligo(A) and a CA microsatellite both in the 3' untranslated region of the cyclin D1 mRNA demonstrate that instability in a tumor, when present, is more prominent for the oligo(A) sequences than for the microsatellite (P < 0.0001). This suggests either that the mechanisms that generate oligo(A) instability are more penetrant than those that generate microsatellite instability, or that the instability at oligo(A) sequences takes place earlier in the development of the tumor and is selected for, thus becoming more prominent. These features of oligo(A) instability suggest that they may be uniquely useful in detecting and quantifying instability in tissues. Further, the presence of repetitive sequence elements in loci overexpressed in colonic mucosa at risk may be related to an extensive literature that demonstrates that a variety of repetitive sequences accumulate in the cellular RNA population during carcinogenesis and in tumors. Such RNA sequences could play a mechanistic role in tumor development.