Figure 3.

Relative abundance of different cell types in normal tissue and tumor may lead to misinterpretation of miR-143 and -145 expression data in solid tumors. In traditional bioptic samples, the amount of miR-143 and -145 that can be measured by RT-PCR is the result of their expression levels in the bulk of the tumor, including both cancer and stromal cells. As shown here, the expression of both microRNAs is much higher in stromal than in epithelial cells in normal tissues, and the relative abundance of the two cell types is inverted in tumors. This leads to reduced microRNA levels in the tumor bulk even when their expression is actually increased in tumor cells, and to the misleading conclusion that miR-143 and miR-145 are always down-regulated in tumors.