Combined laser capture microdissection and serial analysis of gene expression from human tissue samples.

Cell-specific gene expression profiling from heterogeneous human tissues is confounded by cell purification limitations. Here, we describe a technique to generate gene expression profiles of pure populations of prostate cancer cells obtained from fresh-frozen prostatectomy specimens and small initial quantities of RNA by combining laser capture microdissection and microserial analysis of gene expression (LCM-microSAGE). Two microSAGE libraries were obtained from approximately 100,000 laser pulses, estimated to contain fewer than 3 x 10(5) cells and 20-30 ng mRNA. Two libraries were sequenced to a depth of 10,111 and 10,463 unique tags from normal and cancer cells, representing 6453 and 6923 genes, respectively. Most transcripts were expressed at similar levels, but cancer cells compared with normal cells had increased expression of 385 tags and decreased expression of 389 tags. A total of 20 genes were differentially expressed (P<0.05); five of these genes were upregulated and 15 were downregulated in cancer cells. Quantitative reverse transcriptase-polymerase chain reaction results from three selected genes corroborated the existence of cell-specific gene expression in LCM-microSAGE-derived libraries. In conclusion, the LCM-microSAGE approach demonstrates that large-scale expression profiles of known and unknown transcripts can be generated from pure populations of target cells obtained from human tissue samples comprised of heterogeneous mixtures of cell types.