Transcript profiling of small tissue samples using microarray technology

Sammanfattning: Through a number of biological, technological and computational achievements during the 20th century and the devoted work of hundreds of researchers the sequence of the human and other genomes are now available in public databases. The current challenge is to begin to understand the information encoded by the DNA sequence, to elucidate the functions of the proteins and RNA molecules encoded by the genes as well as how they are regulated. For this purpose new technologies within the area of functional genomics are being developed. Among those are powerful tools for gene expression analysis, such as microarrays, providing means to investigate when and where certain genes are used.This thesis describes a method that was developed to enable gene expression analysis, on the transcriptome level, in small tissue samples. It relies on PCR amplification of the 3’-ends of cDNA (denoted 3’-end signature tags). PCR is a powerful technology for amplification of nucleic acids, but has not been used much for transcript profiling since it is generally considered to introduce biases, distorting the original relative transcript levels. The described method addresses this issue by generating uniformly sized representatives of the transcripts/cDNAs prior to amplification. This is achieved through sonication which, unlike restriction enzymes, does not require a specific recognition sequence and fragments each transcript randomly. The method was evaluated using cDNA microarrays, Affymetrix™ oligonucleotide arrays and real-time quantitative PCR. It was shown to perform well, yielding transcript profiles that correlate well to the original, unamplified material, as well as being highly reproducible.The developed method was applied to stem cell biology. The variability in gene expression between different populations of cultured neural stem cells (neurospheres) was investigated. It was shown that neurospheres isolated from different animals or passaged to different degrees show large fluctuations in gene expression, while neurospheres isolated and cultured under identical conditions are more similar and suitable for gene expression analysis. A second study showed that withdrawing epidermal growth factor (EGF) from the culture medium when treating the cells with an agent of interest has profound effects on gene expression, something which should be taken into consideration in future neurosphere studies.