Transcriptional changes in insulin- and lipid metabolism-related genes in the hippocampus of olfactory bulbectomized mice

✍ Scribed by Peter Gass; Fernando Leonardi-Essmann; Maha Zueger; Rainer Spanagel; Peter J. Gebicke-Haerter

• Publisher: John Wiley and Sons
• Year: 2008
• Tongue: English
• Weight: 240 KB
• Volume: 86
• Category: Article
• ISSN: 0360-4012
• DOI: 10.1002/jnr.21753

✦ Synopsis

Abstract

Affymetrix chips were used to perform a hypothesis‐free large‐scale screening of transcripts in the hippocampus of olfactory bulbectomized mice, an established animal model of depression. Because only 11 transcripts were significantly changed, the statistically subsequent 25 transcripts below the significance level were additionally included in a first round of qRT‐PCR evaluations. Furthermore, all 36 genes were then tested for mutual interactions or interactions with other molecules in a physiological context using PathwayArchitect software. Thirty of them were displayed in a network interacting with at least one partner molecule from the list or with other partner molecules known from the literature. All partner molecules from the most prominent 10 molecules of this network were then identified and put together into a new list. On those grounds, the hypothesis was made that metabolic network components of the insulin signaling pathway are perturbed in the disease. This pathway was subsequently tested by a second round of qRT‐PCR, adding also a few additional candidate molecules belonging to this pathway. It turned out that the key target—FABP7—fell into the group of transcripts not significantly regulated within the chip data, and another key target—IRS1—did not show up in the chip experiments at all. In conclusion, our data reveal a problem with adhering to statistical significances in microarray experiments, insofar as molecules important for the disease may fall into the range of statistical noise. This approach may also be useful to find new targets for pharmacotherapy in affective disorders. © 2008 Wiley‐Liss, Inc.