ABSTRACT: Antipsychotic drugs are classified as typical and atypical based on extrapyramidal effects. However, since the frontal cortex is one of the most important regions for antipsychotic actions, this study attempted to classify antipsychotic drugs based on gene expression in the frontal cortex. Chlorpromazine and thioridazine were selected as typical antipsychotics, and olanzapine and quetiapine as atypical antipsychotics. Since these drugs have similar chemical structures, the effect of the basic structure on gene expression can be eliminated. Cluster analysis of microarray experiments showed thioridazine and olanzapine constituted a robust cluster. K-means clustering separated 4-drug-administered mice into chlorpromazine-quetiapine and thioridazine-olanzapine groups. This classification scheme is different from that which is based on criteria currently used to group the typical and atypical drugs and suggests that antipsychotic drugs can be further separated into multiple groups. Experiment Overall Design: Male 13-week-old ddY mice (Japan SLC Co., Hamamatsu, Japan) weighing 38-43 g were used. Animals were housed with free access to standard food in an air-conditioned room under a constant dark-and-light cycle (light: 7:00 a.m. to 7:00 p.m.) at a temperature of 22 to 24°C and 60 to 70% relative humidity. Ether was used for anesthesia during decapitation. All efforts were made to minimize animal suffering and to reduce the number of animals used. The present experiments were carried out after obtaining permission from the Committee of Animal Experimentation of the Graduate School of Medical and Dental Sciences at Kagoshima University. Experiment Overall Design: Doses of 25 mg/kg chlorpromazine HCl (Sigma-Aldrich Co., St. Louis, MO), 25 mg/kg thioridazine HCl (Sigma-Aldrich Co.), 1.25 mg/kg olanzapine (gift from Eli Lilly and Company, Indianapolis, IN) and 18.75 mg/kg quetiapine fumarate (gift from AstraZeneca, Macclesfield, UK) were used in the study. The dosages for these drugs were determined from therapeutically equivalent doses previously reported (Lehman et al, 2003; Woods, 2003). The drugs were dissolved in acetic anhydride and diluted with 0.9% saline, resulting in a final concentration of acetic acid of 0.5%. The drugs were injected intraperitoneally once daily for 28 consecutive days in a volume of 0.1 ml/10 g body weight. Experiment Overall Design: Microarray experiments were performed using an Agilent G4121A Mouse Oligo Microarray Kit (Agilent Technologies, Palo Alto, CA) as per the manufacturerâ??s instructions. The frontal cortex was immediately removed and the RNA was stabilized in RNAlater RNA Stabilization Reagent (Qiagen, Valencia, CA) and stored at -80°C until use. Total RNA was extracted using the RNeasy Mini Kit (Qiagen). The RNA was amplified and labeled by the Low RNA Input Fluorescent Linear Amplification Kit (Agilent). To synthesize cDNA, 200 ng total RNA was used. Vehicle-injected controls were labeled by cyanine 3 (PerkinElmer Life Sciences, Inc., Boston, MA) and drug-injected mice were labeled by cyanine 5 (PerkinElmer Life Sciences, Inc.). Hybridizations to the microarray were performed using the In situ Hybridization Kit Plus (Agilent). Doses of 750 ng cyanine 3-labeled cRNA, 750 ng cyanine 5-labeled cRNA, and control targets were mixed and fragmented in the kitâ??s fragmentation buffer, and then hybridized to the microarrays for 17 hours at 60°C in a hybridization rotator (Agilent) set to rotate at 4 rpm. Microarrays were washed in 6Ã?SSC with 0.005% Triton X-102 at RT for 10 min, followed by 0.1Ã?SSC with 0.005% Triton X-102 at 4°C for 5 min. The slides were dried, and then scanned by the Agilent G2565BA Microarray Scanner System. Data were analyzed using the Agilent Feature Extraction Software version 7.1. A rank consistency filter and LOWESS were used for dye normalization. Control mice and drug-injected mice were processed in parallel. The data discussed in this publication are presented in accordance with the MIAME guidelines (http://www.mged.org/Workgroups/MIAME/miame.html) and were deposited in NCBIs Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/). They can be accessed using the GEO series accession number GSE1501. Experiment Overall Design: Cluster analysis was performed using free software written by M. Eisen (http://rana.lbl.gov/EisenSoftware.htm). The cclust package in the â??Râ?? statistical software system (www.cran.r-project.org) was used to find the most appropriate number of clusters (i.e. â??kâ?? in the k-means clustering).
