Total RNA was isolated from dexamethasone-treated and control TM cells with extraction reagent (TRIzol; Life Technologies, Inc., Rockville, MD), according to the manufacturer’s recommendations. A cDNA probe made from each 2 μg of total RNA was labeled by reverse transcription with either fluorescein or biotin. At probe creation, RNA of unlabeled plant genes was added for normalization. The cDNA probes were then mixed and cohybridized to the microarray for 16 hours. After stringent washes, the hybridized cDNA signal was amplified by the addition of streptavidin horseradish peroxidase (HRP) followed by cyanine 5-tyramide. Anti-fluorescein-HRP and cyanine 3-tyramide was then added sequentially. With a laser detection system (ChipReader; Virtek Vision Inc., Ontario, Canada), the fluorescent signals were scanned to determine differential gene expression between dexamethasone-treated TM cells and control cells. Local background hybridization signals were subtracted before spot intensity comparison and expression profile determination. To remove values from poorly hybridized spots, data were filtered with a cutoff level so that signal intensities at least twice as high as the backgrounds were included. There were 12 control plant gene spots on the microarray. Because an equal aliquot of control plant gene RNA was added to both biotin and fluorescein cDNA reactions, the cyanine-3 and cyanine-5 signal intensity for these genes on the microarray should, in theory, be the same. By making comparisons with the signal intensity of these plant genes, we normalized signal intensity between cyanine-3 and cyanine-5 on each spot of the microarray. Finally, the differential scanning data were processed by computer imaging software (MicroArray Suite; Scanalytics, Inc., Fairfax, VA), to ascertain and profile the genes of interest.