(1) We used Affymetrix oligonucleotide microarrays to determine the effect of chronic administration of valproate, at therapeutically relevant concentrations, on rat brain gene expression. We found 87 down-regulated genes and 34 up-regulated genes of at least a 1.4-fold change in valproate-treated compared to control rats. The experiments were done on five independent samples for each group, each in duplicate. The genes affected are known to be involved in a variety of pathways, including synaptic transmission, ion channels and transport, G-protein signaling, lipid, glucose and amino-acid metabolism, transcriptional and translational regulation, phosphoinositol cycle, protein kinases and phosphatases, and apoptosis. Our results suggest that the therapeutic effect of valproate may involve the modulation of multiple brain signaling pathways. (2) When administered chronically to rats at therapeutically relevant concentrations, the mood stabilizers lithium, valproic acid, and carbamazepine have been shown to downregulate the expression of certain phospholipase A2 (PLA2) enzymes involved in brain arachidonic acid (AA) release and cyclooxygenase (COX)-mediated metabolism. Phase II clinical trials with the anticonvulsant topiramate suggest that this drug may also be effective for bipolar disorder. To see if topiramate had similar effects on brain AA metabolism, we administered topiramate to rats for 14 days at 20 mg/kg, p.o. twice daily. Topiramate treatment did not significantly affect the brain activity or protein level of cytosolic cPLA2, secretory sPLA2, or Ca2+-independent iPLA2. Additionally, brain protein levels of COX-1, COX-2, 5-lipoxygenase, and cytochrome P450 epoxygenase were unchanged. These results suggest that topiramate does not modify the expression of the enzymes involved in brain AA metabolism that have been shown to be targeted by lithium, valproic acid, or carbamazepine. Therefore, if topiramate proves effective in bipolar disorder, it may not act by modulating brain AA metabolism. In view of the proven anticonvulsant effect of topiramate, our results also suggest that the AA cascade is not involved in the anti-seizure properties of the drug. (3) We have previously shown that lithium chloride (LiCl), when given to rats for 6 weeks, decreased brain mRNA, protein, and activity levels of AA-selective cPLA2, without affecting secretory sPLA2 or iPLA2. We investigated whether transcription factors known to regulate cPLA2 gene expression are modulated by treatment with chronic lithium . Male Fischer-344 rats were fed a LiCl-containing diet for 6 weeks to produce a therapeutically relevant brain lithium concentration. Control animals were fed a LiCl-free diet. Using a gelshift assay, we found that LiCl significantly decreased activating protein 2 (AP-2)-binding activity, and protein levels of the AP-2 alpha and AP-2 beta but not of the AP-2 gamma subunits in the frontal cortex. Activating protein 1 (AP-1)-binding activity was increased, whereas glucocorticoid response element, polyoma enhancer activator 3, and nuclear factor kappa B DNA-binding activities were not changed significantly. Since both cPLA2 and AP-2 can be activated by protein kinase C (PKC), we examined the frontal cortex protein levels of PKC alpha and PKC epsilon, as well as AA-dependent PKC activity. The protein levels of PKC alpha and PKC epsilon were decreased significantly, as was AA-dependent PKC activity, in the lithium-treated rats. Our results suggest that the reported decrease in brain gene expression of cPLA2 by chronic lithium may be mediated by reduced AP-2 transcriptional activity, and that decreased expression of PKC alpha and PKC epsilon contributes to lowering the AP-2 activity. (4) Based on the observation that mood stabilizers target AA metabolism to PGE2 in rat brain and that these agents are clinically more effective in the manic phase of bipolar disorder, we hypothesized that PGE2 levels are altered during manic episodes. To test this hypothesis plasma PGE2 levels were measured in both manic bipolar patients, before and after treatment with mood stabilizers, and healthy age- and gender-matched controls. We found that plasma PGE2 levels, measured by ELISA, did not differ significantly in the manic patients compared with the controls. Additionally, there was no significant effect of anti-bipolar treatment (lithium and/or valproate) on plasma PGE2 in the manic patients. This also could be due to the high individual variability and the small sample size. There was no significant correlation of plasma PGE2 with age or gender of the subjects.
