COX-1 and COX-2 produce prostanoids from arachidonic acid (AA) and are thought to have important yet distinct roles in normal brain function. Deletion of COX-1 or COX-2 results in profound differences both in brain levels of prostaglandin E2 (PGE2)and in activation of the transcription factor nuclear factor-B (NF-kB), suggesting that COX-1 and COX-2 play distinct roles in brain arachidonic acid metabolism and regulation of gene expression. To further elucidate the role of COX isoforms in the regulation of the brain transcriptome, we performed microarray analysis of gene expression in the cerebral cortex and hippocampus of mice deficient in COX-1 (COX-1 KO) or COX-2 (COX-2 KO).? Although a majority of the genes (>93%) shown to be differentially expressed in the separate brain regions of COX-1 KO or COX-2 KO mice only occurred in mice null for one isoform and not the other, the expression of some genes was changed in both COX-1 KO and COX-2 KO mice. In the cerebral cortex, there were four genes with altered expression in both COX-1 KO and COX-2 KO mice. Of these four, Rho-GDP dissociation inhibitor exhibited increased expression in both genotypes. The expression of the other three genes (mitochondrial inner membrane protein, GABA transporter GAT3, and ring finger protein 24) changed in opposite directions (downregulated in COX-1 KO and upregulated in COX-2 KO mice), suggesting an isoform-specific effect on expression of these genes. Such an isoform-specific effect was not observed in the hippocampus, where all 12 genes whose expression was altered in both COX-1 KO and COX-2 KO mice exhibited changes in a similar direction in both genotypes when compared with wild-type mice, suggesting that this cohort of genes is responsive to a general alteration in AA metabolism that is not specific to COX-1 or COX-2.? In cerebral cortex of COX-2 KO mice, we identified increased expression of three genes that work in tandem in the final steps of short chain lipid metabolism by oxidation, namely hydroxyacyl-coenzyme A dehydrogenase type II (HADH2), acetyl-coenzyme A acetyltransferase 1 (ACAT1) and ATP citrate lyase (ACLY). Gene expression of ACAT 1, but not that of HADH2 and ACLY, was found to be decreased in the hippocampus of COX-1 KO mice. Expression of methionine adenosyltransferase II (MAT2B) and adenosylhomocysteine hydrolase (AHCY), two genes that work in tandem to metabolize methionine to homocysteine, was increased in cerebral cortex of COX-2 KO mice. Expression of MAT2B, but not of AHCY, was decreased in COX-1 KO hippocampus. Expressions of two genes involved in GABA neurotransmission were altered in cerebral cortex and hippocampus of COX KO mice. Microarray analysis of GABA-A receptor subunit 1 (GABRB1) in the hippocampus of COX-2 KO mice demonstrated downregulation of gene expression. Validation with Q-PCR demonstrated that the expression of GABRB1 was not only decreased in hippocampus of COX-2 KO mice but it was also decreased in cerebral cortex of COX-2 KO mice. GABA transporter (GAT)3 expression in cerebral cortex of COX KO mice exhibited a genotype-specific effect. COX-1 KO mice demonstrated downregulation of gene expression whereas COX-2 KO showed upregulation of mRNA level. Janus kinase (JAK) isoforms 1 and 2 were found to be expressed in a genotype-dependent manner in hippocampus. COX-1 KO mice had increased expression of JAK1 but not of JAK2. On the other hand, JAK-2 expression was decreased in COX-2 KO but not in COX-1 KO mice.? Overall, our findings suggest that ablation of COX activity alters the transcription of many genes, including those involved in oxidation, methionine metabolism, GABA neurotransmission, and cytokine signaling. Although some of the molecular mechanisms underlying these changes are not well understood at this time, these data identify metabolic and signaling pathways that were previously not known to be affected by COX. Because many anti-inflammatory and analgesic treatments, such as nonsteroidal anti-inflammatory drugs, rely on reduction in COX activity for their mechanism of action, the specific alterations observed in this study expand our understanding of the therapeutic and toxicologic consequences of COX inhibition.
