Effects of oral estrogen, raloxifene and arzoxifene on gene expression in serotonin neurons of macaques

The serotonin neural system contributes to cognition and affect, both of which exhibit pathologies with gender bias. We previously showed that estrogen (E) treatment of female macaques via Silastic implant alters gene expression for tryptophan hydroxylase (TPH), the serotonin reuptake transporter (SERT) and the 5HT1A autoreceptor. In addition, we have found that serotonin neurons of macaques express ER beta (ERβ). Together these studies suggest that the serotonin neural system could transduce the action of estrogen via ERβ on aspects of mood and cognition. However, estrogen replacement therapy can increase the risk for breast and uterine cancer. Therefore, we questioned whether the selective estrogen receptor modulators, raloxifene and arzoxifene, act in a manner similar to E on gene expression in serotonin neurons of a nonhuman primate model. Female rhesus macaques were ovariectomized and orally dosed with vehicle, estradiol 17β, raloxifene or arzoxifene once per day by sipper bottles for 30 days. The animals were then euthanized and the midbrains were prepared for in situ hybridization for TPH, SERT and 5HT1A receptor mRNAs followed by densitometric analysis. There was a significant increase in TPH total signal (positive pixelsxOD) with E, raloxifene and arzoxifene, respectively. There was a significant decrease in SERT mRNA optical density with all treatments. 5HT1A autoreceptor mRNA did not change with any treatment. If these changes in gene expression are reflected by similar changes in the functional proteins, then raloxifene or arzoxifene could increase serotonin neurotransmission with little or no negative action in peripheral tissues. In conclusion, the selective estrogen receptor modulators, raloxifene and arzoxifene, act in a manner similar to natural E on TPH and SERT mRNA expression in serotonin neurons. This suggests that raloxifene and arzoxifene are agonists at ERβ in the context of the serotonin neuron. However, the responses to E were more variable and less robust with the oral dosing paradigm compared to a chronic implant paradigm.