Does prolactin act directly on AgRP neurons in the arcuate nucleus?

Abstract

The development of a positive energy balance occurs during pregnancy to support the growth of fetal and maternal tissues, and to increase energy stores in preparation for the metabolic demands of lactation. Leptin levels increase as pregnancy advances, along with appetite and fat deposition. This is unusual, as in the non-pregnant state, leptin decreases food intake and increases energy mobilization. Therefore, leptin is not being recognized during pregnancy. Leptin-responsive neurons, such as Agouti-related peptide (AgRP) and Pro-opiomelanocortin (POMC) neurons in the arcuate nucleus of the hypothalamus, do not respond normally to leptin during pregnancy. It is likely that changes in gestational hormonal profile have a role in mediating leptin resistance. Increased plasma prolactin is observed during pregnancy and, in addition, there are prolactin receptors in the same regions of the brain that control food intake. Prolactin is also known to have an orexigenic effect. We hypothesise that prolactin is a mediator of this increased appetite and resistance to leptin during pregnancy.

The aim of this study was to examine whether prolactin directly targets leptin-responsive neurons in the arcuate nucleus. We examined the AgRP neurons, known to be involved in stimulating food intake. Transgenic mouse models and double-label immunohistochemistry were used to look at whether prolactin action is observed in neurons that express AgRP. In separate experiments, AgRP-Cre-tdTomato mice and AgRP-Cre-tauGFP mice were injected with prolactin and perfused. The brains were used to detect prolactin-induced phosphorylation of STAT5, an indicator of prolactin action. The sections were then double-labelled to identify AgRP neurons with labeling of either the tdTomato transgene or GFP respectively. AgRP neurons were readily detected in the arcuate nucleus. Many pSTAT5-stained nuclei were also detected in the arcuate nucleus of prolactin-treated animals, however, none of the AgRP neurons showed pSTAT5-positive nuclei. To confirm these results, we looked at conditional deletion of the prolactin receptor in AgRP neurons using AgRP-PRLR flox/flox mice. We looked at differences in food intake and body weight between these knockout mice and wildtype littermates, no significant differences were observed. We also used immunohistochemistry to look at GFP expression. These mice would be expected to show Cre-induced expression of GFP if the prolactin receptor was normally expressed in the AgRP neurons. vGAT-Cre-Prlr flox/flox transgenic mice were used as a positive control as previous work suggests that GABA neurons express the prolactin receptor. While GFP was readily detected in the positive controls, no staining of any neurons was visible in the non-pregnant, pregnant or lactating brains from AgRP-PRLR flox/flox mice, indicating that the prolactin receptor gene is not normally expressed in AgRP neurons.

From these experiments we were able to conclude that prolactin-induced pSTAT5 does not colocalise with AgRP neurons, and that the prolactin receptor gene is not expressed in AgRP neurons. Therefore, the results obtained did not support our hypothesis that prolactin-induced increases in food intake are mediated by direct actions of prolactin by AgRP cells, and showed that prolactin does not directly act upon AgRP neurons to mediate leptin resistance during pregnancy.