Magel2 is required for leptin-mediated depolarization of POMC neurons in the hypothalamic arcuate nucleus in mice

Abstract

Prader-Willi Syndrome is the most common syndromic form of human obesity and is caused by the loss of function of several genes, including MAGEL2. Mice lacking Magel2 display increased weight gain with excess adiposity and other defects suggestive of hypothalamic deficiency. We demonstrate Magel2-null mice are insensitive to the anorexic effect of peripherally administered leptin. Although their excessive adiposity and hyperleptinemia likely contribute to this physiological leptin resistance, we hypothesized that Magel2 may also have an essential role in intracellular leptin responses in hypothalamic neurons. We therefore measured neuronal activation by immunohistochemistry on brain sections from leptin-injected mice and found a reduced number of arcuate nucleus neurons activated after leptin injection in the Magel2-null animals, suggesting that most but not all leptin receptor-expressing neurons retain leptin sensitivity despite hyperleptinemia. Electrophysiological measurements of arcuate nucleus neurons expressing the leptin receptor demonstrated that although neurons exhibiting hyperpolarizing responses to leptin are present in normal numbers, there were no neurons exhibiting depolarizing responses to leptin in the mutant mice. Additional studies demonstrate that arcuate nucleus pro-opiomelanocortin (POMC) expressing neurons are unresponsive to leptin. Interestingly, Magel2-null mice are hypersensitive to the anorexigenic effects of the melanocortin receptor agonist MT-II. In Prader-Willi Syndrome, loss of MAGEL2 may likewise abolish leptin responses in POMC hypothalamic neurons. This neural defect, together with increased fat mass, blunted circadian rhythm, and growth hormone response pathway defects that are also linked to loss of MAGEL2, could contribute to the hyperphagia and obesity that are hallmarks of this disorder.

Figure 1. Magel2-null mice have abnormal weight recovery and compensatory refeeding after fasting.

A) Body weights of adult male mice subjected to a 48 h fast and a 72 h refeeding period. Magel2-null mice lost less weight while fasting, and recovered less weight during refeeding (*P<0.05, compared between genotypes by Student's t-test). B) 24 h food intake pre-fast and C) post-fast (*P<0.01). D) Magel2-null mice have a reduced food intake ratio - the ratio of food consumed after fasting to food consumed before fasting - compared to controls (*P<0.05). n = 6 mice of each genotype. Values are means ± SEM.

Figure 2. Peripherally administered leptin fails to reduce food intake in Magel2-null mice.

A) Body weights of 20 week old male mice (*P<0.05, compared between genotypes by Student's t-test), and B) 24 h food intake following leptin or PBS (saline) injections. Leptin reduced food intake in control mice (*P<0.005, compared before and after treatment), while leptin-treated Magel2-null mice showed no significant change in food intake. n = 10 of each genotype. C) Body weights of 6 week old male mice and D) 24 h food intake following leptin or PBS injections. Leptin reduced food intake in control (*P<0.05, compared before and after treatment) but not Magel2-null mice. Controls, n = 7, Magel2-null, n = 6. Values are means ± SEM.

Figure 3. pSTAT3 and c-fos expression in ARC neurons in leptin-treated mice.

A–D) Representative immunohistochemistry (IHC) images showing pSTAT3 immunoreactivity following ip PBS (saline) or leptin injection. Scale bar, 100 µm. E) Magel2-null mice have fewer pSTAT3 positive cells than control following leptin injection (Bregma −1.28, P<0.001; Bregma −1.64, P<0.005; Bregma −2.12, P>0.5, *compared between genotypes by Student's t-test) but both genotypes have more pSTAT3 positive cells after leptin treatment than after PBS (#, P<0.05, compared between treatments by Student's t-test; n = 8 mice of each genotype). F–I) Representative images of c-fos IHC following ip PBS and leptin injection. Scale bar 100 µm. J) Leptin induced c-fos expression in both Magel2-null and control mice compared to PBS (#P<0.05, compared between treatments by Student's t-test). At both baseline and following leptin treatment, Magel2-null mice had significantly fewer c-fos positive cells compared to controls (Bregma −1.28 ∼18% reduction, Bregma −1.64 ∼20% reduction, Bregma −2.12 ∼35% reduction, *P<0.05, compared between genotypes by Student's t-test). Values in E) and J) are means ± SEM.

Figure 4. Magel2-null mice have fewer ARC POMC neurons.

A–B) Representative images of GFP (POMC) IHC in A) control and B) Magel2×POMC^EGFP mice. Scale bar 100 µm. C) Magel2×POMC^EGFP mice have fewer GFP expressing (POMC) cells at all levels of the ARC (Bregma −1.28, 39% reduction, *P<10⁻⁸; Bregma −1.64, 27% reduction, *P<10⁻⁵; Bregma −2.12, 52% reduction, *P<0.01), compared between genotypes by Student's t-test. Values are means ± SEM.

Figure 5. Magel2 is required for the leptin-induced depolarizing response in POMC neurons.

Example of a LepRb^EGFP positive neuron identified for electrophysiological recordings using: A) infrared-differential interference contrast imaging (scale bar, 10 µm) and B) epifluorescence. C) Mean RMP of ARC LepRb+ neurons (n>50 of each genotype, values are means ± SEM). D) Current clamp recording of a LepRb^EGFP neuron showing the hyperpolarizing effect of 300 nM NPY. E) There was no difference in the magnitude of hyperpolarization between control and Magel2-null neurons treated with 300 nM NPY. F–H) Current clamp recordings of typical responses to 100 nM leptin in F) depolarizing neurons (ΔRMP>2 mV over baseline), and G) hyperpolarizing neurons (ΔRMP>2 mV below baseline). H) Changes in RMP with application of 100 nM leptin to ARC LepRb^EGFP neurons. Circles represent individually tested neurons. Depolarizing, hyperpolarizing, and unresponsive neurons were found in control slices, while only hyperpolarizing and unresponsive neurons were found in Magel2-null slices. Difference between genotypes is significant by Fishers Exact Test, P<10⁻⁸. I) Mean RMP of ARC POMC^EGFP neurons (n>40 of each genotype, means ± SEM). J) Changes in RMP caused by application of 100 nM leptin to ARC POMC^EGFP neurons. Depolarizing responses were observed in control but not Magel2-null neurons. Difference between genotypes in the number of depolarizing neurons is significant by Fishers Exact Test, P<10⁻⁸.

Figure 6. Magel2 is not required for leptin-mediated responses in the VMN.

A) While injection of leptin induced a significant increase in pSTAT3 immunoreactivity compared with saline throughout the entire VMN both in control and Magel2-null mice, there was no difference between genotypes in cell numbers demonstrating pSTAT3 immunoreactivity in response to leptin injection. Values are means ± SEM, ^# P<0.05, compared within genotypes for saline vs. leptin treatment (Student's t-test). B) Changes in RMP with application of 100 nM leptin to VMN LepRb^EGFP neurons. Data points represent individually tested neurons (total n = 68 in control, n = 30 in Magel2-null). Depolarizing, hyperpolarizing, and unresponsive neurons were found in both control and Magel2-null slices. There was no significant difference between genotypes in the populations of depolarizing or hyperpolarizing responses to leptin application (Fishers Exact Test, P>0.2).

Figure 7. Effects of intraperitoneal MT-II on food intake.

Food intake following 2.5 mg/kg injection of MT-II or saline in 24 h fasted mice. MT-II significantly reduced food intake for the first 2 h of refeeding in control mice, but this effect was no longer present by 4 h. In Magel2-null mice, MT-II reduced food intake to a greater extent and this effect was still evident 24 h after injection. n = 6 mice of each genotype. Values are means ± SEM. *P<0.05, compared to saline-injected (100%) food intake within each genotype by Student's t-test.