Leptin signaling in astrocytes regulates hypothalamic neuronal circuits and feeding

Abstract

We found that leptin receptors were expressed in hypothalamic astrocytes and that their conditional deletion led to altered glial morphology and synaptic inputs onto hypothalamic neurons involved in feeding control. Leptin-regulated feeding was diminished, whereas feeding after fasting or ghrelin administration was elevated in mice with astrocyte-specific leptin receptor deficiency. These data reveal an active role of glial cells in hypothalamic synaptic remodeling and control of feeding by leptin.

Fig. 1. Cell autonomous impairment of leptin receptor signaling alters astrocyte morphology and reduces astrocytic coverage onto melanocortin cells

(a) Double fluorescence labeling of the astrocyte marker GFAP (red) and leptin receptors (leptin receptor-driven expression of EGFP, green) shows co-localization of GFAP-immunolabeling with EGFP-tagged leptin receptor-containing profiles (white arrows) in the arcuate nucleus (Arc). Scale bar = 100 um. (b) The truncated leptin receptor (exon 17) allele was confirmed by in situ hybridization combined with immunohistochemistry. Red fluorescence indicates GFAP-positive astrocyte and green fluorescence indicates POMC neurons. White dots indicate mRNA signals of the leptin receptor containing exon 17. (c) Bar graphs show the number of astrocytes or POMC cells expressing mRNA of leptin receptor in the Arc (GFAP+LepR: n=6 slices for GFAP-LepR^+/+ (+/+); n=6 slices for GFAP-LepR^−/− (−/−), p=0.0041, t(10)=3.699; POMC+LepR: n=6 slices for GFAP-LepR^+/+; n=6 slices for GFAP-LepR^−/−, p=0.524, t(10)=0.6603). White arrows indicate cells expressing mRNA of leptin receptors. White arrowheads indicate leptin receptor-negative cells. (d) Representative image of GFAP-immunolabeling in the Arc of GFAP-LepR^+/+ and GFAP-LepR^−/− mice. Scale bar = 100 μm. (e) The number of GFAP-positive cells did not differ between GFAP-LepR^+/+ and GFAP-LepR^−/− mice (n=14 slices for GFAP-LepR^+/+; n=12 slices for GFAP-LepR^−/−, p=0.9802, t(24)=0.02503) but (f) the number of primary projections (n=59 cells for GFAP-LepR^+/+; n=39 cells for GFAP-LepR^−/−, p=0.0334, t(96)=2.158) and (g) their length (n=59 cells for GFAP-LepR^+/+; n=37 cells for GFAP-LepR^−/−, p=0.0403, t(94)=2.079) were less in GFAP-LepR^−/− mice compared to GFAP-LepR^+/+ mice. (h) Representative electron micrograph showing astrocyte coverage (green pseudo-color and blue arrows) onto POMC-labeled cells. Scale bar = 1 μm. (i) POMC cells (n=14 cells or GFAP-LepR^+/+; n=14 cells or GFAP-LepR^−/−, p=0.0005, t(26)=3.989) as well as (j) unlabeled neurons (ULN) (n=12 cells for GFAP-LepR^+/+; n=10 cells for GFAP-LepR^−/−, p=0.0316, t(20)=1.967) in their vicinity of GFAP-LepR^−/− mice had less coverage of their perikaryal membranes by astrocytic processes compared to controls. *, p<0.05; **, p<0.01; ***, p<0.001 versus (+/+). Results are means ± the s.e.m. P values for unpaired comparisons were analyzed by two-tailed Student's t-test.

Fig. 2. Impaired leptin receptor signaling in astrocytes increases the number of synapses onto POMC and AgRP neurons

(a) Representative electron micrograph showing astrocyte coverage (green pseudocolor) and synapses (black arrows) onto POMC-labeled cells. Scale bar = 1 μm. (b) POMC cells (n=19 cells for GFAP-LepR^+/+ (+/+); n=15 cells for GFAP-LepR^−/− (−/−), p=0.0097, t(32)=2.751 for Symmetric; p=0.0311, t(32)=2.255 for Asymmetric; p=0.0047, t(32)=3.039 for Total) as well as (c) unlabeled neurons (ULN) (n=12 cells for GFAP-LepR^+/+; n=10 cells for GFAP-LepR^−/−, p=0.0466, t(20)=1.763 for Symmetric; p=0.0352, t(20)=2.259 for Asymmetric; p=0.0297, t(20)=2.341 for Total) in their vicinity of GFAP-LepR^−/− mice had elevated numbers of symmetric, asymmetric and hence, total number of synapses on their perikaryal membrane compared to controls. (d) POMC neurons (identified by POMC-driven GFP labeling) of GFAP-LepR^−/− mice had an elevated frequency of mIPSCs (n=9 cells for GFAP-LepR^+/+; n=9 cells for GFAP-LepR^−/−, p=0.0203, t(16)=2.576) but (e) no changes in frequency of mEPSCs (n=23 cells for GFAP-LepR^+/+; n=25 cells for GFAP-LepR^−/−, p=0.5513, t(45)=0.6003). AgRP neurons (identified by NPY-driven hrGFP labeling) of GFAP-LepR^−/− mice had an elevated frequency of (f) mIPSCs (n=9 cells for GFAP-LepR^+/+; n=9 cells for GFAP-LepR^−/−, p=0.0493, t(16)=2.127) and (g) mEPSCs (n=9 cells for GFAP-LepR^+/+; n=9 cells for GFAP-LepR^−/−, p=0.0164, t(16)=2.681). *, p<0.05; **, p<0.01 versus (+/+). Results are means ± the s.e.m. P values for unpaired comparisons were analyzed by two-tailed Student's t-test.

Fig. 3. Impairment of leptin receptor signaling in astrocytes blunts leptin-induced anorexia and enhances fasting or ghrelin-induced hyperphagia

(a, b) GFAP-LepR^−/− (−/−) mice showed blunted suppression of feeding in response to leptin administration (Fig. 3a: n=5 mice for GFAP-LepR^+/+ (+/+)-vehicle, GFAP-LepR^−/−-vehicle and GFAP-LepR^−/−-leptin; n=6 mice for GFAP-LepR^+/+-leptin, p=0.052, F(1,17)=4.386 for 1 h; p=0.018, F(1,17)=6.873 for 2 h; Fig. 3b: n=6 mice per group, p=0.0095, F(9,36)=2.971). (c) Representative images show double labeled POMC-GFP and Fos cells in the Arc of GFAP-LepR^+/+ and GFAP-LepR^−/− mice. Scale bar = 100 μm. (d) Number of Fos-positive POMC cells induced by leptin treatment was reduced in GFAP-LepR^−/− mice (n=6 slices for GFAP-LepR^+/+-vehicle; n=12 slices for GFAP-LepR^+/+-leptin; n=13 slices for GFAP-LepR^−/−-leptin, p=0.013, t(16)=2.788 for GFAP-LepR^+/+-vehicle versus GFAP-LepR^+/+-leptin; p=0.0056, t(23)=3.055 for GFAP-LepR^+/+-leptin versus GFAP-LepR^−/−-leptin). (e, f) GFAP-LepR^−/− mice showed increased feeding after fasting or ghrelin administration (Fig 3e: n=6 mice for GFAP-LepR^+/+; n=7 mice for GFAP-LepR^−/−, p=0.0378, t(11)=2.361 for 1 h; p=0.0092, t(11)=3.150 for 3 h; Fig 3f: n=5 mice for GFAP-LepR^+/+-vehicle; n=6 mice for GFAP-LepR^−/−-vehicle; n=12 mice for GFAP-LepR^+/+-ghrelin; n=11 mice for GFAP-LepR^−/−-ghrelin, p=0.04, F(1,32)=4.57). (g) Representative images show double labeled AgRP-GFP and Fos cells in the Arc of GFAP-LepR^+/+ and GFAP-LepR^−/− mice. Scale bar = 100 μm. (h) Number of Fos-positive AgRP cells induced by overnight fasting was enhanced in GFAP-LepR^−/− mice (n=12 slices for GFAP-LepR^+/+-fed; n=20 slices for GFAP-LepR^+/+-fasted; n=17 slices for GFAP-LepR^−/−-fasted, p<0.0001, t(30)=6.721 for GFAP-LepR^+/+-fed versus GFAP-LepR^+/+-fasted; p<0.0001, t(35)=6.848 for GFAP-LepR^+/+-fasted versus GFAP-LepR^−/−-fasted). White arrows indicate double-labeled cells. *, p<0.05; **, p<0.01; ***, p<0.001 versus (+/+), leptin or fasted. Results are means ± the s.e.m. P values for unpaired comparisons were analyzed by two-tailed Student's t-test. Two-way ANOVA was performed to detect significant interaction between genotype and treatment (leptin or ghrelin). Two-way repeated measures ANOVA was performed to detect significant interaction between genotype and time (multiple injections of leptin).