doi: 10.1038/s41467-021-26388-4.
Orexin-A and endocannabinoids are involved in obesity-associated alteration of hippocampal neurogenesis, plasticity, and episodic memory in mice
Affiliations
- PMID: 34675233
- PMCID: PMC8531398
- DOI: 10.1038/s41467-021-26388-4
Item in Clipboard
Orexin-A and endocannabinoids are involved in obesity-associated alteration of hippocampal neurogenesis, plasticity, and episodic memory in mice
Nat Commun.
.
Abstract
The mammalian brain stores and distinguishes among episodic memories, i.e. memories formed during the personal experience, through a mechanism of pattern separation computed in the hippocampal dentate gyrus. Decision-making for food-related behaviors, such as the choice and intake of food, might be affected in obese subjects by alterations in the retrieval of episodic memories. Adult neurogenesis in the dentate gyrus regulates the pattern separation. Several molecular factors affect adult neurogenesis and exert a critical role in the development and plasticity of newborn neurons. Orexin-A/hypocretin-1 and downstream endocannabinoid 2-arachidonoylglycerol signaling are altered in obese mice. Here, we show that excessive orexin-A/2-arachidonoylglycerol/cannabinoid receptor type-1 signaling leads to the dysfunction of adult hippocampal neurogenesis and the subsequent inhibition of plasticity and impairment of pattern separation. By inhibiting orexin-A action at orexin-1 receptors we rescued both plasticity and pattern separation impairment in obese mice, thus providing a molecular and functional mechanism to explain alterations in episodic memory in obesity.
© 2021. The Author(s).
Conflict of interest statement
The authors declare no competing interests.
Figures
a Representative scheme and timeline of the novel object recognition test in lean, HFD, and ob/ob mice; n = 10 per group. b NOR in lean, HFD, and ob/ob mice calculated during the habituation phase; n = 10 per group. c Short-term retention phase using similar objects; NOR in lean, HFD, and ob/ob mice calculated after 1.5 h, NOR = −0.04 ± 0.08 and 0.01 ± 0.05 in HFD and ob/ob respectively compared to lean mice with a NOR = 0.31; n = 10 per group, Kruskal−Wallis test and post hoc Dunn’s test, **p < 0.01, Kruskal−Wallis statistic = 15.07. d Long-term retention phase. NOR in lean (gray), HFD (orange), and ob/ob (green) mice calculated after 24 h, NOR 0.25 ± 0.03 in lean NOR = 0.07 ± 0.05 in HFD and NOR = −0.20 ± 0.08, 0.1 in ob/ob; n = 10 mice per group. Kruskal−Wallis test and post hoc Dunn’s test, *p < 0.05 and ***p < 0.001, Kruskal−Wallis statistic = 16.54. e−g Short-term retention phase using distinct objects; NOR in lean, HFD, and ob/ob mice calculated in habituation phase after 1.5 and 24 h for the long-term retention phase, respectively; n = 10 per group. The box plots elements are: center line, median (Q2); square symbol, mean; box limits, 25th (Q1)−75th (Q3) percentiles; whisker length is determined by the outermost data points. Source data are provided as a Source Data file.
a Schematic representation of the water maze protocol (the black and gray dot inside the pool represents the position of the visible and hidden platform, respectively). b, c Latency and path lengths to reach the platform during the visible task and training. ANOVA test and Bonferroni post hoc test; latency *p < 0.0001; F = 24.37; path length *p < 0.0001; F = 22.89; n = 7 lean mice and n = 9 HFD mice. d Percentage of time spent in the target quadrant compared to the others during the probe trial (lean: target quadrant = 44.29% ± 4.1%, other quadrants = 18.57% ± 1.3%; HFD: target quadrant = 51.72% ± 4.3% and other quadrants = 16.07% ± 1.4%. ANOVA test and Bonferroni post hoc test, *p < 0.0001; F = 56.9; n = 7 lean mice and n = 9 HFD mice. e Representative target accuracy in the probe trial was indicated by heat maps. Dark red zones represent a sixfold presence probability. f−h Latency to the first goal entry (lean = 7.9 ± 2.1, HFD = 18.21 ± 3.7; #p < 0.05; F = 4.88), path efficiency (lean = 0.5 ± 0.08, HFD = 0.3 ± 0.05; #p < 0.05; F = 5.55) and goal crossings (#p < 0.05; F = 4.82; lean = 10 ± 0.8, HFD = 7.5 ± 0.7) are all measures related to a precise localization of the platform position in the quadrant during the probe trial. Histograms represent mean ± SEM. *comparison within the groups; #comparison between the groups. n = 7 lean mice and n = 9 HFD mice. Source data are provided as a Source Data file.
a−c Graphs showing the NPCs proliferative and differentiative efficiency by mean of quantitative stereological counting of Ki67-ir (a), or DCX-ir (b) or NeuroD-ir (c) cell profiles in the DG of lean and obese (ob/ob and HFD) mice. Graphs show box-whisker plots (including minima, maxima, and median values, lower and upper quartiles) with single values. Data are from n = 30 slices in n = 6 mice per group; ANOVA test with Bonferroni post hoc, ****p < 0.0001, F = 47.48). d Graph showing dendritic development and complexity of DCX-positive neurons by quantification of PSD95/DCX-ir colocalizing area vs DCX-ir area in the inner (IML) and medial (MML) molecular layers of DG in lean and obese mice. Graphs show box-whisker plots (including minima, maxima, and median values, lower and upper quartiles) with single values. Data are mean of the percentage ± percentage of SEM from n = 30 slices in n = 6 mice per group; ANOVA test with Bonferroni post hoc, ****p < 0.0001, F = 47.48). e−g Confocal images of DCX/PSD95/DAPI immunolabeled of the dentate gyrus (DG) showing the adult-born granule cells DCX-ir (green) which constitute a subpopulation of the granular cell layer with higher PSD95-ir (red) postsynaptic density in the IML and MML of obese (f, g) and lean (e) mice as demonstrated by the high degree of colocalizing DCX/PSD95-ir puncta (yellow-orange) as an index of functional integration of the adult-born granule cells in the DG revealed by the high-density percentage of PSD95/DCX-ir colocalizing area vs DCX-ir area in the ML of obese mice (lean: 19.5% ± 0.9%, HFD: 34.2% ± 1.1%, ob/ob: 29.8% ± 1.1%. n = 30 slices in n = 6 mice per group; ANOVA test with Bonferroni post hoc, ****p < 0.0001, F = 47.48) (scale bar: 100 µm). Immunolabeling was repeated n = 3 times independently in sections from n = 6 mice per group, with similar results. h−j High-power images of typical DCX-ir adult-born granule cells demonstrating the development, length, and branching complexity of DCX-positive neurons in obese than in lean mice (scale bar: 10 µm). k Percentage of basal, proximal, and medial dendrites revealed by PSD95/DCX-ir vs DCX-ir colocalizing area in the respective GCL, IML, and MML (ML) were also significantly different between DG of obese and lean mice. Two-way ANOVA with Tukey post hoc, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data are mean of the percentage ± percentage of SEM from n = 30 slices in n = 6 mice per group: Immunolabeling was repeated n = 3 times independently in sections from n = 6 mice per group, with similar results. l Schematic representation of the cytoarchitectonic organization of the molecular layer (ML) in DG of lean and obese, HFD and ob/ob, mice. GCL: granular cell layer, IML: inner molecular layer, MML: medial molecular layer, GC: granular cell, md: medial dendrites, pd: proximal dendrite, gcld: granular cell layer dendrite. The box plots elements are: center line, median (Q2); square symbol, mean; box limits, 25th (Q1)−75th (Q3) percentiles; whisker length is determined by the outermost data points. Source data are provided as a Source Data file.
a Experimental configuration of LTP recording in lean and ob/ob mice in brain slices. Position of the stimulating and recording electrodes in the molecular layer (ML Rec) and medial perforant path (MPP stim), respectively. The stimulation of the MPP fibers induces a paired-pulse depression. b Representative fEPSPs were recorded before and after the induction of the LTP in the two genotypes. c Plot of the fEPSP slope recorded in brain slices before and after the induction of the LTP (time 0) in the DG of lean and ob/ob mice. Data are presented as mean values ± SEM. The extent of LTP was calculated as a percentage of the baseline between the last 40 and 60 min of recording. n = 10 slices from four lean mice and n = 7 slices from four ob/ob mice. d Box plot of the LTP in lean and ob/ob mice, n = 10 slices from four mice in lean (slope 40−60 min post HFT = 120.5% ± 6.9%) and seven slices from four ob/ob mice (slope 40−60 min post HFT = 143.5% ± 24.37%). e Schematic representation of the in vivo LTP recording, the stimulation electrode was placed in the MPP, the recording site was in the ML. f Representative traces recorded in lean, HFD, and ob/ob before and after the TBS. g Plot of the fEPSP slope recorded before and after the induction of the LTP (time 0) in the DG of n = 5 lean mice, n = 4 HFD mice, and n = 5 ob/ob mice; data are presented as mean values ± SEM, as the extent of LTP was calculated as a percentage of the baseline between the last 40 and 60 min of recording. h Box plot of the LTP in lean, HFD, and ob/ob mice, n = 5 lean mice; n = 4 HFD and n = 5 ob/ob mice. TBS in lean: 251.4% ± 16.09%; 94.01% ± 8.03%, in HFD, 103.55% ± 13.48% in ob/ob, Kruskal−Wallis test and post hoc Dunn’s test, *p < 0.05, Kruskal−Wallis statistic = 9.1. The box plots elements are: center line, median (Q2); square symbol, mean; box limits, 25th (Q1)−75th (Q3) percentiles; whisker length is determined by the outermost data points. Source data are provided as a Source Data file.
a Box plot of the OxA levels in lean + vehicle (lean), lean + OxA (40 µg/kg, i.p), HFD + vehicle, ob/ob, ob/ob + Lep (5 mg/kg). lean + veh: 4.82 ± 0.1 pmol/mg, lean + OxA: 10.10 ± 0.8 pmol/mg, HFD + vehicle = 10.57 ± 0.2 pm/mg, ob/ob + veh: 11 + 0.2 pmol/mg, ob/ob + leptin 5.02 ± 0.1 pmol/mg. n = 21 (7 mice per group in triplicate), ANOVA test and Bonferroni post hoc test, ****p < 0.0001, F = 222.1. b Representative fEPSPs recorded before and after the induction of the LTP in lean, lean + OxA 100 nM, lean + OxA 200 nM, lean + OxA 200 nM + SB 10 μM. c Plot of the fEPSP slope recorded in brain slice before and after the induction of the LTP (time 0) in the DG of lean slices, lean + OxA (100 nM), lean + OxA (200 nM) and lean + OxA + SB (200 nM OxA + SB 10 µM). n = 10 slices from four mice (lean), n = 8 slices from four mice (lean + OxA 100 nM), n = 11 slices from five mice (lean+ 200 nM OxA or lean+ 200 nM OxA+ SB), data are presented as mean values ± SEM. d Box plot of the LTP in the four experimental conditions plotted in (c). n = 10 slices from four mice (lean), slope 40−60 min post HFT: 120.5% ± 6.938% from control slices, n = 8 slices from four mice (lean + OxA 100 nM), slope 40−60 min post HFT: 116.8% ± 6.938%, n = 11 slices from five mice for lean+ 200 nM OxA and lean+ 200 nM OxA +SB groups, slope 40−60 min post HFT: 91.83% ± 6.933% and 116.1% ± 7.023% respectively. *p < 0.05; ANOVA with Bonferroni post hoc test, F = 4.90. e−g Confocal images of DCX/OxA/CB1 immunolabeling of dentate gyrus (DG) showing a wide density of distribution of CB1/DCX-ir (light blue, arrow) or OxA/DCX-ir (yellow-orange, arrowhead) colocalizing puncta in the IML-MML of lean (e) and obese (f, g) mice. e1−g2 High-power fluorescent micrographs of orthogonal stacks are shown for each respective (e−g) area; dotted lines and crosshairs are used to show 3D coordinates and define the area of interest indicated by arrows 1 and arrowheads 2 which correspond to numbered insets in (e−g). Scale bar: 10 μm (e−g) and 2 μm (e1−g2). h−j Graphs showing the percentage of CB1/DCX-ir or OxA/DCX-ir or CB1/OxA/DCX-ir colocalizing area/µm2 of DCX-ir area in the IML-MML (h), or OxA-ir area/µm2 of IML-MML area (i), or CB1/VGluT1 colocalizing area/µm2 of VGluT1-ir area in the IML-MML (j), in lean and obese mice. Graphs show box-whisker plots (including minima, maxima, and median values, lower and upper quartiles) with single values. Data are mean of the percentage ± percentage of SEM from n = 30 slices in six mice per group; Kruskal−Wallis test and post hoc Dunn’s test, ****p < 0.0001). k, l Confocal images of VGluT1/NeuroD/CB1 immunolabeling of DG in lean (k) and obese ob/ob mice (l) showing a wide density of distribution of CB1/VGluT1 (light blue) colocalizing puncta in the IML-MML. Scale bar: 200 µm. Immunolabeling was repeated n = 3 times, independently in different DG sections from n = 6 mice per group, with similar results. The box plots elements are: center line, median (Q2); square symbol, mean; box limits, 25th (Q1)−75th (Q3) percentiles; whisker length is determined by the outermost data points. Source data are provided as a Source Data file.
a Box plot of 2-AG concentrations in lean + vehicle (lean), ob/ob, lean + OxA (40 µg/kg, i.p), lean + OxA (40 µg/kg, i.p) + SB (30 mg/kg, i.p). n = 9 lean mice, n = 10 ob/ob mice, n = 10 lean injected with OxA (40 µg/kg, i.p), n = 10 lean mice injected with OxA (40 µg/kg, i.p) + SB (30 mg/kg, i.p), ANOVA with Bonferroni test, F = 84.35; ****p < 0.0001. The box plots elements are: center line, median (Q2); square symbol, mean; box limits, 25th (Q1)−75th (Q3) percentiles; whisker length is determined by the outermost data points. b−d Confocal images of DCX/PSD95/CB1 immunolabeling of DG in lean and obese mice showing a wide density of distribution of CB1 at PSD95-ir postsynaptic side of DCX-ir dendritic spines. b1−d1 High-power fluorescent micrographs of orthogonal stacks are shown for each area and referred to the box of the respective (b−d) images; dotted with lines and crosshairs are used to show 3D coordinates. Scale bars: 10 μm (b−d) and 2 μm (b1−d1). Immunolabeling was repeated n = 3 times independently in different DG sections from n = 6 mice per group, with similar results. e−h Correlative light and electron microscopy (CLEM) showing OxA/Ox1-R or OxA/DAGLα or Ox1-R/DAGLα or CB1/DAGLα immunogold reactivity at synapses of immature granule-eGFP cells in the MML of DG. e Asymmetrical, putative excitatory, axodendritic synapse between OxA-positive axon terminal (at1) opposite to the dendritic spine (ds1) of immature granule-eGFP cell showing marked Ox1-R immunogold labeling at the edges of the postsynaptic density (arrows). The CLEM analysis revealed the dendritic spines of the newborn-eGFP neurons expressing OX-1R labeling (10 nm immunogold particles at postsynaptic sites of asymmetrical (i.e. putative excitatory)) endings containing OxA dense-core vesicles (20 nm immunogold particles). f Dendritic spines of newborn-eGFP neurons, carrying DAGLα enzyme at the postsynaptic side (10 nm immunogold particles) were found as opposed to OxA-positive puncta (20 nm immunogold particles) at asymmetrical (i.e. putative excitatory) synapses. DAGLα/Ox1-R (g) or DAGLα/CB1 (h) immunogold labeling reveals the expression of DAGLα in the proximity to the Ox1-R dendritic spines (g) or the presynaptic nerve endings carrying CB1 receptors (h) at asymmetrical, i.e. putative excitatory, synaptic cleft. Scale bar 0.1 µm (e−h). Immunolabeling was repeated with similar results n = 3 times independently in n = 6 different ultrathin sections/mouse from n = 3 mice. Source data are provided as a Source Data file.
a Plot of the fEPSP slope recorded in brain slice of lean mice before and after the induction of the LTP (time 0) in lean mice, OxA 200 nM +AM251, OxA 200 nM +O7460, ACEA. n = 10 slices (control), n = 8 slices (OxA + AM251), n = 7 slices (OxA +O7460), n = 9 slices (ACEA) from n = 5 mice/treatment; data are presented as mean values ± SEM. b Box plot of the LTP in the four experimental conditions. n = 10 slices control, n = 8 from OxA +AM251, n = 7 OxA +O7460, n = 9 ACEA from five mice each. Slope 40−60 min post HFT in lean mice = 120.5% ± 6.9%; AM251 (4 µM) 40−60 min post HFT: 135.8% ± 14.83%; O-7460 (10 µM) slope 40−60 min post HFT: 121.1% ± 11.82%; ACEA (125 nM) slope 40−60 min post HFT: 84.73% ± 5.5%. Kruskal−Wallis test and post hoc Dunn’s multiple comparison test, *p < 0.05, Kruskal−Wallis statistic = 11.69. c Plot of the fEPSP slope recorded in vivo before and after the induction of the LTP in HFD mice (time 0), n = 4 mice per group, data are presented as mean values ± SEM. d Box plot of the LTP in the three experimental conditions in HFD mice, HFD + SB (SB—0.1 nmol/0.5 μl) and HFD + AM251 (0.5 nmol/0.5 μl). 195.7% ± 30.39% of the LTP baseline for HFD+SB; 109.4% ± 4.7% of the LTP baseline in HFD + AM251; Kruskal−Wallis test and post hoc Dunn’s multiple comparison test, *p < 0.05, Kruskal−Wallis statistic = 8, n = 4 mice per group. e Plot of the fEPSP slope recorded in vivo before and after the induction of the LTP (time 0) in ob/ob mice. n = 5 vehicle-injected, n = 4 SB-injected and n = 4 AM251-injected ob/ob mice; data are presented as mean values ± SEM. f Box plot of the LTP in the three experimental conditions in ob/ob mice, ob/ob + SB (SB—0.1 nmol/0.5 μl) and ob/ob + AM251 (0.5 nmol/0.5 μl). LTP was 175.6% ± 3.77% in ob/ob + SB and 190% ± 21.5% in ob/ob + AM251. n = 5 ob/ob, n = 4 ob/ob + SB, n = 4 ob/ob + AM251 Kruskal−Wallis test and post hoc Dunn’s multiple comparison test, *p < 0.05, Kruskal−Wallis statistic = 8.69. The box plots elements are: center line, median (Q2); square symbol, mean; box limits, 25th (Q1)−75th (Q3) percentiles; whisker length is determined by the outermost data points. g−i NR2B downregulation in DG of obese mice. Confocal images of DCX/PSD95/NR2B immunolabeling of DG in lean (g), obese HFD (h), and ob/ob (i) mice showing a wide density of distribution of NR2B-ir (blue) at IML-MML DG of lean mice and colocalization with PSD95-ir (red) at postsynaptic side of DCX-ir dendritic spines. Scale bar: 10 µm. Immunolabeling was repeated n = 3 times independently in different DG sections from n = 6 mice per group, with similar results. j, k Measurement of NR2B subunit expression. Effect of leptin on NR2B subunit expression in lean, ob/ob, and HFD mice treated with or without leptin evaluated by western blot analysis. j Representative blot showing the band intensity of NR2B subunit in the aforementioned experimental mice. k Bar graphs showing the quantification of phospho-NR2B subunit normalized to α-tubulin. Data represent the means ± SEM from n = 3 mice/group. Data sets were compared by one-way ANOVA followed by Tukey’s test, *p < 0.05, ***p < 0.001, F = 10.63. Source data are provided as a Source Data file.
a Representative scheme and timeline of the novel object recognition test, animals were injected i.p. 1.5 h before with SB (30 mg/kg, i.p). b NOR in lean, HFD, and ob/ob mice calculated during habituation phase. c Short-term retention phase, NOR in lean, HFD, and ob/ob mice calculated after 1.5 h. d Long-term retention phase, NOR = 0.29 ± 0.06 in lean, NOR = 0.02 ± 0.05 in HFD, NOR = −0.04 ± 0.06 in ob/ob mice calculated after 24 h. **p < 0.01, ANOVA test and Bonferroni post hoc test, F = 8.6, n = 10 mice per group. The box plots elements are: center line, median (Q2); square symbol, mean; box limits, 25th (Q1)−75th (Q3) percentiles; whisker length is determined by the outermost data points. Source data are provided as a Source Data file.
References
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Medical
