Neuroscience

Astrocytic Nonsense-mediated mRNA decay regulates calcium signaling to support synapse function and restrain anxiety

Pablo J Lituma
Aykut Deveci
Estibaliz Barrio-Alonso
Kun Tan
Miles F Wilkinson
Pablo E Castillo
Dilek Colak author has email address

Center for Neurogenetics, Feil Family Brain and Mind Research Institute, Weill Cornell Medical College, Cornell University, New York, United States
Department of Obstetrics, Gynecology, and Reproductive Sciences, School of Medicine, University of California, San Diego, La Jolla, United States
Institute of Genomic Medicine, University of California, San Diego, La Jolla, United States
Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, United States
Department of Psychiatry and Behavioral Sciences, Albert Einstein College of Medicine, Bronx, United States
Gale & Ira Drukier Institute of Children’s Health, Weill Cornell Medical College, Cornell University, New York, United States

https://doi.org/10.7554/eLife.111940.1

Open access
Copyright information

Figures and data

Upf2 cKO astrocytes show cell size adaptations across brain areas.
(A) Schematic of adeno-associated viral delivery of AAV8-GFAP-mCherry-Cre and AAV5-GfaABC1D.PI.Lck-GFP.SV40 reporters to hippocampus (HPC), visual cortex (CTX), and prefrontal cortex (PFC). Image of mCherry and Lck-GFP expression in region of interest-ROI. Representative astrocyte morphology images from CTRL (top) and Upf2 cKO (bottom) across targeted brain areas. (B) Imaris 3D reconstructions of CTRL (top) and cKO (bottom) astrocytes from HPC, CTX, and PFC. (C) Upf2 mutant astrocytes display reduced cell volume as compared to CTRL in HPC, CTX, and PFC (HPC, CTRL: 961 ± 83, 76 cells; cKO: 669 ± 50, 90 cells, p < 0.001, Mann-Whitney test. CTX, CTRL: 1639 ± 123, 54 cells; cKO: 959 ± 60, 48 cells, p < 0.0001, Mann-Whitney test. PFC, CTRL: 1039 ± 91, 77 cells; cKO: 515 ± 35, 96 cells, p < 0.0001, Mann-Whitney test. 3 animals per genotype). (D) Upf2 mutant astrocytes exhibit decreased cell surface area across brain areas as compared to CTRL (HPC, CTRL: 448 ± 25, 76 cells; cKO: 352 ± 17, 90 cells, p = 0.002, Mann-Whitney test. CTX, CTRL: 652 ± 32, 54 cells; cKO: 460 ± 20, 48 cells, p < 0.0001, Mann-Whitney test. PFC, CTRL: 471 ± 26, 77 cells; cKO: 297 ± 13, 96 cells, p < 0.0001, Mann-Whitney test. 3 animals per genotype). Data is presented as mean ± S.E.M. ** denotes p < 0.001 **** denotes p < 0.0001. This figure highlights decreased cell volume and surface area in NMD-deficient astrocytes in hippocampus, cortex, and prefrontal cortex.

Upf2 cKO mice display reductions in both PSD-95 and spine density.
(A) Schematic of tamoxifen dosing and images of ZsGreen1 astrocytes in CTX, HPC, and PFC. (B) Imaris 3D rendering of colocalization signals depicting synaptic engulfment represented by PSD-95⁺/LAMP2⁺ puncta in astrocytic field of views (FOVs) across brain regions. (C) UPF2 deficient animals demonstrated reduced PSD-95 signal across brain areas as compared to CTRL (HPC, CTRL: 1409 ± 239, 10 FOVs; cKO: 342 ± 60, 13 FOVs, p = 0.0014, Unpaired t-test. CTX, CTRL: 963 ± 200, 10 FOVs; cKO: 323 ± 48, 12 FOVs, p = 0.0011, Unpaired t-test. PFC, CTRL: 1069 ± 129, 11 FOVs; cKO: 516 ± 83, 12 FOVs, p = 0.002, Unpaired t-test. 3 animals per genotype. (D) Upf2 mutant astrocytes display minimal synaptic engulfment across brain areas as compared to CTRL (HPC, CTRL: 8.3 ± 0.5, 76 cells, cKO: 3.5 ± 0.3, 90 cells, p < 0.0001, Mann-Whitney test. CTX, CTRL: 4.5 ± 0.4, 54 cells, cKO: 1.9 ± 0.3, 48 cells, p < 0.0001, Mann-Whitney test. PFC, CTRL: 4.0 ± 0.3, 76 cells, cKO: 2.6 ± 0.3, 97 cells, p < 0.0001, Mann-Whitney test. 3 animals per genotype). (E) Schematic of stereotaxic delivery of AAV5-CamKIIα-EGFP virus to label hippocampal CA1 pyramidal neurons (right image) followed by tamoxifen dosing of CTRL and cKO animals. Representative images of dendritic spines in CTRL (top) and cKO (bottom). Scale bar, 10 μm. Astrocytic Upf2 deficiency is associated with reduced spine density of CA1 pyramidal neurons in comparison to CTRL (CTRL: 6.2 ± 0.3, cKO: 3.6 ± 0.2, n = 3 mice per genotype, p < 0.0001, Mann-Whiteny test). Data is presented as mean ± S.E.M. * denotes p < 0.05, ** denotes p < 0.01, **** denotes p < 0.0001. This figure demonstrates reduced PSD-95 levels associated with decreased spine density of CA1 pyramidal neurons in Upf2 cKO brains.

Astrocytic Upf2 absence impairs basal neurotransmission and synaptic plasticity.
(A) Schematic of tamoxifen dosing and electrophysiology in hippocampal area CA1 of acute brain slices. (B) Input-output curve assessment demonstrated reduced synaptic strength in Upf2 cKO mice as compared to CTRL (CTRL: Linear fit 2.04 ± 0.3, r = 0.99; cKO: Linear fit 0.96 ± 0.02, r = 0.99, n = 8 slices, 4 animals per genotype, linear fit slope values: p = 0.011, Unpaired t-test). (C) Presynaptic function assessed by the paired-pulse ratio showed no differences across genotypes (CTRL, 200 ms: 1.3 ± 0.07; cKO, 200 ms: 1.35 ± 0.04, n = 8 slices, 4 animals per genotype, p = 0.4, Unpaired t-test). (D) LTP induced by theta-burst stimulation led to absent potentiation in the cKO condition (CTRL: 138 ± 15%, n = 6 mice, 12 slices; cKO: 78.8 ± 9%, n = 5 mice, 9 slices; p = 0.003, Unpaired t-test). (E) LTD induced by low-frequency stimulation was similar in both genotypes (CTRL: 72.9 ± 5%; cKO: 78.5 ± 4%, n = 5 slices, 3 animals per genotype, p = 0.41, Mann-Whitney test). (F) Chemical-LTD induced by DHPG (50 μM, 5 min) revealed dampened synaptic depression in Upf2 cKO brain slices (CTRL: 55.9 ± 6%, n = 4 mice, 6 slices; cKO: 84.7 ± 9%, n = 5 mice, 7 slices, p = 0.02, Unpaired t-test). Data are presented as mean ± S.E.M. * denotes p < 0.05, *** denotes p < 0.005, ns = not significant. This figure illustrates that Upf2 deficiency in astrocytes alters synaptic function.

Upf2 cKO mice display anxiety-like behavior.
(A) Open field beam breaks display no differences in locomotion amongst CTRL and cKO animals (CTRL: 3892 ± 271; cKO: 4076 ± 239, n = 22 mice per group, p = 0.61, Unpaired t-test). (B) Upf2 cKO mice show reduced time in open arms of elevated plus maze indicating anxiety-like behavior (CTRL: 38.5 ± 3.9; cKO: 23.09 ± 2.6, n = 31 mice per group, p = 0.0036, Unpaired t-test). (C) Astrocytic Upf2 loss is associated with reduced marble burying (CTRL: 16.3 ± 0.6, n = 22 mice; cKO: 13.0 ± 0.8, n = 23 mice, p = 0.0003, Mann-Whitney test). (D) Contextual fear conditioning paradigm revealed differences in contextual memory on Day 2 between CTRL and cKO animals (CTRL: 68.3 ± 3, n = 29 mice; cKO: 77.6 ± 3, n = 24 mice, p = 0.02, Mann-Whitney test). No differences were detected on Day 1 and Day 3 (Day 1: Baseline vs. UC-CS pairings, CTRL vs. cKO, p > 0.59, Sidak’s multiple comparisons test; Day 3: CTRL vs. cKO, p = 0.69, Unpaired t-test). Assessment of baseline freezing revealed Upf2 cKO mice show significant elevation on Day 2 and Day 3 as compared to CTRL (Day 1: CTRL: 1.8 ± 0.6; cKO: 6.1 ± 2.1, p = 0.43, Mann-Whitney test. Day 2: CTRL: 12.4 ± 0.8; cKO: 14.6 ± 0.7, p = 0.048, Unpaired t-test. Day 3: CTRL: 21.2 ± 3.6; cKO: 34.9 ± 4.5, p = 0.03, Mann-Whitney test. CTRL: n = 29 mice, cKO: n = 24 mice). Data are presented as mean ± S.E.M. * denotes p < 0.05, ** denotes p < 0.01, *** denotes p < 0.005, ns = not significant. This figure illustrates that astrocytic UPF2 depletion is associated with avoidance behavior and increased baseline freezing.

Identifying the physiological astrocytic targets of NMD in vivo.
(A) Upf2 wt/wt or fl/fl mice were crossed with the Aldh1L1Cre^ERT2 and Ai6-ZsGreen1 transgenic lines to select CTRL and cKO astrocytes with 488 nm fluorescence after tamoxifen dosing using FACS. Representative image demonstrating ZsGreen1 expression in cortical and hippocampal astrocytes (bottom, left). Side scatter (SSC) and forward scatter (FSC) plot for 488 nm fluorescence selection of ZsGreen1⁺ astrocytes and tamoxifen negative (-TMX) cell suspension showing no 488 nm positivity (bottom, right). (B) Principal component analysis (PCA) of RNA- seq datasets from CTRL and cKO astrocytes, as defined in panel (A). (C) Differentially expressed genes (DEGs; q <0.05, fold change >2) identified from the RNA-seq analysis of CTRL and cKO astrocytes in panel (B). (D) Upregulated DEGs containing at least one NMD inducing feature: downstream Exon-Junction complex-dEJ, upstream Open Reading Frame-uORF, and long 3’-untranslated region-UTR length (>1500 nt). (E) Top biological functions associated with upregulated genes defined in panel (C). Statistical significance [−Log₁₀ (p-value)] is indicated by the bar (bottom). The number of DEGs for a given category is indicated by an “X” (top). (F) Ingenuity Pathway Analysis (IPA) of upregulated genes defined in panel (C). Statistical significance [−Log₁₀ (p-value)] is indicated by the bar (top), with a p-value <0.05 as the cut off. This figure, for the first-time, reveals the physiological astrocytic transcripts regulated by NMD in vivo and their associated biological pathways.

NMD-deficient astrocytes exhibit altered Ca²⁺ activity.
(A) Schematic of adeno-associated viral delivery of AAV8-GFAP-mCherry-Cre/AAV5-GfaABC1D-GCaMP6f reporters to hippocampus, visual cortex, and prefrontal cortex. Representative image of mCherry/GCaMP6f expression in region of interest-ROI. In all panels of this figure: heat map of spontaneous Ca²⁺ activity in CTRL and cKO conditions (left), temporal Ca²⁺ dynamic traces (middle), and peak Ca²⁺ summary data (right). (B) Upf2 mutant astrocytes in hippocampus show elevated peak Ca²⁺ transients of spontaneous activity in comparison to CTRL (CTRL: 3.8 ± 0.6, 34 ROIs, n = 6 mice; cKO: 6.7 ± 1.1, 41 ROIs, n = 8 mice, p = 0.038, Mann-Whitney test). (C) Upf2 deficient astrocytes in visual cortex exhibit enhanced peak Ca²⁺ transients of spontaneous activity as compared to CTRL (CTRL: 3.5 ± 1.3, 14 ROIs, n = 4 mice; cKO: 6.9 ± 2.2, 34 ROIs, n = 5 mice, p = 0.046, Mann-Whitney test). (D) Upf2 cKO astrocytes in prefrontal cortex display increased peak Ca²⁺ transients of spontaneous activity as compared to CTRL (CTRL: 1.1 ± 0.1, 22 ROIs, n = 4 mice; cKO: 3.0 ± 0.7, 36 ROIs, n = 5 mice, p = 0.013, Mann-Whitney test). Data are presented as mean ± S.E.M. * denotes p < 0.05. This figure highlights that UPF2-NMD deficient astrocytes exhibit elevated Ca²⁺ transients in hippocampus, visual cortex, and prefrontal cortex.

CalEx restores basal Ca²⁺ activity in Upf2 cKO astrocytes and rescues neurotransmission along with anxiety.
(A) Scheme for viral delivery and timeline for Ca²⁺ measurement experiments. (B) CalEx diminished Ca²⁺ events to CTRL levels in hippocampus (CTRL: 4.5 ± 1.5, n = 6 mice, 32 ROIs; cKO: 5.1 ± 0.9, n = 8 mice, 68 ROIs; cKO + CalEx: 1.4 ± 0.1, n = 4 mice, 46 ROIs; One-way ANOVA DF = 2, F = 4.802, p = 0.0096, post-hoc Tukey’s multiple comparisons test CTRL vs. cKO + CalEx, p = 0.096). CalEx Ca²⁺ events are comparable to CTRL in visual cortex (CTRL: 4.9 ± 1.7, n = 4 mice, 16 ROIs; cKO: 6.6 ± 2.1, n = 6 mice, 36 ROIs; cKO + CalEx: 2.3 ± 0.4, n = 4 mice, 19 ROIs; One-way ANOVA DF = 2, F = 1.261, p = 0.29). CalEx prefrontal cortex Ca²⁺ events are similar to CTRL (CTRL: 0.8 ± 0.2, n = 4 mice, 15 ROIs; cKO: 2.3 ± 0.4, n = 4 mice, 20 ROIs; One-way ANOVA DF = 2, F = 5.165, p = 0.0093, post-hoc Tukey’s multiple comparisons test CTRL vs. cKO + CalEx, p = 0.542). (C) Scheme for viral delivery and tamoxifen treatment followed by rescue assays for PSD-95 density, neurotransmission, and animal behavior. (D) Total PSD-95 density is not restored to CTRL levels by CalEx manipulation in Upf2 cKO mice (CTRL: 1648 ± 188, n = 24 ROIs, 4 mice; cKO: 459 ± 41, n = 21 ROIs, 4 mice; cKO + CalEx: 771 ± 52, n = 24 ROIs, 4 mice; One-way ANOVA DF = 2, F = 10.28, p < 0.0001, post-hoc Tukey’s multiple comparisons test CTRL vs. cKO + CalEx, p < 0.0001). (E) CalEx elevates synaptic strength in hippocampus of Upf2 cKO mice as compared to CTRL (CTRL: Linear fit 1.88 ± 0.1, r = 0.99, n = 6 slices, 3 mice; cKO: Linear fit 0.84 ± 0.05, r = 0.99, n = 5 slices, 4 mice; cKO + CalEx: Linear fit 2.66 ± 0.24, r = 0.98, n = 5 slices, 3 mice. One-way ANOVA DF = 2, F = 15.08, p = 0.0004, post-hoc Tukey’s multiple comparisons test CTRL vs. cKO + CalEx, p = 0.033). (F) CalEx increases elevated plus maze open arm time in Upf2 cKO mice (CTRL: 70.7 ± 5.7, n = 11 mice; cKO: 25.9 ± 5.2, n = 8 mice; cKO + CalEx: 49.1 ± 8.1, n = 10 mice; CTRL + CalEx: 40.6 ± 6.4, n = 8 mice, One-way ANOVA DF = 3, F = 8.24, p = 0.0003, post-hoc Tukey’s multiple comparisons test CTRL vs. cKO + CalEx, p = 0.086). Data are presented as mean ± S.E.M. * denotes p < 0.05, ** denotes p < 0.01, **** denotes p < 0.0001, ns = not significant. This figure depicts that restoring Ca²⁺ levels in astrocytes of Upf2 cKO mice recovers basal neurotransmission and minimizes anxiety-like behavior.

Sign up for email alerts