Astroglial β-Arrestin1-mediated Nuclear Signaling Regulates the Expansion of Neural Precursor Cells in Adult Hippocampus

Abstract

Adult hippocampal neurogenesis is crucial for preserving normal brain function, but how it is regulated by niche cells is uncertain. Here we show that β-arrestin 1 (β-arr1) in dentate gyrus (DG) regulates neural precursor proliferation. β-arr1 knockout (KO) mice show reduced neural precursor proliferation in subgranular zone (SGZ) which could be rescued by selective viral expression of β-arr1 but not its nuclear-function-deficient mutants under control of hGFAP promotor in DG. Compared with wild type astrocytes, β-arr1 KO astrocytes nurture less neurospheres, and this may be attributed to changed activity of soluble, heat-sensitive excretive factors, such as BMP2. RNA-sequencing reveals that β-arr1 KO DG astrocytes exhibit an aberrant gene expression profile of niche factors, including elevated transcription of Bmp2. Taken together, our data suggest that β-arr1 mediated nuclear signaling regulates the production of excretive factors derived from niche astrocytes and expansion of neural precursors in DG, thus maintaining homeostasis of adult hippocampal neurogenesis.

Figure 1. Decreased Proliferation of Neural Precursors in the SGZ of Adult β-arr1 KO mice.

β-arr1 KO mice (KO) and wild type littermates (WT) of 2–3 month old were used. (A–I) Mice were injected intraperitoneally with 100 mg/kg BrdU (blue arrows) and sacrificed (black arrows) at time points indicated (A,D,G). Sample projected confocal images were shown (B,E,H) and stereological quantification (C,F,I) of BrdU-positive and Ki67-positive (C) cells in SGZ of these mice were performed. Arrowheads indicate the localization of BrdU⁺ cells. Orthogonal views are samples of BrdU⁺ Ki67⁺ cells. The cell number was normalized to the GCL (granular cell layer) volume (in mm³). Data represent mean ± s.e.m.; n = 3 or 4 mice for each genotype; t- test, *p < 0.05. Scale bar, 50 μm; (J–L) After running wheel training (RW) or housed in home cage (HC) for 15 days, mice were injected with BrdU for 3 times (blue arrows) with an 12 h interval and sacrificed (black arrow) 2 h after the last BrdU injection (J). Immunostaining (K) and stereological quantification (L) of BrdU⁺ cells in SGZ were performed. The cell number was normalized to the GCL volume (in mm³). Data represent mean ± s.e.m.; n = 3 mice/group; *p < 0.05 vs. WT/HC, two-way ANOVA, post hoc Tukey test; Scale bar, 50 μm.

Figure 2. Expression of β-arr1 in Neural Precursors and Niche Cells of SGZ.

(A,B) Representative images (A) and data of optical densities (OD) (B) of in situ hybridization using β-arr1 antisense probe on brain sections from P0, P7, P15, P30, or P60 WT mice. Scale bar, 50 μm. (C) In situ hybridization with β-arr1 antisense probe combined with immunostaining for BrdU, GFAP, NeuN, and Sox2 in the DG on brain sections of 3-month-old WT mice. Orthogonal views are shown to confirm colocalization. Arrowheads indicate β-arr1⁺ NeuN⁻ cells. (D) Co-immunostaining for β-arr1 and Nestin. Orthogonal views are shown to confirm colocalization. Arrows indicate the cell bodies of Nestin⁺ RGLs and arrowheads indicate their branches. Scale bar, 50 μm.

Figure 3. β-arr1 in the DG Regulates the Proliferation of Neural Precursors.

One week after injection of lentivirus encoding GFP (LV-GFP), β-arr1 and GFP (LV-β-arr1), or GFP plus β-arr1 shRNA (LV-β-arr1 shRNA) or β-galactosidase shRNA (LV-LacZ shRNA) into the DG, mice were treated daily with BrdU (100 mg/kg i.p.) for 7 days and sacrificed 1 day after the last BrdU injection. (A) Schematic of experimental schedule. (B–D) Expression of LV-β-arr1 and LV-GFP. *p < 0.05 vs. WT/LV-GFP; #p < 0.05 vs. KO/LV-GFP, two-way ANOVA, post hoc Tukey test. (E–G) Expression of LV-β-arr1 shRNA and LV-LacZ shRNA. *p < 0.05, t-test, BrdU⁺ cell and BrdU⁺ GFP⁺ cell number in the SGZ are normalized to the GCL volume (in mm³). Values represent mean ± s.e.m.; n = 3–4 for each group; Scale bar, 100 μm.

Figure 4. Ablation of β-arr1 Reduces NSC Activation and Neuronal Production in Adult DG.

(A,B) Sample projected confocal images and stereological quantification of Nestin⁺ RGLs in DG of P21, 3 and 6 month-old mice. The number of RGLs was normalized to the GCL volume; two-way ANOVA, post hoc Tukey test. (C,D) Sample projected confocal images and stereological quantification of Nestin⁺ GFAP⁺ RGLs and Nestin⁺ GFAP⁺ MCM2⁺ RGLs in the DG of WT and KO mice. Orthogonal images ((C), bottom) are shown to confirm colocalization. (E,F) Confocal images and quantification of Tbr2⁺ IPCs and Tbr2⁺ MCM2⁺ IPCs in SGZ of WT and KO mice. (G) Schematic of experimental schedule. WT and β-arr1 KO mice were injected with BrdU (i.p., 100 mg/kg) for 7 days, and sacrificed 30 days after the last BrdU injection. (H,I) Sample projected confocal images and stereological quantification of BrdU⁺ NeuN⁺ cells in DG. High magnification orthogonal images show the colocalization of BrdU and NeuN. Data are mean ± s.e.m.; n = 5–6 per genotype; *p < 0.05 versus WT, t-test; Scale bar, 50 μm.

Figure 5. Nuclear β-arr1 Regulates the Proliferation of Adult Neural Precursors in DG.

(A,B) Tissue-conditioned media derived from DG of β-arr1 KO mice sustained less WT neurospheres than the complete media and the tissue-conditioned media derived from DG of WT mice. Data are normalized to basal value, n = 3 independent experiments. One-way ANOVA; Scale bar, 100 μm. (C,D) WT neurospheres co-cultured with β-arr1 KO DG astrocytes (KO-Astrocytes) were less than those co-cultured with WT (WT-Astrocytes). Data are normalized to WT, all error bars show s.e.m. of triplicated cultures (3–4 samples per group), t-test; Scale bar, 100 μm. (E–G) The schematic of experimental schedule (E), sample projected confocal images (F), and stereological quantification (G) of BrdU⁺ proliferating cells (arrowheads) in the SGZ of WT mice injected with AAV-hGFAP-mCherry (WT-mCherry) or β-arr1 KO mice injected with AAV-hGFAP-mCherry (KO-mCherry), AAV-hGFAP-β-arr1-mCherry (KO-β-arr1), AAV-hGFAP-β-arr1Q394L-mCherry (KO-Q394L) or AAV-hGFAP-β-arr1K157A-mCherry (KO-K157A). n = 3–6 for each group; Data represent mean ± s.e.m.; *p < 0.05 vs. WT-mCherry; one-way ANOVA, post hoc Tukey test; Scale bar, 50 μm.

Figure 6. RNA-Sequencing Transcriptome Shows Abnormal Gene Profiles in β-arr1 KO Niche Astrocytes.

(A–D) Changes of mRNA levels of different genes in WT and KO niche astrocytes determined by RNA-seq; (E) Comparison of mRNA levels of neural factors in WT and KO niche astrocytes determined by RNA-seq and qRT-PCR; Samples were collected from three mice in each group; # indicates ratio of β-arr1 KO vs. WT >3/2 or <2/3. (F) Protein levels of SHH, BMP2, and WNT5a in KO and WT astrocyte-conditioned media determined by ELISA. n = 3 independent experiments. Data were normalized to WT and represented mean ± s.e.m.; *p < 0.05 vs. WT; t-test.