Astrocyte extracellular matrix modulates neuronal dendritic development

Abstract

Major developmental events occurring in the hippocampus during the third trimester of human gestation and neonatally in altricial rodents include rapid and synchronized dendritic arborization and astrocyte proliferation and maturation. We tested the hypothesis that signals sent by developing astrocytes to developing neurons modulate dendritic development in vivo. We altered neuronal development by neonatal (third trimester-equivalent) ethanol exposure in mice; this treatment increased dendritic arborization in hippocampal pyramidal neurons. We next assessed concurrent changes in the mouse astrocyte translatome by translating ribosomal affinity purification (TRAP)-seq. We followed up on ethanol-inhibition of astrocyte Chpf2 and Chsy1 gene translation because these genes encode for biosynthetic enzymes of chondroitin sulfate glycosaminoglycan (CS-GAG) chains (extracellular matrix components that inhibit neuronal development and plasticity) and have not been explored before for their roles in dendritic arborization. We report that Chpf2 and Chsy1 are enriched in astrocytes and their translation is inhibited by ethanol, which also reduces the levels of CS-GAGs measured by Liquid Chromatography/Mass Spectrometry. Finally, astrocyte-conditioned medium derived from Chfp2-silenced astrocytes increased neurite branching of hippocampal neurons in vitro. These results demonstrate that CS-GAG biosynthetic enzymes in astrocytes regulates dendritic arborization in developing neurons.

Keywords: Astrocyte translatome; Fetal Alcohol Spectrum Disorders (FASD); chondroitin sulfate glycosaminoglycans (CS-GAGs); dendrite development; ethanol; hippocampus.

Figure 1.

Morphometric analysis of CA1 pyramidal neuron apical dendrites from PD7 mice neonatally exposed to ethanol vapor or air control. A: Representative Neurolucida tracings of PD7 CA1 pyramidal neurons (scale bar 50 µm). B: Ethanol exposure increased complexity (p = 0.0004). C: Maximal terminal distance was not altered by ethanol exposure. The sum of terminal orders (D; p = 5.0×10⁻⁵), the length of the apical dendrite (E; p = 0.002), the number of ends (F; p = 0.0006), and the number of nodes (G; p = 0.0007) of the apical dendrite were increased by ethanol exposure.

Figure 2.

Morphometric analysis of CA1 pyramidal neuron basal dendrites from PD7 mice neonatally exposed to ethanol vapor or air control. Ethanol exposure increased basal dendrite complexity (A; p = 8.4 × 10⁻⁶), the maximal terminal distance (B; p = 0.0006), the sum of terminal orders (C; p = 4.2 × 10⁻⁵), the total dendrite length (D; p = 0.0006), the number of ends (E; p = 6.7 × 10⁻⁵); the number of nodes (F; p = 0.0003) and the average dendrite length (G; p = 5.9 × 10⁻⁵). H: The number of basal dendrites was not altered by ethanol exposure.

Figure 3.

Analysis of spine density and dendrite diameter in secondary branches of apical dendrites following neonatal ethanol exposure. A: The number of spines per length of the dendrite (µm) was not altered by ethanol. B: The number of spines per length (µm) per diameter (µm) of the dendrite was increased by neonatal ethanol exposure (p = 0.003). C: Ethanol exposure decreased the diameter of the secondary branches of the apical dendrite from which spine density was determined (p = 9.5 × 10⁻⁵). D: Representative images of Golgi-Cox-stained dendrites analyzed for spine density and dendrite diameter at 100x magnification.

Figure 4.

A: qRT-PCR of cell-type marker genes in the TRAP fractions vs input fractions shows high enrichment of astrocyte markers and depletion of neuronal, microglial, and oligodendrocyte markers. B: Schematic representation of the design of astrocyte TRAP-seq experiments. C: Venn diagrams comparing ethanol regulated (left), ethanol upregulated (center), and ethanol downregulated (right) genes in TRAP and input samples and their overlap.

Figure 5.

Category overrepresentation summary plots of ethanol regulated genes in the TRAP (astrocyte) fraction of genes that were not enriched in astrocytes. A: Ethanol regulated genes not enriched in astrocytes analyzed for Gene Ontology (GO), and pathway overrepresentation using EnrichR. The top 30 gene categories based on p-values are shown. In each plot, the size of the dot represents the number of genes that were significantly regulated, and the color scale represents the p-value, with lighter colors indicating lower p-values. Categories marked with green arrows are related to gene translation. B: Top overrepresented categories derived from GO analysis of genes upregulated by ethanol and not enriched in astrocytes. Gene categories with tan arrows relate to synaptic activity and voltage-gated channels and red arrows relate to dendrite growth. C: Top overrepresented categories derived from GO analysis of genes downregulated by ethanol and not enriched in astrocytes. Gene categories marked with green arrows relate to gene translation.

Figure 6.

Category overrepresentation summary plots of ethanol regulated genes in the TRAP (astrocyte) fraction of genes that were enriched in astrocytes. A. Top overrepresented categories derived from GO analysis of ethanol-regulated genes enriched in astrocytes. Categories marked with blue arrows relate to steroid biosynthesis and with magenta arrows to cell-cell interactions, cell-ECM interactions, and cell adhesion. B: Top overrepresented categories derived from GO analysis of genes upregulated by ethanol and enriched in astrocytes analyzed with EnrichR. Categories with blue arrows relate to steroid biosynthesis and magenta arrows relate to cell-cell interactions, cell-ECM interactions, and cell adhesion. C: Top overrepresented categories derived from GO analysis of genes downregulated by ethanol and enriched in astrocytes analyzed with EnrichR. Gene categories with purple arrows are related to chondroitin sulfate biosynthesis.

Figure 7.

Neonatal ethanol exposure decreases chondroitin sulfate glycosaminoglycan (CS-GAG) biosynthetic enzymes and CS-GAGs levels in the neonatal hippocampus. A: Down-regulation of Chpf2 (left) and Chsy1 (right) in the TRAP fraction by qRT-PCR (Chpf2 p-value = 0.027; Chsy1 p-value = 0.022). B: Chpf2 (top panels) and Chsy1 (bottom panels) expression in astrocytes of the CA1 region of the hippocampus by RNA-FISH. Left panels: 100x objective images of RNA-FISH probes for Chpf2 (green, top) and Chsy1 (green, bottom) and cell-type markers Aldh1l1 (red, astrocytes) and Tubb3 (white, neurons); nuclei are stained by DAPI (Blue; scale bar = 20 µm). Right panels: zoomed region inside the white box of the left panels (scale bar = 10 µm). Staining puncta show proximal presence of Aldh1l1 and Chpf2 (top panels) or Aldh1l1 and Chsy1 (bottom panels) staining. C: Astrocyte enrichment and ethanol regulation of Chpf2 and Chsy1 in TRAP-Seq data. Concentrations (ng/mg tissue) of total CS-GAG disaccharides (D; p = 0.001), CS-4S disaccharides (E; p = 0.002), CS-6S disaccharides (F: p = 0.005), and CS-0S disaccharides (G; p = 0.01) were decreased following ethanol treatment in the neonatal rat hippocampus. CS-GAG disaccharides were quantified by LC/MS and normalized to hippocampus weight.

Figure 8.

ACM from Chpf2 siRNA-treated astrocytes increases neurite outgrowth in hippocampal neurons in vitro. Primary astrocyte cultures were treated with Chpf2 SiRNA or control; ACM was collected and transferred to primary cortical neuron cultures for 48 h. Hippocampal pyramidal neurons incubated in the presence of ACM from astrocytes in which Chpf2 was silenced showed increased neuronal neurite complexity (A; p = 0.03); number of nodes (B; p = 0.05), number of ends (C; p = 0.01), and the length of the minor neurites (D; p = 0.03). E: representative neurons exposed to control ACM (left) or ACM prepared from astrocytes whose Chpf2 expression was silenced.