Glia selectively approach synapses on thin dendritic spines

Abstract

This paper examines the relationship between the morphological modality of 189 dendritic spines and the surrounding astroglia using full three-dimensional reconstructions of neuropil fragments. An integrative measure of three-dimensional glial coverage confirms that thin spine postsynaptic densities are more tightly surrounded by glia. This distinction suggests that diffusion-dependent synapse-glia communication near 'learning' synapses (associated with thin spines) could be stronger than that near 'memory' synapses (associated with larger spines).

Keywords: glia protection; synapses; thin spines.

Figure 1.

Morphology of live astrocytes in dentate gyrus. (a) Characteristic morphologies of individual astrocytes held in whole-cell mode and filled with Alexa Fluor 594 (λ_x = 800 nm); each fluorescence image represents an averaged Z-stack of 70–110 consecutive X–Y two-photon excitation sections taken with 0.5 µm Z-axis steps; patch pipettes are seen. Image on far right, an astrocyte representing a sub-group with prominent dendritic trunks [24]. (b) Input resistance measurement in a typical ‘passive’ astrocyte; whole-cell current clamp mode, a voltage response to a 200 pA current injection pulse is shown (upper trace; also see text and electronic supplementary material, figure S1). (c) In each astrocyte, distribution of the tissue volume fraction occupied by thin, indicator-filled astrocyte processes (excluding areas containing large dendritic trunks) is measured in a Z-series of thin X–Y optical sections provided by two-photon excitation. Left panel, an image stack average (as in (a)). Right inset panels, examples of individual X–Y sections (section number reference is indicated; the effective optical width, approx. 1 µm) containing the soma; dotted line, a sampling segment (example) for the fluorescence brightness profile. Such samples were taken systematically in each X–Y section, normally by rotating the sampling segment in approximately 20° increments around the soma, throughout the Z-stack containing the soma. Plot, an example of measurements in eight different X–Y sections of the same astrocyte (shown on the left): brightness profiles are normalized with respect to the brightness level inside the soma (see text) within each X–Y section; grey, individual profiles; black, average. Gap junctions are blocked with carbenoxolone (Material and methods).

Figure 2.

Three-dimensional reconstruction of contiguous astrocyte fragments with adjacent synapses. (a–d) A fragment of a dentate astrocyte reconstructed in three dimensions (blue, (a,b)) together with the adjacent dendritic spines (grey and dark yellow structures, (b)) equipped with PSDs (red). The spines are unambiguously separated into the sub-groups of thin (dark yellow, shown separately in (c)) and mushroom (grey, (d)). See the electronic supplementary material, figure S2, for examples of the original serial sections with identified structures.

Figure 3.

Astrocyte membranes are much closer to the PSDs occurring on thin, compared to mushroom, dendritic spines. (a) A diagram illustrating automatic measurement of the nearest edge-to-edge distances D_min between all individual PSDs (red mesh) and astrocyte membranes (white mesh) in space, at two levels of detail. Blue segments show D_min determined in 3D automatically; smooth surface rendering (as in figure 2) is omitted for clarity. (b) Distribution (cumulative probability plot) of D_min for individual PSDs occurring on thin (red, n = 136) and mushroom (black, n = 53) dendritic spines. (c) A two-dimensional diagram illustrating a rationale behind the weighting of PSD–glia distances. Distances between PSD (red, either the centroid only or all surface points) and all points on the surface of a neighbouring glial process (green) are measured, with larger distances bearing less weight (indicated by arrow thickness and shade of grey). The weighting reflects a rapid drop of transmitter concentration with distance from the site of release (yellow dot). See Results for details. (d) Grey and white columns, average (±s.e.m.) weighted PSD–astroglia distances D_w for thin and mushroom dendritic spines, respectively, measured with respect to either PSD centroids or all PSD surface points, as indicated (***p < 0.001).