Astroglial Connexin 43 Regulates Synaptic Vesicle Release at Hippocampal Synapses

Abstract

Connexin 43, an astroglial gap junction protein, is enriched in perisynaptic astroglial processes and plays major roles in synaptic transmission. We have previously found that astroglial Cx43 controls synaptic glutamate levels and allows for activity-dependent glutamine release to sustain physiological synaptic transmissions and cognitiogns. However, whether Cx43 is important for the release of synaptic vesicles, which is a critical component of synaptic efficacy, remains unanswered. Here, using transgenic mice with a glial conditional knockout of Cx43 (Cx43-/-), we investigate whether and how astrocytes regulate the release of synaptic vesicles from hippocampal synapses. We report that CA1 pyramidal neurons and their synapses develop normally in the absence of astroglial Cx43. However, a significant impairment in synaptic vesicle distribution and release dynamics were observed. In particular, the FM1-43 assays performed using two-photon live imaging and combined with multi-electrode array stimulation in acute hippocampal slices, revealed a slower rate of synaptic vesicle release in Cx43-/- mice. Furthermore, paired-pulse recordings showed that synaptic vesicle release probability was also reduced and is dependent on glutamine supply via Cx43 hemichannel (HC). Taken together, we have uncovered a role for Cx43 in regulating presynaptic functions by controlling the rate and probability of synaptic vesicle release. Our findings further highlight the significance of astroglial Cx43 in synaptic transmission and efficacy.

Keywords: astrocytes; connexin 43; hippocampus; release probability; synaptic transmission; synaptic vesicle.

Figure 1

Astroglial Cx43 does not alter neuronal morphology. (A) Representative images of a CA1 pyramidal neuron dialyzed with sulforhodamine B via a patch pipette in a wild-type (+/+, left) and a Cx43−/− (−/−, right) hippocampus. The maximum z-projection of the confocal images are shown on the top, and reconstruction of the same neurons are shown at the bottom. (B) The plots show the average of the total length of the basal, apical, and all dendrites. (C) The fraction of the total dendrites at discrete angles around the cell soma in polar histograms, and (D) the fraction of total dendrites intersecting in concentric circles away from the cell soma as shown by Sholl analysis. These analyses show no significant differences in the dendritic organization between pyramidal neurons in +/+ and −/− mice. Mean ± SEM; n = 20 +/+ and 19 −/− cells; and scale bar: 100 µm in (A).

Figure 2

Astroglial Cx43 regulates spine density, but not their properties. (A) Representative images of a dendritic segment of a CA1 pyramidal neuron dialyzed with sulforhodamine B via a patch pipette in a wild-type (+/+, top) and a Cx43−/− (−/−, bottom) hippocampus. The maximum z-projection of the confocal images containing dendritic spines are shown. (B–E) The plots show no significant difference between the averages and cumulative histograms of the spine length (B,C), nor in the spine head diameter (D,E) between neurons in +/+ and −/− hippocampi (F) The relative proportion of the “Thin/Filopodia” or “Stubby/Mushroom” were not different. (G,H) Spine density was significantly higher in −/− overall (G) and also in both “Thin/Filopodia” and “Stubby/Mushroom” spine categories (H). Mean ± SEM; n = 7 +/+ and 6 −/− cells; * p < 0.05, ** p < 0.01; scale bar: 2 µm in (A); and open or filled red arrows denote representative thin/filopodia or stubby/mushroom spines, respectively.

Figure 3

The expression of the key synaptic markers in the hippocampal synaptosomal fractions does not depend on astroglial Cx43 level. (A) Representative Western blots show a comparable protein expression of the synaptic markers in synaptosomal fractions, which were isolated from a wild-type (+/+) and a Cx43−/− (−/−) hippocampus. (B) The quantification of protein expression normalized to +/+ level for each synaptic marker. Mean ± SEM; n = 5 (+/+) and 4 (−/−).

Figure 4

Astroglial Cx43 deficiency disrupts the distribution of synaptic vesicles in presynaptic terminals. (A) Representative electron micrographs of hippocampal presynaptic terminals (boutons) in wild-type (+/+, left) and Cx43−/− (−/−, right) mice. Yellow circles outline the synaptic vesicles; blue lines mark active zones; and red arrowheads denote postsynaptic densities (PSD). (B–D) The plots show no significant difference between the +/+ and −/− mice in terms of the bouton volume (B), PSD surface area (C), or in the number of synaptic vesicles per presynaptic terminal (D). (E,F) However, the distance of synaptic vesicles from the active zone (AZ) was found to be significantly greater in the −/− mice when compared to the +/+ mice, as can be seen when this is plotted in averages (E) and cumulative histograms (F). The distance was determined as the minimal distance between the center of each synaptic vesicles to the corresponding presynaptic membrane opposing the PSD. Mean ± SEM; n = 60 (+/+) and 62 (−/−) terminals in (B,C), as well as the 33 (+/+) and 45 (−/−) terminals in (D–F). From 3 +/+ and 3 −/− mice; * p < 0.05, *** p < 0.001; and scale bar: 0.15 µm in (A).

Figure 5

Astroglial Cx43 contributes to synaptic vesicle recycling in hippocampal neurons. (A) Representative image of the CA1 stratum radiatum region of a hippocampal slice that is loaded with FM1-43 in presynaptic terminals, as is shown in the fluorescently labeled punctate structures. (B) The dye-loaded regions are shown at a higher magnification both before (t = 0) and after (t = 240 s) the dye unloading induced by the 10 Hz stimulation for 2 min. Blue squares mark eight individual synaptic puncta, thereby showing an unloading of FM1-43 dye upon stimulation. (C) Time series images of dye unloading are shown for each synaptic puncta marked in (B). (D) Normalized fluorescence intensity is plotted over time showing the unloading of FM1-43 dye upon stimulation, which is significantly slower in hippocampal slices from Cx43−/− (−/−) mice when compared to wild-type (+/+) (E–G) The plots showing the average of half time (t_1/2, E) and decay time constant (τ, F) of dye unloading, which was found to be significantly higher in −/− when compared to +/+, while, the relative total dye unloaded (G) was not different. Mean ± SEM; n = 3 (+/+) and 4 (−/−) independent experiments; * p < 0.05; *** p < 0.001; and scale bar: 50 µm in (A) and 20 µm in (B).

Figure 6

Astroglial Cx43 regulates the release probability in hippocampal neurons. (A) Representative recordings showing fEPSPs upon the Schaffer collateral paired-pulse stimulations (10–20 µA, 0.1 ms, 40 ms apart). (B) The quantification of the fEPSP amplitude of the second stimulation relative to the first stimulation shows an increase in the paired-pulse ratio in the hippocampal slices from mice that are deficient in astroglial Cx43. These were then rescued by an exogenous application of glutamine (4 mM, −/− Gln, 1–4 h). Glutamine alone did not alter the paired pulse ratio (+/+ Gln). (C) This effect was mimicked by Gap26 in the +/+ slices (+/+ Gap26), but not by a scrambled version of Gap26 (+/+ Gap26Src) and rescued by exogenous glutamine (+/+ Gap26 Gln). Mean ± SEM; n = 19 (+/+), 16 (−/−), 9 (−/− Gln), 8 (+/+ Gln), 17 (+/+ Gap26), 5 (+/+ Gap26 Gln), and 12 (+/+ Gap26Src) independent experiments; * p < 0.05, ** p < 0.01, *** p < 0.001; and scale bars: 0.2 mM, 20 ms in (A).