Neuron-to-glia signaling drives critical period experience-dependent synapse pruning

Abstract

Critical periods enable early-life synaptic connectivity optimization whereby initial sensory experience remodels circuits to a variable environment. In the Drosophila juvenile brain, synapse remodeling occurs within the precisely-mapped olfactory circuit, which has an extensively characterized, manageably short (< 1 week) critical period. In this brain circuit, single receptor olfactory sensory neuron (OSN) classes synapse onto single projection neurons extending to the central mushroom body learning/memory center. Critical period odorant experience drives OSN synapse remodeling, which can only be reversed during this brief interval. Our objective is to dissect intercellular signaling pathways from neurons to glial phagocytes sculpting synapse elimination in response to critical period experience. We find critical period experience causes externalized phosphatidylserine (PS) exposure in activated OSN synaptic glomeruli in an experiential dose-dependent mechanism. We discover that genetic knockdown of phosphatidylserine synthase inhibits critical period experience-dependent pruning of these synaptic glomeruli. We show a genetic interaction in trans-heterozygous mutants of phosphatidylserine synthase and Draper (mammalian MEGF10), the well-conserved glial engulfment receptor that binds phosphatidylserine, with double trans-heterozygotes blocking critical period experience-dependent pruning. This interaction mechanistically links phosphatidylserine signaling to glial phagocytosis synapse elimination. We identify the OSN scramblase that transports phosphatidylserine from the synaptic membrane inner to outer leaflet, and demonstrate phosphatidylserine externalization is rate-limiting for experience-dependent synaptic glomeruli pruning. We discover glial insulin receptors direct experience-dependent glial infiltration phagocytosis. We find activated glial insulin receptor signaling elevates critical period synapse pruning. Together this work identifies coupled intercellular signaling pathways from target neurons to glial phagocytes orchestrating experience-dependent synapse elimination.

Keywords: Draper/MEGF10; Insulin receptor; Insulin-like peptide; Phosphatidylserine.

Fig. 1

Critical period odorant experience triggers synaptic PS externalization. (A) Brain connectome schematic of the EB-responsive Or42a olfactory sensory neurons synapsing onto projection neurons within VM7 glomeruli (https://flywire.ai). Abbreviations: the antennal lobe (AL), mushroom body (MB), lateral horn (LH), and the optic lobe (OL). (B) Representative images of Or42a neurons innervating VM7 synaptic glomeruli with Or42a-Gal4 driven UAS-MApHS-pHluorin::tdTomato after 24-hour exposure to oil vehicle control (oil, left) or ethyl butyrate odorant (EB, right). Experience was applied either during the critical period (0–1 days post-eclosion (dpe); top) or in mature adults (7–8 dpe; bottom). Synaptic glomeruli pruning (red) is temporally-restricted within the critical period (top). Scale bar: 10 μm. (C) Representative images of VM7 glomeruli labeled for astrocyte-like glia (ALG)-Gal4 driven UAS-LactC1C2::GFP (ALG > LactC1C2) following 24-hour critical period exposure to the oil vehicle (top left), 5% EB (top right), 15% EB (bottom left), and 25% EB (bottom right). Synaptic glomeruli PS externalization (green) is dose-dependent. Scale bar: 5 μm. (D) Quantification of LactC1C2-GFP fluorescence intensity within the VM7 glomeruli normalized to the oil vehicle control, compared to 5–25% EB. Scatterplots show all the data points and the mean ± SEM. Significance indicated as not significant (p > 0.05, ns), or significant at p ≤ 0.001 (***) and p ≤ 0.0001 (****).

Fig. 2

Phosphatidylserine synthase is critical for experience-dependent pruning. (A) Representative images of Or42a neurons innervating VM7 synaptic glomeruli with the Or42a promoter driven UAS-mCD8::GFP membrane marker (Or42a-GFP, green) after 24-hour critical period (0–1 dpe) exposure to oil vehicle (top) or EB experience (bottom). The Or42a-GFP/+ genetic background control (left) shows robust experience-dependent synaptic glomeruli pruning (arrows), which fails in trans-heterozygous pss¹⁵/pss³² mutants (right). Scale bar: 10 μm. (B) Quantification of the Or42a OSN innervation volume in VM7 glomeruli in all conditions, normalized to the oil vehicle control. (C) Representative images under the same conditions as above with the Or42a-GFP/+ genetic background control (left) compared to targeted Or42a-Gal4 driven UAS-pss RNAi in pss^∆1/+ heterozygotes (pss RNAi/pss^∆1, right). Scale bar: 10 μm. (D) Quantification of the Or42a OSN innervation volume within VM7 glomeruli in all four conditions, normalized to the oil vehicle control. Scatterplots show all the data points and the mean ± SEM. Significance is indicated as not significant (p > 0.05; ns), or significant at p ≤ 0.05 (*), p ≤ 0.01 (**), and p ≤ 0.0001 (****).

Fig. 3

Draper and PS synthase interact for critical period synaptic pruning. (A) Representative images of the Or42a neuron innervation of the VM7 synaptic glomeruli (Or42a-GFP, green) with 24-hour critical period exposure to oil vehicle (left) or EB (right). Robust innervation pruning occurs in the genetic background control (Or42a-GFP/+, top). Similar VM7 synaptic glomeruli pruning occurs in pss^∆1/+ and draper^∆5/+ heterozygote controls comparing the oil vehicle (left) with EB experience (right, middle two rows). The double trans-heterozygous combination (pss^∆1/+; draper^∆5/+) shows normal innervation in the oil vehicle (bottom, left) and full block of experience-dependent synaptic glomeruli pruning with the EB odorant (bottom, right). Scale bar: 10 μm. (B) Quantification of the Or42a OSN innervation volume in VM7 glomeruli normalized to the genetic background oil vehicle control for all four genotypes and both conditions (8 comparisons). For each matched pair, 24-hour critical period exposure to oil vehicle (left) and EB odorant (right) is shown. Scatterplots show all the data points and the mean ± SEM. Significance is indicated at p ≤ 0.0001 (****), or not significant (p > 0.05; ns).

Fig. 4

Neuronal scramblase is rate-limiting for experience-dependent pruning. (A) Representative images of Or42a neurons innervating the VM7 glomeruli (Or42a-GFP, green) following 24-hour critical period exposure to oil vehicle (top) or 25% EB odorant (bottom). Normal experience-dependent pruning occurs in driver controls (Or42a-Gal4/+, left, arrows), which is impaired by Or42a-Gal4 targeted UAS-subdued scramblase RNAi (Or42a > subdued RNAi, right). Scale bar: 10 μm. (B) Quantification of the Or42a OSN innervation volume in VM7 glomeruli in all conditions, normalized to the oil vehicle control. (C) Representative images under the same conditions as above with the driver control (Or42a-Gal4/+, left) compared to Or42a-Gal4 driven UAS-TMEM16F scramblase over-expression (Or42a > TMEM16F OE, right) in oil vehicle (top) or 15% EB (bottom). Elevated scramblase exacerbates critical period experience-dependent pruning in the reduced EB condition (bottom, right, arrows). Scale bar: 10 μm. (D) Quantification of the Or42a OSN innervation volume in VM7 glomeruli in all conditions, normalized to the oil vehicle control. Scatterplots show all the data points and the mean ± SEM. Significance is indicated as not significant (p > 0.05; ns), or significant at p ≤ 0.01 (**) and p ≤ 0.0001 (****).

Fig. 5

Glial insulin receptors drive experience-dependent glomerulus infiltration. (A) Representative images of Or42a neurons innervating the VM7 synaptic glomerulus (Or42a-GFP, green, top) and the glial repo-Gal4 driven UAS-mCD8::RFP membrane marker (repo-RFP, red, bottom) following 24-hour critical period exposure to the oil vehicle (left) or EB odorant (right). Normal experience-dependent innervation pruning occurs (top, right, green) with robust glial infiltration (bottom, right, red). (B) Representative images under the same conditions as above with repo-Gal4 driven insulin receptor (InR) RNAi targeted to glia (repo > InR RNAi, right). Experience-dependent innervation pruning fails (top right, green) with an absence of glial infiltration into the VM7 synaptic glomerulus (bottom right, red). Scale bar: 5 μm. (C) Normalized quantification of glial membrane fluorescence intensity levels in the repo-Gal4/+ driver control (left) compared to glial-targeted InR RNAi (right). Paired 24-hour critical period exposure to the oil vehicle (left) and EB odorant (right). Scatterplots show all the data points and the mean ± SEM. Significance is indicated at p ≤ 0.0001 (****) or as not significant (p > 0.05; ns).

Fig. 6

Glial insulin receptors bidirectionally control critical period pruning. (A) Representative images of Or42a neurons innervating the VM7 synaptic glomeruli (Or42a-GFP, green) following 24-hour critical period exposure (0–1 dpe) to the oil vehicle (top) or 25% EB odorant (bottom). The glial driver control (repo-Gal4/+) shows normal experience-dependent synaptic glomeruli pruning (left, bottom, arrows), which is strongly impaired by repo-Gal4 glial-targeted InR RNAi (repo > InR RNAi, right). Scale bar: 10 μm. (B) Quantification of the Or42a OSN innervation volume in VM7 glomeruli in all conditions, normalized to the oil vehicle control condition. (C) Representative images under the same conditions as above, except the oil vehicle (top) is compared to a reduced 15% EB exposure (bottom). The glial driver control (repo-Gal4/+, left) is compared to glial-targeted repo-Gal4 driven UAS-InR constitutively-active (CA) construct (repo > InR-CA, right). At 15% EB, the reduced Or42a OSN synaptic glomeruli pruning (left) is exacerbated by activating InR signaling in glia (right, bottom, arrows). Scale bar: 10 μm. (D) Quantification of the Or42a OSN innervation volume in VM7 glomeruli in all conditions, normalized to the oil vehicle control. Scatterplots show all the data points and mean ± SEM. Significance is indicated as p ≤ 0.05 (*), p ≤ 0.0001 (****), or not significant (p > 0.05; ns).