Figure 3. Neu3 is necessary and sufficient to disrupt neuronal network connectivity.

Neurons were labeled with the voltage-sensitive dye BeRST1 and treated with extracellular-vesicle enriched media from either microglia or Neu3 over-expressing cells. Neuronal firing rates and network connectivity were analyzed by fluorescence microscopy. (A) BeRST1 is a membrane-localizing voltage-sensitive fluorophore that undergoes a dramatic increase in fluorescence intensity in response to changes in membrane potential, i.e. upon the depolarization of firing neurons. Representative brightfield and BeRST1 fluorescence of a single field of view and voltage traces of each neuron in a single field of view contain both subthreshold activity and spiking activity. (B) Network connectivity is quantitated by measuring the spike traces for individual neurons within a single field of view, and then looking at the synchronicity of firing by the metric of Shared Variance. (C,D) BV-2 microglia treated with or without LPS and with or without GW4869. The EVs from the conditioned media were enriched and neurons were treated with EV-enriched media, and neuronal activity was measured by voltage imaging with BeRST1. (C) Firing rates of neurons treated with BV-2 EV-enriched media reveal 1.7 Hz decrease in +LPS condition compared to -LPS condition (p=0.048). The effect is partially rescued by inhibition of EV production with GW4869 (+LPS vs. +LPS+GW4869, p=0.062). (D) Treatment with EV-enriched media results in a 29% reduction in subthreshold shared variance per neuron in -LPS vs. +LPS conditions, suggesting that neurons are no longer well-connected to the network (p=0.015). Effect is rescued by addition of GW4869 (+LPS vs. +LPS+GW4869, p=0.029). (E,F) As in (C,D), but using conditioned media from HeLa cells overexpressing either wild-type or loss-of-function (Y369F) Neu3. (E) Firing rates of neurons treated with EV-enriched media of NEU3-overexpressing HeLa cells reveal no significant changes between functional NEU3 and a loss-of-function point mutant (WT: −0.26 Hz, p=0.58; Y369F: −0.14 Hz, p=0.65). (F) Treatment with EV-enriched HeLa media reveals 15% reduction in subthreshold shared variance between WT and Y369F mutant (p=0.046). Coincubation of WT EV-enriched media with DANA prevents this reduction (p=0.028), while coincubation of Y369F media with DANA has no significant effect (p=0.48). For (C,D): n=4 coverslips/condition, 168 neurons total. For (E,F): n=3 coverslips/condition, 331 total neurons. All hypothesis testing was performed by hierarchical permutation tests.