Increased excitation during the embryonic CP alters network stability and motoneuron network synchrony

(A) Schematic demonstrating timeline for picrotoxin (PTX) exposure, and assay of network function. PTX-induced increase in activity during development increases third-instar larval duration of recovery from induced seizure (B) but does not change crawling velocity (C). (D) ROIs captured GCaMP signal generated by forward waves passing through ipsilateral aCC motoneuron axons in the L3 ventral nerve cord, abdominal segments (A6-4). (E) Representative traces recorded in 3 adjacent abdominal segments (A6-A4), during 2 forward waves. Colours represent different segments. Synchrony was measured as the time lag (s) of activity passing between adjacent segments (i.e., between peaks, dotted lines). (F) Embryonic exposure to PTX caused a significant increase in synchronicity across segments A6-A4 versus EtOH controls.

Increased excitation during the CP preferentially increases the strength of excitatory inputs to aCC

(A) Left-hand panel: schematic demonstrating use of optogenetics for testing strength of A18a or A31k synaptic drive on evoked action potential firing in aCC. Right-hand panel: representative traces illustrating the effect of excitatory or inhibitory input on evoked action potential firing in aCC, respectively. (B, C) current-clamp recordings wherein current (∼10pA) was injected to evoke action potential firing in aCC, prior to optogenetic stimulation of premotor input. Amplitude of inputs was quantified as change in frequency of action potential firing (ΔAP Freq. (Hz)) following optogenetic stimulation of A18a or A31k. Change in action potential firing frequency in A18a (p = 0.04), but not A31k (p = 0.15), was significantly potentiated by embryonic exposure to PTX vs vehicle controls (EtOH).

Increasing activity during the embryonic CP increased mini amplitude but not frequency in L3 aCC

(A) Representative trace showing minis (green arrows) in L3 aCC. Events not marked are not mini’s and probably represent single channel events (see Fig S1D for an expanded version of this same trace). (B) Mini frequency, as events observed in the first 20s of a whole-cell voltage-clamp recording (Vm held at −60mV), show no difference between vehicle controls (EtOH) and L3 exposed to picrotoxin (PTX) during embryogenesis. (C) Mini amplitude from the same recordings reveal significantly larger events in larvae that were exposed to PTX during embryogenesis, vs. vehicle controls (EtOH). (D) Cumulative distribution (%) of mini amplitude demonstrates a rightward shift in current size, following activity manipulation.

Increasing activity during the embryonic CP increased A18a-specific mini amplitude but not frequency in L3 aCC

(A) Representative traces showing minis (mEPSPs, green arrows) in L3 aCC (Vm held at −60mV), following optogenetic excitation of A18a. (B) Mini frequency, measured as events observed in the 3s following optogenetic stimulation of A18a (50ms light pulse) was not different between vehicle-only controls (EtOH) and L3 exposed to picrotoxin (PTX) during development. (C) Mini amplitude was significantly larger following exposure to PTX during embryogenesis, versus controls (EtOH). (D) Cumulative distribution (%) of mini amplitudes demonstrates a clear rightward shift in current size, following activity manipulation.

Increasing neural network activity by PTX exposure during development does not alter motoneuron dendritic arbor size or excitatory postsynaptic sites

(A-B) Representative images of aCC motoneurons from larval ventral nerve cords 72 hours after larval hatching (ALH), from control (A) or embryonic exposure to picrotoxin (B). aCC dendritic arbor structure is marked by membrane targeted myr::tdTomato (red), and postsynaptic excitatory sites are marked by a Drep2::YPet fusion protein (cyan). Visualisation of individual aCC motoneurons is achieved by an RN2-enhancer element driving a stochastic FLP-out LexA expression system. (C) tdTomato fluorescence was used for AI-supported 3D reconstruction of the whole aCC dendritic arbor, allowing quantification of total dendrite length (grey segments), branch points (blue spots) and terminals (red spots). (D) Following live imaging, nerve cords were fixed and stained using a fluorophore-tagged nanobody to detect the Drep2::YPet fusion protein (D), enabling high resolution and photostable imaging of Drep2::YPet puncta for quantification of excitatory postsynaptic site number by semi-automated spot detection (D’). (E) The total dendrite length, number of branch points, and number of terminals are shown, for both control and picrotoxin treatments. (F) The number of Drep2::YPet puncta for the larger ipsilateral arbor (‘arbor1’) the smaller contralateral arbor (‘arbor2’) and the combined sum across both arbors (‘aCC’) is shown, for both control and picrotoxin treatments. (G) Synapse density, calculated as the total dendrite length divided by total Drep2::YPet positive postsynaptic sites. In E-G, each point represents a single aCC cell, where the symbol shape denotes the particular cell.

Activating mechanosensory neurons during the CP rescues E:I balance due to increasing activity during embryogenesis

(A) Schematic showing stimulation of mechanosensory chordotonal neurons (ch) by vibration during the CP (17-19h after egg laying, AEL). (B) Stimulating ch neurons during the CP, but not during 15-17h AEL or 19-21h AEL, significantly increases mature larval (L3) recovery duration from electroshock-induced seizure. CS is unmanipulated wild-type (Canton-S) controls. (C-E) Exposure to picrotoxin (PTX) during embryogenesis significantly increases mature larval (L3) recovery duration from electroshock-induced seizure, increases network synchronicity, and increases the strength of excitatory (A18a) inputs to motor neurons (aCC). All these effects are rescued by co-activation of ch neurons during the CP (PTX + Ch).

Restoring E:I balance during the CP rescues seizure phenotype caused by a genetic mutation

(A) The parabss (bss) mutation causes lifelong neural hyperactivity, which is evident in a long duration of recovery from seizure when compared to wild-type controls (Canton-S). This recovery can be rescued by balancing the hyperactivity present with inhibition provided by sufficiently strong stimulation of ch neurons, during the CP (bss + 80dB). Weaker stimulation of the ch neuron-mediated inhibition provides intermediate phenotypes (bss + 60dB, bss + 40dB) that suggest the function of the mature network reflects the sum of excitatory and inhibitory activity during the CP. (B) parabss raised at 21°C experience less excitation than those raised at 25°C (e.g. in (A)) during development, and demonstrate a reduced seizure phenotype compared to the latter. This phenotype is completely rescued by 40dB stimulation of ch neurons (vs. 80dB at 25°C (A)), which supports the idea that it is a balanced sum of excitatory and inhibitory activity during the CP, that enables normal mature network function. Similarly, exposing parabss raised at 21°C to 80dB vibration during the CP results in a significant increase in seizure duration. We hypothesize that at this temperature, inhibition provided by ch neuron activity is greater than the increase in excitation resulting from the parabss mutation.