Panx1 channels promote both anti- and pro-seizure-like activities in the zebrafish via p2rx7 receptors and ATP signaling

The molecular mechanisms of excitation/inhibition imbalances promoting seizure generation in epilepsy patients are not fully understood. Evidence suggests that Pannexin1 (Panx1), an ATP release channel, modulates the excitability of the brain. In this report, we performed electrophysiological, behavioral, and molecular phenotyping experiments on zebrafish larvae bearing genetic or pharmacological knockouts of Panx1a and Panx1b channels, each homologous to human PANX1. When Panx1a function is lost, or both channels are under pharmacological blockade, seizures with ictal-like events and seizure-like locomotion are reduced in the presence of pentylenetetrazol. Transcriptome profiling by RNA-seq demonstrates a spectrum of distinct metabolic and cell signaling states which correlate with the loss of Panx1a. Furthermore, the pro- and anticonvulsant activities of both Panx1 channels affect ATP release and involve the purinergic receptor P2rx7. Our findings suggest a subfunctionalization of Panx1 enabling dual roles in seizures, providing a unique and comprehensive perspective to understanding seizure mechanisms in the context of this channel.


Supplementary Figures
Supplementary Figure 1: The differential localization of Panx1a and Panx1b in the zebrafish larval diencephalon and spinal cord. a) Panx1a immunoreactivity was detected in the neuropil of arborization fields (AFs) in the diencephalon's retinorecipient areas (open arrows). a') shows Panx1a immunoreactive AFs (stars) and a meshwork of potential fiber bundles (X) penetrating layers of densely packed neuronal cells in the optic tectum (TeO), or a commissural fiber between hemispheres (X'). b) Panx1b immunoreactivity was dotted and localized in densely packed neuronal cells layers. b') shows part of the TeO with a region of Panx1b immunoreactivity (X) adjacent to AFs (stars). Note the non-overlapping localization of Panx1a and Panx1b in the retina of 6dpf larvae in both a) & b). c) Like in the diencephalon, Panx1a immunoreactivity was found in the neuropil of the spinal cord (arrows), or d) focal spots between densely packed neuronal cells in the optic tectum (Panx1b, arrows). e) The localization GFAP-positive radial glial cells in frontal sections of the spinal cord. Arrows point at the end-feet of astrocytes. Abbreviations: TeO, tectum opticum, MS, muscle, NC, notochord, SC, spinal cord, GFAP, Glial fibrillary acidic protein, DAPI, 4′,6-diamidino-2phenylindole. Scale bars a,b 100µm; a' 30µm; b' 20µm; c-e 25µm.

Supplementary Figure 2:
One hour baseline recordings of untreated larvae.

Supplementary Figure 3 Comparison of PTZ treatment to baseline activity is indistinguishable.
Spectrograms of baseline activity (top) directly compared to recordings during PTZ treatment (bottom) for DKO (left) and panx1a -/-(right) show that PTZ treatment does not affect neural activity in these genotypes. Sample traces from PTZ recordings at the 1hour mark are included for reference, revealing no major spiking activity.

Supplementary Figure 4 Self-Organizing Maps (SOM) emphasize phenotypic similarities of PTZ treatment in TL and panx1 knockouts.
SOM was applied on behavioral outputs extracted from the ZebraLab software on PTZ treated TL and panx1 knockout larvae (n = 36 per group). The absolute difference between the number of TL and panx1 knockout larvae assigned to each cluster in the SOM were averaged for all clusters and visualized as relative similarity in a form of heatmaps. Smaller value and warmer color represent higher similarity. TL and DKO exhibited the most similar PTZ-induced phenotypes across freezing and normal locomotor behaviors. In general, Panx1a -/and panx1b -/were most distinct from one another, particularly for freezing duration, normal behavior counts and total duration of movement (P < 0.0001). Overall freezing and normal behaviors showed less variation among genotypes compared to general activity and bursting behaviors. See Supplementary Data 3 for details on statistical significance for each behavioral output and genotype comparison. Spectrograms of baseline activity taken after 2hours of recording from TLs with no drug treatment (left) at all and TLs treated with probenecid only (right). This comparison shows that PROB is not inducing a toxicity effect to prevent seizure-like events from occurring as these spectrograms look like baseline activity.

Supplementary Figure 6 Venn diagrams of differentially regulated genes and biological processes.
Venn diagrams highlighting how many genes are regulated in each of our panx1 genotypes and show how many genes fit into biological processes according to FishEnrichR ontology database. They also reveal the overlap in regulated genes.

Supplementary Figure 7 Concentration dependent reduction of PTZ-related hyperactivity in TL larvae by Probenecid.
a) 7dpf TL larvae (n = 30) were treated with 15mM PTZ, or b) incubated with 25µM, 50µM, 75µM, and 100µM Probenecid (PROB; n = 18 per concentration) one hour before treatment with PTZ. PROB significantly increased activity (Δpixel; mean ± s.e.m.) compared to baseline, and significantly reduced PTZ-induced activity above 50µM. PROB applied at 75µM reduced PTZinduced activity, without the evidence of substantial toxicity (activity level 0 within one hour).

Supplementary Figure 8 Linear correlation of larval weight and protein concentration.
Plotting the weight of pooled larvae (n = 50 larvae/sample) in mg against the amount of protein in mg/ml that was measured from the supernatant of the homogenate with a spectrophotometer demonstrates a clear linear relationship (R^2 = 0.9146). Therefore, ATP concentrations were plotted per protein content to account for biological variance of larval weight.

Supplementary Tables
Supplementary Table 1 RT-qPCR values for Immediate Early Gene (IEG) regulation in TL and panx1 knockout larvae treated with PTZ for one hour (Figure 4f). Expression values, respective to non-treated controls, are >1 for upregulated IEGs and <1 for downregulated IEGs. P-values in bold = P < 0.05.