Alternating hemiplegia of childhood associated mutations in Atp1a3 reveal diverse neurological alterations in mice

Abstract

Pathogenic variants in the neuronal Na⁺/K⁺ ATPase transmembrane ion transporter (ATP1A3) cause a spectrum of neurological disorders including alternating hemiplegia of childhood (AHC). The most common de novo pathogenic variants in AHC are p.D801N (∼40 % of patients) and p.E815K (∼25 % of patients), which lead to early mortality by spontaneous death in mice. Nevertheless, knowledge of the development of clinically relevant neurological phenotypes without the obstacle of premature death, is critical for the identification of pathophysiological mechanisms and ultimately, for the testing of therapeutic strategies in disease models. Here, we used hybrid vigor attempting to mitigate the fragility of AHC mice and then performed behavioral, electrophysiological, biochemical, and molecular testing to comparatively analyze mice that carry either of the two most common AHC patient observed variants in the Atp1a3 gene. Collectively, our data reveal the presence but also the differential impact of the p.D801N and p.E815K variants on disease relevant alterations such as spontaneous and stress-induced paroxysmal episodes, motor function, behavioral and neurophysiological activity, and neuroinflammation. Our alternate AHC mouse models with their phenotypic deficits open novel avenues for the investigation of disease biology and therapeutic testing for ATP1A3 research.

Keywords: AHC; ATP1A3; ATPase activity; And Neuroinflammation; Dystonia; Seizure; Spreading depolarization.

Fig. 1.

B6C3 AHC mice develop spontaneous and stress-induced paroxysmal spells. (A) Kaplan-Meier survival curve of B6C3.Atp1a3 ^D801N/+ mice. Wild type littermate controls are shown (black and grey). (B) Kaplan-Meier survival curve of B6C3.Atp1a3 ^E815K/+ mice. Approximately 50 % of B6C3.Atp1a3 ^E815K/+ mice required euthanasia as the study end point (‘death’) due to the low body condition score (BCS). Spontaneous and BCS required deaths are included in the survival curve. Wild type littermate controls are shown (black and grey). (C) Analysis of the ATPase activity of tissues from the hippocampus of B6C3.Atp1a3 ^D801N/+ and B6C3.Atp1a3 ^E815K/+ mice. The data represent the mean ± SD. (D) Hypothermia induced paroxysmal spells (HIP) were induced in B6C3.Atp1a3 ^D801N/+ and B6C3. Atp1a3 ^E815K/+ mice. Dystonia-like events were scored during the recovery period of HIP experiments and the horizontal line (mean) for each group represents the average score. Wild type littermate controls for each mutant strain are shown (black and grey). (E) The occurrence of convulsive-like events of B6C3 AHC mice. The data are shown as the fraction (percent) of mice without (grey) or with (red) convulsion-like defects that were observed during the recovery period of HIP experiments. The exact number of mice without (black font) and with (red font) convulsion-like defects is shown in each bar. Wild type littermate controls for each mutant strain are shown. (F) Latency to regain mobility and movement control is shown. The horizontal line for each group represents the mean. Wild type littermate controls for each mutant strain are shown (black and grey). (G) The data in F are shown as a Kaplan-Meier curve. Wild type littermate controls for each mutant strain are shown (black and grey). M, males; F, females; BCS, body condition score. Mantel-Cox test (A, B, G); Two-way ANOVA was corrected for multiple comparisons using Tukey method (C, D); Fisher’s exact test (E); One-way ANOVA was corrected for multiple comparisons using Tukey method (F). ns, not significant; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. See also Fig. S1. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2.. B6C3 AHC mice show motor and behavioral deficits.

(A) Latency of B6C3 AHC mice to fall off the accelerating rotarod at various ages. The horizontal line for each group represents the mean. Wild type littermate controls for each mutant strain are shown (black and grey). (B) to (G) Naïve B6C3 AHC mice were subjected to open field testing to interrogate exploratory and/or locomotor activity. The horizontal line for each group represents the mean. Wild type littermate controls for each mutant strain are shown (black and grey). (B) Total distance traveled. Note: # D801N mice only showed a “significant” difference in total distance traveled when the data were not corrected for multiple comparisons. p = value of 0.0051 compared to littermate control mice (black) or p = value of 0.0018 when compared to wild type mice (grey). (C) Total movement time. (D) Total rest time. (E) Stereotypic episodes. (F) Vertical episodes. (G) Total time spent in the center or margin of the open field arena. Two-way ANOVA was corrected for multiple comparisons using Tukey method (A, G); One-way ANOVA was corrected for multiple comparisons using Tukey method (B, C, D, E, F). ns, not significant; * p ≤ 0.05; *** p ≤ 0.001; **** p ≤ 0.0001. See also Fig. S2.

Fig. 3.. B6C3 AHC mice show altered hippocampal and cortical network activity.

(A) Schematic representation of LFP recording electrodes placed in motor (M1) and visual (V1) cortex and dorsal CA1 hippocampus. Recordings were performed for 3 consecutive days per time point. Electrodes for reference (Ref.) and ground (Gr.) were placed in cerebellum. (B) Example traces of 5-min AC-LFP traces from the M1, dorsal CA1 hippocampus, and V1 (top-to-bottom) of wild type (black), B6C3.Atp1a3 ^D801N/+ (green) and B6C3.Atp1a3 ^E815K/+ (orange) mice at 14 weeks of age. (C) Example trace of spike-wave discharge (SWD) event in frontal M1 cortical channel recorded in wild type mice at 14 weeks of age. SWDs present as a prominent burst (highlighted by the red dotted line) of negative polarity spikes and a positive polarity wave, exclusively seen in the frontal M1 cortical channel with no corresponding activity in V1 cortical or dHC channels. The horizontal line for each group represents the average SWD frequency per hour. (D) Example traces of hippocampal spike events (red arrowheads) in the dorsal CA1 hippocampus of B6C3.Atp1a3 ^E815K/+ mice at 14 weeks of age. These concerned isolated hippocampal spikes that were not observed in M1 and V1 cortices. The horizontal line for each group represents the average spike frequency per hour. (E) Example traces of hippocampo-cortical giant spike observed in B6C3.Atp1a3 ^E815K/+ mice. A giant spike consists of a simultaneous large spike of varying polarity (indicated by the red arrowhead) seen in all channels that is followed by a slow negative-positive potential shift lasting >0.3 s in the M1 cortical channel and accompanied by a slow positive deflection lasting up to 600 ms in dHC and V1 cortex. The horizon- tal line for each group represents the average giant spike frequency per hour. (F) Average V1 cortical power spectral densities (PSD) during active wakefulness of wild type (black), B6C3.Atp1a3 ^D801N/+ (green) and B6C3.Atp1a3 ^E815K/+ (orange) mice. (G) Leftward shift of θ frequency peak observed in B6C3 AHC mice from 6 weeks onward. The vertical line for each group represents the mean. (H) Absolute power at θ frequency from the V1 cortical PSD during active wakefulness. The horizontal line for each group represents the mean. (I) Absolute power at α frequency from the V1 cortical PSD during active wakefulness. The horizontal line for each group represents the mean. (J) Absolute power at β frequency from the V1 cortical PSD during active wakefulness. The horizontal line for each group represents the mean. Ref, reference electrode; Gr, ground electrode; M1, motor cortex; V1, visual cortex; dHC, dorsal hippocampus CA1; CA1, hippocampal cornu ammonis area 1; SWD, spike-wave discharge. Two-way ANOVA was corrected for multiple comparisons using Tukey method (C, D, E, H, I, J); One-way ANOVA was corrected for multiple comparisons using Tukey method (G). ns, not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. The depicted top view of the mouse skull with electrode configuration and coronal brain views were created using BioRender.com (A). See also Figs. S3 and S4. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4.. Gene expression analysis of different brain regions in B6C3 AHC mice.

(A) UpSet plot of differentially expressed (DE) genes (adj. p ≤ 0.05) comparing AHC mutations and brain regions in mice. The anatomical locations of the analyzed mouse brain regions are shown in the upper, left corner. The total number of differentially expressed (adj. p ≤ 0.05) genes for each brain region of B6C3.Atp1a3 ^D801N/+ (D801N, green) and B6C3.Atp1a3 ^E815K/+ (E815K, orange) mice are shown in the lower left corner. The number of unique genes and the intersection of genes are shown on the right-hand side. Notable gene intersections between genotypes and brain region are highlighted in red. (B) Immunohistochemistry using antibodies to glial fibrillary acidic protein (GFAP, astrocytes, brown) on parasagittal brain sections from 35-week-old wild type, B6C3.Atp1a3 ^D801N/+ and B6C3.Atp1a3 ^E815K/+ mice. Sections were counterstained with hematoxylin and higher magnifications of individual brain regions are shown. (C) Relative intensity of GFAP of wild type, B6C3.Atp1a3 ^D801N/+ and B6C3.Atp1a3 ^E815K/+ mice. Levels are relative to that of controls (wild type) from each cross and brain region. The data represent the mean ± SD. Wild type littermate controls for each mutant strain are shown (black and grey). (D) Longitudinal serum analysis of neurofilament light chain (NFL) of B6C3 AHC mice. The horizontal line for each group represents the mean. Wild type littermate controls for each mutant strain are shown (black and grey). (E) Comparison of serum NFL levels between B6C3.Atp1a3 ^E815K/+ mice that were either deceased at a later age (orange) or remained alive (brown). The horizontal line for each group represents the mean. Scale bars: 150 μm (cortex), 50 μm (CA1), 150 μm (thalamus), 100 μm (cerebellar hemisphere, PML) and 50 μm (cerebellar hemisphere, DCN). DE, differential gene expression; CA1, hippocampal cornu ammonis area 1; ML, molecular cell layer; PL, Purkinje cell layer; GCL, granule cell layer; PML, paramedian lobe; and DCN, deep cerebellar nuclei. Two-way ANOVA was corrected for multiple comparisons using Tukey method (C, D); Student’s t-test (E). ns, not significant; **** p ≤ 0.0001. The depicted mouse brain with specific brain regions was created with BioRender.com (A). See also Figs. S5, S6 and S7. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)