UL12.5 mice brain and UL12.5 neurons were differentiated from 90D iPSCs rna-seq

Mitochondrial DNA (mtDNA) mutations and/or depletion are implicated in epilepsy and many neurodegenerative diseases. However, systematic investigation into how mtDNA alterations relate to epilepsy and neural degeneration is needed. Here, we established a mouse model in which mtDNA depletion is induced by the Herpes Simplex Virus Type 1 (HSV-1) protein UL12.5 in the brain led to an epileptic phenotype characterized by abnormal electroencephalography (EEG) patterns and increased neural excitability in hippocampus. We also found that UL12.5 mediated mtDNA depletion in neurons in vitro (rho-) causes epilepsy–like abnormal EEG. Caloric restriction (CR) or glucose restriction (GR) is a strategy proven to reduce epileptic activity, however GR mimetic 2-deoxy-D-glucose (2-DG), induced degeneration in mtDNA depleted neurons. Mechanistically, mtDNA depletion increased mitochondria-endoplasmic reticulum (ER) contacts, facilitating GR-induced mitochondrial calcium overload. Rho- neurons did not show changes in mitochondrial motility or membrane potential. Our study revealed an unexpected axis of mtDNA depletion, ER-mitochondrial contacts, and calcium overload in the rho- neuron model. Fasting-induced GR causes early motor dysfunction, accelerates epilepsy progression, and worsens neurodegeneration in UL12.5 mice. Importantly, the IP3R inhibitor 2-APB blocks the neurodegeneration induced by fasting. This is the first description of animal and neuronal models of mitochondrial epilepsy. Our findings with these models suggest that GR may not be a viable clinical intervention in patients with mtDNA depletion.

Methods
RNA-seq analysis

Total RNA was extracted with TRIzol. Libraries were prepared using an Illumina TruSeq RNA Sample Prep kit (Illumina, 20020594) following the manufacturer’s instruction. The RNA-seq experiment was performed at Annoroad Gene Technology Co. Ltd. (Beijing, China). The raw reads of RNA-seq were trimmed from 3′ to 5′ ends for each read, and the reads shorter than 50 bp were discarded. The reads that passed the quality control were mapped to mouse genome (mm10) with STAR (version 2.7). Stringtie (version 2.1.4) were then used to estimate transcript abundance in fragments per kilobase per million (FPKM). Genes were annotated as differentially expressed using DESeq2 (version 1.44). The GO analysis and KEGG pathway analysis were performed using R packages.