CHCHD10 mutations induce tissue-specific mitochondrial DNA deletions with a distinct signature

Abstract

Mutations affecting the mitochondrial intermembrane space protein CHCHD10 cause human disease, but it is not known why different amino acid substitutions cause markedly different clinical phenotypes, including amyotrophic lateral sclerosis-frontotemporal dementia, spinal muscular atrophy Jokela-type, isolated autosomal dominant mitochondrial myopathy and cardiomyopathy. CHCHD10 mutations have been associated with deletions of mitochondrial DNA (mtDNA deletions), raising the possibility that these explain the clinical variability. Here, we sequenced mtDNA obtained from hearts, skeletal muscle, livers and spinal cords of WT and Chchd10 G58R or S59L knockin mice to characterise the mtDNA deletion signatures of the two mutant lines. We found that the deletion levels were higher in G58R and S59L mice than in WT mice in some tissues depending on the Chchd10 genotype, and the deletion burden increased with age. Furthermore, we observed that the spinal cord was less prone to the development of mtDNA deletions than the other tissues examined. Finally, in addition to accelerating the rate of naturally occurring deletions, Chchd10 mutations also led to the accumulation of a novel set of deletions characterised by shorter direct repeats flanking the deletion breakpoints. Our results indicate that Chchd10 mutations in mice induce tissue-specific deletions which may also contribute to the clinical phenotype associated with these mutations in humans.

Keywords: CHCHD10; mitochondria; mitochondrial DNA; mtDNA deletions; neurodegeneration.

Figure 1

Deletions in WT and G58R mice. (A) Outline of the experimental setup. Tissue was harvested from WT, G58R and S59L mice (25 hearts, 25 livers, 24 tibialis anterior muscles and 22 spinal cords). Whole DNA was extracted from the tissue samples and mtDNA was amplified by PCR and sequenced. (B) Circle plots representing mtDNA and showing the deletions called through MitoSAlt. Darker shades correspond to higher heteroplasmy (range: 0.5–16%). Deletions shown are compiled from the number of samples indicated in the centre of the circle. (C) Total deletion burden vs. age plots. Each dot indicates a different sample. For (C) an equality of slopes test was performed for WT vs G58R.

Figure 2

The G58R mutation accelerates the accumulation of a liver deletion. (A) Left: Demonstration of the matrix setup used for PCA. Right: PCA for all samples with at least 3 identified deletions. The numbers within the parentheses indicate the percentage of data variance that is explained by each PC. (B) The deletions with the highest (liver predominance) and lowest (heart/muscle predominance) PC1 loading scores. (C) Start and end breakpoints of all the deletions identified from WT and G58R mouse tissues. Each row is an individual deletion, with deletions arranged by size. (D) Heteroplasmy vs. age plot of the 3821 bp deletion in liver samples of WT and G58R mice. (E) Long-range PCR amplifying an ~8 kb region surrounding the minor arc. The bright band around 8 kb is full-length mtDNA, and the fainter band indicated by an arrow is the 3821 bp deletion. For (D) an equality of slopes test was performed.

Figure 3

Chchd10 mutant mtDNA deletions are flanked by shorter direct repeats. (A) Top left: Illustration of an 8 bp direct repeat (microhomology) present in two locations within the mitochondrial genome. Note how one of the repeats is within the range of the deletion and is therefore lost in the deleted molecule. Rest: Length of direct repeats surrounding each deletion. (B) Representative WT and G58R mtDNA deletions. Direct repeats are coloured red, and deletion breakpoints are marked by asterisks. (C) Start and end breakpoints of all the unique deletions identified and coloured by the length of the direct repeats flanking them. Each deletion is shown only once even if it was identified in multiple samples. (D) Total deletion burden vs age plots for samples of the respective tissues, counting only deletions with microhomologies of 8 bp or longer. For (A), one-sided Mann-Whitney tests were performed. For (C), equality of slopes tests were performed.

Figure 4

Chchd10 mutations do not increase mtDNA SNV mutational burdens. (A) Number of SNVs identified for each sample sequenced, annotated by impact of the mutation. (B) Summed SNV heteroplasmy fractions (expressed as percentages) for each sample, with samples stratified into two age groups. For (B), two-sided Mann-Whitney tests were performed.

Figure 5

Proteins implicated in multiple mtDNA deletion syndromes, and known effects of CHCHD10 mutations on some of these proteins. Mutant CHCHD10 causes increased levels of C1QBP, AFG3L2 and SPG7 and cleavage or degradation of OPA1 and ENDOG [36]. In addition, CHCHD10 mutations disrupt IMM stability. These changes could contribute towards mtDNA deletion formation in CHCHD10 mutants.