A Phenotype-Driven Approach to Generate Mouse Models with Pathogenic mtDNA Mutations Causing Mitochondrial Disease

Abstract

Mutations of mtDNA are an important cause of human disease, but few animal models exist. Because mammalian mitochondria cannot be transfected, the development of mice with pathogenic mtDNA mutations has been challenging, and the main strategy has therefore been to introduce mutations found in cell lines into mouse embryos. Here, we describe a phenotype-driven strategy that is based on detecting clonal expansion of pathogenic mtDNA mutations in colonic crypts of founder mice derived from heterozygous mtDNA mutator mice. As proof of concept, we report the generation of a mouse line transmitting a heteroplasmic pathogenic mutation in the alanine tRNA gene of mtDNA displaying typical characteristics of classic mitochondrial disease. In summary, we describe a straightforward and technically simple strategy based on mouse breeding and histology to generate animal models of mtDNA-mutation disease, which will be of great importance for studies of disease pathophysiology and preclinical treatment trials.

Graphical abstract

Figure 1

Breeding Scheme to Generate Mouse Lines Transmitting Pathogenic mtDNA Mutations Females heterozygous for the PolgA^MUT allele are used to generate germline-transmissible mtDNA mutations and mated to wild-type males. Female offspring, which are wild-type at the PolgA locus, are further bred to generate female-derived lines, transmitting the mutated mtDNAs. After establishing maternal lineages to at least the third generation, founder females are sacrificed and their colonic crypts screened. Colonic crypts in founder mice were screened for the presence of mitochondrial dysfunction (some blue crypts on COX/SDH staining). Established mouse lines where the founder showed normal mitochondrial activity (only brown crypts on COX/SDH staining) were discontinued. To identify mtDNA mutations segregating with the mitochondrial dysfunction, the complete mitochondrial genome is sequenced from crypts deficient in mitochondrial function (blue crypts). Using this screening procedure, we identified three distinct lines harboring COX-deficient cells out of the 12 lines. See also Figure S1.

Figure 2

The tRNA^ALA C5024T Mutation Is Identified as a Pathogenic mtDNA Mutation (A) Representative COX/SDH staining of colonic crypts from wild-type mouse (WT) and a mouse that carries the tRNA^ALA C5024T mutation. Black bar represents 100 μm. Crypts, which are brown, have normal COX activity; those that are blue have deficient COX activity. (B) Electropherograms from mtDNA sequences obtained from isolated colonic crypts from WT mice and mice with the C5024T mutation shows the correlation of relative level of C5024T mutation and mitochondrial dysfunction in the colonic crypt. The crypts that are positive for mitochondrial function (COX+) show lower relative levels of the C5024T mutation, and colonic crypts deficient in mitochondrial function (COX−) show higher relative levels of the C5024T mutation. (C) Relative levels of the heteroplasmic C5024T mutation in individually dissected colonic crypts that are either COX positive or COX negative, showing that high levels of the C5024T mutation are present in the COX-negative colonic crypts. Error bars indicate SD. ^∗∗∗p < 0.001; ^∗∗∗∗p < 0.0001 (Mann-Whitney U test). (D) Clover-leaf representations of the tRNA^ALA from humans (left) identifying the positions of known pathogenic mutations and the location of the C5024T mutation in the structure of tRNA^ALA from mice (right). See also Figure S2.

Figure 3

Transmission of the C5024T Mutation in Mice (A) Comparison of the relative levels of the C5024T mutation in 1,105 offspring to heteroplasmic mothers. The red line represents the observed maximum level of the mutation. (B) Four representative tests for neutral segregation of the C5024T mutation using the Kimura distribution (Wonnapinij et al., 2008). Gray bars represent the observed levels of the mutation compared to the expected neutral distribution (orange line). Neutral segregation of the allele was observed until the mother carries 59% of the C5024T mutation, and then the segregation varies from the neutral prediction. All tests are represented in Figure S3.

Figure 4

Abnormal Physiology Caused by Mitochondrial Dysfunction in the tRNA^ALA C5024T Mice (A) Male mice carrying the C5024T mutation are smaller than age-matched wild-type animals when comparing body mass, lean mass, and fat content. n = 10–14 for each genotype and gender. Two independent cohorts were analyzed. ^∗p < 0.05; ^∗∗p < 0.01 (Dunn’s multiple comparison test). (B) Mice with high relative levels of the C5024T mutation show elevated heart mass, indicative of cardiomyopathy. ^∗p < 0.05 (Mann-Whitney U test). (C) The presence of COX-negative smooth muscle fibers in the colonic smooth muscle of mice with high levels of the C5024T mutation after ∼1 year of age. Black bar represents 50 μm. (D) Relative levels of the C5024T mutation from laser-capture dissected COX-positive and COX-negative smooth muscle fibers, showing tight co-segregation of high levels of the mutation and COX-negative phenotype. n = 15–22 experiments per group. ^∗∗∗∗p < 0.0001 (Mann-Whitney U test). For box-and-whisker plots, bars represent data range, + represents mean, line represents median, and box shows 25^th–75^th percentile of the data. See also Figure S4.

Figure 5

Selective Loss of the C5024T Mutation from Blood and the Colonic Epithelium of Aged Mice (A) Relative levels of the C5024T mutation measured from laser-capture dissected colonic smooth muscle versus colonic epithelium from the same mouse reveals a specific loss of the mutation in the colonic epithelium. n = 8. ^∗∗p < 0.01 (Wilcoxon matched-pairs signed rank test). (B) Relative levels of the C5024T mutation from blood cells. Animals with high levels of the mutation (>60% in other tissues) show decreased mutation levels in blood at 50 weeks of age. ^∗∗p < 0.01 (Mann-Whitney U test). For box-and-whisker plots, bars represent data range, + represents mean, line represents median, and box shows 25^th–75^th percentile of the data. See also Figure S5.

Figure 6

tRNA^ALA C5024T Leads to Depletion of tRNA^ALA and Deficiency in Mitochondrial Translation (A) Northern blot analyses of various mitochondrial tRNA, rRNA and mRNA transcripts from heart reveals strong depletion of the steady-state levels of tRNA^ALA, and a mild increase in steady-state levels of some other mitochondrial transcripts. Data are pooled from three independent experiments. WT, n = 12; C5024T, n = 19 (mean age, 65 weeks; C5024T 44%–77%). Error bars represent SD. ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗∗p < 0.0001 (Mann-Whitney U test). (B) Steady-state levels of tRNA^ALA in comparison with the relative levels of mtDNA with the C5024T mutation in heart. n = 31. ^∗p < 0.0001 (linear regression). (C) In organello translation of mitochondria isolated from heart reveal decreased translation capacity in tissues harboring high levels of the mutation and the occasional appearance of low-molecular-weight aberrant translation products (^∗) consistent with prematurely terminated or stalled translation. As a loading control, Coomassie blue staining of proteins after SDS-PAGE is shown. See also Figure S6.