Neither cardiac mitochondrial DNA variation nor copy number contribute to congenital heart disease risk

Abstract

The well-established manifestation of mitochondrial mutations in functional cardiac disease (e.g., mitochondrial cardiomyopathy) prompted the hypothesis that mitochondrial DNA (mtDNA) sequence and/or copy number (mtDNAcn) variation contribute to cardiac defects in congenital heart disease (CHD). MtDNAcns were calculated and rare, non-synonymous mtDNA mutations were identified in 1,837 CHD-affected proband-parent trios, 116 CHD-affected singletons, and 114 paired cardiovascular tissue/blood samples. The variant allele fraction (VAF) of heteroplasmic variants in mitochondrial RNA from 257 CHD cardiovascular tissue samples was also calculated. On average, mtDNA from blood had 0.14 rare variants and 52.9 mtDNA copies per nuclear genome per proband. No variation with parental age at proband birth or CHD-affected proband age was seen. mtDNAcns in valve/vessel tissue (320 ± 70) were lower than in atrial tissue (1,080 ± 320, p = 6.8E-21), which were lower than in ventricle tissue (1,340 ± 280, p = 1.4E-4). The frequency of rare variants in CHD-affected individual DNA was indistinguishable from the frequency in an unaffected cohort, and proband mtDNAcns did not vary from those of CHD cohort parents. In both the CHD and the comparison cohorts, mtDNAcns were significantly correlated between mother-child, father-child, and mother-father. mtDNAcns among people with European (mean = 52.0), African (53.0), and Asian haplogroups (53.5) were calculated and were significantly different for European and Asian haplogroups (p = 2.6E-3). Variant heteroplasmic fraction (HF) in blood correlated well with paired cardiovascular tissue HF (r = 0.975) and RNA VAF (r = 0.953), which suggests blood HF is a reasonable proxy for HF in heart tissue. We conclude that mtDNA mutations and mtDNAcns are unlikely to contribute significantly to CHD risk.

Keywords: congenital heart disease; genome sequencing; mitochondrial copy number; mitochondrial genome.

Figure 1

mtDNAcn versus age in cardiovascular tissue mtDNAcn plotted against age for (A) atrial (n = 45), (B) ventricular (n = 39), and (C) valve/vessel tissues (n = 26) with Pearson correlation values (r) reported.

Figure 2

mtDNAcn versus age in blood mtDNAcn plotted against age for probands (blue; n = 1,830), mothers (red; n = 1,764), and fathers (gold; n = 1,745). Only samples from whole blood were included, with Pearson correlation values (r) reported.

Figure 3

HF in children versus mothers Variant HF in children is plotted against the maternal HF for variants discovered in either the child or the mother for (A) CHD and (B) SFARI. Variants are colored by classification as the control region (D-Loop), non-coding regions (NC), synonymous (Syn), missense (Miss), ribosomal RNA (rRNA), or transfer RNA (tRNA). Variants with HF < 0.1 in both child and mother are considered effectively homoplasmic wild type and are displayed in gray, while those with HF > 0.9 in both child and mother are considered effectively homoplasmic mutant and are displayed in black. The previously unreported variant 16469G is displayed in white.

Figure 4

HF in blood versus tissue Variant HF in tissue was plotted against variant HF in blood for 114 individuals with CHD. Heteroplasmic variants are colored in blue, those with HF < 0.1 in both blood and tissue are considered homoplasmic wild type and are displayed in gray, while those with HF > 0.9 in both blood and tissue are considered homoplasmic mutant and are displayed in black. The Pearson correlation (r) was calculated including only heteroplasmic variants.

Figure 5

Cardiovascular RNA VAF versus blood HF The VAF is plotted against HF of variants discovered in either the blood mtDNA or the cardiovascular mtRNA in (A) CHD-affected probands and (B) GTEx subjects. Variants with both VAF and HF < 0.1 considered effectively homoplasmic wild type and are displayed in gray, while those with both VAF and HF > 0.9 are considered effectively homoplasmic mutant and are displayed in black. The Pearson correlations (r) were calculated including only heteroplasmic variants.