Abstract
The accumulation of somatic mutations contributes to clonal evolution and biological properties of cancers. Acquired mutations in mitochondrial (mt)DNA have been studied, but with the exception of those in isocitrate dehydrogenase genes, no comprehensive assessment of mutations in nuclear mitochondrial genes has been reported in sequential glioblastoma (GBM). We obtained ten pairs of GBM samples at diagnosis (GBM-P) and at recurrence (GBM-R). Extracted DNA was subjected to whole exome and mtDNA sequencing. After filtering out germline variants, bioinformatics analysis was performed using a mitochondrial gene panel of 483 nuclear-encoded, and 37 mtDNA-encoded genes. Variant classification was performed using established clinical- and molecular criteria, integrating population-frequency data, bioinformatic predictions, functional evidence, segregation information, and curated entries from the Mitomap and ClinVar databases. Benign single nucleotide variants in mtDNA-encoded genes of RNR1, RNR2, ATP6, CYB, CO2, TV, ATP8, and ND2 were detected, which changed little over time. However, three variants in TI, ND5 and ND1 with possible or likely pathogenic significance were found in the GBM-R samples. In contrast, pathogenic or likely pathogenic variants in 29 nuclear genes were found in GBM-P and GBM-R samples. Not only the overall number, but also the number of protein-truncating variants in nuclear genes increased over time. Conclusions: This study sheds light on the accumulation of mutations in nuclear genes of mitochondrial proteins in sequential GBM samples. As such variants may influence metabolic, proliferative and invasive properties as well as the necrotic propensity of the tumor, a comprehensive analysis of these genes merits further studies.