Quantitative analyses of structure-function heterogeneity of individual mitochondrial components reveal an early role of small-mitochondrial-networks in priming neoplastic transformation
Lack of quantitative understanding of the marked heterogeneity of multifaceted mitochondria poses challenges in unraveling their translatable role. Here, we hierarchically untangled multilevel heterogeneity of mitochondrial networks and non-networks by simultaneously analysing their structure-function within single components and their units, in cells and tissues. Such mito-SinComp quantitative analyses revealed that redox levels of mitochondrial networks and their dramatic intra-network heterogeneity can be predicted by structural features through a translatable machine-learning approach. Mito-SinComp identified and quantitatively characterized a redox-tuned subpopulation of Small-Mitochondrial-Networks(SMNs) that specifies stemness. These SMNs are generated by severing of nodes of oxidized Hyperfused-Mitochondrial-Networks(HMNs), are ten-folds smaller with specific network complexity and elevated mt-DNA nucleoid abundance. Thus, HMN to SMN conversion supports elevated expression of mtDNA genes in establishing MT-ND1(redox)-KRT15(stemness) transcriptomic interaction, as revealed by coupled scRNA-seq. This stemness priming of non-transformed cells is sustained after neoplastic transformation and is detected in patient transcriptome, demonstrating translational relevance of SMNs.