The Origin and Functional Impact of Mitochondrial DNA Mutations in Acute Myeloid Leukemia

Recent whole-genome sequencing studies have identified somatic mutations in the mitochondrial genome of several human malignancies including acute myeloid leukemia (AML). Although mitochondrial DNA (mtDNA) mutations have previously been shown to promote tumorigenicity in some solid cancers, their role in AML development is unknown. To address this issue, we traced the origin of mtDNA mutations in primary human AML samples and determined their potential impact on mitochondrial respiration and cell fate decisions in hematopoietic stem and progenitor cells (HSPCs).

We first characterized the spectrum of mtDNA mutations in AML by performing hybrid capture-based next generation sequencing of the 13 protein-coding mtDNA genes in 133 AML patient samples collected at the time of diagnosis. We also sequenced 150 recurrently mutated nuclear genes in parallel. We aligned mtDNA sequences to the revised Cambridge Reference Sequence and identified variants with > 5% variant allele frequency. Since matched normal samples were not available, we filtered out variants with a population frequency of > 0.1% in GenBank to enrich for somatic variants. Of the remaining variants, we filtered out ones that were synonymous or caused an amino acid change at a non-conserved residue to enrich for pathogenic mutations. Based on these criteria, we identified 87 different missense mutations in 39% (n=53) of the patients across 12 of the 13 genes. We did not find any hot spot mutations. Interestingly, mtDNA mutations were mutually exclusive with IDH1 R132 mutations (p < 0.05, Chi-square test). Patients with mtDNA mutations had inferior leukemia-free survival compared to those without mtDNA mutations (134 vs 477 days; p = 0.002, Log-rank test).

To trace the origin of these mutations, we developed allele-specific TaqMan assays for 9 of the primary AML samples with known mtDNA mutations. The TaqMan assays were sufficiently sensitive to quantify heteroplasmy at the single cell level. Patient-specific mtDNA mutations were detected in single cells sorted from the hematopoietic stem cell (HSC) compartment (defined as CD45dim/CD33-/CD34+/CD38-/CD90+/CD45RA-) in 8 of the 9 samples. The heteroplasmy levels were generally lower in HSCs than leukemic blasts. Interestingly, mtDNA mutations were also detected at low levels in the T cell compartment of 3 of the 8 samples. Analysis of a matched diagnostic/relapsed pair showed persistence of the mtDNA mutation in HSCs and leukemic blasts. These findings collectively indicate that mtDNA mutations are acquired in HSCs and stable during disease evolution.

To investigate the potential impact of mtDNA mutations on mitochondrial respiration, we generated a pair of transmitochondrial cybrids by repopulating mtDNA-depleted 143B osteosarcoma cells with mitochondria from either KG-1α AML cells or peripheral blood cells of a healthy donor. KG-1α cells harbor a nonsense mtDNA mutation (C13396T) in the MT-ND5 gene which encodes a core subunit of Complex I in the electron transport chain (ETC). The KG-1α cybrid had lower levels of basal and maximal respiration compared with the healthy control cybrid as determined by oxygen consumption rates. To confirm that this defect was due to impairment of Complex I activity, we ectopically expressed the yeast NADH dehydrogenase (NDI1) protein which bypasses endogenous Complex I function. NDI1 expression completely rescued the mitochondrial respiration defect in KG-1α cybrids, whereas expression of the AOX protein which bypasses Complex II and III failed to do so. To determine the potential impact of Complex I dysfunction on cell fate decisions in HSPCs, we transduced human CD34+ cord blood cells with shRNA lentiviral vectors targeting the nuclear genes NDUFV1 and NDUFB11 which encode critical subunits of Complex I. Knockdown of the genes resulted in an accumulation of CD34+ cells in ex vivo culture over time.

In summary, our results demonstrate that mtDNA mutations are commonly found in AML patients and have potential impact on clinical outcomes. The acquisition these mutations appears to be an early event in AML leukemogenesis. Finally, we showed that mtDNA mutations can negatively impact ETC function which we speculate, may in turn drive clonal hematopoiesis by promoting self-renewal and blocking differentiation of HSCs.