Imatinib-Sensitive Chronic Myeloid Leukaemia Is Associated with mtDNA Mutations and Reduced Oxphos Capacity

Introduction: Therapy resistance driven by Leukemic Stem/Progenitor Cells (LSPCs) remains a challenge in Chronic Myeloid Leukemia (CML). Recent studies indicate that Tyrosine Kinase Inhibitor-resistant (TKI-R) cells are more reliant on mitochondrial (mt) Oxidative Phosphorylation (OXPHOS) for survival, highlighting a potential new therapeutic target. Mutations in OXPHOS-related genes are potentially deleterious and can be associated with reduced OXPHOS and activation of apoptosis. We hypothesize that TKI-sensitive (TKI-S) and TKI-R patients may exhibit distinct OXPHOS phenotypes, possibly involving mtDNA mutations.

Methods: TKI-S patients were defined based on the achievement of Major Molecular Response at 12 months of TKI therapy (MMR, BCR::ABL1<=0.1%), and TKI-R included patients who did not achieve MMR at 12 months or developed blast crisis. CD34+ LSPCs cells were isolated from blood of chronic phase (CP) CML patients at diagnosis. Samples were cultured in the presence of physiological growth factors. Mt respiration (OXPHOS) was measured using the MitoStress Test kit on the Seahorse XF96 analyser (4 TKI-S, 5 TKI-R). Mt content was assessed by Mitotracker Green using flow cytometry. MitoTrace software was used to call mtDNA variants from 10X Single Cell (sc) 5' RNA sequencing data (Krishnan V et al Blood 2023) and LoFreq software to call somatic mutations in purified mtDNA from total leukocytes.

Results: TKI-R CD34+ cells showed increased OXPHOS capacity compared with TKI-S, including increased basal oxygen consumption rate (median 82.5 vs 52 pmol/min/10⁵ cells, p=0.043) and increased maximal respiration (median 209.9 vs 115.9 pmol/min/10⁵ cells, p=0.0032). The spare respiratory capacity ("mt fitness") of the TKI-R CD34+ cells was also increased, whilst TKI-S cells showed mt dysfunction (median 126.5 vs 54.9 pmol/min/10⁵ cells, p= 0.012). There was no statistically significant difference in the mt content of CD34+ cells between TKI-R and TKI-S patients (202% vs 151%, normalised on normal CD34+ cells), suggesting that the respiratory deficiency observed in the TKI-S cells might be linked to mt function, rather than decreased mt number.

To investigate whether mtDNA mutations were common in CML and associated with better response to TKI, we sequenced the mtDNA of purified mitochondria from total leukocytes of 80 CP-CML patients. We identified 187 mtDNA mutations in 60 (75%) CP-CML patients (median 2 per patient, range 1-21). TKI-S patients (41/80) showed more mutations at higher variant allele frequency (VAF) than TKI-R (139 vs 48 with median VAF 21% vs 7%, p<0.0001), including non-synonymous protein-coding mutations (49 vs 24 mutations at median VAF 13% vs 6%, p=0.0008). Patients with >=3 mtDNA mutations at diagnosis (n=21/61) had a higher probability of MMR (90% vs 68%, p=0.0038) and MR4.5 (BCR::ABL1<0.0032%; 63% vs 37%, p=0.011) at 24 months of imatinib treatment. A higher frequency of non-synonymous mutations was identified in the MT-COI gene, which is relevant for the assembly and function of the OXPHOS machinery. Mutations in MT-COI were observed in 15 patients. They were associated with better response to imatinib at 24 months, including better MMR (92% vs 67%, p=0.026) and MR4.5 (69% vs 43%, p=0.0413). We linked mtDNA mutation VAF and type with the available functional data, confirming that patients (4 TKI-S) carrying mutations with VAF>7% also had dysfunctional OXPHOS, whilst patients (5 TKI-R) with no mutations or mutations with VAF<=7% had higher mt function (p=0.0079).

We next developed a bioinformatic pipeline to call mtDNA variants in LSPCs from scRNA-sequencing data. We counted the number of mtDNA variants/cell across different cell types and calculated the ratio of number of mtDNA variants in LSPCs relative to T/NK cells (the latter serving as an internal control). Our data showed that TKI-S (8 patients) had a higher ratio of cells carrying non-synonymous mtDNA mutations than TKI-R (12 patients; 3.1 vs 2.2 p=0.02), supporting the data observed in total leukocytes.

Conclusions: This is the first study linking mtDNA mutations to OXPHOS activity and TKI response in CML. Non-synonymous mtDNA mutations may lead to decreased OXPHOS activity and enhanced response to TKI therapy. Our findings suggest that the strategy of pharmacological inhibition of OXPHOS could be exploited to sensitize LSPCs to TKI therapy and mtDNA mutations could be useful as a biomarker of response.