Human mitochondrial DNA (mtDNA) is a circular molecule of 16,569 bp that contains 37 genes coding for 13 polypeptides of the mitochondrial electron respiratory chain, two ribosomal RNAs, and 22 transfer RNAs (tRNA) required for polypeptide synthesis [1,2]. In comparison to the nuclear genome, mtDNA has a modified genetic code, paucity of introns, and lack of histone protection [3,4]. Moreover, the turnover of mtDNA is high, as degradation and replication is a continuous process in the mitochondria during one cell cycle, and mtDNA polymerase does not have proofreading capabilities. Therefore, the accumulation of somatic mutations is greater in mtDNA compared to that in nuclear DNA [1]. The mtDNA molecule has a noncoding region, which includes a unique displacement loop (D-loop) responsible for replication and transcription control. This control region includes a mtDNA production regulation fraction and a hypervariable (HV) region known as a gene mutation “hot spot,” resulting in mtDNA sequence and length heteroplasmies [5]. The sequence heteroplasmies consist of base substitutions and small deletions and insertions. The length heteroplasmies are usually detected by varying numbers of a particular repeated nucleotide, usually poly-C, known as the mtDNA minisatellite instability (mtMSI) [6]. In conjunction with the control region, the tRNAleucine 1 (tRNA^leu1) and cytochrome b (CYTB) genes in the coding region represent approximately 13% of the total mtDNA sequence, and pathological heteroplasmic mtDNA mutations were observed in these regions of various cancer tissues [6,7,8].

The mitochondrion is an important micro-organelle that produces energy for cell development, differentiation, and growth. It also controls cell growth by inducing apoptosis. The mtDNA is comprised of 0.1% to 1.0% of the total DNA in most mammalian cells, and 2 to 10 copies come packaged in each nucleated cell per mitochondrion up to 1,000 mitochondria. The copy number per cell is controlled within a constant range to fulfill the energy requirement of the cell and to maintain normal physiological functions [9,10]. Variations in the mitochondria copy number have been reported to reflect the net results of gene-environmental interactions, and the copy number has been implicated as a potential biomarker for various cancers [11,12].

Since mtDNA is haploid and lacks recombination, specific mutations in the mtDNA genome leading to human diseases arise in particular genetic backgrounds referred to as haplogroups [13]. Human populations usually carry several mtDNA haplogroups defined by unique sets of mtDNA polymorphisms, reflecting mutations accumulated by a discrete maternal lineage [14,15], but the sets and their frequencies differ between populations. Thus, haplogroup association studies have been used to assess the role of mtDNA variants in various diseases and cancers [16].

Acute myeloid leukemia (AML) results from the accumulation of abnormal blasts in the bone marrow. It is likely that many different mutations, epigenetic aberrations, or abnormalities in microRNA expression can produce the same morphological disease; however, these differences may be responsible for the variable responses to therapy, which is a principal feature of AML [17]. In addition to these nuclear genetic changes, non-chromosomal mitochondrial mutations may play a role in the progression and chemosensitivity of leukemias [18,19,20]. In recent years, mtDNA variants within the D-loop or the entire mitochondrial genome in patients with AML and their prognostic impacts have been reported [21,22]. However, few studies have focused on the mtMSI, copy number, or haplogroups in association with AML.

In this study, we analyzed the control region and two coding regions in bone marrow cells of patients with AML to demonstrate the sequence and length heteroplasmies of mtDNA, alterations in mtDNA copy number, haplogroup distribution, and their impacts on patient outcomes.

Several studies on mtDNA genome variations in acute leukemia have been reported. He et al. [26] compared the entire mitochondrial genome from both normal tissue and leukemic samples from 24 leukemia patients. Grist et al. [19] sequenced the D loop (nt 16111–190) of 48 patients with leukemia. Yao et al. [27] analyzed the mtDNA control region sequence variation in single cells from 18 leukemia patients. Sharawat et al. [21] evaluated the entire D loop region in 44 pediatric patients with AML by direct sequencing. Silkjaer et al. [22] investigated the entire mtDNA in 56 AML patients. Finally, Han et al. [28] analyzed mutations and mtMSI in the mtDNA D-loop region in BM cells of 19 patients with acute leukemia.

In this study, we collected 74 patients with AML and 70 control subjects, and studied not only the mtDNA sequence variants, but also the mtMSI, copy number, and haplogroup. The median number of the mtDNA sequence variant was not different in the patients with AML and control subjects. The changed sequence also showed no significant difference between the two groups. He et al. [26] demonstrated that somatic mtDNA mutations were present in approximately 40% of patients, and they found the A15296G mutation as a leukemia-specific marker. In our study, no participants had a mutation on nucleotide (nt) 15296. Silkjaer et al. [22] reported that 21% of the AML patients harbored the T16311C variant, and this variant tended to be associated with chromosomal abnormalities of favorable prognosis. In this study, we found the T16311C variant in six control subjects (8.5%) and three patients with AML (4%). Among the three patients with AML, two patients had normal karyotypes and one had t(15;17). However, the sample size was too small to analyze for the statistical significance of the relationship and chromosome abnormality. We also found the T489C variant, which appeared in a higher proportion (73%) of patients with AML than in control subjects (50%), but it showed no impact on outcomes. All patients of the haplogroup D4 had the T489C variant.

In the present study, most patients with AML had mtMSI in 303 poly-C and 16189 poly-C (Table 3), and the pattern number of the mtMSI was greater in patients with AML than in control cohorts. These greater heteroplasmy lengths in patients with AML might be associated with the clinical implication. Mitochondrial genomic instability has been strongly correlated with a high incidence of circular mtDNA molecules (of multiple lengths) in human leukemic cells, and a positive correlation has been observed between the frequency of these molecules and the disease severity [29].

Meanwhile, Grist et al. [19] found mtDNA mutations in 36% of AML patients. Yao et al. [27] found that some patients at relapse presented a complex shift in major haplotypes. In our study, the haplogroup D4 was found at a higher frequency in patients with AML compared to that in control subjects, and this difference resulted in an OR of 2.408 for AML (P=0.046). Moreover, in the report by Sharawat et al. [21], some variations in HV-1 were associated with inferior EFS. Silkjaer et al. [22] suggested that alterations in cytochrome c oxidase subunit I and II have an adverse impact on prognosis. In addition, they reported that the T16311C variant tended to be associated with chromosomal abnormalities.

There are also limitations in this study. Since we compared the mtDNA from patients with AML with that from control subjects, there could be the possibility that the sequence changes between the patients with AML and controls may simply reflect extensive sequence variation within the population [30]. As shown in this study, there were no significant differences in mtDNA sequence variation between patients with AML and control subjects. However, even though the variants could not be compared with respect to the potential role in malignancy formation, mtDNA mutations might still contribute to genomic instability.

In conclusion, AML cells revealed more heterogeneous patterns in mtMSI markers and had increased mtDNA copy numbers. These findings implicate mitochondrial genome instability in primary AML cells. In addition, the mtDNA haplogroup D4 might be associated with AML risk among Koreans.

Representative sequencing chromatograms revealing mtDNA mutations in tRNA^leu1 gene (AML No.112 patient) and CYTB gene (AML No.50 patient) as nonsynonymous mutations. (A) mtDNA tRNA^leu1 gene mutation-affected amino acid change (Ala→Thr) for AML patient No.112. (B) mtDNA CYTB gene mutation-affected amino acid change (Ala→Thr) for AML patient No.50.

Gene scan analysis of poly C-stretch region at nucleotide position (np) 303–315, 16184–16193, and 514–515(CA)₅ repeats. (A) The mtDNA D-loop HV2 (303 poly-C) region. Capillary electropherogram of the np 303 poly-C region from AML patient No.123 showed 9CT6C as a C insertion type. (B) The mtDNA D-loop HV1 (16189 poly-C) region. Capillary electropherogram of np 16184–16193 poly-C region from samples of AML No.29 patient also disclosed the profile of several mtDNA types (heteroplasmic mutations). (C) The mtDNA D-loop HV2 514–515(CA)₅ repeats region. Gene scan analysis of CA repeats starting at np 514 (AML No.27 patient) also disclosed heteroplasmic mutations, which suggest the coexistence of wildtype and mutant mtDNA in the same patient.