1. Introduction

Mitochondrial DNA (mtDNA) sequence differences within an individual are referred to as heteroplasmy. It manifests itself as a mixture of mtDNA sequences. These sequence differences can be present between tissue types [1, 2] or within a single mitochondrion. The length of a particular sequence may vary [3-5] or differences could exist at a single point or as a mixture of point mutations within a sequence [6-8]. The nature of the mitochondria and its rapid genetic turnover yields a high rate of somatic mutations [9-11].

In the past, detection of mitochondrial heteroplasmy was achieved by denaturing gradient gel electrophoresis, restriction fragment length polymorphisms, Southern blotting, polymerase chain reactions, or other electrophoretic methods [12-20]. Advancements in sequencing methods beyond the traditional Sanger method have made it possible to gain rapidly vast amounts of data with varying amounts of reliability and accuracy. The reported degree of heteroplasmy is mostly due to overestimation of the phenomenon. Most of the over reporting is caused by differences in sensitivity of the Sanger method and the massively parallel sequencing methods [21]. Recently, Just, Irwin [21] provided an important framework for future studies and implored those investigators engaging in this type of research to validate their findings. The future studies they suggested should provide better estimations of position specific mutation rates and determine the probability of heteroplasmy related to mutation rates in the coding region. Technical and interpretative problems exist leading to difficulties distinguishing between heteroplasmy and sequencing errors [22]. However, even a low level of heteroplasmy appears to be important to human health. Besides determining which sequences might lead to pathologies, the number of copies of a mutant sequence leading to illness must be determined [22]. Determining which changes are pathogenic should be a major goal of those researchers in the molecular medicine field. Resources need to be focused on methods to correct or prevent deleterious changes to mtDNA.

2. Methodology Issues

Levels of heteroplasmy reported in the literature vary greatly [23, 24]. Most of the variability is attributed to the established threshold for mtDNA mutation detection. Sequences are compared to a reference genome. The degree of difference between the reference sequence and the investigated sequence is measured and evaluated. The selected threshold is usually a chosen minor allele frequency (MAF) that must be exceeded to declare a sequence to be different. Despite some studies using the same MAF, there are still reported differences in levels of heteroplasmy. This discrepancy may be due to contamination [25, 21]. Discernment between homoplasmic mixtures of individual cells with multiple haplotypes and heteroplasmy within or between cells is extremely difficult [26]. Furthermore, contamination can further complicate the task. Additional differences in levels could be due to data handling mistakes, such as improper sequence alignment [27] and poor data interpretation practices [28]. For example, reports showing high levels of heteroplasmy have been generated by improper use of databases as well as failure to completely search all possible previous reports of so-called novel sequences [28]. Additionally, careful exclusion of possible contamination of sequences with nuclear mitochondrial DNA (NUMT) [29] also may occur. NUMT can be accounted for by scrupulous bioinformatic practices or by proper purification of mitochondria before mtDNA sequencing.

Alignment issues are more likely to occur when length heteroplasmy is present. Chemistry based sources of error in heteroplasmy measurements will continually be minimized as the techniques are improved. Bioinformatic approaches must be carefully vetted before the data they generate are reported [30]. Few standardized methods exist for mtDNA sequence data analysis [31]. Currently, there are no uniform ways of interpreting or reporting length heteroplasmy data [32]. As a result, it is imperative that a specialist is consulted or employed when length heteroplasmy is observed. The suggestion posed by Just, Irwin [21] regarding the creation of a catalog of regions where length heteroplasmy should be followed.

3. Implications

Mutations are likely caused by chronic alcohol consumption [33, 34], cigarette smoking [33, 35], other external environmental insults to the mtDNA, and reactive oxygen species generated by mitochondria. Mutagens could be causing damage at a particular location within the mtDNA. Point mutations that repeatedly occur at the same location within the mtDNA sequence, hotspots, exist in the control region (CR) [6, 36, 37, 5]. Coding region hotspots are rare [38, 39]. There is a need for reliable data to support the existence of this type of hotspot [21]. Many CR hotspots occur in areas that initiate the replication of mitochondria [2, 40, 41], thereby leading to a selective propagation of these mutations. When somatic mutation lead to reduced bioenergetic capacity of the mitochondria the compensatory reaction of the cell is to produce more mitochondria, thus increasing the level of the mutated mtDNA sequences [42]. Consequently, age-related diseases of those organs are observed [43]. Impairment of insulin secretion and reduced insulin sensitivity have been demonstrated to be maternally inherited and associated with tRNA mutations [44, 45]. Atherosclerosis and cardiovascular disease are major complications of untreated type 2 diabetes [46]. Oxidative stress is a hallmark of type 2 diabetes and related metabolic diseases and represents a significant risk factor for atherosclerosis [47]. Recently, evidence for the development of atherosclerosis due to somatic mitochondrial mutations has been presented [48-50]. Psychiatric diseases may result from somatic mutations leading to heteroplasmy [51]. Cancer is the most studied of the diseases that are thought to be caused by mtDNA heteroplasmy [52, 53]. Carcinogenic processes might target normal mitochondrial functioning and cause a disruption of the Krebs cycle and electron transport enzymes [54, 11, 55]. Furthermore, normative mitochondrial function in non-proliferating cells affects relatively high cytosolic ATP/ADP ratios, resulting in functional inhibition of aerobic glycolysis [56]. In contrast, the bioenergetics of the “Warburg” effect that has is linked to the metabolic phenotype of numerous cancer cell types is characterized by enhanced aerobic glycolysis and suppression of mitochondrial aerobic metabolism [57, 58, 54, 59]. Furthermore, aerobic respiration in proliferating cells may lead to the deleterious production of free radicals that can mutate mtDNA. Accordingly, free radical damage could exacerbate compromised mitochondrial functioning and subsequently increase the copy number of mutant mitochondria in cancer cells.

4. Conclusions

The growing body of evidence supports mtDNA mutation involvement in organ-specific disorders and, therefore, they don’t emerge randomly [1, 7, 42, 15, 60, 2]. There are also cases of different locations within an organ having varying degrees of heteroplasmy [61, 49]. The non-random nature of many mtDNA somatic mutations could be explained by looking at the causes of the mutations. In the case discussed, in response to a paper by Bandelt and Salas [25], Prior and colleagues defended their data that demonstrated that the carcinogens in cigarette smoke can mutate mtDNA at specific locations [35]. While some of the arguments and suggestions presented by Bandelt and Salas [25] are valid, an overwhelming amount of data supports that many detected mtDNA mutations are not random [60, 2, 40] and are not due to technical errors or contamination [7]. Once again the focus should be shifted away from arguing about techniques but rather towards understanding why mutations occur at hotspots. We also need to continue to learn which selective processes allow heteroplasmy to exist.

ACKNOWLEDGEMENTS

Thank you to George B. Stefano and Richard M. Kream for their helpful discussions.