Similarities between the processes leading to the formation of new species and new diseases are amongst the first steps towards the justification of applying basic concepts of species evolution to investigate the genetic basis of complex disorders, but also vice versa.3 Importantly, evolutionary (Darwinian) medicine is not offered as an alternative to the old medical inquiry, but rather as a novel vantage point for biomedical phenomena.7
In 2005, Douglas C. Wallace pin-pointed the mitochondria and mitochondrial genetics as reflecting the very center of evolutionary medicine.8 Indeed, constituting a major player in cellular and organism metabolism the mitochondrion is a suitable candidate to respond to changing environments not only in the past but in modern times as well to raise the susceptibility to many complex disorders. This hypothesis received support from James Neel’s idea proposing 40 years ago the involvement of “thrifty genotypes” that were successful in ancient times during conditions of calorie restriction in the emergence of metabolic disorders today.9 Accordingly, a number of research groups, including our own, have demonstrated the association of ancient common mitochondrial DNA (mtDNA) genetic backgrounds with altered susceptibility to diabetes and its complications.10–12 Other complex and age-related disorders were also identified as being associated with mtDNA variation (recently reviewed).13 Not only correlative evidence supports this line of thinking but also experimental findings establishing the functionality of certain human mitochondrial genetic variants, thus revealing them to serve as the “radar” of natural selection.14,15 Therefore, apparently, certain positively selected functional mutations in our phylogenetic history today play a role in disease susceptibility.
The association of ancient genetic variants with disease susceptibility is not unique to the mitochondria but is common to all disease association studies, which are based on the CDCV hypothesis. The uniqueness of mitochondrial involvement in complex disorders stems mainly from the higher magnitude of mutation accumulation in the mtDNA compared to the nuclear DNA. Obviously, this fact results in increased genetic variability due to high fixation rate of mutations thus generating a large mutational repertoire to be sifted through by natural selection. Moreover, mtDNA-encoded factors are in close contact with nuclear DNA-encoded elements, especially within the oxidative phosphorylation and mitochondrial protein translation systems. This epistatic relationship, frequently termed cytonuclear interactions, is directly affected by the large difference in mutation fixation rates of the two genomes, which leads to tight co-evolution of mtDNA and nuclear DNA-encoded factors.16 Thus, cytonuclear interactions were implied to play a major role in adaptive and other evolutionary processes2,16–18 as well as in diseases.19 However, the rapid occurrence rate of mtDNA mutations also results in an increased repertoire of mutated mtDNAs inside the cell during the individual’s lifetime, thus further diversifying the mitochondrial genetic repertoire per cell (heteroplasmy). Hence, mitochondrial genetics is not only affected by its maternal mode of inheritance and high rate of mutation fixation during evolution but also by “intracellular” population genetics.
Heteroplasmy is a known phenomenon in mitochondrial genetics, and different levels of heteroplasmy correlate with disease severity and penetrance.20 Mixed populations of mtDNA molecules could be inherited from the maternal line, though its intracellular variability is thought to be bottleneck-controlled during the maternal germ-line formation,21,22 a mechanism that has recently been challenged.23 In contrast, heteroplasmy due to mutation accumulation during the individual’s lifetime has been supported by multiple lines of evidence, and its contribution to age-related disorders has been highlighted.8 Moreover, mutations may accumulate even faster in certain mitochondrial diseases in which the mtDNA replication and repair mechanisms are impaired24 and in various types of cancer.25,26 Both in the impaired mtDNA repair/replication diseases and in cancer the repertoire of heteroplasmic mutations is expected to be increased.24 This is when natural selection is engaged. We showed that de novo mutational combinations that became fixed in cancer tissues (head and neck squamous cell carcinoma and pancreatic cancer) strikingly tend to recapitulate ancient mtDNA mutational combinations that define established genetic backgrounds – haplogroups.27 Recent deep sequencing (massive parallel sequencing) of whole mtDNAs from colorectal cancer and normal adjacent tissues from 10 different individuals revealed clear differences in the repertoire of under-represented (heteroplasmic) mutations in the normal versus the disease tissue.28 Alternatively, a close inspection of the published heteroplasmic mutation list per individual in the last-mentioned study drew our attention to apparent notable recurrent representations of mutations that recapitulate known fixed common mtDNA variants such as those in nucleotide positions 16126, 4216 (both of which associate with mtDNA haplogroups J and T), and position 72 (which associates with mtDNA haplogroup V) (supplementary table 6 in He et al.28). In that case not only do the principles of evolution apply to the study of complex disorders such as cancer, but the very same mutations could play a role in both malignant and normal evolutionary processes. Although based on the assembly of multiple short (∼50 bp) sequence reads, next-generation sequencing methods provide a high resolution for the inspection of intracellular populations of molecules thus enabling the identification of relatively rare mutations which were previously invisible. Moreover it sets the basis to investigate the process of mutational fixation at the cellular and individual levels prior to their fixation in the species population. This will enable not only the assessment of the roles of natural selection and genetic drift in the mutations fixation process at the cellular level but will also pave the path towards investigating the origin of mitochondrial disease-causing mutations, many of which remain in the heteroplasmic state.
The elevated mutation (fixation) rate in cancer and certain mitochondrial diseases raises the question of the evolutionary advantage of the already high mtDNA mutation rate in healthy conditions. Above I argued that one of the pillars of the evolutionary theory is the continuous formation of genetic variability. Being the most variable coding region in the human genome, the mtDNA was thought to play a role in major evolutionary processes.2 The increased mutation rates in mitochondrial diseases and cancer lead me to hypothesize that the mtDNA mutation rate has a threshold beyond which the capability of the mitochondria to adapt and retain normal activities might be adversely affected. When such a putative threshold is crossed, energy metabolism is affected thus leading either to metabolic disorders, cancer, or aging.29,30 Alternatively, in cancer cells it is possible that the malignant increased mtDNA mutation rate could be part of an adaptive process thus creating novel variants in a rate high enough to allow the accumulation of a large somatic variation and response to the strong selective constraints within the lifetime of a single individual. This prediction will be testable when the cancer de novo mtDNA mutation data sets expand significantly as more mtDNA sequences from cancer and corresponding normal tissues are obtained.
Thus far ∼100 disease-causing mutations were described in the mitochondrial genome and, as mentioned above, the pathological phenotype of which occurs at various levels of heteroplasmy.31 Recently the 3243 A>G mutation causing myoclonic epilepsy and stroke-like episodes (MELAS) was found in low concentration in a notable portion of Caucasians,32 thus raising the possibility that these mutations are formed multiple times but only occasionally reach levels sufficient to cause a phenotype. Is the change in the level of heteroplasmy attributed to random division of the cytoplasm during cell division, i.e. intracellular genetic drift (replicative segregation), or is natural selection involved? The more next-generation sequencing technologies evolve, the more population data could be gathered – thus paving the path towards the construction of a comprehensive map of positions prone to mutagenesis and their tendency to undergo mutation fixation.
Since the repertoire of heteroplasmic mutations varies among different tissues28 (Buchshtav M. et al., in preparation) another dimension is added: tissue specificity. Differences in the proportion of heteroplasmic mutations could distinguish dividing tissues versus post-mitotic cells, such as blood versus muscle, respectively.33,34 Since some mitochondrial diseases exhibit tissue-specific phenotypes, such as visual loss in Leber’s Hereditary Optic Neuropathy (LHON), and since many maternally inherited diseases are caused by mtDNA mutations in a heteroplasmic state, great importance underlies the understanding of the mechanism leading to the formation of such mutations and the principles governing their occasional fixation in the mitochondrial population of different tissues. Next-generation sequencing of whole genomes such as currently generated by the 1000 Genome Project (www.1000genomes.org) and the Cancer Genome Atlas (cancergenome.nih.gov) will provide an indispensable view of the individual and tissue-specific mutational landscape and will pave the path to analysis and the generation of predictions for the functional importance and phenotype future impact of rare and common mutations. As the sequence information generated by next-generation technologies increases, our ability to assess the role of evolutionary principles in diseases becomes clearer.