Commentary: Somatic Cell Evolution: How To Improve With Age

Abstract

A recent article published in this journal illuminates a rare example of somatic evolution where cells improve rather than deteriorate with age. In mitotic intestinal crypts, stem cells with higher levels of a deleterious heteroplasmic germline mitochondrial mutation are purged through time, leading to crypts without the mutation. Similar somatic mitochondrial mutations are not purged from crypts, indicating that special conditions are needed to improve with age.

The article by Su et al [1] documents an interesting example of somatic cell evolution, where a deleterious germline heteroplasmic mitochondrial mutation is frequently lost in mitotic intestinal epithelium but not in adjacent smooth muscle cells. Essentially the epithelium “improves” with age as mutant mitochondrial genomes are purged. Somatic cell evolution is technically difficult to document in normal human tissues, but it is increasingly apparent that mitotic human cells accumulate large numbers of somatic mutations with aging [2–4]. The vast majority of these somatic mutations in normal epithelial tissues appear to be neutral (no apparent selective value), and the current example is a rare case where cells improve with age.

The dynamics of this example are interesting because it illustrates evolution at the organelle, cell, and tissue levels (Fig 1). Each cell contains multiple mitochondria and each mitochondrion contains multiple small circular genomes. At birth, the mitochondrial genomes are a mixture of wild type and mutant genomes (heteroplasmy). The ratios of mutant to reference genomes can change with time because they replicate within a cell and can be partitioned unequally with cell division [5,6]. Much of this segregation can be considered random drift or neutral because of a threshold effect, where a mutant genome does not confer a change in fitness to its cell until it becomes the majority of the genomes (about >80%). The primary mutation in this study is a point mutation in a tRNA (tRNA^Leu(UUR), m.3243A>G), which causes mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS) syndrome and leads to widespread defects in mitochondrial protein synthesis and marked reductions in oxidative phosphorylation once a threshold is exceeded.

Figure.

A cell contains multiple mitochondria that contain multiple circular genomes. A mitochondrial mutation (red “X”) is initially present in only some of the genomes (homoplasmy). At the cell and organelle level, with mitochondrial genome replication, the proportions of mutant genomes can change or drift to higher or lower levels. A defective cell phenotype (red cells) is typically expressed when mutant genomes exceed a threshold. At the tissue level, intestinal crypts contain multiple mitotic stem cells at their bases that randomly turnover through a niche mechanism or neutral drift. The germline mitochondrial mutation m.13243A>G confers strong deleterious defects in oxidative phosphorylation (ox-phos) and therefore stem cells with high levels of the mutation are lost due to negative selection as they are replaced by neighboring stem cells. This loss of germline mutant alleles results in greater tissue fitness with age. However, somatic mitochrondrial mutations that also cause defects in ox-phos often become fixed in crypts [12], indicating that drift can fix even deleterious mutations. The somatic mitochondrial mutations appear to confer milder ox-phos defects [13], indicating that efficient somatic cell selection requires relatively large changes in phenotype.

Evolutionary forces come into play when more offspring are produced than can survive. In non-mitotic muscle cells and neurons, drift within cells can lead to the mutant genomes rising above the threshold, resulting in the major clinical manifestations of MELAS of neuropathy and myopathy in adults [7]. Drift within individual mesenchymal cells can also lead to complete loss of the mutant genomes [8], but negative selection and replacement of the affected neurons and muscle cells does not readily occur in these tissues because of the absence of cell division. Hence, individual cells can improve but these cells do not spread to improve the tissue.

By contrast, in intestinal crypts, stem cells are mitotic and are maintained by a niche mechanism where individual stem cell lineages may either be lost or expand. Most of the time, this stem cell lineage turnover is random or neutral [9,10], and the fates of any neutral or passenger mutations depend on the random survival of their stem cells. However, both positive and negative selection are possible if specific driver mutations alter the fitness of their stem cells. The data in the article by Su et al [1] illustrate that epithelial stem cells that exceed a threshold and have severe oxidative phosphorylation defects likely suffer negative selection and are displaced from the niche by stem cells with lower mutant levels, because entire adult intestinal crypts have lower average levels of the mutant genomes and were generally below the threshold compared to the non-dividing muscle cells. Reductions in the mutant genomes are commonly seen in the blood [11], indicating negative selection also occurs between mitotic stem cells in hematopoietic niches.

These investigators have also studied somatically acquired mitochondrial mutations that similarly lead to oxidative phosphorylation defects. However, in contrast to the germline MELAS mutation, they have inferred that most somatic mitochondrial mutations are effectively neutral and accumulate with age, and therefore can be used as fate markers to study human niche stem cell dynamics [12–15]. This contrast between the germline and somatic mitochondrial mutations further illustrates the blurry probabilistic boundaries between drift and selection.

In evolution “shorthand”, a cell with a positive selective advantage should dominate and a cell with a negative selective advantage should be lost. Taken to the extreme, almost every non-synonymous mutation should result in a change in cell phenotype, with selection able to efficiently discriminate even very small differences in fitness. However, drift is also a powerful evolutionary force, especially within small populations like intestinal crypts, where competition occurs between very limited numbers of neighboring stem cells [16]. Crypt mutations can also be lost or fixed by drift, and for most mutations it is very difficult to distinguish whether their fates depend on chance or selection. It appears that only mutations with very large changes in phenotype efficiently confer detectable somatic cell selection. The germline MELAS mutation in the tRNA likely causes more profound oxidative phosphorylation defects relative to the somatic mutations, that tend to alter single mitochondrial genes [13]. Therefore, negative selection is observed with the germline MELAS mutation, whereas most other somatic mitochondrial mutations may cause less severe defects and therefore are effectively neutral and tend to drift to fixation or loss in normal human crypts.

The accumulation of multiple mutations in a single cell can also lead to neoplasia. Interestingly, even driver mutations with very strong positive selection (Apc, Kras) also contend with drift because they are frequently lost in mouse intestinal crypts [17], further illustrating a probabilistic rather than an absolute efficiency of somatic selection. For example, a single crypt stem cell with a sporadic Apc heterozygous mutation is not fixed 100% of time (deterministic) but rather only about 60% of the time (probabilistic).

The average person is born with hundreds of potentially deleterious mutations [18]. Most of these mutations are heterozygous and potentially we could improve with age by replacing the mutant allele with the wild type or reference allele though mitotic recombination. A recent study [19] of copy number alterations in human blood detected several examples where a rare heterozygous germline variant is either lost or becomes homozygous by mitotic recombination. For example, a presumably defective MPL protein-truncating variant is commonly replaced by the wild type allele. However, examples of loss of the mutant or reference allele were both observed, indicating the unevenness of the direction. In addition, this selection was inefficient and typically occurred in less than 5% of the blood cells. The skin also shows evidence of mitotic recombination where the mutant allele is often replaced by the reference allele in ichthyosis [20].

In summary, the study by Su et al [1] is another example where a defective mutant allele can be replaced by the reference allele. Such examples are still relatively rare and illustrate some of the ingredients needed for such somatic cell evolution---a mutation that changes cell phenotype and confers strong selection, a mitotic microenvironment where competing cells may replace each other, and sufficient time to allow detection. How much of a phenotypic difference is needed to confer detectable somatic cell selection is uncertain, but a recent study of human tumors estimated the fitness advantage must be greater than 20% [21]. Most mutations are either neutral or deleterious, and therefore although many somatic mutations accumulate with aging in mitotic tissues [2–4], we unfortunately seldom improve with age.