The accumulation of somatic mutations over time is a widespread phenomenon that occurs across species and in many different organ systems. Whilst most somatic mutations have no functional consequence, some mutations can lead to altered cellular behavior. For example, stem cells that acquire somatic mutations affecting the function of certain genes can acquire a fitness advantage over their counterparts. Over time, this can lead to an expansion of clones carrying these “driver” mutations, a phenomenon first identified in hematopoiesis but now recognized to occur in several different organ systems. In hematopoiesis, this phenomenon is known as clonal hematopoiesis (CH), with mutations in a relatively small number of genes, including epigenetic regulators (DNMT3A, TET2, ASXL1) and signalling genes (JAK2) accounting for most cases of CH. As these mutations confer only a modest fitness advantage to hematopoietic stem cells (HSCs), the clonal expansion is a slow process that likely occurs over decades, and consequently the prevalence of CH is much higher in older individuals. Furthermore, several lines of evidence also support that many elderly individuals have evidence of CH, which is not associated with presence of a canonical driver mutation, making this phenomenon harder to identify and study.
Presence of somatic mutations provides a unique signature for each HSC that can be used as a powerful tool to study lineage relationships between HSCs. Dr. Emily Mitchell and colleagues exploited this phenomenon to explore the relationship between aging-associated changes in hematopoiesis, clonal HSC expansion, and somatic mutation. The investigators carried out whole-genome sequencing of an impressive 3,579 genomes of single-cell derived colonies from phenotypic HSCs (Lin-, CD34+, CD38-, CD45RA-) across 10 different healthy individuals of different ages from birth to 81 years. Accumulation of mutations occurred in a linear fashion with approximately 17 substitution mutations/HSCs per year after birth, with approximately 0.12 non-synonymous mutations/HSCs per year and approximately 0.7 indels HSCs per year. Telomere length steadily declined with age at a rate of 30bp per year with considerable cell-to-cell heterogeneity in younger adults and some HSCs having unexpectedly long telomeres.
These data allowed reconstruction of phylogenetic trees for each individual. There was a sparsity of branch points at the tips of the phylogenetic trees, representing 10 to 15 years of “molecular time,” which the authors speculate might reflect hematopoietic contribution of short-term HSCs. Hematopoiesis was largely polyclonal in adult persons younger than 65 years, with an estimated 20,000 to 200,000 HSCs contributing to blood cell production. Expanded HSC clones were rare in adults younger than 65 years. In contrast, there was an abrupt and striking reduction in clonal diversity in all three older (>75 years) individuals studied, with a substantial proportion of hematopoiesis derived from a small number of HSC clones (12-18), each typically contributing between 1 percent and 3 percent of HSC-derived colonies sequenced, with some less frequent larger clones. These HSC clones had expanded over decades and, strikingly, only one in five expanded HSC clones carried known CH-associated driver mutations. It was estimated that between one in 34 and one in 12 non-synonymous mutations were drivers, which accumulated at a constant rate in HSCs throughout life. Furthermore, loss of Y chromosome (a phenomenon frequently observed when carrying out karyotype analysis of bone marrow from older men) was also shown to confer a fitness advantage to HSCs. Mathematical models supported that the abrupt switch to oligoclonal hematopoiesis was likely driven by this gradual and inexorable acquisition of driver mutations throughout life, which individually drive only a modest fitness advantage. It takes many decades for substantial clonal selection to occur and consequently this only becomes apparent during old age. Three genes, DNMT3A, ZNF318, and HIST2H3B, were identified as key drivers of this positive selection; all were acquired many decades before emergence of CH. However, the authors also note that the estimated actual number of driver mutations per individual was likely to be far higher than the number of mutations detected in known cancer genes, supporting the observation that other mutated genes or other processes might be important drivers of clonal selection.
In Brief
This study proposes that CH is likely to be a universal and inevitable phenomenon that occurs with aging due to the gradual accumulation of molecular damage by HSCs. Aging is also associated with other alterations in hematopoiesis with an increase in myeloid and platelet-biased HSCs with reduced regenerative potential and increased risk of development of blood cancers; it is also likely to be closely related to the underlying decline in clonal diversity. Canonical CH-associated mutations only explain a minority of the genetic events that gradually accumulate and drive these clonal expansions. Whether heritable epigenetic changes also drive these clonal expansions will be an important area of study, requiring multi-omic approaches at the single-cell or clonal level. This beautiful work provides a paradigm for the future study of how germline genetics, lifestyle, disease-associated, and other factors such as immune cell responses might contribute to or mitigate the development of CH. Applying this approach to large patient cohorts will remain a major challenge, likely to be overcome as technology advances. This will allow correlation of the decline in clonal diversity in hematopoiesis and associated driver events with risk of disease development and overall mortality. The authors speculate that the abrupt impact of this gradual accumulation of somatic mutations and molecular damage, which seems likely to occur across all organ systems, might have profound and far-reaching consequences, ultimately contributing to the sharp increase in mortality rates in humans older than 70 years.
Competing Interests
Dr. Mead indicated no relevant conflicts of interest.
