Pervasive Positive Selection in Blood in 200,618 Individuals and Novel Drivers of Clonal Hematopoiesis

In the last decade, genomic sequencing of blood samples has revealed that clonal hematopoiesis (CH) is common in elderly subjects and associated with hematological malignancies, cardiovascular disease, and all-cause mortality. Initial estimates identified CH in > 10% of those over the age of 70 when screening for mutations in a known set of canonical CH genes (Bick et al. Nature 2020). However, alternative approaches demonstrate the frequency of CH to be grossly underestimated (Zink et al. Blood 2017). Using single-cell sequencing approaches, dozens of parallel clonal expansions can be found in blood in all individuals by the 7th-8th decade of life, with most expansions lacking known driver mutations (Mitchell et al. Nature 2022). It has been estimated that 70% of driver mutations in blood currently remain unknown (Poon et al. Nat Genet 2021).

The UK Biobank (UKBB) provides a large cohort with which to assess gene fitness effects and their associated health effects (Szustakowski et al. Nat Genet 2021). We exploited the idea that a gene can be identified as under positive selection if one finds an enrichment of non-synonymous mutations compared to neutral synonymous mutations within that gene's coding sequence, and comprehensively examined 200,618 UKBB whole blood exomes for the presence of positive selection leading to clonal expansions. We confirmed the discovery of genes under positive selection using two somatic variant calling approaches. We queried the UKBB database to examine the clinical phenotypes and health outcomes associated with these novel drivers of CH. We further validated our findings in more than 10,000 genomes from single cell-derived hematopoietic cells.

We identified 17 new genes, including lymphoid-expressed genes, under positive selection in blood from 681 variants (VAF > 0.1) among 660 individuals. Notably, 93% of these individuals did not otherwise harbour mutations in canonical CH genes, representing an 18% increase in the cohort of individuals with large clone CH in UKBB (VAF > .0.1). As will be demonstrated in detail, newly identified genes show distinctive mutation landscapes, contribute to both small and large clones in blood, and increase in frequency and size with age - effectively demonstrating characteristics of known drivers of CH. Moreover, across 10,837 single cell-derived clonal hematopoietic cell colonies from 50 individuals (including both those with healthy hematopoiesis and a range of hematological malignancies), we identified 150 somatic mutations across our set of novel CH drivers. Over two-thirds of these genes also demonstrated evidence of positive selection and a strikingly similar pattern of non-synonymous mutations to that seen in the bulk whole exome sequencing data from UKBB. CH driven by mutations in novel genes was associated with increased hazards for hematological malignancies (including lymphoid malignancies), infections, and death, with poor health outcomes influenced by clone size (Figure). Mutations in novel genes conferred some of the strongest fitness effects among all CH genes, corresponding to an excess HSC division rate of 15-20% per year.

We believe the expanded list of genetic drivers of CH now accounts for up to half of the full compendium of mutations capable of driving large clonal expansions in blood. Given their clinical relevance, fitness effects, and prevalence, we propose that these novel genes should now be routinely added to diagnostic gene sequencing panels for blood and included in future studies of CH.

Figure. Clinical outcome associations with novel drivers of clonal hematopoiesis (CH). Cox-proportional hazard ratios for a range of poor health outcomes against different CH categories with 95% confidence intervals of log hazard ratios. Novel genes under selection in UKBB are shown in orange; canonical genes under selection in UKBB are shown in blue; other canonical genes associated with CH but not under positive selection in UKBB are shown in green; COPD, chronic obstructive pulmonary disease; MI, myocardial infarction; heme, hematological.