CLONAL HEMATOPOIESIS AND ITS IMPACT ON CARDIOVASCULAR DISEASE

INTRODUCTION

Traditional risk factors of cardiovascular disease (CVD) include hypertension, diabetes mellitus, hyperlipidemia, and smoking. However, the completeness of their predictive value have been questioned by numerous studies^1,2. In light of this, the occurrence of CVD increases exponentially with age, which is traditionally viewed as a non-modifiable risk factor³. Although aging is the major risk factor for CVD, we have poor understanding of how it promotes disease progression. These considerations lead to the question, “Are there age-associated, as-yet-unidentified risk factors for CVD?”

Somatic DNA mutations accumulate in all cells with age such that over time tissues become a mosaic of cells with slightly different genotypes^4–7. In proliferating tissues, mutations in “driver” genes, that provide a selective advantage or “fitness” to cells, can lead to the clonal expansion of these single cells. Previously, there was little information about the role that mutant cell clonal expansion has on age-associated chronic disorders. However, a number of recent epidemiological studies have associated the clonal expansion of hematopoietic cells, a process referred to as clonal hematopoiesis, with increased mortality^8–11. Clonal hematopoiesis increases in the risk of a hematological cancer, but this overall risk is small and cannot account for the increase in mortality in the general population. Surprisingly, these mutations have also been associated with higher rates of CVD^8,12, suggesting a previously unrecognized link between somatic mutations in hematopoietic cells and chronic disease. Thus, a series of experimental studies have been performed to assess the potential causal connection between mutations in hematopoietic system and CVD, and define the mechanisms involved in these processes^12–15. Here, we review recent epidemiological and experimental studies on clonal hematopoiesis that relate to CVD.

Clonal Hematopoiesis: A common condition in the elderly population.

Human bone marrow produces billions of mature blood cells every day. The traditional view is that the progeny blood cells are derived from 10,000 to 20,000 hematopoietic stem and progenitor cells (HSPC)^16,17. Typically, HSPC are equally capable of producing all of the differentiated blood cells, thereby maintaining the healthy state of polyclonal hematopoiesis. In contrast, clonal hematopoiesis is a condition in which substantial proportion of mature blood cells are derived from single dominant HSPC¹⁸. Presumably, this condition arises because HSPC and their immediate progenitors are subjected to Darwinian selective pressures within the stem cell niche, favoring the expansion of clones with somatic mutations that provide a competitive advantaged. Clonal hematopoiesis was initially detected as the skewing of X chromosome inactivation in the white blood cells of women. In one study, 37.9% women over 60 years old displayed skewed inactivation of the X chromosome, while little or no skewing was observed in younger women¹⁹. A notable advance was made when the skewing of X chromosome inactivation was attributed to somatic mutations in the epigenetic regulator TET2²⁰. These data provided the first evidence that clonal hematopoiesis can result from somatic mutations in preleukemic or myelodysplastic syndrome (MDS) “driver” genes. Consistent with these findings, single nucleotide polymorphism (SNP) array analyses detect somatic clonal mosaicism in peripheral blood of healthy individuals, and the frequency of this mosaicism steeply increased in individuals past the age of 50 years old^21,22. More recent studies analyzed exome sequences of more than 32,000 cancer free individuals and found high frequencies of clonal expansion events that were associated with mutations in numerous driver gene candidates that are recurrently mutated in hematologic cancers. The frequency of these somatic mutations increased with age and were most commonly detected included the epigenetic regulators DNMT3A, TET2 and ASXL1^8–10. Despite the clonal expansion of peripheral blood leukocytes with one these mutations, and less frequently two or three somatic mutations, these individuals do not exhibit significant differences in white-cell counts, hemoglobin levels or platelet counts. In this study, men exhibited a higher incidence of clonal hematopoiesis, which is consistent with the difference in the frequency of MDS between genders. Hispanics had lower frequency of clonal hematopoiesis compared to those of European ancestry. Given these considerations, it would be of interest to evaluate the prevalence of clonal hematopoiesis specifically in the Japanese population, which has the second longest life expectancy worldwide (81 years for men and 87 years for women).

Multiple studies have shown that somatic mutation-mediated clonal hematopoiesis is frequent in the elderly populations. In one study, it was found that somatic mutations in driver genes occurred in 10% of individuals over 65 years old and >20% of individuals over 90 years old using methods with a variant allele fraction (VAF) limit of detection of 3.5% and 7.0% for single nucleotide variants (SNV) and small insertions and deletions (indels), respectively⁸. Clonal Hematopoiesis of Indeterminate Potential (CHIP) is a term that is used to describe individuals who harbor a hematologic malignancy-associated somatic mutation in a driver gene that is present at a variant allele fraction (VAF) of 2% or greater in bone marrow of white blood cells, yet the individual does not meet the diagnostic criteria for any detectable hematologic malignancy. This distinction, while convenient, is largely based upon limits of mutation detection by the sequencing methodology rather than being based upon epidemiological data or biological principles. In this regard, the estimated frequency of clonal hematopoiesis in the population doubles when it is deduced from the analysis of “passenger” gene expansion in whole exome sequencing analyses¹⁰. Even more striking, whole genome sequencing analysis of mosaic somatic mutations in peripheral blood cells reveals even higher levels of clonal hematopoiesis in the general population¹¹. In this study, clonal hematopoiesis could be observed in more than 50% of individuals older than 85 years, but only a small proportion could be attributed to a known driver gene candidate or could be detected by non-biased whole exome analysis. Collectively, these studies highlight that genomic instability is prevalent in the hematopoietic cells of elderly individuals, and that the estimated frequencies of clonal hematopoiesis in the population is dependent upon the methodology that is used to detect this condition.

THE ETIOLOGY OF CLONAL HEMATOPOIESIS

In describing the etiology of clonal hematopoiesis, the competition among clones can be viewed as a game of chance among gamblers²³. If the odds favor some players they will become winners while other players will go bankrupt. With regard to clonal hematopoiesis, the odds of “winning” versus “losing” are influenced by many factors including the presence of driver gene mutations, alterations in the bone marrow microenvironment, cis-acting heritable loci, etc. Based upon current knowledge, clonal hematopoiesis can be divided into several categories based on the characteristics of their etiology. We will briefly discuss 3 broad categories: age-associated clonal hematopoiesis, therapy-associated clonal hematopoiesis and neutral drift.

Age-associated clonal hematopoiesis -

As noted previously, numerous studies have shown that there is a strong correlation between age and clonal hematopoiesis. As discussed, an analysis of clonal mutations in 160 candidate driver genes revealed that clonal hematopoiesis occurs in 10% of individuals that are 70 years of age when the sequencing methodology has a detection level of >3.5% for SNVs and >7% for insertions/deletions⁸. However, an analysis of mutational hotspots within a subset of these driver genes using more sensitive detection methods has projected that the mean mutation frequency is >12% of individuals in their 60’s, >19% in their 70’s, >40% in their 80’s and >74% in their 90’s²⁴. A whole-genome sequence analysis revealed that the frequency of clonal hematopoiesis increases from 0.5% in individuals younger than 35 years to more than 50% in those of older than 85 years¹¹. A study employing highly sensitive targeted error-corrected sequencing, which enables the detect clonal hematopoiesis at a very low VAF (~0.03%), revealed that 95% of individuals in their 50’s harbor low levels of mutations in select driver genes²⁵. Collectively, these studies suggest that initiating clonal mutations are ubiquitous by middle age and the expansion of these clones occurs in a portion of individuals at latter ages. These findings are consistent with estimates made from the observed rate of exonic mutation accumulation in human HSPC²⁶. It can be predicted that 44% of healthy 50 year old individuals will possess a HSPC with a randomly generated “oncogenic” TP53 mutation²⁷. While these considerations indicate that nearly everyone will develop mutations in driver genes by middle age, it is unclear why some individuals progress to clonal hematopoiesis with high VAF values whereas others will not display this clonal expansion. Presumably, genetics or environmental factors modulate the kinetics of HSPC clonal expansion. However, longitudinal data on clonal hematopoiesis in individuals is limited. To the extent that it has been examined, it appears that mutant clones are relatively stable over a ~10 year period^25,28,29. These data may indicate that the expansion of clones with an individual is episodic rather than linear, and it is conceivable that clone expansion is triggered by variations in exposure to external stressors that alter the bone marrow niche or change in the demand for hematopoiesis. These hypotheses can be addressed by additional epidemiological and experimental studies.

Therapy-associated clonal hematopoiesis -

Studies have shown that clonal hematopoiesis is associated with prior radiation therapy or chemotherapy^30,31. A study investigating prevalence of clonal hematopoiesis in patients who underwent autologous stem cell transplantation for non-Hodgkin lymphoma reported that 30% of patients harbor driver gene-associated clonal hematopoiesis at a low VAF at the time of transplantation³¹. Correspondingly, this cohort displayed an increased risk of all-cause mortality, later developing therapy-related myeloid neoplasm and a predisposition of the death from CVD. Another study investigating 8,810 individuals with solid tumors identified clonal hematopoiesis in 25% of the patients, among which some individuals harbored the presumptive mutation prior to radiation therapy³⁰. Notably, clonal hematopoiesis resulting from prior exposure to cytotoxic therapy is associated with high frequencies of mutations in TP53 and PPM1D compared with age-related clonal hematopoiesis³². This feature is likely due to the ability of mutated TP53 and PPM1D to confer resistance to genotoxic stress, thus providing HSPC with a survival advantage under these conditions. In support of this notion, it has been reported that patients with therapy-related acute myeloid leukemia (AML)/MDS harbor low levels of TP53-mutated clones before the onset of the disease, and prior to the chemotherapy to treat the primary hematological disease²⁷. These data suggest that TP53 mutations are selected for, but are not induced by the chemotherapy. Evidence for this hypothesis was provided by experiments where mice underwent transplantation with either wild type or p53-haploinsufficient HSPC. It was found that p53-mutant cells preferentially expand after N- ethyl-N-nitrosourea (ENU) administration²⁷. Interestingly, in another study, the fitness advantage of TP53-deficient cells appeared to be non-cell autonomous, such that the outcompeted cells played a role for establishing clonal dominance of mutant cells³³.

PPM1D is a member of PP2C family of Ser/Thr protein phosphatase that functions to suppress the activation of TP53³⁴. While PPM1D is rarely mutated in a primary hematologic maliganancy³⁵, PPM1D mutations are observed in the clonal hematopoiesis that is enriched in patients that have been treated for ovarian^36,37, breast³⁶, prostate³⁸, lung³⁹ and other solid tumors⁴⁰, and in patients that relapse after therapy for hematological malignancy³¹. Like TP53, these data indicate that genotoxic stress promotes the expansion of HSPC that harbor PPM1D mutations^30,31. Consistent with this notion, a recent study showed that PPM1D mutant clones display a competitive advantage under genotoxic stress⁴¹.

Neutral drift -

Neutral drift refers to a phenomenon of clonal selection among HSPC that originally possess an equal proliferative potential. This mechanism is analogous to the scenario of gamblers which have the equal odds of winning at a game of chance²³. In a stochastic manner, a fraction of clones is favored to predominate, and as these clones continue to “win” (i.e. expand), other clones are lost and finally replaced by the “winning” clone. The expansion of HSPC clones in this manner is referred to as “neutral drift-mediated clonal hematopoiesis”. In this scenario, the size of the active HSPC pool will impact the probability of neutral drift and its outcome^11,42. For example, it is much more probable that a clone will outcompete its neighbors if the HSPC pool is small and finite. While this scenario will lead to the “clonal collapse” of the HSPC pool, the pathological consequences of neutral drift-mediated clonal hematopoiesis remains to be elucidated.

CLONAL HEMATOPOIESIS, MORTALITY AND CVD

A number of studies have associated clonal hematopoiesis with an increase in all-cause mortality. The initial reports were published in in 2014^8,10. In these studies, exome sequence analysis was performed on peripheral blood mononuclear cells from individuals who were unselected for cancer or hematologic phenotypes. While these somatic mutations often occurred in the epigenetic regulators DNMT3A, TET2 and ASXL1 and other recognized driver genes, the study by Genovese et al. also reported that mutations in candidate driver gene could only account for half of the observed clonal hematopoiesis based upon a non-biased whole exome sequence analyses¹⁰. These two studies found an increase in mortality during the follow up period that was approximately 40% in individuals that exhibited clonal hematopoiesis^8,10. Marked increases in the frequencies of hematologic cancer were observed in both studies, as would be expected since mutations in these driver genes can be viewed as an early step in the progression to hematologic malignancy. However, the conversion to a hematologic malignancy is low in the general population, and it could not account for the large increases in the clonal hematopoiesis-associated, all-cause mortality in these studies. Expanding on this point, recent studies have employed deep DNA sequence analysis of candidate driver genes to identify features that distinguish individuals that are at particular risk of developing AML from those that exhibit age-related clonal hematopoiesis but do not progress to a hematologic malignancy^28,29. Interestingly, AML was found to be associated with the total number of mutations exhibited by an individual, the magnitude of the VAF and the specific nature of the driver gene mutation. Clonal mutations in DNMT3A and TET2 showed a greater association with clonal hematopoiesis, whereas TP53 and U2AF1 mutations were more prognostic for AML^28,29.

This leads to the question, what can account for most of the elevated mortality in individuals with clonal hematopoiesis? The increased risk of all-cause mortality is most likely associated with a relatively large increase in the risk for CVD. In their original study, Jaiswal et al. (2014) reported that driver gene-associated clonal hematopoiesis corresponded to an increased risk of coronary heart disease (hazard ratio = 2.0) and ischemic stroke (hazard ratio = 2.6) after adjusting for age, sex, type 2 diabetes, systolic blood pressure, and body mass index⁸. In a subsequent study, Jaiswal et al. (2017) reported that clonal hematopoiesis increased the risk of coronary heart disease (hazard ratio = 1.9) in other patient cohorts after adjusting for age, sex, type 2 diabetes mellitus, total cholesterol, high density lipoprotein, smoking, and hypertension¹². This study also reported a significant association between clonal hematopoiesis and early onset (< 50 years of age) myocardial infarction (odds ratio = 4.0) after adjusting for age, sex, type 2 diabetes status, and smoking status. Consistent with these findings, individuals with clonal hematopoiesis exhibited higher coronary artery calcification scores.

Whole genome sequence analysis of 11,262 Icelanders also reported an association between clonal hematopoiesis and mortality¹¹. In this study, clonal hematopoiesis was assessed by SNP-based computational analysis of genomic DNA of peripheral blood. A very high prevalence of age-associated clonal hematopoiesis was detected, but known candidate driver genes accounted for a relatively small fraction of the clonal hematopoiesis in this cohort. Clonal hematopoiesis was associated with an increase in all-cause mortality regardless of whether or not it could be associated with a candidate driver gene mutation. Similarly, the non-biased whole exome sequencing analysis by Genovese et al. also associated clonal hematopoiesis with mortality regardless of whether individuals had a mutation within an identified candidate driver gene¹⁰. While the Icelandic study did not report cardiovascular outcomes, significant associations were observed between clonal hematopoiesis and smoking, smoking-related disease and chronic pulmonary disease¹¹. This study also detected a strong association between clonal hematopoiesis and the mosaic loss of the Y chromosome, suggesting that they are related phenomena.

More recently, a study examined large clonal mosaic chromosomal alterations, ranging from 50kb to 249 Mb, in blood-derived DNA using SNP array data from 151,202 UK Biobank participants⁴³. These chromosome alterations increased with age and were associated with a doubling in the risk for all-cause mortality. Notably, this increase in mortality could only partly be explained by an increase in cancer deaths. While a number of of these loci are predicted to ablate tumor suppressor genes, including DNMT3A and TET2, this study provides evidence that multiple DNA mutations beyond those that target candidate known driver genes can give rise to clonal hematopoiesis with consequences on mortality.

In patients that undergo autologous stem cell transplantation for lymphoma, driver gene-associated clonal hematopoiesis was associated with an increased risk of overall mortality, an increased risk of therapy-related myeloid neoplasm, and a predisposition to death from CVD³¹. In this cohort, hematopoietic stem cells are presumably under extreme hematopoietic stress leading to a distinct mutational spectrum of candidate drivers, including high frequencies of mutations in TP53 and PPM1D. Driver gene-associated clonal hematopoiesis is also found to be enriched in solid tumor patients, and this condition is associated with inferior survival³⁰. In this cohort, clonal hematopoiesis was associated an increased risk of hematologic cancer, but the most common cause of death was the progression of the primary non-hematologic cancers.

EXPERIMENTAL STUDIES ON CLONAL HEMATOPOIESIS

Epidemiological analyses are inherently descriptive, and it is not possible to determine whether clonal hematopoiesis and CVD are causally linked or whether they are epiphenomena of the aging process. Furthermore, it is often difficult to address issues of directionality. For example, it has been suggested that chronic inflammation and other stresses associated with CVD can promote somatic mutagenesis and clonal hematopoiesis⁴⁴. At the current stage of research, animal models have particular utility in addressing questions of causality and directionality.

TET2 -

TET2 is a regulator of HSPC self-renewal and proliferation^45–47. Over 130 TET2 mutations have been reported in cancer-free individuals, most of which are predicted to lead to a loss of function⁴⁸. TET2 is an epigenetic regulator that can activate or repress transcription depending on the gene target and molecular context. TET2 is widely recognized to catalyze the conversion of 5-methylcytosine (5mC) into 5-hydroxymethylcytosine (5hmC), leading to DNA demethylation and transcriptional activation^49–51. TET2 can also promote the repression by pro-inflammatory genes by recruiting histone deacetylases to their promoters⁵².

Murine studies by our group initially revealed a potential causative link between somatic TET2 mutations in blood cells and atherosclerotic CVD¹⁵ and heart failure¹⁴ (Figure). In these studies, a competitive bone marrow transplantation approach was undertaken to simulate the clonal expansion of mutant cells. Tet2-deficient HSPC displayed progressive expansion into all immune cell progeny in the bone marrow, spleen, and blood. A slight myeloid bias was detected with a preferential expansion into the Ly6C^high classical monocyte population, a feature that is more pronounced in humans⁵³. Notably, the expansion of the Tet2-deficient cells did not affect white blood cell numbers, consistent with what is observed in individuals with TET2-mediated clonal hematopoiesis^8,20. In studies of experimental atherosclerosis, Tet2-mediated clonal expansion led to a marked increase in plaque size in hyperlipidemic low density lipoprotein receptor-deficient mice¹⁵. Similarly, an increase in plaque size was observed when cells heterozygous for Tet2-deficiency were transplanted into the atherogenic mouse model. Under these conditions, the degree of clonal expansion was less than what observed with the homozygous-deficient mice and there was a smaller increase in plaque size, indicative of a dose-response relationship. It was also shown that myeloid-specific ablation of Tet2 was sufficient to promote atherosclerosis development¹⁵. Consistent with these observations, it was also reported that full hematopoietic ablation of Tet2 increases plaque size in a murine atherosclerosis model¹².

Figure. TET2-mediated clonal hematopoiesis accelerates CVD.

Somatic Tet2 mutations within HSPC will lead to their clonal amplification, giving rise to myeloid cell progeny that promote disease through excessive production of IL-1β^14,15.

To corroborate and extend these studies, the effects of partial hematopoietic cell Tet2-deficiency were assessed in murine models of heart failure¹⁴. Partial hematopoietic Tet2-deficiency was achieved by bone marrow reconstitution techniques or by myeloid-specific ablation. The inactivation of hematopoietic Tet2 led to increased pathological cardiac remodeling, as indicated by diminished cardiac function and greater myocardial hypertrophy, fibrosis and inflammation in both the left anterior descending (LAD) ligation model of myocardial infarction and in the transverse aortic constriction (TAC) model of pressure overload hypertrophy. In these heart failure models, as in experimental atherosclerosis, accelerated pathology could be observed under conditions of Tet2 haploinsufficiency. Additionally, hematopoietic Tet2-deficiency, achieved by the transplantation of lineage-negative bone marrow cells that underwent lentivirus-mediated, CRISPR/Cas9-mediated gene editing, was shown to promote greater detrimental remodeling in mice that were treated with angiotensin II (AngII) infusion¹³.

Experimental studies have shown that the overactivation of IL-1β signaling contributes to the enhanced cardiovascular pathogenesis that is associated with the expansion of Tet2-deficienct hematopoietic cells^14,15 (Figure). In the atherosclerosis model, it was found that exacerbated IL-1β production by Tet2-deficient plaque cells promotes P-selectin expression and endothelial cell activation in the vascular lesion¹⁵. Similarly, hematopoietic deficiency of Tet2 led to an increase in myocardial expression IL-1β in the models of LAD ligation and TAC¹⁴. In cell culture studies, Tet2-deficient macrophages express higher levels of IL-1β transcript and the inactive precursor protein pro-IL-1β¹⁵. Tet2 deficiency also up-regulates components of the NLRP3 inflammasome, leading to the increased processing and secretion of active IL-1β protein. Mechanistic studies have revealed that Tet2-mediated IL-1β regulation is mediated by its ability to modulate the recruitment of histone deacetylase to the IL-1β promoter^15,52. The functional significance of IL-1β in the Tet2-mediated pathology was indicated by experiments where treatment with an NLRP3 inflammasome inhibitor eliminated the accelerated pathologies associated with partial hematopoietic Tet2 ablation in atherosclerosis and heart failure models^14,15.

The experimental studies on Tet2 regulation of the IL-1β/NLRP3 inflammasome may shed light on the findings of the CANTOS (Canakinumab Anti-inflammatory Thrombosis Outcomes Study) trial. This study found that treatment with an IL-1β neutralizing antibody (Canakinumab) can reduce major adverse cardiovascular events in high-risk patients with a previous history of myocardial infarction and elevated levels of C-reactive protein⁵⁴. Although Canakinumab showed efficacy, its general use in the clinic poses challenges as it increases the risk of infection. Furthermore, it has been reported that a fraction of study participants do not respond to Canakinumab with C-reactive protein reduction, and this subgroup did not show cardiovascular benefits⁵⁵. Based upon experimental studies, one can speculate that individuals with clonal hematopoiesis mediated by mutations in TET2, and perhaps related driver genes, may respond more favorably than the general population to IL-1β/NLRP3-targeted therapies. Thus, an evaluation of clonal hematopoiesis in CANTOS trail participants has merit as it could indicate a path for personalized therapy for the prevention of CVD in individuals carrying somatic mutations in their hematopoietic system.

DNMT3A -

DNA methyltransferase 3A (DNMT3A) is a de novo DNA methyltransferase that catalyzes DNA methylation and modulates gene transcription in various cell types including HSPC^56–60. Accumulating human studies show that DNMT3A is the most prevalent mutated driver gene associated with clonal hematopoiesis in the elderly population^8–11, and DNMT3A-mediated clonal hematopoiesis has been associated with an increased risk of CVD¹². However, whether somatic mutation in Dnmt3a causally contribute to CVD was unknown until recently.

Our group has demonstrated that partial Dnmt3a loss-of-function in hematopoietic stem cells worsens the cardiovascular phenotype in a model of heart failure induced by AngII infusion¹³. To conduct this study, a facile system of mutagenesis was developed where lentiviral vectors were used transduce lineage negative bone marrow cells prior to bone marrow transplantation with Cas9 and a guide RNA to create insertion and deletion mutations in the Dnmt3a gene⁶¹. In contrast to mutations in Tet2, Dnmt3a-disrupted HSPC did not display selective expansion during the 4-month time course of these experiments. This behavior is consistent with previous reports showing that Dnmt3a-null HSPC only expand in aged mice or after sequential bone marrow transplantations^60,62,63, and these findings raise concerns about using murine models as the sole system to study clonal hematopoiesis. Regardless, mice transplanted with Dnmt3a-edited HSPC showed significantly greater cardiac hypertrophy, diminished ejection fraction, and more fibrosis after AngII administration despite a relatively low level of chimerism (approximately 5 to 10%).

From the perspective of mechanism, Dnmt3a has been reported to regulate inflammation in multiple cell types and in multiple ways. In the experimental heart failure studies, partial hematopoietic deficiency of Dnmt3a was associated with greater macrophage accumulation and increased expression of immune cell markers in myocardium, suggesting that this condition impairs the resolution of inflammation¹³. Consistent with this hypothesis, studies with a macrophage cell line showed that Dnmt3a-deficiency promotes inflammation by up-regulating the expression of specific cytokines and chemokines. In other systems, Dnmt3a-deficiency has been shown to accelerate proinflammatory activation of mast cells, dysregulation of T cell polarization, and modulate peritoneal macrophage function^64–67. Taken together, these findings support the concept that hematopoietic cell Dnmt3a-deficiency promotes a pro-inflammatory state. However, given the complex immunomodulatory properties of Dnmt3a, additional experimental studies are warranted to develop a more comprehensive understanding of how hematopoietic cell Dnmt3a-deficiency affects the pathogenesis of CVD.

Other CH-related genes -

Like TET2 and DNMT3a, ASXL1 is an epigenetic regulator. ASXL1 controls epigenetic marks through interaction with polycomb complex proteins and various other transcriptional regulators⁶⁸. ASXL1 gene mutations are detected in variety of myeloid neoplasm, and they are associated with poor prognosis^69–71. These mutations are also frequently observed in clonal hematopoiesis^8–11. Most alterations in ASXL1 are nonsense mutations or frameshift mutations located in or near the last exon that result in a premature stop codon. A truncation in the C-terminus of murine Asxl1 has been reported to lead to an unstable protein that gives rise to a MDS-like phenotype^72–74. However, more recent data suggest that the truncation of Asxl1 may lead to a gain of function in which truncated Asxl1 enhances the activity of BAP1, leading to reduced ubiquitination of H2AK119⁷⁵. Consistent with this notion, truncated ASXL1 protein can be detected in leukemia cell lines⁷⁶. In experimental mouse studies, it was found that the retroviral overexpression of a truncated form of Asxl1 in bone marrow cells leads to an MDS-like disease and a reduction of H3K27me3 levels⁷⁷. More recently, analysis of a “knock-in” mouse model of truncated Asxl1 revealed that mice were void of overt hematological malignancies, a finding that is more consistent with the concept that ASXL1 functions as a clonal hematopoiesis driver gene^78,79. Collectively, current data suggest that the truncated Asxl1 protein can display dominant-negative and gain-of-function effects depending on context. To date, however, it is unknown whether mutations in hematopoietic cell Asxl1 contribute to CVD.

JAK2 associates with a variety of receptors and acts as a signaling kinase⁸⁰. The constitutively active allele JAK2^V617F is present in leukocytes of a majority of patients with myeloproliferative neoplasm (MPN), including polycythemia vera (PV) and essential thrombocytosis (ET), that result from the dysregulated expansion of red blood cells and platelets, respectively⁸¹. Currently, the role of JAK2^V617F in the development of clonal hematopoiesis is controversial, but there is evidence that some JAK2^V617F mutant carriers do not show signs of blood count abnormalities or progress to MPN^{8,10,12,29,82–85}. The phenotypic heterogeneity of JAK2^V617F carriers could be due to the highly heterogeneous nature of HSPC⁸⁶. For example, a JAK2^V617F mutation in platelet-biased HSPC might result in an ET phenotype, whereas a mutation in myeloid-biased HSPC that gives rise to erythroid progenitors might result in PV phenotype. Along these lines, it is possible that JAK2^V617F clonal hematopoiesis occurs when the mutation arises in a more restricted HSPC population. While further mechanistic studies are required, JAK2^V617F−mediated clonal hematopoiesis could impact CVD in multiple ways depending on cell types that express this allele. While the impact of JAK2^V617F mutations on erythroid cells and megakaryocytes has been well studied, its impact on myeloid populations has only recently been investigated. It has been reported that JAK2^V617F will activate b1 and b2 integrin expression in neutrophils and promote thrombus formation in mice^87,88. The JAK2^V617F−mutant neutrophil can also be more prone to neutrophil extracellular trap (NET) formation, a process referred to as NETosis, contributing to thrombosis⁸⁹. Given that neutrophils promote arterial plaque erosion and thrombosis via NETosis^90–92, JAK2^V617F−mediated clonal hematopoiesis may accelerate the onset of ischemic events at culprit lesions. This concept is consistent with the observation of mortality among young individuals with JAK2^V617F−mediated clonal hematopoiesis due to coronary heart disease¹². Also of note, a recent exome-wide study associated the JAK2^V617F allele with lower triglyceride and LDL cholesterol levels despite increased risk of coronary artery disease, suggesting that the JAK2^V617F allele can promote CVD via mechanisms that are independent of lipid metabolism⁹³. While further studies are required to understand the impact of JAK2^V617F−mediated clonal hematopoiesis on CVD, current mouse models are of limited value because the JAK2^V617F mutation in hematopoietic cells leads to strong PV and ET phenotypes. Thus, novel expression systems that target JAK2^V617F to specific cell types may be required to create an appropriate model to study clonal hematopoiesis.

Hematopoietic mutations in TP53 are also associated with clonal hematopoiesis. TP53 encodes a tumor suppressor that is widely recognized to protect against genomic instability through its abilities to regulate DNA repair, cell cycle arrest, and apoptosis⁹⁴. Many studies have examined the link between TP53 and CVD, but these studies have mainly focused on the effects of p53 on cardiac myocytes^95,96, endothelial cells^97,98, and vascular smooth muscle cells⁹⁹, and relatively little is known about the contribution of hematopoietic TP53 mutations to clonal hematopoiesis-associated CVD. However, prior to its appreciation as a clonal hematopoiesis driver gene, it was reported that the transplantation of p53-deficient bone marrow into hyperlipidemic LDLR-KO mice leads to larger atherosclerotic plaques with increased macrophage proliferation in the plaque⁹⁹. Several studies have investigated the effects of p53-deficiency in various inflammatory processes in immune cells. Murine neutrophils and macrophages deficient for p53 express more TNFa, IL-6, and CXCL-2 after LPS stimulation, and p53-deficient neutrophil also upregulate elastase expression^100,101. Given the current interest in clonal hematopoiesis, additional experiments that examine the functional role of TP53 are warranted.

PERSPECTIVES

Hematopoietic cells with somatic mutations are prevalent in the adult population. These mutations generally enhance the fitness of HSPC such that they allow for the clonal amplification of the mutant cell in the absence of changes in blood cell counts. These clonal events increase with age and they are associated with all-cause mortality and CVD. Some of these mutations occur in driver genes that are recurrently mutated in hematologic malignancies. Recent experimental work has delineated how mutations in the Tet2 and Dnmt3a driver genes contribute to pathology in models of heart failure and/or atherosclerosis^12–15. However, many other candidate driver genes have yet to be investigated. Furthermore, epidemiological studies suggest that a large fraction of the observed clonal hematopoiesis cannot be attributed to mutations in candidate hematologic driver genes, suggesting that multiple mechanisms of genome instability can contribute to this condition. Clonal hematopoiesis represents a new mechanism of CVD, and this avenue of research is in its infancy. Continued studies in this burgeoning area may offer new therapeutic opportunities that can be personalized based upon specific gene mutations.

Table.

Reports of clonal hematopoiesis

Study	Method	Sample	Population	Sensitivity	Prevalence of CH
Xie et al. (2014)9	Whole-exome sequencing	Peripheral blood	2728 individuals with first-time primary cancers (selected from TCGA)	0.01 VAF	5–6% of >70 years
Genovese et al. (2014)10	Whole-exome sequencing	Peripheral blood	12,380 Swedish persons unselected for cancer of hematological phenotypes (6245 controls, 4970 with schizophrenia, 1165 with bipolar disorder)	0.01 VAF	1%<50 years 10% of >65 years
Jaiswal et al. (2014)8	Whole-exome sequencing	Peripheral blood	17,182 persons unselected for hematological phenotypes (selected from 22 population-based cohorts)	0.035 VAF	Rare of < 40 years 11% of >70 years
McKerrell et al. (2015)24	Targeted sequencing (15 hotspot analysis)	Peripheral blood Cord blood	4219 individuals (3067 blood donors, 1152 unselected individuals, 32 patients with hematopoietic stem cell transplant, 18 cord blood)	0.008 VAF	0.8% of < 60 years 19.5% of >90 years
Young et al. (2016)25	Targeted sequencing (error-corrected)	Peripheral blood	20 healthy female participants (selected from Nurse’s Health Study)	0.0003 VAF	Mutation in DNMT3A and TET2 in 95% of individuals (56.6–68.1 years old)
Jaiswal et al. (2017)12	Whole-exome sequencing	Peripheral blood	4726 participants with coronary heart disease, 3529 controls (selected from four case-control studies)	0.1 VAF	7% (mean age of 65)
Zink et al. (2017)11	Whole-genome sequencing	Peripheral blood	11,262 Icelanders (in various disease projects at deCODE genetics)	0.1 VAF	0.5% of <35 years >50% of >85 years
Abelson et al. (2018)29	Targeted sequencing (error-corrected)	Peripheral blood	Discovery Cohort: 95 pre-AML cases, 414 controls Validation Cohort: 29 pre-AML cases, 262 controls (selected from EPIC Study)	0.005 VAF	around 30% of >50 years
Desai et al. (2018)28	Targeted sequencing	Peripheral blood	181 age-matched control population against later AML group (selected from WHI)	0.01 VAF	20.75% of<64 years 38.46% of>65 years
Gibson et al. (2017)30	Whole-exome sequencing Targeted sequencing	(pre-and post-ASCT) Peripheral blood and bone marrow Cryopreserved aliquots of autologous stem-cell products	12 patients who underwent ASCT for Hodgkin or non-Hodgkin lymphoma 401 patients who underwent ASCT for non-Hodgkin lymphoma	0.02 VAF	50% 29.9%
Coombs et al. (2017)31	Targeted next generation sequencing	Paired tumor and blood	8810 patients	0.01 VAF	25.1%
Jongen-Lavrencic et al. (2018)32	Targeted next generation sequencing	Bone marrow or peripheral blood	428 patients with a confirmed diagnosis of previously untreated AML, 54 patients who had refractory anemia with excess of blasts	0.02 VAF	89.2%