Hematopoiesis of Indeterminate Potential and Atherothrombotic Risk.

Hematopoiesis is the process of blood production, essential for the continued supply of immune cells and red blood cells. However, the proliferative nature of hematopoietic stem cells (HSCs) renders them susceptible to developing somatic mutations. HSCs carrying a mutation can gain a selective advantage over normal HSCs and result in hematological disorders. One such disorder is termed clonal hematopoiesis of indeterminate potential (CHIP), a premalignant state associated with aging, where the mutant HSCs are responsible for producing a small portion of mature immune cells in the circulation and subsequently in tissues. People with CHIP have been shown to have an increased risk of mortality due to cardiovascular disease (CVD). Why this occurs is under rigorous investigation, but the majority of the studies to date have suggested that increased atherosclerosis is due to heightened inflammatory cytokine release from mutant lesional macrophages. However, given CHIP is driven by several mutations, other hematopoietic lineages can be altered to promote CVD. In this review we explore the relationship between mutations in genes causing CHIP and atherothrombotic disorders, along with potential mechanisms of enhanced clonal outgrowth and potential therapies and strategies to slow CHIP progression. The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution-NonDerivative-NonCommercial License, permitting copying and reproduction so long as the original work is given appropriate credit. Contents may not be used for commercial purposes, or adapted, remixed, transformed or built upon. (https://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

Hematopoiesis is the tightly regulated and hierarchical process of blood cell production. This process originates from hematopoietic stem cells (HSCs) through to lineage-committed progenitors and mature leukocytes, red blood cells, and platelets. The development of specific lineages is regulated through intrinsic factors, including precise combinations of transcription factors and epigenetic modifications, along with extrinsic cues such as cytokines and growth factors, resulting in the expression or repression of gene signatures to shape the morphological and functional capabilities of the mature cell. However, these processes can be altered with aging causing fundamental alterations to the hematopoietic system. 1 One of the irreversible changes that occur in all somatic cells is the acquisition and persistence of mutations due to inefficiencies in DNA repair mechanisms. In HSCs, somatic mutations can accumulate with each division, 2 3 4 with most having no overt functional effect and many resulting in diminished HSC function initiating programmed cell death and clearance of the mutant cell. However, some mutations in HSCs evade clearance and can result in a competitive advantage, characterized by increased self-renewal, proliferation, survival, and biased lineage output. 5 6 7 To put this into numbers, HSCs are thought to acquire one to two mutations per division, which when extrapolated equates to approximately 10 mutations/year, and once we reach adulthood, modeling suggests that most of our HSCs will have two coding mutations and as many as 200,000 noncoding mutations. This suggests that by the time we are adults, each HSC is unique and keeps acquiring mutations. 2 3 4 These mutations can result in a process termed clonal hematopoiesis of indeterminate potential (CHIP), which is a form of clonal hematopoiesis, but is not driven by other mechanisms such as clonal mosaicism. Individuals with CHIP have an increased risk of mortality, which is now linked to cardiovascular disease (CVD). In this review, we will focus on CHIP mutations associated with thrombotic disease, the mechanisms contributing to this pathology, and potential therapies.

Clonal Hematopoiesis of Indeterminate Potential

Somatic mutations can result in several hematological disorders. CHIP occurs when HSCs acquire a somatic mutation providing them with a competitive growth advantage over normal HSCs. This results in a relative increased number of mutated hematopoietic cells in the bone marrow and blood. While the overall abundance of white blood cells (WBCs) is only modestly affected, over time the proportion of mutated cells in the blood grows at the expense of normal WBCs. 6 CHIP is an aging phenomenon, because a key feature of it is insufficient repair of damaged DNA which may then be differentially propagated depending on mutational fitness. 8 CHIP is defined as a variant allele frequency (VAF) of >2% in circulating WBCs (i.e., >4% of WBCs carry the mutation in one allele). 9 10 Largely based on whole exome sequence studies of blood DNA in various datasets, it is estimated that approximately 5% of individuals aged under 60 years display CHIP, which increases to approximately 10% aged over 60 years and is continued to increase with age. 11 12 13 Deep targeted sequencing indicates that hematological somatic leukemogenic mutations at very low VAF (i.e., median 0.2%) are almost ubiquitous in middle age healthy adults. 14 This finding infers that we are all at some point in our lives at risk of developing CHIP and other hematological disorders.

Mutations indicative of CHIP are most commonly observed in the genes DNMT3A , TET2 , ASXL1 , JAK2 , and TP53 . 15 16 Why mutations in the epigenetic modifiers DNMT3A and TET2 are such prevalent drivers of CHIP is not known. However, the relatively open chromatin structure of HSCs suggests that gene expression is largely governed by the methylation status and may reveal why mutations in DNMT3A (methylation) and TET2 (hydroxymethylation) cause such dominant changes in HSCs to promote their outgrowth. 17 Certainly, studies exploring the deletion or loss of function of these genes demonstrate the competitive advantage these mutant HSCs acquire. 18 19 The loss in DNMT3A and thus reduced DNA methylation result in the increased expression of genes involved in HSC proliferation and self-renewal, while the prevention of hydroxymethylation when TET2 is nonfunctional destabilizes key HSC maintenance genes which promotes both hyperproliferation and myeloid skewing.

CHIP and Cardiovascular Disease

CHIP-driving mutations are known to increase the risk of hematologic malignancy and carriers have 10 times the risk of hematologic cancer as those without such mutations do. 13 Initial analysis has found an association of CHIP with increased all-cause mortality, but the increased risk of hematological malignancies of 0.5 to 1% per year is not nearly enough to account for the 40% increase in mortality. 12 13 Further analyses identified a strong association of CHIP with a higher risk of CVD independent of age and other traditional risk factors. 11 13 20 21 22 The direct evidence for causality was first provided by animal studies. Hematopoietic Tet2 or Tet2 markedly increased atherosclerosis in hypercholesterolemic Ldlr mice. 21 23 Mechanistically, Tet2 -deficient macrophages showed increased NLRP3 inflammasome activation and elevated IL-1β production ( Fig. 1A ). 21 23 Concentrations of related biomarkers are also increased among individuals with CHIP. 24 25 An NLRP3 inhibitor selectively reversed the increased atherosclerosis in the Tet2 CHIP model. 23

Fig. 1

Mechanistic features of hematopoietic mutations resulting in increased risk of atherosclerosis and atherothrombotic disease. ( A ) Somatic mutations in TET2 have been shown to accelerate atherosclerosis through inflammatory macrophage signaling driven by the NLRP3/IL-1β axis, which when inhibited reduces atherosclerosis. ( B ) Somatic mutations in JAK2 can result in myeloproliferative neoplasms which influence atherothrombotic disorders, namely venous thrombosis. However, when JAK2 mutations result in CHIP, enhanced activation of the AIM2 inflammasome occurs to release inflammatory cytokines through gasdermin D pores. Whether this causes atherothrombosis is not yet known. ( C ) Germ line mutations in SH2B3 increase the risk of atherothrombotic disease. Inflammatory interactions between neutrophils and platelets occur, particularly driven by platelet release oxidized phospholipids (oxPL), causing NETosis. Interestingly, JAK2 mutations are linked with SH2B3 mutations, where the LNK(R262W, T allele) predisposes individuals to JAK2 VF MPN and CHIP, along with coronary artery disease. CHIP, clonal hematopoiesis of indeterminate potential.

JAK2V617F ( JAK2 ) is less common than the mutations of epigenetic modifiers such as TET2 , DNMT3A , or ASXL1 . 13 21 Nevertheless, a recent study, with innovative deep targeted sequencing that had a screening sensitivity as low as VAF 0.01%, found that the JAK2 VF mutation is detectable in almost 4% of a general European population. 26 Among the JAK2 individuals in this population, approximately 60% had a VAF of >0.1% but most of whom did not have features of myeloproliferative neoplasm (MPN). CHIP-associated JAK2 VF occurs at a younger age than the other CHIP variants 27 and dramatically increases risk of myocardial infarction by as much as 12-fold in younger people. 21 We found that JAK2 VF increases atherosclerotic disease despite lowering LDL (low-density lipoprotein) cholesterol in both mice and humans. 28 29 Interestingly, it has been shown that mouse models of JAK2 VF resembling a MPN (i.e., 100% JAK2 VF bone marrow transplant [BMT]) or CHIP (20% Jak2 VF :80% WT BMT) both increase atherosclerosis. 29 30 JAK2 VF causes altered functionality of multilineage blood cells. 29 Selective expression of JAK2 VF in monocyte/macrophage increases atherosclerosis in association with increased generation of IL-1β and IL-18, the product of inflammasome activation. 30 Unlike Tet2 deficiency, knockout of NLRP3 has little effect, while deletion of AIM2, the essential component of AIM2 inflammasome, reduces the increased atherosclerosis in JAK2 VF models ( Fig. 1B ). Inflammasome activation can lead to programmed cell death that is mediated by the pyroptosis executioner gasdermin D (Gsdmd), leading to release of inflammasome activation products such as IL-1β. Gsdmd reduces atherosclerosis in JAK2 VF mice. 30

Chronic inflammation associated with atherosclerosis has long been thought to mediate atherosclerosis progression. 31 The recent Canakinumab Anti-inflammatory Thrombosis Outcomes Study (CANTOS) largely validated this notion. 32 IL-1β inhibition by canakinumab reduces incident CVD events in individuals with prevalent CVD and elevated high-sensitivity C-reactive protein (hsCRP), a marker of inflammation. 32 However, canakinumab therapy was associated with only a moderate clinical benefit and an increase in infections. Consequently, canakinumab has not been marketed for cardiovascular indications. Similarly, colchicine, which inhibits the microtubule-dependent assembly of the NLRP3 inflammasome and IL-1β secretion, 33 34 appeared to benefit CVD but also increased pneumonia. 35 36

A more precise approach to identify patients who may benefit most from anti-inflammatory therapy, as well as identification of upstream therapeutic targets, could lead to more effective, safer anti-inflammatory treatments for CVD. Enhanced inflammation in NLRP3 or the AIM2 inflammasome/IL-1β axis in TET2 mutant or JAK2 VF atherosclerosis models suggests that individuals with TET2 mutant or JAK2 VF CHIP and increased risk of CVD could benefit more from anti-inflammation therapy. In the preclinical JAK2VF CHIP model, administration of an anti-IL-1β antibody improved readouts of plaque stability, but did not change the overall lesion size and had no apparent effect in control mice. 30 Consistently, in an exploratory preliminary analysis, patients with a TET2 mutant have an improved response to canakinumab (hazard ratio [HR] = 0.38) relative to the response to overall response to canakinumab (HR = 0.93) in the CANTOS trial. 37 In addition to the pharmacological evidence, human genetic studies provide more support for a causal role of inflammation in CVD. IL-6 can be a downstream product of IL-1β signaling and increased IL-6 production and activity are considered to be a common mediator in chronic inflammation associated with CVD. A disruptive IL-6R missense variant is associated with 5% reduced CVD risk in a general population. 38 39 A more recent study indicates that this IL-6R variant attenuates CVD risk in individuals with CHIP by 54%, supporting the notion that individuals with CHIP may benefit most from IL-6 inhibition to reduce the risk of CHIP-associated CVD. 11

CHIP and Thrombosis

Several somatic mutations have been linked to quantitative or qualitative abnormalities in platelets (reviewed by Veninga et al 40 ). Among the common CHIP-driving mutations that affect platelets, the evidence for increased risk of atherothrombosis primarily comes from studies of JAK2 VF -associated MPNs and CHIP. Patients with MPNs have increased risk of arterial and venous thrombosis and thrombotic complications. 41 42 43 The studies of thrombotic risk conferred by JAK2 VF CHIP without apparent MPNs have significantly advanced relatively recently. In a population study of over 10,000 individuals without a known myeloid disorder, JAK2 VF CHIP was associated with an increased incidence of thrombosis 44 ( Fig. 1 ). Polycythemia vera (PV), in which more than 90% of the patients are JAK2 VF positive, 45 has increased hematocrit in association with increased blood viscosity, 46 a major risk factor of thrombosis in PV patients. Phlebotomy has been successfully used to maintain the hematocrit and reduce the risk of thrombosis in PV patients. 43 Thrombotic risk also is increased in JAK2 VF MPN patients without apparently increased hematocrit. JAK2 VF increases myeloid cell production, causing leukocytosis. An increased WBC count has been associated with an increased risk of thrombosis in MPNs. However, many with JAK2 VF CHIP do not show signs of abnormal blood cell counts yet they still have increased risk of thrombosis. 44 JAK2 VF is generally thought to only affect platelets, but platelets can interact with many cells types. In the circulation, neutrophils often form aggregates with platelets which can activate these cells. A consequence of neutrophil activation is the formation of neutrophil extracellular traps (NETs), where neutrophils release their contents leading to the formation of web-like structures (termed NETosis) made of DNA, myeloperoxidase, citrullinated histones and proteases that entrap and kill bacteria. 47 While NETs may help to suppress infections, the formation of NETs in blood vessels can promote atherosclerosis and thrombosis 48 49 ( Fig. 1B ). NETs may serve as biomarkers predicting the risk of thrombosis. 49 Neutrophils from patients with MPNs display some features of enhanced activation. 50 51 Subsequent studies showed that neutrophils from MPN patients are primed for NETosis. 44 However, this is not always observed. 52 JAK2 VF -modeled mice showed increased NET formation and venous thrombosis, which were reduced by DNase treatment or hematopoietic deficiency of peptidyl-arginine deiminase 4 (PAD4), the enzyme essential for citrullination of histones in NET formation. 44

Various mouse models have been generated to assess the impact of JAK2 VF on arterial thrombosis, including mouse JAK2 VF knock-in, human JAK2 VF transgenic, and knock-in of human JAK2 VF cDNA into the mouse Jak2 allele, with constitutive or tamoxifen-induced expression in hematopoietic and endothelial cells. 53 54 55 These models have different levels of mouse or human JAK2 VF expression in different types of cells and generate various thrombotic and hemostasis phenotypes, varying from increased platelet activation, reduced tail vein bleeding time, and increased thrombus formation in a flow chamber assay to decreased platelet activation in response to convulxin or TRAP4, increased tail vein bleeding, and accelerated arterial thrombosis with unstable thrombi in vivo. A common feature of these studies is modeling of MPNs with high JAK2 VF burden but not CHIP. The impact of JAK2 VF CHIP with a lower burden on platelet activity and arterial thrombosis in mice is not known ( Fig. 1B ). We have some evidence that modeling CHIP in mice with a small fraction (2–20%) of hematopoietic cells expressing JAK2 VF shows increased platelet activation to thrombin and accelerated arterial thrombosis, without increased bleeding ( unpublished observations ).

SH2B3/LNK encodes an adaptor protein that acts as a negative regulator of JAK2-mediated hematopoietic cell proliferation. 56 Lnk deficiency promotes multilineage expansion of HSCs in mice. 56 Ldlr mice with hematopoietic LNK deficiency display increased atherogenesis and accelerated atherothrombosis. 57 Lnk deficiency and hypercholesterolemia act synergistically promoting platelet activation and myelopoiesis. 57 More recently, we show that the increased arterial thrombosis in hematopoietic Lnk -deficient mice is due to NETosis in the thrombi and accelerated thrombosis is completely reversed by neutrophil depletion or PAD4 deficiency. 58 Mechanistic studies have identified oxidized phospholipids (OxPLs) released and presented by activated platelets mediate neutrophil activation and NETosis ( Fig. 1C ). Lnk mice show increased plasma OxPL levels and transgenic expression of E06-scFv, which specifically binds and neutralizes OxPL activity, selectively and completely reverses NETosis in thrombi, and accelerates thrombosis in Lnk deficiency. 58 A common SH2B3 polymorphism (p.R262W, c.784T > C) is a loss-of-function LNK variant in association with increased platelet and neutrophil counts and the risk of CVD. 59 Consistent with this, we observed increased NETosis in co-cultures of human-induced pluripotent stem cell (iPSC)-derived neutrophils and activated platelets carrying isogenic LNK(TT) relative to LNK(CC). 58 Interestingly, the LNK(R262W, T allele) predisposes to both JAK2 VF MPN and CHIP. 60 We also confirmed this with data from UK Biobank, showing that individuals with the JAK2 VF mutation only displayed increased CAD when also carrying the LNK(TT) allele.

Accelerators of CHIP

Clone size, as estimated by VAF, is strongly associated with the prognosis of CHIP, whether it be cancer, subclinical atherosclerosis, atherosclerotic CVD, or heart failure. 13 20 21 61 62 Thus, identifying CHIP early will provide a window of opportunity to slow clonal growth and avoid cardiovascular complications. However, there is a scarcity of longitudinal sampling to confidently explore and define drivers of clonal outgrowth over time, with much of our knowledge coming from preclinical models. Thus, a major outstanding question is what drives clonal outgrowth and how can this be halted? Given that CHIP is driven by somatic mutations providing a competitive advantage, one strategy could be to define driver genes and target their expression or protein function to slow proliferative rates or kill the mutant HSCs. This is a complicated option as the major genes mutated in CHIP are epigenetic modifiers and kinases which have a significant impact on a network of genes involved in stem cell maintenance and proliferation. 18 19 We suggest that understanding the interactions between the mutations and environmental drivers of CHIP is key in delaying clonal outgrowth and may reduce the risk of CVD in the context of CHIP.

An alternative approach to exploring the gene regulatory networks that are altered by somatic mutations is to identify the extrinsic drivers of clonal outgrowth. It is slowly emerging that clonal outgrowth is linked with extrinsic factors including comorbidities, diets, smoking along with inflammatory status and infections 13 21 25 63 64 65 ( Fig. 2 ). In the initial studies linking CHIP to mortality due to CVD, an over-representation of individuals with metabolic disorders, namely diabetes, was noted. 13 Indeed, diabetes has been linked with leukemia 66 67 and in murine models diabetes has been shown to cooperate with Tet2 heterozygosity to cause leukemia. 68 TET2 deficiency can also aggravate insulin resistance in mice. 69 Metabolic stressors such as unhealthy diets have now been linked to a higher prevalence of CHIP. 70 71 Additional evidence to support the hypothesis that altered lipid and glucose metabolism often seen in individuals consuming unhealthy diets or with diabetes is linked with clonal outgrowth was seen in the Swedish Obese Subjects (SOS) study. 65 In this preprint manuscript, the authors report that over a 20-year follow-up period, growing clones in the obese individuals were found to correlate with low high-density lipoprotein insulin levels and HOMA index as a readout of insulin resistance ( Fig. 2 ). However, these data were based on a small sample size of <40 individuals and a larger follow-up study is required. Nonetheless, the mechanism(s) responsible for this are unknown, but could relate to low-grade chronic inflammation, which is a known consequence of obesity, providing cytokines that would enhance the proliferation of the HSC-harboring mutations. This hypothesis was also illustrated by Naxerova's group, revealing that disorders underpinned by enhanced hematopoietic activity (i.e., atherosclerosis and sleep fragmentation) accelerated clonal outgrowth. 64 However, it is important to note in the SOS study that hsCRP was not associated with clonal outgrowth. 65 We hypothesize that clonal outgrowth in metabolic disorders is due to an alternative mechanism. 72 First, we have previously shown that obesity and diabetes promote increased hematopoiesis at the level of the common myeloid progenitor and granulocyte-macrophage progenitor, while HSCs are largely unaffected in respect to abundance. 73 74 This suggests that enhanced hematopoiesis at the level of the HSCs due to extrinsic signaling is unlikely to be responsible. Instead, high-energy environments such as obesity and diabetes could reduce the activity of epigenetic modifiers particularly TET2, which is reliant on AMPK activity. 75 Through this mechanism, further loss of TET2 function in TET2 mutant cells (i.e., 1 mutant allele) or combination with mutations in other genes may synergize to cause myeloid skewing and increased clonal outgrowth. This is consistent with hematopoietic TET2 heterozygote mice that display transition to leukemia in a model of hyperglycemia. 68 If this is true, treating individuals with agents such as metformin or novel AMPK activators along with life-style interventions may be effective in slowing the expansion of mutant cells in these individuals. However, the impact on extrinsic stimuli in promoting clonal outgrowth is largely limited to mice and requires large longitudinal clinical studies to address our hypothesis more accurately.

Fig. 2

Accelerators of clonal hematopoiesis. Somatic mutations are ubiquitous in people of middle age. However, what promotes the proliferation of these mutated cells and development of CHIP is relatively unknown. To date, some metabolic disorders/stressors, sleep fragmentation and chronic inflammation, and viral infections have been associated with CHIP. The extrinsic environment could aid in the increase of clonal outgrowth of mutated hematopoietic stem cells (HSCs), influencing VAF and contributing to atherosclerosis. CHIP, clonal hematopoiesis of indeterminate potential; VAF, variant allele frequency.

Interventions for Individuals with CHIP

There is no approved treatment for CHIP-related CVD risk. CANTOS 32 and genetic evidence 11 suggest that anti-IL-1β and anti-IL-6 could be particularly effective and potentially broadly beneficial to CHIP carriers. Targeting NLRP3 or AIM2 inflammasome may need to be tailored to the specific genetic factors responsible for CHIP. While ruxolitinib, a JAK1/2 inhibitor, has been approved for JAK2 VF MPNs, it shows no benefit in atherogenesis in the JAK2 VF -modeled mice. 30 Fedratinib, a selective JAK2 inhibitor, is newly approved for MPN-associated myelofibrosis. 76 We showed that Fedratinib reduced atherogenesis in Apoe mice at least partly by reducing aberrant myelopoiesis 77 but its impact on CHIP-driven atherogenesis is not known. The ASPREE trial indicates that aspirin use in the healthy elderly does not provide benefit against CVD but increases the risk of major hemorrhage, 78 suggesting the need for more targeted therapy. Interestingly, evidence exists suggesting that individuals with JAK2 VF MPNs benefit more from aspirin relative to MPN patients carrying no JAK2 VF . 79 80 Potentially, aspirin could be more effective in JAK2 VF CHIP carriers to reduce thrombosis risk. OxPL has long been considered as a risk factor for CVD. 81 Anti-OxPL therapy could potentially be particularly beneficial for LNK mutant carriers or individuals with LNK risk polymorphism who have increased atherothrombotic risk.

Another important point to consider is surveillance for and managing clonal expansion. It has been shown that individuals with larger clones are at great risk of mortality and that people with small stable clones generally live healthy lives. As discussed above there is growing knowledge surrounding what lifestyle factors and co-morbidities drive clonal outgrowth. 18 34 64 73 We suggest another approach, to avoid small clones becoming problematic, could be to effectively treat comorbidities, alter lifestyles (i.e., diets, cessation of smoking), or activate pathways that might slow the proliferation of the mutated cells, which will likely be dependent on the mutated gene.

Nonetheless, since discovering the link between CHIP and CVD, experimental interventions targeting inflammation may find an indication in individuals with CHIP. With the movement toward precision medicine in the cardiovascular field, it may be important to define the genetic drivers of CHIP to treat these individuals effectively and significantly reduce their risk of CVD.