The spectrum of clonal hematopoiesis

Clonal hematopoiesis (CH) describes the growth of a sub-population of blood cells that have at least one somatic mutation, leading to a greater risk of hematological malignancy and cardiovascular disease, and an increased risk of mortality from non-hematological malignancies.¹

During the Eighth Annual Meeting of the Society of Hematologic Oncology (SOHO 2020) at a ‘Meet the Professor’ session, Koichi Takahashi of the University of Texas MD Anderson Cancer Center, Texas, US, gave a presentation on CH and myeloid malignancy.²

There are multiple terms and definitions for CH, including CH of indeterminate potential (CHIP) and age-related CH (ARCH), which may often be dictated by the method of detection. Detection methods can be observed in Table 1. X-chromosome inactivation tests, using polymerase chain reaction (PCR) methods, were previously used to identify the expanding CH sub-population. Naturally, this identification method limited detection to women only, but demonstrated that development of this sub-population of cells could be detected in an aging population. Subsequently, an alternative method was developed to assess copy number alterations or chromosomal abnormalities, and similarly identified CH in aging populations. This new method also revealed an association between CH and hematological malignancy risk.

More recently, the use of next-generation sequencing led to the identification of gene mutations, such as single nucleotide variants and insertion-deletion events, which are characteristic of CH. Due to a lack of studies utilizing all CH detection methods, it is not yet known whether CH detected using different methods overlap, are mutually exclusive, or whether there is synergy between these different sub-populations.²

Table 1. Detection methods used to identify clonal hematopoiesis.²

aCGH, array Comparative Genomic Hybridization; IG, immunoglobulin; indel, insertion or deletion of bases; NGS, next-generation sequencing; PCR, polymerase chain reaction; SNP, single nucleotide polymorphism; SNV, single nucleotide variant; TCR, T-cell receptor.
Characteristic	Detection Method
Immunophenotype	Flow cytometry
X-chromosome inactivation	Protein isoform/methylation/transcription based
Chromosomal abnormalities/copy number alteration	Karyotyping/aCGH/SNP/NGS
Gene mutations (SNVs, indels)	NGS
TCR rearrangement/IG rearrangement	PCR

Clonal dominance in other tissue

Clonal sub-populations with somatic mutations are not restricted to hematopoietic space but are also seen in other tissues (Table 2). Interestingly, these sub-populations seem to be prevalent in tissues that are subject to external stressors and have some commonalities in terms of driver mutations.

Table 2. Clonal dominance seen in other tissues and associated driver mutations.²

IBD, irritable bowel disease.
Tissue site	Frequently seen driver mutations
Esophagus (Barrett’s esophagus)	NOTCH1, TP53, KMT2D, FAT1, NOTCH3, NOTCH2, ZFP36L2, CHEK2, PPMID, PAX9
Skin	NOTCH1, TP53, NOTCH2, FAT1, NOTCH3, ARID1A, KMT2D, CUL3, AJUBA, PI3KCA, ARID2, TP63, NFE2L2, CCND1
Gut (IBD)	NFKBIZ, TRAS3IP2, ZC3H12A, PIGR, HNRNPF
Lung/bronchus	TP53, NOTCH1, FAT1, CHEK2, PTEN, ARID1A, ARID2, IDH1
Endometrium	PIK3CA, PIK3R1, ARHGAP35, FBXW7, ZFHX3, FOXA2, ERBB2, CHD4, KRAS, SPOP, PPP2R1A, ERBB3

Clinical implications

The clinical implications of CH arise due to its association with:

• Aging
• Premalignancy
  • Myeloid malignancy; myelodysplastic syndrome (MDS), acute myeloid leukemia (AML), myeloproliferative neoplasms (MPN)
  • Lymphoid malignancy such as chronic lymphocytic leukemia (CLL)
  • Therapy-related myeloid neoplasms (t-MN) and secondary malignant neoplasms after aplastic anemia
• Pro-inflammatory conditions: increased GvHD risk and association with type 2 diabetes
• Cardiovascular conditions: increased stroke and heart attack risk

Progression

The evolution of CH to myeloid malignancy seems to be controlled by two factors. The first is the inherent fitness effect conveyed by mutations. A study of this effect using mathematical inference modelling on population level data identified the SRSF2 P95R mutation as conferring a strong fitness advantage enabling CH expansion,² which may or may not equate to increased leukemic potential. The second factor is the selection pressure from external factors such as tobacco, alcohol, UV, cytotoxic chemotherapy, and radiation, which may drive progression to a leukemic state.

Two recent case-control publications studied the leukemic potential of CH mutations and reported that P53, IDH, or spliceosome mutations (i.e., U2AF1, SRSF2) had a strong leukemic potential, whereas DNMT3 and TET2 mutations had a lower leukemic potential.² However, caution is advised as the case-control study results may not be representative of the risk seen in an observational study sampling a larger population of genetic aberrations.

CH and t-MN

Professor Takahashi discussed his group’s work on CHIP and t-MN, which has found an association of CH with cumulative incidence of t-MN, a finding which is supported by observations in other cancer centers. A recent study has also confirmed that TP53 and spliceosome mutations are positively associated with an increase in t-MN risk. This group subsequently created a risk prediction model able to predict outcome based on genetic profile, which is soon to be published. In addition, findings from this study also supported the suggestion that it is the interaction between external stressors and the inherent fitness effect that drives the progression of CH. When considering t-MN, this interaction is particularly important due to the selective pressure of chemotherapy. The PPM1D mutation, for example, seems to convey a selection advantage for CH in patients treated with DNA damage chemotherapy, such as cisplatin, but not when treated with vincristine. Therefore, understanding CH driver mutations could be important in determining appropriate therapy to reduce the risk of t-MN.²

Conclusion

To fully understand the spectrum of CH, an integrated study of copy number and point mutations is needed. The clonal expansion seen in hematopoiesis is also found in other tissue types, all of which are exposed to external stressors. As in other neoplastic syndromes, CH driver mutations may convey a fitness effect which is further shaped by external stressors leading to a malignant state. Identification of these CH drivers and their consequences could provide mounting evidence for disease risk prediction using CH.