Clonal Selection of Mutant HSC in MPN: “What Do I Know?”

Myeloproliferative neoplasms (MPNs) are thought to arise through the acquisition of a JAK/STAT pathway mutation in a single hematopoietic stem cell (HSC). The most common mutation, JAK2 V617F, is detected in the majority of patients with MPNs, and JAK2 inhibitors have shown clinically meaningful benefit for some types of MPNs – even in the absence of the JAK mutation. Furthermore, JAK2 V617F causes an MPN phenotype in mouse models. On the face of it, the relationship between the JAK2 V617F mutation and disease development seems clear. However, several observations suggest the link between JAK2 V617F mutation and development of clinically overt MPN is more complex, with much still to learn. First, the prevalence of JAK2 mutations in the general population in persons with normal blood count parameters — so-called clonal hematopoiesis of indeterminate potential (CHIP) — far exceeds the prevalence of JAK2 mutation–positive MPNs. This suggests that most individuals that acquire a JAK2 V617F mutation never develop clinically overt MPN. Second, the same JAK2 V617F mutation causes a range of different MPN phenotypes, including essential thrombocythemia, polycythemia vera, and primary fibrosis, suggesting that the mutation has a different impact depending on the individual person and cellular context of mutation acquisition. Third, although possible, it is extraordinarily difficult, even in mouse model systems, to induce MPN disease from a single HSC, suggesting that the impact of JAK2 V617F on HSC function is complex and influenced by other unknown factors.

Two recent preprints have provided further cause for reflection with regards to our understanding of MPN development following acquisition of JAK2 mutation in MPN.^1,2  These studies used state-of-the-art lineage tracing approaches to determine the timing of clonal expansion and disease development following acquisition of the JAK2 V617F mutation. With the caveat that peer review is pending on these papers, it is worth noting that both studies made the remarkable and thought provoking observation that the JAK2 V617F mutation is typically acquired decades before disease development; in many cases, the mutation was acquired in utero or in early childhood and yet only caused disease after many decades in adult life. Together with the observation that most patients with JAK2 V617F–positive CHIP never develop MPNs, this suggests that many persons acquire a JAK2 V617F mutation and live with this mutation for decades, and perhaps in many cases, lifelong, without ever developing disease. So why do certain individuals develop MPN when a JAK2 V617F mutation is acquired by an HSC, while others do not?

To confer a disease phenotype, a crucial step in the disease pathogenesis is for the JAK2 V617F-mutant HSC to undergo clonal selection. Typically, persons with JAK2 V617F mutant–positive CHIP only have a very small JAK2 clone in comparison with persons who develop JAK2 V617F-positive MPN. So why does the JAK2 V617F mutation exert a clonal advantage in some individuals to develop MPN but not in others? A recent study by Dr. Erik Bao and colleagues made an important observation regarding the impact of inherited polygenic risk on the function of HSC during MPN development.³  Numerous epidemiologic studies had previously reported the presence of genetic predisposition to MPN development. A first-degree relative with MPN confers an approximately sevenfold risk of getting an MPN compared to the general population.⁴  In 2009, three reports highlighted the existence of a germline JAK2 variant (named JAK2 46/1) that is strongly associated with the acquisition of somatic JAK2 V617F.^5-7  JAK2 46/1 has also been associated with JAK2 V617F-wildtype MPNs.⁸  Other MPN genetic predisposition alleles have subsequently been reported, describing additional risk variants in TERT, MECOM, and HBS1L-MYB. Nevertheless, only a few loci have so far been identified, and the biological mechanisms that link genetic predisposition to MPN development have not been determined.

Dr. Bao and colleagues reported on the results of a very large genomewide association study (3,797 patients with MPNs and 1,152,977 healthy controls). By performing a meta-analysis using publicly available data sets (FinnGen, 23andMe, UK Biobank-UKBB, Million Veteran Program), 17 independent genetic risk alleles reached significance, including seven previously unreported alleles. Altogether, these loci account for around 18.4 percent of the total familial risk for MPNs. Furthermore, additional meta-analysis allowed the investigators to generate a 104-variant polygenic risk score that identifies increased risk of MPNs in high-scoring individuals (up to 2.7 odds ratio in individuals at 90th risk score percentile). The newly generated risk score was able to refine risk stratification even within JAK2 46/1 carriers.

Next, to understand the biological mechanisms that link genetic predisposition to MPN development, the authors cross-referenced the genomic location of their most informative MPN risk variants with chromatin accessibility data obtained from ATAC-seq studies across different hematopoietic populations. This elegant analysis showed that MPN risk alleles were strongly associated with open chromatin in HSCs and early progenitors, suggesting a biological function of these risk variants in HSCs. Furthermore, to gain more insight into the biology of heritable MPN risk, they mapped the variants to target genes: While three alleles were located within exons and caused missense mutations (in ATM, SH2B3, and CHEK2), other alleles falling in noncoding regions could also be assigned to specific genes. Analysis of biological annotations and gene expression data clearly showed that these genes are enriched for HSC expression and function.

Finally, as proof-of-principle, the investigators sought to determine whether some of these variants have a functional effect on HSC biology. First, they demonstrated that mimicking of a CHEK2 hypomorphic variant risk allele by inhibiting CHEK2 conferred to HSC the ability to resist cell death when exposed to irradiation, while also promoting increased cell expansion in long-term cultures. The authors also analyzed a risk allele falling into an enhancer region of GFI1B, concluding that the deleterious variant decreases GFI1B expression as determined by in silico and reporter assays. Deletion of the GFI1B enhancer increased HSC numbers in liquid cultures as well as their replating potential.

In summary, 2020 has been an exciting year for our understanding of the pathogenesis of MPNs. It is now apparent that MPN often develops over a very prolonged period, and disease risk is strongly influenced by genetic risk alleles that are associated with the function and self-renewal of HSC. A common question raised by patients diagnosed with MPN is, “How long have I had it?” Up to now it has been a difficult one to answer; explaining to patients that they might have acquired the JAK2 mutation in utero or childhood will raise many questions that we can currently only provide partial answers to. While the very long latency for MPN development and genetic risk alleles may lend themselves to early detection, the health economic argument for screening and early intervention in MPNs is not straightforward. It also remains unclear how the MPN risk variants modulate HSC biology in the presence or absence of JAK2 V617F mutation. This will require new model systems. The impact of other/cooperating mutations also remains unclear. Do these MPN risk alleles also influence risk for other hematologic malignancies with HSC as the cell of origin (e.g., myelodysplastic syndromes)? Although genetic risk alleles help to explain how germline genetics can influence HSC biology, clearly, other nongenetic factors also influence MPN risk. Notwithstanding these remarkable and unprecedented insights into MPN made in 2020, the reason most persons who acquire JAK2 V617F do not develop MPN still remains largely unknown. One important factor might relate to heterogeneity of HSCs; the relationship between different HSC subtypes, including HSCs early in development, and the impact of JAK2 V617F remains unknown. The role of aging, inflammation, infection, chemotherapy, and other environmental exposures as triggers for clonal expansion of JAK2 V617F–mutated HSC remains an area of intensive research. While we certainly understand a lot about MPN biology, there are many things we have yet to learn, reminding us of the famous motto of Michel de Montaigne, “What do I know?”