Reconstruction of cancer cell lineage history in patients with MPN

The JAK2 V617F mutation is common in the majority of Philadelphia chromosome-negative myeloproliferative neoplasms (MPN), where the activation of JAK2 signaling triggers the overproduction of mature blood cells of the myeloid lineage, ultimately leading to MPN.

However, it is not known when this mutation first occurs and how it affects the differentiation and proliferation of hematopoietic stem cells (HSCs). In this context, Van Egeren and colleagues evaluated patients with newly diagnosed, JAK2 V617F⁺ essential thrombocythemia (ET) and polycythemia vera (PV) to investigate when the JAK2 V617F mutation first arises in stem cells, how the mutation affects the fitness of cells, and how it changes trajectories between JAK2-mutant cells and wild-type (WT) cells. The study was recently published in Cell Stem Cell, and here we report the key findings.¹

Experimental design

The experimental design of this study allowed investigators to evaluate the impact of the JAK2 V617F mutation on differentiation and self-renewal of HSCs and gain insight into the timeframe from the acquisition of the mutation to the complete development of MPN. The investigators analyzed bone marrow aspirates of patients with MPN in two ways:

• Single-cell transcriptomic profiling and amplicon sequencing of somatic mutations of hematopoietic stem and progenitor cells (HSPCs) from seven treatment-naïve patients with MPN (ET, n = 4; PV, n = 3) and two healthy controls was performed, to investigate how the JAK2 V617F mutation affects the differentiation trajectories of HSPCs.
• The phylogenies of the stem cells were reconstructed from the pattern of somatic mutations using Wagner parsimony. To infer the history of MPN development, HSCs and multipotent progenitors were isolated from patients #ET1 and #ET2, expanded in vitro for  ~8 weeks, and then characterized by whole-genome sequencing (WGS) on single-cell colonies. Both mutated and WT colonies were selected to reconstruct the lineage trees with balanced number of cells.
  • #ET1: 22 JAK2-mutant colonies and 20 WT colonies
  • #ET2: 13 JAK2-mutant colonies and 21 WT colonies

The difference in growth rates between JAK2-mutant and WT HSCs was estimated using a mathematical model of stem cell self-renewal based on the Wright-Fisher model. This model considered three parameters: maximum number of mutated stem cells, age at which the disease-initiating JAK2 mutation occurred, and fitness of mutant stem cells.

Results

Single-cell transcriptomic profiling and JAK2 genotyping

Information on patients sampled in the study are shown in Table 1. Out of seven patients, six had the JAK2 V617F mutation and the remaining patient (with ET) had the JAK2 V617L mutation, previously unreported in humans. No additional hematopoiesis-associated mutations (secondary mutations) were observed in patients with ET, while patients with PV carried somatic truncating mutations in TET2 and EZH2. Additional hematopoiesis-associated mutations were detected using next-generation sequencing assay performed on whole white blood cells from peripheral blood.

Table 1. Patients with MPN sampled in the study¹

Full transcriptome profile and genotyping of JAK2 mutations in individual CD34⁺ cells revealed that gene expression profiles were similar between JAK2 V617F and WT cells from the same individuals. Also, a significant fraction of HSCs was mutated in all individuals (range, 5─62%).

The JAK2-mutant fraction varied in myeloid compartments, with a mutation frequency of JAK2 V617F higher in megakaryocyte/erythroid progenitors and lower in lymphoid and granulocyte-macrophage progenitors, suggesting a megakaryocyte-erythroid lineage bias.

Lineage trees

Lineage trees of JAK2 V617F-mutated HSCs were reconstructed in patients #ET1 and #ET2, using the pattern of somatic mutations accumulated in single cells:

• #ET1 (the younger patient) had 713 ± 45 somatic point mutations in individual HSCs/ multipotent progenitors.
• #ET2 (the older patient) had 1,185 ± 75 mutations in each cell.

A constant somatic point mutation rate of 19 ± 1 per year was estimated using the number of point mutations found in each cell and the age of each individual. The number of somatic mutations in JAK2-mutant cells (#ET1, n = 732 ± 26; #ET2, n = 1,209 ± 35) and JAK2 WT cells (#ET1, n = 690 ± 52; #ET2, n = 1,170 ± 89) was comparable, and the difference in mutation rate was not significant, indicating JAK2 V617F mutation had no impact on the somatic mutation rate. However, a shorter average telomere length in the JAK2-mutant cells in both of these patients (p < 0.01) suggested that JAK2-mutant cells might have undergone more cell divisions than JAK2 WT cells.

JAK2 V617F was the only deleterious somatic mutation detected in JAK2-mutant cells, thus this mutation is probably the disease-initiating MPN driver mutation in these two patients.

The phylogenies reconstruction, where two distinct clades defined by the presence or absence of the heterozygous JAK2 V617F mutation were found in each individual, suggested that a single JAK2 mutation event initiated the disease. While no shared somatic mutations were observed across JAK2 WT stem cells in either patient, many somatic mutations were shared across JAK2-mutant stem cells in #ET1 (n = 220) and in #ET2 (n = 398), suggesting a single common ancestor in which the JAK2 mutation firstly occurred. It was estimated, by the inferred somatic mutation rate, that the disease-initiating mutation occurred:

•   ̴25 years prior to sampling in patient #ET1 
•   ̴40 years prior to sampling in patient #ET2

Analyses on the difference in growth rates between JAK2-mutant and WT HSCs by the Wright-Fisher model revealed a selective fitness advantage of JAK2-mutant cells over WT cells, with the JAK2 V617F mutation first occurring at:

• Age 9 ± 2 years with a fitness effect of 63% ± 15% in patient #ET1
• Age 19 ± 3 years with a fitness effect of 44% ± 13% in patient #ET2

Conclusion

Single-cell profiling and WGS of HSPCs from the bone marrow of newly diagnosed patients with MPN revealed that JAK2 V617F acquisition occurred decades before the diagnosis of MPN, expanding exponentially with large fluctuations. At the time of diagnosis, a significant fraction of HSCs (≥5%) were found to be descendants of the original JAK2-mutant HSC. Also, other than affecting the proliferation dynamics, the JAK2 V617F mutation induced a fitness advantage and lineage bias toward the erythroid and megakaryocyte fate. However, there was a significant variation in the fraction of JAK2-mutant cells among different progenitor cell populations in the same individual. 

Limitations of this study included the small cohort of patients and the limited number of HSCs from each individual analyzed with WGS.