Reconstructing the Lineage Histories and Differentiation Trajectories of Individual Cancer Cells in Myeloproliferative Neoplasms

Abstract

Some cancers originate from a single mutation event in a single cell. Blood cancers known as myeloproliferative neoplasms (MPNs) are thought to originate when a driver mutation is acquired by a hematopoietic stem cell (HSC). However, when the mutation first occurs in individuals and how it affects the behavior of HSCs in their native context is not known. Here we quantified the effect of the JAK2-V617F mutation on the self-renewal and differentiation dynamics of HSCs in treatment-naive individuals with MPNs and reconstructed lineage histories of individual HSCs using somatic mutation patterns. We found that JAK2-V617F mutations occurred in a single HSC several decades before MPN diagnosis-at age 9 ± 2 years in a 34-year-old individual and at age 19 ± 3 years in a 63-year-old individual-and found that mutant HSCs have a selective advantage in both individuals. These results highlight the potential of harnessing somatic mutations to reconstruct cancer lineages.

Keywords: JAK2; blood cancer; lineage tree; myeloproliferative neoplasm; single cell sequencing; stem cell dynamics.

Graphical abstract

Figure 1

Experimental Design (A) Individual hematopoietic stem and progenitor cells (HSPCs) from bone marrow aspirates of individuals with MPNs were analyzed in two ways. First, hematopoietic stem cells (HSCs) and multipotent progenitors (MPPs) were expanded in vitro and characterized using WGS. Second, we simultaneously read out the transcriptional profiles and somatic mutations in single HSPCs. (B) Information about the individuals with MPNs sampled in this study. “Allelic burden peripheral blood (PB)” and “secondary mutations” refer to VAFs of JAK2 mutations and other hematopoiesis-associated mutations in PB, respectively. The numbers of JAK2 WT and mutant cells identified in the HSPCs using scRNA-seq are given in the last two rows. See also Figure S1.

Figure 2

Erythroid and Megakaryocyte Progenitors from Individuals with MPNs Are More Likely to Have the JAK2-V617F Mutation than Other CD34+ Bone Marrow HSPCs (A) UMAP of CD34-enriched bone marrow scRNA-seq data from ET 1, colored by cell type. (B) Marker gene expression in ET 1 CD34-enriched bone marrow. (C) Cell type classifications and JAK2 WT/mutant transcript calls in scRNA-seq data from individuals with MPNs (columns). (D) Fraction of JAK2 mutant cells (colors) in different bone marrow cell types from individuals with MPNs. (E) Relationship between the PB VAF and JAK2 V617F mutant transcript fraction in bone marrow HSCs (blue) and erythroid progenitors (red). Error bars indicate 95% confidence intervals. (F) Mean expression of selected marker genes in CD14+ cells that are upregulated in monocyte subsets or were differentially expressed in CD14+ cells between individuals with ET and PV or between individuals with MPNs and healthy controls. See also Figure S2.

Figure 3

Somatic Mutations Can Be Used to Reconstruct the Lineage Trees of WT and Mutant HSCs (A and B) Lineage trees constructed using somatic SNVs for ET 1 (A) and ET 2 (B). The heatmap below the lineage trees shows the relative contribution of the SBS mutational signatures SBS1, SBS2, SBS5, SBS19, SBS23, and SBS32 (STAR Methods) to the mutational spectrum defined by the private mutations detected in each HSC-derived colony. (C) Number of mutant stem cells as a function of time inferred from the ET 1 lineage tree, assuming one generation per year. The dashed lines on the bottom show the times of the coalescent events in the tree. (D) Same as (C) but for individual ET 2 See also Figure S3 and Table S2.

Figure 4

The History of JAK2-Mutant HSC Expansion Is Reconstructed from the Lineage Trees (A) Schematics showing the effect of scaling the number of generations by a factor of 2 while keeping the onset of the disease and fitness the same. As a result, the number of mutant cells doubles because early on the number of mutant cells (shown in white) fluctuates to increase by a factor of 2 to escape stochastic extinction. Increasing the number of generations increases the coalescent rate. Increasing the number of mutant cells decreases the coalescent rate. These effects cancel each other, and the trees are indistinguishable. (B) Green curves (c = 1) represent 50 simulated mutant HSC trajectories that survived extinction with fitness s = 0.8 and a maximum population size of 50,000. Blue curves are similar, except the number of generations and maximum population size are scaled by a factor of 1,000 (c = 1,000). This scaling results in 1,000 times as many mutant HSCs through time (blue) because a larger initial population is needed to escape stochastic extinction. (C) Trees corresponding to the blue and green trajectories in (B) are statistically indistinguishable (STAR Methods). (D) Inference on data from ET 1 and ET 2. For both individuals, we show 50 inferred trajectories of the number of mutant stem cells as a function of time. Heatmaps show the inferred joint distribution of the fitness of the cancer cells and the age when the disease initiating mutation occurred. The marginal distributions are shown as histograms. See also Figure S4.