Analysis of Mechanisms Underlying Clonal Evolution of AML By a New Single-Cell Sequencing Platform

Background

Leukemic cell populations are highly heterogeneous in terms of both gene mutations and gene expression, which is shaped by acquisition of multiple mutations and expansion of adapted clone. This evolutional process is clinically important because it is observed in the contexts of treatment resistance and relapse as well as leukemic transformation, and molecular mechanisms involved in clonal selection can be exploited as a therapeutic target. Nevertheless, direct analysis of such mechanisms in patients' cells is hampered by technical difficulties to characterize both clonal structure and gene expression at a single-cell resolution. On this issue, we have recently developed a new method which enables simultaneously detection of mutations and whole transcriptome information at single-cell level by extensively modifying an existing single cell RNA-seq (Nakagawa et al. ASH abstract 2018). The aim of this study is to understand heterogeneity of clones and to clarify mechanisms behind clonal expansion in AML by longitudinal analysis using our novel single-cell sequencing platform.

Results

In order to estimate clone frequencies and select samples to be analyzed by single-cell sequencing, we first sequenced bulk bone marrow cells from patients with AML. Of interest, we found that AML samples frequently harbored multiple clones having different Ras pathway mutations, most frequently involving NRAS, which exhibited dynamic change in their clone size during the course of AML. These are interesting targets of the analysis of mechanism of clonal evolution of AML. Thus, three patients having multiple (n=3-5) Ras pathway mutations were investigated by sequencing their bone marrow Lin^-, CD34⁺ cells using the newly established single-cell method, which successfully separated distinct clones having distinct mutations, where all of detected Ras pathway mutations were present in independent clones as expected.

In order to examine these independent clones with Ras pathway mutations might show equal or heterogenous cellular phenotypes, proliferation or differentiation statuses as determined from transcriptome data was analyzed for all detected NRAS mutated clones. Among the NRAS mutated clones, some showed significant increase in proliferation-associated gene expression signature (calculated as proliferation score) compared with NRAS wild type clones, and no NRAS mutated clones showed decrease of the score, which is consistent with pro-proliferative function of Ras pathway. Interestingly, such increase in proliferation showed considerable heterogeneity among clones, where some NRAS mutated clones showed greatly increased proliferation scores compared to other NRAS mutated clones. Differentiation statuses of NRAS clones also showed heterogeneity among clones.

In order to examine whether this inter-clone proliferation difference correlates with clone dynamics, we then analyzed longitudinal bone marrow samples for a patient who showed different proliferation between clones. The NRAS mutated clone with highly increased proliferation compared with wild type clone (NRAS p.G12S) had undergone rapid expansion in 3 months (cell frequency 0.08 to 0.74) in spite of continuous azacitidine treatment, while the NRAS mutated clone with less increase in proliferation (NRAS p.G12D) had showed regression (cell frequency 0.72 to 0.14). To investigate the mechanism of this therapy-resistant clonal expansion, we compared transcriptome data of these clones. Unlike the regressed clone, the expanded clone uniquely exhibited increase in expression of genes in PI3K/AKT pathway and unfolded protein response (UPR) pathway, one of cellular stress response pathway. UPR is recently reported to responsible for the promoted survival and competitive advantage in mouse hematopoietic stem cells with Nras mutations (Liu et al. Nat. Cell Biol. 2019). Our data suggest that the enhanced UPR pathway contributes to clonal expansion also in human AML with Ras pathway mutations.

Conclusions

Using a newly developed single-cell sequencing platform, we have successfully characterized gene expression profiles associated with clonal evolution of AML with Ras pathway mutations. Simultaneous measurement of both mutations and transcriptomes at a single-cell level will help understand the mechanism of clonal evolution of AML.