Overview

The Miles Lab focuses on understanding the mechanisms that underlie the evolution and development of acute myeloid leukemia. We aim to understand how mutations synergize to drive disease development, progression, and maintenance with the goal to discover novel therapeutic strategies that improve AML patient outcomes. We aim to answer crucial questions through a combination of single cell evaluations of AML patient samples and multi-allele mouse modeling of AML in the following ongoing projects:

Clonal Evolution of AML

Bulk DNA sequencing of adult AML patient cohorts has suggested that AML develops in a stepwise manner, where a new mutation is acquired within a pre-existing mutant clone. This new mutation presumably alters the cell that harbors it – potentially increasing clonal fitness, impacting cellular signaling or epigenetic regulation, and/or changing its self-renewing and differentiation capacity through cooperation/synergy with the prior existing mutations - i.e. generating a new clone with unique properties.

Until recently, it could be difficult to determine the clonal make-up within an individual AML patient using bulk sequencing. Newer approaches have been developed and optimized that allow for genomic profiling of patient samples at single cell resolution. Single cell DNA sequencing of patient samples now provides a unique opportunity to delineate dynamic changes in clonality, clone fitness, and immunophenotype/cell state during clonal evolution and in response to therapeutic perturbations (Miles, Bowman et al. Nature 2020). These studies have allowed us to begin to connect mutational combinations to cell state and lineage bias within hematopoiesis.

Ongoing projects in the lab continue to generate delineate genotype-immunophenotype relationships using single cell mutli-omic sequencing of adult AML samples. To dissect the underlying mechanisms behind these relationships we uncover in AML patients, we utilize inducible, multi-allele mouse models combining frequent AML mutations. A combination of flow cytometry, stem cell assays, ex vivo co-cultures, and in vivo experiments allow us to examine how co-mutations synergize to drive transformation and leukemic progression.

Current Projects

• Divergent mechanisms in Tet2 and Idh2 mutant AML
• Clonal evolution of Npm1 mutant AML
• Clonal heterogeneity in pediatric AML

Oncogene Plasticity

Large scale sequencing cohorts of patients have given us a great picture of the mutational landscape of adult AML. The identification of frequent AML mutations led to the development and approval of therapies against some of these mutations, i.e. IDH1, IDH2, and FLT3. However, there is a large subset of patients that do not harbor these mutations. Moreover, it is still unclear how dependent leukemic cells are on individual mutations once they have fully transformed. Insight into these dependencies could provide the optimal targets of new therapies for patients who do not harbor mutations targeted by current therapies.

To assess how dependent mutant cells are on a particular mutation, we utilize a novel mouse model originally developed by Drs. Ross Levine (postdoc mentor) and Robert Bowman (collaborator). The model, termed GOLDI-Lox for Governing Oncogenic Loci by Dre Inversion and Lox reversion, utilizes a dual recombinase Dre-on, Cre-off construct in the endogenous loci that allows us to toggle an allele from wildtype (WT) to mutant to WT. Its’ unique design also allows us to combine this allele with any other Cre-inducible alleles to model co-mutations observed in patients. This model provides the ability to assess dependency at any disease state, from pre-leukemic dysregulation to overt AML. By establishing a pre-leukemic or leukemic state and then reverting a mutant allele back to WT, we can determine whether the presence of that mutation or mutant protein is essential and/or sufficient for disease development, progression, and maintenance. Through a combination of techniques including but not limited to flow cytometry, pathologic assessment, and stem cell functional assays, we are able to delineate dependency by the hematopoietic cell compartment and evaluate how co-mutations affect this dependency. Lastly, mechanistic dissection of this dependency provide insights into the critical consequences of a mutation that are requisite for leukemogenesis and has the potential to uncover pathways/functions that are potential targets of new therapies.

Mutation Order

Mutation order has often been inferred from bulk sequencing of AML samples by the variant allele frequency (VAF) of mutations - i.e. the earlier the mutation happened in disease development, the larger that mutant clone will be as it had more time to expand, therefore higher VAF equals earlier mutation. Given the frequent high VAFs of mutations in epigenetic modifying genes, such as DNMT3A, TET2, IDH1/2, these mutations are hypothesized to occur early in disease development. Strenthening this hypothesis, recurrent mutations in a subset of the epigenetic modifier genes are also found in clonal hematopoiesis, where mutant hematopoietic cells are present in healthy individuals without overt disease. Conversely, mutations in signaling genes such as NRAS or FLT3 typically present with lower VAFs and are thought to occur later in disease progression, most likely as a subclone. However, some AML mutations, such as NPM1, present with a wide range of VAFs depending on the patient. Even more, overlapping VAFs between 2 or more mutations also convolute the ability to predict mutation order.

Single cell multi-omic analysis of patient samples has the capability to more optimally determine mutation order, given that we can identify individual clones in a sample. Based on the presence or absence of clones with particular combinations of mutations, we can better hypothesize the order of mutagenesis. Important questions still have yet to be answered, though, including: Does mutation order matter for disease development? Does the order of mutagenesis affect disease phenotype and/or response to therapies? To answer these questions, we have generated dual recombinase mouse models that allow us to (in)activate single mutations in a precise order within the same cell and evaluate the impact of mutation order on clone fitness, disease development/phenotype, and response to therapies. We use a combination of flow cytometry and peripheral blood analysis of in vivo models and ex vivo co-cultures to identify how mutation order affects leukemogenesis.

Current Projects

• Mutation order in the pre-leukemic state
• Mutation order impact in overt AML transformation