High dimensional mapping of temporal evolution within the marrow microenvironment in response to FLT3 inhibitor therapy.

Abstract

7020 Background: The advent of genomic sequencing technologies has revealed underlying genetic alterations, such as FLT3 mutations, that can be targeted in acute myeloid leukemia (AML). However, development of resistance limits the durability of response. Recent data has implicated that factors from the bone marrow microenvironment mediate initial resistance to FLT3 inhibitors (FLT3i) in AML. We combined high dimensional characterization techniques, time-of-flight mass cytometry (CyTOF) and RNA sequencing, to examine sequential marrow stromal samples from a subset of patients with FLT3 mutated AML treated with the FLT3i gilteritinib. Here, we report on the heterogeneity and evolution of cell surface and secreted factors over time. Methods: RNA sequencing of primary FLT3-ITD stromal samples (N = 29) from pre-study and on-treatment patients enabled prioritization of candidate targets for CyTOF. Target-specific purified antibodies were purchased pre-conjugated to metal lanthanides or conjugated in house according to manufacturer protocols (after validation via traditional flow cytometry). Primary stromal cells were cultured ex vivo until confluent and were harvested and stained according to a standardized protocol, and subsequently run on a Helios (Fluidigm) mass cytometer. A computational approach was employed to compensate and visualize data via the CATALYST R package. A total of four pre-treatment and four post-gilteritinib timepoint isolates (N = 8) were analyzed. Results: A 36-target mass cytometry panel revealed protein level differences in patients before and after gilteritinib therapy. Dimensional reduction techniques such as MDS and UMAP showed that samples taken from later timepoints clustered together compared to their earlier counterparts with respect to global protein expression. Inflammatory mediators such as IL1-beta and MCP-1 were upregulated in patient stroma soon after gilteritinib treatment and therefore potentially contribute to early resistance. Novel markers previously implicated in early resistance to targeted therapies in AML such as FGF2 and FGFR1 similarly peaked earlier in treatment, mimicking the clinical course of expression observed in marrow stroma of patients treated with another FLT3 inhibitor quizartinib (Traer et al. Cancer Res. 2016). Conclusions: Our findings show that primary marrow stroma evolves during gilteritinib treatment, and that stromal proteins previously reported to promote early resistance to FLT3i are also upregulated during gilteritinib resistance. The heterogeneity of stromal cell isolates detected by mass cytometry highlights the utility of high dimensional tracking of disease course in patients, and may enable a better understanding of how the temporal evolution of the marrow microenvironment contributes to development of resistance to targeted therapies such as gilteritinib and other FLT3i over time.