Single-Cell Proteomic Interrogation of FLT3-ITD Leukemia Compartment Revealsdistinct Leukemia Cell Populations Resistant to Quizartinib, Decitabine and Venetoclax

Abstract

Triplet therapy (TT) combining Quizartinib(Quiz), Decitabine(Dec) and Venetoclax(Ven) is highly effective against FLT3-ITD acute myeloid leukemia (AML) with high response rates (Yilmaz M, ASCO 2022).Signaling mutations, particularly those involving RAS/MAPK pathway, have emerged as key contributors to primary and secondary resistance to TT. Lack or loss of response to TT is not limited to presence or emergence of clones with signaling mutations and disease relapse occurs in FLT3-ITD AML patients with a variety of other mutations. We hypothesize that resistant leukemia clones have unique proteomic profiles that facilitate survival under therapy pressure and deciphering proteomic profiles at the single-cell level will delineate resistance mechanisms and adaptive responses driven by therapy pressure, and aid in identifying proteomic correlates of response and therapeutic vulnerabilities. To this end, we performed comprehensive single-cell proteomic analysis of serially collected samples (n:162) from patients treated with TT (n:14) and doublet (n:10, DEC+Quiz) therapy, using CyTOF and interrogated leukemia phenotypic profiles, apoptotic networks, signaling pathways, differentiation states and trajectories. First, we performed unsupervised analyses to identify leukemia cells and non-malignant cellular elements of the leukemia compartment in an unbiased manner, interrogated the leukemia proteomic landscape to identify leukemia associated proteomic features that may correlate with clinical response and assessed whether distinct mutation profiles are associated with specific proteomic profiles that may confer resistance to therapy. Since Ven targets BCL2 and FLT3i modulates the expression of anti-apoptotic molecules we assessed the expression of apoptosis regulators across leukemia landscapes and found that leukemia cells with immature phenotype, including those with leukemia stem cell(LSC) phenotype, almost always expressed moderate-high levels of BCL2. Monocytic cells(MC), generally situated in close proximity to leukemia cells in UMAPs, lacked BCL2 expression and had the highest levels of MCL1. Despite expressing moderate levels of MCL1 and BCL-XL, TT was effective in substantially eliminating CD34+ immature leukemia cells (Figure 1). Due to their contrasting apoptosis regulatory networks, we observed a relative enrichment of MCs, either benign or malignant, after TT. Since cells with MC phenotype were inherently resistant to TT we hypothesized that leukemia cell subsets having similar proteomic profiles to MCs will persist after TT. Of note, leukemia cell compartments harbor distinct differentiation hierarchies, the frequencies of which varied across patients. Phenotypic interrogation of leukemia cells (UMAP) revealed that LSCs generally clustered on the opposite pole distant from MCs and leukemia cells on poles facing MCs were more differentiated. Remarkably, FLT3-ITD mutation partners differentially altered and reshaped the proteomic landscape. NPM1 mutant FLT3-ITD AML cells displayed a less diverse leukemia architectural organization with differentiation block. Contrarily, signaling mutation diversified the leukemia landscape into a diverse continuum of differentiation states. Remarkably, we observed that RAS/MAPK mutations in FLT3-ITD AML overcame differentiation block and gave rise to a differentiation continuum encompassing less-differentiated, transitional and differentiated leukemia cells(Figure1). Importantly, we observed that transitional and CD34+ leukemia cells, mapped in close vicinity to MCs, and differentiated MCs preferentially persisted at D28, had active signaling pathways and expressed CD36, which indicates distinct metabolic programs could ensure survival under therapy pressure and also had lower levels of BCL2.The presence of CD34+ leukemia cells at D28, adapted under therapy pressure with persistent signaling at D28, both in CD34+ leukemia or MCs, was indicative of poor clinical outcomes. Conclusion: Single-cell proteomic assessment of FLT3-ITD AML treated with Quiz/Dec/Ven (TT) identifies a unique mode of resistance and proteomic features of surviving cells at D28, and elucidates how RAS/MAPK mutations diversify the leukemia proteomic landscape and modulate the architectural organization which eventually gives rise to emerging clones resistant to therapy. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal