Cellular Plasticity As Mechanisms to Escape from NOTCH1-Inhibition in T-ALL

Abstract

Cellular plasticity has recently emerged as an underlying mechanism for treatment refraction in cancer. Hence, characterizing cell states and the transcriptional changes required for state transitions is critical for the development of efficient targeted therapies. Acute lymphoblastic leukemia (T-ALL) is an aggressive hematopoietic malignancy in children and adolescents that is associated with high rates of treatment failure and early relapse. T-ALL patients frequently harbor NOTCH1 activating mutations as the driving oncogene in this disease. Using T-ALL patient-derived xenograft (PDX) models carrying activating NOTCH1 mutations, we aim to u First, to determine how cell state diversity drives treatment refraction in T-ALL, we performed full-length single-cell transcriptome analysis on 3188 immune cells collected from 2 sensitive and 3 refractory PDX models. PDX models were treated with either with NOTCH inhibitor DBZ (Dipenzazepine; 10 μM/kg every other day intravenously) or vehicle. Profiled cells were collected at both a short (5 days) and a late time point (x-150 days) post-treatment to assess the effects of DBZ. Analysis of early hematopoietic and thymic precursor signatures demonstrate expression of immature hematopoietic signatures and acquisition of alternative lineage identities coexisting within the same cell, predominantly in refractory models. Analyses of pre-treatment samples indicated that disruption of the normal developmental hierarchy was associated with blockade at an early stage of T-cell commitment and can predict response to NOTCH inhibition in vivo. Upon treatment, cells underwent dramatic transcriptional reprogramming resulting in expansion of 2 distinct populations that greatly differ in both differentiation stage and developmental trajectory. One population expresses lymphoid differentiation programs, whereas the other population is enriched in genes associated with RNA processing and more immature hematopoietic signatures. Cell fate trajectories assessment using Monocle 2, revealed that consistent with their immature hematopoietic signature, refractory cells map at the start point of the differentiation trajectory, while sensitive cells clustered at the endpoint. We next sought to characterize the differentiation hierarchies that define the immature states. Developmental potential as assessed by CytoTRACE was higher in refractory cells when compared to their sensitive counterpart. To identify determinants of cell fate in a functionally relevant manner, we inferred gene regulatory network configurations using SCENIC and defined state specific regulon activity. Among the most prominently expressed regulons present in mature states, we identified transcriptional regulators involved in T-cell differentiation such as IKAROS, GATA3, and SOX5, while in immature states we observed high transcriptional regulon activity in transcription factors associated with renewal of hematopoietic progenitors such as ATF4, BCLAF1, and MYC in addition to the chromatin remodelers SMARCA4 and EZH2. To assess whether transcriptional rewiring in refractory cells results from greater chromatin accessibility, we performed single-cell ATAC sequencing (10x Genomics) on splenic cells from 1 sensitive and 1 refractory T-ALL PDX model treated with DBZ or vehicle. After integration with corresponding single-cell RNA-sequencing data, annotated ATAC profiles demonstrated greater chromatin accessibility, predominantly in active promotor and enhancer regions, in the refractory model compared to the sensitive model. Peak to gene linkage analysis revealed widespread alterations in cis-regulatory regions near genes associated with immature hematopoietic precursor signatures such as CD33. Furthermore, greater chromatin accessibility in refractory cells was associated with an increased number of enhancer loops per expressed gene as determined by H3K27ac HiChIP analysis. In conclusion, we demonstrated that presence of highly plastic cellular states defined by aberrant differentiation trajectories, distinct transcriptional circuitries, and remodeled chromatin architecture results in treatment escape in T-ALL.