T-ALL is an aggressive leukemia predominantly affecting adolescents and young adults. T-ALL relapses are characterized by chemotherapy resistance, cellular heterogeneity and dismal outcome (PMID: 33925883). While mutations, changes in gene expression and epigenetic dysregulations have been documented (PMID: 26294725; 27655895; 35585141; 23023710; 27623214), a unifying mechanism explaining the development of relapse remains unknown.

To comprehensively investigate the heterogeneity of T-ALL and gain a better understanding of mechanisms driving T-ALL relapse, we conducted single-cell full-length RNA sequencing utilizing VASA-seq (PMID: 35760914). This analysis included 13 matched pediatric T-ALL patient-derived xenografts (PDX) samples obtained at initial diagnosis and relapse, generating the to date most comprehensive longitudinal single cell study in paired T-ALL samples, along with 5 non-relapsing PDX samples collected at initial diagnosis. Our dataset encompasses 11 TAL1- 3 TLX1/2-, 2 NKX2- and 2 HOXA-driven T-ALLs patients.

While the predominant cell populations in these patients exhibit substantial interpatient heterogeneity, our data unveil a previously elusive subpopulation of T-ALL cells that converges at a gene-regulatory network shared between the majority of patients. This subpopulation is characterized by a molecular stem-like cell phenotype including immaturity in differentiation, persistence in the G1 cell cycling phase, metabolic quiescence and increased cell adhesion.

Among the upregulated genes, we identified several surface markers previously recognized for their role in T-ALL relapse (PMID: 21487112; 20231613) or in stem-like cell maintenance in a murine T-ALL model (PMID: 29781813; 38553571), such as CD44, CD226, CD52, CD97, CD27 and ITGB7. We further identified significant enrichment of the anti-apoptotic proteins BCL-6, BCL-2 and MCL1. By contrast, NOTCH1 is among the downregulated genes in the stem-like cell population when compared to the other leukemic cell populations.

We utilized the SCENIC computational analysis method (PMID: 28991892; 32561888) to identify those transcription factors (TFs) that are driving the expression of most genes in the cluster, implying a particularly relevant biological role. Five TFs were found to be significantly less active in the stem-like cells (Fisher's exact test, padj <0.05), all of which are involved in cell cycling (TFDP1, E2F1, BRCA1, E2F2, E2F3), further supporting the quiescent phenotype of these cells. The 14 TFs with significantly enriched activity in the stem-like cell population (Fisher's exact test, padj < 0.05) are grouped into 6 different TF families: KLF, AP1, NF-κB, ETS, FOXO and PIAS family - all key regulators of the hematopoietic development. Notably, KLF2, one of the significantly enriched TFs, is also a key driver of dormancy, migration and anti-apoptotic signaling in healthy T cells (PMID: 9302292; 11477405; 18246069).

Remarkably, our data show an expansion of the stem-like cell population from 0.46% - 2.16% at initial disease to 10.68% - 43.52% at relapse (paired t-test: p = 6.6e-05) supporting the hypothesis that stem-like cells may originate from pre-leukemic subclones in early phases of the disease (PMID: 38553571; 33414170; 29034206; 26294725). Notably, this subclonal expansion is more prevalent in the TAL1-driven T-ALL consistent with the generally increased subclonal heterogeneity of this subtype compared to other T-ALL subtypes (PMID: 35585141).

Taken together, our scRNA-seq data indicate the presence of an initially small and subsequently expanding subpopulation of cells with a treatment resistance stem-like cell phenotype. We functionally validated this prediction by in-vitro and in-vivo drug testing showing preferential survival of this small subpopulation following exposure to chemotherapy. Our future studies will explore how specific targeting of the transcriptional network of the stem-like cell defined here can be harnessed to overcome treatment resistance in T-ALL.

Disclosures

Schrappe:JazzPharma, Servier, Amgen: Honoraria, Research Funding, Speakers Bureau. Kulozik:Vertex: Honoraria.

This content is only available as a PDF.
2024
Sign in via your Institution