Single-Cell Transcriptome Analysis Reveals the Novel Subtypes and Therapeutic Implications of T-Lymphoid/Myeloid Mixed-Phenotype Acute Leukemia

Introduction: Mixed-phenotype acute leukemia (MPAL) is a rare and heterogeneous group of malignant diseases, accounting for 2%–5% of acute leukemia. MPAL can be further classified as B/myeloid (B/My), T/myeloid (T/My), rare types, acute leukemia of ambiguous lineage, not otherwise specified or acute undifferentiated leukemia based on the immunophenotype, according to the WHO criteria. The treatment of acute leukemia has seen significant advances over the last three decades, partly due to individualized treatment. However, clinical prognosis for MPAL patients is worse than both acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML), with a 5-year survival rate of 47-75% in children and 20-40% in adults. Given the rarity of MPAL, historically poorly understanding in the pathogenesis and biology and a lack of prospective trial data to guide therapy, there are currently no established standard treatment guidelines for MPAL to date. The diversity in treatment sensitivity arises from the heterogeneity and unique molecular characteristics of malignant cells. Therefore, there is an urgent need to clarify the molecular subtypes of MPAL and identify potential therapeutic methods to provide personalized treatment strategies and improve the prognosis of the patients with MPAL.

Methods: We conducted an in-depth analysis of single-cell transcriptome data from nearly 260,000 derived from the bone marrow of patients with AML, T-cell acute lymphoblastic leukemia (T-ALL), or T/myeloid MPAL, comparing these data with healthy donors. We employed various analytical methods, including cell subpopulation analysis, pseudotime analysis, cell type mapping, stemness evaluation, and functional enrichment analysis, to identify novel molecular subtypes associated with T/My MPAL. The findings were subsequently validated through external bulk transcriptome data and immunohistochemistry (IHC) staining. Furthermore, we assessed survival prognosis and drug sensitivity differences among patients with these subtypes.

Results: Our study revealed that T/My MPAL malignant cells exhibited unique biphenotypic characteristics, significantly different from traditional AML or T-ALL. Although clinical differences between biphenotypic and bilineal T/My MPAL may exist, they were not significant at the whole transcriptome level. Using a topological manifold learning algorithm, we identified three subpopulations of MPAL malignant cells at the single-cell level, including AML-like, ALL-like, and a unique subpopulation that highly expressed the HOPX gene and exhibited higher stemness and quiescence. Notably, patients with a high prevalence of the unique subpopulation had significantly poorer prognosis. Additionally, we identified specific molecular markers for each subpopulation and validated these findings in multiple independent public datasets. Furthermore, we screened potential therapeutic drugs for each subtype, with venetoclax showing significant efficacy potential for the unique T/My MPAL subtype.

Conclusions: Our research provided new insights into the molecular heterogeneity of MPAL and offered personalized diagnostic and therapeutic targets for patients. By identifying and validating different molecular subtypes and their specific markers, we screened potential therapeutic drugs for each subtype, with venetoclax showing promising therapeutic potential, particularly for patients with the unique T/My MPAL subtype. These findings are expected to improve the prognosis of MPAL patients and advance the development of personalized treatment strategies.