Single-cell multi-omics reveals co-mutation of TP53 and epigenetic gene driving myeloid transformation in B-ALL following CAR-T therapy

Lineage switch is a rare phenomenon, occurring in less than 1% of leukemia cases. However, with the advent of CD19-targeted immunotherapies such as CAR-T and blinatumomab, it has emerged as a significant complication. Identifying patients at high risk for lineage switch prior to immunotherapy is critical, yet predictive biomarkers remain unclear. In a cohort of 65 B-ALL patients who relapsed after CAR-T therapy, we identified one case of myeloid lineage switch with unknown mechanism. Transcriptomic sequencing across three timepoints: pre-CAR-T, first CAR-T relapse, and second CAR-T relapse (myeloid switch) confirmed a consistent driver gene profile ( ETV6 :: RUNX1 fusion, NF1 , WNK1 and CIC mutations), ruling out a secondary malignancy. To elucidate the mechanism, we performed single-cell DNA-protein sequencing (Mission Bio Tapestri platform) and single-cell RNA sequencing (10x Genomics platform). At first CAR-T relapse, we identified five distinct clones. After CAR-T therapy, three clones were undetectable, whereas two clones persisted and underwent rapid expansion, concomitant with myeloid lineage switch. Notably, both of these clones harboring mutations in TP53 and epigenetic regulator STAG2 . We further discovered a distinct population of lineage-infidelity cells (CD117+CD71+ CD34+CD38+CD56+CD19dim) that co-express myeloid and B-lineage markers, representing a heightened plasticity that may facilitate CAR-T evasion. Projection of our scRNA-seq data onto a normal developmental atlas showed pre-CAR-T sample from the lineage-switch case encompassed a broad spectrum, whereas control case of CD19-negative relapse due to CD19 mutation were confined to the pro-B/pre-B stages. Based on these findings, we hypothesized that co-mutation of TP53 and epigenetic regulators predisposes cells to lineage switch. To validate this finding, we performed a longitudinal flow cytometry analysis. Patients were stratified into two groups: a co-mutation group (n=4) harboring TP53 and epigenetic gene mutations, and a control group (n=2) lacking these mutations. The co-mutation group exhibited significant downregulation of the B-lineage marker CD19 and concurrent upregulation of the myeloid marker following CAR-T treatment. In contrast, the control group showed no notable changes. In conclusion, our data suggests that co-mutations in TP53 and epigenetic genes enhance plasticity in B-ALL. Under the selective pressure of CD19 CAR-T therapy, subclones harboring these mutations are preferentially expanded, leading to lineage switch. We propose monitoring myeloid marker expressing subclones in B-ALL patients carrying co-mutations in TP53 and epigenetic genes during immunotherapy is important to improve their prognosis.