Acute Myeloid Leukaemia (AML) is an aggressive blood cancer with 5-year survival of only 26%, and very poor survival for elderly individuals. Treatment with Venetoclax and chemotherapy has proven effective in fit older patients diagnosed with AML, in particular when their disease carries a mutation in the gene NPM1. However, treatment failure and relapse can occur due to clonal heterogeneity and evolution.
To examine the transcriptional changes that drive treatment resistance we have performed single cell RNA sequencing (scRNA-Seq) of three NPM1-mutated AML patients over several time points (diagnosis, 8 days post Venetoclax monotherapy treatment, at remission and at treatment failure or relapse). The longitudinal patient samples had previously been analysed with bulk exome and Tapestri targeted single-cell DNA sequencing, which identified subclonal populations with differing responses to treatment.
To identify the subclonal populations in the scRNA-Seq libraries, we supplemented conventional short-read sequencing with targeted long-read scRNA-Seq (RaCH-seq) of a panel of 24 genes associated with AML. We genotyped cells for mutations in NPM1, FLT3, TET2 and KIT, and found that targeted long-reads improved genotyping rates several fold compared to untargeted long-reads and short-reads. Through combined short and long-read data analysis, we find subclones consistent with prior data, identify new subclones corresponding to copy number changes, and are able to associate some of these with distinct transcriptional profiles. In each patient we find dynamic changes in expression of genes previously implicated in treatment response. Taken together, the data provides novel insights into the heterogeneity and evolution of AML, and demonstrates the value of performing long-read scRNA-Seq, in combination with conventional scRNA-Seq.