Sensitivity and reliability of DNA-based mutation analysis by allele-specific digital PCR to follow resistant BCR-ABL1-positive cells

Abstract

Mutations in the kinase domain (KD) of BCR-ABL1 are known causes of resistance to tyrosine kinase inhibitor (TKI) treatment of chronic myeloid leukemia (CML) and Ph-positive acute lymphoblastic leukemia (Ph+ ALL), and detection is routinely performed at the transcript level. Sequencing remains the preferential technology and enables the identification of mutations at any position in the KD. Next-generation sequencing (NGS) is nearly 1-log more sensitive than standard sequencing, thus significantly facilitating earlier detection of mutations [1–3]. Moreover, NGS recognizes the complexity of the mutation landscape and can distinguish between polyclonal and compound mutations, if certain conditions are followed [1, 4]. The most frequently used amplicon-based NGS approaches are based on sequencing of shorter overlapping PCR products covering the entire length of the KD, thus distant mutations that may co-occur within one cell cannot be reliably assessed. There is a possibility of sequencing of long-range PCR product of the whole KD of BCR-ABL1, however there were available limited approaches and capacities of NGS platforms [5]. The promising technology for reliable assessment of compound mutations in the KD of BCRABL1 are Oxford Nanopore technology or SMRT PacBio technology. Determination of the level of the BCR-ABL1 KD mutation is important to follow the kinetics of resistant Ph+ cells and to therapeutically prevent progression [6]. However, an expression-based analysis of mutated BCR-ABL1 may not precisely reflect the quantity of the clones. In general, DNA mutational analysis represents an appropriate tool to study the clonality. However, for investigation of mutations in BCR-ABL1, it is necessary to amplify the KD region of ABL1 that is fused to BCR. The use of primers hybridizing to the BCR and ABL1 exons creates amplicons of ~1.400 bp at the mRNA level [7]. For DNA analysis, the position of the patient-specific breakpoint is needed [8]. However, since the KD of ABL1 alone is encoded by exons 4 to 9 corresponding to 17 kbp at the DNA level, it is not feasible to amplify such a long region. Without the amplification step, DNA-based KD sequencing underestimates the level of mutated BCR-ABL1. KD from non-fused ABL1 is unmutated; thus, a co-amplification of the KD from non-fused ABL1 significantly decreases the sensitivity of mutation detection. Due to this technical obstacle, there is a lack of studies on whether the quantity of mutated BCR-ABL1 analyzed at the transcript level corresponds to the number of mutated cells. In this study, allele-specific droplet digital PCR (ASddPCR) technology was used to compare the quantity of the BCR-ABL1 KD mutation measured at the DNA level and transcript level analyzed by NGS and to follow the clonal evolution of resistant BCR-ABL1+ cells using genomic quantification of mutated BCR-ABL1 (gBCR-ABL1). Sixty-eight analyses of individual mutation types from 45 peripheral blood samples from four CML patients (Table S1; patient #1 T315I and Y253H; #2 E225K; #3 M244V and T315I; #4 L248V and T315I) at the time of TKI resistance associated with the KD mutation were performed. KCL-22R-resistant clones (#1 T315I; #2 E255K and #3 Y253H) were prepared by single-cell FACS sorting (Supplementary Methods). Resistant clones were equally mixed to create a culture of polyclonal mutant cells, which * Katerina Machova Polakova [email protected]