Ultradeep-Amplicon Pyrosequencing for Mutation Detection in the Kinase Domain of BCR-ABL Revealed Artificial Low-Level Variants That Need to Be Avoided for Relevant Mutational Data Interpretation

Mutations in the kinase domain (KD) of BCR-ABL are a known mechanism of resistance to the tyrosine kinase inhibitor (TKI) treatment of patients with chronic myeloid leukemia (CML) and Ph+ acute lymphoblastic leukemia (ALL). The gold standard for the routine examination of mutations in patients with therapy failure is the Sanger sequencing of amplified cDNA region of BCR-ABL KD with the sensitivity of ∼15–20%. More than 100 BCR-ABL KD mutations were described in association with development of resistance to imatinib. It was suggested that early ultra-sensitive detection of KD BCR-ABL mutations, at the time of diagnosis or shortly after imatinib start and before switch to 2^nd generation TKI, may predict their subsequent expansion and failure of the treatment if the patient is treated with TKI against which is the type of mutation resistant. Ultradeep amplicon pyrosequencing represents a promising tool to study BCR-ABL mutation development and their expansion under TKI treatment. In our study, the detection of mutations reached the sensitivity of 0.03-0.01%. At this level we may expect detection of artifacts introduced during amplicon library preparation and pyrosequencing.

The aim is to define the threshold for relevant detection of published BCR-ABL mutations (n= 124).

We evaluated 155 CML or Ph+ALL patient samples. As the non-mutated control, the cDNA of KD c-ABL of healthy individuals (n= 9) and cDNA of wild type BCR-ABL from Ph+ cell lines (n=3) were used. To cover cDNA region of kinase domain (950 bp), multiple amplicons with the length up to 400 bp should be prepared to apply 454 technology. The experiments consist of: 1) comparison of two amplicon designs with 3 and 4 fragments covering KD to follow the amplicon-related artifacts; 2) the reproducibility test of 5 samples using 3 and 4 PCR fragments for amplicon library preparation; 3) tests with 2 different reverse transcriptases and Taq polymerases to detect artifactual nucleotide substitutions.

Using 454 technology we detected high number of ultralow-level single nucleotide substitutions (∼ 200) with the frequency < 1 in 100 in all samples analyzed including non-mutated controls using 3 and also 4 amplicon design. Surprisingly, we found that almost all low-level variants were T#x2610;#x0025;C or A#x2610;#x0025;G substitutions (Fig. 1), including those that may result in clinically relevant mutations (V288A, M244V, V289A, F317L, F359L, M351T, E355G, E453G). Evaluation of pyrograms suggested that they were unlikely to be errors of pyrosequencing, because they were not necessarily associated with hompolymer regions. The frequency of T#x2610;#x0025;C or A#x2610;#x0025;G low-level variants was lower when using proof-reading Taq polymerase. The pattern of these nucleotide substitutions is similar to that reported in IGH hypermutation analysis in CLL (Campbell et al. PNAS 2008) showing that T#x2610;#x0025;C or A#x2610;#x0025;G substitutions are errors caused by Taq polymerases. This is supported by the studies on PCR fidelity suggesting that Taq polymerase caused this pattern of base substitution even when using a proof-reading enzyme. Thus, the T#x2610;#x0025;C or A#x2610;#x0025;G substitution in BCR-ABL KD should not be evaluated below the level of 1%. The other but less frequent transitions caused by Taq polymerase are G#x2610;#x0025;A or C#x2610;#x0025;T substitutions that were detected in our study with the frequency < 1 in 1000, thus, the variants detected below 0.1% should not be examined as the true-positive result.

Mutation data from ultradeep amplicon sequencing of BCR-ABL KD below 1% must be evaluated cautiously to avoid misleading interpretations. Attention must be paid to the fact that T#x2610;#x0025;C or A#x2610;#x0025;G substitution may arise spontaneously in vivo, but there is no way to differentiate in vivo spontaneous mutations from Taq polymerase errors in vitro.

View largeDownload slide

View largeDownload slide

Close modal

Soverini:Novartis: Consultancy; Bristol-Myers Squibb: Consultancy; ARIAD: Consultancy. Baccarani:ARIAD, Novartis, Bristol Myers-Squibb, and Pfizer: Consultancy, Honoraria, Speakers Bureau. Martinelli:NOVARTIS: Consultancy, Honoraria, Speakers Bureau; BMS: Consultancy, Honoraria, Speakers Bureau; PFIZER: Consultancy; ARIAD: Consultancy. Machova Polakova:Novartis: Honoraria, Research Funding; BMS: Honoraria.