GENOMIC BCR-ABL1 FUSION CHARACTERISATION AND SNPS IDENTIFICATION UPSTREAM AND DOWNSTREAM OF THE BREAKPOINTS IN BCR AND ABL1 GENES OF CML PATIENTS USING NEXT GENERATION SEQUENCING

ABSSUB-4861

Background: BCR-ABL1 genomic fusion sequences represent clone-specific markers for each patient with chronic myeloid leukemia (CML). Their identification allows more sensitive monitoring of minimal residual disease at DNA level. As BCR-ABL1 is a large fusion gene the characterization of DNA breakpoints is challenging. In this study we used multiplex long-range PCR (mLR-PCR) with primers enabling to amplify large PCR products (˃10 kb) carrying BCR-ABL1 breakpoints. LR-PCR products were directly sequenced using next generation sequencing (NGS). Amplification of large regions upstream and downstream of the breakpoints allows identification of SNPs and other sequence features.

Aims: We aimed to characterize genomic BCR-ABL1 breakpoints and to identify SNPs in CML patients.

Methods: Two BCR-ABL1 positive cell lines (K562; JURL-MK1) and 48 CML patients who achieved a deep molecular response were included in this study. DNA was isolated from leukocytes of peripheral blood at diagnosis. We performed 2 rounds of mLR-PCR. In the 1st round, 1 forward primer located in BCR exon 13 and 10 primers located in intron 1 of ABL1 (Lange et al 1999) were used. LR-PCR products were obtained in 28/49 cases. The 2nd round of mLR-PCR was performed for the rest of samples using 1 BCR and 20 ABL1 primers (Krumbholz et al 2012, Ross et al 2010, Score et al 2010). LR-PCR products were fragmented and rapid library was prepared for 454 NGS technology (Roche Applied Science). NextGENe software (Softgenetics) was used for sequence analysis, breakpoints characterisation and SNP calling. NCBI database and reference sequences for BCR and ABL1 genes were used for breakpoints localization and SNP calling.

Results: Median length of LR-PCR products was 6 kb (range 1.5-10kb). NGS was performed in 41 samples. Characterized breakpoints were confirmed by Sanger sequencing. BCR breakpoints were located in intron 14 in 32/39 patients and in both cell lines, all carrying e14a2 mRNA, and in intron 13 in 7/39 patients with e13a2 mRNA. All ABL1 breakpoints were dispersed across intron 1. In BCR region upstream of the breakpoints, 17 annotated SNPs (1 SNP located in exon 13, 4 in intron 13 and 12 in intron 14) were detected among patients and cell lines with e14a2 mRNA (median 3 SNPs/patient or cell line, range 1-12). The frequency of minor alleles of detected SNPs varies from 0.5% to 30.7% in the 1000Genome phase 1 population (NCBI SNP database). Novel SNPs were identified in 3 patients with e14a2 mRNA in exon 13, intron 13 and exon 14. Interestingly, we did not detect any SNPs upstream of the BCR breakpoints among patients with e13a2 mRNA, except one with 3 SNPs detected. Based on SNP profiles in the BCR region we performed hierarchical clustering analysis showing 3 main distant clusters of patients; the cluster consisted of patients with e13a2 mRNA was distant to other 2 clusters which consisted of patients with e14a2 mRNA.

In the downstream region of the breakpoints in intron 1 of ABL1, 51 different SNPs were identified in 12/39 patients and JURL-MK1.

Summary/Conclusion: Sanger sequencing is not suitable for sequencing of long PCR products, thus an alternative approach had to be sought for BCR-ABL1 fusion identification. The mLR-PCR followed by NGS accelerated the whole process. Moreover, this approach enabled to analyse long sequences downstream and upstream of the BCR-ABL1 fusion and to identify other sequence features. SNPs in introns may alter gene splicing and coding SNPs may alter protein conformation. Interestingly, we identified multiple SNPs upstream of the genomic breakpoints in BCR gene in all CML patients with e14a2 mRNA. No SNPs were observed in all but one patient with e13a2 mRNA. Based on the SNP profile in the BCR region upstream of the breakpoints, 3 main distinct patient clusters were formed. Further studies are warranted to find the potential association between the SNP profile, patient characteristics and outcome.

Supported by IGA/NT11555 and the project for conceptual development of research organization (00023736) from the Ministry of Health of the Czech Republic.

Keywords: BCR-ABL, SNP

http://www.ehaweb.org/