Detection of the Genomic Variants in the Paired De Novo-Relapsed Samples with Pan-Cancer Panel and Whole Exome-Seq in Adult B-Cell Acute Lymphoblastic Leukemia

Background: The relapse rate is around 20-40% in adult B-cell acute lymphoblastic leukemia (B-ALL). The genetic defects are the major reasons for the relapse and poor outcome. We screened the genomic variants with Pan-cancer panel in B-ALL patients and whole exome-seq (WES) in the paired de novo-relapsed B-ALL samples.

Methods: The xGen Pan-Cancer Panel (IDT), which has been designed with the probes targeting 127 significantly mutated genes from the TCGA database across 12 tumor tissue types (Nature, 502:333-339) was used. Agilent SureSelect Human All Exon V4+UTRs (Agilent) was used to target coding exons plus UTRs for the WES analysis. The genomic DNA from 81 Philadelphia chromosome positive (Ph+) B-ALL patient samples (71 de novo and 10 relapsed, 14 to 77 years old) collected between June 2008 and June 2016 at Zhongda Hospital Southeast University were used for the Pan-cancer panel screening. All DNA samples were sheared and generated approximately 260bp DNA fragments. The fragmented DNA was processed into Illumina compatible sequencing libraries using the Kapa Hyper Prep Kit. Each library was uniquely barcoded and captured by either the Pan-cancer panel or WES probes, followed by PCR amplification and sequencing on a HiSeq 2500 (Illumina) with 2x100 bp reads. The sequencing reads were aligned to the human genome (hg19) by following Broad Institute's GATK best practice pipeline to call germline short variants (SNPs and Indels). Called variants were annotated using ANNOVAR (version 2.3). Variants with exonic, nonsynonymous, stopgain, or stoploss, novel SNPs (not covered by dbSNP138), and with predicted deleterious/damaging functions were manually surveyed by IGV to confirm. Two paired de novo-relapse samples from Ph(+) B-ALL patients were performed the WES analysis.

Results: We identified a total of 3945 single nucleotide variants (SNVs), 2222 insertions and deletions (INDELs) in the Pan-cancel panel analysis in 81 Ph(+) B-ALL patients. Among these, 101 genomic variants are with amino acid change, 8 are with stopgain, and 5944 have not been previously reported. We evaluated the frequency and distribution of likely pathogenic variants (PVs) detected in the cohort. Likely PVs were defined by SIFT algorithm which predicts whether an amino acid substitution is likely to affect protein function. Defined by the SIFT's qualitative score 'deleterious', we detected 46 PVs. Among these, PVs were commonly detected in KMT2C, APC, CDKN1A, NSD, BRCA1, EPHA3, and PIK3CG. The PVs were also validated with the Sanger DNA sequencing in the patients. The patients with the likely PVs have significantly higher WBC count (61*10^9/L vs. 24.45*10^9/L, P=0.004). Survival analysis showed that the patients with likely PVs had a worse event-free survival (EFS) and overall survival (OS), the difference was statistically significant (8 months vs. 15 months, P=0.017 and 14 months vs. 25 months, P=0.021). In order to gain an insight to the gene mutations contributing to the disease relapse, we compared the mutants spectrum between the de novo and relapsed samples. We found the genomic variants in NF1, CDK12, mTOR, and USP9X genes appeared mostly in relapsed samples, indicating their roles in the relapse. Using the WES, we further analyzed the genomic variants in two paired de novo and relapsed samples. We detected totally 40354 (de novo) and 16822 (relapsed) genomic variants, and among these 10415 (de novo) and 3082 (relapsed) are with amino acid change in patient 1; likewise, 30130 (de novo) and 14003 (relapse), and among these 7200 (de novo) 2534 (relapsed) with amino acid changes in patient 2. Totally 216 genomic variants with amino acid changes in 162 genes appeared only in the two relapsed samples, among which 10 genomic variants in ADAMTS8, CDK11B, EFCAB4B, FBXL21, HYDIN, IRF2BPL, MIR205HG, PRIM2, ZNF717, ZNF880 appeared in both relapsed samples, revealing their driver roles in relapse. Also, 110 of the 216 genomic variants are not previously reported.

Conclusion: Genomic variants in common human cancer driver genes were also detected in B-ALL patients. The new genomic variants detected in the relapse samples may be involved in the oncogenesis of the relapse, which will be further verified with functional analysis. Our data also suggested the significance of developing more efficient new therapies to prevent the relapse in hematological malignancies.