Genomic Heterogeneity Contributed to the Different Prognosis of Adult and Pediatric Acute Lymphoblastic Leukemia

Abstract

Prognosis of acute lymphoblastic leukemia (ALL) in adults is inferior than children. Hence, ALL is still a challenging disease to be cured in adults population. Aberrant genetic alterations has been observed previously in acute lymphoblastic leukemia (ALL), while the patterns of differential gene alteration have not much been comprehensively determined in the adult and pediatric ALL on a genome-wide scale. This study attempted to investigate the biological differences in genomic profiling between the adults and children with ALL, and the correlation between the genomic heterogeneity and prognosis. We collected the survival informations and surveyed the gene mutations profile using whole-exome sequencing (WES) in samples from 39 adults and 54 children of de novo ALL in our institution.The sequencing revealed 663 high-confidence somatic single-nucleotide variants (SNVs), 25 stopgain and 17 frameshift insertion in 690 genes in the 39 adult ALL samples. While in the 54 pediatric ALL samples, 232 SNVs, 12 stopgain and 10 frameshift insertion in 284 genes were identified. The results showed the similar common mutation types in adult and pediatric ALL. However, the median number of detected gene mutations was 19 (range: 1-53) per sample in adult ALL and 4.5(range: 1-19) in pediatric ALL(P<0.001), indicating the reduced prevalence of genetic alterations in pediatric ALL. A significant correlation between the increased of number of gene mutations and age was found (R2 = 0.3096, P<0.001,FigureC). The most frequently mutated genes were NOTCH1 in 6 samples (15.38%), TTN in 5 samples(12.8%), IKZF1 in 4 samples(10.2%) and NRAS in 4 samples(10.2%) in adult ALL. In pediatric ALL, the most frequently mutated genes were KRAS in 15 samples(27.7%), NOTCH1 in 13 samples (24.1%), NRAS in 9 samples(16.7%) and CREBBP in 6 samples(11.1%). The SETD2 and TTN mutations were significantly enriched in adult ALL. While the KRAS, ARID1A and CREBBP mutations were significantly enriched in pediatric ALL. (P<0.05,FigureA, B). Moreover, TP53 mutation was found in 1(0.03%) adult and 2(0.03%) pediatric ALL samples. 67 and 53 driver genes were identified in adult and pediatric ALL samples, respectively. Analysis of transcriptome sequencing data identified transcripts derived from 41 gene rearrangements in 25 (60.9%) adult ALL samples and 20 gene rearrangements in 15 (27.7%) pediatric ALL samples. The most frequent gene rearrangements were BCR/ABL1 fusions (13 samples), fusions of PAX5/ZCCH7 (4 samples) in adult ALL. The most common gene rearrangements in pediatric ALL were MLL fusions(7 samples), TEL-AML1 fusions (6 samples). Using integrated genomic analysis, we identified 3 functional pathways recurrently mutated in adult ALL: transcriptional regulation, NOTCH1 signaling, Ras signaling, and 5 in pediatric ALL: PI3K-AKT-mTOR signaling, JAK-STAT signaling, NOTCH1 signaling, Ras signaling, microRNA processing. The incidence of relapse was 33.3% and 7.7% in the adult and pediatric ALL, respectively(P=0.003). The overall survival(OS) and relapse free survival (RFS) of adult ALL were poorer than pediatric ALL(P=0.003, P<0.001, respectively,FigureD, E), indicating an unfavorable prognosis. Moreover, the number of gene mutations seems to be related with the decreased of times to relapse(R2 = 0.0571, P=0.39). Patients with different genetic subtypes were assigned to the different subgroups. The signatures may related to the inferior outcome of adults compared to children were identified. Adult ALL patients had more enrichment for alterations of amino acid degradation and transcription misregulation, which may explain in part the disparity in the different responses to treatment of the two populations. The study in genomic profiling across the age spectrum elucidated the genomic heterogeneity between adult and pediatric ALL, including the different in the counts of gene mutation, the frequently mutated genes and the fusion genes, which may be the contributing factors that influence prognosis. This genomic landscape enhanced the understanding of the biological differences of disease between the two populations and provided a clue for novel therapeutic approaches. Disclosures No relevant conflicts of interest to declare.