T-cell acute lymphoblastic leukemia (T-ALL) accounts for 10% to 15% of newly diagnosed cases of childhood acute lymphoblastic leukemia (ALL). Generally, childhood T-ALL patients have a worse prognosis than B cell precursor ALL patients. Recent studies have identified a subtype of T-ALL termed “early T-cell precursor” (ETP) ALL, which is associated with a high risk of treatment failure. In spite of recent improvements of risk stratified multiagent chemotherapy, relapsed patients have a poor prognosis even if they were non-ETP ALL.

Recent genome-wide approach revealed frequent NOTCH1 and FBXW7 oncogenic mutations mutations in T-ALL. In addition, previous whole-exome sequencing disclosed novel CNOT3 mutations in approximately 10% of adult T-ALL cases, and thus, CNOT3 was thought to be one of the novel tumor suppressor gene for adult T-ALL. CNOT3 is part of the CCR4-NOT complex that is the major deadenylase of mRNA. NT5C2, encoding a 5ʹ-nucleotidase was identified as relapse specific mutation, of which mutation is associated with the outgrowth of drug-resistant clones in ALL. However, these mutations have been found in a fraction of childhood T-ALL suggests that the existence of other genetic pathogenesis.

To discover new oncogenic gene mutations which involved in the pathogenesis of relapsed T-ALL and to identify novel prognostic markers of childhood T-ALL, we performed genome-wide analysis using whole-exome sequencing and 250K SNP array analyses in 8 cases with relapsed T-ALL and 16 cases with non-relapsed T-ALL.

The mean coverage in the whole-exome sequencing of tumor and germline samples was 108× and 100× for the 50-Mb target regions, respectively, by which more than 90% of the coding sequences were represented by more than 20 independent reads on average. A mean of nonsilent mutations per sample at presentation was 18, and sample at 1st relapsed was 19. There were only 16 recurrent mutations in 24 cases; however no shared mutation in 8 relapsed cases other than NOTCH1 and FBXW7. NOTCH1 mutations were found in 50% (12/24), and were frequently identified in relapsed cases (6/8). FBXW7 mutations were also frequently found in 6/24 cases, and 60 % (3/6) were compound heterozygous mutations. In those 6 cases, only one case with FBWX7 mutation had a NOTCH1 mutation. CNOT3 mutations were reported to be frequent in adult T-ALL, however we found only two cases with CNOT3 mutations (8.3%). In addition, PHF6 mutation, which is known as X-linked tumor suppressor gene in T-ALL, was recurrent in 3 cases. Other recurrent mutations were shared between 2 cases, respectively. NT5C2 mutation has been reported to a relapse-specific mutation, and we also found NT5C2 mutations in 2 relapsed cases, which detected in only relapsed samples. RPL5 and RPL10 mutations were reported to be found in 10 % of pediatric T-ALL; however there was one mutation in RPL related genes in our study. Furthermore, we found common mutations of acute myeloid leukemia such as TCF7, STAT5A, KIT, RUNX1, and EP300 mutations in a single case. On the other hand, although pediatric T-ALL showed largely normal genomic copy number profiles, homozygous deletions at chromosome 9p21 harboring CDKN2A were frequently detected in our study (17/24 71%). Especially, 9p21 deletions were found in all relapsed cases, suggesting that loss of CDKN2A locus was a critical genetic mechanism of relapsed T-ALL.

In conclusion, our results revealed mutations in several known genes, but overall frequency of recurrent somatic mutations in childhood T-ALL is low, even in relapsed samples. Although loss of CDKN2A locus was detected in all relapsed cases, recurrent relapse-specific mutations could not be identified other than NT5C2. These findings suggest that the majority of relapsed T-ALL may be driven by aberrations of CDKN2A and minor clone variants and/or epigenetic modifications during tumor evolution.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.