TP53 and KRAS Variants at Initial Diagnosis Identify an Ultra-High Risk Group of Pediatric T-Lymphoblastic Leukemia (T-ALL)

Introduction

Patients who suffer a relapse of pediatric T-cell acute lymphoblastic leukemia (T-ALL) face a dismal prognosis. Prognostic molecular biomarkers that reliably predict the risk of relapse at the time of first diagnosis are not available. Inactivating mutations in TP53 were previously detected in approximately 10% of relapsed patients (Hof et al. J Clin Oncol. 2011) and are invariably associated with fatal outcome (Richter-Pechanska et al. Blood Cancer J. 2017). Mutations in other genes were identified to be either specific for relapse (NT5C2 and CCDC88A) or to be associated with a poor prognosis in relapse (IL7R, KRAS, NRAS, USP7, CNOT3 and MSH6) (Meyer et al. Nat Genet. 2013; Richter-Pechanska et al. Blood Cancer J. 2017). We hypothesized that subclones bearing such mutations can give rise to relapse and analyzed these 9 genes at initial diagnosis of T-ALL with targeted ultra-deep sequencing.

Methods

Leukemia samples collected at initial diagnosis of 81 children with T-ALL who later relapsed were analyzed. As a control group, we selected 79 children with T-ALL who remained in first remission for at least three years and were matched with regard to treatment response, treatment, age and sex. Targeted deep sequencing was performed by using the Agilent Haloplex High Sensitivity kit with unique molecular identifiers for reliable detection of mutations with very low allele frequencies (average read depth: 1,012x).

Results

Overall, we detected 75 mutations among 7 targeted genes in 33 / 81 relapsing and 21 / 79 non-relapsing patients. The average allele frequency (AF) of the identified mutations was 25% (0.8% - 83%; SD ± 18%). More than half of the variants (43/75) showed AFs below 30% and were thus classified as subclonal. Interestingly, 7 pathogenic TP53 mutations (subclonal: n=5, clonal: n=2) with AFs of 4.4% - 49.4% were exclusively discovered in 6 patients who experienced a relapse. While 2 of these patients received an allogeneic stem cell transplantation in first remission because of poor treatment response, the remaining 4 patients were treated by chemotherapy in the high-risk (n=1) or medium-risk (n=3) arm. None of the 79 non-relapsing control patients carried TP53 mutations. Consistent with the hypothesis of clonal evolution as a mechanism of relapse in T-ALL, Sanger Sequencing of the relapse sample of one TP53-positive patient confirmed that the subclone harboring the TP53 mutation A159D at initial diagnosis (AF 5.4%) expanded to a major clone (AF 42%) in relapse. The presence of TP53 mutations in two further TP53-positive patients in at least one available post-remission sample is also compatible with clonal selection. However, in a fourth patient the low allele frequency of the TP53 mutation at relapse indicates that the TP53 subclone persisted but did not expand during the development of relapse. In addition to TP53, we identified pathogenic KRAS mutations to be significantly enriched in relapsing patients (9 / 81) compared to non-relapsing patients (2 / 79) at the time of initial diagnosis (chi-squared test, p= 0.032; Table 1).

Conclusion

Subclonal and clonal mutations in TP53 and KRAS at initial diagnosis were enriched in T-ALL patients who later relapsed and identified approximately 17% of patients suffering a relapse. We thus propose that (subclonal) mutations of TP53 and KRAS may define a subgroup of high-risk T-ALL patients already at the time of first diagnosis. The identification of such mutations may complement the current risk stratification which depends on treatment response and may determine a new molecularly defined subgroup of T-ALLs that may benefit from intensified treatment strategies.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal