Mutations of TP53 gene in adult acute lymphoblastic leukemia at diagnosis do not affect the achievement of hematologic response but correlate with early relapse and very poor survival

In adult acute lymphoblastic leukemia (ALL) the evaluation of clinical and biological conventional risk criteria at diagnosis is important but not sufficient to predict clinical outcome. The impact of TP53 mutations has been investigated in a limited number of studies and has still not been defined. A study conducted on 98 newly diagnosed unselected adult ALL patients reported that TP53 mutations were present in 8% of the patients and were associated to a poor response to induction therapy.1 A more recent investigation described the mutational status of TP53 gene in a wide cohort of heterogeneously treated childhood and adult ALL, also including mature/MYC positive B-ALL patients. Next generation sequencing (NGS) data revealed the presence of TP53 mutations in 16% of patients and the incidence increased with age and in the presence of a hypodiploid karyotype. In all cases, TP53 aberrations were independently associated with short survival.32

The aim of this study was to evaluate the impact of TP53 genetic alterations, analyzed by NGS, on the outcome of a cohort of 171 adult Philadelphia-negative ALL [57 T lineage ALL (T-ALL) and 114 B-precursor ALL (Bp-ALL)], enrolled into the NILG-ALL 09/2000 clinical trial (clinicaltrials.gov identifier: 00358072).4 Patients’ characteristics are detailed in Online Supplementary Table S1). TP53 mutational status was evaluated on a total of 171 DNA diagnostic samples isolated from mononuclear cells obtained from bone marrow or peripheral blood, containing at least 30% blasts. The experimental design for the amplification of exons 4–11 was defined in collaboration with the ALL working group of the IRONII consortium co-ordinated by the MLL Munich Leukemia Laboratory (Germany) and Roche Diagnostics (Germany).5 Deep-sequencing was performed using a GS Junior Platform by Roche Diagnostics (Roche Diagnostics, Mannheim, Germany) following the manufacturer’s recommendations and the TP53 mutations were validated by conventional Sanger methodology. The copy number status of TP53 was also evaluated by a quantitative PCR (qPCR) method using hTERT as reference gene for 158 patients as described by the manufacturer (Applied Biosystems, Foster City, CA, USA). Sequencing data analysis and SNP analysis were performed as described in the Online Supplementary Appendix. More details also on statistical and outcome end point analysis are provided in the Online Supplementary Appendix.

The NGS analysis of the TP53 gene allowed identification of 25 different genetic variations present with an allele burden ranging between 97% and 4%, indicating that these alterations can be present at diagnosis also in a minority of leukemic clones. Nineteen of the 25 mutations identified were located in the exonic region and 6 were intronic. All the intronic and 3 exonic variations were recognized as polymorphic variants in dbSNP and IARC databases (Online Supplementary Appendix and Tables S2 and S3).6 The other 16 exonic variants were detected in exons 5–8 corresponding to the DNA-binding domain which is the hotspot region for TP53 mutations.7 These mutations were of different types: 13 were single nucleotide variants (SNV), 2 were duplications (one of 4 and the other of 8 nucleotides) and one was an 11 nucleotides insertion (Table 1). The 13 SNV, identified in 12 patients, were missense mutations. The 3 frameshift mutations identified in 2 samples were not described in the IARC database6 and led to the introduction of a premature stop codon (Table 1). All the alterations with a mutation load higher than 25% were validated by Sanger sequencing.

Table 1.Clinical/biological characteristics of the patients mutated for TP53 gene.

The TP53 copy number status was investigated by qPCR in patients with available DNA material. The analysis revealed that 10 of 158 (6.3%) patients presented one copy of the TP53 gene. Five of these 10 patients belong to the cohort bearing also a TP53 mutation (38.5%), demonstrating a strong correlation between these two genetic alteration (P=0.00049), as previously described.2 However TP53 deletion has no impact on clinical outcome. Furthermore, 2 cases (1.3%) showed 3 copies of the TP53 gene. In 9 patients, qPCR results were also confirmed by SNP analysis. Interestingly, patient BG_4205, affected by p.R273C mutation, showed a TP53 monosomy at diagnosis revealed by qPCR and SNP analysis and suffered the insurgence of two other different cancers during the clinical follow up. In this patient, Sanger sequencing showed the presence of the p.R273C mutated allele also in a sample from hematologic remission with a similar load of wild-type allele (Online Supplementary Figure S1). This finding demonstrated that the alteration was a germline mutation described to be associated with a severe tumor predisposition.6

Overall, we detected TP53 mutations in 14 of 171 ALL patients, corresponding to an 8% incidence. This is similar to data previously published in a cohort of adult ALL, with a comparable median age,1 but lower than the cohort published by Stangel et al., which was characterized by an older median age and included also mature BALL.2 These aberrations were recognized with a higher frequency in Bp-ALL (9.7%) than in T-ALL (5.3%), as reported in the study by Stengel et al.,2 but not in the study by Chiaretti et al.1 This discrepancy can be partially explained by the unselected and prospective collection of our specimens. Furthermore, we analyzed Bp-ALL and T-ALL, excluding Ph ALL, whereas other cohorts also included Ph ALL.21

The univariate analysis indicated a clear relationship with a linear trend between the presence of TP53 mutations and increasing age (P=0.00032) (Online Supplementary Figure S2). Moreover, within the limits of a small number of subjects analyzed, we found a high rate of positivity for TP53 mutation in the nearly triploid cytogenetic subgroup of patients (P=0.0008) (data not shown). No other correlation with clinical features, such as gender, hemoglobin, white blood cell (WBC) count, platelets, percentage of blasts at diagnosis, and clinical risk class, emerged from this analysis, as previously reported.21

In our cohort of patients, median follow up was 20 months and the maximum follow up reached nearly 15 years. The clinical outcome of TP53 mutated patients are summarized in Table 1. All patients carrying a TP53 alteration entered clinical remission after induction therapy, but 93% of them suffered an early relapse (within 15 months from first remission). Indeed, the frequency of relapse in patients mutated for TP53 was significantly higher than in wild-type ones (P=0.019). These data showed that the presence of a mutated TP53 does not itself produce a primary resistance to the induction chemotherapy, but rather leads to a greater susceptibility to relapse. This finding differs to reports by Chiaretti et al. and we can only speculate that a different treatment intensity (such as those used in pediatric patients) could be the reason for such a discrepancy. Nevertheless, most of the TP53 mutated patients of our cohort could not benefit from transplantation in first CR as a consolidation therapy because of advanced age (>60 years) or a very early relapse, which did not allow completion of the induction/consolidation phase and the identification of a suitable donor (Table 1).

Globally, the cumulative incidence of relapse (CIR) was 63% (95%CI: 56%–71%) at four years while the leukemia-free survival (LFS) was 31% (95%CI: 25%–40%) at four years and the overall survival (OS) of the entire cohort was 36% (95%CI: 29%–44%) at four years. We observed that the CIR at four years was significantly higher in TP53 mutated [93% (95%CI: 77%–100%)] compared to wild-type [60% (95%CI: 52%–69%)] patients (P<0.0001) (Figure 1A). In agreement with this, both the 4-year LFS and OS were dramatically shorter in TP53 mutated compared to wild-type patients: LFS was 7% (95%CI: 1%–47%) in mutated versus 34% (95%CI: 27%–43%) in wild-type patients (P=0.0008); OS was 7% (95%CI: 1%–47%) in mutated versus 39% (95%CI: 32%–47%) in wild-type patients (P=0.0013) (Figure 1B and C).

Figure 1.Analysis of the clinical outcome of the TP53 wild-type versus TP53 mutated patients. (A) Cumulative incidence of relapse. (B) Leukemia-free survival. (C) Overall survival.

A univariate analysis was conducted to investigate the correlation of the clinical/biological characteristics of the patients and their outcome in terms of CIR and OS: the presence of the TP53 mutation, age, WBC, central nervous system (CNS) involvement and adverse cytogenetics significantly correlate with a higher CIR and with a lower OS (Table 2). Interestingly, a correlation with minimal residual disease (MRD) status was not possible because, for reasons that are still unclear, in many TP53 mutated patients, no informative clonal rearrangements of immunoglobulin or T-cell receptor genes could be found following the EuroMRD guidelines.1084 By multivariate analysis, age, WBC, CNS involvement and the presence of a TP53 mutation at diagnosis proved to be independently associated with a worse clinical outcome in terms of OS and, except for age, also in terms of CIR (Table 2).

Table 2.Univariate and multivariate analysis for clinical outcome.

To our knowledge, this study is the first in which a large cohort of adult ALL patients with a long follow up (up to 15 years) enrolled in the same clinical protocol was analyzed for the presence of TP53 mutations at diagnosis. Since the presence of TP53 alterations at diagnosis defines a group of patients with a very poor outcome, the definition of the mutational status of this gene must be included in the diagnostic workup of adult ALL. The broad range of mutation load we found in our diagnostic specimens underlines the importance of evaluating samples with a highly sensitive technique, such as NGS methodologies, also at disease presentation.