Prognostic value and oncogenic landscape of TP53 alterations in adult and pediatric T-ALL

TO THE EDITOR:

Among hematologic malignancies, T-cell acute lymphoblastic leukemia (T-ALL) represents a class of aggressive tumors with dismal clinical outcome in cases of relapsed or refractory disease. Although the intensification of multiagent chemotherapy protocols has dramatically improved prognosis, refractory and relapsed cases are clinically challenging because of limited therapeutic options.^1,2 p53 is a transcription factor and a master tumor suppressor gene frequently altered in cancer.³ In contrast to carcinomas and other hematologic malignancies, TP53 alterations (TP53^Alt), encompassing mutations (TP53^Mut), and/or pan-exon deletions (TP53^Del) are remarkably rare at diagnosis in T-ALL, and their clinical implication remains elusive.^4-6 Critically, TP53^Alt have been reported to be acquired in up to 20% of the relapsed T-ALL cases, where they convey a deleterious prognosis.^7-9 Here, we produce the first comprehensive analysis of TP53^Alt and the associated oncogenetic landscape in an extensive cohort of 476 patients newly diagnosed with T-ALL.

We investigated the clinical characteristics of TP53^Alt in 476 patients including 215 adults and 261 children. Adult patients were enrolled in the GRAALL-2003-2005 trial (GRAALL-2003, #NCT00222027; GRAALL-2005, #NCT00327678), and pediatric patients were enrolled in the FRALLE 2000 trial (supplemental Figure 1, available on the Blood website). Based on the DNA availability for molecular analysis, 215 adult patients out of 337 and 261 pediatric patients out of 427 were included in this study. No difference in clinical outcomes was observed between the included patients and the entire cohort (supplemental Figures 2 and 3; supplemental Table 1). Diagnostic peripheral blood or bone marrow samples were collected after informed consent was obtained, according to the Declaration of Helsinki. All samples contained ≥80% blasts, immunophenotypic of T-ALL samples, minimal residual disease (MRD) assessment, and multiplex ligation–dependent probe amplification analysis (P383 T-ALL, MRC Holland) were performed as previously described.^10-12 

Genomic analysis was performed using pan-exon targeted next-generation sequencing of DNA extracted from diagnostic samples, and DNA libraries were prepared using the Nextera XT kit (Illumina) and sequenced on a MiSeq (Illumina). The next-generation sequencing panel included 63 genes known to be mutated in T-ALL (supplemental Table 2). Genetic lesion co-occurrences and mutual exclusions were computed using the DISCOVER R package. We performed a computational approach previously described for the detection of copy number variants from next-generation sequencing data,^13,14 including a systematic analysis of the depth of TP53 gene coverage. This method is based on variations in the depth of coverage of the aligned sequence reads using a locally developed algorithm. The copy number variants detected were confirmed by high-resolution comparative genomic hybridization and/or multiplex ligation–dependent probe amplification analysis (kits P037-CLL-1 and P038-CLL-2, MRC Holland). Diagnostic DNA was hybridized on a Cytogenetics Whole-Genome 2.7M Array (Affymetrix, Santa Clara, CA) (comparative genomic hybridization array), according to the manufacturer’s recommendations. Data analysis was performed using Chromosome Analysis Suite software (Affymetrix). Gene copy number aberrations were compared with the Database of Genomic Variants¹⁵ to study only somatic aberrations.

Comparisons of categorical and continuous variables between subgroups were performed using Fisher exact test and the Mann-Whitney test, respectively. Overall survival (OS) was calculated from the date of diagnosis to the last follow-up date, censoring patients alive. The cumulative incidence of relapse (CIR) was calculated from the complete remission date to the date of relapse, and censoring patients alive without relapse at the last follow-up date. Relapse and death in complete remission were considered to be competitive events. Univariate and multivariate analyses to assess the impact of categorical and continuous variables were performed using the Cox model. The proportional-hazards assumption was checked before conducting the multivariate analyses. In the univariate and multivariate analyses, age and log10 (white blood count) were considered continuous variables. All the analyses were stratified during the trial. Variables with P < .1 in univariate analysis were included in the multivariable models. Statistical analyses were performed using the STATA software (STATA 12.0 Corporation, College Station, TX). All P values were 2-sided, with P < .05 denoting statistical significance. Additional details are included in the supplemental Methods.

The incidence of TP53^Alt at the time of diagnosis in this cohort was 4% (21/476). TP53^Mut was detected in 9 patients (6 adults and 3 pediatric cases) (Figure 1A; supplemental Tables 3 and 4), TP53^Del was identified in 15 patients (7 adult and 8 pediatric cases), and 3 patients harbored both TP53^Mut and TP53^Del. Patients with TP53^Alt did not significantly differ from patients with TP53 wild-type (TP53^WT) regarding sex, age, and central nervous system involvement (Table 1), but were associated with an immature phenotype (combining IM0 T-ALL; T-cell receptor δ [TCRδ] and TCRγ germ line, IMδ T-ALL; TCRδ rearranged but not TCRγ and IMg T-ALL; both TCRδ and TCRγ rearranged)¹² (8/20 [40%] vs 81/399 [20%], P = .048). The oncogenetic landscape of TP53^Alt was comparable to that of TP53^WT T-ALLs (Figure 1B-C; supplemental Figures 4-6; supplemental Table 5). To investigate the prognostic value of TP53^Alt, survival analyses were performed on the series of 476 patients. TP53^Alt did not confer an increased poor prednisone response, defined by a peripheral blood blast count >1.0 × 10⁹/L at the end of the induction phase (38% vs 45%, P = .7) (Table 1). Although TP53^Alt did not significantly influence the morphological complete response rate at the end of the induction course (81% vs 93%, P = .07), patients harboring TP53^Alt were associated with delayed early medullary blast clearance, as confirmed by the end of induction MRD1 assessment, with more positivity (≥10^–4) in TP53^Alt than in TP53^WT cases (75% vs 35%, P = .01). Patients (both adult and pediatric) with TP53^Alt had an inferior outcome compared with TP53^WT (Table 1; Figures 1D-E; supplemental Figure 7), with an increased CIR (5-year CIR: 65% vs 27%; specific hazard ratio [HR], 3.1; 95% confidence interval [CI], 1.67-5.78; P < .001) and a shorter OS (5-year OS: 48% vs 72%; HR, 2.34; 95% CI, 1.30-4.24; P = .005). In multivariate analysis, TP53^Alt predicted a statistically lower OS (HR, 2.87; 95% CI, 1.56-5.26; P = .001) and higher CIR (specific HR, 2.90; 95% CI, 1.55-5.44; P = .001) even after adjustment on the 4 genes NOTCH1/FBXW7/RAS/PTEN (NFRP) classifier, which identified patients with a poor prognosis in both GRAALL and FRALLE trials.^10,16 

The limited number of MRD1 assessments available for patients with TP53^Alt T-ALL (12/21, 57%) did not allow us to include this end point in the multivariate analysis, justifying further studies to establish whether this alteration remains an independent prognostic factor when combined with MRD1 status in T-ALL.

This study provides the largest comprehensive analysis of TP53^Alt in T-ALL, describing, for the first time, their clinical profile and, most importantly, the extremely poor prognostic impact associated with TP53^Alt at diagnosis in T-ALL, urging the need to develop innovative targeted therapies for patients harboring TP53^Alt.

Approximately 25% of pediatric patients and 50% of adult patients with T-ALL relapse, with a 5-year survival rate of less than 20% in both age groups.^1,2 The sole therapeutic approaches with curative potential for T-ALL relapsed cases are limited to conventional chemotherapy or hematopoietic stem cell transplantation. Molecular genetic analyses and sequencing studies have recently led to the identification of recurrent T-ALL alterations associated with prognosis, allowing for the refinement of the stratification of relapse risk.^10,14,17,18 However, a significant proportion of T-ALL relapses remain unpredicted, underlining the need for new predictive markers.

Negative outcomes observed in TP53^Alt T-ALL are likely to be related to p53-induced therapeutic resistance previously described for other malignancies, combined with loss of p53 tumor suppressor activity and acquisition of novel functions that disrupt the DNA damage response pathway.^19,20 APR-246 (eprenetapopt), a small molecule interacting with mutated p53 protein to restore its WT conformation,^21,22 has recently shown promising results in myeloid malignancies and B-cell ALL,^23-26 justifying further studies to explore the potential efficacy of this new drug in T-ALL TP53^Mut.