PTEN abnormalities predict poor outcome in children with T‐cell acute lymphoblastic leukemia treated according to ALL IC‐BFM protocols

Pediatric T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive malignancy derived from T-cell precursors and represents 10% to15% of childhood acute lymphoblastic leukemia, the most common cancer in children. In T-ALL many recurrent molecular lesions, for example, alterations affecting NOTCH1, PI3K/PTEN/AKT/mTOR, or JAK-STAT signaling pathways, hold promise for genetic-based risk stratification and targeted therapy. Yet, the prognostic value of these alterations is still an area of active debate and probably depends on the treatment context. For instance, PTEN abnormalities, the negative regulator of the PI3K/AKT pathway, were associated with unfavorable long-term outcome in some pediatric and adult T-ALL series,1-3 yet not all.4 The ambiguous impact of PTEN might be linked to the type of genetic abnormality, for example, mutation or deletion, although few studies, comprising small groups of patients, reported on both types of alterations. Here, we analyzed the prognostic value of mutations and copy number alterations of recurrently mutated T-ALL oncogenes and tumor suppressor genes (NOTCH1, FBXW7, PTEN, WT1, IL7R, STAT5B, FLT3, RUNX1, DNMT3A, SIL-TAL1, LEF1, CASP8AP2, MYB, EZH2, CDKN2A/B, MLLT3, NUP214-ABL1, LMO1, LMO2, NF1, SUZ12, PTPN2, and PHF6) in a large cohort of 162 children, consecutive patients diagnosed with T-ALL, treated with International Berlin-Frankfurt-Munster Study Group trials, ALL IC-BFM 2002 (n = 91, 56%) and ALL IC-BFM 2009 (n = 71, 44%). Clinical characteristics of patients are given in Supporting Information Table S1. Using combined analysis of point mutations and copy number alterations (Supporting Information Table S2), we detected a total of 130 single nucleotide variants and small insertions/deletions (indels) and 303 copy number alterations (Supporting Information Figure S1 and Supporting Information Table S3), including 13 PTEN mutations in 9% of patients (12/140) and 21 PTEN deletions in 16% of patients (21/131). PTEN abnormalities (PTEN.ABN, either deletion or mutation) were present in 20% of patients (26/131). Bi-allelic inactivation of PTEN, as a result of co-occurrence of point mutations and deletions, was observed in 8% of patients (10/131). PTEN mutations only with no co-occurring PTEN deletions (PTEN.MUT) were detected in 4% of patients (5/121), while PTEN deletions only with no co-occurring PTEN mutations (PTEN.DEL) in 11% of patients (13/121). Of all genetic aberrations studied, we found only PTEN alterations to significantly predict for the unfavorable outcome of T-ALL children treated with ALL IC-BFM protocols. PTEN.ABN status associated with poor PR (≥1000 blasts/mL of peripheral blood on day 8; OR: 3.06; 95%CI: 1.22-7.67, P = 0.02) (Supporting Information Tables S4 and S5). In patients with PTEN.ABN, the probability of five-year EFS was estimated at 47%, compared with 78% in patients without any PTEN alterations (PTEN.WT), P = 0.00019 (Figure 1A). Consequently, PTEN.ABN patients had higher risk of relapse (pCIR of 38% vs 16%, P = 0.0095) (Supporting Information Figure S2). When we looked into how the type of PTEN genetic abnormality determined the clinical outcome, both PTEN deletions, irrespective of mutational status, and PTEN mutations, irrespective of deletional status, associated with worse survival (Supporting Information Figure S3A,B). However, looking into PTEN mutation only (PTEN.MUT) or PTEN deletion only (PTEN.DEL), we observed that patients with PTEN.DEL had five-year EFS of 45% vs 78% in PTEN.WT patients, P = 0.00076 (Figure 1B). The risk of relapse was estimated at 42% in PTEN.DEL patients vs 16% in PTEN.WT patients, P = 0.023 (Supporting Information Figure S4). Of note, in patients with PTEN.MUT, the association with EFS was not statistically significant; EFS was estimated at 60% vs 78% in PTEN.WT patients, P = 0.2 (Figure 1C). Consequently, patients with PTEN.MUT presented with nonsignificant increased risk of relapse (pCIR of 40% vs 16% in PTENWT patients, P = 0.07; Supporting Information Figure S5). In the separate analysis of patients from ALL IC-BFM 2002 and ALL IC-BFM 2009 trials, PTEN.ABN, PTEN.DEL, and PTEN.MUT significantly predicted unfavorable outcome of patients from ALL IC-BFM 2009 trial only (Supporting Information Figure S6A-C). In contrast, none of PTEN alterations significantly associated with shorter EFS in patients treated according to ALL IC-BFM 2002 protocol (Supporting Information Figure S7A-C). Of note, PTEN.ABN and PTEN.DEL resulted in shorter EFS in both ALL IC-BFM protocols, but what differed between these two cohorts was the outcome of PTEN.WT patients, which was better in ALL IC-BFM 2009 protocol. These results suggest comparably unfavorable outcome of patients with PTEN abnormalities in both protocols but a more favorable outcome of PTEN.WT patients from ALL IC-BFM 2009 protocol. Thus, our results render PTEN abnormalities candidate prognostic markers in T-ALL, in agreement with FRALLE2000T2 and AIEOP-BFM1 reports, but in contrast to MRC UKALL 2003 4 report. The summary and comparison of studies on prognostic significance of PTEN abnormalities in pediatric and adult T-ALL are presented in Supporting Information Table S6. Because in ALL IC-BFM 2002 trial, T-ALL patients were randomized to receive different types of delayed intensification, while in the ALL IC-BFM 2009 trial (ongoing), patients were stratified to receive different types of early intensification: standard IB or IB Augmented (in Poland SR and IR groups received IB protocol and HR group IB Augmented protocol), our results suggest a beneficial role of IB Augmented remission induction and HR blocks together with protocol II in reinduction phase (ALL IC-BFM 2009) in PTEN.WT patients. However, the overall impact of ALL IC-BFM 2009 protocol will be evaluated in larger patients groups, because the trial is still ongoing. Next, we determined whether PTEN status provides additional prognostic information to minimal residual disease (MRD) groups defined by flow cytometry (FC-MRD) and predict for the outcome. We divided ALL IC-BFM patients into FC-MRD risk groups according to FC-MRD levels at day 15, using criteria that predict the risk of relapse (<0.1% blast cells for SR, 0.1-<10% for IR, and ≥ 10% for HR group).5 Interestingly, in ALL IC-BFM patients classified as FC-MRD-IR group at day 15, PTEN.ABN associated with lower EFS. Five-year EFS was of 50% in PTEN.ABN cases vs 85% in PTEN.WT patients, P = 0.03 (Figure 1D). A similar tendency was observed in ALL IC-BFM patients with PTEN.ABN and belonging to FC-MRD-HR group at day 15 (Supporting Information Figure S8) and in patients from ALL IC-BFM 2009 trial only (in which FC-MRD at day 15 was obligatory for risk stratification) belonging to FC-MRD-IR and FC-MRD-HR groups at day 15 (Supporting Information Figure S9A,B). Altogether these data indicate that PTEN.ABN status can divide patients from FC-MRD-IR and FC-MRD-HR groups into favorable and unfavorable prognosis, thus holds potential for the improvement of FC-MRD-based risk assessment. This is of particular importance for the management of patients from FC-MRD-IR group, in which group the majority of relapses occur. The Cox regression model was applied to evaluate the prognostic significance of PTEN abnormalities (PTEN.ABN) after adjusting for major known prognostic factors in T-ALL, such as FC-MRD at day 15, response to prednisone, WBC count, and age. Although none of the tested variables proved to be independent outcome predictors in the studied cohort, FC-MRD at day 15 was related to an 86% increase in the hazard ratio (HR = 1.86, P = 0.21) and PTEN. ABN status to 70% increase in the hazard ratio (HR = 1.7, P = 0.36) (Supporting Information Table S7). Interestingly, in ALL IC-BFM-treated children, we did not observe the favorable effect of NOTCH1 mutations, or combined NOTCH1, or FBXW7 mutations (Supporting Information Figures S10 and S11). In light of previous findings on the favorable prognostic significance of NOTCH1/FBXW7 mutations restricted to patients without RAS/PTEN alterations in children6 treated with FRALLE and protocols, we tested a similar combination of oncogenic synergy between mutated genes studied in our panel. We compared the outcome of patients with NOTCH1/FBXW7 mutations and PTEN.WT (defined as genetic low risk; gLoR) with the rest of the cohort (defined as genetic high risk; gHiR). While FRALLE and GRAALL study groups could define an oncogenetic classifier on the basis of NOTCH1/FBXW7/RAS/PTEN mutation profile, a similar combination of genetic alterations, based on NOTCH1/FBXW7/PTEN did not predict differences in the outcome of children treated according to ALL IC-BFM protocols analyzed here (Supporting Information Figure S12), in agreement with data from MRC UKALL 2003 study.4 Taken together, we showed that the presence of PTEN alterations (deletions and/or mutations) identified a subgroup of T-ALL children with poor response to corticosteroid prophase therapy and with a particularly unfavorable clinical outcome assessed in the context of ALL IC-BFM treatment. Importantly, the PTEN status added prognostic information to risk group stratification based on the level of MRD assessed by flow cytometry. Inclusion of PTEN status discriminated between patients with favorable and unfavorable outcome in the FC-MRD-IR group at day 15 of therapy. Therefore, we argue for the prognostic value of PTEN inactivation in the context of BFM-type treatment, and we conclude that screening of PTEN could help refine the currently applied risk-adapted therapy strategies. Lastly, these results rationale the current research on the development and clinical evaluation of many PI3K inhibitors. This work was supported by grants from the National Science Center, Poland: 2013/11/N/NZ5/03730, the National Centre for Research and Development, Poland: grant number STRATEGMED3/304586/5/NCBR/2017. BSZ was supported by short-term fellowship from European Molecular Biology Organization (ASTF 50-2015) and received financial resources as part of financing the doctoral scholarship from the National Science Center (ETIUDA scholarship number: 2017/24/T/NZ5/00359). Authors express their appreciation to all the Heads of the Polish Pediatric Leukemia/Lymphoma Study Group centers for providing data and samples from study patients. PVV was supported by grants from the Research Foundation Flanders (FWO) and the Ghent University Special Research Fund (BOF). The authors declare that they have no conflict of interest with the contents of this article. BSZ and MD conceived the study; BSZ, MK, MD, and MKH performed molecular analyses; BSZ, JBK, ORB, and MD analyzed and interpreted data; ŁS and TS provided FC-MRD data; KZ and PB performed statistical analyses; JK and TS coordinated acquisition of samples and clinical data; BSZ and RJ designed and created figures; PVV supervised sWGS analyses; AEK supervised MLPA analyses; MD supervised the study and MW was in charge of overall direction; BSZ and MD wrote the manuscript; all authors approved the manuscript. Supporting Figure S1. Overview of the mutations identified in the panel of analyzed genes. Each type of mutation is indicated by a different color. Genes are grouped into functional pathways linked to the pathogenesis of T-ALL. Gender and information regarding the relapse are also shown for each patient (bottom table). Supporting Figure S2. Cumulative incidence of relapse (pCIR) according to PTEN.ABN status. PTEN.ABN patients had pCIR of 38% (95%CI: 21%-66%) vs 16% (95%CI: 9%-26%) in PTEN.WT patients, P=0.0095. Supporting Figure S3. Survival rates according to PTEN.DEL, irrespective of mutational status and PTEN.MUT, irrespective of deletional status, in the ALL IC-BFM cohort. A. In patients with PTEN.DEL irrespectively of mutational status the probability of EFS was estimated at 41% (95%CI:23%-71%), compared with 77% (95%CI:69%-86%) in those without PTEN deletions. The HR for poorer EFS in the PTEN.DEL group was 3.81 (95% CI, 1.31-11.1; P=0.00008). B. In patients with PTEN.MUT, irrespective of deletional status, the probability of 5-year EFS was estimated at 49% (95%CI:26%-88%), compared with 74% (95%CI:66%-83%) in patients without PTEN mutations. The hazard ratio (HR) for poorer EFS in the PTEN.MUT group was 2.36 (95% CI, 0.7-7.99; P=0.047). Supporting Figure S4. Cumulative incidence of relapse (pCIR) according to PTEN.DEL status in the ALL IC-BFM cohort. PTEN.DEL patients had pCIR of 42% (95%CI: 18%-94%) vs 16% (95%CI: 9%-26%) in PTEN.WT patients, P=0.0231. Supporting Figure S5. Cumulative incidence of relapse (pCIR) according to PTEN.MUT status in the ALL IC-BFM cohort. PTEN.MUT patients had pCIR of 40% (95%CI: 12%-100%) vs 16% (95%CI: 9%-26%) in PTEN.WT patients, P=0.07. Supporting Figure S6. Survival rates according to PTEN.ABN, PTEN.DEL and PTEN.MUT in ALL IC-BFM 2009. A. pEFS according to PTEN.ABN in ALL IC-BFM 2009. The probability of 5-year EFS was estimated at 42% (95%CI:22%-79%) vs 89% (95%CI:81%-99%) in PTEN.WT patients. The HR for poorer EFS was 7.9 (95%CI:1.85-33.79; P=0.00002). B. pEFS according to PTEN.DEL in ALL IC-BFM 2009. The probability of 5-year EFS was estimated at 42% (95%CI:18%-94%) in PTEN.DEL patients, compared with 89% (95%CI:81%-99%) in PTEN.WT patients. The HR for poorer EFS was 7.79 (95%CI:1.11-54.87; P=0.00011). C. pEFS according to PTEN.MUT in ALL IC-BFM 2009. The probability of 5-year EFS was estimated at 33% (95%CI:6%-100%) in PTEN.MUT patients, compared with 89% (95%CI:81%-99%) in PTEN.WT patients. The HR for poorer EFS was 9.52 (95%CI:0.22-411.1; P=0.001). Supporting Figure S7. Survival rates according to PTEN.ABN, PTEN.DEL and PTEN.MUT in ALL IC-BFM 2002. A. pEFS according to PTEN.ABN in ALL IC-BFM 2002. The probability of 5-year EFS was estimated at 53% (95%CI:30%-94%) vs 70% (95%CI:59%-84%) in PTEN.WT patients. The HR for poorer EFS was 1.87 (95%CI:0.55-6.32; P=0.22). B. pEFS according to PTEN.DEL in ALL IC-BFM 2002. The probability of 5-year EFS was estimated at 50% (95%CI:19%-100%) in PTEN.DEL patients, compared with 70% (95%CI:59%-84%) in PTEN.WT patients. The HR for poorer EFS was 2.32 (95%CI:0.28-18.98; P=0.25). C. pEFS according to PTEN.MUT in ALL IC-BFM 2002. The probability of 5-year EFS was estimated at 100% in PTEN.MUT patients, compared with 89% (95%CI:81%-99%) in PTEN.WT patients, P=0.45. Supporting Figure S8. Event-free survival (pEFS) according to the status of PTEN.ABN in patients from FC-MRD-HR group at day 15. The probability of 5-year EFS was estimated at 47% (95%CI:23%-94%) in PTEN.ABN cases vs 78% (95%CI:64%-95%) in PTEN.WT patients, P=0.03. Supporting Figure S9. Event-free survival (pEFS) according to the status of PTEN.ABN in ALL IC-BFM 2009 patients from FC-MRD-IR and FC-MRD-HR groups at day 15. A. pEFS in patients from FC-MRD-IR group according to PTEN.ABN status. The probability of 5-year EFS was estimated at 50% (95%CI:19%-100%) in PTEN.ABN cases vs 89% (95%CI:76%-100%) in PTEN.WT patients, P=0.049. B. pEFS in patients from FC-MRD-HR group according to PTEN.ABN status. The probability of 5-year EFS was estimated at 25% (95%CI:5%-100%) in PTEN.ABN cases vs 89% (95%CI:77%-100%) in PTEN.WT patients, P<0.001. Supporting Figure S10. Event-free survival (pEFS) according to NOTCH1 status. The probability of 5-year EFS was estimated at 77% (95%CI:69%-88%) in NOTCH1 mutated patients vs 70% (95%CI:67%-88%) in NOTCH1.WT, P=0.39. Supporting Figure S11. Event-free survival (pEFS) according to NOTCH1 and FBXW7 status. The probability of 5-year EFS was estimated at 78% (95%CI:69%-88%) in NOTCH1/FBXW7 mutated patients vs 69% (95%CI:59%-80%) in NOTCH1 and FBXW7.WT patients, P=0.24. Supporting Figure S12. Event-free survival (pEFS) according to the gLoR and gHiR status. gLoR - NOTCH1/FBXW7 mutations with no PTEN alterations; gHiR - no NOTCH1/FBXW7 mutations with no PTEN alterations; or no NOTCH1/FBXW7 mutations and the presence of PTEN alterations; or NOTCH1/FBXW7 mutations and PTEN alterations. The probability of 5-year EFS was estimated at 78% in gHoR patients vs 67% in gLoR patients, P=0.22. Supporting Table S1 Clinical characteristics of patients treated according to ALL-IC BFM 2002 and ALL-IC BFM 2009 protocols. Supporting Table S2 Characteristics of genomic regions covered by mutational and copy number analyses. Supporting Table S3 List of all changes detected by Sanger sequencing, including SNP, SNV and novel variants. Supporting Table S4 Characteristics of T-ALL patients according to their PTEN status. Supporting Table S5 Associations between mutated genes and clinico-biological characteristics of leukemia (univariate analysis). Supporting Table S6 Overview of PTEN prognostic significance in pediatric and adult T-ALL. Supporting Table S7 Cox regression model applied to evaluate the significance of PTEN.ABN after adjusting for major known prognostic factors in T-ALL. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.