In this issue of Blood, Xu-Monette and colleagues report the results of TP53 mutational profiles in a cohort of 506 de novo diffuse large B-cell lymphoma patients treated with rituximab-CHOP, and they concluded that TP53 mutation was an independent adverse factor for survival.1
Diffuse large B-cell lymphoma (DLBCL) is the most common type of lymphoma and is characterized by an aggressive clinical course. However, DLBCL exhibits considerable heterogeneity in terms of clinical, morphologic, molecular, and cytogenetic features. Gene expression profiling (GEP) studies have identified 2 primary molecular subgroups of DLCBL: germinal center B (GCB) cells and activated B cells (ABCs).2 The GCB subgroup shows better survival than the ABC subgroup, independent of the international prognostic index (IPI). In the past decade, the introduction of the humanized monoclonal anti-CD20 antibody rituximab (R) to the combination chemotherapy of cyclophosphamide, hydroxydaunorubicin, vincristine, and prednisone (CHOP) has clearly improved the outcome of DLBCL.3 The improvement is mostly observed in patients with a low or intermediate risk based on the IPI and in the GCB subtype.4 The search for new biologic markers in neoplastic diseases is of major interest because they can provide a better understanding of the biology of the disease. They can also suggest the probability of treatment failure. In relapsed DLBCL patients, among those with early relapse and prior treatment with rituximab, the response rate of salvage therapy is only 46%, with a 25% 3-year progression-free survival (PFS).5 It is clear that a biomarker could be affected by new treatment,6 but conducting studies for designing more targeted treatment is essential. However, such a marker needs to be validated in a large cohort of patients. It should also be an independent parameter from clinical data and be much easier and less expensive to collect.
The choice for this study was the TP53 tumor suppressor gene, which plays an important role in the regulation of the cell cycle, cell proliferation, apoptosis, and genomic integrity. The p53 protein mediates cell-cycle arrest when cells experience stressful challenges, such as DNA damage, hypoxia, or oncogene activation, whereas mutant p53 protein results in cell-cycle dysregulation, genomic instability, and the uncontrolled proliferation of damaged cells (see figure). The presence of TP53 mutations has been associated with drug resistance, poor response to treatment, and short survival in several cancers, including DLBCL.7
Patterns of TP53 mutations in 506 cases of DLBCL patients treated with R-CHOP. (A) Proportion of classified point mutations. (B) Proportion of mutations based on mutation impact on the p53 protein sequence. (C) Proportion of mutations based on mutation effect on the p53 function. (D) Distribution of mutation numbers according to TP53 exons. The numerals at the top are numbers of mutations in exons, and the numerals at the bottom are numbers of mutations in splicing sites. (E) Codon distribution of TP53 mutations. Codons with mutations of high frequency in the cohort are marked. See Figure 1 in the article by Xu-Monette et al that begins on page 3986.
Patterns of TP53 mutations in 506 cases of DLBCL patients treated with R-CHOP. (A) Proportion of classified point mutations. (B) Proportion of mutations based on mutation impact on the p53 protein sequence. (C) Proportion of mutations based on mutation effect on the p53 function. (D) Distribution of mutation numbers according to TP53 exons. The numerals at the top are numbers of mutations in exons, and the numerals at the bottom are numbers of mutations in splicing sites. (E) Codon distribution of TP53 mutations. Codons with mutations of high frequency in the cohort are marked. See Figure 1 in the article by Xu-Monette et al that begins on page 3986.
In DLBCL patients treated with CHOP, the same group described that TP53 mutations were an adverse prognostic factor for survival but was restricted to patients with GCB-DLBCL.8 The focus of this new study was also TP53 but in a larger cohort of patients treated with rituximab and CHOP. To overcome some of the limitations of the technical aspects, they used gene sequencing microarray data from DNA extracted from formalin-fixed, paraffin-embedded (FFPE) tissues, gene expression data from RNA extracted from FFPE tissues, fluorescence in situ hybridization results for TP53 gene deletion, and immunohistochemical staining. Bioinformatics and database support were widely used for the analysis.
The incidence of TP53 mutations was 21%, and there was no difference between the 2 subgroups. GCB-DLBCL had a significantly better survival than ABC-DLBCL. There was a significantly better survival for patients with wild-type (WT) TP53 compared with mutated (MUT) TP53 for both subtypes, GCB- and ABC-DLBCL. In multivariate analysis, an IPI score of more than 2, TP53 mutations, the ABC subtype, and B symptoms were independent factors that predicted worse survival and PFS. As previously noted, distinctive molecular types can be reflected in IPI parameters. Mutations were associated with a high proliferation index (Ki-67 > 70%). Protein expression with a cut-off of more than 50% by standardized immunohistochemistry (IHC) was a surrogate marker for TP53 mutation status. From a practical perspective, if gene mutation data are not available, immunohistochemistry staining can be used to stratify patients between GCB and ABC subtypes. Further stratification can be achieved according to p53 protein expression between WT TP53 and MUT TP53. The data were validated in this study with sophisticated technology and bioinformatics, and IHC could be standardized for clinical study. This biomarker would mostly be interesting for the GCB subtype with better prognosis, in which the response to treatment can be related to the type of chemotherapy. In the relapsed DLBCL patients from the Collaborative trial in Relapsed Aggressive Lymphoma (CORAL) study,9 51% of the patients had the GCB subtype, and 49% had the ABC subtype according to the Hans algorithm.10 Patients with GCB DLBCL who were treated with rituximab, dexamethasone, aracytine, and cisplatin (R-DHAP) had a better PFS than patients with non-GCB DLBCL (3-year PFS = 52% vs 32%, respectively, P = .01). Patients treated with rituximab, ifosfamide, carboplatin, and etoposide (R-ICE) had a poor PFS, with no significant difference between GCB and non-GCB types. Multivariate analysis showed an independent prognostic impact of the following parameters: GCB/non-GCB interaction with treatment (P = .04), prior rituximab exposure (P = .0052) and secondary IPI (P = .039). The incidence of the GCB subtype in relapsed patients suggests that other parameters in addition to an IPI score of more than 2 may play a role. TP53 mutations were found in 26% of de novo GCB-DLBCL patients and should be considered as a key factor. In addition, fluorescence in situ hybridization has demonstrated that the survival of DLBCL patients is influenced by 8q24/MYC overexpression. Although most current treatments rely on the dose modulation of cytotoxic drugs, targeted therapy that may act independently of the DNA damage pathway is becoming available and should be evaluated according to the tumor molecular profile.
The present study is a great example of how we can make progress in studying a large cohort of patients through a collaborative effort. It demonstrates that the introduction of rituximab to chemotherapy did not change the prognostic value of TP53 mutation. However, it also shows that translation to more accessible procedures is possible in future clinical trials.
Conflict-of-interest disclosure: The author declares no competing financial interests. ■
