Molecular Profiling in Lymphoma Comes of Age

The emergence of gene-expression analysis has added a powerful new weapon to the oncologist’s armamentarium in the classification of neoplastic diseases and in the identification of prognostic indicators. The basic premise underlying this approach is that tumors that appear to be identical from a purely morphologic standpoint may in fact harbor highly disparate genetic profiles. A corollary of this notion is that individual genetic patterns may be associated with specific biologic behaviors and may result in divergent responses to chemotherapy and/or in overall survival.

One of the first attempts to validate this concept occurred in the case of diffuse large B-cell lymphoma (DLBCL), an aggressive but potentially curable form of non-Hodgkin lymphoma. Molecular profiling of DLBCL helped to identify three distinct subtypes.¹ These include germinal center (GC)-DLBCL, which arises from germinal center B cells, and activated B-cell (ABC)-DLBCL, which arises from post-germinal center B cells. A third subtype, which is relatively rare, is the primary mediastinal B-cell lymphoma (PMBL). Most notably, ABC and GC-DLBCL display numerous differences from each other, including distinct karyotypic profiles, abnormalities in miRNA expression, and dys-regulation of various signaling pathways. One of the most striking differences between the ABC and GC subtypes is that the latter cells are characterized by constitutive activation of NF-κB and appear to be dependent upon activation of this transcription factor for survival. This is based on results from preclinical studies indicating that ABC-DLBCL cells are significantly more sensitive to agents that interrupt the NF-κB pathway than their GC counterparts. Significantly, ABC-DLBCL lymphomas are significantly less responsive to conventional chemotherapy than the GC and PBML subtypes and are associated with a significantly shorter five-year survival.^{2, 3}

Until recently, molecular profiling provided potentially useful prognostic information for patients with DLBCL but gave little specific guidance in the optimal selection of agents or regimens. However, this situation may have changed with this recent report by Dunleavy et al., from NCI and the Roswell Park Cancer Center. In a large multi-institutional trial, they compared responses of patients with refractory ABC- or GC-DLBCL to a regimen consisting of dose-adjusted EPOCH (etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin) with bortezomib. The latter agent was selected, because in preclinical studies bortezomib has been shown to inhibit NF-κB by preventing degradation of IκBα, a protein that traps NF-κB in the cytoplasm, thus preventing its nuclear translocation and activation of cytoprotective NF-κB-dependent genes. While bortezomib has significant activity in multiple myeloma and mantle cell lymphoma, it has little singleagent activity in DLBCL. The authors hypothesized that an NF-κB antagonist such as bortezomib might be particularly active against the ABC-DLBCL subtype.

The results of this study, which involved a total of 49 patients, were quite striking. Although in controls, patients with ABC- and GC-DLBCL had equally poor outcomes, responses of ABC-DLBCL patients to bortezomib-containing chemotherapy were significantly better (P<0.001) than those of patients with GC-DLBCL (83%; 5 CR and 5 PR) versus 13 percent (1 CR and 1 PR), as was survival (10.8 months vs. 3.4 months). The authors concluded that bortezomib-containing regimens are most appropriate for relapsed DLBCL with the ABC disease subtype.