Cluster Analysis of Apoptosis-Associated bcl2 Family Proteins in Diffuse Large B-cell Lymphomas. Relations with the Apoptotic Index, the Proliferation Profile and the B-cell Differentiation Immunophenotypes

Background: There is evidence that apoptotic mechanisms mediated by bcl2 family proteins are involved in the pathogenesis of diffuse large B-cell lymphomas (DLBCL). In order to gain further insight into the apoptosis profile of DLBCL, 79 cases were investigated to determine whether distinct clusters of the combined expression levels of bcl2 family proteins can be identified in these lymphomas. Materials and Methods: The combined immunohistochemical expression levels of the proteins bax, bak, bad, bid, bcl2 and bcl-xl were evaluated by cluster and discriminant analysis. The produced clusters were analyzed in relation to the apoptotic index, the proliferation profile and the B-cell differentiation immunophenotypes. Results: Cluster analysis produced: a) a low expression (69/79 cases) and a high expression pro-apoptotic cluster (10/79 cases) for the combined expression levels of the pro-apoptotic proteins bax, bak, bad and bid and b) a low expression (37/76 cases) and a high expression anti-apoptotic cluster (39/76 cases) for the combined expression levels of anti-apoptotic proteins bcl2 and bcl-xl. The decreasing order of discriminant power for the percentages of tumor cells expressing pro-apoptotic and anti-apoptotic proteins was % bax + cells>% bak+ cells>% bid+ cells>% bad+ cells and % bcl2+ cells>% bcl-xl+ cells, respectively. The high expression pro-apoptotic cluster was significantly associated with higher mean values of Ki67 (p=0.047) and cyclin A (p=0.033) expression. The high expression pro-apoptotic cluster was significantly associated with the germinal center B-cell bcl6/CD10/MUM1/CD138 differentiation immunophenotype (p=0.043). Conclusion: This study identified distinct clusters of DLBCL with respect to the combined expression levels of the apoptosis-associated bcl2 family proteins. These findings, taken together with our previous observations that distinct clusters with respect to the apoptotic index and the proliferation profile are identified in DLBCL, indicate that subgroups with distinct cellular kinetic properties can be defined in these lymphomas. The cluster analysis approach might be useful for the identification of subgroups of DLBCL with different clinical behavior since increased proliferation and apoptosis were reported to be associated with aggressive tumor behavior in these lymphomas.