Identification of Genetic and Expression Markers in DLBCL Cell Lines That Influence the In Vivo Response of DLBCL to Doxorubicin.

Diffuse large B-cell lymphoma (DLBCL) is the largest aggressive group of non-Hodgkin’s lymphoma (~40%) and exhibits great clinical and pathological heterogeneity. While DLBCL represents a chemosensitive disease, ultimately, only approximately 50% of patients are cured with current treatment options. Standard front-line therapy comprises CHOP and more recently Rituximab-CHOP. CHOP commonly contains doxorubicin as the curative anthracycline. Controversy currently exists in the use of genetic and expression markers in predicting response to treatment and outcome of DLBCL, thus, we examined the cellular responses of a panel of eight DLBCL cell lines to doxorubicin in order to define molecular markers that influence the response. Expression profiling and subsequent classification of the cell lines revealed that six of the lines were of a germinal center B-cell-like (GCBL) subtype and two (Ly3 and Ly10) were of a non-GCBL subtype. Ly3 and Ly10 exhibited a rapid apoptotic response within 24 hours after exposure to doxorubicin as evaluated by Hoescht staining. Expression profiling of the response of Ly10 cells (in triplicate) revealed the up-regulation of transcripts consistent with mediation of the cellular response via a p53-dependent pathway. Mutation analysis and Western blotting revealed wild type TP53 in Ly3 and Ly10 with an induction of p53 levels after doxorubicin. The remaining cell lines exhibited mutated TP53 sequences and no induction of p53 by doxorubicin. Therefore, in DLBCL cell lines, wild type p53 effects an early rapid apoptotic response to doxorubicin. In the six non-GCBL cell lines, a decrease in viability after doxorubicin was observed at much later times (up to 48 hours after treatment) by both apoptotic (exemplified by Ly2 and Ly8) and non-apoptotic mechanisms (exemplified by Ly1). Expression profiling of these three cell lines revealed alterations in expression of transcripts consistent with a G2/M cell cycle delay induced by doxorubicin, which was confirmed by follow-up flow cytometric analysis. An induction of transcripts consistent with NF-κB and AP-1 activation was observed in Ly1, but not in Ly2 and Ly8, which was confirmed by electrophoretic mobility shift assays. In order to determine if this cellular response plays a role in determining the response of DLBCL to doxorubicin in vivo, the expression profiles of untreated Ly1 and Ly2 were compared, and the resulting 160 differentially expressed transcripts were used to cluster 41 newly-diagnosed DLBCL specimens which had previously been profiled and had a known response to doxorubicin-containing therapy. This analysis revealed a marked correlation between capacity to induce NF-κB and AP-1 with in vivo treatment failure. Overall, these experimental studies have begun to outline genetic and biologic features that together contribute to the response of DLBCL to doxorubicin in vivo.