Targeting drug resistance to close the gap in diffuse large B-cell lymphoma

The ability to predict which patients with diffuse large B-cell lymphoma (DLBCL) will fail treatment with R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone) would be extremely valuable, since this is one of the most commonly encountered types of lymphoma in clinical practice. The past decades have witnessed major advances in understanding the biology of DLBCL with the use of molecular biological techniques. The clinical heterogeneity of this disease, seen at the bedside, has been clarified in particular by the seminal studies of Shipp et al. [1] and Rosenwald and co-workers [2], whose gene expression profiles proved that DLBCL prognosis was a function of gene expression of hundreds if not thousands of genes. Shortly thereafter, similar studies revealed that the heterogeneity seen at the clinical and histological levels correlated with variations at the cellular level [3]. These insights have enabled sharpening the predictive power of the clinicians, and have even been distilled to the expression of smaller numbers of genes [4]. However, although there were high hopes over a decade ago that these insights would pave the way for an “armamentarium of molecularly targeted therapies, which would be an integral component of clinical management” [5], this has not yet proven to be the case. In fact a paradox arises from considering the predictive genes that have been identified in molecular profiling of lymphomas. Since most patients with DLBCL die of chemotherapy resistance, why is it that gene expression profiling studies do not identify high expression of drug resistance genes as possible prognostic markers? One possibility is that gene expression profile analysis begins by seeking genes with high variance. However, drug resistance genes do not need high expression to confer a poor prognosis, since the therapeutic index of most chemotherapeutic drugs is very small. Even low levels of overexpression can cause failure of chemotherapy, before causing excessive toxicity. This paradox epitomizes the current dichotomy between delineating prognostic biomarkers and determining markers of clinical efficacy, and provides a partial explanation of why gene profiling has not yet led to tangible breakthroughs in treatment.

In this issue of Leukemia and Lymphoma, Nobili and colleagues attempt to make inroads into achieving the latter goal using pharmacogenetic analysis to confirm genes that are associated with failure of the R-CHOP14 regimen in DLBCL [6]. The authors elected to study five genes that would be expected to cause resistance to anthracyclines, which are essential in the success of CHOP-like regimens. Interestingly they have found, in a relatively small group of 54 patients, that high expression of GSTP1 and topoisomerase 2α were both associated with poor survival following R-CHOP14 treatment, although interestingly, not linked with objective response. While this study did not find that the expression of these genes could outperform either the International Prognostic Index (IPI) score or the germinal center–non-germinal center histological categorization in predicting survival, their results are important nonetheless for emphasizing potential therapeutic targets.

Topoisomerases are nuclear enzymes that resolve topological problems associated with DNA during various genetic processes such as replication and transcription. High topoisomerase 2α has already been found in DLBCL and correlates with poor survival [7]. More importantly, in DLBCL, topoisomerase 2α was significantly associated with response to chemotherapy [7]. Topoisomerase inhibitors are among the standard agents used to treat DLBCL, and include anthracyclines and epipodophyllotoxins. Anthracyclines tend to be topoisomerase 2α specific while epipodophyllotoxins tend to target topoisomerase 2β [8]. Topoisomerase inhibitors are effective agents but have a very narrow therapeutic index due to their limited target cell selectivity, such that they can be associated with severe dose-limiting toxicity. Pharmacological approaches to increase the therapeutic index have been proposed [8]. Alternateivly, new, less toxic agents can be found and perhaps utilized, such as resveratrol, a natural compound found in grapes and wine, which has been found to have topoisomerase 2α inhibiting activity [9]. Glutathione S-transferases (GSTs) are a family of enzymes that catalyze the conjugation of many hydrophobic and electrophilic compounds with reduced glutathione. GST enzymes are variable in expression, and have often been examined for their effect on predisposition to malignancy rather than for their effect on chemotherapy resistance. GSTP1 expression has been found to be particularly high in mantle cell lymphoma, and high GSTP1 expression was found to predict poor survival in a number of lymphoma cell types, including DLBCL [10]. Over 1000 such compounds are known that inhibit GSTP1, but it appears that none are sufficiently potent to have a role in anticancer therapy [11].

Both topoisomerase 2α and GSTP1 are targets for drug delivery. The fact that they were discovered to be predictive in a small number of patients does not detract from the significance of the findings, since current therapeutic use of many neoplastic drugs is selective, based on histological or molecular analysis showing evidence which predicts response to a particular group of agents in a subset of patients. The findings of Nobili et al. [6] suggest that in the future, selection of patients for targeted therapy based on drug resistance gene expression which is specific to their DLBCL could perhaps result in improved treatment outcome. Although as far back as 18 years ago it was predicted that the gap between biology and clinical management of lymphomas was closing [12], this has not yet been realized. Practical approaches such as targeting drug metabolism and resistance genes could indeed help close this gap in DLBCL.

Potential conflict of interest

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.