Vive la résistance: stem cells in mantle cell lymphoma

Much interest in cancer stem cells is inspired by the belief that they resist standard therapies, spawning recurrent tumors in which the bulk population has become resistant as well [1]. Although attractive as an explanation for clinical observations, and not without mechanistic underpinnings, this belief is largely a hypothesis, with limited supporting data in most types of cancer. In this issue of Leukemia and Lymphoma, McCarty and colleagues address this in mantle cell lymphoma (MCL), extending their earlier work on a CD45+CD19− subpopulation termed MCL-initiating cells (MCL-ICs) which can reproduce the tumor in immunodeficient mice, satisfying the stem cell criterion of self-renewal [2]. They now report that as compared to the CD45+CD19+ bulk of MCL cells, MCL-ICs are substantially more resistant in vitro to combinations of standard agents modeling regimens frequently used in patients, with IC₅₀ (concentration of a drug required for 50% inhibition in vitro) values exceeding concentrations achievable in vivo [3]. This chemotherapy-resistant MCL-IC population may explain why MCL has generally been the least curable form of non-Hodgkin lymphoma, despite an adequate rate of achieving complete remission [4]. Jung et al. also provide two insights into resistance by MCL-ICs: (1) standard agents are synergistic when used in combination versus CD45+CD19+ MCL cells, but not (or are even antagonistic) versus MCL-ICs, and (2) inhibition of the ABCB1 drug transporter, highly expressed by MCL-ICs, increases their sensitivity to vincristine. Drug transporters have often been cited as a potential reason for drug resistance in cancer stem cells, but this is specific evidence that suggests therapeutic application.

This report has significant limitations, but some are typical of studies of this type. Toxicity was determined after only 6 h of drug exposure, because Jung et al. found that both bulk MCL cells and MCL-ICs began to show evidence of spontaneous apoptosis if cultured for longer periods. Clearly this period of drug exposure is shorter than what prevails in vivo, although intravenously administered drugs may be rapidly eliminated, and it raises the question of whether differences in IC₅₀ may have been due only to differences in the rate at which drug-induced apoptosis was executed. Brief exposures with viable cells are probably more valid than commonly used longer exposures with cells that are already beginning to die, but neither is ideal, and both are necessitated by the lack of a microenvironment that can maintain cell viability and model the in vivo state. Standard in vitro testing cannot assess agents whose mechanisms of action depend on, or affect, the microenvironment; lenalidomide, a promising new agent against MCL, is a prime example that was not tested. Rituximab was tested, but its effect mediated by antibody-dependent cellular cytotoxicity was not assessed, due to the absence of normal immune cells. When utilizing serum without heat-inactivation, rituximab was shown to be toxic only to CD45+CD19+ MCL cells, even though both they and MCL-ICs express CD20. This is of clinical and mechanistic significance, since it indicates that resistance by MCL-ICs does not depend solely on drug transporters. The short duration of the cultures precluded modeling of regimens in which agents are given at different times, which may affect their efficacy; Hyper-CVAD (hyperfractionated cyclophosphamide, vincristine, doxorubicin, dexamethasone) is an effective (although non-curative) regimen in which fractionated cyclophosphamide is alternated with methotrexate and cytarabine [5]. Other limitations were the lack of tests using bortezomib or other proteasome inhibitors, which have efficacy alone or in combinations against MCL in patients, and the use of cyclophosphamide (a pro-drug requiring hepatic microsomes for activation) rather than its active congener 4-hydroperoxycyclophosphamide [6].

Despite its limitations, this report provides two important interdependent conclusions: (1) MCL-ICs are more resistant to standard therapy than are bulk MCL cells, and (2) this difference can be demonstrated in an experimental setting. The first conclusion should be validated in vivo, and the report suggests one way how it could be done. The frequency of MCL-ICs among all MCL cells, as determined by low staining for CD19 or Rhodamine 123 (extruded by drug transporters), increased with combination therapy. If MCL patient samples were examined before and after therapy (at least before regrowth occurs), it would be supportive to find that CD19−CD20+ MCL-ICs were relatively increased. Anticipating a positive result, the fundamental challenge remains: how to make therapy more effective versus MCL-ICs. Jung et al. suggest that quiescence of MCL-ICs may promote resistance to standard therapy; no evidence is provided other than that MCL-ICs are predominantly in G0 phase, but strategies are now emerging to activate quiescent cancer stem cells [7]. We need more detailed insight into the biology of stem cells in specific cancers; agents that can target their distinctive features; and testing in preclinical models that replicate the in vivo state and allow longer-term testing for the multi-log killing of stem cells required for greater efficacy or cure. Given the magnitude of the problem, investments in genomic profiling of MCL-ICs, and the use of immunodeficient mouse models that support growth of primary MCL cells [8], are warranted.

Potential conflict of interest:

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