Most successful treatments invite inquiry — for example, why they work and whether there are ways to improve upon them. Rituximab, in the treatment of B-cell lymphoma and autoimmunity, is no exception, and the relative therapeutic contributions of Fc receptor-γ (FcRγ)-mediated cellular cytotoxicity, complement fixation, and direct cell surface signaling remain contentious. The characterization of two different families of anti-CD20 antibody has highlighted this: type I (rituximab-like) antibodies move the antigen into lipid rafts and are effective complement activators, whereas type II (tositumomab-like) do not redistribute antigen, but instead mediate homotypic adhesion and direct cell killing.1 A group of researchers led by Glennie and Cragg from the University of Southampton, UK, has been investigating the underlying differences between these two types using a mouse model in which B cells express the human CD20 antigen. In this model, infusion of type II antibodies produced a much longer lasting depletion of B cells than did type I antibodies. Using mice genetically deficient in complement components, they showed the effect was independent of complement, and, using Bcl-2 overexpressing cells, they showed that resistance to apoptosis had no effect either. However, in FcRγ-chain null mice, the therapeutic effect was lost for both types of antibody, as it was when Fab fragments were used, confirming the role of FcR engagement. An examination of expression on human CD20 transgenic B cells transferred into FcR null mice showed that the antigen was apparently lost from the cell surface in response to type I antibody treatment, but not after type II. Using a fluorescence quenching assay, the authors went on to show that the loss of cell surface CD20 was due to internalization within early endosomes and lysosomes. This was accompanied in vivo by a significant shortening of the antibody half-life for type I but not type II and a consequent reduction in phagocytosis by macrophages in immunocompetent hosts.
CD20 has previously been thought of as a highly stable cell surface molecule, based upon short-term studies of B-cell lines and small numbers of patient samples. Indeed, this was put forward as one reason for the success of anti-CD20 therapeutics — that the antigen did not modulate. There have been a few reports of apparent loss of CD20 expression in recurrent lymphoma following antibody treatment.2 In this paper, however, further studies of both normal peripheral blood B cells and primary malignant B cells ex vivo showed that type I anti-CD20 antibodies produce significant modulation of the antigen after six hours, an effect which was not seen with type II reagents. Interestingly, there was significant heterogeneity between different malignancies, with the relatively rituximab-sensitive germinal center lymphomas showing very little modulation, but CLL and mantle cell lymphoma showing much more, a finding which may relate to the differential effects seen in the clinic.
In Brief
Although these experiments to elucidate the mechanisms of action of anti-CD20 antibodies may seem distant from routine clinical practice, they have important messages about how we might improve the results of treatment. First, there are several type II antibodies in clinical trials already, and it may be that these will be preferable for the treatment of those malignancies that modulate CD20 rapidly. Whether they are likely to prove superior in non-modulating lymphoma types is less clear. Second, these observations may provide a rationale for the apparently greater efficacy of high-dose type I anti-CD20 in CLL, where rapid modulation may shorten the effective half-life in many cases. Finally, if we can unravel the process by which modulation is mediated on the cell surface, we may be able to devise strategies to block it, and thereby reduce resistance to therapy.
References
Competing Interests
Dr. Johnson indicated no relevant conflicts of interest.
