Response: Complement in antibody therapy for lymphoma: both a help and a hindrance?

Beers, Cragg, and Glennie have made major contributions to our understanding of the mechanisms of action of anti-CD20 antibodies, and we appreciate their thoughtful response to our manuscript.¹  As they have cited, there are animal models that suggest complement-dependent lysis contributes to the therapeutic effects of anti-lymphoma antibodies. The primary point of our paper was to raise the possibility that, in some circumstances, complement may hinder antibody-dependent cellular cytotoxicity (ADCC). Thus, when it comes to antibody therapy for lymphoma, we agree it is unclear whether complement is a “help or hindrance.”

Several points deserve specific comment. Beers et al refer to our prior work demonstrating that granulocytes contribute to the antitumor effects of antibody therapy in the 38C13 model.²  In that publication, we reported that both granulocytes and NK cells contribute to the therapeutic response. Activated natural killer (NK) cells produce cytokines that secondarily activate granulocytes.^3,4  Thus, enhancing NK-cell activation may improve the therapeutic response even if granulocytes play a role.

Beers et al correctly point out that cobra venom factor (CVF) may have effects other than “simple complement depletion” and can cause inflammatory lung damage.⁵  This effect has been shown to be mediated by the C5a anaphylatoxin, which binds and activates neutrophils, leading to sequestration in the lungs. HC3-1496 was more effective than CVF in our model, yet, in contrast to CVF, HC3-1496 does not activate C5.⁶  In primate studies, intra-arterial injection of high doses HC3-1496 resulted in no impairment of pulmonary function.⁷  These data suggest that induction of an inflammatory response was not responsible for the observed effects in our model. Nevertheless, we are currently exploring depletion of C3 using other approaches and in other models. Rechallenge of cured mice demonstrated no protection (S. Wang, G.J.W., unpublished data, June 2009); thus, we do not believe the therapeutic response we observed with the combination of CVF or HC3-1496 and antibody was due to development of an active immune response.

The data presented by Beers et al demonstrate that FcγR, but not complement, is necessary for B-cell clearance. This speaks to the importance of FcγR, but does not specifically address the role of complement in ADCC. Depletion of B cells by anti-CD20 in their studies was complete in both wild-type and C1q knockout mice; thus, any enhanced effect in the absence of complement could have been lost. Cure of a highly aggressive lymphoma, such as 38C13, may require a greater degree of NK-cell activation than is needed to demonstrate transient clearance of benign B cells from the spleen.

It may well be that the answer to the “helps or hinders” question depends upon the scenario, with complement-mediated lysis being more important in some circumstances and complement-inhibiting ADCC more important in others. The real answer will come both from further studies in animal models and from clinical trials.