To the editor:
There remains considerable uncertainty regarding the role of complement as an effector system for anti-cancer monoclonal antibodies (mAbs) such as rituximab and alemtuzumab. We were, therefore, interested to read the intriguing findings of Wang et al,1 indicating that complement activation can be detrimental to antibody immunotherapy. They suggest that breakdown products of the complement component C3 block anti-lymphoma mAb activity and render natural killer (NK) cells less active in antibody-dependent cellular cytotoxicity and the treatment of the 38C13 lymphoma with anti-idiotype mAb.1 Clearly, these are important observations, well supported by in vitro experimentation. However, we feel it is important to stress that the beneficial effect of complement inactivation in vivo is not the general experience with therapeutic mAb, where the absence of complement typically either reduces efficacy2,3 or makes no difference4,5 (reviewed in Lim et al6 ).
By way of example, Figure 1 shows that depletion of normal B cells with anti-CD20 mAb is equally effective regardless of whether the mice are deficient in C1q or C3 or the mAb is engineered not to engage C1q.
Several possibilities exist to explain these apparently contradictory findings. The first is that the anti-Id/38C13 lymphoma model is unusually dependent on NK cells, which, in turn, are particularly sensitive to blocking by C3b, consistent with the interpretation given by Wang et al. However, this seems unlikely, since previous work from this group has shown that granulocytes also play a role in this model.8 In addition, successful mAb immunotherapy in mice, including those with anti-CD20 mAb, commonly require an intact mononuclear phagocytic system, but not NK cells (reviewed in Lim et al6 ).
This leaves 2 further potential explanations. The first is that idiotype may be an atypical target for mAb. In support of this, Golay et al have demonstrated that therapy of a hCD20+ variant of 38C13 with CD20 mAb was completely dependent upon complement activity, with no role for NK cells or neutrophils.3 These findings are clearly counter to the results described by Wang et al, perhaps indicating that impact of complement on therapy is determined by the degree of complement activation, with the high levels evoked through CD20 mAb usually being beneficial to therapy, and lower levels, as occur with anti-idiotype mAb, being detrimental as observed. Second, it must be considered that cobra venom factor (CVF) treatment has functions other than simple complement depletion. Rather than removing complement passively, CVF acts as an unregulated analog of C3b, resulting in uncontrolled activation of C3 with systemic release of breakdown products that impact on innate and acquired immune activation. Hence, CVF treatment provides a useful model of acute respiratory distress syndrome where it leads to acute organ damage.9 The release of cytokines and chemokines and accumulation of neutrophils after CVF are precisely the inflammatory conditions that might influence the growth of a small number (5000) of passaged tumor cells, particularly with an immunogenic lymphoma such as 38C13.10 In accordance with this supposition, the results from Wang et al show that CVF alone slows the growth of 38C13 to that seen after treatment with anti-Id mAb. Therefore, the synergistic activity of anti-Id mAb and CVF leading to long-term survival is consistent with the development of acquired immunity. Thus, we feel that an adjuvant effect of CVF must be considered as a plausible alternative explanation of the current in vivo data, especially given the weight of prevailing evidence showing that complement is not deleterious to direct targeting antibody immunotherapy.
Acknowledgments: The authors thank the following for the provision of mice: Prof Mark Shlomchik (Yale School of Medicine, New Haven, CT) for hCD20 Tg; Prof Marina Botto (Imperial College, London, United Kingdom) and Dr Aras Kadioglu (University of Leicester, Leicester, United Kingdom) for C1q−/−; and Prof Jan van de Winkel (University Medical, Utrecht, The Netherlands) for FcγR−/−.
Conflict-of-interest disclosure: The authors declare no competing financial interests.
Correspondence: Martin J. Glennie, Tenovus Research Laboratory, Cancer Sciences Division (mailpoint 88), University of Southampton School of Medicine, Tremona Rd, Southampton, United Kingdom SO16 6YD; e-mail: email@example.com.