We recently conducted the first phase I dose escalation study of intraventricular rituximab (anti-CD20 antibody) in patients with recurrent central nervous system (CNS) non-Hodgkin’s lymphoma (NHL). The goals of this study have been to define the safety and pharmacokinetics of intraventricular rituximab monotherapy and to gain insight into the molecular basis of rituximab efficacy and resistance in the treatment of high-grade lymphomas. We observed a rapid lymphocytotoxic effect of intraventricular rituximab administration in six out of ten subjects treated, resulting in four complete responses at the completion of five weeks of therapy. Cytologic responses were evident within one hour of rituximab treatment and were not likely related to the action of corticosteroids as these were held stable, tapered or withheld in the week before intrathecal therapy. Three potential mechanisms have been proposed to explain the basis for rituximab efficacy in the treatment of NHL:

  1. direct induction of apoptosis;

  2. complement-dependent cytotoxicity (CDC);

  3. antibody dependent cellular cytotoxicity (ADCC).

Given the rapid rate of response detected in the leptomeningeal compartment and the relative paucity of Fc-receptor bearing natural killer cells detected in the CSF, we tested the hypothesis that complement activation is responsible for the rapid rituximab-mediated tumor cell lysis which we observed. Pretreatment gene expression profile analysis of meningeal lymphoma cells isolated from CSF demonstrated that RNA levels of expression of CD46, CD55, and CD59 were similar in both responding and non-responding patients. However microarray analysis suggested a possible correlation between low levels of complement receptor 2 (CD21) and immediate rituximab resistance. We focused on the potential generation of C3a anaphylatoxin in the CSF after rituximab administration as C3 is the central component in the classical and alternative complement pathways. Quantitative measurement of C3a concentration in CSF by ELISA demonstrated rapid activation of the complement cascade within one hour of intrathecal anti-CD20 antibody administration. C3a concentrations increased between three-to-eight-fold up to peak concentrations greater than 300 ng/ml. These findings were confirmed by immunoblot analysis of C3a in CSF. Acquired rituximab resistance was associated with marked decreases in the magnitude of complement activation in two patients who initially had exhibited complete responses to intraventricular rituximab monotherapy. These results suggest a mechanistic basis for the rapid lymphocytoxic effects observed and provide a possible explanation for the development of acquired rituximab resistance.

Author notes

Disclosure:Research Funding: We are discussing results from a research study funded by Genentech. Membership Information: Dr. Rubenstein is a member of the Genentech Speakers Bureau. Off Label Use: We will describe results of an off-label use of rituximab.