Binding of Submaximal C1q to B Cells Opsonized with Anti-CD20 Mabs Ofatumumab (OFA) or Rituximab (RTX) Promotes Complement Dependent Cytotoxicity (CDC), and Considerably Higher Levels of CDC Are Induced by OFA Than by RTX.

The CD20 mAb OFA is more effective than RTX in promoting lysis of opsonized B cells via classical pathway CDC, likely due to the fact that OFA targets a unique small loop epitope of CD20 close to the surface of the B-cell. Classical pathway CDC is initiated by binding of C1q to aggregated IgG bound to cells. Here we sought to determine the role of C1q in the dynamics of complement (C) activation and CDC of B cells reacted with the CD20 mAbs OFA or RTX. We first examined binding and cellular co-localization of C1q with the CD20 mAbs using high resolution digital imaging of cells by flow cytometry (Amnis ImageStream). Daudi cells were incubated in media with Al488 C1q and either Al647 OFA or Al647 RTX and evaluated for co-localization. Similarity bright detail scores, calculated based on a log transformation, were 2.5 to 3 for C1q with OFA, indicating substantial co-localization of C1q with cell-bound OFA. Comparable scores for cells opsonized with RTX were significantly less (1.5 to 2), indicating less co-localization. Greater co-localization of captured C1q with cell-bound OFA may promote a higher level of CDC. Therefore, we next used flow cytometry to measure C1q binding and CDC. Varying concentrations of unlabeled C1q and saturating concentrations of the Al647-labeled CD20 mAbs (10 μg/ml), were reacted with B cells at 37°C in media or in C1q-depleted serum. Time course studies showed that >90% binding of the mAbs was achieved in 6 minutes, but C1q binding (measured by secondary development with an anti- C1q antibody) to the opsonized cells was slow, requiring more than 1 hour for maximal binding to be achieved at C1q inputs of 35 μg/ml. Dose-response studies indicated that small amounts of C1q were bound to the cells after 1 hour for C1q input concentrations of 1–2 μg/ml, but much greater binding of C1q was achieved for input concentrations of 35 μg/ml. Moreover, under all conditions tested, much higher levels of C1q binding were achieved for cells reacted with OFA compared with cells reacted with RTX. For example, after 20 minutes’ incubation at 2 μg/ml C1q, binding of C1q to Raji cells in the presence of OFA corresponded to 13,600 molecules of equivalent soluble flurochrome (MESF) units, while binding in the presence of RTX corresponded to only 1,340 MESF units. These differences between CD20 mAbs were demonstrable in CDC assays. Brief incubation of cells with OFA or RTX in C1q-deficient serum supplemented with small amounts of C1q were shown to be adequate to promote substantial CDC, even though the amount of C1q bound to the cells corresponded to only ~5% of maximal C1q binding. In a representative experiment, Daudi cells were incubated for 5 minutes at 37°C in C1q-depleted serum supplemented with varying concentrations of C1q, in the presence of either 10 μg/ml OFA or 10 μg/ml RTX. CDC (based on propidium iodide staining) for cells reacted with OFA and C1q was 39, 56 and 59%, for C1q concentrations of 2, 5 and 10 μg/ml, respectively. CDC for cells similarly reacted with RTX and C1q was 9, 22 and 32%, respectively. Moreover, even in cases in which C1q binding was comparable for OFA- and RTX-opsonized cells (use of higher input concentrations of C1q were required for RTX-opsonized cells), CDC was still greater for OFA-opsonized cells. These results indicate that binding of very small amounts of C1q to mAb-opsonized cells, far below the maximal C1q binding capacity of the cells, is sufficient to promote CDC. Greater CDC induced by OFA compared with RTX may be due in part to a higher level of binding of C1q to OFA-opsonized B cells. Also, cell-bound C1q is more closely associated with OFA (based on the co-localization studies), likely functioning more effectively to activate the classical C pathway and promote CDC.