Neutrophils: positive or negative?

In this issue of Blood, Valgardsdottir et al demonstrate that human neutrophils do not kill chronic lymphocytic leukemia (CLL) B cells opsonized with the CD20 monoclonal antibody (mAb) rituximab (RTX) or obinutuzumab (OBZ); instead, the neutrophils remove both CD20 and bound mAbs from B cells by trogocytosis.¹ 

The CD20 mAb RTX was approved 20 years ago for treatment of relapsed or refractory, low-grade or follicular, B-cell non-Hodgkin lymphoma, and since that time, several next-generation CD20 mAbs have been approved or are under development for a variety of B-cell–associated malignancies.²  Although therapies that include CD20 mAbs (usually in combination with chemotherapy) have proven to be effective, in almost all cases they are not curative.³  In fact, often patients’ tumors become refractory to additional CD20 therapy, and in several instances, this has been attributed to loss of CD20 from targeted B cells.⁴  Therefore, there is a real need to enhance the potency of CD20 mAbs based on clearly delineating their mechanisms of action and the reasons for decreases in their efficacy. These issues have been the subject of intense investigations and some controversy.⁵  The report by Valgardsdottir et al clarifies the role of neutrophils in CD20 immunotherapy.

There is now consensus that CD20 mAbs must make use of immune effector functions to eliminate targeted B cells.^2,5-8  mAb-opsonized B cells are subject to killing by cells that express Fcγ receptors (FcγRs), which allows these effector cells to engage the immune complexes composed of mAb immunoglobulin G molecules chelated to CD20 on targeted B cells. This initiates activating and signaling cascades specific to effector cells. Natural killer (NK) cells eliminate B cells by antibody-dependent cell-mediated cytotoxicity (ADCC); the B cells are directly lysed and killed as a result of NK-cell secretion of granzymes and perforin into the B cells. In addition, opsonized B cells are subject to phagocytosis and elimination by macrophages. Complement can mediate clearance and direct killing of mAb-opsonized cells,⁷  but in the present context, this mechanism is not relevant to the work of Valgardsdottir et al.

FcγR-mediated trogocytosis is a reaction in which effector cells that are capable of killing CD20 mAb-opsonized cells instead form an immunological synapse with opsonized cells based on chelation of B-cell–bound immune complexes by FcγRs on effector cells. After formation of the synapse, the B-cell–bound CD20 mAbs, associated CD20, and nearby B-cell plasma membrane are transferred from the B-cells and taken up and internalized by acceptor cells via their FcγRs.⁸  The B cells then emerge alive but with considerably reduced CD20 target. Obviously this reaction can substantially decrease the efficacy of CD20 mAb–mediated therapy, and several clinical investigations and in vitro models have indeed demonstrated that after ADCC and complement are exhausted, this reaction can lead to almost complete loss of CD20 on CLL cells when patients are treated with the usual high doses of RTX or ofatumumab.⁹  The role of neutrophils in these processes is uncertain, but the findings of Valgardsdottir et al clearly place neutrophils firmly on the negative side of the ledger with respect to CD20 mAb–mediated immunotherapy. Although the investigators’ previous investigations as well as those of other groups suggested that neutrophils could kill CD20 mAb–opsonized B cells, the current analyses, based on combinations of time-lapse microscopy, confocal microscopy, and flow cytometry, reveal otherwise. Neither RTX nor OBZ mediate killing by neutrophils; only trogocytosis is observed.

The careful experimental design and interpretation of results in this report are important and informative and provide a cautionary note for future investigations of phagocytosis in similar systems. The investigators recognize that flow cytometry experiments can confuse phagocytosis with trogocytosis. That is, if membranes of mAb-opsonized donor cells are labeled with a fluorescent dye, and these opsonized cells are reacted with acceptor cells, then transfer of relatively small amounts of dye to acceptor cells (compared with no transfer with controls), observed over several time points, is actually diagnostic for trogocytosis. If full phagocytosis of donor cells by the acceptor cells were to occur, then acceptor cells would be expected to have fluorescence signal intensities approximately equal to those of donor cells. The investigators compared trogocytosis with another reaction, internalization, which can result in reduction of CD20 on mAb-opsonized cells. They found that over 3 hours, neutrophils promoted substantial loss of CD20 from RTX-opsonized B cells, but internalization (conducted in the absence of neutrophils) was quite modest, in excellent agreement with findings of Beum et al,¹⁰  who reported that monocyte-mediated trogocytosis of CD20 on opsonized B cells was much faster than internalization.

FcγR-mediated trogocytosis has been demonstrated for several other mAbs and their cellular targets,⁸  thus extending the importance of this report. Therefore, several research directions are indicated, with the goal of developing strategies that preserve cytotoxic mechanisms but prevent trogocytosis. First, during mAb therapy, after effector functions are exhausted, it will be important to determine which immune effector cells play the most important role in trogocytosis. Candidates now include circulating monocytes, neutrophils, NK cells, fixed tissue macrophages, and liver sinusoidal endothelial cells (LSECs).⁸  In the case of CLL, circulating cells in blood samples obtained during and after CD20 mAb therapy should be examined. In particular, after trogocytosis, can CD20 and/or RTX be found inside neutrophils or other effector cells? If appropriate mouse models can be developed, analyses can be extended to include macrophages and LSECs. Although type II mAb OBZ did not promote killing of B cells by neutrophils, its ability to mediate trogocytosis was considerably less than that of RTX. To our knowledge, there have been no investigations as to whether OBZ mediates trogocytosis of CLL cells when it is used in the clinic, and this question should be addressed. Moreover, laboratory studies that compare RTX or OBZ with respect to trogocytosis mediated by other effector cells, including monocytes, macrophages, and NK cells, may be most informative.

Valgardsdottir et al have rigorously asked and answered an important question with respect to the role of neutrophils in mediating trogocytosis of CD20 mAb–opsonized cells. Their findings may have important clinical implications and set the stage for future intriguing studies both in the laboratory and in the clinic.