Antibody-dependent cell cytotoxicity in monoclonal antibody-mediated tumor immunotherapy

Abstract

Antibody-dependent cell cytotoxicity (ADCC) is critical in monoclonal antibody (mAb)-mediated cancer therapy. We recently showed that a tumor-specific mAb in combination with cyclophosphamide inhibited tumor cell growth and induced ADCC-synapses between tumor and effector cells in vivo, opening perspectives to enhance anti-tumor responses by manipulating the immune system.

Monoclonal antibody (mAb) therapy treats certain hematologic and solid tumors.¹ Therapeutic mAbs targeting specifically tumor-associated cell surface antigens currently instigate growing interest and intensive development. However, the mechanisms by which these therapeutic mAbs inhibit tumor growth are still incompletely understood. Most mAbs display direct anti-tumor effects, inducing apoptosis or blocking receptor signaling upon binding to their specific antigen. Moreover, therapeutic mAbs can also destroy tumor cells indirectly by mobilizing the immune system. The role of antibody-dependent cell cytotoxicity (ADCC) in tumor rejection is suggested by the requirement of functional activating Fc receptors for IgG (FcγR) in humans and mice, but this phenomenon has not been directly demonstrated in vivo.^2-5

In a recent paper,⁶ we investigated the role of ADCC in mAb-mediated cancer immunotherapy in a syngenic mouse tumor model. We took advantage of the properties of the Chi-Tn mAb, a chimeric mouse/human IgG1 Fc engineered mAb recognizing the glycopeptidic Tn antigen (N-Acetyl-Galactosamin-O-serin/threonin epitope) at the plasma membrane of a wide variety of epithelial tumor cells, but undetected in normal cells.⁷ The Chi-Tn mAb does not exert direct cytotoxic effects on epithelial tumor cells in vitro or in vivo, which made possible the analysis of the anti-tumor effects of the mAb mediated by the immune system. The Tn-positive mouse breast tumor cell line TA3Ha grafted intraperitoneally (i.p.) was rejected in more than 80% of BALB/c mice after treatment with a single dose of cyclophosphamide (CTX) and two weekly injections of the Chi-Tn mAb (6 total injections at 20 mg/kg). Tumor rejection in vivo was abolished in mice deficient for the Fc receptors-associated γ chain (FcRγ), lacking activating FcγR, indicating that the therapeutic effects observed depend on FcγR and ADCC.

TA3Ha cells grow in suspension in the peritoneal cavity, and could be easily recovered together with cells from the tumor microenvironment. Using three-dimensional deconvolution microscopy, we found that GFP-transfected TA3Ha tumor cells formed stable conjugates with hematopoietic cells present in the peritoneal infiltrate of CTX+Chi-Tn mAb-treated mice but not of mice injected with CTX and a control mAb, as soon as 4 h after mAb injection (Fig. 1A). The effector cells thought to be responsible for ADCC in vivo varies depending on the studies, and macrophages, neutrophils or NK cells could be involved. In our model, tumor cells formed conjugates with macrophages (F4/80⁺), neutrophils (Ly6G⁺), as well as B cells (CD19⁺), at equal frequency.⁶ Interestingly, orthogonal projections revealed that the contact zone between ADCC effectors and tumor cells was organized into multifocal supra-molecular activation clusters (SMACs), similarly to T-cell immunological synapses. Indeed, F-actin formed “actin rings” at the interface between host peritoneal cells and tumor cells, and the activating FcγRIII accumulated within these actin “wholes” (Fig. 1B). The accumulation of tyrosine phosphorylated proteins at the interface and the polarization of LAMP-1+ vesicles toward the contact zone suggested that active signals were transduced, and that these structures represent functional what we called “ADCC-synapses”⁶.

Figure 1:

ADCC synapse formation in mice treated by the Chi-Tn mAb. BALB/c mice were grafted i.p. with TA3Ha-GFP cells (10⁶/mouse) on day 0. PBS or CTX (50 mg/kg) was given on day 1, and Chi-Tn mAb (n = 3) or trastuzumab (n = 3) were injected at 20 mg/kg on day 2. Peritoneal cells were harvested 4 h after mAbs injection. (A) Peritoneal cells from wild-type (WT) mice or from FcR-γ chain-deficient (FcR-γ^−/−) mice treated with the indicated combination were labeled with phalloidin (yellow) and DAPI (blue). Green, TA3Ha-GFP cells. Synapses were observed in all experimental conditions where the Chi-Tn mAb was injected. Tumor rejection occurred in WT mice only when CTX was associated to Chi-Tn mAb treatment. FcR-γ^−/− mice treated with CTX+Chi-Tn did not display tumor rejection. (B) Peritoneal cells were labeled using the indicated primary antibody against FcR-γ chain or FcγRIII (1st Ab, red), and phalloidin (yellow). TA3Ha-GFP cells are in green, and nuclei labeled with DAPI are in blue. Merge images represent close contacts between TA3Ha-GFP+ and immune cells. One representative analysis of several conjugates is shown. Synapses between FcγRIII+ cells and TA3Ha-GFP+ cells were studied by three-dimensional analysis of the interface in the z, x plane. Yellow, phalloidin; red, FcγRIII. One representative experiment is shown.

A critical question was to determine whether ADCC was necessary for tumor cell killing, or if merely cross-linking the Chi-Tn mAb by the FcγR would be sufficient to eliminate the tumor cells. Unexpectedly, synapses were still observed in experimental conditions where no anti-tumor effect was observed (Fig. 1A, lower panels), such as in FcR-γ-deficient mice (lacking activating FcγR) after CTX+Chi-Tn mAb treatment or in wild-type mice treated with Chi-Tn mAb in the absence of CTX. Thus, Chi-Tn mAb cross-linking was not sufficient to kill TA3Ha cells in vivo, accordingly to the results of de Haij et al.,⁴ suggesting that activation of the effector cells is required for the elimination of tumor cells through FcR-γ-chain dependant ADCC.

In conclusion, our study is in accordance with the previously reported studies suggesting that ADCC is involved in mAb-mediated anti-tumor responses,^2-5 and now shows for the first time the occurrence of anti-tumor ADCC synapses in vivo. The most intriguing question yet to be resolved is the necessity to combine CTX with the Chi-Tn mAb to get a potent anti-tumor effect. Synergy of some chemotherapeutics with therapeutic mAbs is now well demonstrated clinically, but the molecular basis of this phenomenon is largely unclear.¹ An explanation would be that CTX could directly potentiate tumor cell death by Tn cross-linking or by ADCC, however, these events were not observed in vitro.⁶ Another attractive hypothesis is that CTX could induce the production of cytokines as previously proposed,⁸ which in turn would prime and/or activate the effectors of the immune system for ADCC. This idea is substantiated by our observations that synapses are not sufficient by themselves in the absence of CTX to induce tumor cell death.⁶ Another interesting point to analyze is the schedule of injection of CTX and the Chi-Tn mAb, which would be critical in modifying the tumor infiltration by immune cells and in eliciting anti-tumor responses.⁹ Several mechanisms by which chemotherapeutics interfere with anti-tumor immune responses have been recently reported.¹⁰ Understanding the molecular mechanisms by which CTX participated to the anti-tumor effect of the Chi-Tn mAb in vivo will open new exciting ways to improve mAb-mediated cancer therapy. Of particular interest will be the characterization of the molecular events induced by CTX and the way by which they stimulate the effectors of ADCC.

Glossary

Abbreviations:

monoclonal antibody

mAb

intraperitoneally

i.p.

antibody-dependent cell cytotoxicity

ADCC

Fc receptors for IgG

FcγR

Fc receptors-associated γ chain

FcRγ

cyclophosphamide

CTX

supra-molecular activation clusters

SMACs