The binding of an anti-PD-1 antibody to FcγRΙ has a profound impact on its biological functions

Abstract

Antibodies targeting PD-1 have been demonstrated durable anti-cancer activity in certain cancer types. However, the anti-PD-1 antibodies are less or not efficacious in many situations, which might be attributed to co-expression of multiple inhibitory receptors or presence of immunosuppressive cells in the tumor microenvironment. Most of the anti-PD-1 antibodies used in clinical studies are of IgG4 isotype with the S228P mutation (IgG4S228P). The functional impact by the interaction of anti-PD-1 IgG4S228P antibody with Fc gamma receptors (FcγRs) is poorly understood. To assess the effects, we generated a pair of anti-PD-1 antibodies: BGB-A317/IgG4S228P and BGB-A317/IgG4-variant (abbreviated as BGB-A317), with the same variable regions but two different IgG4 Fc-hinge sequences. There was no significant difference between these two antibodies in binding to PD-1. However, BGB-A317/IgG4S228P binds to human FcγRI with high affinity and mediates crosslinking between PD-1 and FcγRI. In contrast, BGB-A317 does neither. Further cell-based assays showed that such crosslinking could reverse the function of an anti-PD-1 antibody from blocking to activating. More importantly, the crosslinking induces FcγRI+ macrophages to phagocytose PD-1+ T cells. In a mouse model transplanted with allogeneic human cancer cells and PBMCs, BGB-A317 showed significant tumor growth inhibition, whereas BGB-A317/IgG4S228P had no such inhibition. Immunohistochemistry study revealed an inverse correlation between FcγRI+ murine macrophage infiltration and the density of CD8+PD-1+ human T cells within tumors in the BGB-A317/IgG4S228P-treated group. These evidences suggested that FcγRI+ binding and crosslinking had negative impact on the anti-PD-1 antibody-mediated anti-cancer activity.

Keywords: Antibody; Cancer therapy; FcγRI; Macrophages; PD-1.

Conflict of interest statement

Conflict of interest

All authors have ownership interest in BeiGene. Tong Zhang, Lanlan Xu, Qi Liu., Jing Song and Kang Li are inventors on a patent covering BGB-A317 described in this study.

Research involving human participants and/or animals

All procedures performed in studies involving human participants were in accordance with the ethical standards of BeiGene and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. All procedures performed in studies involving animals were in accordance with the ethical standards of BeiGene where the studies were conducted.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Figures

Fig. 1

Comparison of BGB-A317 and BGB-A317/IgG4S228P binding profiles to human PD-1, FcγRI, and FcγRIIIA. a, b Real-time SPR sensorgrams of BGB-A317 and BGB-A317/IgG4S228P binding to human PD-1 assayed using BIAcore. Y-axis, response unit (RU). X-axis, reaction time course, seconds. c Binding of BGB-A317 and BGB-A317/IgG4S228P to HuT78/PD-1 cells assayed by FACS. Mean fluorescence intensity (MFI) was determined by flow cytometry. d Differential binding of BGB-A317 and BGB-A317/IgG4S228P to FcγRI assayed by FACS using HEK293/FcγRI cells. e Both BGB-A317 and BGB-A317/IgG4S228P have low binding activity for FcγRIIIA, determined by ELISA. The anti-PD-1 antibodies were incubated with secondary antibodies to form immune complexes before adding to FcγRIIIA-coated 96-well plates. Bound immune complexes were detected with a chemiluminescence substrate. HuIgG was used as a positive control

Fig. 2

BGB-A317/IgG4S228P mediates crosslinking between PD-1 and FcγRI receptors determined by biochemical and cell signaling assays. a Overlaid sensorgrams of BGB-A317 and BGB-A317/IgG4S228P binding to PD-1 followed by binding with FcγRI. The PD-1 antibody (either BGB-A317 or BGB-A317/IgG4S228P) was injected onto the surface of human PD-1/His-coated CM5 chip, followed by FcγRI association and dissociation. The arrows indicated the time points when the analytes were injected. b Cartoon interpretation of the SPR assay in (a) showing that BGB-A317/IgG4S228P mediates crosslinking between PD-1 and FcγRI, while BGB-A317 does not. c Comparison of PD-1 antibody functional activities by P3Z assay. In the assays, detection of IL-2 secretion served as a quantitative indicator of PD-1 signaling. The dose–response was assayed either in the “two-cell line” co-cultures (HuT78/P3Z and HEK293/PD-L1) presented by solid symbols or in the “three-cell line” system (HuT78/P3Z, HEK293/PD-L1 and FcγRI+ THP-1 cells) presented by open symbols. Ab conc. Antibody concentration. d Rabbit Anti-CD64 (FcγRIα) polyclonal antibody (pAb) restored the inhibitory effect of BGB-A317/IgG4S228P on PD-1-signaling in the “three-cell” co-culture system. THP-1 cells were pre-treated with the rabbit Anti-CD64 pAb before co-culturing with HuT78/P3Z and HEK293/PD-L1 cells in the presence of BGB-A317/IgG4S228P. The placebo contained antibody buffer solution only

Fig. 3

BGB-A317/IgG4S228P induces primary M2 macrophages to phagocytose PD-1+ T cells (ADCP) and to express IL-10 gene. a CD64 (FcγRI) and CD32 (FcγRII) expression on the in vitro differentiated M2 macrophages as determined by FACS. b ADCP assay using M2 macrophages. Primary M2 macrophages were co-cultured with CFSE-labeled PD-1+ HuT78/PD-1 cells overnight in the presence of the indicated antibodies. HuIgG was used as negative control. Representative dot plots of three independent experiments are shown. M2 Mac: M2 macrophage. c Bar graphs summarized the results of three independent experiments. The % of ADCP was determined as described in the “Materials and methods”. The mean + SD represents triplicate data points. ##P < 0.01, comparing BGB-A317/IgG4S228P versus huIgG. **P < 0.01, comparing BGB-A317/IgG4S228P versus BGB-A317. d Rabbit anti-CD64 polyclonal antibody neutralizes FcγRI-mediated ADCP. Rabbit IgG and placebo were used as negative controls. e Crosslinking of FcγRI+ M2 macrophages to plate-coated PD-1 by BGB-A317/IgG4S228P induces IL-10 gene expression. M2 macrophages were added to PD-1-coated 96-well plates in the presence of anti-PD-1 Abs and cultured overnight. The IL-10 gene expression was assayed by real-time PCR. IL-10 expression in huIgG-treated M2 macrophages was set as a baseline. The mRNA levels of placebo (antibody buffer solution), BGB-A317 or BGB-A317/IgG4S228P-treated M2 macrophages were normalized against the baseline. The results from three independent experiments are shown as mean + SD of duplicate data points. HuIgG is a mixture of human IgG1, IgG2, IgG3 and IgG4 (Invitrogen)

Fig. 4

Anti-tumor activities of BGB-A317 and BGB-A317/IgG4S228P in an allogenic xenograft model. a The anti-tumor activity of BGB-A317 at doses of 1 and 10 mg/kg, QW, i.p. was assessed in A431 allogenic xenograft model, in which PBMCs from healthy donors and A431 cancer cells were co-injected subcutaneously into NOD/SCID mice. Each treatment group had ten mice. b The anti-tumor activities of BGB-A317 (n = 11) and BGB-A317/IgG4S228P (n = 13) were compared at the same dose of 10 mg/kg, QW and in the same model as in (a). c PK curves of BGB-A317 (1 and 10 mg/kg, n = 3–4) and BGB-A317/IgG4S228P (10 mg/kg, n = 4) with a single dose treatment in NOD/SCID mice is shown. The PK parameters are summarized in Suppl. Table 2

Fig. 5

The effect of anti-PD-1 antibody treatment on tumor-infiltrating T cells. a Representative images of hCD8 and hPD-1 immunofluorescence staining of tumor tissues from BGB-A317- or BGB-A317/IgG4S228P-treated mice. b, c Quantified result of tumor-infiltrating hCD8 and hPD-1 positive cell numbers in each indicated group (n = 5). Ten images were taken for each tumor tissue sample by Vectra. The relative numbers of immunofluorescence-positive cells were quantified using inForm software

Fig. 6

Crosslinking of T cells and FcγRI positive cells in vivo. a Representative low power (40×) images of multiplex immunofluorescence staining of whole mount tumors treated with BGB-A317 or BGB-A317/IgG4S228P. The xenograft tumor model was generated by subcutaneously implanting the allogenic human epidermoid cancer cells (A341 cell line) and PBMCs in NOD/SCID mice. Representative high power images (200×) were shown in (b). The biomarkers are indicated by green (hCD8), red (mCD64) and blue (nucleus), respectively. c–e Quantified results of hCD8, mCD64 staining intensity and co-localization of hCD8 and mCD64 in each indicated group (n = 5). Six to ten images in each tumor sample were captured by Vectra, and the staining intensity and co-localization was quantified using inForm software. (Note: CD64 = FcγRIα)