Targeting Human-Cytomegalovirus-Infected Cells by Redirecting T Cells Using an Anti-CD3/Anti-Glycoprotein B Bispecific Antibody

Abstract

The host immune response to human cytomegalovirus (HCMV) is effective against HCMV reactivation from latency, though not sufficient to clear the virus. T cells are primarily responsible for the control of viral reactivation. When the host immune system is compromised, as in transplant recipients with immunosuppression, HCMV reactivation and progressive infection can cause serious morbidity and mortality. Adoptive T cell therapy is effective for the control of HCMV infection in transplant recipients. However, it is a highly personalized therapeutic regimen and is difficult to implement in routine clinical practice. In this study, we explored a bispecific-antibody strategy to direct non-HCMV-specific T cells to recognize and exert effector functions against HCMV-infected cells. Using a knobs-into-holes strategy, we constructed a bispecific antibody in which one arm is specific for CD3 and can trigger T cell activation, while the other arm, specific for HCMV glycoprotein B (gB), recognizes and marks HCMV-infected cells based on the expression of viral gB on their surfaces. We showed that this bispecific antibody was able to redirect T cells with specificity for HCMV-infected cells in vitro In the presence of HCMV infection, the engineered antibody was able to activate T cells with no HCMV specificity for cytokine production, proliferation, and the expression of phenotype markers unique to T cell activation. These results suggested the potential of engineered bispecific antibodies, such as the construct described here, as prophylactic or therapeutic agents against HCMV reactivation and infection.

Keywords: HCMV; T cell activation; T cell immunity; bispecific antibody; therapeutic agent.

FIG 1

Humanization of a rabbit HCMV gB-specific antibody and detection of gB expression on the surfaces of HCMV-infected cells. (A) Sequence alignment of the closest human germ lines (IGHV3-53*04), rabbit antibody 272.7, and the humanized antibody (hu272.7). The combined CDRs determined are boxed. Antibody humanization was performed by CDR grafting. (B) The humanized antibody maintained affinity and specificity for gB. The rabbit 272.7 and hu272.7 antibodies in titration were tested for binding to gB protein by ELISA. EC₅₀s were deduced from four-parameter curve fitting. The statistical significance of differences between the rabbit 272.7 and hu272.7 antibodies was analyzed by two-way ANOVA. n.s., not significant (P > 0.05). (C) Detection of gB expression on the surfaces of HCMV-infected ARPE-19 cells by a flow cytometry assay. The mean fluorescence intensities ± SD of gB-specific signals from triplicate samples are shown. The data are representative results from two independent experiments. Statistical significance was determined by the unpaired two-tailed t test. **, P < 0.01; ***, P < 0.001.

FIG 2

Design and characterization of the bispecific antibody (BsAb). (A) Schematic depiction of the bispecific antibody with gB and CD3 specificities. (Gly₄Ser)₃ linkers were constructed between the V_H and V_L domains of their respective scFvs. (B) Bispecific-antibody expression in HEK293F cells. Single-chain controls and the bispecific antibody were expressed by transfection of the scFv-K-Fc or scFv-H-Fc construct, or both, and were affinity purified with protein A. The constructs were then evaluated by SDS-PAGE under either reducing conditions (lanes 1 to 3) or nonreducing conditions (lanes 4 to 6). Lanes 1 and 4, single-chain hu272.7 scFv-K proteins; lanes 2 and 5, OKT3 scFv-H; lanes 3 and 6, the bispecific antibody. K, knob monomer; H, hole monomer; K/H, knob or hole mixed monomer; KK, knob homodimer; HH, hole homodimer; KH, knob-into-hole heterodimer. The densitometry values of the bands were analyzed by ImageJ software and are shown below the bands in lanes 4 to 6. The percentages of dimers were calculated and are shown below the gel. (C) Specificity of the BsAb for immobilized gB as determined by ELISA. Statistical significance was determined by two-way ANOVA. (D) Testing of the specificities of the BsAb (red) and the hu272.7 antibody (green) for T lymphocytes by flow cytometry. The binding signal is detected by a PE-conjugated anti-human Fc antibody.

FIG 3

The bispecific antibody activates T cells, as shown by the proliferation and secretion of Th1-type cytokines. (A) Dual specificity of the bispecific antibody. Biotinylated gB protein was incubated with CD4 or CD8 T cells pretreated with the bispecific (red) or hu272.7 (green) antibody. PE-conjugated streptavidin was used to detect the binding signal as an indication of dual specificity. (B) Antigen-specific proliferation of T cells mediated by the bispecific antibody. T cells were prestained with CFSE and were then incubated with the bispecific antibody in culture plates coated with recombinant gB protein (left) or BSA (right) for 2 days. The proliferation of T cells in response to the antibody treatment was then determined by measuring the CFSE signal by flow cytometry. (C) Activation of T cells as shown by cytokine production. T cells were incubated with HCMV-infected (circles) or uninfected (squares) ARPE-19 cells in the presence of the bispecific antibody in titration. The supernatants were collected 2 days later for the measurement of TNF-α and IFN-γ production. Statistical significance was determined by two-way ANOVA. (D to F) Two days after the incubation of HCMV-infected (red) or uninfected (green) ARPE-19 cells with the BsAb, T cells were analyzed by flow cytometry for activation markers CD25 (D) and CD69 (E) and for degranulation marker CD107a (F).

FIG 4

Light microscopy of T cells incubated with HCMV-infected or uninfected ARPE-19 cells in the presence of the bispecific antibody. T cells and the bispecific antibody were added to a 6-well plate that was seeded with HCMV-infected (as indicated by GFP) or uninfected ARPE-19 cells. Fluorescence microscopy images were taken 48 h later. (A′ and B′) T cells migrated toward HCMV-infected ARPE-19 cells. (D′) T cells were evenly dispersed in the wells with uninfected ARPE-19 cells. Magnification, ×10 for panels A, B, C, D, and E; ×40 for panels A′, B′, C′, and D′.

FIG 5

The bispecific antibody mediated T-cell-dependent inhibition of cell growth. (A) Bispecific antibodies in titration were incubated with T cells and HCMV-infected ARPE-19 cells. The electric impedance was monitored in real time for 60 h in order to map the cell growth curve, and the cell index curve was normalized to the growth of the control (HCMV-infected ARPE-19 cells, with no T cells). (B) The percentage of inhibition of the HCMV-infected cell varies proportionally with the antibody concentration. The percentage of T-cell-dependent inhibition of cell growth in response to the antibody was calculated using the following formula: [(cell index with T cells only) − (cell index of T-cell-plus-antibody treatment)]/(cell index with T cells only) × 100%.