Antibody-dependent-cellular-cytotoxicity-inducing antibodies significantly affect the post-exposure treatment of Ebola virus infection

Abstract

Passive immunotherapy with monoclonal antibodies (mAbs) is an efficacious treatment for Ebola virus (EBOV) infections in animal models and humans. Understanding what constitutes a protective response is critical for the development of novel therapeutic strategies. We generated an EBOV-glycoprotein-pseudotyped Human immunodeficiency virus to develop sensitive neutralizing and antibody-dependent cellular cytotoxicity (ADCC) assays as well as a bioluminescent-imaging-based mouse infection model that does not require biosafety level 4 containment. The in vivo treatment efficiencies of three novel anti-EBOV mAbs at 12 h post-infection correlated with their in vitro anti-EBOV ADCC activities, without neutralizing activity. When they were treated with these mAbs, natural killer cell (NK)-deficient mice had lower viral clearance than WT mice, indicating that the anti-EBOV mechanism of the ADCC activity of these mAbs is predominantly mediated by NK cells. One potent anti-EBOV mAb (M318) displayed unprecedented neutralizing and ADCC activities (neutralization IC₅₀, 0.018 μg/ml; ADCC EC₅₀, 0.095 μg/ml). These results have important implications for the efficacy of antiviral drugs and vaccines as well as for pathogenicity studies of EBOV.

Figure 1. pHIV–ZGP–Fluc construction and cell sensitivity.

(A) Procedural flow chart of pHIV–ZGP–Fluc construction. 293T cells were co-transfected with pEBOV-ZGP and pSG3.Δenv.cmv.Fluc. After 48 h, pHIV–ZGP–Fluc was collected from the supernatant, concentrated, purified, and used to infect the target cells. pHIV–ZGP–Fluc-infected cells were incubated for 48 h before the luciferase activity assay. (B) Optimization of the HIV framework plasmid. pEBOV-ZGP and different framework plasmids were co-transfected to generate pHIV–ZGP–Fluc. (C) Optimization of the GP expression vector. pSG3.Δenv.cmv.Fluc and different expression vectors (EBOV-ZGP-7A or EBOV-ZGP-8A) were co-transfected to generate pHIV–ZGP–Fluc. (D) Optimization of transfection reagents. pCDNA3.1–EBOV-ZGP-8A and pSG3.Δenv.cmv.Fluc were co-transfected with different transfection reagents. (E) Optimization of the proportion of plasmids. Different proportions of pCDNA3.1–EBOV-ZGP-8A and pSG3.Δenv.cmv.Fluc were tested. (F) Cell tropism of pHIV–ZGP–Fluc. Different cell lines were infected with pHIV–ZGP–Fluc. The relative light units of the infected cells were measured.

Figure 2. Construction and identification of a mouse model of pHIV–ZGP–Fluc infection.

(A) Four-week-old and 8-week-old female KM, NIH, BALB/c, and C57BL/6 mice were inoculated with pHIV–ZGP–Fluc by IP injection (1 × 10⁷ TCID₅₀/mouse). The relative levels of bioluminescence are shown in pseudocolours, with red and blue representing the strongest and weakest photon fluxes, respectively. Values of total flux for each group at 4 dpi are shown on the right. (B) BALB/c mice were inoculated with pHIV–ZGP–Fluc and monitored for the duration of bioluminescence. Bioluminescent images were superimposed on grey-scale photographs at 6 h, 2 days, 4 days, 6 days, 10 days, and 14 days post-infection. Values for total flux at different time points are shown on the right. Each data point is a mean value (n = 5). (C) Photon fluxes in infected BALB/c mice and their dissected tissues were measured at 4 dpi. Different organs and tissues shown are 1) thymus, 2) heart, 3) liver, 4) spleen, 5) kidney, 6) lung, 7) lymph node, 8) muscle, 9) skin, 10) ovary, 11) brain, and 12) intestine. (D) Correlation of pHIV–ZGP–Fluc viral loads in various tissues with bioluminescence intensity at 4 dpi. (E) Histopathological analysis of heart, liver, spleen, lung, kidney, thymus, and muscle at 7 dpi. Paraffin-fixed tissue sections were stained with haematoxylin and eosin. Arrows indicate lesion sites. Scale bar, 20 μm.

Figure 3. In vitro neutralizing and ADCC assays of human and murine anti-EBOV mAbs.

(A) Procedural flow chart of the pHIV–ZGP–Fluc ADCC assay. (B) EBOV antigen presented on the cell surface was measured with flow cytometry. pHIV–ZGP–Fluc and pHIV (strain SF162) viruses were used to infect CEM cells. The non-virus group was used as a control, and antibody M401-FITC was used to detect the EBOV antigen. (C) The M401-FITC binding activity at different time points (2, 4, 8, 12, and 16 hpi) was measured as in (B). (D–F) Optimization of the pseudovirus-based neutralization assay. (D) Optimization of HEK293T cell density. The x-axis indicates different numbers of HEK293T cells. Dots show relative light units after incubation for 48 h. (E) Optimization of incubation time before detection. The x-axis indicates the incubation time, and the y-axis indicates the relative light units. (F) Dose–response between pHIV–ZGP–Fluc input and IC₅₀ values of mAbs. Serial dilutions of mAbs directed against EBOV GP were pre-incubated with pHIV–ZGP–Fluc, and the percentage inhibition and R² were determined. (G,H) In vitro neutralization activities (G) and ADCC activities (H) of the MIL77 monoclonal antibodies, MIL77–1, MIL77–2, and MIL77–3. Calculated IC₅₀ and EC₅₀ values are shown next to each curve. (I–L) In vitro neutralization and ADCC activities of four anti-EBOV-GP mAbs, M001 (I), M401 (J), M318 (K), and M501 (L). Calculated IC₅₀ and EC₅₀ values are shown next to each curve.

Figure 4. In vivo prevention and treatment efficacy assays.

(A) Four murine anti-EBOV mAbs, M001, M401, M318, and M501, were injected to evaluate their prevention efficacies (upper) and treatment efficacies (lower). The mAbs were injected at 3 days or 4 h before the inoculation of pHIV–ZGP–Fluc to determine their preventive efficacies, and at 12 h or 3 days after viral inoculation to determine their treatment efficacies. Photon flux was measured at 4 dpi. The first column shows the control pHIV–ZGP–Fluc-infected BALB/c mice that did not receive any mAb treatment. (B–C) Preventive efficacy (B) and treatment efficacy (C) were detected as the total photon flux in pHIV–ZGP–Fluc-infected mice treated with various mAbs compared with that in similarly infected mice without mAb treatment (first column). *p < 0.05; **p < 0.01. (D–I) ELISA binding curves for recombinant GP proteins, including recombinant GP (D) and its receptor-binding domain (E) from EBOV strain H.sapiens-wt/GIN/2014/Kissidougou-C15 and recombinant GP (F), receptor-binding domain (G), GP2 (H), and GP1 (I) from EBOV strain Mayinga 1976.

Figure 5. Identification of NK-cell-mediated ADCC against pHIV–ZGP–Fluc infection.

(A) NK inhibitor, mAbs, and pHIV–ZGP–Fluc virus were injected as shown in the procedural flow chart. BALB/c mice were treated with the NK inhibitor at baseline and 5 days. The mice were infected with pHIV–ZGP–Fluc at 3 days and treated with mAbs at 12 h and 2 dpi. Bioluminescent images of BALB/c mice were analysed at 4 dpi with the IVIS system. (B) Percentage of NK cells in mouse blood. Flow cytometry was performed at 0, 2, 3, and 7 days to confirm the deletion of NK cells. mCD3⁻mNK1.1⁺ cells were recognized as NK cells, and the percentage of NK cells was calculated as shown. (C) Viral clearance rates mediated by mAbs M401 and M501 were evaluated after NK deletion compared with those in normal BALB/c mice. (D) Immunohistochemical staining of pHIV–ZGP–Fluc in thymus and spleen. Sections of BALB/c mice (naïve), pHIV–ZGP–Fluc-infected BALB/c mice treated with M401 (pHIV–ZGP–Fluc + mAb), and NK-deleted mice with the same treatment (pHIV–ZGP–Fluc + NK⁻ + mAb) were stained with monoclonal mouse anti-EBOV antibody. Scale bar, 20 μm.