Anti-severe acute respiratory syndrome coronavirus spike antibodies trigger infection of human immune cells via a pH- and cysteine protease-independent FcγR pathway

Abstract

Public health measures successfully contained outbreaks of the severe acute respiratory syndrome coronavirus (SARS-CoV) infection. However, the precursor of the SARS-CoV remains in its natural bat reservoir, and reemergence of a human-adapted SARS-like coronavirus remains a plausible public health concern. Vaccination is a major strategy for containing resurgence of SARS in humans, and a number of vaccine candidates have been tested in experimental animal models. We previously reported that antibody elicited by a SARS-CoV vaccine candidate based on recombinant full-length Spike-protein trimers potentiated infection of human B cell lines despite eliciting in vivo a neutralizing and protective immune response in rodents. These observations prompted us to investigate the mechanisms underlying antibody-dependent enhancement (ADE) of SARS-CoV infection in vitro. We demonstrate here that anti-Spike immune serum, while inhibiting viral entry in a permissive cell line, potentiated infection of immune cells by SARS-CoV Spike-pseudotyped lentiviral particles, as well as replication-competent SARS coronavirus. Antibody-mediated infection was dependent on Fcγ receptor II but did not use the endosomal/lysosomal pathway utilized by angiotensin I converting enzyme 2 (ACE2), the accepted receptor for SARS-CoV. This suggests that ADE of SARS-CoV utilizes a novel cell entry mechanism into immune cells. Different SARS vaccine candidates elicit sera that differ in their capacity to induce ADE in immune cells despite their comparable potency to neutralize infection in ACE2-bearing cells. Our results suggest a novel mechanism by which SARS-CoV can enter target cells and illustrate the potential pitfalls associated with immunization against it. These findings should prompt further investigations into SARS pathogenesis.

Figures

Fig. 1.

Susceptibility of hematopoietic cell lines to infection by SARS-CoV Spike pseudoparticles (SARS-CoVpp). (A) SARS-CoVpp were incubated in the presence or absence of different dilutions (1/1,000, 1/2,000, and 1/4,000) of either control (solid bars) or anti-Spike (hatched bars) serum for 1 h prior to addition to the cells. At 3 days postinfection, luciferase substrate reagent was added to wells, and the luminescence was measured. The data were normalized to control conditions, viz., cells incubated with SARS-CoVpp in the absence of any serum (taken as unity), and are expressed as the fold changes in luminescence. Because the results for SARS-CoVpp with or without control serum were virtually identical with all cell lines, for the sake of clarity only one dilution (1/4,000) is shown. The results are the means ± the SD of six measurements from two independent experiments. Statistical significance was assessed by comparing the appropriate dilutions of control and anti-Spike serum (‡, P < 0.05; *, P < 0.001 [unpaired Student t test]). (B to D) HIV Gag-normalized lentiviral particles (0.1 ng of p24 protein/μl) pseudotyped with the envelope glycoprotein of SARS-CoV Spike (SARS-CoVpp [B]) or vesicular stomatitis virus (VSVpp [C]) or lacking any viral envelope protein (Δenv.pp [D]) were incubated in the presence or absence of a 1/1,000 dilution of either control (solid gray bars) or anti-Spike (hatched bars) serum for 1 h prior to addition to the cells. At 3 days postinfection, luciferase substrate reagent was added, and the luminescence was measured. The results are the means ± the SD of nine measurements from three independent experiments. When not visible, the SD values were contained within the size of the symbols. Anti-Spike serum either significantly decreased (VeroE6) or increased (THP-1, Raji, and Daudi) entry of SARS-CoVpp. *, P < 0.001 (unpaired Student t test).

Fig. 2.

Anti-Spike antibodies trigger SARS-CoV infection of human Raji B cells. VeroE6 or human Raji cells were infected (MOI=1) with SARS-CoV strain HK39849 (SCoV) in the presence of either control or Anti-Spike serum for 1 h. Samples were fixed with 4% paraformaldehyde at 15 h postinfection. A mouse monoclonal antibody specific for the SCoV nucleocapsid (N) and rabbit polyclonal antibodies specific for SCoV membrane (M) protein were visualized using TRITC-conjugated goat anti-mouse and FITC-conjugated goat anti-rabbit antibodies, respectively. Cell nuclei were labeled with DAPI, and separate channel images were acquired. (A) As expected, SCoV infection of VeroE6 (I) was completely blocked by anti-Spike serum (1/500, II). (B) In contrast, SCoV viral proteins were detected in Raji cells only when they were incubated with anti-Spike (1/500, IV) but never with control serum (III). Quantification of infection (right-hand portions in panels A and B) was done by counting positive cells in 30 randomly chosen fields (630×, final magnification), using Metamorph as described in Materials and Methods. The results are shown as means ± the SD of the number of fields.

Fig. 3.

Endpoint and real-time PCR detection of SARS-CoV genes in ADE-infected Raji cells. (A) VeroE6 and human Raji cells were incubated for 1 h in the absence (mock) or presence of SARS-CoV strain HK39849 (SCoV), with either control or anti-Spike (S) serum as indicated. The cells were lysed at 15 h postinfection, and total RNAs were extracted for RT-PCR amplification as described in Materials and Methods. Amplicons were visualized by ethidium bromide staining after agarose gel electrophoresis. A 1-kb molecular weight standard is shown on the left lane of each gel. (B) From the cDNA samples prepared in panel A, the levels of SCoV nucleocapsid and SCoV ORF1b (not shown) genes were determined by real-time quantitative RT-PCR. Raji cells were consistently infected in the presence of serial dilutions (1/250 to 1/4,000 in 2-fold dilutions) of anti-Spike serum that, in contrast, inhibited in dose-dependent fashion viral entry into the susceptible VeroE6 cells. The data are shown as the means ± the SD of four measurements from duplicate cultures and are representative of three independent experiments with similar results. **, P < 0.001 (unpaired Student t test).

Fig. 4.

Abortive replication of SARS-CoV in Raji cells following antibody-mediated infection. (A) Human Raji cells were incubated for 1 h in the absence or presence of SARS-CoV strain HK39849 (SCoV), with 2-fold serial dilutions (1/250 to 1/4,000) of either control or anti-Spike (S) serum. Cells were lysed at 6, 24, 48, and 72 hpi, and total RNAs were extracted and reverse transcribed into sense-specific cDNA species as described in Materials and Methods. The levels of positive (+) and negative (−) SCoV nucleocapsid strands were determined by real-time quantitative RT-PCR. For the sake of clarity, only one serum dilution (1/1,000) is shown. At an early time point (6 hpi), a high viral load was consistently detected in cells infected in the presence of anti-Spike but not in control serum. At later times (24 to 72 hpi), no further difference among the treatments was noticeable. The data are shown as means ± the SD of four measurements. When not visible, the SD values were contained within the size of the symbols. **, P < 0.001 (unpaired Student t test). (B) Human Raji cells infected as indicated in panel A were fixed with 4% paraformaldehyde, and costaining for SCoV nucleocapsid and cell nuclei was performed as described in Materials and Methods. Quantification of the infected and uninfected cells was done by counting 10 randomly chosen fields using Metamorph software. For the sake of clarity, results for only one serum dilution (1/1,000) are shown. Cell numbers and related percentage of infection are shown as means ± the SD of the number of observations. SCoV infection of Raji cells occurred only in the presence of anti-Spike serum, but the percentage of infected cells did not significantly increase over time.

Fig. 5.

Blockade of FcγRII abrogates antibody-mediated infection of human immune cell lines. (A) Total RNAs from the indicated cell lines were extracted for RT-PCR amplification as described under Materials and Methods. cDNAs and negative RT controls (RT−) were probed for detection of FcεR1γ-chain, FcγRIA (CD64a), FcγRIIA (CD32a), FcγRIIB (CD32b), FcγRIIIA (CD16a), ACE2, and GAPDH. DNA plasmids containing the coding sequence of the human Fc receptors or the human ACE2 (104 and 106 copies) were used as positive controls. Amplicons were visualized by ethidium bromide staining after agarose gel electrophoresis. For reference, a 1-kb molecular-weight standard is also shown. Due to the murine origin of the macrophage-like P388D1 and J774A.1 cell lines and because the primers were designed for the detection of human FcR (Table 2), no FcγR transcripts were detected in P388D1 and J774A.1 cells. (B) Hematopoietic cell lines displaying a detectable level of FcγR transcript in panel A were subjected to flow cytometry. Cell lines were stained with either isotype-matched control antibody (gray-filled histograms) or FcγR-specific mouse monoclonal antibody (open histograms) as indicated, followed by secondary FITC-conjugated goat anti-mouse antibodies. Labeled samples were then read by flow cytometry and post-acquisition analysis was performed using FlowJo software. Insets show the median of fluorescence intensity of the anti-FcγR staining. When not visible, the open histograms were superimposed on the gray-filled ones. The data are representative of two independent experiments with similar results. (C) Effect of FcγR blockage on susceptibility of the cells to ADE infection. Prior to infection, cells were treated with anti-FcγR blocking antibodies. At 3 days postinfection, luciferase substrate reagent was added to wells, and the luminescence was measured. The data were normalized to control conditions, viz., cells infected in the presence of IgG isotype-matched control antibody, and expressed as a percentage of the changes in luminescence. For all conditions, values of background luminescence (never exceeding 3% of control) were subtracted prior to normalization. The results are the means ± the SD of nine measurements from three independent experiments. ‡, P < 0.05; *, P < 0.001 (unpaired Student t test).

Fig. 6.

Human FcγRIIA and B1, but not huFcγRIA nor huFcγRIIIA, are responsible for ADE of SARS-CoV pseudoparticles (SARS-CoVpp) infection. (A) Human Raji, parental, and monoclonal FcγR-transduced human ST486 B cell lines were stained with either isotype-matched control antibody (gray-tinted histogram) or FcγR-specific mouse monoclonal antibody (open histogram) as indicated, followed by secondary FITC-conjugated goat anti-mouse antibodies. Labeled samples were then read by flow cytometry and postacquisition analysis was performed using FlowJo software. The inset shows the median of fluorescence intensity of the anti-FcγR staining. When not visible, the open histograms were superimposed on the gray-filled ones. (B) Potency of FcγR(s) to trigger antibody-mediated entry of SARS-CoVpp was assessed by using ST486 cell lines stably expressing a single FcγR. SARS-CoVpp were incubated in the presence or absence of different dilutions (1/1,000, 1/2,000, and 1/4,000) of either control (solid bars) or anti-Spike (hatched bars) serum for 1 h prior to addition to the cells. At 3 days postinfection, luciferase substrate reagent was added to wells, and the luminescence was measured. The data were normalized to control conditions, viz., cells incubated with SARS-CoVpp in the absence of any serum (black bar, taken as unity) and expressed as fold changes in luminescence. Because results for SARS-CoVpp ± control serum were virtually identical with all cell lines, for the sake of clarity only one dilution (1/4,000) is shown. The results are the means ± the SD of nine measurements from three independent experiments. Statistical significance was assessed by comparing the appropriate dilutions of mock and anti-Spike serum. ‡, P < 0.05; *, P < 0.01 (unpaired Student t test).

Fig. 7.

Antibody-mediated infection is independent of acidic pH and cysteine-protease activity of the endosomal/lysosomal compartment. (A, C, and D) pH and protease requirements for ACE2- versus antibody-mediated entry of SARS-CoVpp were investigated with different pharmacological treatments. Prior to infection, VeroE6 and Raji cells were preincubated with the indicated concentrations of ammonium chloride (NH4Cl) (A) for 1 h or the cysteine protease inhibitor E-64d (C) and cathepsin L inhibitor III (D) for 3 h. SARS-CoVpp, with or without anti-Spike serum (1/2,000), were then added to the cells in the continuous presence of the drugs. At 3 days postinfection, luciferase substrate reagent was added to wells, and the luminescence was measured. The data were normalized to the appropriate control conditions (taken as unity) with SARS-CoVpp with or without anti-Spike serum for Raji and VeroE6 cells, respectively. The results are shown as means ± the SD of nine measurements from three independent experiments. Statistical differences between VeroE6 and Raji cells were assessed by the unpaired Student t test. **, P < 0.005. (B) Cell surface expression of FcγRII on Raji cells after NH4Cl treatment. Cells were incubated with the indicated concentrations of the drug for 1, 3, and 5 h, labeled with anti-huFcγRII antibody, and subjected to flow cytometry. The data were normalized to the mean fluorescence intensity of controls (no drug). The results are shown as means ± the SD of six measurements from two independent experiments. When not visible, the SD values were contained within the size of the symbols.

Fig. 8.

Occurrence of ADE of SARS-CoV infection after immunization with different vaccine candidates. (A) Schematic illustration of the different forms of recombinant SARS-CoV Spike protein used for vaccination of BALB/c mice. Positions of the different regions are indicated according to Swiss-Prot accession no. P59594. (B and C) Neutralizing and enhancing abilities of antisera elicited after immunization with different SARS vaccine candidates. SARS-CoVpp were incubated in the presence of a 2-fold serial dilution (1/1,1000 to 1/16,000) of either control or anti-Spike serum for 1 h prior to addition to VeroE6 cells (B) (neutralization assay) or Raji cells (C) (ADE assay). At 3 days postinfection, luciferase substrate reagent was added to the wells, and the luminescence was measured. Five independent experiments were performed in triplicates, and data are presented as the means ± the SD of a representative experiment. (D) IgG composition of the immune serum of mice vaccinated with the indicated immunogens. The binding activity toward immobilized recombinant SARS-CoV Spike protein of pooled serum (1/2,000 dilution) was measured by ELISA, and bound IgGs were revealed with specific polyclonal secondary antibodies. A representative result from three independent experiments is shown. The results are expressed as means ± the SD of triplicate measurements. When not visible, the SD values were contained within the size of the symbols.