High neutralizing potency of swine glyco-humanized polyclonal antibodies against SARS-CoV-2

Abstract

Heterologous polyclonal antibodies might represent an alternative to the use of convalescent plasma or monoclonal antibodies (mAbs) in coronavirus disease (COVID-19) by targeting multiple antigen epitopes. However, heterologous antibodies trigger human natural xenogeneic antibody responses particularly directed against animal-type carbohydrates, mainly the N-glycolyl form of the neuraminic acid (Neu5Gc) and the α1,3-galactose, potentially leading to serum sickness or allergy. Here, we immunized cytidine monophosphate-N-acetylneuraminic acid hydroxylase and α1,3-galactosyl-transferase (GGTA1) double KO pigs with the Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike receptor binding domain to produce glyco-humanized polyclonal neutralizing antibodies lacking Neu5Gc and α1,3-galactose epitopes. Animals rapidly developed a hyperimmune response with anti-SARS-CoV-2 end-titers binding dilutions over one to a million and end-titers neutralizing dilutions of 1:10 000. The IgG fraction purified and formulated following clinical Good Manufacturing Practices, named XAV-19, neutralized spike/angiotensin converting enzyme-2 interaction at a concentration <1 μg/mL, and inhibited infection of human cells by SARS-CoV-2 in cytopathic assays. We also found that pig GH-pAb Fc domains fail to interact with human Fc receptors, thereby avoiding macrophage-dependent exacerbated inflammatory responses and a possible antibody-dependent enhancement. These data and the accumulating safety advantages of using GH-pAbs in humans warrant clinical assessment of XAV-19 against COVID-19.

Keywords: COVID-19; SARS-CoV-2; pig; polyclonal antibodies; spike.

Figure 1

RP‐UPLC chromatograms of pig IgG and the clinical batch of glyco‐humanized pig IgG showing absence of Neu5Gc residues on glyco‐humanized pig IgG. Sialic acids quantification was performed by DMB‐labeled released sialic acid analysis by RP UPLC‐FLD/ESI‐MS. Labeled monosaccharides were separated by UPLC using a BEH C18 column and the quantification was performed with a fluorescent detector coupled to the chromatography. The procedure was performed following standard procedures from Waters (Application Note UPLC/FLR/QT of MS Analysis of Procainamide‐Labeled N‐glycans). Panel A: WT pig IgG with Neu5GC (left filled histogram) and Neu5AC (right filled histogram) peaks. Panel B: Glyco‐humanized pig IgG with one Neu5AC peak. Filled histograms height denotes the relative amount of each N‐glycan in the sample. Data are from a single analysis from one representative sample of each condition.

Figure 2

Reactivity of human IgG against rabbit, pig, and CMAH/GGTA1 DKO pig IgG. Serum from healthy volunteers (n = 59; 100‐fold dilutions) was incubated on ELISA plates coated with WT pig IgG, GGTA1 KO pig IgG, CMAH/GGTA1 DKO pig IgG, or rabbit IgG. Binding of human IgG was revealed with a labeled secondary antibody against human IgG. Symbols represent individual sera, and error bar represents mean ± SEM. Data are from three independent experiments. Nonparametric statistics were used for pairwise comparisons using Kruskal–Wallis tests. NS: not significant. ***p < 0.001.

Figure 3

High complement activation with Neu5Gc and αGal‐negative pig IgG. GGTA1/CMAH double KO pigs (n = 4) were immunized with human T cells and the hyperimmune IgG fraction purified from pooled sera by Protein A chromatography. Rabbit anti‐human T cells used here was Thymoglobulin. The two IgG preparations were differentially diluted in nonimmune IgG of the respective species to obtain similar binding titer and intensity on target human T cells by flow cytometry. A fluorescent Protein‐G reagent has been used instead of species‐specific secondary antibodies to ensure similar, reagent‐independent revelation (A). Target human T cells were used in a complement‐mediated cytotoxicity (CDC) assay where rabbit IgG and GGTA1/CMAH double KO pig IgG directed against human T cells were compared (B). Results shown here are one representative experiment of three independent assays.

Figure 4

SARS‐CoV‐2 spike binding ELISA. GGTA1/CMAH double KO pigs were immunized with SARS‐CoV‐2 spike RBD recombinant proteins and blood was collected after two (A and C) or three to five (B and D) immunizations. Serum (A and B) and the IgG fraction (C and D) were tested by ELISA in a single experiment for binding to recombinant SARS‐CoV‐2 spike molecules. In (A) (n = 10), (B) (n = 12), and D (n = 4), curves represent data from individual animals (one animal per experiment). In (C), IgGs were extracted from a serum pool from animals depicted in (A) (means of triplicates ± SD, assessed by ELISA in a single experiment). In (D), the orange curve corresponds to one of the best responding animals and the red one to the pool of sera selected for further GMP processing and clinical usage (n = 10).

Figure 5

Assessment of inhibition of SARS‐CoV‐2 spike/ACE‐2 interaction. Samples analyzed in Fig. 4 were tested in an ELISA interaction assay where ACE‐2 was immobilized on plastic and Spike‐Fc ligand binding to ACE‐2 was revealed with a secondary antibody against Fc. 100% inhibition represents absence of spike/ACE‐2 interaction. (A) Mean ± SD of four individual sera collected after two immunizations, assessed in a single ELISA. (B) Mean (n = 4) ± SD of individual IgG fractions shown in (A). Black square symbols: IgG extracted from a pool of pre‐immune serum (n = 4). (C) Triangles represent data from a serum pool from six animals obtained after three to five immunizations. Square symbols: nonimmune serum. (D) Individual profiles (means of duplicate measures for each point) of six individual IgG fractions from the six animals shown in (C).