Cross-reactivity of glycan-reactive HIV-1 broadly neutralizing antibodies with parasite glycans

Abstract

The HIV-1 Envelope glycoprotein (Env) is the sole target for broadly neutralizing antibodies (bnAbs). Env is heavily glycosylated with host-derived N-glycans, and many bnAbs bind to, or are dependent upon, Env glycans for neutralization. Although glycan-binding bnAbs are frequently detected in HIV-infected individuals, attempts to elicit them have been unsuccessful because of the poor immunogenicity of Env N-glycans. Here, we report cross-reactivity of glycan-binding bnAbs with self- and non-self N-glycans and glycoprotein antigens from different life-stages of Schistosoma mansoni. Using the IAVI Protocol C HIV infection cohort, we examine the relationship between S. mansoni seropositivity and development of bnAbs targeting glycan-dependent epitopes. We show that the unmutated common ancestor of the N332/V3-specific bnAb lineage PCDN76, isolated from an HIV-infected donor with S. mansoni seropositivity, binds to S. mansoni cercariae while lacking reactivity to gp120. Overall, these results present a strategy for elicitation of glycan-reactive bnAbs which could be exploited in HIV-1 vaccine development.

Keywords: CP: Immunology; HIV-1; Schisotosoma mansoni; glycan epitope; neutralizing antibody; vaccine.

Graphical abstract

Figure 1

Binding of HIV-1 bnAbs to S. mansoni parasite-derived glycans on glycan microarrays (A) Synthetic glycan microarray displaying S. mansoni-derived N-glycans and fragment structures as previously described (Brzezicka et al., 2015; Echeverria et al., 2018). No high-mannose glycan structures are present on this array. Glycans bound by the HIV-1 bnAbs PGT121, PGT151, and PGT123 are displayed. Mammalian “self” and “non-self” glycans are shown in gray and blue, respectively. Binding is reported in relative fluorescence units (RFU). Each histogram represents the average RFU values for four replicates spots and the error bars represent the SD of the mean. A full list of glycan structures present on the synthetic glycan array can be found in Figure S1. (B) Shotgun glycan microarray of S. mansoni-extracted glycans as previously described (De Boer et al., 2007; van Diepen et al., 2012, 2015). Binding of HIV-1 bnAbs PGT121, PGT151, and PGT128 determined. Each line represents the average mean fluorescence intensity (MFI) values for three replicates spots. Glycans bound are displayed on the right and mammalian “self” glycans and non-mammalian “non-self” glycan motifs are indicated. The life stage and type of glycan (N-, O-, or glycosphingolipid [GSL] glycan) are reported. Glycan structures with “s” indicate structures only found on the shotgun glycan microarray. Man, green circle; GlcNAc, blue square; GalNAc, yellow square; Gal, yellow circle; sialic acid, purple diamond; fucose, red triangle; xylose, star.

Figure 2

Binding mode of PGT121 to a xylosylated parasite N-glycan with terminal LacDiNAc and glycan-array binding of PGT121 alanine mutants (A) Model after the solved STD-NMR structure of parasite N-glycan 28 with terminal LacDiNAc occupying the secondary binding pocket of PGT121, which is usually occupied by a bi-antennary complex-type N-glycan. Interactions were determined by STD-NMR (see Figures S2A–S2C and S3; Table S1). The conserved core of the N-glycan is indicated in orange. The part contacted by PGT121 is indicated in blue. (B) Molecular dynamics and superposition of the parasite N-glycan 28 occupying the secondary binding pocket of PGT121. PDB files of representative frames of the MD simulation represented in (B) are found Data S1–S3. (C) Structure of the complex N-glycan 144SIA, bound by the secondary binding pocket of PGT121 (PDB: 4FQC) (Mouquet et al., 2012). The conserved core of the N-glycan is indicated in orange. The part contacted by PGT121 is indicated in blue. (D) Structure of glycan 144SIA bound in the secondary, open face of PGT121 (PDB: 4FQC) (Mouquet et al., 2012) with paratope residues color-coded according to alanine mutants used in (E). (E) Alanine-knockout mutants covering the complex-glycan binding, open face of IgG PGT121 were screened on the synthetic glycan array for parasite glycan recognition. Binding is reported in relative fluorescence units (RFU). Each histogram represents the average RFU values for four replicates spots and the error bars represent the SD of the mean. Man, green circle; GlcNAc, blue square; GalNAc, yellow square; Gal, yellow circle; sialic acid, purple diamond; fucose, red triangle; xylose, star.

Figure 3

Cross-reactivity of HIV-1 bnAbs with S. mansoni (A) Western blot of S. mansoni-soluble cercariae antigen (SCA) and adult worm antigen (AWA) (both at 25 μg per lane) with glycan-binding bnAbs 2G12, PGT128, PGT121, and PGT151 and V3-specific mAb F425-b4e8 (all at 50 μg/mL). A control western blot stained with sera from S. mansoni-infected humans or rabbits immunized with SCA/AWA is shown in Figure S4A. Binding of HIV-1 bnAbs to recombinant gp120 is shown in Figure S4B. (B) Optical density at 450 nm of 2G12, PGT128, PGT121, and PGT151 (at 50 μg/mL) determined by ELISA using S. mansoni-soluble egg antigen (SEA), SCA, and AWA. V3-specific mAb F425-b4e8 and CD4-bnAb VRC01 (at 50 μg/mL), as well as healthy human serum, and S. mansoni-positive human serum as positive control. Error bars represent data from repeated experiments. The horizontal dotted line indicates the cut-off for the positive control provided with the assay kit (see STAR Methods). Bar graphs show the mean of the OD values from at least 3 repeat experiments, and the error bars represent the SD. (C–F) Binding of HIV-1 bnAbs to S. mansoni cercariae. HIV-1 bnAbs binding was detected with Alexa Fluor 488 (green), actin was detected with rhodamine (red), and DNA was stained with DAPI (blue). Black and white images show projections of bnAb staining only, while colored images show overlays. (C) Binding of 2G12 to S. mansoni cercariae. 2G12 bound to putative ciliated sensory papillae (1), superficial neural projections (2), and the duct opening on the oral sucker and dotted fractions of glycoprotein on the oral sucker (3). (D) Binding of PGT128 to S. mansoni cercariae. PGT128 recognized spines, ciliated sensory papillae, and basement membrane/tegument/glycocalyx structures (1), similar to ConA (Figure S5A). PGT128 also recognized surface motifs (spines, ciliated sensory papillae, and basement membrane/tegument/glycocalyx structures) above the actin layer (2) and the duct opening on the oral sucker and dotted fractions of glycoprotein on the oral sucker (3). (E) Binding of PGT121 to S. mansoni cercariae. PGT121 recognizes the pre- and post-acetabular gland system, which is sits interior of the parasite (1) and surface motifs of ciliated sensory papillae and basement membrane/tegument/glycocalyx structures above the actin layer (2). The green arrow indicates ciliated sensory papillae reaching outside while being anchored to the outer surface by actin (red arrow). PGT121 also bound the duct opening on the oral sucker and dotted fractions of glycoprotein, co-locating with actin on the oral sucker (3). (F) Binding of PGT151 to S. mansoni cercariae. PGT151 recognized surface motifs of ciliated sensory papillae above the actin layer with high intensity (1) and ciliated tufts of the flame cells and on the protone phridial tubules (2). Binding of glycan-independent HIV-1 bnAbs F425-b4e8 and VRC01 is shown in Figures S6A and S6B. Confocal microscopy staining was performed twice on different cercariae preparations. Representative images are shown from one experiment.

Figures 4

Analysis of HIV-1 neutralization and S. mansoni seroprevalence in IAVI Protocol C donors (A) Comparison of neutralization score between donors with S. mansoni SEA seroreactivity (SH1) and no seroreactivity (SH0) at different time points after HIV-1 infection (TP2, TP3, and the time point with highest neutralization score [Best]). Neutralization score (Score) is a numerical representation of neutralization breadth and potency against a HIV-1 cross-clade virus panel and categorized by “TOP” (score ≥ 1), “MEDIUM” (score < 1 and ≥0.5), and “WEAK” (score < 0.5) as determined by Landais et al. (Landais et al., 2016). Groups were compared using a chi-square test, but no statistical differences were observed. (B) Distribution of neutralizing epitopes targeted in sera (as determined by serum mapping studies; Landais et al., 2016) across S. mansoni status is displayed in a bar chart with the sum of each specificity overall. Single specificities are displayed in orange, while the distribution of mixed specificities is stacked on top in blue. (C) Comparison of geometric mean ID₅₀ at TP3, the number of viruses neutralized (breadth) at TP3, neutralization score at TP3, best neutralization score (from all time points in the study) (Landais et al., 2016), CD4 count at setpoint (cells per microliter), viral load (log RNA copies per milliliter) at setpoint, and time to ART in years from setpoint between donors with S. mansoni seroreactivity (SH1) and no S. mansoni seroreactivity (SH0). p values (Mann-Whitney test) are reported as follows: not significant (ns), ^∗p < 0.0332, ^∗∗p < 0.0021, ^∗∗∗p < 0.0002, and ^∗∗∗∗p < 0.00001. Additional cohort analysis can be found in Figures S7A–S7C.

Figure 5

Cross-reactivity of N332/V3 PCDN76-lineage mAbs with S. mansoni isolates (A) Binding of PCDN76 38A and 38B and their less mutated precursors UCA-P, UCA-S, 22A, and 22C to recombinant JR-CSF gp120 was measured using ELISA. Experiments were performed in duplicate and repeated twice. A representative dataset is shown. Error bars represent the range of the value for experiments performed in duplicate (not shown when smaller than symbol size). (B) Binding of PCDN76 38A and 38B and their less mutated Ab precursors to S. mansoni SEA, SCA, and AWA on ELISA (at 50 μg/mL). N-glycan-independent HIV-1 mAbs F425-b4e8 and VRC01 were used as negative controls and S. mansoni-positive human serum as positive control. Error bars represent data from repeated experiments. Bar graphs show the mean of the OD values from at least 3 repeat experiments, and the error bars represent the SD. (C–H) Binding of less mutated PCDN76 precursors (UCA-P, UCA-S, 22A, 22C, and 38A) to S. mansoni cercariae. HIV-1 mAbs were detected with Alexa Fluor 488 (green), actin was detected with rhodamine (red), and DNA was stained with DAPI (blue). Black and white images show projections of mAb staining only, while colored images show overlays. (C) UCA-P recognizes superficial neural projections and ciliated sensory papillae (1), similar to 2G12. UCA-P recognized motifs of superficial neural projections and ciliated sensory papillae are located above the actin layer (2). Similar to 2G12, UCA-P displays binding to duct opening on the oral sucker (3). (D) UCA-S recognizes surface motifs (spines, ciliated sensory papillae, and basement membrane/tegument/glycocalyx structures) above the actin layer (1), and basement membrane/tegument/glycocalyx structures similar to ConA and PGT128 (2). (E) 16A binds to glycoproteins on the body, which could be ciliated tufts of the flame cells and/or ciliated regions on the protone phridial tubules and to ciliated sensory papillae on the tail and the body (1). Furthermore, binding to the duct opening on the oral sucker is observed (2). (F) 22A recognizes spines, ciliated sensory papillae and basement membrane/tegument/glycocalyx structures, similar to ConA and PGT128 (1). 22A recognizes surface motifs (spines, ciliated sensory papillae and basement membrane/tegument/glycocalyx structures) above the actin layer (2). 22A binds dotted fractions of glycoprotein on the oral sucker (3). (G) 22C recognizes spines, ciliated sensory papillae, and basement membrane/tegument/glycocalyx structures (1). (H) 38A recognizes spines, ciliated sensory papillae, and basement membrane/tegument/glycocalyx structures. Confocal images showing staining of PGT128 and PCDN76 38B are shown in Figures S6C and S6D, respectively. Confocal microscopy staining was performed twice on different cercariae preparations. Representative images are shown from one experiment.