Fusion of the molecular adjuvant C3d to cleavage-independent native-like HIV-1 Env trimers improves the elicited antibody response

Abstract

An effective HIV vaccine likely requires the elicitation of neutralizing antibodies (NAbs) against multiple HIV-1 clades. The recently developed cleavage-independent native flexibly linked (NFL) envelope (Env) trimers exhibit well-ordered conformation and elicit autologous tier 2 NAbs in multiple animal models. Here, we investigated whether the fusion of molecular adjuvant C3d to the Env trimers can improve B- cell germinal center (GC) formation and antibody responses. To generate Env-C3d trimers, we performed a glycine-serine- based (G₄S) flexible peptide linker screening and identified a linker range that allowed native folding. A 30-60- amino- acid- long linker facilitates Env-to-C3d association and achieves the secretion of well-ordered trimers and the structural integrity and functional integrity of Env and C3d. The fusion of C3d did not dramatically affect the antigenicity of the Env trimers and enhanced the ability of the Env trimers to engage and activate B cells in vitro. In mice, the fusion of C3d enhanced germinal center formation, the magnitude of Env-specific binding antibodies, and the avidity of the antibodies in the presence of an adjuvant. The Sigma Adjuvant System (SAS) did not affect the trimer integrity in vitro but contributed to altered immunogenicity in vivo, resulting in increased tier 1 neutralization, likely by increased exposure of variable region 3 (V3). Taken together, the results indicate that the fusion of the molecular adjuvant, C3d, to the Env trimers improves antibody responses and could be useful for Env-based vaccines against HIV.

Keywords: C3d; HIV-1; immunogenicity; molecular adjuvant; vaccine.

Figure 1

Activation by complement system, design, and schematic representation of HIV-1 Env-C3d fusion Trimers. (A) Cartoon model depicting enhanced activation of the innate immune system by simultaneous engagement of HIV-1 Env-C3d trimers with naïve B- cell receptors (BCRs) and CR2, CD19, and CD81 receptors of the complement system, respectively. (B) Further activation of the adaptive immune system in germinal centers by the engagement of CD2 receptors on follicular dendritic cells (FDCs) with BCR- attached Env-C3d trimers. (C) Model of Env-C3d (16055 NFL TD CC T569G, PDB 5UM8; human C3d, PDB: 3OED). The three Env gp120 subunits are shown in blue; the gp41 ectodomain subunits in cyan to form the gp140 trimers; mouse C3d in magenta. The three C3d subunits were manually fitted to the C-termini of the NFL trimer by visual inspection in PyMOL. The dotted lines represent the linker connecting the C-termini of NFL Env gp140 to the N-termini of C3d. (D) Linear schematic diagram of JRFL-C3d trimers. Mouse C3d was fused to the C-termini of NFL gp140 trimers by a flexible linker (G₄S) of varying lengths (14, 30, and 60 amino acids to yield JRFL-L14-C3d, etc.). JRFL NFL TD CC+ (namely, JRFL) was used for the linker-length screening. The native signal peptide sequence of JRFL (or 426c NFL trimers) was replaced by the CD5 leader sequence to increase secretion in mammalian cells, as previously performed.

Figure 2

Biochemical and biophysical characterization of JRFL native-like-C3d fusion trimers. (A) SEC profile of JRFL-C3d trimers following lectin affinity purification and F105 negative selection. (B) Blue-native PAGE (BN-PAGE) analysis of JRFL-C3d trimers. (C) Western blot analysis of JRFL-C3d trimers. Protein identification was determined by anti-Env antibodies (2G12 and VRC01) or mouse C3d-specific antibody. Purified JRFL NFL native-like trimers and mouse C3d-foldon proteins were used as controls. (D) 2D averages from NS- EM of JRFL-C3d trimers revealed a native-like well-ordered structure. (E) DSC measurements of JRFL-C3d trimers. The T_m values are shown on top of each temperature peak.

Figure 3

The JRFL-C3d trimers display a favorable antigenic profile as determined by Bio-layer interferometry (BLI). (A) JRFL-C3d trimers bind to the mouse C3d-specific antibody, which was immobilized on streptavidin sensors. Binding was not detected for JRFL trimer lacking the C3d domain. (B) JRFL and JRFL-C3d trimers are recognized by bNAbs and are not recognized by non-NAbs. Data are representative of at least two independent experiments.

Figure 4

JRFL-C3d trimers bind and activate mouse B cells in vitro. (A) Flowcytometry analysis of JRFL-C3d trimers binding to mouse B cells with or without PGV04 expression on their surface. Percentage of maximum cell count is shown on the Y-axis, and relative fluorescence of Alexa Flour 680 (AF680) dye- conjugated trimers is shown on the X-axis. (B) Calcium flux analysis of B- cell activation by JRFL-C3d trimers. Mouse B cells expressing PGV04 on their surface were used. The resulting Ca²⁺ flux was measured over a period of 120 s. Representative data from two independent experiments is shown. JRFL trimers served as control for both the assays.

Figure 5

B cells from germinal centers in mice are better activated by JRFL-C3d trimers compared with those lacking C3d. (A) Comparison of activated GL7⁺ B cells measured by FACS after immunization with adjuvant- only, JRFL NFL, and JRFL-C3d trimers. Percent of activated GL7⁺ B cells were measured in draining popliteal and inguinal LNs and spleen 2 weeks after priming immunization. (B) Percent of activated GL7⁺ B cells in popliteal LN 6 weeks after two immunizations. The second immunization was performed at week 4. ns indicates non-significant. *, indicate P ≤ 0.05; **, indicate P ≤ 0.01; ***, indicate P ≤0.001.

Figure 6

Immunization schedule and comparison of Env-specific binding antibody responses elicited in mice. (A) Immunization groups, antigens, and vaccine regimen. (B) JRFL trimer-specific IgG-binding titers were determined using 2G12-capture ELISA. JRFL-C3d fusion trimers elicit statistically significant higher binding antibodies than trimers lacking the C3d domain. The ED₅₀ of each animal plotted, with geometric mean and min–max values shown. The box plots illustrate the following: horizontal line, median; plus, mean; box, interquartile range; whiskers, min/max. Statistical differences were evaluated by two-tailed Mann–Whitney U test. * indicates P ≤ 0.05; ** indicates P ≤ 0.01.

Figure 7

Avidity of the anti-Env IgG in mice after four immunizations. JRFL-specific NaSCN-displacement ELISA was used to measure the avidity index of anti-Env IgG in mouse sera at week 19 (2 weeks after the fourth immunization). The avidity index was calculated as the percentage of endpoint titers for NaSCN-treated samples to those of PBS-treated samples. The box plots illustrate the following: horizontal line, median; plus, mean; box, interquartile range; whiskers, min/max. Statistical differences were evaluated by two-tailed Mann–Whitney U test. * indicates P ≤ 0.05; ** indicates P ≤ 0.01. Data shown are representative from two independent experiments. JRFL-C3d trimers result in a statistically significant higher avidity index in the SAS adjuvant than in JRFL trimers alone in PBS and SAS adjuvant.

Figure 8

Design, antigenic characterization, and immunogenic analysis of 426c-ΔGly3-C3d trimers. (A) Schematic representation of 426c-ΔGly3-C3d fusion trimers. (B) 426c-ΔGly3-C3d trimers display favorable antigenic profile against Env-specific bNAbs, non-bNAbs, and mouse C3d-specific Ab. (C) Immunization groups and vaccine regimen. (D) ED₅₀ binding Ab titers with geometric mean and geometric standard derivation against 426c-ΔGly3 trimers (captured by 2G12) in VRC01^gHL mice after two immunizations. Data shown are an average of two independent measurements. Pre-immune sera from mice had no detectable specific IgG response and are not shown. Fusion of C3d domains to 426c-ΔGly3 trimers resulted in statistically significant improved binding titers in mice *, indicate P ≤ 0.05; **, indicate P ≤ 0.01; ***, indicate P ≤ 0.001. ns indicates non-significant..