Signatures of VH1-69-derived hepatitis C virus neutralizing antibody precursors defined by binding to envelope glycoproteins

Abstract

An effective preventive vaccine for hepatitis C virus (HCV) remains a major unmet need. Antigenic region 3 (AR3) on the E1E2 envelope glycoprotein complex overlaps with the CD81 receptor binding site and represents an important epitope for broadly neutralizing antibodies (bNAbs) and is therefore important for HCV vaccine design. Most AR3 bNAbs utilize the V_H1-69 gene and share structural features that define the AR3C-class of HCV bNAbs. In this work, we identify recombinant HCV glycoproteins based on a permuted E2E1 trimer design that bind to the inferred V_H1-69 germline precursors of AR3C-class bNAbs. When presented on nanoparticles, these recombinant E2E1 glycoproteins efficiently activate B cells expressing inferred germline AR3C-class bNAb precursors as B cell receptors. Furthermore, we identify critical signatures in three AR3C-class bNAbs that represent two subclasses of AR3C-class bNAbs that will allow refined protein design. These results provide a framework for germline-targeting vaccine design strategies against HCV.

Fig. 1. Binding and neutralization by germline AR3C-class bNAbs.

a Top: linear representation of full-length E1E2 with the hypervariable region 1 (HVR1), transmembrane domains (TMD) and hydrophobic external region (MPER) indicated. Bottom: schematic representation of the permuted E2E1 trimer design. E1 and I53-50A.NT1 (I53-50A) are separated by a Gly-Ser linker (GSGGSGGSGGSGGS). Numbering is based on standard H77 polyprotein numbering. E2 and E1 are separated by a Gly-Ser linker (GGSGGSGGSGGSGGS) followed by a furin cleavage site (RRRRRR). The protein contains a strepII-tag (WSHPQFEK) used for affinity purification. b Binding of AR3C-class bNAbs and their inferred germline predecessors against seven different E2E1 trimers from HCV genotypes 1-6. ELISA results are represented as area under the curve (AUC). A no-binding threshold of 0.001 is set by the dashed line. A paired two-sided Wilcoxon signed-rank test was used for the comparison (****P = 3.5e−15). c Representative ELISA binding curves for mature and inferred germline versions of AR3A, AR3C, and HEPC74 against H77, UKNP4.1.1, and UKNP5.2.1 E2E1 trimers. All experiments were performed in duplo (indicated by error bars). d BLI analysis of antibody binding to UKNP4.1.1 and UKNP5.2.1 E2E1-I53-50A. The indicated bNAbs were immobilized onto protein A biosensors and incubated with 500 nM of E2E1 trimers. The representative binding curves from two independent experiments. e Potency (in midpoint neutralization concentrations (IC₅₀)) of germline and mature AR3A, AR3C, and HEPC74 against HCVpp clones derived from seven different strains (genotypes 1-6). A paired two-sided Wilcoxon signed-rank test was used for the comparison (****P = 9.5e−07).

Fig. 2. Activation of B cells expressing mature and germline AR3C and HEPC74 by E2E1-I53-50A trimers and E2E1-I53-50NPs.

B cells expressing inferred germline (top) or mature antibodies (bottom) of AR3C (a) or HEPC74 (b) as BCRs were stimulated with either H77 or UKNP4.1.1 E2E1-I53-50A trimers (red), E2E1-I53-50 nanoparticles (purple), ionomycin (positive control, orange) or no stimulus (blue). The experiments were performed with 10 or 100 µg/mL E2E1-I53-50A or equimolar amount of E2E1-I53-50A in nanoparticles.

Fig. 3. Shared characteristics of AR3C-class bNAbs.

a Superposition of structures of AR3C-class bNAbs AR3C, AR3A, AR3B, AR3D, HEPC3, HEPC74, and AT1209 (PDB entries: 4MWF, 6BKB, 6BKC 6BKD 6MEI, 6MEH, 7T6X, respectively) on the E1E2 complex (PDB: 7T6X) which bind to the same AR3 epitope on the neutralizing face of E2. Only the variable heavy (VH) regions of the bNAbs are shown for clarity. Left panel: superposition of the CDRH2 (Kabat: 50-65). The exceptionally long CDRH2 of AT1209 is highlighted in orange. Right panel: superposition of the CDRH3 loops (Kabat: 91-104) of mature AR3C-class bNAbs with a bent (AR3A, AR3B, AR3C, AR3D, and AT1209) and straight CDRH3 (HEPC3 and HEPC74). b Length distribution of human CDRH2 (left) and CDRH3 (right). Human CDRH2 lengths were extracted from the online abYsis system (http://www.bioinf.org.uk/abysis/) using the Kabat numbering scheme Kabat and National Institutes of Health (U.S.). Highlighted are the relative positions of the AR3C-class antibodies.

Fig. 4. Impact of the heavy and light chains on AR3C-class antibody binding and neutralization.

To determine the importance of the heavy chain (HC) and light chain (LC) of AR3C-class bNAbs for binding and neutralization we generated chimeric mAbs by pairing AR3C-class HCs with the LC of the heterologous anti-HIV-1 VRC01 mAb. a Binding and neutralization by chimeric (gl-)AR3C HC – VRC01 LC mAbs in ELISA (left) and HCVpp neutralization assays (right). b Similar as (a), ELISA binding (left) and HCVpp neutralization (right) by chimeric (gl-)HEPC74 HC – VRC01 LC to a panel of HCV strains. Binding signal is expressed as area under the curve (AUC) and neutralization activity is expressed as IC₅₀ (μg/mL). Maximum concentration tested for each chimeric antibody was 50 μg/mL. Paired two-sided Wilcoxon signed-rank test were used for all calculations (*P < 0.05, ns=non-significant). ELISA experiments where performed in duplo while neutralization assays were performed in triplo.

Fig. 5. Polymorphisms in V_H1-69 alleles impact binding of germline AR3C-class bNAbs.

a Sequence alignment of CDRH3 of gl-AR3C and gl-HECP74 engrafted in V_H1-02*02 and V_H3-23*01. The CDRH1, 2, and 3 are highlighted. Sequences are numbered using Kabat numbering. Heat map summarizes the ELISA binding of the engrafted antibodies against different E2E1 trimers. Combinations that were not tested because of low production yield, are indicated in gray. b Similar to (a), sequence alignment of CDRH3 of gl-AR3C and gl-HECP74 engrafted in different V_H1-69 alleles. Heat map summarizes ELISA binding results. Combinations that were not tested because of low production yield, are indicated in gray or in the case of gl-HEPC74, V_H1-69*01 engrafted sequence is the same as WT sequence.

Fig. 6. Study of the F54 and G50 polymorphisms in binding of selected AR3C-class bNAbs.

a Binding of AR3C-class antibodies displaying the CxGGxC motif (left, germline; right, mature) with either F54 (F) or L54 (L) to all E2E1-I53-50A trimers used in this work. Note that mature HEPC74, contains S54 instead of F54 and is mutated here to S54L. A two-sided Wilcoxon matched-pairs signed-rank test was used for the comparison (ns: non-significant). All experiments were performed twice independently. b ELISA binding of gl-AR3A, gl-AR3C and gl-HEPC74 or the same mAbs with CDRH2 mutations F54L, G50R or F54L + G50R to UKNP2.2.1 and UKNP5.2.1 E2E1 trimers. Binding is represented as AUC and the dotted line represents the detection limit. c Top left panels: Alphafold2 predicted structure of gl-AR3A or gl-AR3A G50R (in orange) superimposed on the available structure of mature AR3A (white) in complex with of E2 (dark gray) (PDB: 6BKB). Bottom left panels: Alphafold2 predicted structure of gl-AR3C or gl-AR3C G50R (in blue) superimposed on the available structure of mature AR3C (white) in complex with of E2 (dark gray) (PDB: 4MWF). Top right panels: Alphafold2 predicted structure of gl-HEPC74or gl-HEPC74 G50R (red) superimposed on the available structure of mature HEPC74 (white) with E2 (dark gray) (PDB: 6MEH). For all G50R panels, the arginine at position 50 mutation is highlighted in sticks.

Fig. 7. In silico analysis of the predicted frequency of AR3C-class bNAb precursors in healthy human individuals.

a AR3C-class bNAb precursors in the naive B cell repertoire. The frequencies at which the shared characteristics of AR3C-class bNAb precursors are found in the naive B cell repertoire (data available at https://github.com/briney/grp_paper) are shown for the different donors: all V_H1-69 (left), V_H1-69 combined with a long CDRH3 (>19 amino acids, middle), V_H1-69 with a long CDRH3 and CxGGxC motif (right). b Comparison of the V_H gene usage in the complete and naive B cell repertoires of eight individuals. V_H gene usage was calculated as a percentage of the naive (upper panel) or complete (bottom panel) B cell repertoire. c V_H gene usage in the complete B cell repertoire represented as the fold change compared to the naive repertoire calculated for each of the eight individuals. Gray bars represent the median values per V_H gene. d Comparison of levels of somatic hypermutation (SHM) between the 7 V_H1-69-derived AR3C-class mature HCV bNAbs and 10 V_H1-02-derived VRC01-class HIV-1 bNAbs at the amino acid level. The crossbars represent median percentage identity to germline. A two-sided Mann-Whitney U test was used for the comparison (**P = 0.0054).