Elicitation of Neutralizing Antibodies Targeting the V2 Apex of the HIV Envelope Trimer in a Wild-Type Animal Model

Abstract

Recent efforts toward HIV vaccine development include the design of immunogens that can engage B cell receptors with the potential to affinity mature into broadly neutralizing antibodies (bnAbs). V2-apex bnAbs, which bind a protein-glycan region on HIV envelope glycoprotein (Env) trimer, are among the most broad and potent described. We show here that a rare "glycan hole" at the V2 apex is enriched in HIV isolates neutralized by inferred precursors of prototype V2-apex bnAbs. To investigate whether this feature could focus neutralizing responses onto the apex bnAb region, we immunized wild-type rabbits with soluble trimers adapted from these Envs. Potent autologous tier 2 neutralizing responses targeting basic residues in strand C of the V2 region, which forms the core epitope for V2-apex bnAbs, were observed. Neutralizing monoclonal antibodies (mAbs) derived from these animals display features promising for subsequent broadening of the response.

Keywords: HIV; SOSIP trimer; V2 apex; bnAbs; germline targeting; glycan hole; immunizations; neutralizing antibodies; rabbit; vaccine.

Graphical abstract

Figure 1

Sensitivity of Select Envs to iP Versions of V2-Apex Prototype bnAbs (A) Env iP neutralization table. HIV isolates neutralized by three prototype V2-apex bnAb iP Abs and the CAP256 UCA. Neutralization IC₅₀ values against each pseudovirus are in micrograms per milliliter. Envelope features, including the strand-C lysine-rich basic patch sequence and V1 V2 regions, are also noted. The presence of PNG sites at specific sequences are marked by a plus (+) sign (two glycans are ++) and the total number of PNG sites in the V1V2 loop are given. NXT/S sequons are indicated by a bold red N in sequence alignments. Envs that were successfully generated as soluble native trimers are in blue font. (B) Trimer V2-apex model. The BG505 trimer (PDB: 5T3X) highlighting residue 130 (blue spheres) and the base of the V2′ hypervariable region loop. Sequence variation in this region between select Envs is shown in the insert on the right. Residues between the last structured amino acids in the V2′ region loop (185 and 187; HXB2) are highlighted in yellow with NXS/T sequons underlined. 130 and V2′ sites lie close to each other and to the V2-apex bnAb core epitope, including the strand-C basic residues N156 and N160 (pink). (C) iP antibody neutralization of V2-apex glycan variants. Shown are neutralization IC₅₀ values (in micrograms per milliliter) for mature or iP PG9 and CH01 antibodies against virus variants with V2-apex PNG sites at 130 or within the V2′ region either added or removed. (D) Biolayer interferometry (BLI) showing the binding and dissociation curves for PG9 and CH01 iPs and for CAP256 UCA to PGT145-purified SOSIPs. Binding was measured by incubation with the indicated concentrations of SOSIP (180–300 s) measured in delta (nanometers) or response units. Dissociation of SOSIP from the IgG in PBS buffer was further recorded.

Figure 2

Biophysical and Antigenic Properties of iP-Binding Soluble Trimers (A) Negative-stain EM. Class average, negative-stain EM images of CRF250, C108, WITO, and MGRM8 native trimers in descending order. (B) SOSIP glycan composition. Elution time of enzymatically released N-glycans from SOSIP immunogens in HILIC-UPLC. Area of eluted peaks represents the quantity of a particular glycoform (indicated above the peak). Pie diagrams indicate glycoform ratios for each of the immunogens. (C) Differential scanning calorimetry (DSC) showing the thermal stability of the SOSIP immunogens and their sequence-matched gp120s. Protein denaturation is measured in kilocal/mol/°C. Melting peak temperature (T_m) is indicated above the peak. (D) mAb binding to SOSIPs. Maximal binding in response units (RUs) of 500 nM CRF250 (red), C108 (blue), WITO (orange), and MGRM8 (purple) SOSIPs (colored circles) as well as their sequence-matched gp120s (black circles) to a panel of 85 mAbs measured by BLI. The Abs are divided by epitope as indicated on the far left y axis. RU values are given in Figure S1C. (E) Neutralization activities. Pseudovirus neutralization IC₅₀ values plotted against SOSIP-binding response units show a strong correlation (two-tailed Spearman R).

Figure 3

Rabbit Trimer Immunizations and Elicited Neutralizing Responses (A) Immunization schedule. Six groups of 4 female New Zealand white rabbits were immunized with mixtures or individual SOSIPs as indicated by the colored circles. 50 μg total protein and 100 U Iscomatrix were given intramuscularly (IM) at 0, 1, 3, and 6 months. Serum or plasma was collected at times indicated by red arrows. (B) Autologous and tier 1 neutralization responses in immunized rabbits. Serum IgG from rabbit pre-bleeds or at 14 days post-immunization (B1, B2, B3C, and B4C) was assessed for neutralizing ability against the sequence-matched pseudovirus for the SOSIP immunogens used in this study (colored boxes) or tier 1 pseudoviruses (gray boxes) on the left side of the table. Data are arranged by immunization group from left to right, and the immunogen most recently given for the time point in each group is indicated across the top of the table by a colored circle. Pseudovirus neutralization by rabbit immune serum IgG is given as the IC₅₀ (in micrograms per milliliter) of IgG purified from rabbit serum. Samples are indicated by rabbit identification numbers to the left of IC₅₀ values. Higher neutralizing responses are colored. (C) Neutralization of heterologous viruses by SOSIP mixture immunized rabbit, P5672. Virus neutralization as a function of rabbit immune serum IgG concentration is shown for five tier 2 clade AE viruses (92TH is an unconfirmed tier 2 virus), neutralized by the cocktail immunized rabbit P5672. The average percent neutralization and SDs from triplicate experiments are given. Orange, blue, and red curves show neutralization of these same viruses after absorption of the P5672 IgG with WITO (orange), C108 (blue), or CRF250 SOSIPs, respectively.

Figure 4

Specificities of Neutralizing Responses (A and B) Trimer dependence of neutralizing abs for the 10 rabbits with high titer neutralizing responses in this experiment are shown. D368R autologous SOSIP or gp120 protein was used to adsorb IgG from bleed “3C” samples before neutralization in the TZMB-l assay. The plots in (A) are neutralization data for 9 rabbits that raised responses to the autologous virus only so the sequence-matched SOSIP or gp120 was used to adsorb IgG in each sample. The plots in (B) are neutralization data for the three viruses that were neutralized by SOSIP mixture immunized rabbit, P5672, after the second boost. Absorption of neutralization by WITO, C108, and CRF250-gp120 or SOSIP was assessed for each virus. Error bars represent SD between replicate data points. (C) Envelope mutants of the viruses neutralized by rabbit immune IgG were generated as indicated in the left-hand column (HXB2 numbering). These include N160/N156 glycan knockouts, addition of PNG sites at predicted glycan holes, and mutation of the strand-C V2-apex bnAb core epitope of basic residues (169 and 171). In addition, wild-type viruses grown in the presence of kifunensine (Kif) were included. Neutralization IC₅₀s for these different pseudoviruses are given in μg/ml of rabbit IgG (identification number is given across the top). The limit of detection is 100 μg/mL. Titration curves are shown in Figure S4. (D) BG505 trimer side view (PDB: 5XT3) showing the glycan hole targeted at 234/276 in some immunized rabbits (compare with C). Glycans are absent at positions 234 (blue) and 276 (purple) in CRF250 trimer and 234 in MGRM8. resulting in glycan holes.

Figure 5

Neutralizing Monoclonal Rabbit Antibodies Derived from Three Different Immunized Animals (A) Heavy and light (kappa) chains including genetic identity to their most similar O. cuniculus germline genes (from the IMGT database), CDR lengths, and CDR3 sequences. YD motifs in the heavy CDRH3 or kappa CDRL3 loops are highlighted. (B) Neutralization IC₅₀ values of three C108-specific mAbs (in micrograms per milliliter) are given to the right of values for the parent rabbit immune serum IgG IC₅₀ values for a variety of C108 variants (left column). 3903-1 and 3903-2 are members of a clonal lineage (Figure S5A). (C) Neutralization IC₅₀ values of CRF250-specific mAbs derived from two different animals. As in (B), parent immune serum IgG IC₅₀ values are given to the left of the values for the mAbs against the virus variants indicated in the left column.