Molecular evolution of broadly neutralizing Llama antibodies to the CD4-binding site of HIV-1

Abstract

To date, no immunization of humans or animals has elicited broadly neutralizing sera able to prevent HIV-1 transmission; however, elicitation of broad and potent heavy chain only antibodies (HCAb) has previously been reported in llamas. In this study, the anti-HIV immune responses in immunized llamas were studied via deep sequencing analysis using broadly neutralizing monoclonal HCAbs as a guides. Distinct neutralizing antibody lineages were identified in each animal, including two defined by novel antibodies (as variable regions called VHH) identified by robotic screening of over 6000 clones. The combined application of five VHH against viruses from clades A, B, C and CRF_AG resulted in neutralization as potent as any of the VHH individually and a predicted 100% coverage with a median IC50 of 0.17 µg/ml for the panel of 60 viruses tested. Molecular analysis of the VHH repertoires of two sets of immunized animals showed that each neutralizing lineage was only observed following immunization, demonstrating that they were elicited de novo. Our results show that immunization can induce potent and broadly neutralizing antibodies in llamas with features similar to human antibodies and provide a framework to analyze the effectiveness of immunization protocols.

Figure 1. 3E3, A14, B9 and B21 neutralization breadth and potency.

(A, B, C, D) Each spoke represents a strain of HIV neutralized by B9 (A), B21 (B) A14 (C) and 3E3 (D) in the TZM-bl assay described in the Materials and Methods. Strains from different clades and CRF are color-coded according to the labels. The outer circle represents an IC50 of <1 µg/ml, the inner circle <5 µg/ml and the centre of the circle 50 µg/ml. legend. (E) Median IC50 µg/ml generated in the TZM-bl assay for each clade/CRF shown on the X-axis for the four VHH indicated in the color-coded legend. All viruses were assayed in duplicate to generate IC50 values.

Figure 2. A14, B9, B21 binding to HIV ENV.

VHH binding to immunogens (A) clade B gp140 R2, (B) clade C gp140 96ZM651.02, (C) clade A 92UG037 gp140, (D) CRF BC CN54 gp140, (E) YU2 WT and D368R gp120, and (F) RSC3 mutant and RSC3 delta mutant gp120, was assessed by ELISA as described in the Materials and Methods. The positive control for gp140 binding was J3 (McCoy et al. 2012) that for gp41 binding 2H10 . The positive control for both RSC3 gp120 proteins was D47 a non-neutralizing RSC3-specific VHH (L McCoy unpublished data). All binding assays were carried out in duplicate and error bars represent standard deviation. These data are representative of at least three independent experiments.

Figure 3. VHH mutation alters affinity, potency and breadth of neutralization.

(A) VHH binding to clade B gp120 Bal.26 in ELISA detected by their C-terminal Myc tag. (B) VHH neutralization of Bal.26 HIV pseudovirus in the TZM-bl assay. Mutant VHH were generated by site-directed mutagenesis as detailed in the Materials and Methods. (C) The fold change in IC50 µg/ml values for B9 S54W relative to B9, indicating the increase potency of the mutant are plotted on the Y-axis. The IC50 µg/ml values for B9 against each virus in the legend are plotted on the X-axis, indicating the baseline potency of B9. All assays were carried out in duplicate and error bars represent standard deviation. These data are representative of at least three independent experiments.

Figure 4. Combined breadth and potency of anti-CD4bs VHH.

(A) IC50 values in µg/ml against the HIV strains indicated in the left-hand column on TZM-bl cells. VHH were titrated five-fold in duplicate both individually (from 10 µg/ml) and in combination with one another in the same plate for each virus. IC50 values of less than 0.1 are color-coded in dark red, those between 0.1 and 1 in red, those between 1 and 10 in orange and those above 10 in yellow. (B) IC50 values in µg/ml against each of the HIV strains indicated on the X-axis are shown for J3 or the combined VHH according to the color-coded legend. (C) Each spoke represents a strain of HIV neutralized by J3 or the combined VHH. Strains from different clades and CRF are color-coded according to the labels. The outer circle represents an IC50 of <1 µg/ml, the inner circle <5 µg/ml and the centre of the circle 50 µg/ml.

Figure 5. VHH repertoires of naïve and protein-immunized llamas.

VHH sequences from immunized llamas 8 and 9 and seven naïve llamas were amplified from their respective phagemid libraries by PCR using primers specific to the 5′ and 3′ conserved regions of the VHH and subjected to 454 sequencing. (A) Unique sequences generated from the indicated llama phage library were used to build end-joining network diagrams with significantly more linkages (P = 0.001) in the naive llamas versus (B) immunized llamas. (C) Shared percentage identities with neutralizing VHH J3 and divergence from its inferred V gene Vt were calculated for all unique sequences from the control naïve llamas, and the J3-source llama 8. The left hand panel shows percentage identity for all sequences from naïve 3 plotted against divergence from Vt. The right panel shows percentage identity for all sequences from llama 8 plotted against divergence from Vt. The horizontal dotted line on each panel indicates the percentage identity shared by Vt and J3 and the vertical dotted line indicates the divergence of J3 from Vt, (D) Shared percentage identities with neutralizing VHH 3E3 and divergence from its inferred V gene Ve were calculated for all unique sequences from the control naïve llamas, and the 3E3-source llama 9. The left hand panel shows percentage identity for all sequences from naïve 3 plotted against divergence from Ve. The right panel shows percentage identity for all sequences from llama 9 plotted against divergence from Ve. The horizontal dotted line on each panel indicates the percentage identity shared by Ve and 3E3 and the vertical dotted line indicated the divergence of 3E3 from Ve.

Figure 6. VHH repertoires post DNA/VLP priming and gp140 boosts.

VHH sequences from immunized llama 1 at both t = 54 and t = 174 were amplified from their respective phagemid libraries by PCR using primers specific to the 5′ and 3′ conserved regions of the VHH and subjected to 454 sequencing. (A) Unique sequences generated from the llama 1 were used to build end-joining network diagrams with no significant difference in the number of network linkages between time points (B) Shared percentage identities with neutralizing VHH B9 its inferred V genes Vg were calculated for all unique sequences from both time points. Percentage identity with B9 for all sequences are plotted against divergence from Vg for t = 54 in the left hand panel and t = 174 in the right hand panel. (C) Unique sequences generated from the llama 3 were used to build end-joining network diagrams with no significant difference in the number of network linkages between time points (D). Shared percentage identities between neutralizing VHH B21 and its inferred V genes Vu were calculated for all unique sequences from both time points. Percentage identity with B21 for all sequences is plotted against divergence from Vu for t = 54 in the left hand panel and t = 174 in the right hand panel. In both (B) and (D) the horizontal dotted line on each panel indicates the percentage identity shared by the mature and germ line VHH and the vertical dotted line indicates the divergence of the mature VHH from its putative germ line precursor.

Figure 7. Anti-HIV activity of germ line V gene VHH.

GL VHH comprising inferred germ line V gene paired with the mature VHH CDR3 and J region were produced. GL VHH binding to (A) clade B immunogen R2 gp140 and (B) clade C immunogen 96ZM9651.02 gp140 was assessed by ELISA as per the materials and methods. GL VHH neutralization of (C) R2 pseudovirus and (D) 96ZM651.02 pseudovirus was assessed in TZMbl assay as per the materials and methods. (E) Unique sequences from llamas 1 and 3 at both timepoints were filtered for B9, A14 and B21 germ line V gene usage (Vg and Vu) respectively. The y-axis shows the percentage of sequences within each subset with identical residues to each VHH at each individual CDR3 position (and the preceeding three V gene residues −1,−2,−3 = CAT/CNA) indicated on the x-axis. (F, G, H) The number of unique sequences within the subsets which share exact runs of CDR3 residues are plotted on the y-axis against runs of increasing CDR3 length on the x-axis for (F) B9, (G) A14 and (H) B21 at the time points indicated in the legend.