Identification of HIV gp41-specific antibodies that mediate killing of infected cells

Abstract

Antibodies that mediate killing of HIV-infected cells through antibody-dependent cellular cytotoxicity (ADCC) have been implicated in protection from HIV infection and disease progression. Despite these observations, these types of HIV antibodies are understudied compared to neutralizing antibodies. Here we describe four monoclonal antibodies (mAbs) obtained from one individual that target the HIV transmembrane protein, gp41, and mediate ADCC activity. These four mAbs arose from independent B cell lineages suggesting that in this individual, multiple B cell responses were induced by the gp41 antigen. Competition and phage peptide display mapping experiments suggested that two of the mAbs target epitopes in the cysteine loop that are highly conserved and a common target of HIV gp41-specific antibodies. The amino acid sequences that bind these mAbs are overlapping but distinct. The two other mAbs were competed by mAbs that target the C-terminal heptad repeat (CHR) and the fusion peptide proximal region (FPPR) and appear to both target a similar unique conformational epitope. These gp41-specific mAbs mediated killing of infected cells that express high levels of Env due to either pre-treatment with interferon or deletion of vpu to increase levels of BST-2/Tetherin. They also mediate killing of target cells coated with various forms of the gp41 protein, including full-length gp41, gp41 ectodomain or a mimetic of the gp41 stump. Unlike many ADCC mAbs that target HIV gp120, these gp41-mAbs are not dependent on Env structural changes associated with membrane-bound CD4 interaction. Overall, the characterization of these four new mAbs that target gp41 and mediate ADCC provides evidence for diverse gp41 B cell lineages with overlapping but distinct epitopes within an individual. Such antibodies that can target various forms of envelope protein could represent a common response to a relatively conserved HIV epitope for a vaccine.

Fig 1. Results of BAMA assay for QA255 antibodies.

(A) The binding of the QA255 mAbs indicated at the top of each column are shown in relation to the antigen tested in the BAMA assay, which is indicated in the first two columns. The first two mAbs bind to different epitopes on gp120 [65] and served as controls. Binding to each antigen is defined by the fold increase over background (results with HIV negative plasma) and the binding is color coded as indicated to the right, with increasing shades of red indicating more binding. Gray indicates binding was not detected above background (<2-fold). (B & C) Binding defined by ELISA to MN gp41 and ZA.1197 gp41 ectodomain proteins. Mean absorbance +/- SD is shown in relation to antibody concentration; the dotted line indicates the limit of detection. Data are representative of at least two independent experiments. The key for the mAbs tested is shown to the right.

Fig 2. gp41-specific QA255 mAbs mediate ADCC activity.

Percent ADCC activity for target cells coated with either (A) clade B MN gp41 protein, (B) clade C ZA.1197 gp41 ectodomain or (C) clade A Q461.e2 TAIV gp140 protein is shown on the y-axis. The key to the mAbs tested is shown in the lower right corner, with QA255 gp41-specific mAbs shown by blue bars, a gp120-specific mAb by green bars, and a control Influenza mAb Fi6_v3 in grey. Results are the average of two replicates +/- SD. The results are representative of studies with PBMCs from two different donors (S1 Fig).

Fig 3. Results of competition ELISAs with mAbs that target known epitopes in gp41.

(A) The mAbs used for competition experiments and their epitope targets are shown with a schematic of gp41 below. (B-E) Results of competition experiments reported as biotinylated mAb (B) QA255.006, (C) QA255.016, (D) QA255.067, (E) QA255.072 binding in the presence versus absence of the competitor mAbs. The tested Bt-Ab is indicated at the top of each panel, and the competitor mAb is defined by separate colors as specified in the legend to the right of the graphs. The results shown represent the mean relative binding (+/- SD), are from technical duplicates in the same experiment and are representative of at least two biological replicates.

Fig 4. Peptides enriched in phage display immunoprecipitation with gp41 mAbs and their variation in natural sequences.

(A) The top two panels show the peptides that were enriched by phage display immunoprecipitation for QA255.067 and QA255.072, with the mostly highly enriched peptides shown at the top of the list. The common sequences among all the enriched peptides are highlighted in gray. (B) This third panel shows a summary of the core sequences identified for these mAbs and compared to 240-D. (C) Logo plot of 5,471 sequences of HIV from the LANL database across the epitopes defined for these mAbs. Colors indicate relative hydrophobicity.

Fig 5. Binding to and ADCC activity against pre- and post-fusion gp41 intermediates.

(A-B) Average binding (+/- SD) to the (A) 6-Helix and (B) 5-Helix peptide is measured on the y-axis. Monoclonal antibody D5, which binds to the NHR region, was used as a positive control. (C) ADCC activity is measured against the 6-Helix protein using a modified version of the RF-ADCC assay. Average ADCC activity (+/- SD) is shown for each of the mAbs indicated by the key on the right. Data are representative of at least three independent experiments.

Fig 6. Infected cell recognition and ADCC susceptibility to Cluster I and Cluster II antibodies.

(A and C) Binding or (B and D) ADCC activity was measured against cells infected with a wildtype NL4.3 virus construct expressing the ADA envelope (pNL43/ADA/WT), the construct with a deficient nef (pNL43/ADA/N-) or vpu gene (pNL43/ADA/U-), the construct with both nef- and vpu- deficient genes (pNL43/ADA/N-U-), or the construct with both deficient genes and containing the D368R mutation in the ADA envelope. In C and D cells were treated with IFNα as described in the Methods. Data represent the average +/- SD of 5 (A and B) and 4 (C and D) independent experiments.

Fig 7. Cell based ELISA to detect Env recognition at the cell surface.

(A) Binding to 293T cells transfected with an empty pcDNA3.1 plasmid or increasing concentrations of a plasmid expressing HIV-1_JRFLΔCT Env as described in the Methods. The key to the antibody tested is shown to the right. (B) Binding to cells pre-incubated in the presence or absence of sCD4 (10μg/mL for 1 hr at room temperature) before addition of the different anti-Env Abs. The concentration of plasmid expressing HIV Env used in the transfection corresponds to the 1X condition in panel A. For both panels, signals obtained with the empty pcDNA3.1 plasmid (negative control) were subtracted from signals obtained from Env-transfected cells. Results are presented as the average +/- SD of relative luminescence units (RLU). Results are representative from three independent experiments performed in quadruplicate.