Novel biopanning strategy to identify epitopes associated with vaccine protection

Abstract

Identifying immune correlates of protection is important to develop vaccines against infectious diseases. We designed a novel, universally applicable strategy to profile the antibody (Ab) repertoire of protected vaccine recipients, using recombinant phages encoding random peptide libraries. The new approach, termed "protection-linked (PL) biopanning," probes the Ab paratopes of protected vaccinees versus those with vaccine failure. As proof of concept, we screened plasma samples from vaccinated rhesus macaques (RMs) that had completely resisted multiple mucosal challenges with R5-tropic simian-human immunodeficiency viruses (SHIVs). The animals had been immunized with a multicomponent vaccine (multimeric HIV-1 gp160, HIV-1 Tat, and SIV Gag-Pol particles). After PL biopanning, we analyzed the phagotopes selected for amino acid homologies; in addition to the expected Env mimotopes, one recurring motif reflected the neutralizing Ab epitope at the N terminus (NT) of HIV-1 Tat. Subsequent binding and functional assays indicated that anti-Tat NT Abs were present only in completely or partially protected RMs; peak viremia of the latter was inversely correlated with anti-Tat NT Ab titers. In contrast, highly viremic, unvaccinated controls did not develop detectable Abs against the same epitope. Based upon the protective effect observed in vivo, we suggest that Tat should be included in multicomponent HIV-1 vaccines. Our data highlight the power of the new PL-biopanning strategy to identify Ab responses with significant association to vaccine protection, regardless of the mechanism(s) or targets of the protective Abs. PL biopanning is also unbiased with regard to pathogens or disease model, making it a universal tool.

Fig 1

Protection-linked (PL) biopanning to identify Ab epitopes associated with protection. Plasma samples from three vaccine-protected rhesus macaques (RMs) (RRi-11, RTr-11, and RGe-11) (15) were screened for Ab responses specific for protection using a peptide phage display-based approach (PL biopanning). Each PL-biopanning round consisted of (1) positive selection, (2) negative selection, and (3) amplification of phages selected. Paramagnetic beads (in brown) were used to capture anti-RM IgGs. Red, the Ab Fc portion; light gray, the Fab domain of anti-RM IgGs immobilized on beads via Fc. Positive selection used plasma from one of the three protected animals (week 0). Abs induced by immunization and corresponding phages bound to Abs linked to protection are shown in dark blue. Positively selected phages were counter-selected with plasma from a nonprotected, highly viremic vaccinee (RDo-11 [15], enrolled in same study as monkeys RRi-11, RTr-11, and RGe-11 that had been used as positive selectors) (same time point). Nonprotective Abs from this negative selector and corresponding bound phages are shown in yellow; purple or red indicates unspecific phages bound to anti-RM Ab or beads, respectively. Unspecific phages and phages recognized by Abs that were not linked to protection (in yellow) are depleted during the negative selection, whereas phages representing epitopes of Abs linked to protection are enriched during the positive selection (in blue). Phages eluted after the third round of positive selection (PL mimotopes) were sequenced and grouped according to peptide motifs.

Fig 2

Spiking plasma of a vaccinated RM with a known nMAb: can PL biopanning identify the expected epitope? (A) The sequences of the consensus clade C V3 crown and the 33C6-specific mimotope used for Ab isolation (Tc.2) are shown at the top of the panel, as indicated. The linear alignment shows the sequences of eight unique 33C6-specific mimotopes (33C6 PL-mimes) isolated by spiked PL biopanning (Table 2, experiment I). Homologies between them and Tc.2 are shaded in dark gray, and additional homologies to the consensus clade C V3 crown are shaded in light gray. (B) The eight 33C6 PL-mimes from panel A were tested in a phage ELISA for binding with MAb 33C6, a second anti-V3 Ab, HGN194 (29), the spiked plasma (RGe-11 plus MAb 33C6), and the two plasma samples used for PL biopanning (positive selector, RGe-11; negative selector, RDo-11). The original 33C6-specific mimotope Tc.2 was the positive control, and the wild-type (WT) phage was the negative control. One plasma Ab-specific PL mimotope (RGe-11 PL-mime) isolated with experiment I was included as well. Dashed line, OD signals 10 times higher than signals detected with the wild-type phage.

Fig 3

Sequence alignment and mimotope location on HIV-1 Tat. (A) The sequence of the immunogen (HIV-IIIB) is illustrated in gray in the header. Homologies between mimotopes and the parental strain are shaded in gray. Alignment shows sequences for five protection-linked Tat mimotopes isolated from RGe-11 (PL-Tat mimes) and eight Tat mimotopes isolated using plasma from RAt-9 (RAt-9 Tat mimes). The three different biopanning experiments are indicated (Table 2, experiments II to IV). (B) Sequence logo of 12 mimotopes (in bold in panel A) representing the NT of Tat using WebLogo, version 3. Bits represent the relative frequency of amino acids. (C) Three-dimensional location of the NT epitope on Tat protein (PDB code 1JFW). The five conserved amino acid residues are highlighted in blue (LEPWK). The figure was prepared with Chimera (28).

Fig 4

Cross-reactivity profile of Tat mimotopes. For each animal, a time point before the first live-virus encounter (week 0 or week −2) and a time point after SHIV-C exposures (week 7 or week 30) were tested; numbers below monkey names indicate weeks tested. Binding patterns are shown in the form of a heat map, and OD mimotope signals relative to those of the wild-type phage control are color coded as indicated. White squares, binding signals below the cutoff. Negative control (N), preimmune plasma pool of vaccinees. The animals were ranked from left to right according to increasing peak vRNA loads. (A) Binding profile for protection-linked Tat mimotopes (PL-Tat mimes) isolated using the PL-biopanning approach. (B) Binding profile for Tat mimotopes isolated using plasma from RAt-9 (RAt-9 Tat mimes).

Fig 5

Epitope mapping using anti-Tat MAbs. Two mouse anti-Tat MAbs (2A4.1 and NT3 2D1.1) were tested for binding to linear NT peptides and anti-Tat neutralization activity as well as their potential to recognize the Tat mimotopes. (A) Binding ELISA results using the anti-Tat MAbs. The sequences of the three overlapping peptides containing the same conserved NT motif as the Tat mimotopes are shown (peptides 5113 to 5115). Negative controls, human IgG1 MAb trastuzumab and a scrambled C-terminal gp120 peptide (control; sequence shown in inset). (B) HLM1 cell-based viral rescue assay to determine the potential of anti-Tat MAbs to neutralize Tat transactivation. Both MAbs were used at 40 μg/ml. The height of the bars illustrates the percentage of transactivation compared to the transactivation measured without any MAbs added (values for the cells plus Tat considered 100% transactivation). (C) Both MAbs were tested for their reactivity with the Tat mimotopes. Negative controls, MAb trastuzumab and helper phage without a peptide insert (WT, wild type). For all three panels, the height of each bar represents the result from at least two independent assays. The error bars represent the standard errors of the means.

Fig 6

Purified IgG from vaccinees neutralizes Tat transactivation in vitro. IgG was purified from RJr-11 and RBr-11 (plasma pool from weeks 2 and 3) and tested for binding to linear NT Tat peptides as well as for Tat neutralization. Control IgG purified from two naïve RMs was included in all assays (naïve IgGs). (A) Binding ELISA results using linear Tat peptides (pool of peptides 5113 to 5115). (B) Purified IgG was tested in HLM1 cell-based viral rescue assay. The nMAb NT3 2D1.1 was included as a positive control (data not shown). The height of the bars represents the average results of two independent assays, testing each IgG fraction in triplicates. Error bars represent the standard errors of the means. The percentage of transactivation between naïve and immune IgG was compared using an unpaired, two-tailed Student's t test (significance after Bonferroni correction, P < 0.025).

Fig 7

Epitope mapping using a consensus B Tat peptide library. Polyclonal plasma taken at week 1 (vaccinees RDo-11, RJr-11, and RBr-11) was tested for binding to peptides of an overlapping HIV-1 clade B Tat library (peptides 5113 to 5135). Negative controls, a scrambled C-terminal gp120 peptide (control) (sequence shown in Fig. 5A, inset) and a pool of the preimmune plasma of the same three vaccinees. A signal 20 times higher than the signal detected with the control peptide was considered positive (dashed line). The height of the bars represents the average results of two independent assays, testing each peptide in duplicate. Error bars represent the standard errors of the means. For better orientation, the overlapping peptides were organized into six protein domains.

Fig 8

Abs targeting HIV-1 Tat NT are linked to a reduced risk of SHIV-C acquisition. The anti-Tat binding titers of polyclonal plasma Abs in 12 vaccinated and 10 control RMs (all derived from the same immunization study [15]) were determined. Vaccinees were ranked in ascending order of peak viremia after the five low-dose challenges. Aviremic RMs (RRi-11, RTr-11, RGe-11, and RFo-11) were assigned a vRNA load of 49 copies/ml (34) (red dots, peak vRNA loads). All four animals received the high-dose challenge, after which RGe-11 and RFo-11 became viremic. One of the control animals from this immunization study (RAk-11) remained aviremic during the low-dose challenges (15) but became viremic after the high-dose challenge (asterisk and red dot, right y axis). X, mean peak plasma viremia of unvaccinated controls; W, week of peak viremia. The binding titers were determined at week 0 (2 weeks after the last immunization but before virus challenges; striped bars) and week 7 (after virus challenges; black bars). (A) Binding Ab titers against Tat NT peptides (pool of peptides 5113 to 5115). Inverse correlation of binding titers with peak viremia was assessed using Spearman correlation analysis. The Spearman's rank correlation coefficient (r) and the P value are shown for the vaccine-induced Ab titers (week 0, striped bars). (B) Binding Ab titers against full-length Tat protein. (C) Binding Ab titers against Tat NT peptides in 10 unvaccinated but SHIV-1157ipEL-p-challenged control animals (15). RAk-11 is the aviremic control animal (in red).