The type-specific neutralizing antibody response elicited by a dengue vaccine candidate is focused on two amino acids of the envelope protein

Abstract

Dengue viruses are mosquito-borne flaviviruses that circulate in nature as four distinct serotypes (DENV1-4). These emerging pathogens are responsible for more than 100 million human infections annually. Severe clinical manifestations of disease are predominantly associated with a secondary infection by a heterotypic DENV serotype. The increased risk of severe disease in DENV-sensitized populations significantly complicates vaccine development, as a vaccine must simultaneously confer protection against all four DENV serotypes. Eliciting a protective tetravalent neutralizing antibody response is a major goal of ongoing vaccine development efforts. However, a recent large clinical trial of a candidate live-attenuated DENV vaccine revealed low protective efficacy despite eliciting a neutralizing antibody response, highlighting the need for a better understanding of the humoral immune response against dengue infection. In this study, we sought to identify epitopes recognized by serotype-specific neutralizing antibodies elicited by monovalent DENV1 vaccination. We constructed a panel of over 50 DENV1 structural gene variants containing substitutions at surface-accessible residues of the envelope (E) protein to match the corresponding DENV2 sequence. Amino acids that contribute to recognition by serotype-specific neutralizing antibodies were identified as DENV mutants with reduced sensitivity to neutralization by DENV1 immune sera, but not cross-reactive neutralizing antibodies elicited by DENV2 vaccination. We identified two mutations (E126K and E157K) that contribute significantly to type-specific recognition by polyclonal DENV1 immune sera. Longitudinal and cross-sectional analysis of sera from 24 participants of a phase I clinical study revealed a markedly reduced capacity to neutralize a E126K/E157K DENV1 variant. Sera from 77% of subjects recognized the E126K/E157K DENV1 variant and DENV2 equivalently (<3-fold difference). These data indicate the type-specific component of the DENV1 neutralizing antibody response to vaccination is strikingly focused on just two amino acids of the E protein. This study provides an important step towards deconvoluting the functional complexity of DENV serology following vaccination.

Figure 1. Surface-accessible residues that differ between DENV1 and DENV2 identified for mutagenesis.

The flavivirus E protein contains three distinct domains (DI–III) and forms antiparallel dimers on the mature virus particle. Shown is the crystal structure of the soluble ectodomain of the DENV2 E protein dimer (PDB 1OAN) as viewed from the top (top panels) and side (bottom panels). (A) Ribbon diagram of the E protein with DI, II, and III colored in red, yellow, and blue, respectively . (B) The 130 amino acid residues in the soluble ectodomain that differ between the DENV1 and DENV2 components of the NIAID candidate tetravalent vaccine are highlighted in blue on one E protein; the second E protein of the dimer is shaded in green. (C) Surface accessibility was estimated using solvent accessible surface areas of the residues determined from the crystal structure (UCSF Chimera package), with a cut-off value of 30 Ų . Residues selected for study were restricted to the top of the dimer. The 68 residues identified as surface-accessible differences between DENV1 WP and DENV2 NGC are shown in red.

Figure 2. Type-specific neutralization by DENV immune sera from a clinical vaccine trial.

Pooled immune sera from a DENV1 or DENV2 vaccine study were tested for their capacity to neutralize DENV1 and DENV2 RVPs. Sera pooled from multiple vaccinees (two and three for DENV1 and DENV2, respectively) were used in neutralization experiments by incubating RVPs with serial dilutions of immune sera for one hour at room temperature, before addition to Raji-DCSIGNR cells. After incubation at 37°C for two days, infection was measured using flow cytometry. Dose response curves for (A) DENV1 sera and (C) DENV2 sera are expressed relative to the infectivity of the RVPs in the absence of serum. The concentration of sera indicated on the x-axis is expressed as Log₁₀ (dilution factor of serum). Error bars represent the standard error of duplicate infections. Dose response curves shown in (A) and (C) are representative of 11 and nine independent experiments, respectively, performed using at least three independent RVP preparations. Neutralization titer (NT50) values were determined by nonlinear regression analysis using Prism software (GraphPad), and are summarized for (B) DENV1 sera and (D) DENV2 sera. Error bars represent standard error of the mean. ***p<0.0001.

Figure 3. Impact of mutations on the neutralization potency of DENV1 immune serum.

A panel of 54 DENV1 RVP variants containing single, double, or triple amino acid changes was constructed by site-directed mutagenesis. This panel represented all surface-accessible residues identified in Figure 1c . (A) The infectious RVP titer for each variant was determined concurrently with WT DENV1 using Raji-DCSIGNR cells. Values are the mean relative titer as compared to WT DENV1 infectivity measured in parallel from at least two independent RVP preparations; error bars represent standard error of the mean. Variants located in domains I, II and III of the E protein are colored in red, yellow, and blue, respectively. The sensitivity of the 54 DENV1 variants to neutralization by DENV1 immune serum was compared to WT DENV1 as described in Figure 2 . (B–D) Examples of the three patterns of neutralization by DENV1 immune sera observed are shown. Error bars represent the standard error of duplicate infections. (E) Neutralization sensitivities of all DENV1 variants to DENV1 immune sera are depicted as the mean fold increase in neutralization sensitivity ([NT50 variant]/[NT50 WT]); error bars represent standard error of the mean of 2–5 independent experiments.

Figure 4. Combined effect of DENV1 mutations E126K/E157K on the neutralization potency of DENV1 immune serum.

(A) The location of residues E126 and E157 are highlighted on the E protein crystal structure as cyan and green spheres, respectively. E protein domains are colored as in Figure 1 . (B) Infectious titer of DENV1 E126K/E157K RVPs harvested at four time points post-transfection was determined in parallel studies with WT DENV1 using Raji-DCSIGNR cells; error bars represent the standard error of the mean of 2–4 independent experiments. (C and D) DENV1 E126K/E157K RVPs were tested for sensitivity to neutralization by the DENV1 immune serum. (C) Representative dose-response curves for the single and double mutants are shown; error bars represent the standard error of duplicate infections. (D) The neutralization sensitivity is summarized as the fold-increase in NT50 from WT DENV1 for the single mutants E126K and E157K (n = 10), the E126K/E157K double mutant (n = 11), and DENV2 (n = 11); error bars represent the standard error of the mean. ***p<0.0001 for a comparison of the Log NT50 values to WT DENV1 by an ANOVA followed by Tukey's multiple comparisons test.

Figure 5. Characterization of the E126K/E157K DENV1 variant.

(A and B) WT DENV1 and DENV1 E126K/E157K RVPs were evaluated for sensitivity to neutralization by DENV1 DIII-binding mAbs. Within each panel, representative dose response curves for each antibody are shown on the left; error bars represent the standard error of duplicate infections. Plots on the right show the EC50 values obtained from independent experiments; error bars represent standard error of the mean. The antibodies tested were (A) mAb E103 (n = 8, p = 0.14), and (B) mAb E105 (n = 5, p = 0.57). (C) WT DENV1 and E126K/E157K RVPs were analyzed by Western blot with an anti-E mAb and an anti-prM mAb. The efficiency of prM cleavage was evaluated on blots normalized by loading equivalent E protein. (D) WT DENV1 RVPs were produced using standard methods (Std), in the presence of high levels of human furin expression (furin), or in cells treated with furin inhibitor (FI) and then tested for sensitivity to neutralization by mAb E60. Three independent experiments were performed; representative dose response curves are shown. Error bars represent the standard error of duplicate infections. (E) DENV1 E126K/E157K was evaluated for sensitivity to neutralization by mAb E60 as compared to WT DENV1. Representative dose response curves are shown on the left; error bars represent the standard error of duplicate infections. EC50 values from independent experiments are shown in the right panel; error bars represent standard error of the mean (n = 8, p = 0.68). (F) Furin- and FI-DENV1 RVPs were tested for sensitivity to neutralization by sera from DENV1 vaccine recipients. Error bars represent standard error from three independent experiments. Statistical evaluation using ANOVA followed by a Šidák correction for multiple comparisons failed to identify a difference between Furin- and FI-DENV1 RVPs (p>0.05 for each pair). (G) DENV1 E126K/E157K was evaluated for sensitivity to neutralization by pooled sera from DENV2 vaccine recipients. A representative dose response curve is shown on the left; error bars represent the standard error of duplicate infections. NT50 values obtained from eight independent experiments are shown on the right; error bars represent standard error of the mean (p = 0.08).

Figure 6. Longitudinal analysis of the effects of DENV1 E126K/E157K mutations on serum neutralizing activity.

Sera collected from three DENV1 vaccine recipients at five times post-vaccination were tested for a capacity to neutralize WT DENV1, DENV2, and DENV1 E126K/E157K RVPs. (A) Dose-response curves for immune sera from one subject are shown. Error bars represent the standard error of duplicate infections. (B) NT50 values for each curve were determined by nonlinear regression analysis using Prism software (GraphPad), and are summarized for the three subjects.

Figure 7. Cross-sectional analysis of the contribution of the E126K/E157K epitopes on TS-neutralization.

Sera from an additional 21 DENV1 vaccine recipients collected on days 42 and 222 (when available) were tested for their capacity to neutralize WT DENV1, DENV2 and DENV1 variant E126K/E157K RVPs. (A) The mean neutralization potency (Log NT50) of the sera against each virus is shown; error bars represent one standard deviation of the mean. ***p<0.0001; ns, p = 0.15. (B) The neutralization titers (NT50) of the sera against WT DENV1, DENV2 and DENV1 E126K/E157K RVPs are presented. The fold-difference in sensitivity between DENV1 variant E126K/E157K RVPs and DENV2 RVPs was determined by the equation (NT50 DENV1 E126K/E157K)/(NT50 DENV2). Red shading indicates a <3 fold difference, yellow shading corresponds to a 3–6 fold difference, and green shading reflects a >6 fold difference.