A tetravalent live attenuated dengue virus vaccine stimulates balanced immunity to multiple serotypes in humans

Abstract

The four-dengue virus (DENV) serotypes infect several hundred million people annually. For the greatest safety and efficacy, tetravalent DENV vaccines are designed to stimulate balanced protective immunity to all four serotypes. However, this has been difficult to achieve. Clinical trials with a leading vaccine demonstrated that unbalanced replication and immunodominance of one vaccine component over others can lead to low efficacy and vaccine enhanced severe disease. The Laboratory of Infectious Diseases at the National Institutes of Health has developed a live attenuated tetravalent DENV vaccine (TV003), which is currently being tested in phase 3 clinical trials. Here we report, our study to determine if TV003 stimulate balanced and serotype-specific (TS) neutralizing antibody (nAb) responses to each serotype. Serum samples from twenty-one dengue-naive individuals participated under study protocol CIR287 (ClinicalTrials.gov NCT02021968) are analyzed 6 months after vaccination. Most subjects (76%) develop TS nAbs to 3 or 4 DENV serotypes, indicating immunity is induced by each vaccine component. Vaccine-induced TS nAbs map to epitopes known to be targets of nAbs in people infected with wild type DENVs. Following challenge with a partially attenuated strain of DENV2, all 21 subjects are protected from the efficacy endpoints. However, some vaccinated individuals develop post challenge nAb boost, while others mount post-challenge antibody responses that are consistent with sterilizing immunity. TV003 vaccine induced DENV2 TS nAbs are associated with sterilizing immunity. Our results indicate that nAbs to TS epitopes on each serotype may be a better correlate than total levels of nAbs currently used for guiding DENV vaccine development.

Fig. 1. DENV antibody depletion assay to estimate proportions of serotype-specific and cross-reactive-neutralizing antibodies in a subject who received TV003.

Subject 6 was tested 6 months after vaccination to measure levels of TS binding and neutralizing antibodies to each DENV serotype. The sample was incubated with immobilized bovine serum albumin (control depleted), DENV2 antigen (DENV2 depleted) and a mix of DENV1, 3 and 4 antigens (DENV1, 3, 4 depleted) to remove specific populations of antibodies. The DENV2 depletion resulted in the removal of DENV cross-reactive and DENV2 TS antibodies, while retaining any DENV1, 3, or 4 TS antibodies in the sample. The DENV1, 3, 4 depletion resulted in the removal of DENV cross-reactive and DENV1, 3, and 4 TS antibodies, while retaining any DENV2 TS antibodies in the sample. The sample depleted of specific antibody populations was tested by a ELISA to detect any DENV1, 2, 3, or 4 type specific binding antibodies. Data was analyzed from subject 6 and represented as bars from a single independent experiment and expressed as mean ODs with corresponding data points shown. b Neutralization test to measure levels of TS-neutralizing antibodies to each DENV serotype. Data was analyzed from subject 6 and represented as dots from a single independent experiment and expressed as 50% neutralization titers.

Fig. 2. Proportion and level of homotypic-neutralizing antibodies to each serotype in DENV naive subjects who received TV003.

The antibody depletion assay was used to measure proportions of DENV TS (homotypic)-neutralizing antibodies induced by TV003 in (n = 21) DENV naive subjects. a The percentage of the total-neutralizing antibody response to each serotype contributed by the serotype-specific antibodies for each subject. b The absolute level of homotypic-neutralizing antibody to each serotype for each subject. c Number and % of subjects that developed an absolute TS nab titer >20 is depicted. Data were analyzed from a total number of 21 subjects indicated as dots from a single independent experiment and expressed as mean titers ± 95% confidence intervals (a, b). Nonparametric one-way Anova with Friedman Dunn’s multiple comparisons test was used for determining the statistical differences between the level of TS nAbs across serotypes (b). Each p-value was adjusted to account for multiple comparisons. (****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05). The serotypes with significantly high levels of TS nabs compared to other serotypes is shown (b) (DENV3: **p = 0.0092; CI:148.4–481.5; DENV4: **p = 0.0061; CI: 172.9–469.8). The dotted line indicates the cut off for the absolute type specific titer.

Fig. 3. Characterization of DENV4 126 and 131 epitope transplant rDENV (rDENV2/4) used for mapping the DENV4 responses.

a Recombinant DENV2 containing EDI/II hinge, EDII residues from DENV4 that includes epitopes of the DENV4-126 and DENV4-131 mAbs. The figure depicts a model of DENV2 E protein dimer, with the mutated residues colored in blue. b Amino acid alignment of DENV2, rDENV2/4, DENV4, protein sequences in the EDI/II hinge; EDII regions with mutated residues in rDENV2/4 colored in blue. c Monoclonal antibodies (2D22, 3F9, 4J23: DENV2 TS mAbs, 5H2, 126,131: DENV4 TS mAbs, EDE1 C8, EDE2 B7 pan DENV specific mAbs) neutralization of rDENV2/4 compared to each parental strain’s is shown. Data shown here was from a single independent experiment and represented as mean neutralization titers. d Characterizing the neutralizing properties of DENV2 (n = 3) and DENV4 (n = 3) polyclonal natural infection sera against rDENV2/4 and the parental viruses. Data were obtained from a single independent experiment and expressed as mean titers.

Fig. 4. TV003 induce varying levels of DENV4 serotype-specific responses against known DENV4-neutralizing epitopes.

a DENV2 wt (green), DENV4 wt (blue), rDENV2/4 with DENV4-specific 126, 131 mAb epitopes transplanted into DENV2 (shown in blue) are depicted. b Panel of 19 sera (n = 19) with evidence of DENV4 TS responses were subjected to depletions against DENV2 to map the DENV4 TS nAb epitopes using the rDENV2/4 virus. This is a gain of function virus. Figure shows the 50% neutralization antibody titers for the BSA depleted and the DENV2 depleted serum samples against DENV2 wt, DENV4 wt and the rDENV2/4 viruses. Data were analyzed from a total number of 19 subjects indicated as dots from a single independent experiment and expressed as geometric mean titers ± 95% confidence intervals of the grouped samples (b).

Fig. 5. TV003 induce varying levels of DENV1 serotype-specific responses against a known DENV1-neutralizing epitope.

a DENV1wt (yellow), DENV2wt (green), rDENV2/1 with DENV1-specific 1F4 mAb epitope transplanted into DENV2 (shown in yellow) are depicted. b Panel of 8 TV003 sera (n = 8) with evidence of DENV1 TS responses were subjected to depletions against DENV2 to map the DENV1 TS nAb epitopes using the rDENV2/1 virus. Figure shows the 50% nAb titers for the BSA (control) depleted and the DENV2 depleted serum samples against DENV1 wt, DENV2 wt and the rDENV2/1 viruses. Data were analyzed from a total number of eight subjects indicated as dots from a single independent experiment and expressed as geometric mean titers ± 95% confidence intervals of the grouped samples (b).

Fig. 6. TV003 induce varying levels of DENV3 serotype-specific responses that do not track with known DENV3-neutralizing epitope.

a DENV3 wt (orange), DENV4 wt (blue), rDENV4/3 with DENV3-specific 5J7 mAb epitope transplanted into DENV4 (shown in orange) are depicted. b Panel of 8 TV003 sera (n = 8) with evidence of DENV3 TS responses were subjected to depletions against DENV4 to map the DENV3 TS nAb epitopes using the rDENV4/3 virus. Figure shows the Neut₅₀ titers for the BSA depleted and the DENV4 depleted serum samples against DENV3 wt, DENV4 wt and the rDENV4/3 viruses. Data were analyzed from a total number of eight subjects indicated as dots from a single independent experiment and expressed as geometric mean titers ± 95% confidence intervals of the grouped samples (b).

Fig. 7. TV003 induce varying levels of DENV2 serotype-specific responses against known DENV2-neutralizing epitope domain.

a DENV2 wt (green), DENV4 wt (blue) rDENV4/2 with the entire DENV2 EDIII domain transplanted into DENV4 (shown in green) are depicted. b Panel of 16 TV003 sera (n = 16) with evidence of DENV2 TS responses were subjected to depletions against DENV4 to map the DENV2 TS nAb epitopes using the rDENV4/2 virus. Figure shows the Neut₅₀ titers for the BSA depleted and the DENV4 depleted serum samples against DENV2 wt, DENV4 wt and the rDENV4/2 viruses. Data were analyzed from a total number of 16 subjects indicated as dots from a single independent experiment and expressed as geometric mean titers ± 95% confidence intervals of the grouped samples. Nonparametric one-way Anova with Friedman Dunn’s multiple comparisons test (b) was used for the epitope mapping data to determine the loss or gain in neutralization to the recombinant chimeric viruses compared to the wild type parental backbone viruses. Each p-value was adjusted to account for multiple comparisons. (****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05). The significant gain in neutralization of DENV2 TS nAbs against rDENV4/2 is indicated (***p = 0.0007; CI: 95.0–535.7).