Isolation of DENV4 neutralizing human monoclonal antibodies.

To further characterize the B cell response to DENV4 infection, we transformed B cells from two DENV4-immune individuals (subjects 002 and 112) and isolated hMAbs, as previously described (27). The transformed B cell culture supernatants were screened for binding to DENV4. On the basis of the number of positive wells and the number of transformed B cells tested (determined by average colony counts in transformed wells), the frequencies of DENV-specific B cells in circulation were estimated to be 0.19% and 0.20% of transformable B cells for subjects 002 and 112, respectively. Previously, it was reported that there may be a long-term set point frequency of about 0.1% to 0.2% DENV-specific B cells in the circulating memory B cell pool following DENV infection (28), and the frequencies of 0.19% and 0.20% in the two subjects studied here are consistent with those previous estimates. Note that, even though the frequencies of DENV4-specific B cells determined for subjects 002 and 112 were similar, their serum titers were different. As mentioned above, the B cell response to primary DENV infection is a mix of DENV serotype cross-reactive and type-specific antibodies. Moreover, a relatively small fraction of the virus-specific B cells produce neutralizing Abs (29,–31). The serum neutralizing Abs are derived from long-lived plasma cells (LLPCs) in the bone marrow and not the memory B cell compartment. Previous studies have also reported a weak and inconsistent correlation between DENV-specific or other viral antigen-specific antibody titers in the serum and the frequency of antigen-specific memory B cells in the blood (32,–34).

We also determined the DENV serotype specificity of all the positive B cell culture supernatants from subject 112. Of the 34 DENV antigen-reactive supernatants, antibodies in 32% bound only to DENV4 (type-specific) and in 68% bound to two or more serotypes (cross-reactive). From the EBV-transformed B cell lines secreting DENV antigen-reactive antibodies, we isolated 8 human B cell hybridoma cell lines, as previously described (28). Two cell lines, designated D4-126 and D4-131, secreted DENV4 type-specific and neutralizing hMAbs with Neut₅₀ values (defined as the dilution factor required to neutralize 50% of infection) of 0.54 μg/ml and 0.43 μg/ml, respectively (Fig. 2A). Both of these antibodies were isolated from subject 002. To characterize the binding properties of D4-126 or D4-131 hMAbs further, we performed binding assays with whole DENV4 virions, recombinant E (rE) or rEDIII proteins of DENV4, and increasing concentrations of hMAb D4-126 or D4-131. The two hMAbs bound to whole DENV4 particles similarly (Fig. 2B). HMAb D4-126 did not bind to rE protein, whereas D4-131 exhibited low levels of binding to rE protein at high concentrations (>10 μg/ml) (Fig. 2B). The hMAbs did not bind to rEDIII protein (Fig. 2B). These studies revealed that neutralizing hMAb D4-126 or D4-131 bound best to intact DENV4 virions. However, it is also possible that MAb D4-126 or D4-131 could bind to E dimers formed from rE monomers at higher concentrations as demonstrated by the D4-131 MAb binding to rE at higher concentrations.

Open in a separate window

FIG 2

DENV4 neutralizing hMAb D4-126 or D4-131 preferentially binds to epitopes displayed on intact virions. (A) U937+DC-SIGN flow cytometry-based neutralization assays were conducted to determine the neutralization of DENV1 to DENV4 by D4-126 or D4-131. Results are from 2 technical replicates in an experiment. Each point represents the mean neutralization value from the two replicates, and the error bars depict the standard deviations. Neutralization experiments were repeated at least twice. (B) Binding to DENV4 whole virions, recombinant E protein ectodomain (rE), and recombinant E protein domain III (rEDIII) by D4-126 or D4-131 was tested in an ELISA. Results are from 2 technical replicates in one experiment. Each point represents the mean from the two replicates, and the error bars depict the standard deviations OD, optical density.

Additionally, these antibodies were not cross-reactive to the Zika virus (H/PF/2013 and MR 766 strains) and remained type specific to DENV4.

hMAb neutralization of different DENV4 strains.

To determine if hMAbs D4-126 and D4-131 neutralized diverse strains of DENV4, we used a panel of recombinant isogenic DENV4 expressing the E protein from different DENV4 genotypes and laboratory strains (Table 3). The hMAbs equally neutralized all variants tested, except for a Cambodia 2010 genotype I (GI) strain, which was not neutralized by D4-126 and was neutralized less efficiently by D4-131 than by other strains tested (Fig. 3).

Open in a separate window

FIG 3

Neutralization of different DENV4 strains by hMAb D4-126 or D4-131. U937+DC-SIGN flow cytometry-based neutralization assays were conducted to compare the abilities of D4-126 (A) and D4-131 (B) hMAbs to neutralize different laboratory-adapted or clinical strains of DENV4. The MAbs neutralized all variants except for the Cambodia 2010 genotype I strain, which was not neutralized by D4-126 and was neutralized less efficiently by D4-131 than by the other strains. Results are from 2 technical replicates in one experiment. Each point represents the mean neutralization value from the two replicates, and the error bars depict the standard deviations.

Mapping the epitopes of DENV4 neutralizing hMAbs.

Human DENV1, DENV2, and DENV3 type-specific neutralizing antibodies often bind to quaternary structure epitopes centered on the EDI/II hinge (17,–19, 21, 38) and/or the EDIII region (2, 20, 21, 39, 40). Recently, we demonstrated that it is possible to recover recombinant chimeric DENVs displaying E protein domains or epitopes from viruses of two different serotypes (38, 39). We used a rDENV4 strain with a mutated EDI/II hinge region and EDII region (rDENV4/3) to map the binding sites of hMAbs D4-126 and D4-131 (Fig. 4A and ​andB).B). The recombinant virus has 25 amino acid replacements at different positions that are highlighted in green (Fig. 4A and ​andB)B) (unpublished data). We did not detect any binding or neutralizing activity for the hMAb D4-126 or D4-131 with the rDENV4/3, indicating that the DENV4 EDI/II hinge residues are part of the epitope recognized by these MAbs (Fig. 4C and ​andDD).

Open in a separate window

FIG 4

The EDI/II hinge region is important for the binding and neutralization of DENV4 by hMAbs D4-126 and D4-131. (A) Amino acid alignment of DENV4 and rDENV4/3 E protein sequence in the EDI/II hinge; EDII regions with mutated residues are colored in green. (B) Recombinant DENV4 containing EDI/II hinge; EDII residues from DENV3 (rDENV4/3) used to map the epitopes of hMAbs D4-126 and D4-131 are shown. The figure depicts a model of the DENV4 E protein dimer, with the mutated residues colored in green. (C) Binding of hMAbs D4-126 and D4-131 to DENV4, DENV3, or rDENV4/3. (D) DENV4 and rDENV4/3 neutralization by hMAbs D4-126 and D4-131. Results are from 2 technical replicates in an experiment. Each point represents the mean from the two replicates, and the error bars depict the standard deviations. Mapping of the critical residues for D4-126 and D4-131 was performed by shotgun mutagenesis by screening for loss of MAb binding to a DENV4 alanine scan mutation array. Critical residues are visualized as magenta spheres on the DENV E dimer structure (PDB identifier [ID] 1OAN). (E) Six critical residues were identified for D4-126 in the E protein DII and DI/II hinge region. (F) Four critical residues were identified for D4-131 in the E protein DII and DI/II hinge region.

As an alternative approach to mapping the epitopes of D4-126 and D4-131, both hMAbs were screened by shotgun mutagenesis against a comprehensive mutation library in which nearly every residue within the precursor membrane (prM) and E was individually mutated to alanine, as described previously (35). Residues were identified as critical to binding of the DENV4 hMAb if they did not bind to the DENV4 hMAb but did bind to other conformation-dependent MAbs. Six amino acids (K51, V53, K124, L135, K200, and K234) in the EDI/II hinge and EDII regions were critical for binding of D4-126 (Fig. 4E). K51, V53, K124, and K200 were among the residues that were altered in rDENV4/3. Four amino acids (K51, K124, K200, and K202) within the EDI/II hinge and EDII region were critical for binding of D4-131 (Fig. 4F). K51, K124, and K200 were among the residues that were altered in rDENV4/3. These data validated our observations that the EDI/II hinge region residues are critical for binding of D4-126 and D4-131 and also indicated that the epitopes differ slightly between the two hMAbs.

Ethics statement.

Convalescent DENV immune blood samples were collected from travelers visiting countries in which dengue is endemic. The protocol for obtaining dengue immune blood samples was approved by the Institutional Review Board of the University of North Carolina at Chapel Hill (protocol 08-0895). Written informed consent was obtained from all subjects before their participation in the study.

DENV immune sera.

DENV serum samples used in this study were obtained from the previously existing dengue traveler collections as mentioned above. The samples were not from acute clinical patient samples and hence were not PCR confirmed. DENV4 immune sera were also obtained from people who had received a live attenuated monovalent DENV4 vaccine (25, 26) developed by the US National Institutes of Health (NIH) and were provided by Anna Durbin and Stephen Whitehead. All samples were coded and analyzed anonymously.

Viruses, rE, and rEDIII.

DENV1 (American genotype; strain West Pac74), DENV2 (Asian genotype; strain S-16803), DENV3 (Asian genotype; strain CH-53489), and DENV4 (American genotype; strain TVP-376) (provided by Robert Putnak, Walter Reed Army Institute of Research, Silver Spring, MD) were used for both binding enzyme-linked immunosorbent assays (ELISAs) and neutralization assays. All viruses used in the neutralization assays were grown in C6/36 Aedes albopictus mosquito cells at 32°C, as previously described (48). DENV4 was purified in our laboratory as previously described (16). Purified live DENV2 was purchased from Microbix Biosystems, Inc. (Mississauga, Ontario, Canada). Recombinant envelope (rE) proteins (80% of E protein) from each of the four serotypes were produced within our laboratory or purchased from Hawaii Biotech, Inc. (49). Recombinant EDIII proteins were produced within our laboratory as described previously (50).

Depletion of DENV4-specific antibodies from human immune sera collected from subjects with prior DENV4 infection or vaccination.

Purified DENV was absorbed onto 4.5-μm-diameter Polybead polystyrene microspheres (Polysciences, Inc.) at a bead (microliters) to ligand (micrograms) ratio of 5:2. Polystyrene beads were washed three times with 0.1 M borate buffer (pH 8.5) and incubated with the relevant purified DENV (DENV4 for homotypic depletions and DENV2 for heterotypic depletions) overnight at room temperature (RT). Control beads were incubated overnight with an equivalent amount of bovine serum albumin (BSA). The control and virus-adsorbed beads were blocked with BSA (10 mg/ml)–borate buffer for 30 min at RT three times and washed four times with phosphate-buffered saline (PBS). DENV4 immune sera from naturally infected individuals or NIH vaccine candidate recipients were depleted of virus-specific antibodies by incubating the samples with virus-adsorbed beads for 1 h at 37°C with end-over-end mixing. Samples were subjected to at least three sequential rounds of depletions before successful removal of the respective antibodies was confirmed by ELISA. The ability of the depleted samples to neutralize viruses of all of the four serotypes was tested after the confirmation ELISA.

Generation of DENV4-specific MAbs.

Previously cryopreserved peripheral blood mononuclear cells (PBMCs) isolated from blood samples collected as part of the dengue traveler studies were thawed rapidly in a 37°C water bath and washed prior to transformation with Epstein-Barr virus (EBV) contained in the clarified supernatants from cultures of B95.8 cells (ATCC VR-1492) and incubated with CpG and additional supplements, as described previously (27, 51). Cultures were incubated at 37°C with 5% CO₂ for 10 days prior to screening for DENV4-reactive cell lines with ELISA. The minimal frequency of DENV4-reactive B cells was estimated on the basis of the number of wells with DENV4-reactive supernatants compared to the total number of lymphoblastoid cell line (LCL) colonies in the transformation plates as follows: (number of wells with DENV4-reactive supernatants)/(number of LCL colonies in the plate). Cells from wells with supernatants reacting in the DENV4 capture ELISA were subjected to cytofusion, as previously described (27, 51). Following cytofusion, hybridomas were selected for growth in hypoxanthine-aminopterin-thymidine (HAT) medium containing ouabain. Wells containing hybridomas producing DENV4-reactive antibodies were cloned biologically by 3 rounds of limiting dilution plating. Once clonal, the cell lines were used to produce MAb immunoglobulin G (IgG) in cell supernatants, using serum-free medium, followed by protein G column purification.

Virus, rE, and rEDIII ELISA.

Equivalent quantities of DENV (as previously titrated by ELISA) were captured by anti-E mouse MAb 4G2, or rE proteins were directly coated (rE, 100 ng/well; rEDIII, 200 ng/well) on ELISA plates overnight at 4°C (39). Plates were blocked with 3% (vol/vol) normal goat serum (Gibco, Thermo Fisher, USA)–Tris-buffered saline (TBS)–0.05% (vol/vol) Tween 20 (blocking buffer). Primary antibodies were diluted serially to generate a range of concentrations. Alkaline phosphatase-conjugated secondary antibodies were used to detect binding of primary antibodies with p-nitrophenyl phosphate substrate, and reaction color changes were quantified by spectrophotometry.

Blockade of binding assays.

Assays for blockade of binding were performed as described previously (52, 53). Briefly, DENV4 was captured using mouse anti-E MAb 4G2 and was blocked as described above. Serial dilutions of DENV serum were added to the DENV4-coated plates and incubated at 37°C for 1 h. The plates were washed, and alkaline phosphatase-conjugated DENV4 hMAbs were added and incubated at 37°C for 1 h. P-nitrophenyl phosphate substrate was added, and reaction color changes were quantified by spectrophotometry. Percentages of blockade of binding were calculated as follows: [100 − (optical density of sample/optical density of control) × 100].

Flow cytometry-based U937+DC-SIGN neutralization assay.

Neutralization potentials of the DENV4 immune sera and hMAbs were measured using a flow cytometry-based neutralization assay with U937+DC-SIGN cells as previously described (17). Briefly, virus and antibody mixtures or sera were incubated for 1 h at 37°C, prior to the addition of U937+DC-SIGN cells. After 2 h of incubation, cells were washed twice with infection media. Cells then were fixed and permeabilized 24 h after infection and probed with 2H2 (anti-prM MAb) conjugated to Alexa-Fluor 488, and the percentage of infected cells was determined using a Guava flow cytometer (FC) (EMD Millipore). Stained cells were analyzed to calculate 50% neutralization titers.

Construction of rDENV strains.

rDENV4/3 was generated as previously described (39). Briefly, the DENV4 genome (Sri Lanka 1992; GenBank accession no. KJ160504.1) (36) was split into 4 separate plasmids (plasmids A, B, C, and D), allowing production of genomic cDNA. Plasmids were digested, and genome fragments were ligated together into a full-length cDNA genome from which RNA transcripts were derived. These transcripts were electroporated into cells, and cell culture supernatant containing viable virus was harvested. Virus was passaged two times on C6/36 cell monolayer cultures and stored at −80°C. To generate rDENV, the nucleotide sequence of the E glycoprotein was changed to alter the amino acid residues. rDENV4/3 contains EDI/II hinge residues (with 25 amino acids altered) from DENV3 (Fig. 4A) (Sri Lanka 1989, designated UNC3001; NCBI accession no. JQ411814.1) (43).

Generation of DENV4 strains displaying diverse E glycoproteins.

In order to examine genetic diversity within a serotype, a panel of near-isogenic rDENV4 was generated by replacing the E gene of WT genotype II infectious clone virus (Sri Lanka 1992; accession no. KJ160504.1) (36) with E glycoprotein genes representing diverse strains within DENV4. Subgenomic A plasmids encoding E protein genes only were synthesized (all other proteins were encoded by genes based on DENV Sri Lanka 1992) representing genotype I (GI) (Cambodia 2010; NCBI accession no. KF543272.1) (V. Duong, S. Lay, and P. Buchy, unpublished data) or genotype II (GII) (Puerto Rico 1999; NCBI accession no. FJ882599.1), available in GenBank database (http://www.ncbi.nlm.nih.gov/nuccore/228539113) or a sylvatic E protein sequence (Malaysia 1973; NCBI accession no. JF262780) (37). Recombinant subgenomic A plasmids were synthesized, and then viral assembly and rescue were performed as described above for generation of rDENV4.

We thank all the dengue-immune travelers who participated in the study. We also thank members of the de Silva laboratory for their assistance.

This research was supported by funding from U.S. National Institutes of Health grants R01 AI107731, (principal investigator [PI], Aravinda M. de Silva), R01 AI125198 (PI, Aravinda M. de Silva), U19 AI109761 (PI, Ralph S. Baric), and P01 AI106695 (PI, Eva Harris) and the Bill and Melinda Gates Foundation (PI, Anna P. Durbin).