Antibody Responses to Zika Virus Infections in Environments of Flavivirus Endemicity

Abstract

Zika virus (ZIKV) infections occur in areas where dengue virus (DENV), West Nile virus (WNV), yellow fever virus (YFV), and other viruses of the genus Flavivirus cocirculate. The envelope (E) proteins of these closely related flaviviruses induce specific long-term immunity, yet subsequent infections are associated with cross-reactive antibody responses that may enhance disease susceptibility and severity. To gain a better understanding of ZIKV infections against a background of similar viral diseases, we examined serological immune responses to ZIKV, WNV, DENV, and YFV infections of humans and nonhuman primates (NHPs). Using printed microarrays, we detected very specific antibody responses to primary infections with probes of recombinant E proteins from 15 species and lineages of flaviviruses pathogenic to humans, while high cross-reactivity between ZIKV and DENV was observed with 11 printed native viruses. Notably, antibodies from human primary ZIKV or secondary DENV infections that occurred in areas where flavivirus is endemic broadly recognized E proteins from many flaviviruses, especially DENV, indicating a strong influence of infection history on immune responses. A predictive algorithm was used to tentatively identify previous encounters with specific flaviviruses based on serum antibody interactions with the multispecies panel of E proteins. These results illustrate the potential impact of exposure to related viruses on the outcome of ZIKV infection and offer considerations for development of vaccines and diagnostics.

Keywords: Zika; cross-reactivity; flavivirus; humoral immunity; protein microarray.

FIG 1

Phylogenetic relationships and recognition of microarrayed antigens by virus-specific antibody standards. (A) The phylogenies of flaviviruses examined in this study were inferred from an alignment of amino acid sequences from envelope (E) proteins. (B) Microarrays of E, nonstructural protein 1 (NS1), and premembrane (pM) proteins probed with mouse polyclonal antibodies generated against each virus shown (centered labels above each row of bar graphs). Antibody binding data are shown as log₁₀-transformed mean fluorescence intensities (±standard errors of the means [SEM] [error bars]), and the arrows indicate the virus-specific antigens. Heterologous antigens that exhibit increased recognition compared to the virus-specific antigen are labeled with an asterisk (P < 0.05, one-way ANOVA with Tukey's range test). Virus abbreviations: YFV, yellow fever virus; SLEV, St. Louis encephalitis virus; DENV, dengue virus; DENV1, dengue virus serotype 1; POWV, Powassan virus; TBEV-E, tick-borne encephalitis virus, Eastern strain; TBEV-EUR, tick-borne encephalitis virus, European strain; MVEV, Murray Valley encephalitis virus; WNV, West Nile virus; ZIKV, Zika virus; ZIKV-AFR, ZIKV from Africa; ZIKV-AS, ZIKV from Asia; JEV, Japanese encephalitis virus; ROCV, Rocio virus.

FIG 2

Specificity and kinetics of the humoral immune response to ZIKV. (A and B) ZIKV-challenged nonhuman primate (NHP) IgG recognition of ZIKV particles harvested early (48 h) or late (144 h) postinfection of HEK293T cells (A) and ZIKV proteins (envelope [E], nonstructural protein 1 [NS1], and premembrane protein [pM]) (B) from five Asian (AS) and six African (AFR) lineages (Table 1). ZIKV-specific antibody responses are denoted by scatter plots with center horizontal lines representing the mean binding of serum antibodies from NHPs challenged with either an AFR (n = 3) (circles) or AS (n = 3) (squares) lineage ZIKV at 0 to 2 days postinfection (dpi) (open symbols) and 21 to 28 dpi (filled symbols). Error bars indicate SEM. Statistically significant differences between mean antibody binding of all ZIKV-challenged NHPs to ZIKV antigens at 0 to 2 dpi and 21 to 28 dpi were calculated using a one-tailed Student's t test (*, P < 7.5e−5; ns, not significant), while no significant differences were observed between mean antibody binding of ZIKV-AS- and ZIKV-AFR-challenged groups to AS and AFR ZIKV antigens at 21 to 28 dpi (two-tailed Student's t test). (C) IgM and IgG binding profiles to ZIKV particles (harvest at 144 h) and ZIKV E protein are compared to viral load (Zika Open-Research Portal [https://zika.labkey.com]) from preinfection (day 0) to 28 dpi for ZIKV-challenged NHPs (n = 9). Second-order (IgM), third-order (IgG), and fourth-order (viral load) polynomial curves were fitted to the data, with fitted lines and shading under the curve consistent with data point colors.

FIG 3

Differentiation of nonhuman primates challenged with ZIKV or DENV by specific IgG binding to E antigens. (A) Binding of convalescent-phase serum antibodies from nonhuman primates (NHPs) challenged with either an Asian (H/PF) (n = 3) (red) or African (MR-766) (n = 3) (royal blue) lineage ZIKV, or DENV (n = 4 each for the DENV1 [black], DENV2 [green], DENV3 [orange] groups; n = 3 for the DENV4 group [magenta]) to whole viruses (144 h) and E proteins. Values shown are antibody binding signals relative to the virus used for challenge (±SEM). (B) Principal-component analyses of relative IgG binding to E proteins and viruses (144 h) by NHP antibodies. Individual data points and virus-specific clusters are colored according to the challenge virus as in panel A. PC1, principal component 1.

FIG 4

Antibody specificity of primary and secondary flavivirus infections. Relative binding (±SEM) of convalescent-phase serum antibodies from nonhuman primate (NHP) and human flavivirus infections to 15 flavivirus E proteins is shown. (A) Sera from primary infections are indicated by color as follows: gray, DENV-challenged NHPs (individual data for each NHP group are overlaid in a scatter plot; n = 4 each for the DENV1 [black], DENV2 [green], and DENV3 [orange] groups and n = 3 for the DENV4 group [magenta]); green, human (Hu) rDEN2Δ30 (n = 8) (primary infection); red, pooled African and Asian lineage ZIKV NHPs (n = 6); white, YFV-vaccinated NHPs (n = 3). (B) Sera from confirmed human flaviviral infections with unknown infection histories are indicated by color as follows: gray, DENV (individual data are overlaid in a scatter plot; the colors correspond to the most recent DENV infection); green, DENV2 (n = 5); orange, DENV3 (n = 2); red, ZIKV (n = 4); white, YFV vaccination (n = 13); cyan, WNV (n = 20). (C) Predicted infection histories of human secondary DENV (gray in panel B) and primary ZIKV (red in panel B) infections, based on a supervised SVM classifier. Individual human sera are shown at the bottom (Z for ZIKV, D2 for DENV2, and D3 for DENV3; virus followed by serum identification [ID] number), with probability values for each viral class (left) gradient colored from low to high (white to royal blue) (right). Predicted infection histories are designated by colored bars above serum ID (DENV1 [black], DENV4 [magenta], no prediction [no bar]).

FIG 5

Quantitative comparisons of antibodies directed to the infecting virus versus all other flaviviruses. Antibody recognition of microarrayed E proteins displayed as mean fluorescence intensity (±SEM). Antibodies from primary flavivirus infections of NHPs (ZIKV, DENV1 to DENV4, and YFV) and humans (rDEN2Δ30) exhibited significantly decreased recognition of heterologous E antigens compared to virus-specific E (dark gray) (P < 0.05 by one-way ANOVA with Tukey's range test). DENV E proteins are separated from all other flavivirus E proteins (including YFV, SLEV, POWV, TBEV, MVEV, WNV, JEV, and ROCV) to show DENV antibody cross-reactivity between serotypes.

FIG 6

Overlap in rising and waning antibody responses to independent infections. The primary infection of a flavivirus-naive individual with dengue virus occurs at day 0 (solid black arrowhead). Levels of virus-specific antibody (gray bars and shading) begin to increase shortly after the acute phase of infection, peak after convalescence, and subside thereafter. A second infection with Zika virus (solid red arrowhead) is followed by an increase in virus-specific antibody (red bars and shading), resulting in detection of a mixture of anti-dengue virus and anti-Zika virus antibodies that will vary with time from infections. The ratio of dengue virus-to-Zika virus antibodies, as shown, will be further increased if the secondary infection results in a less potent activation of serological immune responses.