Zika seroprevalence declines and neutralizing antibodies wane in adults following outbreaks in French Polynesia and Fiji

Abstract

It has been commonly assumed that Zika virus (ZIKV) infection confers long-term protection against reinfection, preventing ZIKV from re-emerging in previously affected areas for several years. However, the long-term immune response to ZIKV following an outbreak remains poorly documented. We compared results from eight serological surveys before and after known ZIKV outbreaks in French Polynesia and Fiji, including cross-sectional and longitudinal studies. We found evidence of a decline in seroprevalence in both countries over a two-year period following first reported ZIKV transmission. This decline was concentrated in adults, while high seroprevalence persisted in children. In the Fiji cohort, there was also a significant decline in neutralizing antibody titres against ZIKV, but not against dengue viruses that circulated during the same period.

Keywords: Zika; antibody response; dengue; epidemiology; global health; human; infectious disease; microbiology; seroprevalence.

Figure 1.. Dynamics of ZIKV seroprevalence following outbreaks in Fiji and French Polynesia.

(A) Seroprevalence by MIA in Fiji. Red, seroprevalence and 95% confidence intervals for children (aged ≤16 years). Orange, seroprevalence and 95% confidence intervals for adults (aged >16 years). Solid lines, trends in data collected from the same individuals. Dotted line indicates the first confirmed ZIKV case. (B) Epidemiological dynamics in Fiji between 2013 and 2018. Coloured bars show number of PCR-confirmed samples of different DENV serotypes and ZIKV in Fiji; black lines show reported prolonged fever in Fiji from the Pacific Syndromic Surveillance System (World Health Organization, 2019). There was a major outbreak of DENV-3 outbreak in 2013–14 (Kucharski et al., 2018a) with a smaller DENV-2 outbreak in early 2017 (Aubry et al., 2019). (C) Seroprevalence by MIA in French Polynesia. Dashed lines, trends in seroprevalence between population representative cross-sectional surveys. Note that the pre-outbreak samples were collected between July 2011 and October 2013; for brevity, the latest possible collection date is used in the plot. (D) Epidemiological dynamics in French Polynesia between 2013 and 2018. Solid black line shows reported symptomatic dengue cases; dashed lines showed reported symptomatic Zika cases. In French Polynesia, between the sampling periods, there were no reported DENV outbreaks for serotypes 2,3,4, and there was hyper-endemic DENV-1 circulation. In April 2019, a DENV-2 outbreak was declared, the first since 1997 (Aubry et al., 2019).

Figure 1—figure supplement 1.. Seroprevalence against DENV-1 in Fiji and French Polynesia, by age group.

Figure colour scheme and data characteristics are same as in Figure 1.

Figure 1—figure supplement 2.. Seroprevalence against DENV-2 in Fiji and French Polynesia, by age group.

Figure colour scheme and data characteristics are same as in Figure 1.

Figure 1—figure supplement 3.. Seroprevalence against DENV-3 in Fiji and French Polynesia, by age group.

Figure colour scheme and data characteristics are same as in Figure 1.

Figure 1—figure supplement 4.. Seroprevalence against DENV-4 in Fiji and French Polynesia, by age group.

Figure colour scheme and data characteristics are same as in Figure 1.

Figure 2.. Waning of neutralizing antibody responses against ZIKV and DENV-3 in Fiji for participants who were seronegative to each virus in 2013 and seroconverted in 2015.

(A) Histogram of change in neutralization assay log titre against DENV-3 (n = 19) and ZIKV (n = 31) between 2015–2017 for individuals who seroconverted to these respective viruses between 2013–2015 (i.e. log titre < 2 in 2013 and log titre ≥ 2 in 2015). (B) Histogram of change in log titre against DENV-3 and ZIKV for individuals who seroconverted to these respective viruses during 2013–2015.

Figure 2—figure supplement 1.. Waning of neutralizing antibody responses against ZIKV and DENV-3 in Fiji for participants who had a four-fold rise between 2013 and 2015.

Histogram of change in neutralization assay log titre against DENV-3 (n = 25) and ZIKV (n = 35) between 2015–2017 for individuals who had a rise in log titre of at least two to these respective viruses between 2013–2015. (B) Histogram of change in log titre between 2013–2015 against DENV-3 and ZIKV for individuals who had a rise of at least two during this period.

Figure 2—figure supplement 2.. Waning of neutralizing antibody responses against DENV-1 and DENV-2 in Fiji for participants who were seronegative in 2013 and seroconverted in 2015.

Histogram of change in neutralization assay log titre against DENV-1 (n = 26) and ZIKV (n = 18) between 2015–2017 for individuals who seroconverted between 2013–2015 (i.e. log titre <2 in 2013 and log titre ≥2 in 2015). (B) Histogram of change in log titre against DENV-1 and DENV-2 for individuals who seroconverted to these respective viruses during 2013–2015.

Figure 3.. Relationship between mean DENV log neutralization titre across the four serotypes in 2013 and ZIKV seroprevalence using different assays, in a subset of 45 participants.

(A) Seroprevalence by MIA, shown in grey, and neutralization test (NT), shown in orange, for sera collected in 2013. Line shows prediction from GAM fitted to each dataset, with shaded region showing 95% CI, and points show raw data. (B) Seroprevalence for sera collected from the same participants in 2015. (C) Seroprevalence for sera collected from the same participants in 2017.

Figure 4.. Distribution of ZIKV neutralization titres in the Fiji serosurveys.

Results shown for 45 participants who had samples available from 2013, 2015, and 2017. Dashed line shows the threshold used to define seropositivity.

Figure 4—figure supplement 1.. Individual-level neutralization log titres against the four DENV serotypes and ZIKV in Fiji.

Points show assay results in the 2013, 2015 and 2017 sample collections for each participant, coloured by virus (n = 45).

Figure 4—figure supplement 2.. Correlation between rise in DENV and ZIKV neutralization log titres between 2013–2015 for participants who were initially seronegative (i.e. log titre <2) to all five viruses in 2013 (n = 20).

There is significant correlation between DENV-1 and DENV-3 viruses (top row, p=0.0012), suggesting likely cross-reactive responses. However, changes in ZIKV titres were not associated with responses to any of DENV viruses (far right column), which strongly indicates that the ZIKV results were genuine infections..

Figure 4—figure supplement 3.. Correlation between change in DENV and ZIKV log titres between 2013–2015 for participants who were initially seropositive (i.e. log titre ≥2) to at least one DENV virus in 2013 (n = 49).

There is significant correlation between ZIKV and other DENV viruses, suggesting likely cross-reactive responses. However, there was limited circulation of viruses such as DENV-1 and DENV-2 during 2013–15 (Figure 1), suggesting that it was infection with ZIKV that generated a cross-reactive response against these viruses, rather than the other way around.