Repeated exposure to dengue virus elicits robust cross neutralizing antibodies against Zika virus in residents of Northeastern Thailand

Zika virus (ZIKV) and dengue virus (DENV) are antigenically related mosquito-borne flaviviruses. ZIKV is becoming increasingly prevalent in DENV-endemic regions, raising the possibility that pre-existing immunity to one virus could modulate the response to a heterologous virus, although whether this would be beneficial or detrimental is unclear. Here, we analyzed sera from residents of a DENV-endemic region of Thailand to determine the prevalence of DENV-elicited antibodies capable of cross-neutralizing ZIKV. Sixty-one participants who were asymptomatic and unselected for viral serostatus were enrolled. Among them, 52 and 51 were seropositive for IgG antibody against DENV or ZIKV E proteins (ELISA assay), respectively. Notably, 44.23% (23/52) of DENV seropositive participants had serological evidence of multiple exposures to DENV, and these subjects had strikingly higher titers and broader reactivities of neutralizing antibodies (NAbs) against ZIKV and DENV heterotypes compared with participants with serological evidence of a single DENV infection (25/52, 48.1%). In total, 17 of the 61 participants (27.9%) had NAbs against ZIKV and all four DENV serotypes, and an additional 9 (14.8%) had NAbs against ZIKV and DENV1, 2, and 3. NAbs against DENV2 were the most prevalent (44/61, 72.1%) followed by DENV3 (38/61, 62.3%) and DENV1 (36/61, 59.0%). Of note, anti-ZIKV NAbs were more prevalent than anti-DENV4 NAbs (27/61, 44.3% and 21/61, 34.4%, respectively). Primary ZIKV infection was detected in two participants, confirming that ZIKV co-circulates in this region. Thus, residents of DENV-endemic regions with repeated exposure to DENV have higher titers of NAbs against ZIKV than individuals with only a single DENV exposure.


Results
Seroprevalence of anti-DENV and anti-ZIKV IgG. Blood samples were collected from 64 participants and tested for anti-DENV IgG and anti-ZIKV E IgG by ELISA. Of the 64 donors, 61 were healthy Thai subjects who resided in Nakhon Ratchasima, Thailand, and were unselected for prior virus exposure or serostatus. Blood samples were collected from these individuals between April and July 2016, at which time ZIKV had been documented in the region for more than a decade 2 . The remaining three blood samples (negative controls) were obtained from European donors who had never traveled to Southeast Asia.
Selected demographic features and the DENV and ZIKV serostatus of the 61 Thai subjects are shown in Table 1. The group consisted of 25 men and 36 women with a median age of 31 years (SD 9.9, range 18-69 years). Fifty-two subjects (85.3%) were seropositive for DENV IgG, indicating that most participants had been infected with DENV at least once ( Fig. 1A and Table 1). A breakdown of DENV seroprevalence indicated a positive association between prior exposure to DENV and age. Thus, 100% of participants aged 38 years and older were DENV seropositive (n = 21), compared with 80% of participants aged 28-37 years (n = 16) and 75% of those aged 18-27 years (n = 15) ( Table 1). A total of 31 women (86.1%) and 21 men (84.0%) were DENV seropositive, indicating a lack of significant association between sex and seroprevalence. The three control sera were DENV seronegative, as expected (Fig. 1A).
Fifty-one of the 61 (83.6%) Thai subjects were ZIKV seropositive (Table 1), and as observed for DENV serostatus, there was no significant association between sex and ZIKV seroprevalence, with 86.5% of men (31/36) and 80.0% of women (20/25) being ZIKV seropositive. ZIKV seropositivity was positively associated with age: all 21 (100%) participants aged ≥ 38 years were ZIKV seropositive, compared with 16/20 (80%) aged 28-37 years and 14/20 (70%) of those aged 18-27 years ( Fig. 1B and Table 1). Nine of the 61 individuals (4 men aged 19-31 years and 5 women aged 26-33 years) were seronegative for both DENV and ZIKV. Thus, of the 61 Thai subjects tested, 51 were DENV and ZIKV seropositive, 1 was seropositive for DENV only, and 9 were DENV and ZIKV naïve. There were no significant differences in the proportion of sera from the 25 men and 36 women positive for NAbs; thus, ~ 59%, ~ 72%, ~ 62%, ~ 34%, and ~ 44% of sera from both sexes contained NAbs against DENV1, DENV2, DENV3, DENV4, and ZIKV, respectively. Not surprisingly, the prevalence of NAbs against each DENV serotype and ZIKV increased significantly with age. For example, the proportion of participants in each age group    3). Among the 25 participants whose serology was consistent with primary DENV infection, most were DENV2 infections, followed by DENV1, and DENV3 (n = 13, 7, and 5, respectively). In contrast, none of the serological profiles were consistent with a primary DENV4 infection (Table 3). Only 4 of the 25 (16%) participants (no. 40, 34, 21, and 5) with primary DENV infection had high NAb titers against DENV, but they exhibited cross-neutralization of ZIKV (Table 3), suggesting a correlation between the presence of anti-ZIKV NAbs and a high level of anti-DENV NAbs. In contrast, 19 (82.6%) of the 23 participants with secondary DENV infection had anti-ZIKV NAbs. The four negative samples had low titers of NAbs against DENV1-4, suggesting that the donors mounted weak NAb responses to all flaviviruses/serotypes tested. Thus, nearly all of the participants showing evidence of repeated infection with DENV contain Abs with robust ZIKV-neutralizing capacity. Only two participants showed serological evidence of primary ZIKV infection. One of these (no. 15) had a low PRNT 90 for ZIKV and was negative for DENV NAbs, and the second (no. 30) had a very high titer of ZIKV NAbs and lower but marked neutralizing activity against all four DENV serotypes. These results indicate that repeat exposure to heterotypic DENV elicits a strong cross-NAb response against ZIKV.
ZIKV-neutralizing activity of DENV-elicited antibodies. The 61 sera were assigned to one of six groups according to their DENV1-4 and/or ZIKV NAb serostatus and analyzed for their reactivity to ZIKV NS1 and E proteins: group 1, DENV − ZIKV − (n = 9); group 2, DENV + (one to three serotypes) ZIKV -(n = 20); group 3, DENV + (one to three serotypes) ZIKV + (n = 12); group 4, DENV + (all four serotypes) ZIKV + (n = 19); and group 5, DENV − ZIKV + (n = 1). Notably, all of the DENV seropositive samples also contained anti-ZIKV E ( Fig. 4A) and/or anti-ZIKV NS1 ( Fig. 4B) Abs, and the binding patterns for ZIKV E and NS1 were similar for each group. Although group 2 Abs were defined as negative for anti-ZIKV NAbs, Abs against ZIKV E and NS1 antigens could be detected. This result might be due to cross-reactivity by anti-DENV Abs. The anti-ZIKV Ab endpoint titer of sera with NAbs against all four DENV serotypes (group 4) was not significantly higher than that of the sera with NAbs against fewer DENV serotypes (group 3) (P > 0.99), indicating that the anti-DENV NAb response influenced the magnitude of the cross-reactive anti-ZIKV Ab response. We then compared ZIKV cross-neutralizing activity between sera from the past primary and past secondary DENV infection groups, and we found that higher titers of ZIKV cross-NAbs correlated with past secondary DENV infection serostatus (Fig. 5). Collectively, these results indicate that both the quantity and quality of DENV and ZIKV NAbs are influenced by prior exposure to the heterologous virus/serotype.

Discussion
In this study, we sought to determine whether healthy residents of a DENV-endemic region in Thailand showed evidence of cross-NAbs to ZIKV. DENV has been endemic in the area since 1958 28 , and although there have been no outbreaks of ZIKV in Thailand, its presence has been documented since 2002 2 . These observations suggest that most residents of the region have likely been infected with DENV, if not ZIKV, at least once. This suggestion is supported by our results here showing that 52 of the 61 participants in our study were DENV seropositive, 25 of whom showed evidence of repeat infections based on NAbs. Similarly, 51 participants were ZIKV seropositive, 4 of whom were confirmed as having past primary or past secondary ZIKV infection based on NAbs. Thus, a substantial proportion of the participants showed evidence of exposure to DENV but not ZIKV, yet their sera contained strongly cross-neutralizing activity against ZIKV.  29 . The authors of that study suggested that protection against ZIKV arises from prior exposure to ZIKV, not DENV, based on the finding that Abs to ZIKV NS1 were detected only in serum samples with PRNT 90 ≥ 20 and not those with PRNT 90 < 20 for ZIKV NAbs 30 . This discrepancy could be due to the different regions in Thailand from which the samples were obtained, together with the higher circulation of DENV compared with ZIKV throughout the country. Although the population density of Nakhon Ratchasima is much lower than that of Bangkok (129 vs 3592 people/km 2 ), public health access and preventive measures such as mosquito nets are less accessible to residents of the more sparsely populated areas. Thus, the relative exposure to DENV is likely higher and exposure to ZIKV is likely lower for residents of Nakhon Ratchasima compared with Bangkok 2 . In support of this hypothesis, a 2019 study of field-caught mosquitoes by Phumee and colleagues found a lower prevalence of ZIKV-carrying mosquitoes in the Northeastern region of Thailand (which includes Nakhon Ratchasima) than in the Central region (including Bangkok) 31 . The results of the present study thus indicate that ZIKV seroprevalence levels in Thailand are most likely dependent on geography, and that multiple exposures to DENV of residents of the Northeastern region likely results in acquisition of ZIKV cross-NAbs.  32 . They found that patients with primary DENV infection showed no evidence of ZIKV NAbs, and only a low frequency of those with secondary DENV infection had ZIKV NAbs (23%, based on the IC 50 in a neutralization assay). That study was performed with samples collected ≥ 6 months after infection, suggesting that the NAb response had likely become more specific to a particular virus/serotype 32 . The discrepancy in findings between our study and that of Collins et al. may be due to differences in residency (Thailand vs United States) and/or the time elapsed after primary infection (undefined vs ≥ 6 months). Recent cohort studies of patients from Latin America and Asia, including Thailand, showed that convalescent sera collected ≥ 5 months after secondary DENV infection failed to cross-neutralize ZIKV 33 . In contrast, our results support the findings of a previous study that healthy but repeatedly infected residents of regions  www.nature.com/scientificreports/ hyperendemic for DENV can maintain a repertoire of broadly cross-NAbs against the four DENV serotypes 34 . Furthermore, our results are consistent with a previous cohort study of patients with confirmed dengue in Southern Vietnam, which showed that individuals with multiple exposures to DENV carried higher cross-NAb titers to ZIKV than individuals who had been exposed to DENV once 35 . A recent study of memory B cells from dengue patients in Nicaragua showed that DENV infection induced ZIKV cross-reactive memory B cells, particularly in patients with multiple exposures to DENV. Moreover, the ZIKV cross-neutralizing capacity of Abs generated by the memory response was higher than that of NAbs generated after primary DENV infection 36 . Another study suggested that cross-reactive immunity resulting from prior ZIKV exposure was responsible for the decline in dengue incidence in the Americas in 2017 37 . These results suggest that both the number and sequence of prior ZIKV and/or DENV serotype infections influence the risk of severe dengue disease 26 . Interestingly, recent study of ZIKV phylogenetic and in silico analysis indicated that while DENV and ZIKV share immunodominant B cell epitopes, ZIKA T-cell epitopes were derived from Culex-borne flaviviruses such as JEV, which might induce cross-protective T-cell responses against ZIKV. This could explain why ZIKA did not cause major epidemics in Thailand and other Asian countries where JEV is widespread 38 .
In summary, our study provides insights into the prevalence and potency of anti-DENV1-4 and ZIKV cross-NAbs in healthy adults living in a DENV-endemic region of Thailand. We found that an equally high proportion of participants (~ 85%) were DENV or ZIKV seropositive. Nearly all of the DENV seropositive individuals showed evidence of cross-NAbs against ZIKV and/or heterologous DENV serotypes, and about two-thirds of ZIKV-seropositive individuals had NAbs against at least one DENV serotype. All participants with secondary DENV infection contained anti-ZIKV E-and NS1-reactive Abs. Collectively, these results show that repeated exposure to DENV1-4 elicits an Ab response that strongly cross-neutralizes heterologous DENV serotypes as well as ZIKV. These observations not only have important implications for the surveillance of residents of DENV-endemic regions where ZIKV co-circulates, and vice versa, but also inform the design of safe, effective, and potentially cross-protective vaccines.

Methods
Study participants and sample collection. The study included blood samples collected from 61 healthy volunteers (25 men, 36 women, aged 18-69 years) who were seen at the Suranaree University of Technology (SUT) Hospital, Nakhon Ratchasima, Thailand. The inclusion criteria were no prior hospitalization or hospitalization > 5 years ago for DENV infection. All subjects were residents of Nakhon Ratchasima, were enrolled nonconsecutively between April and July of 2016, and provided a single (5 mL) blood sample. The cohort demographics are representative of the population in the Northeastern region of Thailand. As controls, blood samples were also obtained from three female healthy European volunteers (mean age 22 years) who had never traveled to Southeast Asia and were confirmed here to be ZIKV and DENV seronegative. The study was performed in accordance with the Declaration of Helsinki. The study protocol was approved by the Ethics Committee of SUT (protocol no. EC-59-10) and the Institutional Review Board of the La Jolla Institute for Immunology (protocol no. IB-189-1118). Written informed consent was obtained from all participants before blood collection. Organization reference strains DENV1 16007, DENV2 16681, DENV3 16562, and DENV4 1036 were kindly provided by Dr. Duncan R. Smith (Mahidol University, Thailand). ZIKV Asian lineage strain SD001 was isolated from a donor who had visited Venezuela in 2016 39 . All viral stocks were propagated in Aedes albopictus C6/36 cells (ATCC no. CRL-1660). Virus-containing supernatants were clarified by low-speed centrifugation, concentrated by ultracentrifugation, and supplemented with fetal bovine serum to a final concentration of 20%. All viral stocks were stored at − 80 °C. Viral titers were determined using a focus forming assay with baby hamster kidney (BHK)-21 cells (ATCC no. CCL-10) as previously described 40 . In vitro neutralization assay. To quantify anti-DENV1-4 and anti-ZIKV neutralizing Abs (NAbs), a plaque reduction neutralization test (PRNT) was performed with LLC-MK 2 cells as previously described 29 . For ZIKV NAbs, the standard PRNT was modified by using DMEM (Capricorn Scientific, Germany) containing 1% FBS and 0.75% NaHCO 3 as the sample and virus diluent solution. Inoculated cells were gently overlaid with 2X nutrient and 1.6% UltraPure agarose (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA) and maintained at 37 °C in a 5% CO 2 atmosphere for 7 days. Plaque numbers after incubation with (test) or without (control) test sera were counted and the percentage neutralization was calculated as 100 × ([control plaque number -test plaque number]/control plaque number). The World Health Organization recommendation for DENV PRNT suggests that an PRNT 90 cutoff value should be used for samples obtained in flavivirus-endemic areas 41 . For both viruses, PRNT 90 was defined as the reciprocal of the serum dilution resulting in a 90% reduction of viral infectivity. PRNT 90 was calculated by non-linear regression using Prism 8 software (GraphPad, La Jolla, CA). DENV and ZIKV ELISA. Serum levels of IgG against DENV E antigen serotypes (mixture of 1, 2, 3, and 4) were quantified using an indirect ELISA kit (E-DEN-01G, Panbio, Brisbane, QLD, Australia), according to the manufacturer's instructions. DENV-specific IgG is expressed as Panbio units, which were calculated as: ([optical density; OD value of the test specimen]/[mean cutoff calibrator OD value] × 10). According to the kit manufacturers' recommendations, samples with > 11 Panbio units were considered positive. Serum levels of IgG against recombinant ZIKV E and NS1 proteins were quantified by ELISA, as previously described 40 . Anti-ZIKV IgG levels are expressed as the endpoint titer, calculated as the reciprocal of the highest serum dilution giving an OD value ≥ 2-fold the OD value of the blank sample (bovine serum albumin in phosphate-buffered saline). Anti-ZIKV E IgG positivity was defined as endpoint titers > 1 standard deviation (SD) above the value for the serum negative control value.