Antigenic Relationships between Sylvatic and Endemic Dengue Viruses

Sylvatic dengue viruses (DENVs) are both ecologically and evolutionary distinct from urban DENVs and are transmitted in an enzootic cycle most likely between non-human primates and arboreal Aedes spp. mosquitoes. In Southeast Asia, isolations of zoonotic DENV-1, DENV-2, and DENV-4 suggest an association between Macaca and Presbytis spp. monkeys with Aedes niveus mosquitoes as the primary vectors.¹ Sylvatic DENV-3 has not been isolated but is presumed to exist in Malaysian sylvatic cycle on the basis of seroconversions in sentinel monkeys.² Sylvatic DENV cycles have also been documented in west Africa, where DENV-2 circulates among Erythrocebus patas and possibly other monkeys, and sylvatic Aedes mosquitoes, including Ae. taylori, Ae. furcifer, and Ae. luteocephalus, in a sylvatic focus near Kedougou, Senegal.^3–5 Limited spillover transmission of sylvatic DENV to humans has been documented in west Africa,^6,7 but until recently, the available data suggested that these sylvatic viruses were confined to forest habitats. However, recent phylogenetic and seroepidemiologic evidence suggests that spillover epidemics do occasionally occur into urban settings.^8,9

Phylogenetic analyses suggest that the recent emergence of endemic, human DENV from sylvatic progenitors in Asia occurred within the past 1,000 years.¹⁰ However, hyperendemic, urban DENV probably arose in Southeast Asia and eventually globally after the invasion of the highly efficient, anthropophilic vector Ae. aegypti aegypti. This process may also have been facilitated by adaptation of DENV from the arboreal canopydwelling, ancestral Aedes spp. vectors (other than Ae. albopictus and Ae. aegypti) to the peridomestic, an-thropophilic mosquitoes Ae. aegypti and Ae. albopictus.¹¹ However, recent analyses using a larger number and diversity of isolates have not supported this hypothesis (Vasilakis N, unpublished data, and Hanley K, unpublished data). Recent reports have also shown that the forest canopydwelling mosquito Ae. furcifer is highly susceptible to sylvatic DENV infection.¹² Aedes furcifer shows a pattern of movement into villages in eastern Senegal,⁵ which suggests that this species may act as a bridge vector for exchange of DENV between forest and peridomestic habitats. Furthermore, the ability of Ae. aegypti and Ae. albopictus to transmit sylvatic DENV-2¹² indicates that little or no change in vector infectivity accompanied endemic emergence from sylvatic progenitors. Finally, the ability of sylvatic DENV-2 to replicate efficiently in two model systems for human infection, monocyte-derived, human dendritic cells (maDCs), and mice engrafted with human hepatoma cells,¹³ as well as recent evidence of rapid sylvatic DENV turnover caused by their high nucleotide substitution rates across the genome,¹⁴ suggest the risk for sylvatic DENV-2 re-emergence into an urban cycle may be high. This finding could limit the potential for eradicating the human transmission cycle with the dengue vaccines now under development.^15–18

We assessed the likelihood of current sylvatic DENV re-emergence in the face of immunity to current urban strains or vaccine candidates by evaluating the neutralizing capacity of sera from DENV vaccinees and convalescent patients against geographically and genetically diverse sylvatic and urban DENV strains (Table 1). This information may allow us to predict the long-term success of candidate vaccines currently under development, and evaluate the degree to which pre-existing antibodies to urban DENV-2 in the human population could prevent the re-emergence of sylvatic DENV-2. In other words, if endemic DENV cycles were eradicated by vaccination, would cross-immunity against sylvatic strains prevent their re-emergence as long as herd immunity remained above a threshold level needed to prevent efficient urban transmission?

First, we used a modified plaque reduction neutralization test (PRNT) to evaluate the neutralizing capacity of sera from 26 (18 test and 8 placebo), 27 (19 test and 8 placebo), and 24 (20 test and 4 placebo) DENV vaccinees^16,19,20 collected 42 days post-vaccination with the rDEN1Δ30, rDEN2/4Δ30, and rDEN4Δ30 vaccine candidates, respectively. The vaccine candidate rDEN4Δ30, a live recombinant DENV-4 virus, was derived from the 814669 (Dominica/81) strains and contains a 30-nucleotide deletion in the 3′ non-coding genome region.²¹ Strain rDEN1Δ30 was derived from the Nauru/74 strain and contains a 30-nucleotide deletion in the 3′ non-coding genome region.^19,22 Strain rDEN2/4Δ30 was derived from the rDEN4Δ30 virus and the prototypic DEN2 strain New Guinea C. The premembrane and envelope proteins of DENV-4 were replaced with those of NGC.²³

The PRNT was performed in 12-well, Vero-microplate-cell cultures using a fixed virus inoculum (approximately 800 focus-forming units) against varying serum dilutions (1:20–1:2,560). Serum samples were diluted in minimal essential medium (MEM) containing 2% fetal bovine serum. Virus was mixed with an equal volume of each serum dilution (1:20–1:2,560) and the mixture was incubated for one hour at 37°C. A 250-μL volume of the serum–virus mixture was then placed into Vero cultures and incubated for one hour at 37°C. A 1-mL volume of 4% methycellulose in OPTIMEM-I overlay (GIBCO BRL, Gaithersburg, MD) was placed in each well and the plates were incubated for four days at 37°C. The plates were then fixed with 1:1 methanol:acetone and foci were stained immunologically and counted to determine the level of virus neutralization as described previously.¹³ The PRNT titers were scored as the reciprocal of the highest dilution of serum that inhibited 80% of the foci (PRNT₈₀; data not shown). Although some DENV studies use lower reduction percentages to report titers, we used the 80% endpoint to estimate conservatively in vivo protection.

Sera from all eight placebo-vaccinated persons were unable to neutralize any DENV tested (PRNT₈₀ < 20). Our data indicated that rDEN2/4Δ30 vaccination or DENV-2 infection elicited a robust, homotypic virus neutralization responses against both urban and sylvatic DENV-2 strains (PRNT₈₀ = 20–1,280 and 20–640, respectively). Only one sample from a DENV-2–vaccinated person exhibited any heterotypic virus-neutralizing activity (Table 2).

All but four sera from persons vaccinated with rDEN1Δ30 exhibited a homotypic virus neutralization response (PRNT₈₀ = 40–640) against the endemic DENV-1 strain. Although only one of these volunteers failed to seroconvert in a previous study,¹⁹ this discrepancy is caused by the stringency of our assay (PRNT₈₀ versus PRNT₆₀ used previously). Sera from 7 of these 18 rDEN1Δ30–vaccinated persons failed to neutralize the sylvatic DENV-1 strain P72-1244 at the sensitivity range of our test system (PRNT₈₀ < 20; Table 3). Sera from seven of these vaccinated persons exhibited weak heterotypic virus neutralization to DENV-2 and DENV-4 (PRNT₈₀ = 20–40) (Table 3).

In a similar manner to sera from DENV-1 vaccinees, sera from rDEN4Δ30 vaccinees exhibited relatively weak homotypic neutralization (PRNT₈₀ = 20–40) (Table 4). Sera from three of these vaccinated persons exhibited heterotypic virus neutralization to DENV-2 (range of reciprocal PRNT₈₀ = 40–160), whereas serum from one vaccinated person exhibited a relatively strong virus neutralizing response DENV-2 (PRNT₈₀ = 320–> 1,280) (Table 4). For the latter person, we obtained a similar profile by hemagglutination inhibition assay; data not shown. Although volunteer participation in this vaccine trial included stringent eligibility criteria, including absence to prior exposure to flaviviruses,¹⁹ the volunteer’s robust neutralization titer (> 1,280) to DENV-2 suggests a possible secondary antibody response (previous DENV-2 infection) and “original antigenic sin.”^24,25 This concept was proposed by Halstead and others²⁵ as a mechanism of DENV immune response involved in sequential infections, where response to a secondary infection is dominated by the proliferation of cross-reacting memory cells induced by the primary infection, which may be of lower affinity for the secondary challenging antigen. However, it is very unlikely that volunteer no. 12 would have been exposed to any flavivirus because this person was a Baltimore resident with no travel history to DENV-endemic countries. This volunteer was thoroughly screened and found to be negative for previous flavivirus infection. The person had day 0 (prior to vaccination) DENV PRNT₆₀ titers < 10 for all four DENV, yellow fever virus, and West Nile virus. In addition, hemagglutination inhibition titers to St. Louis encephalitis virus, West Nile virus, and Japanese encephalitis virus were also negative prior to vaccination. The postvaccination titer was not unusual for DENV infection.

We next examined the ability of sera from convalescent patients after primary infection with DENV-2 or DENV-3 to neutralize sylvatic DENV-2. Primary cases were defined by positive virus identification by reverse transcription–polymerase chain reaction (DENV-2 or DENV-3)²⁶ and IgG antibody titer ≤ 1:80 during the acute phase of disease (0–5 days after onset of symptoms). Paired, convalescent serum specimens obtained 12–20 days after the onset of symptoms were used for neutralization assays. These specimens were obtained from routine surveillance, and were de-identified and approved for research studies under institutional review board exemption 4797 at the Centers for Disease Control and Prevention. Because no sylvatic DENV-3 have been isolated,² we only tested for the neutralization capacity of DENV-3 sera against endemic DENV-3. Furthermore, the limited volume of available sera prohibited us from evaluating their neutralization capacity with a large collection of DENV strains or genotypes.

Sera from convalescent patients with primary DENV-2 infection exhibited robust homotypic neutralization capacity for both urban and sylvatic DENV-2 (PRNT₈₀ = 80–> 1,280) and no heterotypic virus neutralization of DENV-3 (Table 5). Similarly, sera from primary DENV-3 infection exhibited robust homotypic neutralization for DENV-3 (PRNT₈₀ = 80–320) and no heterotypic neutralization for either urban or sylvatic DENV-2 (Table 5). Control sera from two consenting healthy volunteers with no history of DENV infection were unable to neutralize any DENV (PRNT₈₀< 20; data not shown).

Collectively, our findings demonstrate the capacity of sera from DENV-vaccinated persons and convalescent patients after primary infection with DENV-2 and DENV-3 serotypes to neutralize geographically and genetically diverse sylvatic and urban DENV-2 strains because of strong homotypic immunity. Furthermore, our data suggest limited heterotypic virus neutralization activity in sera of DENV-vaccinated persons, consistent with Sabin’s observation that DENV infection produces protection for up to 12 weeks against disease caused by heterotypic infection.²⁷ Thus, should a licensed vaccine lead to the eradication of the urban dengue transmission cycle in the future, re-emergence of sylvatic strains into an urban cycle may be limited by homotypic humoral immunity. However, the danger of endemic DENV re-emergence will remain as long as vaccine coverage is incomplete and Ae. aegypti and/or Ae. albopictus remain in sufficient numbers.

Several lines of evidence suggest that re-emergence of current sylvatic DENV into the urban human transmission cycle is quite likely,^12–14 and could pose a major public health problem. If DENV can readily re-emerge from sylvatic cycles not amenable to intervention, the reduction and ultimate eradication of dengue from human populations solely by vaccination campaigns could be, at best, short-lived. Successful eradication of dengue will hinge on sustained vaccination of the susceptible populations at risk to introduction of sylvatic DENV, as well as control of the urban mosquito vectors.