Materials and Methods

Rhesus monkeysYoung adult rhesus monkeys of Indian origin, bred at the veterinary facility of the US Army Medical Component, Armed Forces Research Institute of the Medical Sciences (USAMC-AFRIMS), Bangkok, Thailand, were used for this study. The USAMC-AFRIMS Animal Care and Use Program is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International (AAALAC). Monkeys were housed and handled in accordance with institutional guidelines. All monkeys were skin-test negative for tuberculosis and seronegative for cercopithecine herpesvirus 1 (B virus), simian retrovirus, simian immunodeficiency virus, and simian T cell leukemia virus type 1, by standard commercial assays. Before entry into the study, all monkeys tested negative for DENV serotypes 1, 2, 3, and 4 (DENV-1, -2, -3, -4) and for Japanese encephalitis virus antibody by hemagglutination-inhibition and neutralizing-antibody assay [16, 17]. The monkeys were inbred locally and therefore not screened for yellow fever virus antibody. Research was in compliance with the Animal Welfare Act and with principles of the National Research Council [18]. The protocol was approved by the institute’s Animal Care and Use Committee and by the Animal Use Review Division of the US Army Medical Research and Material Command. This study was performed in 1999

Vaccine virusesThe 4 DENV serotype vaccine strains were each derived from a human clinical isolate. Viral attenuation by serial passages in primary dog kidney cell cultures and then fetal rhesus monkey lung cells was performed as described elsewhere [19]. Each vaccine was produced under Good Manufacturing Practice regulations as a single lot; the same lots have been used in clinical trials [20]. The vaccine virus strains were designated DENV-1 45AZ5 PDK20, DENV-2 S16803 PDK50, DENV-3 CH53489 PDK20, and DENV-4 341750 PDK20. Each lyophilized monovalent vaccine was reconstituted with sterile water and administered in a volume of 0.5 mL. The doses of DENV-1, -2, -3, and -4 were 10⁶, 10⁶, 10⁵, and 10⁵ pfu/mL, respectively. The TDV dose was prepared by mixing 0.25 mL of each reconstituted monovalent component and was administered in a total volume of 1 mL—0.5 mL administered subcutaneously in each upper arm. The dose of the TDV was 1.1×10⁶–2.8×10⁶ pfu/mL

Challenge virusesThe 4 challenge virus strains were DENV-1 West Pac 74, DENV-2 S16803 PDK10, DENV-3 CH53489, and DENV-4 341750. DENV-1, -3, and -4 are parent viruses of the vaccines. The DENV-1, -2, and -4 strains have previously consistently produced viremia in rhesus monkeys: 4 of 4 monkeys (mean, 6.8 days of viremia), 4 of 4 monkeys (mean, 5 days), and 3 of 3 monkeys (mean, 4.7 days), for the DENV-1, -2, and -4 strains, respectively [19]. The present study is the first in which the DENV-3 CH53489 strain has been used in a rhesus monkey challenge

Vaccination and challenge scheduleMonkeys were randomly assigned to 1 of 3 groups: validation, vaccine, and control, consisting of 12, 20, and 8 monkeys, respectively. The validation group had 2 purposes: (1) to ensure that the challenge viruses consistently induced viremia and (2) to compare the 2 viremia detection methods. The 12 monkeys in the validation group were divided into 4 subgroups of 3 monkeys each. Each subgroup received 1 of the 4 challenge viruses. After successful validation of each challenge virus, the 20 monkeys in the vaccine group were administered TDV on days 0 and 32, and the 8 controls were inoculated with saline placebo on the same study days. The study veterinarian and caretakers were blinded to the test article assignments. At 4.5 months after the second vaccination, the 20 vaccinated monkeys were randomly divided into 4 subgroups of 5 monkeys each, and each subgroup was challenged with 1 of the 4 challenge viruses. Concurrently, the 8 monkeys in the control group, all of which were flavivirus naive, were divided into 4 subgroups of 2 monkeys each, and each subgroup was randomly assigned and challenged with 1 serotype of the challenge viruses. All challenges were administered subcutaneously in a shaved medial upper arm

Blood collectionApproximately 5.0 mL of blood was collected from each monkey 6 weeks before the first inoculation, to ascertain flavivirus antibody status. For the validation group, 2.0 mL of blood was collected on study day 0 (day of challenge) and then on study days 1–12 and 21. For the vaccine and control groups, 2.0 mL of blood was collected on study days 0 (before inoculation), 21, 167 (4.5 months after dose 2, on the day of challenge), then daily for 12 days (study days 168–179), and then finally on study day 194 (4 weeks after challenge). All blood specimens were coded so that the laboratory studies were performed in blinded fashion

Virus isolation and quantitationEnd-point titrations on LLC/MK₂ cell monolayers were performed on all residual vaccine and challenge inocula, for verification of titers. Plasma collected after challenge from the validation-group monkeys was used to compare the sensitivity of 2 viral isolation methods: intrathoracic inoculation into Toxorhynchites splendens [21] mosquitoes and a nested reverse-transcription polymerase chain reaction (RT-PCR) assay, described elsewhere [22, 23]. Briefly, for each plasma specimen, 20 T. splendens mosquitoes were each injected with 0.17 mL of plasma, using a microcapillary pipette. Inoculated mosquitoes were reared at room temperature for 2 weeks and then killed, and the heads were tested for DENV by a DENV-specific immunofluoresence antibody (IFA) method. The bodies of mosquitoes testing positive for DENV by the IFA method were then ground up, and viral isolation on Vero cells was completed as described elsewhere [24, 25]. The nested-primer RT-PCR used was a DENV serotype–specific virus RNA detection assay, performed using a modified Lanciotti method [23]. Oligonucleotide primers used to amplify and type DENV were D1 (5′-TCA ATA TGC TGA AAC GCG CGA GAA ACC G-3′), D2 (5′-TTG CAC CAA CAG TCA ATG TCT TCA GGT TC-3′), TS1 (5′-CGT CTC AGT GAT CCG GGG G-3′), TS2 (5′-CGC CAC AAG GGC CAT GAA CAG-3′), TS3 (5′-TAA CAT CAT CAT GAG ACA GAG C-3′), and TS4 (5′-CTC TGT TGT CTT AAA CAA GAG A-3′). The nested RT-PCR was performed on serum to measure viremia after challenge

Plaque-reduction neutralization testDENV-specific neutralizing antibody titers were determined using previously described methods [16]. Briefly, a monolayer of LLC/MK₂ cells was inoculated with 50 pfu of reference virus. The reference virus strains were DENV-1 16007, DENV-2 16681, DENV-3 16562, and DENV-4 1036. All samples were were coded, heat inactivated, and assayed in a blinded fashion in duplicate at 4-fold serial dilutions of 1:10, 1:40, 1:160, 1:640, and 1:2560. The 50% plaque-reduction neutralization titer (PRNT₅₀) was calculated by use of a log probit regression method and reported as a reciprocal titer. A titer of ⩾1:10 was considered to be positive. A value of 1 was used for titers <1:10 in the calculation of geometric mean titer (GMT). Titers of >1:640 or >1:2560 indicate that no higher titration was performed; those values were used in GMT calculations

Statistical analysisDays of viremia were compared using a 2-sample t test. Onset of viremia was compared using analysis of variance (ANOVA). Differences in seroconversion rates between groups were compared using the χ² test. Differences in fold increases in antibody titers were compared using Mann-Whitney analysis. Statistical analysis was performed using SPSS software (version 10.0; SPSS). For all analyses, P⩽.05 was considered to be significant

Results

Validation of challenge viruses: viremia and neutralizing antibody in unvaccinated monkeys after challengeAll 4 challenge viruses produced viremia in all validation-group monkeys, as assessed by both mosquito inoculation and nested RT-PCR methods. Table 1 shows that, in the validation group, more days of viremia were detected by the nested RT-PCR method than by mosquito inoculation, with means of 6.1 days and 4.3 days, respectively (P=.05). The differences were especially marked with DENV-3 and -4. On the basis of these results, the nested RT-PCR was considered to be more sensitive and was used for determining viremia in the subsequent vaccine and control group experiments

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Reciprocal 50% plaque-reduction neutralization titers and viremia after challenge in flavivirus-naive rhesus monkeys

Results from all 20 flavivirus-naive monkeys in the validation and control groups are shown together in table 1. Nineteen of these 20 monkeys that were given challenge developed viremia. Only 1 animal challenged with DENV-4, DA681, did not develop viremia and was not included in the analysis. The mean durations of viremia were 5.4, 7.4, 5.4, and 6.0 days for DENV-1, -2, -3, and 4, respectively. DENV-3 caused the earliest onset of viremia. The median days of onset of viremia were days 4, 4, 2, and 4.5, respectively. No significant differences were noted between serotypes in total days of viremia (P>.05, ANOVA)

All monkeys in the validation and control groups seroconverted to the challenge viruses, including monkey DA681, which had no viremia detected after DENV-4 challenge. As expected for these flavivirus-naive monkeys, the challenge virus induced the highest antibody titer against the homologous serotype. GMTs were 3 log₁₀ for DENV-1, -2, and -3 and were 2 log₁₀ for DENV-4. GMTs against the homologous viruses were all at least 10-fold higher than those against heterologous viruses after challenge. Cross-reactive, heterotypic antibodies were elicited in all monkeys except DA681. In the flavivirus-naive monkeys, cross-reactivity of neutralizing antibodies varied among serotype combinations. For example, recipients of DENV-1 had relatively high levels of cross-reactive DENV-2 antibody, with a DENV-1:DENV-2 GMT ratio of 10, and, reciprocally, recipients of DENV-2 had a DENV-2:DENV-1 GMT ratio of 30. In contrast, DENV-3 induced no cross-reactivity with DENV-4 at all, and, reciprocally, DENV-4 had little cross-reactivity with DENV-3. Notably, antibody cross-reactivity was not necessarily reciprocal. For example, monkeys challenged with DENV-2 developed very little cross-reactive antibody to DENV-4, with a DENV-2:DENV-4 GMT ratio of 599, but the DENV-4:DENV-2 GMT ratio was only 10

Neutralizing antibody response to tetravalent vaccinationsThe antibody responses in the 20 monkeys vaccinated with TDV are shown in table 2. Seroconversion rates after a single dose of TDV were 95%, 100%, 70%, and 15% to DENV-1, -2, -3, and -4, respectively (table 2). Seropositivity rates 4.5 months after the second vaccination were 100%, 100%, 90%, and 70%, respectively. Thirteen (65%) of 20 monkeys developed neutralizing antibodies to all 4 serotypes after 2 vaccinations. Six (30%) of 20 monkeys developed trivalent responses, 2 lacked a response to DENV-3, and 6 lacked a response to DENV-4. Altogether, 19 (95%) of 20 monkeys developed a tri- or tetravalent neutralizing antibody response after a 2-dose vaccination schedule. On the basis of seroconversion rates and GMTs, the DENV-2 component appeared to be the most immunogenic, followed by DENV-1, -3, and -4, in descending order

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 2

Reciprocal 50% plaque-reduction neutralization titers in flavivirus-naive rhesus monkeys after 2 doses of tetravalent dengue virus (DENV) vaccine at 0 and 1 month

Viremia and antibody response in vaccinated monkeys after challengeTable 3 shows each monkey’s neutralizing antibody status in relation to viremia. The durations of viremia were significantly reduced in the vaccinated monkeys (table 4), except for those challenged with DENV-3 (P=.07). All 5 vaccinated monkeys challenged with DENV-2 were completely protected, as is shown by an absence of DENV viremia. None of these 5 monkeys showed any significant (⩾4-fold) increase in DENV-2 antibody titer after challenge. Complete protection against DENV-1, -3, and -4 challenge was observed in 4 of 5, 4 of 5, and 3 of 5 vaccinated monkeys, respectively. Even in the vaccinated monkeys with breakthrough viremia, the durations were shorter than in the control monkeys. The lone exception was monkey DA650, whose duration of DENV-3 viremia was 6 days, compared with the controls’ mean duration of 5.4 days. All 16 vaccinated monkeys that were completely protected against challenge had measurable neutralizing antibodies to the homologous challenge serotype at the time of challenge. Three of 4 vaccinated monkeys that were incompletely protected had measurable antibody to the homologous challenge serotype. GMTs in the protected and unprotected monkeys were 165 and 19, respectively. One of 5 monkeys (DA575) challenged with DENV-1 developed viremia of 2 days’ duration. The prechallenge antibody titer in this monkey was 1:20, which was the lowest titer among the 5 monkeys. The single monkey (DA650) that developed viremia after DENV-3 challenge had a prechallenge antibody titer of 1:119. Interestingly, another monkey (DA651) protected from challenge with DENV-3 had a prechallenge titer of 1:13. After DENV-4 challenge, 2 of the 5 vaccinated monkeys (DA658 and DA661) developed viremia of 1 and 4 days’ duration, respectively. These 2 monkeys had titers of 1:10 and <1:10 at the time of challenge, respectively. Among the vaccinated animals, the highest boosts in postchallenge homologous antibody titers were seen in the 4 animals that developed viremia. The 5 completely protected monkeys in the DENV-2 challenge group had the lowest ratios of postchallenge to prechallenge DENV-2 antibody titer (1.0–2.8-fold), suggesting that there was little, if any, viral replication. The mean fold increase of postchallenge antibody titers in all vaccinated monkeys with breakthrough viremia, compared with that in all protected monkeys, was 120-fold versus 7-fold (P<.006)

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 3

Reciprocal 50% plaque-reduction neutralization titers and viremia in rhesus monkeys that received the tetravalent dengue virus (DENV) vaccine

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 4

Viremia in rhesus monkeys after challenge

Discussion

Overall, our 4 challenge viruses consistently produced viremia of 3–10 days duration in rhesus monkeys. One exception was a control-group monkey that did not develop viremia after DENV-4 challenge. However, this monkey did seroconvert to DENV-4, indicating that DENV-4 infection did occur. These challenge viruses have also induced reproducible viremia in cynomolgus monkeys [26]. Not surprisingly, each challenge virus induced almost a 1 log₁₀ higher GMT than the vaccine viruses, which is consistent with our experience that the more attenuated vaccine viruses elicit lower levels of neutralizing antibodies

Our study provides evidence that the live attenuated TDV protected rhesus monkeys against DENV challenge. Eckels et al. previously showed that a single dose of each of the monovalent components of the TDV induced 100% seroconversion in rhesus monkeys [19]. In that study, the titers of vaccine virus were similar to, or 1 log₁₀ lower than, those used in this study. In our study, the antibody responses to DENV-1 and -2 after a single dose of TDV were similar to those in rhesus monkeys that were administered the monovalent vaccine viruses. This was not the case with DENV-3 and -4, for which the seroconversion rates after a single dose of TDV were 14 (70%) of 20 and 3 (15%) of 20, much less than the 100% observed when these viruses were given alone. Similarly, the GMTs after 1 dose of TDV were highest for DENV-1 and -2 and lowest for DENV-4. Thus, the DENV-1 and -2 components appeared to suffer no loss of immunogenicity when given as TDV, whereas DENV-3 and -4 elicited lower antibody responses when given as TDV, suggesting viral interference. Possibly, the immunogenicity of TDV depends not only on the dose but also on the dose ratios among the individual components. Such dose-dependent interference has been observed with other live attenuated TDVs in cynomolgus and rhesus monkeys that were given the ChimeriVax-DENV TDV and the TDV containing chimeric and 3′-untranslated region deletion mutants [26, 27]. Sabin made similar observations, using volunteers; he observed that the simultaneous administration of equal doses of DENV-1 and -2 viruses resulted in suppression of or delay in the development of DENV-1 antibody [28]. When DENV-1 was administered after DENV-2 at an interval of <6 weeks, no DENV-1 antibody developed. Apparent interference among TDV serotypes in humans was also observed with the Mahidol/Aventis vaccine, in which DENV-3 was immunodominant [29]. We anticipated that a second dose of TDV vaccination would be required to overcome interference. A second dose of TDV administered 32 days after the first dose appeared to have boosted seroconversion to DENV-3 and -4 in the monkeys. A parallel control group that does not receive a second vaccine would be needed to confirm this. Interestingly, when human subjects were given TDV by use of the same 0- and 1-month schedule, there were no additional seroconversions from the second vaccination [20]. This difference between rhesus monkeys and humans suggests that optimization of the vaccination schedule should be done in clinical trials. Like the monovalent DENV inoculations in Sabin’s experiments, the TDV will likely require an interval longer than 1 month between vaccinations, to induce maximal antibody responses

Natural DENV infection results in protection against heterotypic DENV for ∼6 months but in decades-long protection against the homotypic DENV [28, 30]. DENV neutralizing antibody after natural infection is thought to be the marker of protective immunity. By inference, serotypic neutralizing antibodies are considered to be markers for vaccine-induced protection, which may be embodied as T cell and B cell memory. Previous studies showed that DENV infection produces a diversity of antibodies that are protective but also may include enhancing antibodies that mediate DHF [31–33]. There may be qualitative differences between antibodies from natural infections and those from TDV vaccination. Each single virus delivered in a tetravalent formulation might induce a panoply of homotypic and heterotypic, protective and nonprotective, and neutralizing and nonneutralizing antibodies. Indeed, all of the challenge viruses in this study elicited both homotypic and lower cross-reactive heterotypic neutralizing antibodies at 28 days. Although it is likely that most of the prechallenge antibodies 4.5 months after vaccination represent serotype-specific protective antibodies rather than heterotypic cross-reactive antibodies, the PRNT assay does not distinguish these 2 types of antibodies

Monkey challenge studies attempt to address the question of whether antibodies induced by TDV will protect humans against disease. In work done by Halstead et al., rhesus monkeys given a single dose of near-wild-type virus seroconverted and were protected against homologous DENV challenge [14]. We evaluated the protection afforded by 4 attenuated viruses in combination. The TDV afforded variable protection against the 4 serotypes. In general, fully protected animals (no viremia) had antibody levels higher than those in unprotected animals. However, although neutralizing antibody as detected by PRNT₅₀ was necessary, it did not appear to be sufficient for full protection; 3 of 4 vaccinated monkeys developed viremia from challenge, despite having neutralizing antibody. One monkey had a prechallenge titer of 1:119 and still developed viremia from the challenge virus. This finding suggests that immunity to DENV infection is multifactorial and imperfectly predicted by our neutralization test. Our observation in nonhuman primates—that the mere presence of neutralizing antibody may not fully correlate with protection—is also supported by a recent epidemiologic study [34]. In this prospective study of 2000 Thai schoolchildren, a subgroup of children with DENV-3 DHF had DENV-3 neutralizing antibody before the onset of disease, although the presence of antibody was associated with milder disease. The PRNT assay cannot distinguish homologous protective antibodies from cross-reactive, nonprotective antibodies. Direct evaluation of TDV in a field setting will be required to clearly elucidate the role of antibody in host protection

Viremia is definitive evidence of disseminated viral replication. The TDV provided sterile protection against DENV-2 and at least partial protection against DENV-1, -3, and -4. Even the monkeys with breakthrough viremia tended to have more-abbreviated viremia. The nested PCR assay we used to measure viremia was highly sensitive and able to detect as little as 0, 1, 2, and 2 log₁₀ pfu/mL of DENV-1, -2, -3, and -4, respectively (V. Vassilev, personal communication). No reliable quantitative PCR assay was available to us at the time of this study. A weakness of the nested PCR assay was that it could not distinguish between viable infective virus and neutralized circulating virus. It is possible that, in humans, TDV may at least attenuate, if not totally prevent, disease. Finally, in the 4 vaccinated monkeys with viremia, there was no evidence of antibody enhancement as measured by duration or onset of viremia. On the basis of the results of this study and the results of studies of humans [20], a more balanced immunogenicity will require a more immunogenic DENV-4 component and the TDV to be given with a 2-dose schedule with a longer dose interval

In summary, the results of our study suggest that the TDV conferred protection against all 4 DENV parental challenges in rhesus monkeys. Viral interference among the 4 serotypes of DENV appeared to result in lower immunogenicity for the DENV-3 and -4 components. The imbalance in antibody response can be overcome by revaccination. Clinical studies will be needed to optimize the TDV to achieve a balanced protective immune response