Am. J. Trop. Med. Hyg., 80(2), 2009, pp. 302-311
Copyright © 2009 by The American Society of Tropical Medicine and Hygiene
Evaluation of Interferences between Dengue Vaccine Serotypes in a Monkey Model
Bruno Guy*,
Veronique Barban,
Nathalie Mantel,
Marion Aguirre,
Sandrine Gulia,
Jeremy Pontvianne,
Therese-Marie Jourdier,
Laurence Ramirez,
Veronique Gregoire,
Christophe Charnay,
Nicolas Burdin,
Rafaele Dumas, AND
Jean Lang
Research and Discovery Departments, Sanofi Pasteur, Marcy l’Etoile, France
 |
ABSTRACT
|
Interferences between different antigens in the same vaccine formulation have been reported for some vaccines (e.g., polio vaccines, live attenuated dengue vaccine candidates). We examined interferences between the four serotypes of ChimeriVax dengue vaccines (CYDs) in a monkey model when present within a tetravalent formulation in equal concentrations (TV-5555). Immunoassays of vaccinated non-human primates showed that serotype 4 (DEN-4), and to a lesser extent, DEN-1 were dominant in terms of neutralizing antibody levels. Parameters that affected the interferences were identified, including 1) the simultaneous administration of two complementary bivalent vaccines at separate anatomical sites drained by different lymph nodes; 2) the sequential administration of two complementary bivalent vaccines; 3) the establishment of heterologous flavivirus pre-immunity before subsequent tetravalent immunization; 4) the adaptation of formulations by decreasing the dose of the immunodominant serotype; and 5) the administration of a 1-year booster. The applicability of these data to human responses is discussed.
|
INTRODUCTION
|
Dengue disease is caused by four closely related but antigenically distinct serotypes (DEN-1–DEN-4) of a virus of the genus Flavivirus. It is spread by Aedes species mosquitoes (A. aegypti and A. albopictus) in tropical and subtropical areas. Although often asymptomatic, dengue infection can cause a spectrum of disease ranging from non-specific viral syndrome (dengue fever [DF]) to life-threatening hemorrhagic disease (dengue hemorrhagic fever [DHF] and dengue shock syndrome [DSS]). An estimated 5–10% of dengue cases result in DHF and DSS, predominantly among children. The overall morbidity, mortality, and economic impact caused by dengue are substantial.1
Although protection against homotypic reinfection is complete and probably lifelong, cross-protection between serotypes is limited.2 Patients frequently experience a secondary, heterotypic infection with a greater risk of developing DHF and DSS. The pathogenesis of DHF/DSS is multifactorial and includes not only the viral strain and the patient’s age and genetic predisposition, but also their immune status and dengue infection history. Inappropriate, pre-existing, heterologous humoral and cellular immune responses are thought to be involved, highlighting the importance of prior immunity and the existence of positive or negative interference between serotypes.3–5
A major challenge in the development of a tetravalent dengue vaccine is to induce immune responses against all four serotypes to prevent any theoretical risk of developing more serious disease. Previous clinical studies have shown that antibodies are sometimes preferentially solicited against one or two dominant serotypes.6–8 Both intrinsic differences in immunogenicity of the serotypes and inter-serotype interference are thought to be involved. This study examined the involvement of some factors in serotype interference/dominance. Several questions were addressed: 1) the importance of administrating all four serotypes together in the same formulation; 2) the role of prior immunity against one or several dengue serotypes or other flaviviruses; 3) the role of the respective dose of each serotype; and 4) the role of booster vaccinations.
Some non-human primates (NHPs), including Rhesus (Macaca mulatta) and Cynomolgus monkeys (Macaca fascicularis), are susceptible to infection by dengue and yellow fever (YF) viruses. These species are recognized by WHO as good models to assess the neurotropism and the viscerotropism of attenuated YF vaccines and can provide valuable information on the immunogenicity and viremia induced by dengue vaccine candidates.9–12 We used Cynomolgus macaques to study immunogencity and viremia induced by different schedules and formulations of dengue candidate vaccines with respect to serotype interference. Schedules and formulations were 1) the simultaneous administration of two bivalent vaccines at separate anatomical sites drained by different lymph nodes; 2) the sequential administration of two complementary, bivalent vaccines; 3) the influence of heterologous flavivirus priming on subsequent tetravalent immunization; 4) the adaptation of formulations by decreasing the dose of the immunodominant serotype; 5) the effect of a 1-year booster dose. The vaccines studied were live-attenuated DEN-1 and DEN-2 vaccines (VDV1 and VDV2), derived in Vero cells from previous live attenuated (LAV) Mahidol vaccines,6 and the ChimeriVax DEN1–4 vaccine candidates based on the YF 17D virus vaccine (ChimeriVax-DEN1–4; CYD1–4; Sanofi Pasteur, Marcy l’Etoile, France), which are currently in clinical evaluation. 13,14
|
MATERIALS AND METHODS
|
Animals
Experiments were performed with Cynomolgus macaques (Macaca fascicularis; mean weight, 3.5 kg; weight range, 2.4–4 kg; males, 2–3 years old; obtained from Noveprim, Le Vallon, Mauritius) and group housed at Sanofi Pasteur. Protocols were submitted to the Sanofi Pasteur internal animal care committee, and the animals were housed and manipulated according to the European Directives relating to animal experimentation.
Vaccines and immunization schedules
Groups of four animals were allocated to receive different combinations of mono-, bi-, or tetravalent vaccine candidates in a series of immunization schedules performed in several independent experiments, designed to examine the influence of certain factors on serotype interference (Table 1
). VDV1 and VDV2 vaccines have been derived in Vero cells from previous Mahidol LAV1 and LAV2 vaccines obtained in PDK cells6 and correspond to clinical batches in these experiments. The CYD1–4 vaccine candidates based on the YF 17D virus vaccine were prepared from either clinical bulk or development Phase 2 batches, cultured in serum-free conditions. YF vaccine was the licensed vaccine Stamaril (Sanofi Pasteur). CYD and VDV vaccines contained 5 log10 CCID50 of each serotype. All immunizations were by subcutaneous injection in a final volume of 0.5 mL in one or both arms (deltoid region). Second doses were given 2 months later in Groups 5–14. Animals in Groups 6 and 8 received two bivalent vaccines on the same day in separate arms. Animals were observed for 30 minutes after each injection for immediate local and systemic reactions and daily for the next 12 days. Treatment groups and vaccination schedules are detailed in Table 1 and are described below.
View this table:
[in this window]
[in a new window]
|
TABLE 1 Summary of the various vaccination schedules studied and vaccine candidates given to each group of four Cynomolgus macaques
|
|
The immunogenicity of each individual CYD vaccine serotype was first assessed (Groups 1–4). Potential interference when using tetravalent vaccines was examined in three independent experiments (Groups 5, 7, and 9). To study whether interference occurs at the early stages of viral replication and antigen presentation at the injection site and local draining lymph nodes, and whether it can be modulated by targeting separate lymph nodes, two bivalent vaccines were given concomitantly in different arms. Two combinations of bivalent vaccines were tested and compared with tetravalent vaccines in two independent experiments: in the first combination, each bivalent vaccine contained one dominant and one non-dominant serotype (i.e., CYD1,2 and CYD3,4; Group 6). In the second, they contained either two dominant or two non-dominant serotypes (i.e., CYD1,4 and CYD2,3; Group 8). Animals received primary and booster immunizations with the same combination of vaccines. Heterologous prime-boost tetravalent vaccination strategies were used to study whether prior immunity against one or several CYD or DEN serotype(s) or against another flavivirus such as YF17D vaccine affected serotype interference and whether it could be decreased by temporal separation of dengue serotypes into complementary bivalent formulations (Groups 10 and 11). The affect of prior anti-dengue/flavivirus immunity on subsequent tetravalent dengue vaccination was also examined (Groups 12 and 13).
The effect of decreasing/increasing the relative dose of dominant/non-dominant serotypes on interference was examined by using a 2-log decrease (from 5 to 3 log10) in the immunodominant CYD4. This formulation (TV-5553, Group 14) was tested head-to-head with the classic tetravalent TV-5555 formulation (Group 7). Finally, we examined if a 1-year booster of TV-5555 formulation could induce a significant response against all serotypes by complementing the weaker responses. This was done in monkeys previously immunized twice with the TV-5555 or TV-5553 formulations or immunized concomitantly in two separate sites with two bivalent vaccines (Groups 7, 8, and 14).
Analyses
Viremia was monitored from D2 to D10 after each injection, and immunogenicity was assessed on D0 and D28 after each vaccination. Blood was kept on crushed ice throughout the animal manipulation pending analysis. Sera were kept on ice during sampling before being stored at –70°C pending analysis. All serum samplings were carried out under a laminar flow hood.
For practical and ethical reasons, it was not possible to perform experiments using a larger number of animals per group, and statistical comparisons were thus not done between groups. Results were descriptive, and only tendencies were analyzed and discussed. We have nevertheless performed analyses to define what we considered as a detectable x-fold difference between our reference group (tetravalent vaccine with 5 log10 CCID50 of each serotype) and the other immunization regimens, performed in parallel within each independent test. We proceeded as follows: we compared the log10 GMT obtained for the reference group after the second immunization in five different tests. When no detectable levels of neutralizing antibodies were observed in SN50 assay, the log10 GMT of these samples was by convention set at 0.7 log10 for calculation (one half of the minimum detectable titer: 10 SN50 = 1.0 log10). Titers against serotypes 1, 3, and 4 were found very reproducible (SD = ±0.15–0.22 log10), whereas DEN2 neutralization titers exhibited more variability (SD = ±0.5 log10). This latter observation can be explained by the large number of monkeys that did not respond against DEN2 in these experiments (9 of 20), introducing a bias in calculation. The variability of the SN50 assay itself partly accounts for the variability of the data: serum samples tested in two independent experiments exhibited a mean SD of ±0.15 log10. Based on these calculations, we considered a difference of at least 2 SD as detectable. For serotypes 1, 3, and 4, this represents at least a 3-fold variation in the GMT value. For DEN2 GMT, we also considered a 3-fold variation (
1 SD) as detectable, because of the bias introduced by the number of negative samples. We indicate in the tables by an asterisk the serotypes for which such a detectable difference was established. When the difference was higher than 3 SD (6-fold differences), a dagger is used.
Seroneutralizing antibody assays
Dengue neutralizing antibody levels were measured by a high-throughput, liquid 50% seroneutralization assay (SN50), a limit dilution assay based on the CCID50 test. To define the correlation between classic 50% plaque reduction assay (PRNT50) and SN50 assay, a subset of sera from the studies presented in this report was evaluated by both techniques. Data obtained in the two assays were correlated (R2 0.92), whichever the considered serotype. In such comparisons, PRNT50 values were 2- to 3-fold higher than SN50 values, showing this latter assay to be more stringent. To briefly describe the SN50 assay: six 4-fold serial dilutions of each decomplemented (56°C, 30 minutes) serum sample were prepared in IMDM (Iscove modified Dulbecco medium; Invitrogen, Paisley, Scotland) and 4% fetal calf serum (FCS) medium and distributed in a 96-deep well plate. The diluted serum was added to 225 PFUs of virus diluted in IMDM and 4%FCS in a final volume of 0.9 mL, and the mixtures were incubated for 1 hour at 37°C. Each dilution mix (6 x 0.1 mL) was distributed in 6 wells of a 96-well plate seeded 3 days before with 8,000 Vero cells per well. After 6 days of incubation at 37°C with 5% CO2, cells were fixed with a mixture of ethanol/acetone 70%/30% and incubated with pan-flavivirus mAb 4G2 for 1 hour at 37°C. After development of lysis plaques with BCIP/NBT (5-bromo-4-chloro-. 3-indolylphosphate. nitroblue tetrazolium) colorimetric substrate (Sigma, Steinheim, Germany), the neutralizing titer was calculated using the Karber formula 15: [log10 SN50 = d + f/N (X + N/2)], where d is the dilution giving 100% neutralization (six negative replicates), f is the log10 of the dilution factor (dilution factor 1:4, f = 0.6), N is the number of replicates/dilution (N = 6), and X is the total number of wells exhibiting no sign of infection, excluding dilution d. In our conditions, the lowest measurable titer was 10 SN50 (20 PRNT50).
Viral strains used for neutralization were Mahidol strains DEN-1 16007, DEN-2 16681, DEN-3 16562, and DEN-4 1036 For controls, initial viral dilutions were back-titrated.
Viremia
Post-vaccinal viremia was measured by real-time, quantitative reverse transcriptase-polymerase chain rection (qRT-PCR). 16 Two sets of primers and probes located in DEN-1 and DEN-2 NS5 genes were used to measure VDV1 and VDV2 RNA levels, respectively. A third set of primers and probe, located in the YF NS5 gene, was used to quantify both YF17D and CYD RNA. In addition, four sets of serotype specific primers and probes, located at the junction DEN E/YF NS1 genes, were designed for identification of CYD serotype in YF NS5 RNA-positive samples. Seven plasmids containing the region targeted by each PCR under the control of a T7 promoter were transcribed in vitro to generate a set of synthetic RNA that was included as standards in each qRT-PCR assay to allow quantification of copy numbers (unit: GEQ, genomic equivalents). The limits of viral RNA quantification were 2.7 log10 GEQ/mL, for DEN-NS5 qRT-PCR and 3.3 log10 GEQ/mL for YF-NS5 and DEN-E/YF-NS1 qRT-PCR (500 and 2,000 GEQ/mL; 2.5 and 10 GEQ/reaction, respectively).
Correlation between infectious titer and qRT-PCR values was evaluated by spiking known infectious quantities of a control virus in 0.140 mL of negative monkey serum samples (D0), processed in parallel with test serum samples. Control sera were prepared at two dilutions containing 1 and ~100 PFU in 5 µL (2.3 and 4.3 log10 PFU/mL, respectively). The following correlation was established: YF and CYD assays, 1.0 log10 PFU = 3.1 log10 GEQ (CYD1), 3.3 log10 GEQ (CYD2), 2.7 log10 GEQ (CYD3 and CYD4); VDV2 assay, 1.0 log10 PFU = 3.5 log10 GEQ.
|
RESULTS
|
Animals presented no clinical signs in any study performed with VDV or CYD vaccine candidates.
As detailed in the Materials and Methods, vaccine-induced viremia was evaluated in a viral genome quantification assay, 16 providing data in genomic equivalents (GEQ), whereas the induction of neutralizing antibodies was assessed in a high-throughput liquid assay; this new seroneutralization 50 (SN50) assay was more stringent in our conditions than the standard PRNT50 assay, giving 2- to 3-fold lower titers. Thus, whereas these results generated in parallel in identical assays can be compared between schedules and groups, potential differences between assays should be kept in mind when comparing these data with previous published studies.
Monovalent vaccinations
A single vaccination with monovalent CYD vaccines (Groups 1–4) induced immune responses in the majority of animals, especially CYD1 and CYD4. The CYD2 and CYD3 vaccines seemed to be less potent in monkeys, inducing responses in only three of four animals (Table 2). No correlation was observed between immunogenicity and viremia, because viremia was detected for CYD3 and CYD4 only (Table 2), despite detectable levels of CYD1-induced antibodies. These results were confirmed in two independent studies in which monovalent CYD1 and CYD4 were dominant serotypes, whereas CYD2 and CYD3 were non-dominant (results not shown).
Tetravalent vaccinations
In three independent experiments using two different lots of vaccines (Groups 5, 7, and 9), TV immunization induced immune responses that were comparable between groups and in line with those described above with monovalent vaccines, although responses against CYD2 and CYD3 were further decreased or suppressed, as judged by the number of non-responders (Tables 3 and 5). Although quantitative variations were seen in the levels of anti-CYD1 or -CYD4 neutralizing antibodies after TV administration, the serotype dominance was similar in the three experiments: moderate to high immune responses were seen only against CYD1 and CYD4, even after the second vaccination. Responses to CYD2 and CYD3 remained low or undetectable. Viremia was consistently moderate and detectable only after priming and almost exclusively for serotype 4, except for one animal (AM645), in which low CYD1 viremia was also detected on 1 day only. In contrast to observations with CYD3 monovalent vaccine, no CYD3 viremia was detected after vaccination with tetravalent vaccine (Tables 4). and 6).
View this table:
[in this window]
[in a new window]
|
TABLE 3 Immunogenicity of simultaneous administration of bivalent vaccines in separate sites in comparison to tetravalent vaccination in a single site
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 4 Viremia induced on simultaneous administration of bivalent vaccines in separate sites in comparison to tetravalent vaccination in a single site
|
|
Tetravalent immunization by concomitant administration of two bivalent vaccines
No clear benefit was seen after priming with CYD1,2 and CYD3,4 at separate sites (Group 6), whereas the booster clearly had a positive effect for serotype 2 against which all four animals responded (Table 3
). Compared with CYD1–4, the CYD1,2 + CYD3,4 boosters also resulted in higher responses against the dominant DEN-1 and DEN-4 serotypes. As with the tetravalent vaccine, CYD4 viremia predominated after vaccination with the two bivalent formulations, but with higher peak values. One animal (AP817) also had detectable CYD1 viremia on 1 day only. No viremia was detected after the booster (Table 4). In a subsequent independent experiment, concomitant vaccination with the second combination of bivalents (CYD1,4 and CYD2,3) at separate sites induced a detectable and higher response after the booster against all four serotypes in all monkeys (Group 8), clearly showing the advantage of this strategy compared with vaccination with a tetravalent vaccine (Group 7) (Table 3): seroneutralizing titers were more evenly distributed between serotypes and were higher for DEN-1, DEN-2, and DEN-3. In terms of viremia, serotype 4 was again predominant. However, there was some limited and early viremia for both serotypes 1 and 3 in the bivalent group (Table 4).
Heterologous prime-boost immunizations
As noted above, two immunizations with CYD1–4 (Group 9) induced an immune response against DEN-1 and DEN-4 but lower responses against DEN-2 and DEN-3. In contrast, priming with either CYD1,2 or VDV1,2, followed by a heterologous boost with complementary serotypes CYD3,4 (Groups 10 and 11) induced seroneutralizing responses against all four serotypes at even higher levels than those observed with the concomitant administration of two bivalent vaccines in Group 8 described above (Table 5). In particular, all animals responded against DEN-2 and DEN-3. Priming with VDV1,2 led to a dominant response against DEN-2 after the CYD3,4 booster, whereas priming with CYD1,2 led to a well-balanced booster response against all four serotypes. Interestingly, after either the VDV1,2-prime or the CYD1,2-prime, the CYD3,4-booster enhanced the responses induced against DEN-1 and DEN-2. As previously observed, viremia was predominantly caused by CYD4, with no differences observed in its level depending on whether CYD4 was included in the primary (Group 9) or booster vaccine (Groups 10–13). The VDV2 and VDV1,2 vaccines induced VDV2 viremia only (i.e., no VDV1 viremia was detected in the Group 11 animals; Table 6). The VDV2-prime/CYD1–4-boost vaccination schedule elicited strong responses against all four serotypes (particularly DEN-2). To a lesser extent, the YF-prime/CYD1–4-boost schedule also elicited responses against all four DEN-serotypes, which were higher than after two immunizations with CYD1–4.
Adapted formulations with decreased dose of dominant serotype
The TV-5553 formulation induced a predominant response against DEN-1 after primary immunization compared with TV-5555 responses in which DEN-1 and DEN-4 were dominant, with no immediate benefit for DEN-2 or DEN-3 (Tables 7 and 8). However, the D56 booster induced responses against all four serotypes in almost all animals. No serotype 4 viremia was seen after primary immunization with TV-5553, although some viremia similar to the one seen for serotype 4 after TV 5555 administration was detected after the booster, but its serotype specificity could not be determined in this experiment.
Tetravalent vaccine with 1-year booster
Table 9 shows that titers observed before the 1-year booster (D370) were stable compared with those measured after the first booster at D56 (Tables 3 and 7); some initially weak anti-DEN-2 responses became undetectable in some animals. We observed a dramatic increase for all serotypes in all groups after the 1-year booster, inducing significant responses against all serotypes in almost all animals, including those in Group 7 that received two TV-5555 primary immunizations. No viremia was detected after this 1-year booster.
View this table:
[in this window]
[in a new window]
|
TABLE 9 Influence of a 1-year booster with TV 5555 vaccine in monkeys previously immunized twice with TV5555, TV5553, or by simultaneous bivalents in separate sites
|
|
|
DISCUSSION
|
We observed in a monkey model interference between CYD serotypes when administered as a tetravalent formulation at equal concentrations (TV-5555). The interference amplified the hierarchy of neutralizing antibody titers observed in this species for monovalent vaccines (DEN-4 > DEN-1 > DEN-3/DEN-2). This interference on neutralizing antibody titers after immunization could have at least two origins: higher infectivity of one or several serotypes, possibly resulting in competition occurring at the initial replication level, and/or immune competition linked to immunodominance of T- and/or B-cell epitopes. In the first case, the serotype(s) initially replicating at the highest levels in tissues and secondary lymphoid organs will have a competitive advantage over the other serotypes, which can be further enhanced if immunodominant epitopes for them exist. It remains to be explored whether the resulting MOI (multiplicity of infection) at the injection site is high enough to create a competitive environment; further in vitro and/or in vivo experiments with graded doses of vaccine are required to address this point.
In monkeys, the dominance of CYD4 could be primarily linked to its higher replication rates, as shown by its predominance in cases of viremia. However, intrinsic immunogenicity could also be involved because CYD3, which also induces viremia when administered in a monovalent vaccine, does not induce such high levels of neutralizing antibodies. On the other hand, CYD1 induced consistently high neutralizing antibody levels in the absence of viremia. The viremia hierarchy observed in monkeys (CYD4 > CYD3 > CYD1/2) parallels to some extent the hierarchy in replication titers previously observed in human dendritic cells in vitro.17 The neutralizing antibody and viremia titers observed in this study after immunization with 5555 or 5553 formulations were different from those obtained previously with CYDs in monkeys. 12 These discrepancies may be linked to different culture conditions, animal origin, and assays used to assess viremia and immunogenicity. Phase II lots produced in serum-free conditions were used in this study, whereas the former vaccines were produced at smaller scale in presence of serum. In addition, as indicated in the Materials and Methods, we used a more stringent SN50 assay to monitor neutralizing antibodies: PRNT50 values were 2- to 3-fold higher than SN50 values. Finally, our viremia assay is based on real-time quantification of viral genome (qRT-PCR; see Materials and Methods and Ref. 16).
In any case, whereas serotype 2 was consistently low in monkeys in this study, this is not the case in humans, at least when used as a monovalent vaccine. 14 It remains possible that both replication-linked and immunodominance-linked events contribute in some part to the interference observed in these studies. Although serotype interference was consistent and reproducible with TV-5555 formulations, some parameters such as pre-immune status, administration at different anatomical sites, or inclusion of a 1-year booster dose were able to modulate the final immune response and induce significant responses against all four serotypes in immunized animals.
Vaccinating at different anatomical sites was first evaluated > 30 years ago by Halstead and others. 18 These authors showed that injecting monkeys with wild-type dengue viruses at two or four different anatomical sites provided protection against subsequent infection. However, the superiority of separate administrations compared with administration in a single site was not shown with regard to protection against viremia after challenge. It is possible that the wild-type isolates used in this experiment retained their full ability to replicate and did not need the replication advantage afforded by separating the serotypes, whereas the attenuated viruses used in the current vaccine formulations do. Separating attenuated vaccine candidates may provide non-dominant serotypes the opportunity to replicate to higher levels, thus decreasing interference in replication and antigen presentation occurring at the injection site and the draining lymph nodes. Consistent with this, we observed at some time points detectable levels of CYD1 and CYD3 viremia with the separate administration of bivalent vaccines but not with the tetravalent vaccine. Furthermore, responses induced on separation seemed to be more homogeneous when non-dominant serotypes (CYD2 and CYD3) were not associated in the same bivalent vaccine with the more dominant serotypes (CYD1 and CYD4).
Heterologous prime-boost dengue vaccination strategies have also been tested in the past. They aimed to induce protection or cross-protection against the four serotypes through sequential infection with two serotypes (e.g., protection against DEN-3 and DEN-4 after sequential infection with DEN-1 and DEN-2), but this approach proved to be unsuccessful. 19–21 In contrast, in our prime-boost experiments, all animals were eventually vaccinated with all four vaccine viruses, because we did not look for heterologous cross-protection, but to see whether we could modulate interferences between the serotypes. The immunological mechanism involved in heterologous prime/boost strategies (17D or CYD1,2 priming followed by TV CYD or CYD3,4, respectively) might be the result of different and possibly complementary mechanisms. Potent cellular responses against the common YF-NS backbone may be recalled on tetravalent or bivalent CYD boosting, respectively, which would assist specific responses against the envelope proteins from booster serotypes through a positive bystander effect. Priming with VDV1,2 would also provide such bystander help thanks to flavivirus cross-reactive responses against NS proteins recalled by the CYD3,4 booster. Additionally or alternatively, cross-reactive, envelope-specific memory B cells—known to be potent antigen presenting cells at boosting—may have contributed to the response against heterologous booster serotypes after capturing and presenting them to T cells. The heterologous help provided by one immunization to the other worked in both directions, because a CYD3,4 booster also increased the anti-1,2 responses induced after priming, eventually resulting in strong responses to all four serotypes. This positive booster effect was paradoxically not seen in a homologous, tetravalent prime-boost, because it recalled predominantly the anti E 1/4-dominant responses induced after priming. The positive effect of YF 17D priming agrees with previous observations in monkeys and humans. 14 Although we did not study immunization sequences beginning at CYD3,4 and thus have no information regarding CYD4 viremia after primary bivalent immunization, it should be noted that heterologous prime-boost schedules did not induce higher CYD4 viremia after the CYD3,4 booster compared with levels observed after a primary vaccination with monovalent CYD4 (Table 2) or tetravalent vaccine (Table 6). This agrees with observations that dengue immunization provides some degree of cross-protection against heterologous challenge performed 4 months after the immunization or 6–8 weeks after immunization. 18,22
Considering heterologous priming further, pre-immunity against other flaviviruses or against one or several dengue serotypes before tetravalent vaccination may represent a situation similar to those encountered in areas in which YF vaccination is routinely performed or in dengue-endemic areas where contact with a least one dengue serotype early in life is frequent. The fact that such pre-immunity allows induction of a broader and stronger response after tetravalent vaccination may predict that a better response would be induced after only one or two immunizations in such areas.
Based on mathematical modeling, a recent report proposed serotype separation as a strategy to induce broad CD8 responses against dengue epitopes. 23 According to this model, the optimum approach would be to prime with a non-dominant CD8 epitope and boost with epitopes from the four dengue serotypes at four different anatomical sites. We found, however, that priming with a potent live attenuated dengue vaccine (VDV2), both in terms of immunogenicity and viremia, resulted in a strong antibody response against all four serotypes after a boost with a single tetravalent vaccine.
We also confirmed that serotype interference could be reduced by decreasing the relative dose of the dominant serotype, an approach previously applied with other vaccine candidates. 6,12,24 The decreased load of the dominant serotype, resulting in undetectable replication after the primary immunization, would leave more opportunity for the other serotypes to replicate at higher levels, at least locally. This was observed for serotype 1, which became dominant after priming, despite undetectable viremia, but two immunizations were still needed to get a detectable response against serotypes 2 and 3. Relative dominance of one or several serotypes has previously been observed on vaccination with live attenuated TV vaccine in rhesus monkeys, 25 in which serotype 4 induced the weakest response after both priming and boost. In this study, the dose of serotype 4 vaccine was 1 log lower than serotypes 1 and 2 (6655 formulation), which may have contributed to the low response induced; however, in our study using the 5553 formulation, whereas CYD4 induced almost no response after priming, anti-CYD4 responses were still dominant over CYD2 and CYD3 serotypes after boost, suggesting that it could be intrinsically immunodominant in monkeys. This may still be linked predominantly to its higher infectivity, even at low doses, although we have not been able to identify the serotype linked to viremia after boost with the 5553 formulation.
Finally, it seemed that whatever the initial schedule and formulation, a booster performed 1 year after primary immunization and 10 months after the first boost could eventually induce a significant response against all four serotypes. Even if responses were still better in groups initially immunized twice with the TV-5553 formulation or with complementary bivalents in different anatomical sites, it worked also for the TV-5555 formulation. It is important that immunizations are done several months apart to prevent negative interference, possibly because of short-lived antibodies such as IgM and/or to innate immunity, because we observed in monkeys that three immunizations performed 2 months apart were not effective at inducing responses against all four serotypes (results not shown).
The nature of the immunity induced by CYD vaccines on different immunization regimens may be different in monkeys and in humans, depending in particular on the nature of the dominant serotype(s). Nevertheless, the immunological concepts shown here may be applicable to both species; some regimens described in this paper could be evaluated in humans, keeping in mind practical aspects and theoretical issues in the field. For instance, the need for inducing a broad response against all four serotypes may raise practical and/or immunological questions about sequential bivalent immunizations; simultaneous polytopic bivalent vaccinations or an adapted 5553 formulation could be applied more easily, should they confer a significant advantage over a 5555 tetravalent immunization. In this respect, and as stated above, vaccination of flavivirus-naive travelers versus pre-immune subjects in endemic zones may correspond to different situations and different schedules, and this will require further studies in humans.
In conclusion, we showed that there are interferences in a monkey model between the four CYD serotypes linked to different in vivo replication abilities and/or to epitope dominance. The elucidation of the exact mechanisms will require further studies, and for instance, recombinant/non-replicating antigens could be used to differentiate between replication and/or epitope-linked interference. In any case, we showed in this model the different ways to modulate these interferences to eventually induce in animals immune responses to all four dengue serotypes.
Received May 13, 2008.
Accepted for publication August 26, 2008.
Acknowledgments: The authors thank Gee Marsh and Simon Jones for critical help in the preparation of the manuscript; Franck Raynal, Celine Vaure, and the animal facility team for excellent care and expertise; and Jeffrey Almond and Emanuelle Trannoy for constant support and helpful discussions.
Disclosure: The authors are employed by Sanofi Pasteur. This statement is made in the interest of full disclosure and not because the authors consider this a conflict of interest.
* Address correspondence to Bruno Guy, Sanofi Pasteur, 69280 Marcy l’Etoile, France. E-mail: bruno.guy{at}sanofipasteur.com 
Authors’ addresses: Bruno Guy, Veronique Barban, Nathalie Mantel, Marion Aguirre, Sandrine Gulia, Jeremy Pontvianne, Therese-Marie Jourdier, Laurence Ramirez, Veronique Gregoire, Christophe Charnay, Nicolas Burdin, Rafaele Dumas, and Jean Lang, Sanofi Pasteur, 69280 Marcy l’Etoile, France.
|
REFERENCES
|
- Clark DV, Mammen MP Jr, Nisalak A, Puthimethee V, Endy TP, 2005. Economic impact of dengue fever/dengue hemorrhagic fever in Thailand at the family and population levels. Am J Trop Med Hyg 72: 786–791.[Abstract/Free Full Text]
- Adams B, Holmes EC, Zhang C, Mammen MP Jr, Nimmannitya S, Kalayanarooj S, Boots MC, 2006. Cross-protective immunity can account for the alternating epidemic pattern of dengue virus serotypes circulating in Bangkok. Proc Natl Acad Sci USA 103: 14234–14239.[Abstract/Free Full Text]
- Green S, Rothman A, 2006. Immunopathological mechanisms in dengue and dengue hemorrhagic fever. Curr Opin Infect Dis 19: 429–436.[Web of Science][Medline]
- Rothman AL, 2004. Dengue: defining protective versus pathologic immunity. J Clin Invest 113: 946–951.[Web of Science][Medline]
- Vaughn DW, Green S, Kalayanarooj S, Innis BL, Nimmannitya S, Suntayakorn S, Endy TP, Raengsakulrach B, Rothman AL, Ennis FA, Nisalak A, 2000. Dengue viremia titer, antibody response pattern, and virus serotype correlate with disease severity. J Infect Dis 181: 2–9.[Web of Science][Medline]
- Sabchareon A, Lang J, Chanthavanich P, Yoksan S, Forrat R, Attanath P, Sirivichayakul C, Pengsaa K, Pojjaroen-Anant C, Chokejindachai W, Jagsudee A, Saluzzo JF, Bhamarapravati N, 2002. Safety and immunogenicity of tetravalent live-attenuated dengue vaccines in Thai adult volunteers: role of serotype concentration, ratio, and multiple doses. Am J Trop Med Hyg 66: 264–272.[Abstract]
- Sun W, Edelman R, Kanesa-Thasan N, Eckels KH, Putnak JR, King AD, Houng HS, Tang D, Scherer JM, Hoke CH Jr, Innis BL, 2003. Vaccination of human volunteers with monovalent and tetravalent live-attenuated dengue vaccine candidates. Am J Trop Med Hyg 69 (Suppl): 24–31.[Abstract/Free Full Text]
- Kitchener S, Nissen M, Nasveld P, Forrat R, Yoksan S, Lang J, Saluzzo JF, 2006. Immunogenicity and safety of two live-attenuated tetravalent dengue vaccine formulations in healthy Australian adults. Vaccine 24: 1238–1241.[Web of Science][Medline]
- World Health Organization, 1998. Requirements for yellow fever vaccine (Requirements for Biological Substances no. 3, revised 1995). WHO Tech Rep Ser 872: 31–68.
- Eckels KH, Dubois DR, Putnak R, Vaughn DW, Innis BL, Henchal EA, Hoke CH Jr, 2003. Modification of dengue virus strains by passage in primary dog kidney cells: preparation of candidate vaccines and immunization of monkeys. Am J Trop Med Hyg 69 (Suppl): 12–16.[Abstract/Free Full Text]
- Putnak RJ, Coller BA, Voss G, Vaughn DW, Clements D, Peters I, Bignami G, Houng HS, Chen RC, Barvir DA, Seriwatana J, Cayphas S, Garçon N, Gheysen D, Kanesa-Thasan N, McDonell M, Humphreys T, Eckels KH, Prieels JP, Innis BL, 2005. An evaluation of dengue type-2 inactivated, recombinant subunit, and live-attenuated vaccine candidates in the rhesus macaque model. Vaccine 23: 4442–4452.[Web of Science][Medline]
- Guirakhoo F, Pugachev K, Zhang Z, Myers G, Levenbook I, Draper K, Lang J, Ocran S, Mitchell F, Parsons M, Brown N, Brandler S, Fournier C, Barrere B, Rizvi F, Travassos A, Nichols R, Trent D, Monath T, 2004. Safety and efficacy of chimeric yellow fever-dengue virus tetravalent vaccine formulations in non-human primates. J Virol 78: 4761–4775.[Abstract/Free Full Text]
- Guirakhoo F, Pugachev K, Arroyo J, Miller C, Zhang ZX, Weltzin R, Georgakopoulos K, Catalan J, Ocran S, Draper K, Monath TP, 2002. Viremia and immunogenicity in nonhuman primates of a tetravalent yellow fever-dengue chimeric vaccine: genetic reconstructions, dose adjustment, and antibody responses against wild-type dengue virus isolates. Virology 298: 146–159.[Medline]
- Guirakhoo F, Kitchener S, Morrison D, Forrat R, McCarthy K, Nichols R, Yoksan S, Duan X, Ermak TH, Kanesa-Thasan N, Bedford P, Lang J, Quentin-Millet MJ, Monath TP, 2006. Live attenuated chimeric yellow fever dengue type 2 (ChimeriVax-DEN2) vaccine: phase I clinical trial for safety and immunogenicity: effect of yellow fever pre-immunity in induction of cross neutralizing antibody responses to all 4 dengue serotypes. Hum Vaccin 2: 60–67.[Web of Science][Medline]
- Kärber G, 1931. Beitrag zur kollektiven behandung pharmacologischer reihenversuche. Collect Treatment Ser Pharmacol Stud 162: 480–484.
- Mantel N, Aguirre A, Gulia S, Girerd-Chambaz Y, Colombani S, Moste C, Barban V, 2008. Standardized quantitative RT-PCR assays for quantitation of yellow fever and chimeric yellow fever-dengue vaccines. J Virol Methods 151: 40–46.[Medline]
- Deauvieau F, Sanchez V, Balas C, Kennel A, De Montfort A, Lang J, Guy B, 2007. Innate immune responses in human dendritic cells upon infection by chimeric yellow-fever dengue vaccine serotypes 1–4. Am J Trop Med Hyg 76: 144–154.[Abstract/Free Full Text]
- Halstead SB, Shotwell H, Casals J, 1973. Studies on the pathogenesis of dengue infection in monkeys. I. Clinical laboratory responses to primary infection. J Infect Dis 128: 7–14.[Web of Science][Medline]
- Halstead SB, Shotwell H, Casals J, 1973. Studies on the pathogenesis of dengue infection in monkeys. II. Clinical laboratory responses to heterologous infection. J Infect Dis 128: 15–22.[Web of Science][Medline]
- Scherer WF, Breakenridge FA, Dickerman RW, 1972. Cross-protection studies and search for subclinical disease in new world monkeys infected sequentially with different immunologic types of dengue viruses. Am J Epidemiol 95: 67–79.[Abstract/Free Full Text]
- Whitehead RH, Chaicumpa V, Olson LC, Russell PK, 1970. Sequential dengue virus infections in the white-handed gibbon (Hylobates lar). Am J Trop Med Hyg 19: 94–102.[Abstract/Free Full Text]
- Price WH, 1968. Sequential immunization as a vaccination procedure against dengue viruses. Am J Epidemiol 88: 392–397.[Abstract/Free Full Text]
- Zhou H, Deem MW, 2006. Sculpting the immunological response to dengue fever by polytopic vaccination. Vaccine 24: 2451–2459.[Web of Science][Medline]
- Edelman R, Wasserman SS, Bodison SA, Putnak RJ, Eckels KH, Tang D, Kanesa-Thasan N, Vaughn DW, Innis BL, Sun W, 2003. Phase I trial of 16 formulations of a tetravalent live-attenuated dengue vaccine. Am J Trop Med Hyg 69 (Suppl): 48–60.[Abstract/Free Full Text]
- Sun W, Nisalak A, Gettayacamin M, Eckels KH, Putnak JR, Vaughn DW, Innis BL, Thomas SJ, Endy TP, 2006. Protection of rhesus monkeys against dengue virus challenge after tetravalent live attenuated dengue virus vaccination. J Infect Dis 193: 1658–1665.[Medline]
TOP
ABSTRACT
INTRODUCTION
