Live Chimeric and Inactivated Japanese Encephalitis Virus Vaccines Differ in Their Cross-Protective Values against Murray Valley Encephalitis Virus

Abstract

The Japanese encephalitis virus (JEV) serocomplex, which also includes Murray Valley encephalitis virus (MVEV), is a group of antigenically closely related, mosquito-borne flaviviruses that are responsible for severe encephalitic disease in humans. While vaccines against the prominent members of this serocomplex are available or under development, it is unlikely that they will be produced specifically against those viruses which cause less-frequent disease, such as MVEV. Here we have evaluated the cross-protective values of an inactivated JEV vaccine (JE-VAX) and a live chimeric JEV vaccine (ChimeriVax-JE) against MVEV in two mouse models of flaviviral encephalitis. We show that (i) a three-dose vaccination schedule with JE-VAX provides cross-protective immunity, albeit only partial in the more severe challenge model; (ii) a single dose of ChimeriVax-JE gives complete protection in both challenge models; (iii) the cross-protective immunity elicited with ChimeriVax-JE is durable (≥5 months) and broad (also giving protection against West Nile virus); (iv) humoral and cellular immunities elicited with ChimeriVax-JE contribute to protection against lethal challenge with MVEV; (v) ChimeriVax-JE remains fully attenuated in immunodeficient mice lacking type I and type II interferon responses; and (vi) immunization with JE-VAX, but not ChimeriVax-JE, can prime heterologous infection enhancement in recipients of vaccination on a low-dose schedule, designed to mimic vaccine failure or waning of vaccine-induced immunity. Our results suggest that the live chimeric JEV vaccine will protect against other viruses belonging to the JEV serocomplex, consistent with the observation of cross-protection following live virus infections.

Murray Valley encephalitis virus (MVEV) is a mosquito-borne flavivirus belonging to the Japanese encephalitis virus (JEV) serocomplex which can cause severe, sometimes fatal, disease in humans (reviewed in references 30, 31, 32, and 42). The virus is endemic in northern Australia and Papua New Guinea, where it causes a small number of human cases of encephalitis in most years. In symptomatic patients the case fatality rate is ∼20%, and among those who recover a large number (∼50%) will suffer from neuropsychiatric sequelae. Cases of Murray Valley encephalitis are more common in children or visitors in areas of endemic disease than in adult residents, who have preexisting immunity (7, 42, 46). Sporadically, MVEV spreads to central or southern regions of Australia (e.g., the Murray Valley of southeastern Australia) and causes epidemic viral encephalitis in humans (32). There are no vaccines or antiviral agents available against MVEV, and given the relatively small number of human cases, it is unlikely that a MVEV-specific vaccine for human use will be produced. However, it has been known for many years that at least in animal models, live viral infection with other members of the JEV serocomplex will give cross-protective immunity against heterologous viruses belonging to this group (10, 17, 33, 48, 52). MVEV is genetically and antigenically closely related to JEV (82% amino acid sequence identity in the envelope [E] protein), the most important encephalitic flavivirus in terms of human disease incidence and severity (reviewed in reference 4). A number of live and inactivated JEV vaccines have been licensed or are under development (reviewed in references 2, 16, and 34). If effective and long-lasting cross-protective immunity against MVEV was induced by one of the JEV vaccines, a strong case could be made for its prophylactic use in populations at risk of MVEV infection in Australia. A further reason for investigating the suitability of JEV vaccines in the Australian context is the recent emergence of JEV in northern Australia (18, 19, 41). This has raised the prospect that JEV may become established in enzootic cycles on the Australian mainland, necessitating the use of JEV vaccines in regions where MVEV is also endemic. The impact of MVEV infection in JEV vaccine recipients in terms of disease outcome remains unknown.

In contrast to its protective value against heterologous flaviviruses, cross-reactive flavivirus immunity has also been associated with infection- and/or disease-enhancing consequences in natural and laboratory settings (1, 9, 20, 39). Antibody-dependent enhancement of infection is thought to account for the more severe forms of dengue sometimes associated with secondary, heterologous dengue virus infections by a mechanism putatively involving the increased uptake of virus bound with nonneutralizing antibody into Fc receptor-bearing cells (14, 15). For the MVEV/JEV pair, it has been reported that transfer of subneutralizing concentrations of JEV-immune serum or sera from mice suboptimally immunized with inactivated JEV vaccine (JE-VAX; Biken, Japan) can prime recipient mice for a more severe disease when challenged with MVEV (3, 50). We have demonstrated this potentially detrimental effect for the first time in the context of the full complement of the vaccine-primed immune response: the administration of an experimental UV-inactivated MVEV vaccine at a suboptimal dose greatly increased the susceptibility of mice (up to 75% mortality) to challenge with a dose of JEV, which was sublethal in unvaccinated animals (29). It is not clear if this phenomenon is an inherent property of inactivated vaccines, which provide relatively poor immunity in terms of quality, magnitude, and duration in comparison to live virus infections. Here we investigate the protective value and risk of disease potentiation of a recombinant, live JEV vaccine candidate (ChimeriVax-JE) and a licensed, inactivated JEV vaccine (JE-VAX) in mouse models of MVEV and West Nile virus (WNV) encephalitis. ChimeriVax-JE is constructed from yellow fever virus 17D vaccine cDNA by replacement of the viral structural prM and E proteins with those of an attenuated JEV strain; it has been shown to protect mice and monkeys from JEV challenge (12, 36) and has undergone phase 2 and phase 3 trials for safety and efficacy in humans (35, 37).

MATERIALS AND METHODS

Virus.

Working stocks of MVEV (strain MVE-1-51), JEV (strain Nakayama), and WNV (strain Kunjin MRM61C) were 10% suckling mouse brain homogenates in Hanks' balanced salt solution containing 0.2% bovine serum albumin and HEPES (pH 8.0; 20 mM final concentration) (HBSS-BSA). ChimeriVax-JE (12) was propagated in Vero cells grown in M199 medium (Invitrogen, Carlsbad, CA) supplemented with 5% fetal bovine serum and antibiotics. The clarified infected culture fluid had a titer of 1 × 10⁷ Vero cell PFU and was frozen in single-use aliquots. Accordingly, the ChimeriVax-JE vaccine used in this investigation differs from that used in the phase 2 and phase 3 vaccine trials by one additional Vero cell passage, although this is unlikely to have resulted in phenotypic changes, in view of the reported genetic stability of the virus during sequential passage in Vero cells (12). Chimeric vaccine and virus stock dilutions for inoculation into mice were in HBSS-BSA.

Mice, vaccination, and challenge.

C57BL/6 (B6) mice, alpha interferon (IFN-α) receptor knockout (IFN-α-R^−/−) mice (38), and IFN-α/γ-R^−/− mice (49) (both knockout mouse strains are on the 129 background) were bred under specific-pathogen-free conditions and supplied by the Animal Breeding Facility at the John Curtin School of Medical Research, ANU, Canberra. Groups of 6-week-old male or female mice were immunized by the subcutaneous (s.c.) route with JE-VAX (Sanofi Pasteur Inc., Swiftwater, PA), which was prepared as recommended by the supplier. JE-VAX immunization schedules involved priming and two booster vaccinations given at 2-week-intervals (reviewed in references 16 and 34). Vaccination with ChimeriVax-JE was by the intravenous (i.v.) or intraperitoneal (i.p.) route, which differed from that (s.c. route) used for immunization with ChimeriVax-JE in preclinical and clinical studies with humans and nonhuman primates (35-37). The vaccine doses used are given in the figure legends and ranged from 10 to 10⁵ PFU of ChimeriVax-JE and ∼6 to 600 ng, or 1/1,000 to 1/10 of a human dose, of JE-VAX. Mice were challenged at 14 weeks or at a later age by i.v. or i.p. injection of virus doses as described in the corresponding figure legends. A high dose (10⁸ PFU, i.v.) of MVEV was used for challenge of B/6 mice, while a low virus dose (10² to 10³ PFU) was used for challenge of IFN-α-R^−/− and IFN-α/γ-R^−/− mice with JEV, MVEV, or WNV (strain Kunjin). The latter dose was previously shown to consistently produce 100% mortality in IFN-α-R^−/− mice (25, 28).

Serological tests.

For titration of MVEV- and JEV-reactive antibody in mouse sera, enzyme-linked immunosorbent assays (ELISA) were performed with horseradish peroxidase-conjugated goat anti-mouse immunoglobulin (Dako, Carpinteria, CA) and the peroxidase substrate 2,2′-azino-di(3-ethyl-benzthiasoline sulfonate) (Kirkegaard and Perry Laboratories, Gaithersburg, MD) as described previously (6). JEV and MVEV NS1 protein-specific ELISA and 50% plaque reduction neutralization assays (PRNT₅₀) were as described previously (29). For determination of ELISA end point titers, optical density (OD) cutoff values were established as the mean OD of eight negative control wells containing sera of naïve mice plus 3 standard deviations. OD values of test sera were considered positive if they were equal to or greater than the OD cutoff and end point titers calculated as log₁₀ of the reciprocal of the last dilution giving a positive OD value. For preparation of MVEV NS1 ELISA antigen, CV1 cells in a 175-cm² tissue culture flask were infected (multiplicity of infection of ∼0.5) with the MVEV NS1 protein-expressing vaccinia virus recombinant VV-NS1 (13) for 2 days and lysed with 1 ml of NP-40 lysis buffer (0.5% Nonidet P-40, 50 mM Tris-HCl [pH 7.5], 150 mM NaCl, and 2 mM EDTA, containing 20 μg of phenylmethylsulfonyl fluoride per ml) for 30 min on ice. Nuclei were removed by centrifugation at 5,000 × g for 5 min at 4°C, NaN₃ was added, and the lysate was stored at 4°C. It was used at a 1/200 dilution for coating ELISA trays. Negative control antigen for use in MVEV NS1-specific ELISA was produced by infection of CV1 cells with wild-type vaccinia virus and processing of infected cell lysates as described above. Immune sera did not show reactivity on negative control plates coated with vaccinia virus-infected cell lysates.

Virus titration.

The virus content in sera and brains of infected mice was determined by plaque assay on Vero cell monolayers as described previously (27).

Transfer experiments.

Six-week-old IFN-α-R^−/− mice were vaccinated with 1 × 10⁵ PFU ChimeriVax-JE and were sacrificed a month later for aseptic removal of spleens and blood extraction. Spleens were placed in cold minimal essential medium plus 5% fetal calf serum (FCS) after dissection. Serum samples were isolated after centrifugation for 10 min at 10,000 rpm, heat inactivated for 30 min at 56°C, and stored at −70°C. Immune serum transfer into 10-week-old IFN-α-R^−/− mice was by the i.p. route with the volumes of serum and times of transfer shown in Table 3.

TABLE 3.

Humoral and cellular immune responses contribute to heterotypic protection by ChimeriVax-JE against MVEV

Challenge route and treatmenta	No. of survivors/total	ATD, days (P)b
i.v.
None	0/8	5.9
Immunization (105 PFU ChimeriVax-JE)	5/5	NA
Immune serum (300 μl)c	0/4	9.5 (0.003)
Immune splenocytes (1 × 107 cells)	0/8	7.4 (0.0004)
Immune splenocytes (5 × 107 cells)	0/6	7.3 (0.001)
s.c.
None	0/5	6.8
Immunization (105 PFU ChimeriVax-JE)	11/11	NA
Immune serum (300 μl)d	0/6	9.7 (0.006)
Naive splenocytes (2 × 107 cells)	0/10	6.8
Immune splenocytes (2 × 107 cells)	0/8	7.8 (0.002)
Immune serum (300 μl)d plus splenocytes (2 × 107 cells)	4/10	12.8 (0.0006)
Immune serum (300 μl)d plus T cellse (1 × 107 cells)	2/6	9.0 (0.002)
Immune serum (300 μl)d plus T cellsf (5 × 106 cells)	1/5	7.8 (0.01)

^aSix-week-old IFN-α-R^−/− mice were immunized with 10⁵ PFU of ChimeriVax-JE i.p. or left untreated. These donor mice were sacrificed 4 weeks later, and serum and spleen cells were harvested and transferred to 10-week-old IFN-α-R^−/− recipients, which were challenged with 10² PFU MVEV i.v. or by injection into the footpad (s.c.). Age-matched naïve and ChimeriVax-JE-immunized mice were included as controls.

^bThe ATD was analyzed for significance using the Wilcoxon signed-rank test. For mice challenged i.v., the ATD was compared to that of the naive group. For mice challenged s.c., the effect of serum transfer only on ATD was compared to that of the naïve group, while the effect of splenocyte transfer with or without immune serum on ATD was compared to that of naïve splenocyte recipient mice. NA, not applicable.

^cImmune serum transfer schedule: 150 μl on day 1 prechallenge and 150 μl on day 1 postchallenge.

^dImmune serum transfer schedule: 100 μl on day 1 prechallenge and 50 μl on days 0, 1, 3, and 5 postchallenge.

^eThe transferred cell population contained 90% T cells with 5% B-cell contamination.

^fThe transferred cell population contained 95% T cells with 1% B-cell contamination.

Single-cell splenocyte suspensions were prepared by pressing the spleen tissue gently through a fine metal mesh tissue sieve. Erythrocyte lysis was by suspension of the splenocyte pellet in 4.5 ml distilled water followed immediately by the addition of 0.5 ml 10× phosphate-buffered saline (PBS). Lysed cells were discarded after centrifugation at 400 × g for 5 min. Splenocytes were resuspended in 1× PBS and injected through the lateral tail veins of 10-week-old IFN-α-R^−/− recipient mice. Recipient mice were challenged a day later with 1 × 10² PFU MVEV via footpad injection. For T-cell enrichment, 1.2 × 10⁸ splenocytes suspended in HBSS plus 10% FCS were loaded onto 10-ml nylon wool columns and incubated for 45 min at 37°C. Effluent from the columns was collected, and cells were pelleted by centrifugation at 400 × g for 5 min. This procedure resulted in a population of 90% T cells with 5% B-cell contamination. To further improve the purity of T cells eluted from nylon wool columns, the T-cell-enriched splenocyte population was incubated with anti-CD45R/B220 supernatant (RA2-3A1) in minimal essential medium plus 5% FCS for 30 min at 4°C, followed by incubation with rabbit serum complement (Cedarlane Laboratories Ltd.) for 30 min at 37°C. Cells were washed twice with PBS before transfer into recipient mice. This procedure gave rise to a population of 95% T cells with 1% B-cell contamination.

Statistics.

Differences in survival ratios in mouse challenge experiments were assessed using Fisher's exact test. The Mann-Whitney and Wilcoxon signed-rank tests were applied to assess differences in antibody titers and average time to death (ATD), respectively, for significance.

RESULTS

Immunization against JEV protects against a high-dose challenge with MVEV in B6 mice.

In a previous study we have shown that an inactivated experimental MVEV vaccine can elicit increased disease severity in mice challenged with JEV and that this phenomenon becomes apparent in low-dose vaccine recipients (29). Here, the converse combination of JEV vaccination and MVEV challenge was addressed using the licensed, inactivated JEV vaccine JE-VAX (reviewed in references 16 and 34) and the live, experimental yellow fever virus/JEV prM-E chimeric vaccine ChimeriVax-JE (12). Six-week-old B6 mice were immunized three times at 2-week intervals with either undiluted (corresponding to 1/10 of a recommended human dose) or 100-fold-diluted JE-VAX or, in a second experiment, inoculated once with a high dose (10⁵ PFU) or a low dose (10 PFU) of ChimeriVax-JE and challenged with 10⁸ PFU of MVEV i.p. at 14 weeks of age (Fig. 1A). Both JEV vaccines provided protection against MVEV in the high-dose groups: in each case, mortality was reduced to 7%, in comparison to 25% and 60% mortality in groups of MVEV-challenged naïve mice used as controls for JE-VAX (P = 0.33) and ChimeriVax-JE (P = 0.01) immunizations, respectively. The experiments in the left and right panels of Fig. 1A were not carried out concurrently, and the difference in mortality in MVEV-infected naïve control groups (and as a consequence the difference in significance in the protective values of the two vaccines) is a reflection of the variability of MVEV and JEV challenge models in adult immunocompetent mice, as also reported previously (29), due to factors which remain unclear. To overcome this variability, a more stringent MVEV challenge model using mice deficient in type I IFN responses, which gives reproducible mortality outcomes, was developed and used subsequently (see below). The cross-protective values of the two vaccines were dose dependent, given that the low-dose vaccination regimens did not significantly reduce mortality (Fig. 1A). No disease potentiation in terms of increased mortality was noted among the low-dose vaccine recipients, although direct and indirect measures of viral load revealed such an effect in low-dose JE-VAX-immunized mice (see below).

FIG. 1.

Susceptibility of B6 mice immunized with low and high doses of JE-VAX or ChimeriVax-JE to challenge with MVEV. Groups of 6-week-old B6 mice were immunized with three doses of undiluted (∼0.6 μg) or 100-fold-diluted JE-VAX s.c., given at 2-week intervals, or with one dose of 10⁵ or 10¹ PFU of ChimeriVax-JE i.v. At 14 weeks of age, mice were challenged with 10⁸ PFU of MVEV i.v. (A) Mortality. Groups of mice were monitored twice daily for morbidity and mortality for 21 days. (B) Anti-MVEV NS1-specific antibody titers in prechallenge sera and sera of surviving mice at 28 days post-MVEV challenge as an indirect measure of viral load. Horizontal bars show mean end point titers, the dotted line indicates the detection limit, and asterisks indicate significance of difference in antibody titer relative to the unvaccinated group of mice (*, P < 0.03; **, P < 0.003).

Homologous and cross-reactive antibody responses in B6 mice vaccinated against JEV.

A single inoculum of ChimeriVax-JE (10⁵ PFU) was superior to three doses of undiluted JE-VAX in producing homologous and MVEV cross-reactive antibody responses with virus neutralizing activity, with the proviso that different routes for delivery of the two vaccines were used (Table 1). It is noteworthy that only the high dose of ChimeriVax-JE provided a detectable neutralizing antibody response against MVEV, which was at a threshold considered to be indicative of protection from disease (22).

TABLE 1.

Antibody responses in B6 and IFN-α-R^−/− mice vaccinated with high or low doses of JEV vaccines

Mice and vaccinea	Dose (no./group)	Mean log10 antibody titer (range)b		Mean PRNT50 titerc
		Anti-JEV	Anti-MVEV	Anti-JEV	Anti-MVEV
B6
JE-VAX	Undiluted (15)	3.8 (2.9-4.1)	3.2 (2.9-3.8)	40	<10
	1/100 (15)	<2.0 (<2.0)	<2.0 (<2.0)	<10	<10
ChimeriVax-JEV	105 PFU (14)	4.7 (2.9-5.0)	3.8 (2.6-4.1)	160	10
	10 PFU (15)	<2.0 (<2.0-2.3)	<2.0 (<2.0)	10	<10
IFN-α-R−/−
JE-VAX	Undiluted (10)	3.5 (2.6-3.8)	3.1 (2.6-3.5)	20	NT
	1/100 (10)	<2.0 (<2.0-2.0)	<2.0 (<2.0)	<10	NT
ChimeriVax-JEV	105 PFU (8)	5.3 (4.4-5.6)	3.7 (3.5-4.1)	320	80
	103 PFU (10)	4.7 (4.1-5.0)	3.6 (2.9-4.1)	320	20
	10 PFU (9)	4.4 (2.9-4.7)	2.2 (<2.0-2.6)	20	10

^aSix-week-old mice were immunized with three doses of JE-VAX s.c., given 2 weeks apart, or with one dose of ChimeriVax-JE i.v.

^bOD cutoff values were established as the mean OD of eight negative control wells containing sera of naïve mice plus 3 standard deviations. OD values of test sera were considered positive if they were equal or greater than the OD cutoff and end point titers calculated as log₁₀ of the reciprocal of the last dilution giving a positive OD value.

^cThe percent plaque reduction with pooled sera was calculated relative to virus controls incubated with naïve mouse serum, and titers are given as the reciprocals of serum dilutions which resulted in >50% reduction of the number of plaques. NT, not tested.

Effect of JEV vaccines on viral load following MVEV challenge.

Antibody against the NS1 protein of the challenge virus is indicative of infection in vaccinated individuals (6, 24, 45) if the vaccine lacks NS1 (in case of JE-VAX, which consists of purified, formalin-inactivated virion particles) or encodes NS1 from a distantly related flavivirus (in case of ChimeriVax-JE, which encodes the yellow fever virus nonstructural proteins). The magnitude of the NS1-specific antibody response is a measure of the viral burden following virus challenge (6, 29). Figure 1B shows the anti-MVEV NS1 antibody titers in pre- and 4-weeks-post-MVEV challenge sera from groups of B6 mice vaccinated against JEV or left untreated. Immunization with three doses of undiluted JE-VAX or with 10⁵ PFU ChimeriVax-JE did not provide sterilizing immunity against MVEV; however, both immunizations markedly reduced the viral load in surviving mice following challenge as reflected in the significant reduction of NS1 antibody titers in comparison to those for naïve mice (P = 0.009 and 0.002, respectively). This was consistent with the levels of protection against MVEV found for the two groups (Fig. 1A). In contrast, recipients of low-dose (1/100) JE-VAX vaccine, which did not succumb to challenge, showed evidence of enhancement of MVEV infection, given that NS1-specific antibody titers were significantly higher than in untreated mice (P = 0.02). Vaccine-induced potentiation of MVEV infection was not apparent following immunization with a low dose (10 PFU) of ChimeriVax-JE, and anti-MVEV NS1 antibody titers in prechallenge sera from JEV vaccine recipients were at or below the detection threshold. Replication of the challenge virus was needed for the high levels of MVEV NS1-specific antibodies observed after MVEV challenge: NS1 protein in the MVEV inoculum contributed only to a minor extent to the antibody titers measured, given that control vaccination with 10⁸ PFU of UV-inactivated MVEV (29) produced low (≤2.3 log units) NS1-specific antibody end points. Attempts to measure viral loads in sera of MVEV-infected immunized or naïve B6 mice at days 2 and 3 postchallenge by nested real-time reverse transcription-PCR (detection limit = 2,000 copies/0.1 ml serum) were unsuccessful.

Protective value of JEV vaccines against MVEV in mice deficient in IFN responses.

Mice defective in type 1 IFN responses (IFN-α-R^−/− mice) (38) are highly susceptible to infection with encephalitic flaviviruses, where low-dose virus inocula are uniformly lethal independent of the age of the animals (25, 26, 28, 43). Accordingly, these mice represent a severe challenge animal model for evaluation of the protective efficacy of flavivirus vaccines. To demonstrate that vaccination against JEV can give protection against the homologous virus in IFN-α-R^−/− mice, 6-week-old animals were immunized with three doses of undiluted JE-VAX at 2-week intervals and challenged 4 weeks after the second booster immunization with 100 PFU of JEV (Fig. 2A). All vaccinated animals survived the virus challenge, while all naïve mice died by day 7 postinfection (p.i.), confirming that the deficiency in the IFN response did not prevent the establishment of protective immunity in animals vaccinated against JEV.

FIG. 2.

Comparison of the protective values of inactivated and live JEV vaccines against homologous and heterologous virus challenge in mice defective in IFN responses. Groups of 6-week-old mice were immunized with JE-VAX (three immunizations with different antigen concentrations, as indicated, given at 2-week intervals) or ChimeriVax-JE (one immunization with different virus doses, as indicated) or left untreated and challenged at 14 weeks of age with 10² PFU of JEV, MVEV, or WNV i.v. (A) IFN-α-R^−/− mice immunized with JE-VAX and challenged with JEV; (B) IFN-α-R^−/− mice immunized with JE-VAX and challenged with MVEV; (C) IFN-α-R^−/− mice immunized with ChimeriVax-JE and challenged with MVEV; (D) IFN-α/γ-R^−/− mice (deficient in both type I and type II IFN responses) immunized with ChimeriVax-JE and challenged with MVEV; (E) IFN-α-R^−/− mice immunized with ChimeriVax-JE and challenged with WNV (strain Kunjin). Mice were monitored twice daily for morbidity and mortality for 21 days.

Only limited cross-protective immunity was elicited by the high-dose JE-VAX immunization regimen in IFN-α-R^−/− mice (Fig. 2B): 3 out of 10 mice survived MVEV challenge, and 10- and 100-fold-lower doses of vaccine gave even poorer protection. In contrast, vaccination with ChimeriVax-JE established excellent cross-protective immunity against MVEV in IFN-α-R^−/− mice (Fig. 2C). Single doses of 10⁵ or 10³ PFU of the vaccine gave 100% and 90% protection (P < 0.0001), and the lowest dose of ChimeriVax-JE used allowed one out of nine animals to survive. Finally, we tested the cross-protective value of ChimeriVax-JE against MVEV in mice defective in both type I and type II IFN responses (due to the knockout of the IFN-α and -γ receptors; IFN-α/γ-R^−/− mice) (49). Infection with 10² PFU of MVEV i.v. was uniformly lethal in these mice, and the ATD was 1 to 2 days earlier than that for IFN-α-R^−/− mice infected with the same dose of MVEV (Fig. 2D). When 6-week-old IFN-α/γ-R^−/− mice were immunized with 10⁵ PFU of ChimeriVax-JE, all animals were protected against MVEV challenge at 6 weeks after the immunization (Fig. 2D) (P = 0.003). It is noteworthy that ChimeriVax-JE is sufficiently attenuated such that no signs of disease were apparent following infection of the immunodeficient mice with the recombinant virus.

Antibody responses in IFN-α-R^−/− mice vaccinated against JEV.

ChimeriVax-JE elicited significantly higher antibody titers against JEV and MVEV than JE-VAX in IFN-α-R^−/− mice (Table 1). Notably, low doses of the live vaccine (10³ and 10 PFU) were more or equally effective in eliciting homologous and heterologous flavivirus immunity in comparison to the high-dose immunization regimen with the inactivated vaccine. The efficient induction of humoral immunity with low-dose ChimeriVax-JE vaccinations was more pronounced in IFN-α-R^−/− mice than in immunocompetent B6 mice (Table 1); for instance, 10³ PFU of the vaccine induced twofold-higher homologous and cross-reactive PRNT₅₀ titers in the type I IFN response-defective mice than a 100-fold-higher vaccine dose in B6 mice. This effect was most likely due the more vigorous and prolonged growth of the live vaccine in the immunodeficient mice than in B6 mice.

Vaccination with ChimeriVax-JE protects against lethal WNV challenge.

To investigate whether the live, chimeric JEV vaccine also elicits cross-protective immunity against a more distantly related virus belonging to the JEV serocomplex, immunized IFN-α-R^−/− mice were challenged with the Kunjin strain of WNV (Fig. 2E). The WNV strain is more virulent than JEV and MVEV in the type I IFN-deficient mice: infection with 10² PFU of WNV Kunjin i.v. resulted in earlier mortality (ATD = 4.4 days; 100% mortality) than infection with the same dose of JEV or MVEV (6.4 and 8.0 days, respectively) (Fig. 2). All mice vaccinated with a single dose of 10⁵ PFU of ChimeriVax-JE survived the lethal WNV challenge. The shorter survival time of IFN-α-R^−/− mice challenged with WNV in comparison to MVEV is consistent with the more efficient growth of WNV than MVEV in extraneural mouse tissues, reflected by up to 1,000-fold-higher viremia (25, 28), and provides an explanation for our failure to detect MVEV in extraneural tissues of infected, immunocompetent B/6 mice (see above and reference 27), in contrast to the case for WNV (51).

Durability of protective immunity against MVEV in ChimeriVax-JE-immunized mice.

To investigate whether the cross-protective immune response elicited with ChimeriVax-JE is long lived, IFN-α-R^−/− mice were immunized with 10⁵ PFU ChimeriVax-JE, immunized with three high doses of JE-VAX, or left untreated and challenged with JEV or MVEV at 23 or 19 weeks after termination of the vaccination regimens, respectively (Table 2). Both vaccines gave complete protection against JEV. However, only ChimeriVax-JE provided complete protection against MVEV (P = 0.008), while JE-VAX-immunized mice were not significantly resistant to lethal challenge with the heterologous virus. The prechallenge antibody ELISA titers against JEV were ≥10-fold greater in mice that had been inoculated with one dose of ChimeriVax-JE than in those that had received three doses of JE-VAX (P = 0.0001) (Fig. 3); similarly, the prechallenge PRNT₅₀ titers of pooled sera from ChimeriVax-JE-immunized mice were ∼10-fold higher than those for JE-VAX-immunized mice against JEV (1,280 and 160, respectively) and MVEV (160 and 20, respectively). Both vaccines elicited humoral immunity, which was relatively stable for several months (compare antibody titers in Fig. 3 with those in Table 1). ChimeriVax-JE induced durable immunity against JEV and MVEV in IFN-α-R^−/− mice, which to a great extent suppressed virus growth following challenge: NS1 protein-reactive antibodies were mostly undetectable in JEV-challenged mice, while all JE-VAX-immunized mice produced readily detectable levels of antibody against NS1 (P = 0.011) (Fig. 3B), suggesting significant replication of the challenge virus in the latter group. Similarly, at 3 weeks after MVEV challenge of ChimeriVax-JE-immunized mice, anti-MVEV NS1 protein-reactive antibody was found in only one mouse (ELISA end point titer, 2.9), while the other mice in the group (n = 4) did not show an increase in antibody titer against MVEV NS1 relative to those found in the prechallenge sera (ELISA end point titers, ≤2.0).

TABLE 2.

Durability of the protective immunity induced with JE-VAX and ChimeriVax-JE in homologous and heterologous virus challenges

Vaccinea	% Mortality (ATD, days) after challenge withb:
	JEV	MVEV
JE-VAX	0	80 (7.3)
ChimeriVax-JE	0	0
None	100 (5.0)	100 (6.2)

^aSix-week-old IFN-α-R^−/− mice were immunized with one dose of 10⁵ PFU of ChimeriVax-JE i.v. or with three (∼0.3-μg) doses of JE-VAX, given at 2-week intervals, s.c. or were left untreated.

^bAt 23 or 19 weeks after completing the immunization schedule with ChimeriVax-JE or JE-VAX, respectively, groups (n = 5) of mice were challenged i.p. with 10³ PFU of JEV or MVEV, and morbidity and mortality were monitored twice daily for 21 days.

FIG. 3.

Durability of humoral immunity induced with JE-VAX and ChimeriVax-JE. Groups of 6-week-old IFN-α-R^−/− mice were immunized three times at 2-week intervals with high doses (∼0.3 μg) of JE-VAX s.c. or with one dose of 10⁵ PFU of ChimeriVax-JE i.v. (A) At 28 weeks of age, prechallenge serum samples from individual mice in each group (n = 10) were collected and JEV-specific antibody titers determined; mean titers calculated from ELISA end points ± standard deviations are shown. (B) At 29 weeks of age, groups of five mice were challenged with 10³ PFU of JEV i.p. (protection data are given in Table 2), and 3 weeks later, sera from surviving mice were collected. The anti-JEV NS1 protein-specific antibody responses in prechallenge (n = 10) and postchallenge (n = 5) sera were measured, and mean titers calculated from ELISA end points ± standard deviations are shown.

Humoral and cellular immune responses contribute to the heterotypic protection by ChimeriVax-JE against MVEV.

The efficient neutralizing antibody responses against JEV and MVEV elicited in ChimeriVax-JE-immunized IFN-α-R^−/− mice (Table 1 and Fig. 3) are consistent with a primary role of humoral immunity in protection. To investigate whether cell-mediated immune responses add to the cross-protective value of the vaccine against MVEV, ChimeriVax-JE immune splenocytes were transferred to naïve mice with or without the additional transfer of immune serum. Recipient mice were challenged with MVEV (10² PFU) by the i.v. or s.c. route (Table 3); the ATD of naïve mice infected with MVEV via the footpad was delayed by ∼1 day in comparison to that following i.v. challenge. Transfer of ChimeriVax-JE immune splenocytes (1 × 10⁷ or 5 × 10⁷ cells) did not provide protection from lethal challenge by either of the two routes but significantly prolonged survival by 1 to 2 days in the recipient groups relative to untreated mice (Table 3). Naïve splenocyte transfer did not alter ATD and mortality. Immune serum transfer (300 μl total given at intervals pre- and postchallenge [Table 3]) also failed to protect animals from lethal MVEV infection but delayed the ATD by ∼3 days in comparison to that in the untreated controls. PRNT₅₀ titers of pooled immune sera used in the transfer experiments were in the ranges of 320 to 640 against JEV and 20 to 40 against MVEV and thus were similar to the values shown in Table 1. Mice were partially protected against lethal MVEV challenge following transfer of immune serum plus splenocytes, demonstrating the beneficial role of cell-mediated immunity elicited with ChimeriVax-JE in heterotypic protection. This was at least in part mediated by T cells, given that B-cell-depleted splenocytes maintained some protective value when transferred in the presence of immune serum, although this effect was marginal.

Evidence for immune enhancement of MVEV infection in low-dose JE-VAX-immunized mice.

We have provided evidence that the viral load in B6 mice challenged with MVEV was significantly higher in recipients of 100-fold-diluted JE-VAX than in unvaccinated mice, using the correlate of magnitude of antibody responses specific for NS1 of the challenge virus (Fig. 1B). However, given the relatively low and variable mortality rate and the absence of detectable MVEV titers in extraneural tissues and sera of infected adult B6 mice (27), it was not possible to obtain direct proof of vaccine-induced infection enhancement reflected in increased viral burden. In contrast to the case for B6 wild-type mice, viremia and extraneural virus growth can be observed in adult IFN-α-R^−/− mice, even when they are infected with a low dose of MVEV (28). Thus, IFN-α-R^−/− mice were immunized with three doses of undiluted or 100-fold-diluted JE-VAX at 2-week intervals (or left untreated) and challenged 4 weeks later with 100 PFU of MVE i.v., and the magnitude and kinetics of viremia and virus growth in the brain were measured (Fig. 4). Earlier viremia was found in mice vaccinated with low doses of JE-VAX than in unvaccinated mice: on day 1 p.i., viremia ranging from 10² to 10³ PFU/ml serum was found in 6 out of 10 vaccine recipients but remained undetectable in the unvaccinated control group; on day 2 p.i., all mice in the JE-VAX (1/100) group had detectable viremia ranging from 6 × 10² to 3 × 10³ PFU/ml serum, while 6 out of 14 unvaccinated mice had no detectable viremia (P = 0.007); and on day 4 p.i., viremia was seen in all mice from the unvaccinated and the low-dose JE-VAX-immunized groups but the mean virus titer was significantly higher in the former than in the latter group (1.8 × 10⁵ and 3 × 10³, respectively; P = 0.04), suggesting earlier antibody-mediated removal of virus from the circulation. Viremia was cleared to mostly undetectable levels in all groups of mice by day 6 p.i.; immunization with high doses of JE-VAX prevented the occurrence of detectable viremia in all mice except in one animal on day 4 p.i. (Fig. 4A).

FIG. 4.

Enhancement of MVEV infection in IFN-α-R^−/− mice suboptimally immunized with JE-VAX. (A) Serum; (B) brain. Groups of 6-week-old IFN-α-R^−/− mice were immunized three times at 2-week intervals with undiluted (∼0.6 μg) or 100-fold-diluted JE-VAX, or left untreated, and challenged with 10² PFU of MVEV i.v. at 14 weeks of age. Blood and brains were harvested at the times indicated and processed as described previously (27), and virus content was measured by plaque assay on Vero cells. Dotted lines indicated detection limits; each symbol corresponds to an individual mouse. For the calculation of mean titers, which are represented by horizontal bars, serum and brain samples with values below the detection limit were set at 20 and 200 PFU/ml or g, respectively.

In the low-dose JE-VAX-immunized IFN-α-R^−/− mice, the kinetics of virus entry into the brain was more rapid and virus titers were significantly higher than in the untreated group of mice following infection with MVEV (Fig. 4B), although this was not reflected in a markedly reduced average survival time (Fig. 2B). At 5 days p.i., 83% of mice in the JE-VAX (1/100) group showed virus in the brain (titers always exceeded those found in blood), in contrast to only 55% of mice in the unvaccinated group, and mean titers were significantly higher in the former (1.2 × 10⁷/g tissue) than in the latter group (7.6 × 10³ PFU/g tissue) (P = 0.005). This difference remained apparent on day 6 p.i.: mean brain titers for the JE-VAX (1/100) and unvaccinated groups were 1.5 × 10⁸ and 2.9 × 10⁷, respectively (P = 0.03), where 6 of 13 mice in the latter group remained free of detectable virus in the brain. We found high levels of immunoglobulin M antibody against MVEV at days 5 and 6 p.i. in all mice for which results are presented in Fig. 4B, showing that the animals were infected. Consistent with its cross-protective value against MVEV, the high-dose vaccination regimen with JE-VAX mostly prevented virus invasion and growth in the brain (Fig. 4B). In groups of IFN-α-R^−/− mice immunized with 10¹ (n = 6), 10³ (n = 3), or 10⁵ (n = 6) PFU of ChimeriVax-JE, MVEV challenge (100 PFU, i.v.) at 7 weeks after immunization did not result in detectable virus titers in the brain, except in one mouse in the 10³ PFU ChimeriVax-JE group (6 × 10³ PFU on day 6 p.i.).

DISCUSSION

This study has addressed protective and potentially disease-enhancing properties of the cross-reactive immunity induced by live and inactivated JEV vaccines in the immunized mouse when challenged with a second virus belonging to the JEV serocomplex. Insights into this phenomenon are applicable to public health to guide recommendations on whether (i) JEV vaccines that are licensed or in clinical development can be used for immunization against closely related flaviviruses for which vaccines are not available and (ii) the cross-reactive immunity elicited by a given JEV vaccine is a risk factor for disease potentiation when vaccine recipients are infected with a second JEV serocomplex flavivirus, a complication documented for secondary dengue virus infection (14, 15). We used two mouse models of flaviviral encephalitis to investigate these questions: high-dose extraneural challenge of adult B6 mice (27, 51) and low-dose challenge of mice deficient in type I IFN responses (28).

High-dose challenge of 14-week-old B6 mice with MVEV resulted in mortality between days 6 and 12 p.i., suggesting that in some mice virus in the inoculum directly breached the blood-brain-barrier, as shown previously (27), giving rise to early deaths, while in others virus replication in extraneural tissues occurred before virus entered the central nervous system, causing deaths at a later time. Both the JE-VAX and ChimeriVax-JE vaccines gave dose-dependent protection against MVEV, suggesting that, at least in some B6 mice, virus cross-reactive antibody was present at sufficient magnitude to prevent virus entry into the brain. Notably, only ChimeriVax-JE produced neutralizing antibody against MVEV, of low but detectable titers. Cross-reactive memory B and CD4⁺ T cells may also have played a protective role, whereas CD8⁺ T cells are detrimental in recovery from MVEV infection in B6 mice (27). Three immunizations with high doses of JE-VAX (∼0.6 μg, which corresponds to 1/10 of the dose recommended for use in humans) were needed to protect mice against MVEV. A 10-fold reduction in antigen dose resulted in loss of detectable neutralizing antibody against JEV as well as protection from homologous virus challenge (data not shown). A level of cross-protection which was at least comparable to that of the high-dose JE-VAX immunization schedule was achieved with a single inoculum of 10⁵ PFU of ChimeriVax-JE. A two-dose schedule of ChimeriVax-JE would be expected to further increase and broaden the cross-reactive immunity against MVEV and other related flaviviruses (11). This raises the prospect that the live JEV vaccine may be suitable for protection of humans against MVEV. A significant increase in the recommended JE-VAX antigen dose or addition of additional booster immunizations to the vaccination regimen in order to augment the level of cross-protective immunity elicited with the inactivated vaccine would not seem practical.

A pronounced difference in the cross-protective values of JE-VAX and ChimeriVax-JE against MVEV was found in the more severe IFN-α-R^−/− mouse model. While both vaccines provided complete protection in homologous challenge with JEV, only immunization with ChimeriVax-JE elicited solid and durable protective immunity against heterologous challenge with MVEV and WNV: even at 23 weeks after immunization with a single dose of ChimeriVax-JE, IFN-α-R^−/− mice were resistant to MVEV challenge, with persistence of high neutralizing antibody titers, which greatly restricted or even prevented growth of the challenge viruses (based on absent or low levels of anti-JEV and -MVEV NS1 antibody after challenges). This was not the case for mice vaccinated with JE-VAX on a high-dose schedule, which were susceptible to MVEV challenge but remained resistant to JEV infection at 19 weeks postimmunization; the level of antibody against JEV in prechallenge sera was ∼10-fold lower than that in ChimeriVax-JE-immunized mice and did not give sterile immunity against JEV. The ability of ChimeriVax-JE to replicate in IFN-α-R^−/− mice more efficiently than in immunocompetent mice most likely explains the very high antibody titers against JEV and MVEV in the immunodeficient animals and could be considered to favor the immunogenicity of live vaccines over that of killed vaccines in this animal model. However, given that ChimeriVax-JE produces detectable viremia in humans and nonhuman primates (35-37) but not in mice, the immunogenicity of the vaccine in IFN-α-R^−/− mice may resemble that in human vaccine recipients more closely than that in immunocompetent mice would.

The magnitude of vaccine-induced humoral immunity in IFN-α-R^−/− mice correlated with cross-protection, where an ∼10-fold-greater antibody titer was found in ChimeriVax-JE-immunized mice, which were protected against MVEV and WNV challenge, than in JE-VAX-immunized mice, which remained susceptible to heterologous flavivirus challenge. This conclusion was consistent with the finding that transfer of ChimeriVax-JE immune serum failed to protect against MVEV challenge, given that the final anti-JEV antibody concentration in the recipient animals would have been ∼1/10 of that in the vaccinated donor mice and therefore at levels achieved with high-dose JE-VAX immunization, which did not protect. In addition to the vital role of antibody, cell-mediated immunity elicited with ChimeriVax-JE also contributed to cross-protection against MVEV. While ChimeriVax-JE immune splenocyte transfer significantly prolonged survival of lethally challenged mice, cotransfer of total and, to a lesser extent, B-cell-depleted splenocytes plus serum from ChimeriVax-JE-vaccinated donor mice provided partial protection of recipient mice. This finding is consistent with a recent publication demonstrating a contributory role of T cells in supplementing the humoral immune response elicited with inactivated or DNA-based vaccines against WNV in protection against homologous virus challenge (44).

The IFN-α-R^−/− mouse challenge model unambiguously differentiated between the cross-protective values of ChimeriVax-JE and JE-VAX. Its advantage over other mouse models is the exquisite sensitivity of adult IFN-α-R^−/− mice to low-dose virus challenge by an extraneural route. However, the mice are fully or partially resistant to infection with attenuated strains of JEV, MVEV, and WNV (5, 25, 26), and the mouse model corroborates the high level of virulence attenuation of ChimeriVax-JE (12). A dose of 10⁵ PFU of the vaccine i.v. did not produce morbidity or mortality in 6-week-old mice, even when both type I and II IFN responses were defective, although younger, 4-week-old IFN-α/γ-R^−/− mice were partially susceptible to extraneural infection with the vaccine (two of three mice died on day 9 p.i. [data not shown]).

Live virus infection has been shown to elicit cross-protective immunity between viruses belonging to the JEV serocomplex in hamsters (17, 48), pigs (52) monkeys (10), and birds (8). This contrasts with the risk of infection enhancement in secondary, heterologous dengue virus infection (14, 15). The amino acid sequence divergence in the E protein, the dominant antigen which harbors most of the epitopes recognized by virus neutralizing antibodies, is slightly greater within the dengue virus (62 to 77% identity) than within the JEV serocomplex (72 to 93% identity) (4). However, it is uncertain whether the level of genetic difference alone accounts for the potential of disease enhancement in one case but not in the other. On the other hand, we have recently shown evidence of enhancement of infection and disease in secondary JEV infection in mice suboptimally immunized with inactivated MVEV (29). This potentially detrimental effect of an inactivated vaccine was reproduced in this study using JE-VAX and MVEV challenge. We documented a significantly increased virus burden in suboptimally immunized B6 mice and earlier viremia with more rapid virus entry into the central nervous system accompanied by higher brain virus titers in vaccinated type I IFN-deficient mice, relative to untreated animals, following challenge with MVEV. Importantly, the infection enhancement became apparent only when vaccine failure, which occurs in a small portion of human recipients of JE-VAX (21), or waning of vaccine-induced immunity was mimicked by using a low-antigen dose vaccination schedule. While Broom et al. (3) also observed infection enhancement using transfer of JE-VAX-immune serum into mice subsequently challenged with MVEV, a second group demonstrated partial protection against WNV challenge in JE-VAX-immunized animals (47). However, the outcome of the latter study does not contradict our observations, because low-dose JE-VAX immunizations were not attempted. In humans, the recommended three-dose immunization series with JE-VAX did not give detectable neutralizing antibody against WNV (23) unless primary vaccination was coadministered with a second, live vaccine against yellow fever virus (53), which could have had an adjuvant effect (40). It remains uncertain if our laboratory observation of infection enhancement and disease potentiation upon heterologous virus exposure of mice immunized with inactivated JEV or MVEV antigen is also a risk factor in vaccinated humans.

In summary, our study with mice suggests that vaccination against JEV can cross-protect against lethal MVEV challenge and raises the prospect that Australian encephalitis in humans may be preventable by vaccination against JEV. Comparisons of the cross-reactive immunities elicited with ChimeriVax-JE and JE-VAX in mouse models of flaviviral encephalitis predict that the live chimeric vaccine will be safer and more effective in averting severe MVEV infection in humans than the inactivated vaccine.