assessment of a vesicular stomatitis virus–based vaccine by use of the mouse model of ebola virus...
TRANSCRIPT
S404 • JID 2007:196 (Suppl 2) • Jones et al.
S U P P L E M E N T A R T I C L E
Assessment of a Vesicular Stomatitis Virus–BasedVaccine by Use of the Mouse Model of Ebola VirusHemorrhagic Fever
Steven M. Jones,1,2,3 Ute Stroher,1,3 Lisa Fernando,1 Xianggou Qiu,1 Judie Alimonti,1 Pasquale Melito,1 Mike Bray,4
Hans-Dieter Klenk,5 and Heinz Feldmann1,3
1Special Pathogens Program, National Microbiology Laboratory, Public Health Agency of Canada, 2Department of Immunology and 3Departmentof Medical Microbiology, University of Manitoba, Winnipeg, Manitoba, Canada; 4Biodefense Clinical Research Branch, National Institutesof Health, Bethesda, Maryland; 5Institute of Virology, Philipps-University, Marburg, Germany
Background. In humans and nonhuman primates, Ebola virus causes a virulent viral hemorrhagic fever forwhich no licensed vaccines or therapeutic drugs exist. In the present study, we used the mouse model for Ebolahemorrhagic fever to assess the safety and efficacy of a vaccine based on a live attenuated vesicular stomatitis virusexpressing the Zaire ebolavirus (ZEBOV) glycoprotein.
Methods. Healthy mice were given the vaccine in various doses, decreasing from to 2 plaque-forming42 � 10units (pfu), with both systemic and mucosal vaccination routes used. Mice were challenged with to lethal3 610 10doses of mouse-adapted ZEBOV. Severely immunocompromised mice were injected with pfu, which is 1052 � 10times greater than the immunization dose normally used, to test vaccine safety.
Results. Two plaque-forming units of the vaccine protected against lethal challenge, and mucosal immunizationwas found to be as protective as systemic injection. The replicating vaccine was never detected in the immunizedanimals, nor were there clinical signs after immunization. Immunization of severely immunocompromised micewith 200,000 pfu of vaccine resulted in no clinical symptoms.
Conclusions. Our data suggest that the vaccine is highly potent and safe and that it very rapidly induces“sterile” immunity in mice. The potential for mucosal delivery, if confirmed in nonhuman primates, makes it anexcellent candidate for mass immunization during outbreaks or in the event of intentional release.
Zaire ebolavirus (ZEBOV) causes highly virulent hem-
orrhagic fever in humans, with a fatality rate approach-
ing 90% [1]. ZEBOV is listed on the Centers for Disease
Control and Prevention’s “A” list as a pathogen that
can be considered to be a bioterrorism threat [2]. No
antiviral therapy or licensed vaccine exists for this dev-
astating disease [3]. It has been shown that Ebola virus
Potential conflicts of interest: none reported.Presented in part: XIIth International Congress of Virology, 27 July to 1 August
2002, Paris, France (abstract V-291); XIIIth International Congress of Virology, 23–28 July 2005, San Francisco, California (abstract 89-V); Filoviruses: RecentAdvances and Future Challenges, International Centre for Infectious DiseasesSymposium, Winnipeg, Manitoba, Canada, 17–19 September 2006.
Financial support: Public Health Agency of Canada; Canadian Institute of HealthResearch (CIHR; grant MOP 43921 to H.F.). Supplement sponsorship is detailed inthe Acknowledgments.
Reprints or correspondence: Dr. Steven M. Jones, 1015 Arlington St., Winnipeg,Manitoba, Canada R3E 3R2 ([email protected]).
The Journal of Infectious Diseases 2007; 196:S404–12� 2007 by the Infectious Diseases Society of America. All rights reserved.0022-1899/2007/19610S2-0040$15.00DOI: 10.1086/520591
(EBOV) is infectious after oral and ocular exposure of
nonhuman primates (NHPs) [4], as well as after aerosol
exposure [5]. Therefore, an effective ZEBOV vaccine
should be capable of rapidly inducing both systemic
and mucosal immunity, because the intentional release
of ZEBOV would probably result in mucosal infection
by small-particle aerosol dispersion [3, 6].
EBOV vaccine research has been reviewed extensively
[1, 7–9]; in general, many of the approaches reviewed
resulted in protective efficacy in rodent models but
failed to protect primates [8]. Studies of immunization
of NHPs by use of a DNA prime/adenovirus boost
approach [10] and, subsequently, a single dose of ad-
enovirus [11] were highly successful. In both of these
studies [10, 11], the genes for the ZEBOV glycoprotein
(GP) and nucleoprotein (NP) were expressed as the
virus-specific immunogens. There are, however, limi-
tations associated with these approaches. First, the ini-
tial vaccination strategy involved a very lengthy DNA
prime/adenovirus boost protocol [10], and, although
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from
VSV Ebola Vaccine Trials in Mice • JID 2007:196 (Suppl 2) • S405
the accelerated vaccination strategy [11] provided protection
within 28 days after immunization with a single dose of ade-
novirus vaccine, there remain concerns related to the significant
rate of preexisting immunity to adenoviruses in humans (up
to 40%–50%) [12–14] and the high dose of recombinant ad-
enovirus required for protection [15–17]. Furthermore, the im-
munization route for both of these vaccines was intramuscular
injection, which is less favorable for rapid mass vaccination
than either oral or intranasal delivery.
Recombinant vesicular stomatitis virus (VSV) has potential
as a vaccine vector because it is capable of high-titer growth
in vitro and is relatively easy to genetically manipulate. Fur-
thermore, it induces strong cellular and humoral immune re-
sponses in vivo [18, 19] and has the potential to elicit both
mucosal and systemic immunity [20]. There is a low incidence
of mostly asymptomatic or mild human infections leading to
a negligible level of preexisting immunity in the world popu-
lation [21]. The efficacy of recombinant VSVs as vaccine vectors
against several infectious agents has been demonstrated else-
where [21–25], including in a study demonstrating the pro-
tective efficacy of live attenuated VSV expressing the HIV en-
velope and core proteins in rhesus monkeys challenged with a
highly pathogenic simian HIV (SHIV89.6P) [26].
We previously described the generation of recombinant live
attenuated vaccines for Ebola, Marburg, and Lassa viruses caus-
ing hemorrhagic fever [27], and we showed that these vaccines
can protect NHPs from otherwise lethal challenge [28–30]. Very
recently, we showed that postexposure administration of the
MARV vaccine protects all NHPs against lethal infection [31]
and that 50% of animals survive when the EBOV vaccine is
given after challenge [32].
Although the NHP is indeed the reference standard model for
efficacy testing of EBOV vaccines, here we demonstrate that,
contrary to previous reports, the mouse can be extremely useful
as an alternative model for testing candidate vaccines for EBOV.
We demonstrated that protection in mice [33] required only a
single dose of as little as 2 pfu. Immunization just 7 days before
an otherwise lethal challenge conferred complete specific pro-
tection, and the vaccine did not result in disease, even in severely
immunocompromised NOD-SCID mice. Importantly, protec-
tion could also be achieved for the first time by mucosal im-
munization, which makes the vaccine vector easy to administer
and, therefore, a promising candidate for mass immunization.
MATERIALS AND METHODS
Vaccine vectors and viruses. The recombinant VSV express-
ing the GP of Zaire EBOV Mayinga (ZEBOV-May) was gen-
erated as recently described elsewhere [27], by use of the in-
fectious clone for VSV Indiana serotype [27]. In brief, the
appropriate open reading frame for the ZEBOV GP was gen-
erated using polymerase chain reaction (PCR), cloned into the
VSV genomic vectors lacking the VSV GP gene, sequence con-
firmed, and rescued using the method described elsewhere [27,
34]. The virus was designated “VSVDG/ZEBOVGP.” Mice were
challenged using a mouse-adapted (MA) variant (9 passages in
mice) derived from ZEBOV-May (MA-ZEBOV-May) [33].
Animal studies. Female BALB/c mice, 5–6 weeks of age, were
purchased from Charles Rivers (Quebec, Canada). Mice (in
groups of 5) were routinely immunized intraperitoneally at 2
sites with pfu of VSVDG/ZEBOVGP (total dose, pfu).4 410 2 � 10
Control animals were immunized with either recombinant wild-
type VSV (rVSVwt) (given in the same dose and by the same
route) or PBS. Mice were challenged intraperitoneally 28 days
after immunization with LD50 of MA-ZEBOV-May.310
The vaccine dose remained constant, independent of the im-
munization route used; however, the volume administered was
altered as follows: for the intranasal route, 20 mL was split between
both nostrils; for the intramuscular route, 100 mL was split be-
tween 2 sites; for the subcutaneous route, 100 mL was admin-
istered in the scruff of the neck; and, for the oral route, 100 mL
was administered by oral gavage (number of animals/group, 10).
To determine the minimum effective vaccine dose stock, VSVDG/
ZEBOVGP was repeatedly diluted 10-fold in unsupplemented
Dulbecco’s modified Eagle medium to attain desired immuni-
zation doses decreasing from to 2 pfu.42 � 10
All animals were weighed for the first 11 days after vaccination
and challenge and were kept 3 times longer than the length of
time to the death of the last control animal (i.e., at least 28 days),
to confirm that protection was complete and that death was not
merely delayed by immunization. NOD-SCID mice ( ) thatn p 5
were 6–8 weeks of age were immunized intraperitoneally at 2
sites with pfu (total dose, pfu) of VSVDG/ZEBOVGP5 510 2 � 10
or VSVDG/MARVGP or were left untreated. Because of the need
for sterile handling conditions, NOD-SCID mice were not
weighed every day; however, they were carefully observed every
day for 14 days and had their clinical symptoms scored.
All animal work was performed according to “Animal Use
Document H-01-001 and H-04-019,” approved by the Animal
Care Committee, Canadian Science Centre for Human and An-
imal Health (Winnipeg, Manitoba), according to the guidelines
of the Canadian Council on Animal Care (Ottawa, Ontario), and
was performed in the biosafety level 4 facility at the National
Microbiology Laboratory.
Virus detection. RNA was isolated from blood and organs
by use of the appropriate RNA isolation kits from Qiagen (Mis-
sissauga, Ontario, Canada). For the detection of VSV, we used
a reverse-transcription (RT)–PCR assay targeting the matrix
gene (nt 2355–2661; GenBank accession no. NC_001560).
ZEBOV RNA was detected using a primer pair targeting the L
gene (nt 13349–13622; GenBank accession no. AF 272001). Live
virus was titrated by means of a TCID50 assay performed on
Vero E6 cells obtained from blood and select tissue samples.
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from
S406 • JID 2007:196 (Suppl 2) • Jones et al.
Figure 1. Protective efficacy of 1-dose vs. 2-dose immunization regi-mens. Mice were immunized intraperitoneally with either 2 doses (solidline with triangles) of pfu of VSVDG/ZEBOVGP (recombinant42 � 10vesicular stomatitis virus in which the native glycoprotein [GP] gene hasbeen deleted and functionally replaced with the Zaire ebolavirus Mayinga[ZEBOV-May] GP gene) given 28 days apart or a single dose (dashed linewith triangles) or were mock immunized with recombinant wild-type ve-sicular stomatitis virus (rVSVwt) (dashed line with squares) and challengedwith LD50 of mouse-adapted ZEBOV Mayinga 28 days after the final310immunization. Weight was recorded for 11 days after immunization andchallenge, and surviving animals were monitored for at least 3 timeslonger than the length of time to the death of the last control animal.
Immune responses. IgG and IgM antibodies against ZE-
BOV were quantified using a modification of an ELISA [35]
using purified virus-like particles [36] as an antigen and horse-
radish peroxidase–conjugated, isotype-specific antibodies to de-
tect mouse antibodies to EBOV. A standard neutralization assay
(PRNT80) was performed as previously described, by measuring
VSVDG/ZEBOVGP plaque reduction in a constant virus:serum
dilution format in a modification of a protocol described else-
where [37, 38]. For this assay, a standard amount of VSVDG/
ZEBOVGP (∼100 pfu) was incubated with serial 2-fold dilu-
tions of the serum sample for 60 min. The mixture was used
to inoculate Vero E6 cells for 60 min. Cells were overlayed with
an agar medium and incubated for 3 days, and plaques were
counted for 48 h after crystal violet staining.
T cell depletion. Mice were immunized intraperitoneally
once with the standard dose of the VSVDG/ZEBOVGP vaccine.
Challenge with LD50 of MA-ZEBOV-May was performed310
28 days after immunization. Mice were injected on days 3 and
1 before challenge and subsequently every 3 days with 100 mL
(0.5 mg/dose) of anti-CD8 antibody (YTS 169.4), in a manner
similar to a protocol described elsewhere [39]. Depletion of
CD8 T cells was confirmed by flow cytometry performed using
antibody that binds to an epitope on CD8 T cells different from
the epitope used for depletion.
Passive transfer of immune serum. On days 1 and 28, mice
were immunized intraperitoneally with a standard dose of
VSVDG/ZEBOVGP vaccine or rVSVwt (in the same dose and
by the same route), and age-matched vaccine-naive animals
were left untreated. Sixty days after the final immunization, the
mice were killed by exposure to terminal anesthesia, and blood
was removed by cardiac puncture. Serum samples obtained
from the 3 groups of mice were injected intraperitoneally into
vaccine-naive mice (0.5 mL/animal) 24 h before challenge with
LD50 of MA-ZEBOV-May.310
RESULTS
Provision of complete protection against EBOV challenge with
106 LD50 given in a single-dose vaccination. For initial effi-
cacy testing, groups of 5 mice were immunized either once (on
day 0) or twice (on days 0 and 14) with rVSVwt or VSVDG/
ZEBOVGP. Mice were challenged 28 days after initial immu-
nization with LD50 of MA-ZEBOV-May. Untreated control310
animals and animals immunized with rVSVwt rapidly lost
weight and died. However, mice immunized with VSVDG/
ZEBOVGP were completely protected against MA-ZEBOV-May
infection, regardless of the number of immunizing doses re-
ceived (figure 1).
To define the absolute upper limit of protection by increasing
the challenge dose in 10-fold increments, mice were immunized
with either rVSVwt or VSVDG/ZEBOVGP on day 0 and were
challenged with MA-ZEBOV-May in the range of to3 610 10
LD50 28 days later. All control animals died after having lost
up to 18% of their body weight in 6 days. In contrast, the
animals immunized with VSVDG/ZEBOVGP were all protected
against MA-ZEBOV-May infection, regardless of the challenge
dose (figure 2A).
Because EBOV is transmitted by close contact and blood
contamination, it was essential to demonstrate that immuni-
zation prevents the establishment of viremia as well as severe
disease or death. Mice were immunized with either rVSVwt or
VSVDG/ZEBOVGP and were challenged 28 days later. Each
day after challenge, 3 mice from each treatment group were
killed, and blood samples were obtained for EBOV isolation
and detection of EBOV-specific RNA by RT-PCR. Animals im-
munized with rVSVwt remained without viremia for 2 days
after challenge but then rapidly developed viremia, which
reached a maximum level of TCID50 before death (6 days710
after infection). At no time after challenge did the mice im-
munized with VSVDG/ZEBOVGP have detectable viremia (fig-
ure 2B).
Protection against EBOV infection provided by very low
doses of vaccine and in the absence of detectable vaccine
replication. Although efficacy is an important concern after
immunization, the safety of the vaccine and the potential for
serious side effects are as important. In general, lower doses of
vaccine are preferred. We determined the minimum dose of
VSVDG/ZEBOVGP required for complete protection of mice.
Animals were immunized with 10-fold dilutions of the vaccine,
starting at pfu and decreasing to 2 pfu, and they un-42 � 10
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from
VSV Ebola Vaccine Trials in Mice • JID 2007:196 (Suppl 2) • S407
Figure 2. Protective efficacy and induction of “sterile” immunity of immunization with a single dose of VSVDG/ZEBOVGP (recombinant vesicularstomatitis virus in which the native glycoprotein [GP] gene has been deleted and functionally replaced with the Zaire ebolavirus Mayinga [ZEBOV-May] GP gene). Mice were immunized intraperitoneally with pfu of VSVDG/ZEBOVGP and were challenged 28 days thereafter with increasing42 � 10doses of mouse-adapted ZEBOV-May (MA-ZEBOV-May). Weight was recorded for a minimum of 11 days after challenge, and surviving animals weremonitored for at least 3 times longer than the length of time to the death of the last control animal. A, Vaccine resulted in protection against highchallenge doses. Animals mock immunized with recombinant wild-type vesicular stomatitis virus (rVSVwt) were not protected from lethal challengewith LD50 (group 1). They rapidly lost weight and died by day 6 after infection with MA-ZEBOV-May. In contrast, mice immunized with VSVDG/210ZEBOVGP and subsequently challenged with increasing doses of MA-ZEBOV-May (group 2, LD50; group 3, LD50; group 4, LD50; and group3 4 510 10 105, LD50 ) did not lose any weight, and all survived. Each group consisted of 5 animals. B, Vaccination induced sterile immunity. Each day for 6610days after challenge, 3 animals from each treatment group were killed to determine levels of viremia. In mice immunized with VSVDG/ZEBOVGP(black bars), no viremia was detectable. In contrast, in rVSVwt-immunized controls (gray bars), infectious MA-ZEBOV-May was detectable by day 3after challenge, and viremia rapidly increased until day 6 (3 animals/group each day). The limit of detection was 2.3 log10 TCID50 and is denoted bythe dotted line. Please note that reverse transcription–polymerase chain reaction also failed to detect MA-ZEBOV-May in blood and tissues samplesobtained from the immunized mice (data not shown).
derwent challenge with LD50 28 days later. All immunized310
animals were protected from the lethal challenge, even at an
ultra-low immunization dose (figure 3A). Because replication—
not only the initial dose—may affect safety, we attempted to
determine the extent of VSV replication in immunized animals.
Immunized and control mice were culled each day between
days 1 and 7 and then every 7 days after immunization until
day 28. Blood, spleen, liver, and kidney samples were obtained
for viral nucleic acid detection and virus isolation. We were
unable to detect virus by means of either RT-PCR or virus
isolation at any time after immunization with either rVSVwt
or VSVDG/ZEBOVGP (data not shown).
Provision of complete protection against systemic challenge
by means of mucosal immunization. The intraperitoneal
route of immunization is safe and relatively easy to perform
under experimental conditions with rodents in high contain-
ment, but it is not a practical route for immunization of hu-
mans. To investigate the effect of different immunization routes
on efficacy, mice were immunized with pfu of VSVDG/42 � 10
ZEBOVGP or rVSVwt by the intramuscular and intranasal
routes. After 28 days, groups of animals were challenged with
10-fold increasing doses of MA-ZEBOV-May (between and310
LD50). All mice were completely protected, with no evidence610
of differences in clinical symptoms or weight changes noted in
any of the mice immunized with VSVDG/ZEBOVGP (figure
3B). All vaccine-naive control or rVSVwt control immunized
mice died within 6 days. Intranasal immunization significantly
increases the potential for rapid mass immunization. The oral
route is the only route with greater potential than the intranasal
route for mass vaccination of susceptible populations. Mice
were immunized with the VSVDG/ZEBOVGP or rVSVwt vac-
cine twice (on days 0 and 28), and animals were challenged 28
days later with MA-ZEBOV-May. All control mice died within
4 days of infection, whereas the immunized animals were pro-
tected (figure 3C).
Cellular and humoral immunity. When human volunteers
cannot ethically be used to test effectiveness of vaccination
strategies, as in the case of EBOV infection, it is possible, under
US Food and Drug Administration regulations, to use animal
models. This requires that immune correlates of protection be
determined in the animal model and that similar immune re-
sponses can be detected in human volunteers in phase 1 or
phase 2 clinical trials. The mechanism for protection against
EBOV infection has not been definitively determined for any
of the animal models; however, there is evidence suggesting
that different pathological mechanisms are important in ro-
dents and primates [8, 40, 41]. To determine the role of an-
tibody and cytotoxic T cell responses in mice, we performed
T cell depletion and serum transfer studies (table 1). Complete
depletion of cytotoxic T cells by use of a specific antibody
against CD8 T cells did not compromise protection, indicating
that, in immunized mice, CD8+ cytotoxic T cells are not an
absolute requirement for protection. Passive transfer of im-
mune serum 60 days after immunization resulted in 80% pro-
tection, but transfer of vaccine-naive or rVSVwt immune serum
failed to provide protection from lethal challenge with MA-
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from
S408 • JID 2007:196 (Suppl 2) • Jones et al.
Figure 3. Protective efficacy depending on dose and route of immunization. Animals were immunized intraperitoneally with decreasing doses ofthe vaccine, as well as by different routes. The challenge was performed with increasing doses of mouse-adapted Zaire ebolavirus Mayinga (MA-ZEBOV-May) 28 days thereafter. A, Ultra-low-dose immunization did not compromise protective efficacy. Mice immunized with recombinant wild-typevesicular stomatitis virus (rVSVwt) (group 1) lost weight and died by day 6 after challenge. In contrast, mice immunized with VSVDG/ZEBOVGP(recombinant vesicular stomatitis virus in which the native glycoprotein [GP] gene has been deleted and functionally replaced with the Zaire ebolavirusMayinga GP gene) in decreasing doses (group 2, pfu; group 3, pfu; group 4, pfu; group 5, pfu; group 6, pfu)4 3 2 1 02 � 10 2 � 10 2 � 10 2 � 10 2 � 10all survived infection with no evidence of clinical disease or weight loss after challenge. Each group consisted of 5 animals, and challenge wasperformed with LD50 of MA-ZEBOV-May. B, Mucosal immunization was as effective as systemic injection. Mice (5 animals/group) were immunized310with a dose of pfu of rVSVwt (group 2) or VSVDG/ZEBOVGP by intraperitoneal injection (group 3), intramuscular injection (group 4), or intranasal42 � 10instillation (group 5). Vaccine-naive controls were left untreated (group 1). Subsequently, mice were challenged with LD50 of MA-ZEBOV-May 28610days after immunization. All rVSVwt-immunized and vaccine-naive control mice displayed clinical symptoms, lost weight, and died. In contrast, VSVDG/ZEBOVGP-immunized mice did not show any clinical symptoms or weight loss. All animals survived, irrespective of challenge dose or route ofimmunization. C, Oral immunization is also effective. Mice (10 animals/group) were immunized with 2 doses of pfu of rVSVwt (group 1) or42 � 10VSVDG/ZEBOVGP (group 2) by oral gavage on days 0 and 28. Subsequently, mice were challenged with LD50 of MA-ZEBOV-May 28 days after310immunization. All rVSVwt-immunized mice displayed clinical symptoms, lost weight, and died within 4 days. In contrast, VSVDG/ZEBOVGP-immunizedmice did not show any clinical symptoms or weight loss.
ZEBOV-May (table 1). Antibody is sufficient to protect mice
from infection, a finding that concurs with results published
elsewhere [42, 43].
After showing that passive transfer of immune serum results
in protection (table 1), we further characterized the humoral
immune response. We used a quantitative ELISA to determine
the concentration of ZEBOV GP–specific antibodies in the im-
munized mice and a plaque reduction neutralization test
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from
VSV Ebola Vaccine Trials in Mice • JID 2007:196 (Suppl 2) • S409
Table 1. Correlates of immunity in BALB/c mice (5 animals/group).
Survivalrate
at day 28,a
%Time to death,
mean days
CD8 T cell depletion 100 NA
No T cell depletion 100 NA
Passive transfer
Of VSVDG/ZEBOVGP immune serum 80 10b
Of rVSVwt immune serum 0 5.0
Of vaccine-naive serum 0 5.33
NOTE. NA, not applicable; rVSVwt, recombinant wild-type vesicular sto-matitis virus; VSVDG/ZEBOVGP, recombinant vesicular stomatitis virus in whichthe native glycoprotein (GP) gene has been deleted and functionally replacedwith the Zaire ebolavirus Mayinga GP gene.
a After challenge.b .n p 1
(PRNT80) to detect neutralizing antibodies. Neutralizingantibody
titers detected by PRNT80 were equivocal, despite the clear pro-
tection afforded by passive immunization (table 1). ELISA data
(table 2) showed that (1) within 7 days of immunization, the
mean concentration of GP-specific IgM was 1.3 mg/mL, whereas
that of IgG2a was 152 ng/mL; and (2) by the time of challenge
in most studies (28 days after immunization), IgG2a levels had
increased to a mean of 781 ng/mL. However, IgM levels had
decreased to 66 ng/mL. The levels of the other IgG isotypes
remained negligible throughout the study (data not shown).
Rapid immunity and long-term protection. The induction
of rapid immunity from challenge is essential if the VSVDG/
ZEBOVGP vaccine vector is to be useful in natural outbreaks
or in the event of an intentional release. To determine the
minimum time required for protective immunity to develop,
mice were immunized at 7-day intervals starting 28 days before
challenge. All VSVDG/ZEBOVGP-immunized animals treated
between days 28 and 7 before challenge were completely pro-
tected from lethal challenge (figure 4A). In contrast, all of the
rVSVwt control animals died of the infection (figure 4B).
Rapid immunity is an important characteristic of an effective
vaccine, but the longevity of protection is also very important.
To test this, we immunized groups of mice with a single dose
of pfu of rVSVwt or VSVDG/ZEBOVGP vaccine. The42 � 10
mice were challenged 3, 6, or 9 months later, along with age-
matched vaccine-naive controls. At each time point, all
VSVDG/ZEBOVGP-vaccinated animals were completely pro-
tected (data not shown).
Infection of NOD-SCID mice with VSV vaccines. Im-
munization of the NOD-SCID mice with 10 times the normal
dose of vaccine given to healthy mice appeared to result in a
self-limiting, asymptomatic infection. At no time were clinical
signs, such as ruffling, hunching, or lack of mobility, observed,
and the infected animals were indistinguishable from the un-
treated control mice (data not shown).
DISCUSSION
In the present study, we utilized the mouse model for ZEBOV
infection to characterize and evaluate the suitability of a re-
combinant live attenuated vaccine candidate (i.e., VSVDG/
ZEBOVGP). The mouse was the only feasible animal model
for such extensive studies [33]. When used appropriately, the
mouse model provides a very useful tool for screening vaccine
candidates and is the only practical tool available for fully ex-
ploring the limits of protection afforded by a vaccine strategy.
Indeed, virtually all observations made in the mouse model
have been confirmed in subsequent studies of NHPs [28, 29,
31, 32].
Animals can be protected from a challenge dose of 1 million
LD50, which is the highest dose ever used in a study of protection
from ZEBOV. This protection was noted even when the vaccine
was delivered mucosally. Importantly, oral and intranasal im-
munization with VSVDG/ZEBOVGP provided protection from
a systemic infection with ZEBOV. This is significant for 2 rea-
sons. First, these are the routes of choice for mass vaccination.
Second, these routes may stimulate mucosal as well as systemic
immunity. Any deliberate attempt to release ZEBOV is likely
to involve the generation of aerosols or contamination of food
supplies. After such an event, protection at the mucosal surfaces
would be important in reducing the dose of virus entering the
body. Because natural infections may also occur by the oral
route or by ocular inoculation, mucosal immunity could be
important in containing transmission in both natural and de-
liberate outbreaks.
We have shown that CD8+ T cells are not required for the
protection of mice and that passive transfer of serum results
in significant specific protection. Further studies involving both
the mouse and NHP models will be required to identify possible
immune correlates of protection. Our data on CD8 T cell de-
pletion contrasts with recently published data, which showed
that (1) after subcutaneous injection of mice with MA-ZEBOV-
May, CD8 T cells were essential for protection, and (2) under
certain experimental conditions, they are sufficient for protec-
tion [44]. ZEBOV infection is an extremely acute disease in
mice, with rapid progression to death (mode time to death, 6
days). CD8 T cell responses generally require 7–12 days to reach
maximum efficiency, and early induction of potent antibody
responses are likely to be essential for effective vaccination.
Despite equivocal in vitro neutralization data, passive transfer
of immune serum to vaccine-naive mice provides protection,
indicating that either neutralizing antibodies are generated or
alternative antibody dependent immune mechanisms are more
important in protection; antibody responses are required and
sufficient for protection. We purposely maintained the actively
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from
S410 • JID 2007:196 (Suppl 2) • Jones et al.
Table 2. IgM and IgG2a antibody response kinetics after immunization of 6 mice.
Antibody, day afterimmunization
Serum antibody level, ng/mL
Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 Mouse 6 Mean � SD
IgM7 1321.4 1355.3 1567.5 1181.3 990.4 … 1238.2 � 214.214 341.4 448.9 299.0 387.4 403.7 … 376.1 � 57.828 23.8 60.3 39.8 167.3 15.6 93.9 66.8 � 56.7
IgG2a7 162.9 151.1 140.0 175.0 133.2 … 152.5 � 16.914 336.5 486.5 1059.6 1404.9 691.7 … 795.9 � 435.328 613.0 958.4 671.2 1154.1 509.0 780.8 781.1 � 238.7
Figure 4. Time required to develop protective immunity. Groups of 5 mice were immunized with either recombinant wild-type vesicular stomatitisvirus (rVSVwt) (dotted lines) or VSVDG/ZEBOVGP (recombinant vesicular stomatitis virus in which the native glycoprotein [GP] gene has been deletedand functionally replaced with the Zaire ebolavirus Mayinga [ZEBOV-May] GP gene) (solid lines) on the following days before challenge with mouse-adapted ZEBOV-May (MA-ZEBOV-May) ( LD50): day 28 (A and B; groups 2 and 6), day 21 (A and B; groups 3 and 7), day 14 (A and B; groups 4310and 8), and day 7 (A and B; groups 5 and 9). Vaccine-naive mice (group 1) were not treated before challenge. All mice immunized with the VSVDG/ZEBOVGP vaccine were completely and specifically protected when immunized as soon as 7 days before challenge (A and B).
immunized mice for 60 days after immunization, to preclude
the transfer of nonspecific immune mediators, resulting from
infection with VSV, in serum.
There has been significant debate over the safety of the VSV
live attenuated recombinant viral platform, compared with the
adenovirus alternative. This debate has focused mainly on the
possibility that VSV infection will lead to significant patholog-
ical findings in immunocompromised individuals. To assess this
possibility, we used the most severely immunocompromised
animal model. The NOD-SCID mouse has no functional T or
B cell responses and is considered to have a nonleaky phenotype
(i.e., T cell and B cell functions are not partially restored over
time, as is seen with some SCID mouse strains). These mice
completely controlled the VSV infection, even when given 10
times the normal vaccine dose. These vaccines apparently are
safe as well as efficacious. In a recent study, toxicological as-
sessment of VSV-based HIV vaccine vectors showed that they
were indeed attenuated in NHPs, and this finding should clear
the way for phase 1 clinical trials in humans [45].
Furthermore, the adenovirus-based ZEBOV vaccine required
the delivery of – particles for effective immunization of10 1210 10
mice and NHPs, respectively [11]. In contrast, mice can be
protected with doses as low as 2 pfu, and NHPs can be protected
with pfu [29] with the VSVDG/ZEBOVGP. The safety of710
adenovirus vectors has been called into question by the death
of 1 person from systemic inflammatory response syndrome
in a gene therapy trial. The steep toxicity curve for replication-
defective adenovirus vectors and the substantial subject-to-sub-
ject variation in host responses to systemically administered
vectors are important [46]. We believe that reducing the initial
dose of vaccine will help to minimize possible adverse effects.
ZEBOV causes the most virulent form of viral hemorrhagic
fever, and control of an epidemic resulting from a maleficent
act or a natural outbreak would be greatly assisted by a vaccine
capable of rapidly inducing strong immunity and an application
route that allows for mass vaccination. The DNA/adenovirus
vaccine and a single-dose adenovirus vaccine delivered the GP
and NP from ZEBOV and induced protective immunity in the
cynomolgus macaque model [10, 11]. There are limitations with
these strategies—in particular, with the length of the immu-
nization scheme associated with the initial DNA prime/ade-
novirus boost protocol [10] and the fact that the preexisting
immunity associated with the accelerated vaccine [11] might
limit the utility of these approaches. The VSV platform has
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from
VSV Ebola Vaccine Trials in Mice • JID 2007:196 (Suppl 2) • S411
advantages, compared with the existing successful vaccination
approaches, including the very low incidence of immunity to
VSV in human populations; the lack of the VSV GP, which is
the main target for preexisting neutralizing antibodies; the ease
of vaccine production; the short time to immunity; and the
potential for mucosal delivery and postexposure treatment.
Last, the potential for use as a mucosal vaccine, particularly by
the oral route of administration, may offer some hope of pro-
tecting the great apes, which are currently under threat from
ZEBOV [47, 48].
Acknowledgments
We thank Daryl Dick, Friederike Feldmann, Allen Grolla, and AndreaPaille for technical assistance and assistance with animal care. We also thankAllison Groseth and Toni Marchione for assistance in preparation of themanuscript. We are grateful to Thomas Geisbert (US Army Medical Re-search Institute of Infectious Diseases, Fort Detrick, Frederick, Maryland),for support and discussions; John Rose (Yale University, New Haven, Con-necticut), for kindly providing the vesicular stomatitis virus infectious clonesystem; and Kent HayGlass (Immunology Department, University of Man-itoba, Winnipeg, Manitoba, Canada), for kindly providing the antibodyused for T cell depletion.
Supplement sponsorship. This article was published as part of a sup-plement entitled “Filoviruses: Recent Advances and Future Challenges,”sponsored by the Public Health Agency of Canada, the National Institutesof Health, the Canadian Institutes of Health Research, Cangene, CUH2A,Smith Carter, Hemisphere Engineering, Crucell, and the International Cen-tre for Infectious Diseases.
References
1. Feldmann H, Jones S, Klenk HD, Schnittler HJ. Ebola virus: fromdiscovery to vaccine. Nat Rev Immunol 2003; 3:677–85.
2. Centers for Disease Control and Prevention. Bioterrorism agents/dis-eases. Available at: http://www.bt.cdc.gov/Agent/Agentlist.asp.Accessed11 October 2007.
3. Borio L, Inglesby T, Peters CJ, et al. Hemorrhagic fever viruses asbiological weapons: medical and public health management. JAMA2002; 287:2391–405.
4. Jaax NK, Davis KJ, Geisbert TJ, et al. Lethal experimental infection ofrhesus monkeys with Ebola-Zaire (Mayinga) virus by the oral andconjunctival route of exposure. Arch Pathol Lab Med 1996; 120:140–55.
5. Johnson E, Jaax N, White J, Jahrling P. Lethal experimental infectionsof rhesus monkeys by aerosolized Ebola virus. Int J Exp Pathol 1995;76:227–36.
6. Bazhutin NB, Belanov EF, Spiridonov VA, et al. The effect of the meth-ods for producing an experimental Marburg virus infection on thecharacteristics of the course of the disease in green monkeys [in Rus-sian]. Vopr Virusol 1992; 37:153–6.
7. Geisbert TW, Jahrling PB. Towards a vaccine against Ebola virus. ExpertRev Vaccines 2003; 2:777–89.
8. Geisbert TW, Pushko P, Anderson K, Smith J, Davis KJ, Jahrling PB.Evaluation in nonhuman primates of vaccines against Ebola virus.Emerg Infect Dis 2002; 8:503–7.
9. Hart MK. Vaccine research efforts for filoviruses. Int J Parasitol2003; 33:583–95.
10. Sullivan NJ, Sanchez A, Rollin PE, Yang ZY, Nabel GJ. Developmentof a preventive vaccine for Ebola virus infection in primates. Nature2000; 408:605–9.
11. Sullivan NJ, Geisbert TW, Geisbert JB, et al. Accelerated vaccination
for Ebola virus haemorrhagic fever in non-human primates. Nature2003; 424:681–4.
12. Chirmule N, Propert K, Magosin S, Qian Y, Qian R, Wilson J. Immuneresponses to adenovirus and adeno-associated virus in humans. GeneTher 1999; 6:1574–83.
13. Kass-Eisler A, Falck-Pedersen E, Elfenbein DH, Alvira M, Buttrick PM,Leinwand LA. The impact of developmental stage, route of adminis-tration and the immune system on adenovirus-mediated gene transfer.Gene Ther 1994; 1:395–402.
14. Kass-Eisler A, Leinwand L, Gall J, Bloom B, Falck-Pedersen E. Cir-cumventing the immune response to adenovirus-mediated gene ther-apy. Gene Ther 1996; 3:154–62.
15. Letvin NL, Huang Y, Chakrabarti BK, et al. Heterologous envelopeimmunogens contribute to AIDS vaccine protection in rhesus monkeys.J Virol 2004; 78:7490–7.
16. Ophorst OJ, Kostense S, Goudsmit J, et al. An adenoviral type 5 vectorcarrying a type 35 fiber as a vaccine vehicle: DC targeting, cross neu-tralization, and immunogenicity. Vaccine 2004; 22:3035–44.
17. Reyes-Sandoval A, Fitzgerald JC, Grant R, et al. Human immunode-ficiency virus type 1-specific immune responses in primates upon se-quential immunization with adenoviral vaccine carriers of human andsimian serotypes. J Virol 2004; 78:7392–9.
18. Wagner RR, Rose JK. Rhabdoviridae: the viruses and their replication.In: Knipe DM, Howley PM, eds. Fields virology. 3rd ed. Vol. 1. Phil-adelphia: Lippincott Williams & Wilkins, 1996:1121–59.
19. Zinkernagel RM, Adler B, Holland JJ. Cell-mediated immunity to ve-sicular stomatitis virus infections in mice. Exp Cell Biol 1978; 46:53–70.
20. Roberts A, Kretzschmar E, Perkins AS, et al. Vaccination with a re-combinant vesicular stomatitis virus expressing an influenza virus hem-agglutinin provides complete protection from influenza virus challenge.J Virol 1998; 72:4704–11.
21. Roberts A, Buonocore L, Price R, Forman J, Rose JK. Attenuated ve-sicular stomatitis viruses as vaccine vectors. J Virol 1999; 73:3723–32.
22. Schlereth B, Rose JK, Buonocore L, ter Meulen V, Niewiesk S. Successfulvaccine-induced seroconversion by single-dose immunization in thepresence of measles virus-specific maternal antibodies. J Virol 2000;74:4652–7.
23. Grigera PR, Marzocca MP, Capozzo AV, Buonocore L, Donis RO, RoseJK. Presence of bovine viral diarrhea virus (BVDV) E2 glycoproteinin VSV recombinant particles and induction of neutralizing BVDVantibodies in mice. Virus Res 2000; 69:3–15.
24. Kahn JS, Roberts A, Weibel C, Buonocore L, Rose JK. Replication-competent or attenuated, nonpropagating vesicular stomatitis virusesexpressing respiratory syncytial virus (RSV) antigens protect miceagainst RSV challenge. J Virol 2001; 75:11079–87.
25. Buonocore L, Blight KJ, Rice CM, Rose JK. Characterization of vesicularstomatitis virus recombinants that express and incorporate high levelsof hepatitis C virus glycoproteins. J Virol 2002; 76:6865–72.
26. Rose NF, Marx PA, Luckay A, et al. An effective AIDS vaccine basedon live attenuated vesicular stomatitis virus recombinants. Cell 2001;106:539–49.
27. Garbutt M, Liebscher R, Wahl-Jensen V, et al. Properties of replication-competent vesicular stomatitis virus vectors expressing glycoproteinsof filoviruses and arenaviruses. J Virol 2004; 78:5458–65.
28. Daddario-DiCaprio KM, Geisbert TW, Geisbert JB, et al. Cross-pro-tection against Marburg virus strains by using a live, attenuated re-combinant vaccine. J Virol 2006; 80:9659–66.
29. Geisbert TW, Jones S, Fritz EA, et al. Development of a new vaccinefor the prevention of Lassa fever. PLoS Med 2005; 2:e183.
30. Jones SM, Feldmann H, Stroher U, et al. Live attenuated recombinantvaccine protects nonhuman primates against Ebola and Marburg vi-ruses. Nat Med 2005; 11:786–90.
31. Daddario-DiCaprio KM, Geisbert TW, Stroher U, et al. Postexposureprotection against Marburg haemorrhagic fever with recombinant ve-sicular stomatitis virus vectors in non-human primates: an efficacyassessment. Lancet 2006; 367:1399–404.
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from
S412 • JID 2007:196 (Suppl 2) • Jones et al.
32. Feldmann H, Jones SM, Daddario-DiCaprio KM, et al. Effective post-exposure treatment of Ebola infection. PLoS Pathog 2007; 3:e2.
33. Bray M, Davis K, Geisbert T, Schmaljohn C, Huggins J. A mouse modelfor evaluation of prophylaxis and therapy of Ebola hemorrhagic fever.J Infect Dis 1998; 178:651–61.
34. Schnell MJ, Buonocore L, Kretzschmar E, Johnson E, Rose JK. Foreignglycoproteins expressed from recombinant vesicular stomatitis virusesare incorporated efficiently into virus particles. Proc Natl Acad Sci USA1996; 93:11359–65.
35. Bullifent HL, Griffin KF, Jones SM, Yates A, Harrington L, Titball RW.Antibody responses to Yersinia pestis F1-antigen expressed in Salmonellatyphimurium aroA from in vivo-inducible promoters. Vaccine 2000;18:2668–76.
36. Wahl-Jensen V, Kurz SK, Hazelton PR, et al. The role of Ebola virussecreted glycoproteins and virus-like particles in activation of humanmacrophages. J Virol 2005; 79:2413–9.
37. Jahrling PB. Filoviruses and arenaviruses. In: Murray PR, ed. Manualof clinical microbiology. Washington, DC: ASM Press, 1999:1125–36.
38. Jahrling PB, Geisbert TW, Geisbert JB, et al. Evaluation of immuneglobulin and recombinant interferon-a2b for treatment of experimen-tal Ebola virus infections. J Infect Dis 1999; 179(Suppl 1):S224–34.
39. Cobbold SP, Martin G, Qin S, Waldmann H. Monoclonal antibodiesto promote marrow engraftment and tissue graft tolerance. Nature1986; 323:164–6.
40. Bray M, Hatfill S, Hensley L, Huggins JW. Haematological, biochemicaland coagulation changes in mice, guinea-pigs and monkeys infected
with a mouse-adapted variant of Ebola Zaire virus. J Comp Pathol2001; 125:243–53.
41. Gibb TR, Bray M, Geisbert TW, et al. Pathogenesis of experimentalEbola Zaire virus infection in BALB/c mice. J Comp Pathol 2001; 125:233–42.
42. Gupta M, Mahanty S, Bray M, Ahmed R, Rollin PE. Passive transferof antibodies protects immunocompetent and immunodeficient miceagainst lethal Ebola virus infection without complete inhibition of viralreplication. J Virol 2001; 75:4649–54.
43. Parren PW, Geisbert TW, Maruyama T, Jahrling PB, Burton DR. Pre-and postexposure prophylaxis of Ebola virus infection in an animalmodel by passive transfer of a neutralizing human antibody. J Virol2002; 76:6408–12.
44. Gupta M, Mahanty S, Greer P, et al. Persistent infection with Ebolavirus under conditions of partial immunity. J Virol 2004; 78:958–67.
45. Johnson JE, Nasar F, Coleman JW, et al. Neurovirulence properties ofrecombinant vesicular stomatitis virus vectors in non-human primates.Virology 2007; 360:36–49.
46. Raper SE, Chirmule N, Lee FS, et al. Fatal systemic inflammatoryresponse syndrome in a ornithine transcarbamylase deficient patientfollowing adenoviral gene transfer. Mol Genet Metab 2003; 80:148–58.
47. Leroy EM, Rouquet P, Formenty P, et al. Multiple Ebola virus trans-mission events and rapid decline of central African wildlife. Science2004; 303:387–90.
48. Vogel G. Conservation biology: can great apes be saved from Ebola?Science 2003; 300:1645.
at Aston U
niversity on September 6, 2014
http://jid.oxfordjournals.org/D
ownloaded from