Download Self-Replicative RNA Vaccines Elicit Protection against Influenza A

1395
CONCISE COMMUNICATION
Self-Replicative RNA Vaccines Elicit Protection against Influenza A Virus,
Respiratory Syncytial Virus, and a Tickborne Encephalitis Virus
Marina N. Fleeton,1 Margaret Chen,1,3 Peter Berglund,1
Gary Rhodes,4 Suezanne E. Parker,5 Marie Murphy,6
Gregory J. Atkins,6 and Peter Liljeström1,2
1
Microbiology and Tumorbiology Center, Karolinska Institutet,
and 2Department of Vaccine Research, Swedish Institute
for Infectious Disease Control, Stockholm, and 3Department
of Immunology, Microbiology, Pathology, and Infectious Diseases,
Huddinge University Hospital, Karolinska Institutet, Huddinge, Sweden;
4
Department of Medical Pathology, University of California, Davis,
5
and Vical, San Diego, California; 6Department of Microbiology, Moyne
Institute of Preventive Medicine, Trinity College, Dublin, Ireland
In genetic vaccination, recipients are immunized with antigen-encoding nucleic acid, usually
DNA. This study addressed the possibility of using the recombinant alpha virus RNA molecule, which replicates in the cytoplasm of transfected cells, as a novel approach for genetic
vaccination. Mice were immunized with recombinant Semliki Forest virus RNA–encoding
envelope proteins from one of 3 viruses: influenza A virus, a tickborne flavivirus (louping ill
virus), or respiratory syncytial virus (RSV). Serologic analyses showed that antigen-specific
antibody responses were elicited. IgG isotyping indicated that predominantly Th1 type immune responses were induced after immunization with RSV F protein–encoding RNA, which
is relevant for protection against RSV infection. Challenge infection showed that RNA immunization had elicited significant levels of protection against the 3 model virus diseases.
The well-documented efficacy of live virus vaccines against
vaccinia, polio, measles, mumps, and rubella demonstrated the
potency of in vivo antigen synthesis for induction of protective
immune responses. The discovery that naked plasmid DNA
could elicit broad humoral and strong cellular immunity has
prompted its use in the design of new vaccines against many
viruses for which no vaccines currently exist [1]. In vitro–made
mRNA also has been used for gene delivery in vivo and has
been injected directly in naked form, administered by gene gun
or in complex with lipids [2–4]. However, immune responses
induced after vaccination with nonreplicative mRNAs have
been weaker than responses after plasmid immunization.
Recently, naked RNA transcribed in vitro from an attenuated
tickborne encephalitis (TBE) cDNA clone was used for vaccination. Since this RNA constituted a full-length infectious
genomic molecule, vaccination led to production and spread of
attenuated virus, thus mimicking a flavivirus infection [5]. One
Received 23 October 2000; revised 29 January 2001; electronically published 30 March 2001.
Financial support: Swedish Medical Research Council, Swedish Council
for Engineering Sciences, European Union Biotechnology Programme, and
Wellcome Trust and BioResearch Ireland.
Reprints or correspondence: Dr. Peter Liljeström, Microbiology and
Tumorbiology Center, Karolinska Institutet, S-171 77 Stockholm, Sweden
([email protected]).
The Journal of Infectious Diseases 2001; 183:1395–8
䉷 2001 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2001/18309-0012$02.00
consequence is that the safety of this approach must be considered equivalent to that of a live attenuated virus vaccine.
In this study, we endeavored to combine the safety of mRNA
immunization with the efficiency of live virus vaccination by
using a recombinant (r) self-replicating RNA, based on the
Semliki Forest virus (SFV) replicon, as a vaccine. The bipartite
division of SFV-coding sequences into nonstructural and structural regions allows the replacement of viral structural genes
with the gene coding for the heterologous antigen while retaining the self-replicative capacity of the molecule. The suicidal
nature of alpha viral replicons provides for transient antigen
production and leads to rapid removal of the nucleic acid from
tissue [6]. A preliminary study showed that humoral responses
could be elicited after immunization with rSFV RNA [7]. Here
we tested the ability of self-replicating rSFV RNA to elicit
antibody responses to 3 virus spike proteins. We examined
whether vaccination with rSFV RNA induced protection in
mice against challenge with the corresponding viruses: respiratory syncytial virus (RSV), influenza A virus (FLU), or louping ill virus (LIV), a tickborne flavivirus belonging to the TBE
virus complex.
Materials and Methods
Plasmid templates. Plasmids pSFV-LacZ, pSFV-hemagglutinin
(HA), and pSFV-prME, encoding Escherichia coli b-galactosidase
protein, FLU HA protein (strain PR/8), and LIV prME protein,
respectively, have been described elsewhere [8, 9]. Plasmid pSFVF, which encodes the fusion (F) envelope protein from RSV, was
1396
Fleeton et al.
constructed by subcloning the F gene of RSV from the plasmid
pGEM-LF1 (provided by J. A. Melero, Centro Nacional de Biologı́a Fundamental, Instituto de Salud Carlos III, Madrid) into the
SFV vector.
Production of rSFV RNA and formalin-inactivated (FI) RSV for
vaccination. In vitro transcription of RNA from linearized plasmid vectors has been described elsewhere [8]. To confirm antigen
expression in cell culture, BHK-21 cells were electroporated with
recombinant RNAs, followed by indirect immunofluorescence with
specific monoclonal antibodies. Antibody specific for FLU HA
protein was purchased from Virostat. Antibody 4.2 specific for LIV
prME protein has been described elsewhere [9]. Antibody F47 specific for RSV F protein was provided by J. A. Melero. We prepared
FI-RSV in aluminum hydroxide gel adjuvant (Superfos Biosector)
by using purified RSV, according to the procedures used in the
1969 vaccine trial [10].
Immunization of mice. We used female Balb/c mice (6–10 weeks
old) in all experiments (MTC breeding unit, Karolinska Institutet;
Harlan UK). Mice were injected intramuscularly with 50 mL per
hind leg with 10 mg in total of recombinant RNA in 150 mM NaCl.
Negative control mice were immunized with saline or rSFV-LacZ
RNA. For FI-RSV, mice were immunized intramuscularly, as described elsewhere [10], and received booster immunizations 2 weeks
after priming. Serum samples were taken from immunized mice 2
weeks after priming and after booster immunizations, for determination of antibody titers. Challenge experiments were done 2
weeks after booster immunization.
Serologic assays. Serum samples were tested by ELISA for
antigen-specific IgG. We used purified b-galactosidase protein
(Boehringer Mannheim) or heat-inactivated virus from an egggrown influenza A (PR/8) virus stock to detect anti-LacZ [11] or
anti-HA [8] antibodies, respectively. Antibodies specific for LIV
prME proteins were detected by whole cell ELISA, as described
elsewhere [9]. We used 5 mg/mL sucrose gradient–purified RSV
(Long strain) to detect anti–F antibodies by RSV ELISA. Alkaline
phosphatase or horseradish peroxidase–conjugated secondary antibodies were used to determine titers of total IgG or IgG subtypes
(IgG1 or IgG2a), respectively. The assay cutoff value was determined by adding 3 times the SD to the mean optical density of
serum samples from control mice.
FLU challenge. Mouse-adapted influenza virus strain A/PR/8
(FLU) was propagated in lungs of adult Balb/c mice. For intranasal
challenge, mice were anesthetized with methofane and were inoculated on the nares with 50 mL of a suspension containing a previously determined LD90 for the virus [12]. Mice were monitored
daily for 30 days after challenge.
LIV challenge. LIV strain LI/31 was prepared from clarified
brain homogenates obtained from intracerebrally inoculated neonatal Balb/c mice [9]. The mice were challenged subcutaneously
with 100 LD50 of challenge virus and were monitored daily for 3
weeks after challenge.
RSV challenge. RSV (Long strain) was propagated in MA104
cells [10]. Mice were challenged intranasally with 2.2 ⫻ 10 6 pfu of
RSV. On day 4, the mice were killed, and their lungs were homogenized in Dulbecco’s MEM (Gibco-BRL) containing 10% fetal
calf serum, 2 mM glutamine, 20 mM HEPES, nonessential amino
acids, penicillin, and streptomycin sulfate in a 10% (wt/vol) suspension. The cleared homogenates were serially diluted and were
JID 2001;183 (1 May)
monitored by plaque assay on MA104 cells, as described elsewhere
[10]. The detection level of the RSV plaque assay was 200 pfu/g
of tissue. To calculate geometric mean titers, any sample containing
undetectable RSV levels was given a designated value of 200.
Statistical methods. We used Fisher’s exact probability test (2tailed) to compare survival data from the FLU and LIV challenge
experiments and Student’s t test (2-tailed) to compare RSV lung
titers between experimental versus control groups in the RSV challenge experiment.
Results
Antibody response after immunization with naked rSFV
RNA. Recombinant RNAs encoding 3 viral spike proteins
(HA of FLU, prME envelope proteins of LIV, or F envelope
protein of RSV) were injected intramuscularly into mice. A
single inoculation of 10 mg of naked rSFV RNA resulted in
induction of significant serum antibody titers in ∼50% of immunized mice. All mice immunized with RNA seroconverted
after 2 immunizations (figure 1A–1C). To determine whether
the plasmid DNA template in the inoculum could lead to an
antibody response, control mice were immunized with transcription mixture, including the plasmid DNA template coding
for E. coli LacZ but lacking the SP6 RNA polymerase enzyme.
These mice did not seroconvert, which indicates that the transcribed RNA, not the plasmid DNA template, was the active
immunogen (data not shown).
To determine the serum antibody isotype profile induced by
vaccination with rSFV-F RNA, we analyzed serum levels of
IgG2a and IgG1 (figure 1D). We detected significantly more
IgG2a than IgG1 (P ! .001). This contrasted with mice immunized with FI-RSV, which had high levels primarily of IgG1
antibodies in serum samples.
Protection from viral challenge.
Balb/c mice immunized
twice with naked rSFV RNA were challenged with LIV, FLU,
or RSV. Significant levels of protection were observed in each
of the 3 viral challenge models (figure 2). Of mice immunized
with rSFV-prME, 70% were protected from a LIV challenge
dose that killed 100% of mice in control groups (P p .0015).
Of mice immunized with rSFV-HA, 90% were protected from
a FLU challenge that killed 90% of control mice (P p .0011).
Mice immunized with rSFV-F RNA were significantly protected
from RSV challenge (P ! 10⫺7). A 32-fold reduction in RSV
lung titers in rSFV-F RNA–immunized animals (geometric
mean RSV titer, 821) was observed, compared with mice immunized with rSFV-HA RNA or saline (geometric mean RSV
titer, 26,868).
Discussion
This study describes a generic strategy to vaccinate with recombinant RNA. This self-replicating alpha viral RNA molecule consists of 2 open-reading frames, one encoding the antigen of interest, the other encoding the alpha viral replicase.
JID 2001;183 (1 May)
Self-Replicative RNA Vaccines
1397
be used for vaccination by mode of infection [8, 9]. Layered
DNA/RNA plasmid vectors based on alpha viral replicons have
also been used for immunization studies [11, 13]. Vaccination
with such recombinant particles or plasmid vectors resulted in
strong humoral and cellular responses, sustained immunologic
memory, and protection from disease or death. In this study,
microgram amounts of naked RNA were sufficient to induce
significant antibody responses in vaccinated mice against 3 struc-
Figure 1. A–C, Serum IgG antibody responses induced in individual
mice after intramuscular injection of 10 mg of recombinant Semliki
Forest virus (rSFV) RNA coding for viral spike proteins prME of
louping ill virus (LIV), hemagglutinin (HA) protein of influenza A virus
(FLU), or fusion (F) protein of respiratory syncytial virus (RSV), respectively. 䡺, IgG titers 2 weeks after immunization 1; 䡵, IgG titers
in same mice 2 weeks after immunization 2 (2 weeks after priming).
Values below dashed lines are reciprocal serum IgG titers below the
detection limit of the ELISA (!50). Mice immunized with saline or
rSFV-LacZ RNA were seronegative for viral antigens tested. D, Analysis of serum IgG1 and IgG2a responses in mice immunized twice with
rSFV-F RNA or formalin-inactivated RSV (FI-RSV). Bars are ratios
between IgG2a and IgG1 ELISA titers in serum samples of individual
mice immunized with RNA (black bars) or from pooled serum samples
of 5 mice immunized with FI-RSV (white bar).
The replicase functions to propagate the antigen-encoding
mRNA inside the transfected host cell. This replication process
affects an immune response in several ways. First, as a result
of higher antigen production in vivo, it may allow for the induction of effective immune responses with lower input doses
of RNA than those required for conventional mRNA vaccination [2]. Second, the amplification process, which mimics viral
infection, results in the synthesis of double-stranded RNA intermediates that may operate as a natural adjuvant to the immune system.
Previous studies of recombinant alpha viral RNAs used helper
cells to package the RNA replicons into virus particles that could
Figure 2. Viral challenge of mice immunized twice with recombinant
Semliki Forest virus (rSFV) RNA. A, Mice immunized twice with
rSFV-prME RNA (䡵), rSFV-LacZ RNA (䡺), or saline (䢇) were challenged with louping ill virus (LIV) 2 weeks after last immunization. B,
Mice immunized twice with rSFV-hemagglutinin (HA) RNA (䡵) or
saline (䡺) and challenged with influenza A virus (FLU) 2 weeks after
last immunization. C, Mice immunized twice with rSFV-F RNA (䡵),
rSFV-HA RNA, or saline (䡺) were challenged 3 weeks later with respiratory syncytial virus (RSV). Four days after challenge, RSV titers
in lung tissue were analyzed. Data are logarithmic reciprocal RSV titers
in tissue of individual mice. Detection limit of the plaque assay (dashed
line) was 200 pfu/g of lung tissue. FLU challenge was done 3 times
and RSV and LIV experiments 2 times—all with results shown.
1398
Fleeton et al.
turally different viral spike proteins: prME, HA, or RSV-F. We
observed that levels of serum antibodies elicited after vaccination
were not as high as those elicited when mice were immunized
with rSFV particles or rSFV DNA coding for the same antigens
[8, 9, 11, 13]. However, the responses induced were sufficient to
induce protective immunity from peripheral challenge with LIV
and intranasal challenge with FLU or RSV. Similar to vaccination with other alpha viral vaccine vectors, a predominantly
Th1 type immune response was induced, as determined by levels
of IgG2a and IgG1 antibody isotypes in serum samples. This is
particularly important for vaccination against RSV, in which a
Th2 type immune response, such as that elicited after FI-RSV
vaccination, is associated with immunopathology and more severe lung disease after infection [14].
In the present study, humoral immune responses were demonstrated; however, the possible involvement of cell-mediated
immunity in protection from disease should not be excluded.
Immunization with rSFV RNA can also induce cytotoxic T
lymphocyte (CTL) responses, and significant splenic CTL responses were observed after immunization of mice with RNA
coding for influenza virus nucleoprotein (data not shown). Recently, vaccination with rSFV RNA was used successfully in a
cancer therapy model [15], in which protection was probably
mediated by CTL activity.
Similar to immunization with either conventional plasmid or
recombinant viral vaccine vector, rSFV RNA vaccination results in antigen expression in vivo. However, the RNA vaccines
used in this study possess interesting features in terms of biosafety. In rSFV RNA, both replication and amplification of the
RNA occur exclusively in the cytoplasm of transfected cells.
This means that concerns such as genomic integration of vector
sequences and cell transformation, which are theoretical problems for DNA vaccination are avoided [1]. The alpha viral
replicon is cytolytic for cells, and thus the rSFV RNA vaccine
is intrinsically transient and self-eliminating in nature [11]. This
provides an additional level of safety, compared with conventional plasmid DNA vaccination strategies.
Compared with viral vectors, the most obvious difference
with the rSFV RNA vaccine approach is the absence of viral
structural components. Because viral structural proteins are often highly immunogenic, the lack of such genes and/or antigens
in the vaccine formulation is of importance in situations where
booster immunizations are desired. It also allows for the same
basic RNA “backbone” to be used in the same vaccine recipient
for later immunizations against other infectious agents. Furthermore, issues related to reversion of the vector to a pathogenic virus within a vaccinated recipient are avoided with the
rSFV RNA strategy, as there is no production of virions. This
contrasts with the recent vaccine approach described elsewhere
JID 2001;183 (1 May)
for TBE [5], in which reversion of the full-length RNA vaccine
to a pathogenic form in vivo remains a theoretical possibility.
In conclusion, our results demonstrate the utility of the naked
RNA approach and suggest a novel DNA-free strategy for
generic genetic vaccine design.
Acknowledgments
We thank J. A. Melero (Centro Nacional de Biologı́a Fundamental,
Instituto de Salud Carlos III, Madrid, Spain) for respiratory syncytial
virus F protein cDNA and anti-F monoclonal antibody.
References
1. Gurunathan S, Klinman DM, Seder RA. DNA vaccines: immunology, application and optimization. Annu Rev Immunol 2000; 18:927–74.
2. Conry RM, Lo Buglio AF, Wright M, et al. Characterization of a messenger
RNA polynucleotide vaccine vector. Cancer Res 1995; 55:1397–400.
3. Qiu P, Ziegelhoffer P, Sun J, Yang NS. Gene gun delivery of mRNA in situ
results in efficient transgene expression and genetic immunization. Gene
Ther 1996; 3:262–8.
4. Hoerr I, Obst R, Rammensee HG, Jung G. In vivo application of RNA
leads to induction of specific cytotoxic T lymphocytes and antibodies. Eur
J Immunol 2000; 30:1–7.
5. Mandl CW, Aberle JH, Aberle SW, Holzmann H, Allison SL, Heinz FX. In
vitro–synthesized infectious RNA as an attenuated live vaccine in a flavivirus model. Nat Med 1998; 4:1438–40.
6. Johanning FW, Conry RM, LoBuglio AF, et al. A Sindbis virus mRNA
polynucleotide vector achieves prolonged and high level heterologous gene
expression in vivo. Nucleic Acids Res 1995; 23:1495–501.
7. Zhou X, Berglund P, Rhodes G, Parker SE, Jondal M, Liljeström P. Selfreplicating Semliki Forest virus RNA as recombinant vaccine. Vaccine
1994;12:1510–4.
8. Berglund P, Fleeton MN, Smerdou C, Liljeström P. Immunization with recombinant Semliki Forest virus induces protection against influenza challenge in mice. Vaccine 1999; 17:497–507.
9. Fleeton MN, Sheahan BJ, Gould EA, Atkins GJ, Liljeström P. Recombinant
Semliki Forest virus particles encoding the prME or NS1 proteins of
louping ill virus protect mice from lethal challenge. J Gen Virol 1999; 80:
1189–98.
10. Li X, Sambhara S, Li CX, et al. Protection against respiratory syncytial virus
infection by DNA immunization. J Exp Med 1998; 188:681–8.
11. Berglund P, Smerdou C, Fleeton MN, Tubulekas I, Liljeström P. Enhancing
immune responses using suicidal DNA vaccines. Nat Biotechnol 1998; 16:
562–5.
12. Ulmer JB, Donnelly JJ, Parker SE, et al. Heterologous protection against
influenza by injection of DNA encoding a viral protein. Science 1993;
259:1745–9.
13. Fleeton MN, Liljeström P, Sheahan BJ, Atkins GJ. Recombinant Semliki
Forest virus particles expressing louping ill virus antigens induce a better
protective response than plasmid-based DNA vaccines or an inactivated
whole particle vaccine. J Gen Virol 2000; 81:749–58.
14. Hussell T, Openshaw P. Recent developments in the biology of respiratory
syncytial virus: are vaccines and new treatments just round the corner?
Curr Opin Microbiol 1999; 2:410–4.
15. Ying H, Zaks TZ, Wang RF, et al. Cancer therapy using a self-replicating RNA
vaccine. Nat Med 1999;5:823–7.