Enhanced Influenza Virus-Like Particle Vaccination with a Structurally Optimized RIG-I Agonist as Adjuvant Vladimir Beljanski, Cindy Chiang,* Greg A. Kirchenbaum,* David Olagnier,* Chalise E. Bloom,* Terianne Wong,* Elias K. Haddad,* Lydie Trautmann,* Ted M. Ross,* John Hiscott* Vaccine & Gene Therapy Institute of Florida, Port St. Lucie, Florida, USA

ABSTRACT

The molecular interaction between viral RNA and the cytosolic sensor RIG-I represents the initial trigger in the development of an effective immune response against infection with RNA viruses, resulting in innate immune activation and subsequent induction of adaptive responses. In the present study, the adjuvant properties of a sequence-optimized 5=-triphosphate-containing RNA (5=pppRNA) RIG-I agonist (termed M8) were examined in combination with influenza virus-like particles (VLP) (M8-VLP) expressing H5N1 influenza virus hemagglutinin (HA) and neuraminidase (NA) as immunogens. In combination with VLP, M8 increased the antibody response to VLP immunization, provided VLP antigen sparing, and protected mice from a lethal challenge with H5N1 influenza virus. M8-VLP immunization also led to long-term protective responses against influenza virus infection in mice. M8 adjuvantation of VLP increased endpoint and antibody titers and inhibited influenza virus replication in lungs compared with approved or experimental adjuvants alum, AddaVax, and poly(I·C). Uniquely, immunization with M8-VLP stimulated a TH1-biased CD4 T cell response, as determined by increased TH1 cytokine levels in CD4 T cells and increased IgG2 levels in sera. Collectively, these data demonstrate that a sequence-optimized, RIG-I-specific agonist is a potent adjuvant that can be utilized to increase the efficacy of influenza VLP vaccination and dramatically improve humoral and cellular mediated protective responses against influenza virus challenge. IMPORTANCE

The development of novel adjuvants to increase vaccine immunogenicity is an important goal that seeks to improve vaccine efficacy and ultimately prevent infections that endanger human health. This proof-of-principle study investigated the adjuvant properties of a sequence-optimized 5=pppRNA agonist (M8) with enhanced capacity to stimulate antiviral and inflammatory gene networks using influenza virus-like particles (VLP) expressing HA and NA as immunogens. Vaccination with VLP in combination with M8 increased anti-influenza virus antibody titers and protected animals from lethal influenza virus challenge, highlighting the potential clinical use of M8 as an adjuvant in vaccine development. Altogether, the results describe a novel immunostimulatory agonist targeted to the cytosolic RIG-I sensor as an attractive vaccine adjuvant candidate that can be used to increase vaccine efficacy, a pressing issue in children and the elderly population.

A

nnual vaccination with the trivalent inactivated influenza vaccine (TIV), quadrivalent inactivated influenza vaccine (QIV), or the live attenuated influenza vaccine (LAIV) are the primary strategies for reducing the morbidity and mortality associated with human influenza infection (1, 2). The protection provided by the TIV is strong in young adults, but its efficacy decreases in the elderly due to immunosenescence, which is characterized by decreases in effector cell number and function, as well as alterations in the production of inflammatory and antiviral cytokines (3–5). Additionally, individual immune responses vary dramatically because of multiple factors, including the immunogen, route of administration, age, and virus type. Thus, an additional immune stimulation is frequently necessary to enhance vaccine efficacy. With increasing emphasis on subunit- and/or peptide-based immunization and the ultimate need to develop a universal influenza vaccine, new approaches to improve vaccine efficacy are warranted (6). Virus-like particles (VLP) are an attractive alternative to moretraditional live attenuated or split vaccines (7). VLP mimic the virus in structure and morphology but are noninfectious and thus possess a high safety profile that enhances their potential for future vaccine development against highly pathogenic strains (7). As with live, attenuated virus vaccination, VLP stimulate the immune

10612

jvi.asm.org

system, leading to both humoral and cellular immune responses. An effective VLP-based vaccine typically includes a strong immunogen (i.e., VLP expressing viral surface glycoprotein such as

Received 12 June 2015 Accepted 4 August 2015 Accepted manuscript posted online 12 August 2015 Citation Beljanski V, Chiang C, Kirchenbaum GA, Olagnier D, Bloom CE, Wong T, Haddad EK, Trautmann L, Ross TM, Hiscott J. 2015. Enhanced influenza virus-like particle vaccination with a structurally optimized RIG-I agonist as adjuvant. J Virol 89:10612–10624. doi:10.1128/JVI.01526-15. Editor: A. García-Sastre Address correspondence to John Hiscott, [email protected]. * Present address: Cindy Chiang and John Hiscott, Istituto Pasteur-Fondazione Cenci Bolognetti, Rome, Italy; Greg A. Kirchenbaum, Chalise E. Bloom, Terianne Wong, and Ted M. Ross, Center for Vaccine Development, Department of Infectious Diseases, University of Georgia, Athens, Georgia, USA; David Olagnier, Lady Davis Institute, Jewish General Hospital, Montreal, Canada; Elias K. Haddad, Department of Medicine, Division of Infectious Diseases and HIV Medicine, Drexel University, Philadelphia, Pennsylvania, USA; Lydie Trautmann, Cellular Immunology Section, U.S. Military HIV Research Program (MHRP), HJF, WRAIR, Silver Spring, Maryland, USA. Copyright © 2015, American Society for Microbiology. All Rights Reserved. doi:10.1128/JVI.01526-15

Journal of Virology

October 2015 Volume 89 Number 20

Adjuvant Activity of a RIG-I Agonist

hemagglutinin) and a potent adjuvant for inducing antiviral signals (8, 9). Importantly, VLP can be genetically engineered to express vaccine antigens that represent a population of sequences and elicit cross-protective immune responses against multiple pathogens (10). Influenza VLP can be formed following coexpression of just three viral proteins—matrix (MA), hemagglutinin (HA), and neuraminidase (NA)—in a mammalian expression system; VLP express the major surface influenza proteins in the same conformation as found in the influenza virion and have been shown to stimulate a potent immune response (7). Addition of an adjuvant is a key strategy that (i) enhances immunogenicity of the antigen, (ii) permits a reduction in the amount of viral epitope per vaccine (termed “antigen sparing”), and (iii) stimulates immune responsiveness in the elderly, thus increasing vaccine efficacy in this population (11). The most commonly used FDA-approved adjuvant is aluminum salts (alum), although it is not included in any of the current influenza formulations in the United States. In addition to alum, vaccines can be formulated with adjuvants such as MF59, AF03, and AS03, vaccine antigen delivery vehicles virosomes (12), and adjuvant combination AS04 (13), all of which create an antigen depot, activate antigen-presenting cells, and trigger the innate immune response by stimulation of danger signals (14). Given the wide range of vaccine strategies currently under study, a high priority for the development of influenza vaccines is to identify novel adjuvants that elicit a broad and robust immune response to increase immunogenicity of antigens and to enhance the antigen-sparing effect (14, 15). Influenza infection is sensed by RIG-I (retinoic acid-inducible gene I), a cytosolic sensor that detects viral RNA during replication through its helicase domain (16). RIG-I also possesses an effector caspase activation and recruitment domain (CARD) that forms a complex with the mitochondrial adaptor MAVS (mitochondrial antiviral signaling) (17). MAVS serves as a signaling platform for protein complexes that activates transcription factors NF-␬B, interferon regulatory factor 3 (IRF3), and IRF7 (18, 19) that induce the expression and production of type I interferon (IFN), as well as proinflammatory cytokines and antiviral proteins (20). A secondary response involving hundreds of IFN-stimulated genes (ISGs) is induced by secreted IFN binding to the type I IFN receptors on adjacent cells and tissues, leading to amplification of the antiviral immune response via the JAK-STAT (JAK stands for Janus kinase, and STAT stands for signal transducer and activator of transcription factor) pathway (21). This stimulation of the innate antiviral response also contributes to the maturation of dendritic cells (DC) and clonal expansion of CD4⫹ and CD8⫹ T cells (22), all of which contribute to the establishment of long-term adaptive immunity against infection. Therefore, an attractive strategy for the development of broad-spectrum antivirals and vaccine adjuvants involves the use of natural or synthetic 5=triphosphate-containing RNA (5=pppRNA) that activate innate immunity via RIG-I. We previously demonstrated that intravenous injection of a short, double-stranded RNA containing a 5=-triphosphate moiety, derived from the 5= and 3= untranslated regions of the vesicular stomatitis virus (VSV) genome (wild-type [WT] 5=pppRNA) stimulated innate responses in lungs and protected mice from lethal H1N1 influenza virus challenge (23). Transcriptional profiles of lung epithelial cells treated in vitro with WT 5=pppRNA identified overlapping and unique transcriptional signatures associated with genes capable of mobilizing multiple arms of innate

October 2015 Volume 89 Number 20

and adaptive responses. Specific modifications in the WT 5=pppRNA structure—modifications to the primary RNA sequence that eliminated mismatch in the double-stranded RNA (dsRNA) region, removed the panhandle structure, and introduced additional bases—resulted in a 59-nucleotide RNA structure (M5) with improved antiviral properties compared to those of WT 5=pppRNA. Further improvement of M5 antiviral properties was achieved by the addition of AU base pairs to increase the length of the dsRNA stem structure (24). The resulting sequenceoptimized agonist, M8, further increased the breadth and magnitude of the antiviral and inflammatory response observed in primary human dendritic cells compared to WT 5=pppRNA or M5. As an antiviral agent, M8 efficiently inhibited influenza and dengue virus replication in vitro and decreased both Chikungunya and influenza virus replication in vivo (24). Because of its capacity to stimulate a potent antiviral and inflammatory response, we hypothesized that M8 may also function as an adjuvant in a VLP-based influenza vaccine formulation. In the present study, we demonstrate that M8-VLP (VLP with M8 as adjuvant) increased antibody levels, protected against lethal H5N1 influenza virus challenge, stimulated formation of germinal center B cells, and induced a TH1-predominant cellular immune phenotype upon vaccination. These results illustrate that agonistspecific stimulation of the RIG-I pathway can be utilized to adjuvant influenza VLP vaccination and dramatically improve humoral and cellular mediated protective responses against influenza virus challenge. MATERIALS AND METHODS In vitro synthesis of 5=pppRNA. RIG-I agonists were synthesized using an in vitro RNA transcription kit (Ambion) as previously described (23). RNA transcripts were digested with DNase for 15 min at 37°C and then purified using miRNeasy kit (Qiagen). The integrity of the purified 5=pppRNA was analyzed on a denaturing 15% Tris-borate-EDTA (TBE)urea polyacrylamide gel (Bio-Rad) and compared to the 5=pppRNA digested with RNase A or DNase (Ambion). Isolation and transfection of MDDC. Human peripheral blood mononuclear cells (PBMC) were isolated from buffy coats from healthy volunteers in a study approved by the institutional review board (IRB) and by the Vaccine & Gene Therapy Institute of Florida (VGTI-FL) Institutional Biosafety Committee (2011-6-JH1). Written informed consent approved by the VGTI-FL ethics review board was provided to study participants. Research conformed to ethical guidelines established by the ethics committee of the Oregon Health & Science University (OHSU) VGTI and Martin Health System. PBMC were isolated using the FicollPaque Plus medium (GE Healthcare Bio). CD14⫹ monocytes were isolated by positive selection using CD14 microbeads (Miltenyi Biotec) and a magnetic cell separator. Purified CD14⫹ monocytes were cultured for 7 days in in 10 ml of complete monocyte differentiation medium (Miltenyi Biotec). On day 3, the medium was replenished with fresh medium. Only cells with the CD14lo CD1ahi DC-SIGNhi phenotype after 5 to 7 days differentiation were used in subsequent experiments. Monocyte-derived dendritic cells (MDDC) were transfected with various amounts of wildtype (WT), M5, or M8 5=pppRNA or poly(I·C) for 24 h using HiPerfect transfection reagent (Qiagen) for 24 h. MDDC were stained for 5 min with human TruStain FcX (Biolegend) followed by staining with CD83-phycoerythrin (PE) (Biolegend), CD86-Pacific Blue (Biolegend), CD80-PE, or CD40-PE, for 15 min at 4°C. Cells were analyzed on an LSRII flow cytometer (Becton Dickinson). Calculations and population analyses were done using FACSDiva software. Quantitative real-time RT-PCR. Total RNA was isolated from cells using RNeasy kit (Qiagen), and RNA was reverse transcribed using the SuperScript VILO cDNA synthesis kit (Invitrogen). PCR primers were

Journal of Virology

jvi.asm.org

10613

Beljanski et al.

designed using Roche’s Universal ProbeLibrary Assay Design Center (www.universalprobelibrary.com). Quantitative reverse transcriptionPCR (RT-PCR) was performed on a LightCycler 480 Probes Master (Roche). All data are presented as a relative quantification with efficiency correction based on the relative expression of target gene versus glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the invariant control. The sequences of forward (F) and reverse (R) primers used as well as their complementary probes (numbers in parentheses) are as follows: CD40 (54) F, 5=-GCAGGGGAGTCAGCAGAG-3=; R, 5=-AGAGGCAGACGAA CCATAGC-3=; CD74 (43) F, 5=-CCCAAGCCTGTGAGCAAG-3=; R, 5=CATACTTGGTGGCATTCTGC-3=; CD80 (4) F, 5=-TTGCCCTTTACGT ATCTGCTC-3=; R, 5=-GCTACTTCTGTGCCCACCAT-3=; CD83 (25) F, 5=-GGGGTGTGCCTGTCTGTTAC-3=; R, 5=-GATGCCATCTTCAGCG TAGG-3=; CD86 (15) F, 5=-ACAGCAGAAGCAGCCAAAAT-3=; R, 5=-G AATCTTCAGAGGAGCAGCAC-3=; 4-1BB (39) F, 5=-GCTCTCGAT ATCCGGTAGGA-3=; R, 5=-GCCTGACCTAGCTAAGACACTTCT-3=; HLA-DRA (14) F, 5=-AGCACTGGGAGTTTGATGCT-3=; R, 5=-GGCAC ACACCACGTTCTCT-3=; HLA-DQA (68) F, 5=-ACCAAGGGCCATTG TGAAT-3=; R, 5=-AATCGGGCCAGAGAATAGTG-3=. Virus propagation and challenge. Influenza reassortant mouseadapted H5N1 virus (H5N1) expressing H5 hemagglutinin (HA) A/Vietnam/1203/2004 and neuraminidase (NA) A/Thailand/1(KAN-1)/2004 and internal viral genes from mouse-adapted A/Puerto Rico/8/1934 (PR8), was propagated using MDCK cells as previously described (25). In animal challenge experiments, anesthetized female BALB/c mice were infected intranasally with a lethal dose (5 ⫻ 103 PFU in 50 ␮l phosphatebuffered saline [PBS]) of H5N1; this dose represents approximately 50 LD50 (50% lethal doses) in mice. Virus-like particle propagation. The H5N1 virus-like particles (VLP) were purified from HEK293T cells which were transfected using Lipofectamine 2000 (Invitrogen) with 5 ␮g of each plasmid DNA expressing H5N1 A/Vietnam/1203/2004 HA and H5N1 A/Thailand/1(KAN-1)/2004 NA (codon optimized) and 10 ␮g of plasmid DNA expressing HIV Gag protein. Cells were incubated for 72 h at 37°C, supernatants containing VLP were collected and sterile filtered, and VLP were purified by centrifugation at 100,000 ⫻ g through a 20% glycerol cushion and resuspended in PBS. Total protein was quantified using bicinchoninic acid (BCA) protein assay (Thermo Fisher Scientific), and VLP were aliquoted in PBS and stored at ⫺80°C. HA content was quantified by densitometry as described previously (26). Immunization. BALB/c mice (6 to 8 weeks of age, Jackson Laboratories) were housed in cage units, fed ad libitum, and cared for under USDA guidelines for laboratory animals. For immunization, mice were anesthetized with IsoSol (Patterson Veterinary) and immunized via the intramuscular route (i.m.) with 0.5 ␮g to 2 ␮g (based on HA content) of purified VLP (in 50 ␮l PBS) with or without 0.1 ␮g to 5 ␮g 5=pppRNA as adjuvant and then challenged at week 3. The 5=pppRNA was delivered with in vivo-jetPEI (PolyPlus, France) at a nucleoprotein/phosphoprotein (N/P) ratio of 8 according to the manufacturer’s instructions. Imject alum (Fisher Science, Pittsburgh, PA) and AddaVax (Invivogen, San Diego, CA) were added to 50% volume of VLP in PBS solution per the manufacturer’s recommendation. Animals were monitored for survival and morbidity weekly during the immunization regimen and each day during the viral challenge. Blood samples for serological analysis were collected from anesthetized mice via the retro-orbital sinus. Blood was allowed to clot at room temperature, and serum was removed and frozen at ⫺80°C after centrifugation. All procedures were performed in accordance with the National Resource Council (NRC) Guide for the Care and Use of Laboratory Animals (27) and the Animal Welfare Act (28). Sickness score. After the mice were infected, they were monitored daily for weight loss, disease signs, and death for up to 21 days after infection. Individual body weights, sickness scores, and death were recorded for each mouse after inoculation. The sickness score was generated by evaluating activity (0 for normal, 1 for reduced, and 2 for severely reduced), hunched back (0 for absent and 1 for present), and ruffled fur (0

10614

jvi.asm.org

for absent and 1 for present) as described previously (29). The final score was the addition of each individual score resulting in the minimum score of 0 for a healthy mouse and 1 to 4 for a sick mouse. Animals that lost 20 to 25% of their original weight were sacrificed, and the weight loss curves are reported for data with statistical difference between treatment groups. Virus plaque assay. Lungs were isolated from mice postmortem and snap-frozen in dry ice/ethanol bath. Dulbecco’s modified Eagle medium (DMEM) (10 volumes of DMEM to grams of tissue) was then added to the tissue placed in a 0.7-␮m cell strainer in a petri dish, and the sample was muddled until the tissue was fully disaggregated. The remaining liquid was collected from the petri dish in a sample tube for stock lung homogenate sample. Confluent MDCK cells plated in 6-well tissue culture plates were inoculated with 0.2 ml of virus or lung cell supernatant serially diluted (1:10) in DMEM. Virus was adsorbed for 1 h, with shaking every 15 min. Wells were overlaid with 1.6% (wt/vol) Bacto agar (BD Diagnostic Systems) mixed 1:1 with L-15 medium (Cambrex) containing 1% penicillin (Pen)-streptomycin (Strep) (Life Technologies), with 0.6 mg/ml tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-trypsin (SigmaAldrich). The plates were incubated for 2 days at 37°C, and plaques were visualized with crystal violet. Hemagglutination inhibition activity. The hemagglutination inhibition (HAI) assay was used to measure functional anti-HA antibodies that were able to inhibit agglutination of horse red blood cells. The procedure was adapted from the Centers for Disease Control and Prevention influenza surveillance manual (30) and performed as previously described (25). Serological assays. Enzyme-linked immunosorbent assay (ELISA) plates (BD Biosciences) were coated with 25 ng per well of purified recombinant hemagglutinin encoded by A/Vietnam/1203/2004 in carbonate buffer (pH 9.4) containing 5 ␮g/ml bovine serum albumin (BSA) fraction V at 4°C overnight. The plates were then blocked with ELISA blocking buffer (PBS containing 5% BSA fraction V, 2% bovine gelatin, and 0.05% Tween 20) for more than 90 min at 37°C. Serial dilutions of immune sera were incubated for more than 90 min, and the plates washed five times with PBS. Total HA-specific IgG was detected using horseradish peroxidase-conjugated goat anti-mouse IgG (␥ chain specific) (Southern Biotech, Birmingham, AL) diluted to 1:2,500 in ELISA blocking buffer. Following additional PBS washes, 2,2=-azino-bis(3-ethylbenzothiazoline6-sulfonic acid) (ABTS) substrate was added, and the plates were incubated 10 to 20 min at 37°C. Colorimetric conversion was measured as the optical density (OD) (at 414 nm) by a spectrophotometer (BioTek, Winooski, VT, USA). OD values for wells coated with only BSA were subtracted to determine HA-specific reactivity. HA-specific IgG subclasses were measured in serum samples diluted to 1:50 in ELISA blocking buffer and detected using horseradish peroxidase-conjugated goat anti-mouse IgG1, IgG2a, IgG2b, and IgG3 antibodies (Southern Biotech, Birmingham, Alabama) diluted to 1:1,000 in blocking buffer. Assessment of T and B cell responses. Spleens from mice immunized intraperitoneally (i.p.) 8 days prior with 1 ␮g VLP (based on HA content) were harvested in cold RPMI 1640. The spleens were mechanically dissociated through 70-␮m cell strainers, and the resulting single-cell suspension was washed in cold RPMI 1640. The cell suspensions were then incubated with ACK (ammonium chloride potassium) lysing buffer to lyse red blood cells (RBC) and counted following a washing step. For assessment of B cell responses, 2 ⫻ 106 to 5 ⫻ 106 splenocytes were stained directly ex vivo. For assessment of T cell responses, 5 ⫻ 106 splenocytes were cultured in RPMI 1640 containing vehicle, VLP (1 ␮g/ml), or the HA-derived peptide (IYSTVASSL) (10 ␮g/ml) for 21 h. Brefeldin A (eBioscience) was added 6 h prior to harvest. The following monoclonal antibodies were used: fluorescein isothiocyanate (FITC)-labeled anti-mouse tumor necrosis factor alpha (TNF-␣) (MP6-XT22), phycoerythrin (PE)labeled anti-mouse interleukin 2 (IL-2) (clone JES6-5H4), PE-Cy7-labeled anti-mouse gamma interferon (IFN-␥) (clone XMG1.2), Pacific Blue-labeled anti-mouse CD8a (clone 53-6.7), allophycocyanin (APC)labeled anti-mouse CD4 (clone GK1.5), PE-Dazzle594-labeled anti-

Journal of Virology

October 2015 Volume 89 Number 20

Adjuvant Activity of a RIG-I Agonist

mouse B220 (clone RA3-6B2), Pacific Blue-labeled anti-mouse CD44 (clone IM7), peridinin chlorophyll protein (PerCP)-labeled anti-mouse CD19 (clone eBio1D3), FITC-labeled anti-mouse CD38 (clone 90), and Alexa Fluor 647-labeled anti-mouse IgG1 (clone RMG1-1; diluted 1:500) (all from BioLegend). Additional antibodies were biotin-labeled antimouse CD95 (clone Jo2) and PE-labeled anti-CD138 (clone 281-2) (both from BD Biosciences). Both LIVE/DEAD fixable dead cell staining (Invitrogen) and intracellular staining using BD Cytofix/Cytoperm kit were performed according to the manufacturers’ instructions. Histology. Lungs were harvested and prepared for immunohistochemistry using modified protocol (31). After euthanasia, the chest cavity was opened, and the lungs were gently inflated intratracheally with 4% paraformaldehyde in PBS at 4°C, removed, and immersed in 4% paraformaldehyde at 4°C overnight. The next day, the solution was replaced with 70% ethanol, and tissues were kept at 4°C for up to 2 weeks. Tissue sections were embedded in paraffin, sectioned into slices (⬃5 ␮m in thickness), and stained with hematoxylin and eosin (H&E) or left unstained. Tissue sections were imaged with a Qimaging Micropublisher 5.0 RTV digital camera on an Olympus BX61 fluorescence microscope. To quantify the number of apoptotic lung cells, representative sections were deparaffinized and rehydrated in xylene and graded alcohols, respectively, by using standard procedures, and terminal deoxynucleotidyltransferasemediated dUTP-biotin nick end labeling (TUNEL) assay was performed according to the manufacturer’s instructions (Roche, Mannheim, Germany). The percentages of TUNEL-positive cells within the tissue sections were determined by counting at least 100 cells each from eight randomly selected fields. Statistical analyses. Values are expressed as means ⫾ standard errors of the means (SEMs), and statistical analysis was performed by one-way analysis of variance (ANOVA) using PRISM software (GraphPad software), followed by Tukey posthoc test to determine significance.

RESULTS

M8 potentiates DC maturation and antiviral signaling in human MDDC. The most efficient antigen-presenting cells are mature, immunologically competent dendritic cells (DC) (32), and it is now accepted that the adjuvant component of vaccines contributes to vaccine efficacy by triggering DC maturation and antigen presentation (33). To determine whether selected 5=pppRNA sequences differ in their ability to induce DC maturation ex vivo, we first examined the mRNA levels of selected genes involved in DC maturation and activation. Chemokine CCL4 (CC chemokine ligand 4), activation and/or costimulation markers CD40, CD80, CD83, and CD86, costimulatory receptor 4-1BB, major histocompatibility complex (MHC) class II molecules HLA-DRA and HLADQA, and a regulator of antigen presentation, CD74, were all dramatically upregulated by M8 treatment in primary DC cultures (Fig. 1A). We then confirmed these data by examining the surface expression levels of several DC activation/maturation markers in MDDC treated with WT 5=pppRNA, M5, and M8 (Fig. 1B); the well-characterized synthetic RNA-based Toll-like receptor 3 (TLR3) agonist poly(I·C) was also used for comparison (34). Consistent with the observed increases in RNA levels, M8 treatment led to an ⬃5-fold increase in the expression of costimulatory molecules and/or activation markers CD40 and CD86, an ⬃3-fold increase in the expression of CD83, and an ⬃2-fold increase in the expression of CD80, compared to cells treated with WT 5=pppRNA, M5, or poly(I·C). For surface marker CD86, WT 5=pppRNA and M5 induced higher mRNA levels compared to the protein levels, likely indicating higher turnover of mRNA induced by these two molecules compared to M8-transfected cells. Collectively, these data indicated that M8 was a potent inducer of DC maturation and activation.

October 2015 Volume 89 Number 20

FIG 1 Transfecting monocyte-derived dendritic cells (MDDC) with WT, M5, or M8 5=pppRNA or poly(I·C) increases expression of activation and differentiation markers and their mRNA levels. MDDC were isolated from peripheral blood mononuclear cells (n ⫽ 4), differentiated, and transfected with 10 ng WT, M5, M8, or poly(I·C) using HiPerFect transfection reagent for 24 h. (A) Gene expression analysis using the Fluidigm BioMark platform for several genes (indicated on the right) in MDDC transfected with 20 fmol of WT, M5, M8, or poly(I·C) for 24 h. (B) MDDC surface expression of activation and differentiation markers as assessed by flow cytometry (mean plus standard error of the mean [SEM] [error bar]). Values that are significantly different are indicated by bars and asterisks as follows: *, P ⱕ 0.05; **, P ⱕ 0.01; ***, P ⱕ 0.005.

M8 potentiates influenza VLP immunogenicity. To determine whether M8 possessed adjuvant activity in an in vivo model of vaccination, BALB/c mice were vaccinated by intramuscular injection with VLP coexpressing the HA and NA from H5N1 and HIV Gag protein, in combination with M8, M5, or poly(I·C) (which was previously shown to potentiate responses to influenza virus antigens [35]). Three weeks later, mouse sera were collected and analyzed for HAI activity (hemagglutination inhibition assay [HAI] or receptor blocking titers); mice immunized with VLP adjuvanted with M8, M5, or poly(I·C) displayed a 2- to 3-fold increase in HA-specific IgG titers (data not shown) and 2- to 3-fold-higher HAI antibody titers (P ⬍ 0.005) (Fig. 2A) than mice immunized with VLP alone. M8 was a more potent stimulator of HA-specific IgG and HAI antibody titers than M5 or poly(I·C) (Fig. 2A). Next, vaccinated mice were challenged with a lethal dose of

Journal of Virology

jvi.asm.org

10615

Beljanski et al.

FIG 2 Protective efficacy of the VLP vaccine adjuvanted with M5, M8, or poly(I·C). Mice (n ⫽ 5) were immunized intramuscularly with 2 ␮g of VLP alone or combined with 5 ␮g M5, M8, or poly(I·C) as a 50-␮l injection, and 3 weeks later, they were challenged with 5,000 PFU of H5N1 influenza virus. (A) Hemagglutination inhibition (HAI) antibody titers in immunized mice prior to infection were determined by hemagglutination inhibition assay using horse red blood cells. (B) Assessment of viral replication in lungs of infected animals 3 days postinfection by a plaque assay. (C) TUNEL-positive (apoptotic) lung cells in infected mice were quantified by a TUNEL assay as described in Materials and Methods. All values in panels A to C are expressed as means plus SEMs. Values that are significantly different are indicated by bars and asterisks as follows: *, P ⱕ 0.05; ***, P ⱕ 0.005. (D) H&E staining of paraffin-embedded lung cross sections from mice 3 days after challenge. The yellow arrows indicate the airways of mice that were vaccinated with M8 only (top) or M8-VLP (bottom).

H5N1 influenza virus, and lungs were harvested for assessment of influenza replication and histopathological analysis 3 days postinfection (Fig. 2B). Immunization with VLP alone resulted in a 4-log-unit decrease in viral plaques, while adjuvantation of VLP with M8, M5, or poly(I·C) resulted in an additional 1-log-unit reduction in viral titer (P ⬍ 0.05). Compared to M5-VLP or poly(I·C)-VLP, mice immunized with M8-VLP had the lowest

10616

jvi.asm.org

virus titer in the lungs. These data were further corroborated by histopathological examination and assessment of lung tissue apoptosis (TUNEL staining); the number of apoptotic cells was reduced approximately 3-fold in VLP-immunized mice from the control mice, and it was further reduced 5- to 10-fold when M8, M5, or poly(I·C) was used as adjuvant (P ⬍ 0.05) (Fig. 2C). Mice immunized with M8-VLP had ⬃50% fewer apoptotic cells than

Journal of Virology

October 2015 Volume 89 Number 20

Adjuvant Activity of a RIG-I Agonist

FIG 3 Dose-response assessment of VLP and M8. (A and B) Mice (n ⫽ 5) were immunized with decreasing doses of VLP (2 ␮g to 0.5 ␮g) in combination with 5 ␮g of M8; 3 weeks later, they were challenged with H5N1 influenza virus, and lungs from infected animals were harvested 3 days postchallenge. HAI antibody titers in immunized mice prior to infection were determined by HAI assay (A), and viral replication in lungs was assessed by plaque assay (B). (C to F) Mice were immunized with 0.5 ␮g of VLP with 0.1 to 5 ␮g of M8. HAI antibody titers were determined by HAI assay (C), and viral replication was determined by plaque assay (D). Weight loss (E) and survival (F) were determined as well. Values are expressed as means plus SEMs. Values that are significantly different are indicated by bars and asterisks as follows: *, P ⱕ 0.05; **, P ⱕ 0.01; ***, P ⱕ 0.005.

M5-VLP- or poly(I·C)-VLP-immunized mice. Similarly, H&E staining of lung cross sections indicated complete absence of edema and inflammation in mice immunized with VLP with M8 as adjuvant, while nonvaccinated mice, mice vaccinated with adjuvant only, and to a lesser extent, mice vaccinated with VLP had

October 2015 Volume 89 Number 20

signs of inflammation around airways (Fig. 2D). Altogether, these data indicate that VLP adjuvanted with M8, M5, or poly(I·C) efficiently blocked influenza virus replication, decreased virus-induced apoptosis, and decreased inflammation in the lungs of vaccinated animals. Of the adjuvants tested, M8-VLP elicited the

Journal of Virology

jvi.asm.org

10617

Beljanski et al.

FIG 4 Adjuvant comparison strategy and antibody immune responses for M8, alum, AddaVax, and poly(I·C)-adjuvanted VLP vaccine. (A) Strategy for adjuvant

comparison. 3d, 3 days. (B to D) Mice were immunized with 0.5 ␮g of VLP in combination with 5 ␮g of M8 or poly(I·C) or in combination with 50% volume of alum or AddaVax, and 5 days and 3 weeks after immunization, sera were collected. HA-specific IgG antibodies (B) and influenza virus HAI antibody titers (C) were determined 3 weeks after immunization by ELISA and HAI assay, respectively. (D) HA-specific IgM antibodies were determined 5 days after immunization by ELISA. Values are expressed as means plus SEMs. Values that are significantly different are indicated by bars and asterisks as follows: *, P ⱕ 0.05; **, P ⱕ 0.01. O.D., optical density.

highest level of HAI antibody titers and demonstrated the lowest level of influenza-induced lung tissue damage, highlighting the improved activity of sequence-optimized M8. Antigen-sparing capacity of M8 in combination with VLP. Antigen sparing is regarded as a important parameter for pandemic vaccine development (36) where vaccine supplies are likely to be limited. Because adjuvantation is an important antigensparing strategy (36), mice were immunized with various concentrations of VLP (0.5, 1, or 2 ␮g, based on HA content) alone or in combination with a fixed quantity of M8 (5 ␮g). Analysis of the HAI antibody titers 3 weeks later revealed that 0.5 ␮g VLP plus 5 ␮g M8 generated titers similar to those generated by 2 ␮g VLP (Fig. 3A). Next, mice were challenged with a lethal inoculum of H5N1 influenza virus (5,000 PFU/mouse). Consistent with the results observed with HAI antibody titers, lung virus titers measured 3 days postchallenge were comparable between mice immunized with VLP alone (2 ␮g) or VLP (0.5 ␮g) adjuvanted with M8 (Fig. 3B). Next, to determine the minimal dose of M8 that was able to elicit a protective immune response, BALB/c mice were immunized with a single low-dose inoculum of VLP (0.5 ␮g) and with increasing concentrations of M8 (0.1 to 5 ␮g). Three weeks

10618

jvi.asm.org

postimmunization, sera were collected to determine antibody titers, and mice were challenged with H5N1; at 3 days after challenge, lungs were collected to determine virus titers. Addition of M8 (0.5 ␮g) increased HAI antibody titers by 1.5-fold compared to VLP alone (Fig. 3C), and further increases in antibody titers were observed in a dose-dependent manner with increasing M8. A reciprocal relationship was observed with lung virus titers—increasing M8 decreased lung viral load and was maximal in mice immunized with 5 ␮g M8-VLP (Fig. 3D). Finally, mice immunized with 0.5 to 5 ␮g M8 in combination with VLP all survived the H5N1 challenge, although limited weight loss was observed at lower M8 concentrations (Fig. 3E and F), indicating that M8 (0.5 ␮g or higher) functioned as an antigensparing adjuvant in protecting BALB/c mice from a lethal influenza virus challenge. Adjuvant properties of M8, alum, AddaVax, and poly(I·C). The adjuvant activity of M8 was next assessed relative to the FDAapproved adjuvant alum, AddaVax (an MF59-like adjuvant) (37), and to poly(I·C) by examining antibody titers, influenza virus lung titers, sickness score, weight loss, and survival (illustrated schematically in Fig. 4A). Assessment of HA-specific IgG by ELISA

Journal of Virology

October 2015 Volume 89 Number 20

Adjuvant Activity of a RIG-I Agonist

FIG 5 Protective efficacy of 0.5 ␮g of VLP in combination with 5 ␮g of M8 or poly(I·C) or in combination with 50% volume of alum or AddaVax. (A to C) Three weeks after vaccination, mice (n ⫽ 8) were challenged with the lethal dose of H5N1 (5,000 PFU), and their survival (A), weight (B), and sickness score (C) were monitored. (D) Viral replication in lungs was assessed by plaque assay in a separate group of immunized animals (n ⫽ 5) 3 days postinfection. Values are expressed as means ⫾ SEMs. Values that are significantly different (P ⱕ 0.005) are indicated by bars and asterisks (***).

revealed ⬃2-fold-higher antibody levels in mice immunized with M8-VLP than in mice immunized with alum-VLP, AddaVaxVLP, or poly(I·C)-VLP (Fig. 4B). Similarly, the HAI assay revealed an ⬃1.5-fold-higher HAI antibody titer in M8-VLP-vaccinated mice (Fig. 4C) than in alum-VLP-, AddaVax-VLP-, or poly(I·C)VLP-vaccinated mice. Furthermore, the levels of HA-specific IgM were also higher in M8-VLP-immunized animals 5 days after vaccination (Fig. 4D). Strikingly, all animals immunized with adjuvanted VLP survived lethal H5N1 influenza virus challenge (Fig. 5A); furthermore, animals immunized with VLP together with M8 showed no signs of weight loss following challenge, whereas a temporary weight loss was observed in animals immunized with VLP plus alum, AddaVax, or [poly(I·C)] (Fig. 5B). In contrast, the median survival for nonimmunized animals or animals immunized with adjuvant only was 9 to 10.5 days (data not shown); animals immunized with VLP had a median survival of 12 days (three animals survived influenza virus challenge). Postimmunization, the control VLP-immunized and adjuvant-immunized animals all lost weight (Fig. 5B and data not shown); however, the animals, as determined by the sickness score, did have a lower sickness score (indicating better health) for animals immunized with M8-VLP than the animals immunized with alum-VLP, AddaVax-VLP, or poly(I·C)-VLP (Fig. 5C). Additionally, M8-VLP treatment re-

October 2015 Volume 89 Number 20

sulted in an ⬃1-log-fold decrease in viral titers 3 days postinfection compared to mice immunized with alum-VLP, AddaVaxVLP, or poly(I·C)-VLP (Fig. 5D). Overall, M8 adjuvantation protected mice completely from a lethal influenza virus challenge, as did the other approved or experimental adjuvants; however, M8-VLP induced higher HAI antibody titers and decreased influenza viral loads in the lungs of infected animals, compared to animals immunized with alum-VLP, AddaVax-VLP, or poly(I·C)-VLP. M8 stimulates long-term immune responses. The same immunization schedule was used to determine whether VLP vaccine adjuvanted with M8, alum, AddaVax, or poly(I·C) induced a longterm memory response against influenza virus challenge. To determine changes in HAI antibody titers over time, sera were collected at 4 weeks and again at 4 months after immunization. After the second serum sample collection, the mice were challenged with H5N1 influenza virus, and their weight, survival, and sickness score were assessed (Fig. 6). At 4 months after immunization, all animals immunized with VLP with adjuvant had detectable antibody titers, although a 5- to 10-fold decrease in endpoint titers and ⬃50% decrease in HAI antibody titers were observed in all cases (Fig. 6A and B). At 4 months, mice immunized with VLP adjuvanted with M8, alum, AddaVax, or poly(I·C) all survived challenge, with only a 10 to 15% weight loss (Fig. 6C and D); animals also had comparable increases in sickness score during the

Journal of Virology

jvi.asm.org

10619

Beljanski et al.

FIG 6 Long-term protective responses in mice immunized with 0.5 ␮g of VLP in combination with 5 ␮g of M8 or poly(I·C) or in combination with 50% volume of alum or AddaVax. Mouse sera (n ⫽ 8) were collected 4 weeks (white bars) and 16 weeks (black bars) postvaccination to determine HA-specific IgG antibodies (ELISA) (A) and HAI antibody titers (B). Upon challenge with a lethal H5N1 dose (5,000 PFU) weight (C), survival (D), and sickness score (E) were monitored for 3 weeks. Values are expressed as the means ⫾ SEMs.

weight loss period (Fig. 6E). In contrast, VLP-immunized mice had no detectable antibody titers and were not protected from the lethal challenge. Altogether, these data indicate that M8, similar to alum, AddaVax, and poly(I·C), stimulated a prolonged immune response capable of protecting mice against homologous influenza virus challenge. M8-VLP promotes GC formation. Next, the capacity of M8 and other adjuvants to drive germinal center (GC) formation when coadministered with VLP was determined. Initially, mice were immunized i.m., but it was not possible to determine increases in GC formation between different groups and controls (data not shown). Therefore, for this experiment only, mice were immunized by intraperitoneal inoculation with VLP (1 ␮g, based on HA content), together with M8 or poly(I·C) (5 ␮g), or mixed 1:1 with alum or AddaVax. Splenocytes from immunized mice were harvested on day 8 and evaluated by flow cytometry ex vivo for the presence of B220hi CD19hi CD95hi CD38lo GC B cells (Fig. 7A). Compared to mice immunized with VLP alone, a 2.2-fold increase in the frequency of GC B cells was observed for M8 (P ⬍ 0.05), 3.5-fold for alum (P ⬍ 0.01), 4.5-fold for AddaVax (P ⬍ 0.005), and 1.5-fold for poly(I·C) (Fig. 7B). In addition, the number of IgG1⫹ GC B cells was also determined by intracellular staining. The number of IgG1⫹ GC B cells was elevated in mice immunized with alum-VLP (P ⬍ 0.005) or AddaVax-VLP (P ⬍ 0.005) than in mice immunized with VLP alone (Fig. 7C), indicating that these two adjuvants induced a TH2-biased response.

10620

jvi.asm.org

M8-VLP biases CD4ⴙ T cell effector function. To establish whether M8 induces TH1 or TH2 responses, the levels of four IgG subclasses (IgG1, IgG2a, IgG2b, and IgG3) were determined after immunization using HA-coated ELISA plates and respective IgG secondary antibodies. Immunization of mice with VLP in combination with either alum or AddaVax induced higher levels of IgG1, while immunization with M8-VLP induced higher levels of IgG2 (Fig. 8A), indicating that M8 preferentially induced a TH1-biased response, whereas alum or AddaVax preferentially induce a TH2biased response (38). To further examine whether M8 promoted TH1/Tc1 cytokine secretion, splenocytes from immunized mice were incubated with either HA518 –526 (HA from amino acid positions 518 to 526) (10 ␮g/ml) (IYSTVASSL) peptide or VLP (1 ␮g/ml, based on HA content). B220-negative cells were segregated based on surface CD4 and CD8 expression and intracellular staining for IFN-␥, IL-2, and TNF-␣ (Fig. 8B to E). In vitro restimulation of splenocytes with VLP stimulated secretion of IFN-␥ in CD8 T cells (Fig. 8B) and induced the secretion of all three cytokines in CD4 T cells (Fig. 8C to E). The highest induction of these three cytokines was observed in splenocytes originating from M8VLP-immunized animals, followed by alum-VLP-, AddaVaxVLP-, and poly(I·C)-VLP-immunized animals (P ⬍ 0.05 comparing M8-VLP- versus alum-VLP-immunized animal values and P ⬍ 0.005 comparing M8-VLP- versus AddaVax-VLP-immunized animal values for IL-2 and TNF-␣, respectively). These data indicate that M8-VLP promoted TH1 priming of VLP-stimula-

Journal of Virology

October 2015 Volume 89 Number 20

Adjuvant Activity of a RIG-I Agonist

FIG 7 Quantification of germinal center (GC) B cells from spleens of i.m. immunized mice (n ⫽ 5) by flow cytometry. (A) Gating strategy for quantification of GC B cells; (B) percentage of GC B cells in B220hi splenocytes; (C) quantification of IgG1⫹ GC B cells. Values are expressed as means plus SEMs. Values that are significantly different are indicated by bars and asterisks as follows: *, P ⱕ 0.05; **, P ⱕ 0.01; ***, P ⱕ 0.005.

ted CD4 T cells compared to alum-VLP, AddaVax-VLP, and poly(I·C)-VLP. Additionally, we observed low intracellular levels of TH2 cytokine IL-10 in animals vaccinated with M8-VLP (Fig. 8F), indicating that M8 does not induce a TH2-biased response. Interestingly, restimulation of splenocytes from M8-VLP-immunized mice with HA518 –526 (IYSTVASSL) peptide did not elicit cytokine production by CD8 T cells (data not shown), indicating that i.p. vaccination with M8-VLP did not promote cross-presentation (39). DISCUSSION

There is a significant need to enhance immunogenicity for many vaccine formulations, particularly vaccines with low immunogenicity, such as peptide-based or VLP-based vaccines (40). VLP are

October 2015 Volume 89 Number 20

a versatile vaccine platform, capable of engaging the immune response by trafficking antigen into draining lymph nodes, stimulating antigen-presenting cells, and enhancing a robust adaptive immune response (7). The success of the versatile VLP vaccine platform is perhaps best illustrated with the human papillomavirus (HPV) vaccine which consists of HPV VLP as antigens, adjuvanted with alum (41). Stimulation of pattern recognition receptors (PRRs), including the cell surface- and endosome-associated TLRs or the cytosolic RIG-like receptors, is an attractive target for the development of vaccine adjuvants (42), since stimulation of these PRRs will differentially promote synthesis of antiviral effectors, inflammatory mediators, cytokines, and chemokines with roles in the priming, expansion, and polarization of the immune responses. We previously demonstrated that the 5=pppRNA sequence can be rationally modified to maximize the RIG-I response, leading to high expression of many proinflammatory and antiviral genes (24). Among the PRR-stimulating adjuvants, 5=pppRNA molecules, TLR agonists, defective interfering RNA produced by Sendai virus, and baculovirus-based vaccination vectors have all been utilized as adjuvants in several vaccine formulations, yielding promising data in preclinical studies (35, 43–47). Unfortunately, widespread implementation of PRR-based adjuvants in vaccines has been hampered by safety concerns, as infections and innate responses can sometimes be linked to autoimmune diseases (15). However, innate stimulation with adjuvants typically lasts less than 24 h (48) and enhances the immune response to non-selfantigens. Indeed, large human vaccination studies did not link increases in autoimmune responses or infection with MF59- or AS04-adjuvanted vaccines (15). In this proof-of-principle study, we evaluated adjuvant properties of a sequence-optimized, self-complementary dsRNA agonist in a murine model of influenza vaccination using influenzaVLP as antigen. Initially, M8 was shown to maintain the specificity for the RIG-I cytosolic sensor and also increased expression of DC activation and maturation markers ex vivo; previously, we demonstrated that the 5=pppRNA response was dependent on RIG-I signaling in vivo. Several features of the structure of M8 may contribute to specificity. (i) The dsRNA character of M8 (49 bp) is too short to stimulate the cytosolic Mda5 sensor or dsRNA-dependent, endosomal TLR3. (ii) Similarly, because of the extensive AU-rich base pairing of M8, no single-stranded RNA (ssRNA) is generated to activate endosomal TLR7/8. (iii) Several studies, including our own, demonstrate that RIG-I sensing requires the 5= triphosphate moiety which is specific for RIG-I sensing (16, 23, 24, 49–51). M8-VLP was evaluated in these studies as a single, prime dose vaccination, and the adjuvant properties of M8 were compared to the clinically used adjuvants alum, AddaVax, and experimental adjuvant poly(I·C). One i.m. immunization with M8-VLP increased the HAI antibody titers in response to VLP immunization and protected mice against lethal H5N1 influenza virus challenge, thus eliminating the need for a second, boost vaccination. M8VLP immunization also provided long-term protection against influenza infection at 4 months postimmunization, compared to animals immunized with VLP alone. Furthermore, M8 adjuvantation of VLP increased endpoint and HAI antibody titers and further decreased influenza virus replication in lungs, compared with VLP adjuvanted with alum, AddaVax, and poly(I·C). Using VLP as immunogen, alum and AddaVax had similar adjuvanting

Journal of Virology

jvi.asm.org

10621

Beljanski et al.

FIG 8 Quantification of IgG subclasses from i.m. vaccinated animals (n ⫽ 8) and intracellular cytokine levels in T cells isolated from the spleens of i.p. vaccinated animals (n ⫽ 5) after 24 h of VLP stimulation. (A) IgG subclasses, IgG1, IgG2a, IgG2b, and IgG3, from sera of vaccinated animals (as indicated below the bars) were determined by ELISA using HA-coated plates. (B to F) The percentages of cytokine-secreting cells were obtained by subtracting the numbers of cytokinesecreting cells that were not stimulated with VLP. The percentages of IFN-␥⫹ CD8hi cells (B), IL-2⫹ CD4hi cells (C), TNF-␣⫹ CD4hi cells (D), IFN-␥⫹ CD4hi cells (E), and IL-10⫹ CD4hi cells (F) are shown. All values are expressed as means plus SEMs. Values that are significantly different are indicated by bars and asterisks as follows: *, P ⱕ 0.05; **, P ⱕ 0.01; ***, P ⱕ 0.005.

properties, a result that is in agreement with a recent study that utilized amyloid beta protein as an immunogen (52). M8 clearly manifested potent adjuvant activity in combination with influenza VLP, with a mechanism that appears distinct from alum or AddaVax. M8 induced a potent proinflammatory TH1biased immune response, whereas both alum-VLP and AddaVaxVLP have been shown to induce a stronger TH2 type response, and thus may be less efficient in vaccine formulations against patho-

10622

jvi.asm.org

gens that require TH1 cell-mediated immunity such as influenza virus (53). Such limitations can potentially be circumvented using an adjuvant such as M8, and the cumulative evidence indicates that TH1 memory T cells can be sustained in an antigen-independent manner for many years or even decades after vaccination (54). A role for T cells in M8-VLP vaccination was also indicated by the increased expression of TH1 cytokines in CD4 T cells, which was demonstrated following stimulation of splenocytes with VLP;

Journal of Virology

October 2015 Volume 89 Number 20

Adjuvant Activity of a RIG-I Agonist

this result suggests that VLP antigens are presented via MHC class II-restricted epitopes upon i.p. vaccination (55). Most vaccines provide protection by generating long-lived antibody-secreting plasma cells that impair establishment of an infection. The germinal center is a specialized microenvironment formed in secondary lymphoid tissues that produce memory B cells and antibody-secreting plasma cells after infection or immunization and thus provide prolonged protection against infection (56). After immunization via an intraperitoneal route, increased levels of GC B cells were observed in the spleens of animals immunized with AddaVax-VLP and alum-VLP. This observation likely reflects differences in the route of administration—i.p. administration favors adjuvants that create an antigen depot, whereas i.m. administration of M8 generates a strong but transient response in muscle cells, leading to a potent adaptive immune response to M8-VLP. Future studies will explore the relative contribution of plasma cells and T cells in the long-term protective responses generated in response to M8-VLP vaccinations. Altogether, these results highlight the efficacy of a novel RIG-I agonist functioning as a vaccine adjuvant to enhance the adaptive immune response to influenza virus. The strategy of combining M8 with VLP rescued animals from a lethal challenge with H5N1 influenza virus and stimulated humoral and cell-mediated responses against influenza virus compared to currently available adjuvants. However, the CD4 T cell response was TH1 biased, resulting in increased IgG2 levels, an observation not seen for alum or AddaVax. Furthermore, while the complex mechanism of action for both alum and AddaVax remain unclear, M8 specifically and transiently stimulates a well-characterized antiviral signaling pathway. This stimulation is similar to both de novo viral infection, as well as vaccination with LAIV in which viral replication transiently activates an acute immune response, leading to a long-term immunological memory. However, LAIV contain live attenuated organisms, thus raising safety and stability concerns. Thus, we propose that VLP-based vaccines adjuvanted with a PRR agonist(s) such as M8, may offer a viable alternative to the use of LAIV. Additionally, increase of immunogenicity with M8 is not only limited to VLP-based antigens but can be utilized to boost efficacy of other vaccine formulations, including TIV as well as other vaccines with low immunogenicity.

5.

6.

7. 8. 9.

10. 11.

12. 13.

14. 15. 16.

17.

ACKNOWLEDGMENTS This research was supported by grants from the National Institutes of Health (AI108861 to J.H.) and the Vaccine & Gene Therapy Institute of Florida to J.H. We thank Corey Crevar for technical assistance, as well as Ruben Donis and Richard Webby for providing the H5N1 reassortant viruses. We declare that we have no conflicts of interest.

REFERENCES

19. 20. 21.

1. Hannoun C. 2013. The evolving history of influenza viruses and influenza vaccines. Expert Rev Vaccines 12:1085–1094. http://dx.doi.org/10.1586 /14760584.2013.824709. 2. McKeage K. 2013. Inactivated quadrivalent split-virus seasonal influenza vaccine (FluarixR quadrivalent): a review of its use in the prevention of disease caused by influenza A and B. Drugs 73:1587–1594. http://dx.doi .org/10.1007/s40265-013-0114-3. 3. Atmar RL, Keitel WA. 2009. Adjuvants for pandemic influenza vaccines. Curr Top Microbiol Immunol 333:323–344. http://dx.doi.org/10.1007 /978-3-540-92165-3_16. 4. Baldwin SL, Bertholet S, Reese VA, Ching LK, Reed SG, Coler RN. 2012. The importance of adjuvant formulation in the development of a

October 2015 Volume 89 Number 20

18.

22.

23.

tuberculosis vaccine. J Immunol 188:2189 –2197. http://dx.doi.org/10 .4049/jimmunol.1102696. Metcalf TU, Cubas RA, Ghneim K, Cartwright MJ, Grevenynghe JV, Richner JM, Olagnier DP, Wilkinson PA, Cameron MJ, Park BS, Hiscott JB, Diamond MS, Wertheimer AM, Nikolich-Zugich J, Haddad EK. 2015. Global analyses revealed age-related alterations in innate immune responses after stimulation of pathogen recognition receptors. Aging Cell 14:421– 432. http://dx.doi.org/10.1111/acel.12320. Goff PH, Hayashi T, Martinez-Gil L, Corr M, Crain B, Yao S, Cottam HB, Chan M, Ramos I, Eggink D, Heshmati M, Krammer F, Messer K, Pu M, Fernandez-Sesma A, Palese P, Carson DA. 2015. Synthetic Tolllike receptor 4 (TLR4) and TLR7 ligands as influenza virus vaccine adjuvants induce rapid, sustained, and broadly protective responses. J Virol 89:3221–3235. http://dx.doi.org/10.1128/JVI.03337-14. Schneider-Ohrum K, Ross TM. 2012. Virus-like particles for antigen delivery at mucosal surfaces. Curr Top Microbiol Immunol 354:53–73. http://dx.doi.org/10.1007/82_2011_135. Alving CR, Peachman KK, Rao M, Reed SG. 2012. Adjuvants for human vaccines. Curr Opin Immunol 24:310 –315. http://dx.doi.org/10.1016/j .coi.2012.03.008. Osterholm MT, Kelley NS, Sommer A, Belongia EA. 2012. Efficacy and effectiveness of influenza vaccines: a systematic review and metaanalysis. Lancet Infect Dis 12:36 – 44. http://dx.doi.org/10.1016/S1473 -3099(11)70295-X. Giles BM, Ross TM. 2012. Computationally optimized antigens to overcome influenza viral diversity. Expert Rev Vaccines 11:267–269. http://dx .doi.org/10.1586/erv.12.3. Leroux-Roels I, Borkowski A, Vanwolleghem T, Drame M, Clement F, Hons E, Devaster JM, Leroux-Roels G. 2007. Antigen sparing and crossreactive immunity with an adjuvanted rH5N1 prototype pandemic influenza vaccine: a randomised controlled trial. Lancet 370:580 –589. http: //dx.doi.org/10.1016/S0140-6736(07)61297-5. Schwendener RA. 2014. Liposomes as vaccine delivery systems: a review of the recent advances. Ther Adv Vaccines 2:159 –182. http://dx.doi.org /10.1177/2051013614541440. Lee YN, Kim MC, Lee YT, Hwang HS, Cho MK, Lee JS, Ko EJ, Kwon YM, Kang SM. 2014. AS04-adjuvanted virus-like particles containing multiple M2 extracellular domains of influenza virus confer improved protection. Vaccine 32:4578 – 4585. http://dx.doi.org/10.1016/j.vaccine .2014.06.040. Reed SG, Orr MT, Fox CB. 2013. Key roles of adjuvants in modern vaccines. Nat Med 19:1597–1608. http://dx.doi.org/10.1038/nm.3409. Coffman RL, Sher A, Seder RA. 2010. Vaccine adjuvants: putting innate immunity to work. Immunity 33:492–503. http://dx.doi.org/10.1016/j .immuni.2010.10.002. Jiang F, Ramanathan A, Miller MT, Tang GQ, Gale M, Jr, Patel SS, Marcotrigiano J. 2011. Structural basis of RNA recognition and activation by innate immune receptor RIG-I. Nature 479:423– 427. http://dx.doi.org /10.1038/nature10537. Komuro A, Bamming D, Horvath CM. 2008. Negative regulation of cytoplasmic RNA-mediated antiviral signaling. Cytokine 43:350 –358. http://dx.doi.org/10.1016/j.cyto.2008.07.011. Belgnaoui SM, Paz S, Hiscott J. 2011. Orchestrating the interferon antiviral response through the mitochondrial antiviral signaling (MAVS) adapter. Curr Opin Immunol 23:564 –572. http://dx.doi.org/10.1016/j.coi .2011.08.001. Loo YM, Gale M, Jr. 2011. Immune signaling by RIG-I-like receptors. Immunity 34:680 – 692. http://dx.doi.org/10.1016/j.immuni.2011.05.003. Takeuchi O, Akira S. 2010. Pattern recognition receptors and inflammation. Cell 140:805– 820. http://dx.doi.org/10.1016/j.cell.2010.01.022. Schoggins JW, Wilson SJ, Panis M, Murphy MY, Jones CT, Bieniasz P, Rice CM. 2011. A diverse range of gene products are effectors of the type I interferon antiviral response. Nature 472:481– 485. http://dx.doi.org/10 .1038/nature09907. Longhi MP, Trumpfheller C, Idoyaga J, Caskey M, Matos I, Kluger C, Salazar AM, Colonna M, Steinman RM. 2009. Dendritic cells require a systemic type I interferon response to mature and induce CD4⫹ Th1 immunity with poly IC as adjuvant. J Exp Med 206:1589 –1602. http://dx.doi .org/10.1084/jem.20090247. Goulet ML, Olagnier D, Xu Z, Paz S, Belgnaoui SM, Lafferty EI, Janelle V, Arguello M, Paquet M, Ghneim K, Richards S, Smith A, Wilkinson P, Cameron M, Kalinke U, Qureshi S, Lamarre A, Haddad EK, Sekaly RP, Peri S, Balachandran S, Lin R, Hiscott J. 2013.

Journal of Virology

jvi.asm.org

10623

Beljanski et al.

24.

25.

26.

27. 28. 29.

30. 31. 32. 33. 34.

35.

36. 37. 38. 39. 40. 41. 42.

Systems analysis of a RIG-I agonist inducing broad spectrum inhibition of virus infectivity. PLoS Pathog 9:e1003298. http://dx.doi.org/10 .1371/journal.ppat.1003298. Chiang C, Beljanski V, Yin K, Olagnier D, Ben Yebdri F, Steel C, Goulet ML, DeFilippis VR, Streblow DN, Haddad EK, Trautmann L, Ross T, Lin R, Hiscott J. 2015. Sequence-specific modifications enhance the broad-spectrum antiviral response activated by RIG-I agonists. J Virol 89:8011– 8025. http://dx.doi.org/10.1128/JVI.00845-15. Bright RA, Carter DM, Crevar CJ, Toapanta FR, Steckbeck JD, Cole KS, Kumar NM, Pushko P, Smith G, Tumpey TM, Ross TM. 2008. Crossclade protective immune responses to influenza viruses with H5N1 HA and NA elicited by an influenza virus-like particle. PLoS One 3:e1501. http://dx.doi.org/10.1371/journal.pone.0001501. Giles BM, Ross TM. 2011. A computationally optimized broadly reactive antigen (COBRA) based H5N1 VLP vaccine elicits broadly reactive antibodies in mice and ferrets. Vaccine 29:3043–3054. http://dx.doi.org/10 .1016/j.vaccine.2011.01.100. National Research Council. 2011. Guide for the care and use of laboratory animals, 8th ed. National Academies Press, Washington, DC. US Congress. 1966. Public Law 89-544 Act of August 24, 1966. United States Code, Title 7, Sections 2131-2156. Feng Q, Langereis MA, Olagnier D, Chiang C, van de Winkel R, van Essen P, Zoll J, Hiscott J, van Kuppeveld FJ. 2014. Coxsackievirus cloverleaf RNA containing a 5= triphosphate triggers an antiviral response via RIG-I activation. PLoS One 9:e95927. http://dx.doi.org/10.1371 /journal.pone.0095927. Centers for Disease Control and Prevention. August 2011. Chapter 6: influenza. Manual for the surveillance of vaccine-preventable diseases. Centers for Disease Control and Prevention, Atlanta, GA. Olagnier D, Hiscott J. 2014. Type I and type III interferon-induced immune response: it’s a matter of kinetics and magnitude. Hepatology 59:1225–1228. http://dx.doi.org/10.1002/hep.26959. Schraml BU, Reis e Sousa C. 2015. Defining dendritic cells. Curr Opin Immunol 32:13–20. http://dx.doi.org/10.1016/j.coi.2014.11.001. Palucka K, Banchereau J. 2012. Cancer immunotherapy via dendritic cells. Nat Rev Cancer 12:265–277. http://dx.doi.org/10.1038/nrc3258. Yu M, Levine SJ. 2011. Toll-like receptor, RIG-I-like receptors and the NLRP3 inflammasome: key modulators of innate immune responses to double-stranded RNA viruses. Cytokine Growth Factor Rev 22:63–72. http://dx.doi.org/10.1016/j.cytogfr.2011.02.001. Goff PH, Eggink D, Seibert CW, Hai R, Martinez-Gil L, Krammer F, Palese P. 2013. Adjuvants and immunization strategies to induce influenza virus hemagglutinin stalk antibodies. PLoS One 8:e79194. http://dx .doi.org/10.1371/journal.pone.0079194. Dormitzer PR, Galli G, Castellino F, Golding H, Khurana S, Del Giudice G, Rappuoli R. 2011. Influenza vaccine immunology. Immunol Rev 239:167–177. http://dx.doi.org/10.1111/j.1600-065X.2010.00974.x. Mbow ML, De Gregorio E, Valiante NM, Rappuoli R. 2010. New adjuvants for human vaccines. Curr Opin Immunol 22:411– 416. http://dx .doi.org/10.1016/j.coi.2010.04.004. Visciano ML, Tagliamonte M, Tornesello ML, Buonaguro FM, Buonaguro L. 2012. Effects of adjuvants on IgG subclasses elicited by virus-like particles. J Transl Med 10:4. http://dx.doi.org/10.1186/1479-5876-10-4. Nazmi A, Mukhopadhyay R, Dutta K, Basu A. 2012. STING mediates neuronal innate immune response following Japanese encephalitis virus infection. Sci Rep 2:347. http://dx.doi.org/10.1038/srep00347. Roldão A, Mellado MC, Castilho LR, Carrondo MJ, Alves PM. 2010. Virus-like particles in vaccine development. Expert Rev Vaccines 9:1149 – 1176. http://dx.doi.org/10.1586/erv.10.115. Palmer AK, Harris AL, Jacobson RM. 2014. Human papillomavirus vaccination: a case study in translational science. Clin Transl Sci 7:420 – 424. http://dx.doi.org/10.1111/cts.12166. Szabo A, Rajnavolgyi E. 2013. Collaboration of Toll-like and RIG-I-like

10624

jvi.asm.org

43.

44.

45. 46.

47.

48.

49.

50.

51.

52.

53. 54. 55.

56.

receptors in human dendritic cells: tRIGgering antiviral innate immune responses. Am J Clin Exp Immunol 2:195–207. Ichinohe T, Watanabe I, Ito S, Fujii H, Moriyama M, Tamura S, Takahashi H, Sawa H, Chiba J, Kurata T, Sata T, Hasegawa H. 2005. Synthetic double-stranded RNA poly(I:C) combined with mucosal vaccine protects against influenza virus infection. J Virol 79:2910 –2919. http: //dx.doi.org/10.1128/JVI.79.5.2910-2919.2005. Kulkarni RR, Rasheed MA, Bhaumik SK, Ranjan P, Cao W, Davis C, Marisetti K, Thomas S, Gangappa S, Sambhara S, Murali-Krishna K. 2014. Activation of the RIG-I pathway during influenza vaccination enhances the germinal center reaction, promotes T follicular helper cell induction, and provides a dose-sparing effect and protective immunity. J Virol 88:13990 –14001. http://dx.doi.org/10.1128/JVI.02273-14. Martinez-Gil L, Goff PH, Hai R, Garcia-Sastre A, Shaw ML, Palese P. 2013. A Sendai virus-derived RNA agonist of RIG-I as a virus vaccine adjuvant. J Virol 87:1290 –1300. http://dx.doi.org/10.1128/JVI.02338-12. Maisonneuve C, Bertholet S, Philpott DJ, De Gregorio E. 2014. Unleashing the potential of NOD- and Toll-like agonists as vaccine adjuvants. Proc Natl Acad Sci U S A 111:12294 –12299. http://dx.doi.org/10.1073 /pnas.1400478111. Strauss R, Huser A, Ni S, Tuve S, Kiviat N, Sow PS, Hofmann C, Lieber A. 2007. Baculovirus-based vaccination vectors allow for efficient induction of immune responses against Plasmodium falciparum circumsporozoite protein. Mol Ther 15:193–202. http://dx.doi.org/10 .1038/sj.mt.6300008. Calabro S, Tortoli M, Baudner BC, Pacitto A, Cortese M, O’Hagan DT, De Gregorio E, Seubert A, Wack A. 2011. Vaccine adjuvants alum and MF59 induce rapid recruitment of neutrophils and monocytes that participate in antigen transport to draining lymph nodes. Vaccine 29:1812– 1823. http://dx.doi.org/10.1016/j.vaccine.2010.12.090. Loo YM, Fornek J, Crochet N, Bajwa G, Perwitasari O, MartinezSobrido L, Akira S, Gill MA, Garcia-Sastre A, Katze MG, Gale M, Jr. 2008. Distinct RIG-I and MDA5 signaling by RNA viruses in innate immunity. J Virol 82:335–345. http://dx.doi.org/10.1128/JVI.01080-07. Olagnier D, Scholte FE, Chiang C, Albulescu IC, Nichols C, He Z, Lin R, Snijder EJ, van Hemert MJ, Hiscott J. 2014. Inhibition of dengue and Chikungunya virus infections by RIG-I-mediated type I interferonindependent stimulation of the innate antiviral response. J Virol 88:4180 – 4194. http://dx.doi.org/10.1128/JVI.03114-13. Pichlmair A, Schulz O, Tan CP, Naslund TI, Liljestrom P, Weber F, Reis e Sousa C. 2006. RIG-I-mediated antiviral responses to singlestranded RNA bearing 5=-phosphates. Science 314:997–1001. http://dx .doi.org/10.1126/science.1132998. Mantile F, Trovato M, Santoni A, Barba P, Ottonello S, De Berardinis P, Prisco A. 2014. Alum and squalene-oil-in-water emulsion enhance the titer and avidity of anti-Abeta antibodies induced by multimeric protein antigen (1-11)E2, preserving the Igg1-skewed isotype distribution. PLoS One 9:e101474. http://dx.doi.org/10.1371/journal.pone.0101474. Petrovsky N, Aguilar JC. 2004. Vaccine adjuvants: current state and future trends. Immunol Cell Biol 82:488 – 496. http://dx.doi.org/10.1111 /j.0818-9641.2004.01272.x. Foulds KE, Wu CY, Seder RA. 2006. Th1 memory: implications for vaccine development. Immunol Rev 211:58 – 66. http://dx.doi.org/10 .1111/j.0105-2896.2006.00400.x. Herkel J, Jagemann B, Wiegard C, Lazaro JF, Lueth S, Kanzler S, Blessing M, Schmitt E, Lohse AW. 2003. MHC class II-expressing hepatocytes function as antigen-presenting cells and activate specific CD4 T lymphocytes. Hepatology 37:1079 –1085. http://dx.doi.org/10.1053/jhep .2003.50191. Victora GD, Nussenzweig MC. 2012. Germinal centers. Annu Rev Immunol 30:429 – 457. http://dx.doi.org/10.1146/annurev-immunol -020711-075032.

Journal of Virology

October 2015 Volume 89 Number 20