key: cord-0688823-nb6c8buz
authors: Kwak, Hye Won; Park, Hyo-Jung; Ko, Hae Li; Park, Hyelim; Cha, Min Ho; Lee, Sang-Myeong; Kang, Kyung Won; Kim, Rhoon-Ho; Ryu, Seung Rok; Kim, Hye-Jung; Kim, Jae-Ouk; Song, Manki; Kim, Hun; Jeong, Dae Gwin; Shin, Eui-Cheol; Nam, Jae-Hwan
title: Cricket paralysis virus internal ribosome entry site-derived RNA promotes conventional vaccine efficacy by enhancing a balanced Th1/Th2 response
date: 2019-08-23
journal: Vaccine
DOI: 10.1016/j.vaccine.2019.07.070
sha: 53ad783bd651dab3f1ff08e681a286db611075aa
doc_id: 688823
cord_uid: nb6c8buz

An ideal adjuvant should increase vaccine efficacy through balanced Th1/Th2 responses and be safe to use. Recombinant protein-based vaccines are usually formulated with aluminum (alum)-based adjuvants to ensure an adequate immune response. However, use of alum triggers a Th2-biased immune induction, and hence is not optimal. Although the adjuvanticity of RNA has been reported, a systematic and overall investigation on its efficacy is lacking. We found that single strand RNA (termed RNA adjuvant) derived from cricket paralysis virus intergenic region internal ribosome entry site induced the expression of various adjuvant-function-related genes, such as type 1 and 2 interferon (IFN) and toll-like receptor (TLR), T cell activation, and leukocyte chemotaxis in human peripheral blood mononuclear cells; furthermore, its innate and IFN transcriptome profile patterns were similar to those of a live-attenuated yellow fever vaccine. This suggests that protein-based vaccines formulated using RNA adjuvant function as live-attenuated vaccines. Application of the RNA adjuvant in mouse enhanced the efficacy of Middle East respiratory syndrome spike protein, a protein-subunit vaccine and human papillomavirus L1 protein, a virus-like particle vaccine, by activating innate immune response through TLR7 and enhancing pAPC chemotaxis, leading to a balanced Th1/Th2 responses. Moreover, the combination of alum and the RNA adjuvant synergistically induced humoral and cellular immune responses and endowed long-term immunity. Therefore, RNA adjuvants have broad applicability and can be used with all conventional vaccines to improve vaccine efficacy qualitatively and quantitively.

Although inactivated and protein-based vaccines have long been used to prevent the spread of various pathogens [1] successfully, some of them elicit a weak immune response, short immune duration, and have several safety issues [2, 3] . Therefore, research has focused on identifying antigen components responsible for immunogenicity in order to generate safer and more effective vaccines. However, purified recombinant antigens generally show poor immunogenicity [2] . To overcome this, immunostimulatory agents (i.e., adjuvants) are added to vaccines to induce robust immune responses against vaccine antigens [4] [5] [6] . Abbreviations: APCs, Antigen presenting cells; DCs, dendritic cells; Th2, T helper 2; TLR, Toll-like receptor; IFN, interferon; ssRNAs, single-stranded RNAs; IL, interleukin; Tfh, follicular helper T; RIG, retinoic acid-inducible gene; IRESs, internal ribosome entry sites; CrPV, Cricket paralysis virus; IGR, intergenic region; MERS, Middle East respiratory syndrome; CoV, coronavirus; HPV, human papillomavirus; dLN, draining lymph node; NAbs, neutralizing antibodies; VLP, virus-like particle; Tg, transgenic; WT, wild-type; ELISA, enzyme-linked immunosorbent assay; PRNT, plaque-reduction neutralization test; PFUs, plaque-forming units; ELISPOT, enzyme-linked immunospot; hPBMCs, human peripheral blood mononuclear cells; TNF, tumor necrosis factor; Myd88, myeloid differentiation primary response 88.

Aluminum hydroxide (alum) is widely used as an adjuvant in several human vaccines [7] . It has been reported that alum injection leads to the rapid accumulation of neutrophils and other cells forming lumps composed of neutrophil extracellular traps. This further induces lysosome-mediated lysis to distort the plasma membrane of antigen presenting cells (APCs) and/or rapidly induces the synthesis cytokines/chemokines, finally activating the NALP3 inflammasome and triggering the secretion of IL-1b. This process might contribute to the adjuvant effect of alum. However, further confirmation is required [8] . In addition, alum induces a profound T helper 2 (Th2) response along with antibody production. However, alum cannot induce Th1 responses and sometimes stimulates autoimmunity by upregulating IgE presentation [9] . Since classical adjuvants solely induce Th2 responses, new adjuvants capable of promoting Th1 responses are required to increase the cell-mediated response [8] . Non-classical adjuvants, such as Toll-like receptor (TLR) ligands, can stimulate the production of type I interferon (IFN) and proinflammatory cytokines [10] . Type I IFN facilitates antigen presentation to major histocompatibility complex class I/II, leading to an enhanced cytotoxic CD8+ T cell response, as well as a CD4+ T cell response [11] .

Since its first use as an adjuvant with the inactivated Japanese encephalitis virus (JEV) vaccine to efficiently promote neutralizing antibodies (NAbs), IFN production, and resistance to JEV infection, poly I:C and its derivates have been widely used in the development of various vaccines [12, 13] . In addition, RNase-resistant poly I:C stabilized with poly-L-lysine (poly ICLC) was shown to have strong adjuvanticity by microbial mimicking for inducing innate immune responses in humans through TLR3 and MDA5 [14] . Despite the strong adjuvanticity of poly I:C, there are concerns about its induction of adverse and excessive immune responses, such as autoimmune diseases and difficulty in standardization for mass production [15] . In previous studies, poly I:C-induced massive type I IFNs were shown to be critical for the pathogenesis of lupus [16] . Moreover, repeated administration of poly I:C derivate, poly ICLC, and carboxymethylcellulose, showed severe pathogenicity and resulted in the development of tuberculosis in mice [17] .

Single-stranded RNAs (ssRNAs) function as vaccine adjuvants and promote Th1 responses through the induction of IFN-a and interleukin (IL)-12 in DCs [18] . Sendai virus-derived defective interfering RNA activates the innate immune response, leading to increased efficacy of inactivated influenza vaccine [19] . Additionally, nucleoside-modified influenza mRNA vaccines showed adjuvant effects by inducing antigen-specific follicular helper T (Tfh) cells and germinal center B cell responses [20] . Recently, ssRNA formulated with carriers showed immune-enhancement effects through TLR and retinoic acid-inducible gene (RIG)-like helicase signaling pathways [21, 22] . Although these studies demonstrated that ssRNAs could function as adjuvants, systematic investigations of their effects and applicability to protein-based or inactivated vaccines are lacking, thereby preventing their widespread clinical adoption.

We previously developed novel ssRNA-expression platforms from several viral internal ribosome entry sites (IRESs) [23] . In this study, we investigated the adjuvant effects of ssRNAs (RNA adjuvant) derived from the Cricket paralysis virus (CrPV) intergenic region (IGR) IRES [23] . We formulated the RNA adjuvant with protein-subunit vaccine for Middle East Respiratory Syndrome (MERS)-coronavirus (CoV) and virus-like particle vaccine for human papillomavirus (HPV) and showed that the RNA adjuvant induced a balanced Th1/Th2 response and innate immunity. Moreover, as expected, the RNA adjuvant worked through TLR7 and recruited APCs to increase antigen uptake into the injected site and their retention in the draining lymph node (dLN). Furthermore, we showed that the RNA adjuvant enhanced humoral and cellular immune responses by promoting antigen-specific neutralizing antibodies (NAbs) and the Th1 response and ultimately endowing long-term immunity. The results demonstrated that an RNA adjuvant could be developed as a general, safe, and effective adjuvant applicable to protein-based vaccine types.

For immunization with MERS S soluble protein vaccine and HPV L1 virus-like particle (VLP) vaccine, female BALB/c and C57BL/6 mice were purchased from Dae-Han Bio-Link. For MERS-CoV challenge, hDPP4-transgenic (Tg) mice (5-7weeks old) were maintained in a biosafety level 3 facility at the Korea Zoonosis Research Institute of Chonbuk National University and followed the guidelines of the Institutional Animal Care and Use Committee (2018-045).

Soluble MERS spike (S) protein vaccine was provided by the International Vaccine Institute (Seoul, Korea). This MERS S protein expressed in insect cells has no transmembrane domain and comprises the first 1296 amino acids of MERS-CoV EMC/2012 strain (GenBank #AFS88936.1). VLPs containing HPV L1 (a mixture of 10 types of L1 - 6, 11, 16, 18, 31, 33, 35, 45, 52 , and 58 L1) were obtained from SK Bioscience (Seoul, Korea) and expressed in insect cells. Due to patent issues, detailed information on the HPV vaccine was unavailable.

The DNA platform was designed using the IGR IRES and SV40 late-polyadenylation signal sequences ( Supplementary Fig. 1 ) [23] . DNA templates were linearized with NotI. In vitro transcription was performed using the EZ T7 high yield in vitro transcription kit (Enzynomics, Daejeon, Korea) and HiScribe T7 Quick high yield RNA synthesis kit (New England Biolabs, Ipswich, MA, USA). See Supplementary Methods for more details. 

For MERS S protein vaccine studies, C57BL/6 WT and hDPP4-Tg mice (6-weeks old) were inoculated intramuscularly into the upper thigh twice a week at 2-week intervals with the following formulations: (1) 1 mg MERS S protein vaccine with/without 20 mg RNA adjuvant or 500 mg alum (Thermo Fisher Scientific, Waltham, MA, USA) for wild-type (WT) mice; and (2) 1 mg MERS S protein vaccine with/without 20 mg RNA adjuvant or 24 mg alum (Brentanne, Frederikssund, Denmark) for hDPP4-Tg mice. For VLP-HPV-L1 vaccine studies, BALB/c mice (6-weeks old) were inoculated by intramuscular injection into the upper thigh three times a week at 2-week intervals with 6 mg VLP-HPV-L1 with/without 20 mg protamineformulated RNA adjuvant or 5 mg alum.

Antigen-specific IgG1, IgG2a, and IgG2c in mouse serum were measured by ELISA. The 96-well plates (Corning, Corning, YN, USA) were coated with 50 ng/well MERS S protein and 100 ng/well VLP-HPV-L1 vaccine and incubated overnight at 4°C. See Supplementary Methods for more details.

Serum from MERS-CoV infected hDPP4-Tg mice were serially diluted from 10-to 5120-fold with serum-free medium, and the virus-serum mixture was prepared by mixing 100 plaqueforming units (PFUs) of MERS-CoV using the diluted serum samples and incubated at 37°C for 1-h. HPV-specific NAb titration was performed as described previously [24] . See Supplementary Methods for more details.

Splenocytes from immunized mice and GPs at the end of the experiments were stimulated with 0.125-1 lg/well of antigens for 48 h at 37°C. ELISPOT for the detection of IFN-c-secreting T cells was performed according to manufacturer instructions (Mabtech, Stockholm, Sweden).

For surface staining, splenocytes and isolated immune cells from muscle and lymph nodes were stained with the following antibodies for 15 min at room temperature; CD4 (Clone GK1.5, eBioscience; Clone H129.19, Bio Legend), CD8 (Clone 53-6.7, BD Biosciences; Clone 53-6.7, Invitrogen), CD69 (Clone H1.2F3, BD Biosciences), CD44 (Clone IM7, Invitrogen), CD62L (Clone MEL 14, BD Biosciences), CD11b (Clone M1/70, Bio Legend), F4/80 (Clone BM8, Invitrogen), CD86 (Clone GL1, BD Biosciences), and CD11c (Clone N48, eBioscience). Cells were fixed with 1% paraformaldehyde, analyzed using a FACS Canto II flow cytometer (BD Biosciences), and the data were analyzed using FlowJo (TreeStar). For determining polyfunctional T cells, isolated splenocytes were re-stimulated with 1 lg/well MERS spike protein or 100 ng/well of the peptide mixture. To evaluate the cytotoxic degranulation activity of T cells, anti-CD107a-BV421 (Clone 1D4B, BD Biosciences) was added into the culture medium. Brefeldin A (Golgi-Plug, BD Biosciences) and monensin (GolgiStop, BD Biosciences) were added 2 h later. After another 10 h incubation, splenocytes were first stained with ethidium monoazide (Sigma) and then stained with anti-CD3-BV510, anti-CD4-APCH7, and anti-CD8-PE-cy7 (Clone 145-2C11, Clone GK1.5, and Clone 53-6.7) (BD Biosciences). The stained cells were permeabilized using Cytofix/Cytoperm kit (eBioscience) and then stained with anti-IFN-c-APC, anti-TNF-a-FITC, and anti-IL-2-PE (Clone XMG1.2, BD Biosciences; Clone MP6-XT22, Invitrogen; Clone JES6-5H4, eBioscience). Cells were fixed with 1% paraformaldehyde, analyzed using an LSRII flow cytometer (BD Biosciences), and T cells positive for the various combinations of cytokines and degranulation were analyzed and quantified using a Boolean gating function in FlowJo (TreeStar).

A laser-scanning intravital confocal microscope (IVM-C, IVIM Technology) was used to visualize dendritic cells and monocyte recruitment. CX3CR1-GFP mice (Jackson Laboratory, Bar Harbor, ME, USA, stock #005582) that endogenously express GFP in dendritic cells and monocytes were used. Mice were deeply anesthetized with intraperitoneal injections of zoletil (30 mg/kg) and rompun (10 mg/kg). During imaging, mouse body temperature was maintained at 37°C with a homeothermic controller (Physio-Suite; RightTemp; Kent Scientific). A mixture of Alexa Fluor 647 (A20006, Thermo Fisher Scientific) and either RNA adjuvant or PBS was subcutaneously injected into the mouse ear skin using a 31G microinjector (1700 series syringe, Hamilton). To label vasculatures, an anti-CD31 antibody (553369, BD Biosciences) conjugated with Alexa Flour 555 (A20009, Thermo Fisher Scientific) was injected intravascularly 1 h before imaging. Using the distinctive vasculature and the fluorescence signal of Alexa Fluor 647 as a landmark for the repetitive imaging of the injected site, the recruitment of CX3CR1-GFP + dendritic cells (irregular outlines) and monocytes (round outlines) was imaged in vivo at 20 min and 24 h after the injection. The recruitment of dendritic cells and monocytes was quantified by calculating the area occupied by CX3CR1-GFP + cells using ImageJ plugins (NIH).

After stimulating the human peripheral blood mononuclear cells (hPBMCs) with the RNA adjuvant for 24 h, concentrations of tumor necrosis factor (TNF)-a, IFN-c, IL-2, IL-12, IL-6, and IL-10 were analyzed using the Magnetic Luminex screening assay kit (R&D Systems, Minneapolis, MN, USA) according to manufacturer instructions.

To examine the effects on T cells, C57BL/6 mice (6-weeks old) were inoculated by intramuscular injection twice at 6-day intervals with 1 mg MERS S protein vaccine with 20 mg RNA adjuvant, followed by inoculation by intraperitoneal injection with 10 mg/ kg anti-mouse CD4 (YTS 191; BioXcell, West Lebanon, NH, USA) and/or 10 mg/kg anti-mouse CD8 (YTS 169.4; BioXcell) every 3days after the first immunization.

One-way analysis of variance was used to assess significant differences among treatment groups. For each significant treatment effect, the Tukey HSD test was used to compare multiple group means. In addition, all histomorphometrical values are expressed as means ± standard deviation (SD). Multiple comparison tests for the different treatment groups were conducted. If the Levene test indicated no significant deviations from variance homogeneity, the data obtained were analyzed by a least-significant differences multi-comparison test. If significant deviations from variance homogeneity were detected using the Levene test, the nonparametric Kruskal-Wallis H-test was conducted. When a significant difference was detected by the Kruskal-Wallis H-test, a Mann-Whitney U-test was conducted as a post hoc analysis. Statistical analyses were conducted using SPSS for Windows (Release 14.0K, SPSS Inc.). To assess significant differences between two groups, Student's t-test was used. Differences were considered significant at P < 0.05.

3.1. The RNA adjuvant affects genes related to the innate immune responses, T cell activation, and chemotaxis similar to poly(I:C) in hPBMCs

To investigate the adjuvant effect, we treated hPBMCs with the RNA adjuvant and used poly(I:C) as a positive control, finding that the RNA adjuvant induced Th1-and Th2-related cytokine release similar to poly(I:C) (Fig. 1A) . We then compared the transcriptome profiles of RNA adjuvant-and poly(I:C)-treated hPBMCs by RNA sequencing. Control, poly(I:C), and RNA adjuvant treatment formed a separate cluster without any bias (Fig. 1B) , with scatter plots of gene profiling confirming that changes in geneexpression between the RNA adjuvant and poly(I:C) groups were not greater than those between the control and RNA adjuvant groups at 6-and 24-h post-treatment (p 0.05; fold-change ! 2) (Fig. 1B) .

To investigate the biological pathways associated with genes affected by the RNA adjuvant, we analyzed gene-expression patterns using the Kyoto Encyclopedia of Genes and Genomes database (https://www.genome.jp/kegg/mapper.html) (Fig. 1C ) and the network of genes involved in each pathway based on the STRING database (Fig. 1D) . After 6-h, similar biological pathways were affected by RNA adjuvant and poly(I:C) treatment (Supplementary Fig. 2) , and heat maps of innate-immune-related genes [type 1 and 2 IFNs and TLR-myeloid differentiation primary response 88 (Myd88)], as well as T cell activation and leukocyte chemotaxis, indicated similar expression patterns between the two groups (Fig. 1C) . TLR2, TLR3, TLR7, IRF7, and NF-jB among the TLR related genes were induced, and IL-6 and IL-15 among T cell activation related genes and STAT 1 and 3 among multiple transcriptional factors function as key players in the RNA adjuvant group (Fig. 1D) . 

Yellow fever vaccine (YF17D) is the most successful liveattenuated vaccine, as its duration of protective immunity after single immunization is >35 years [25] . We compared the transcription profile induced in RNA adjuvant-treated hPBMCs (6-h posttreatment) with that of PBMCs obtained from a YF17Dimmunized patient (at 7-days post-immunization) [26] . Approximately 80% of the innate immune response and 100% of type I IFN-related genes, which are key factors contributing to a superior vaccine [14, 27] , showed similar changes in expression pattern between the two samples ( Fig. 2A and B) . The inductions of IRF7 and STAT1 are important upstream immune transcription factors targeting different effectors of the immune response to the YF17D vaccine [26] . Interestingly, increases in IRF7 and STAT1 levels in response to the RNA adjuvant were the same as those in response to YF17D (Fig. 2C) , indicating that the RNA adjuvant affected important transcription factors that play key roles in YF17D-vaccine efficacy.

The RNA adjuvant increased APC retention in the dLN to promote selection of antigen-specific T and B cells by epitopepresenting APCs in a process similar to that observed following treatment with poly(I:C) (Fig. 3A) . Furthermore, in vivo timelapse imaging clearly showed that the RNA adjuvant summoned immune cells, mostly DCs (irregular outlines), to the injection site ( Fig. 3B and C) , which was consistent with the heat map describing upregulated leukocyte chemotaxis (Fig. 1C) and indicating that the RNA adjuvant enhanced antigen uptake by DCs at injection sites and promoted epitope presentation in the dLN.

To confirm adjuvant effects in vivo, we formulated it with the MERS spike (S) protein and used it to immunize WT C57BL/6 mice with/without alum (Fig. 4A) . At 2-weeks after first immunization, G5 and G4 showed the highest IgG1 level (indicating a Th2 response). After boosting (at 2 weeks after second immunization), G2 to G5 showed similar IgG1 levels. Conversely, IgG2c (indicating a Th1 response) was only induced in G4 and G5 (Fig. 4B) , suggesting that the RNA adjuvant induced a Th1 response. Moreover, G5 showed the highest NAb level (indicating a strong Th2 response), and G3 and G4 showed similar levels (Fig. 4B) . Interestingly, G5 and G4 showed higher numbers of IFN-c secreting cells following stimulation with the MERS S protein as compared with G2 and G3 (Fig. 4C) , indicating that the RNA adjuvant induced MERS S protein-specific Th1 responses. Moreover, we observed elevations in the number of polyfunctional CD4+ T cells in G5 (Fig. 4D) . These results suggested that the RNA adjuvant promoted CD4+ T cell responses, especially Th1 responses, consequently inducing antigen-specific cellular immune responses. Furthermore, administering alum along with the RNA adjuvant generated a synergistic effect, leading to an increase in NAb levels by stimulating balanced Th1/Th2 responses.

To investigate the protective effect of immunization with the MERS S protein formulated with the RNA adjuvant plus alum (Fig. 5A) , we immunized hDPP4-Tg mice, a surrogate mouse model for MERS challenge [28] . Although the body weights of G2 and G3 were immediately reduced (but quickly recovered) after challenge (Fig. 5B) , we observed 100% survival (Fig. 5C) , possibly due to a sufficient NAb level derived from boosting immunization, although G2 showed lower NAb levels relative to G3 in hDPP4-Tg mice (Supplementary Fig. 3) .

To investigate the long-term immune effect of the RNA adjuvant, we immunized mice with a 10-value HPV vaccine as a VLP vaccine, an RNA adjuvant-containing 16 HPV-L1 gene (RNA adjuvant-HPV; Supplementary Fig. 1 ), protamine (used as a condensing component of RNA-cationic peptide complex [29] ), and alum ( Fig. 6A) . At 2-weeks after the first immunization, the IgG1 levels of G2 and G3 were similar, whereas the IgG2a level of G3 was higher than that of G2 (Fig. 6B) . After three doses (7-27 weeks after the first immunization), G2 and G3 showed similar IgG1 and IgG2a levels (Fig. 6B) ; however, G3 showed higher NAb levels against 11, 16, 18, 35, 45 , 52, and 58 HPVs, similar levels against 6 and 33 HPVs, and lower levels against 31 HPVs than G2 (Fig. 6C) . Furthermore, the frequency of IFN-c-secreting cells among splenocytes from G3 after stimulation with 16, 18, or all HPV-L1 proteins was significantly higher than that of G2, suggesting that G3 maintained an active HPV-L1-specific Th1 response, even at 27-weeks post-immunization (Fig. 6D ). This indicated that the RNA adjuvant-HPV helped maintain long-lasting humoral and cellular immune responses.

Our results indicated that the RNA adjuvant induced T cell activation and expression of TLR7-related genes, as well as a balanced Th1/Th2 response. To investigate this process, Tlr7-KO and WT mice were immunized with the MERS S protein, RNA adjuvant, and alum. As expected, IgG1 levels of in the KO mice were lower than those of WT mice at 2-weeks after first immunization (Fig. 7A) . However, IgG1 was increased, while IgG2c was not, in KO mice at 2-weeks after second immunization (Fig. 7A) , suggesting that the RNA adjuvant might have partially overcome the loss of TLR7 for Th2 induction but not Th1 via boosting. Mouse embryonic fibroblasts treated with the RNA adjuvant or a positive control (R848 as TLR7/8 agonist and poly I:C) showed the induction of Myd88, RIG-I, melanoma differentiation-associated protein 5, phosphorylated interferon regulatory factor 3, and TANK-binding kinase 1/NF-jB activating kinase, which are the downstream molecules of TLR7 and RIG-Ilike receptor pathways [30] (Supplementary Fig. 4) , suggesting that, while Th1 activation by RNA adjuvant may be critically dependent on the TLR7 pathway, Th2 activation by RNA adjuvant may be induced by multiple mechanisms that bypass the TLR7 signaling pathway, such as the pathway involving RIG-Ilike receptor, which is one of other pattern recognition receptors (PRRs).

To investigate the necessity of CD4+, CD8+, or both T cells to activate the Th1 response, these were respectively depleted in the mouse spleen ( Fig. 7B and Supplementary Fig. 5 ). CD4depleted mice (G3) showed significant reductions in IgG1 and IgG2c levels, whereas CD8-depleted mice (G4) showed no reduction in IgG1 levels but had reduced IgG2c levels. Moreover, mice lacking both CD4 and CD8 (G5) showed reductions in both IgG1 and IgG2c levels (Fig. 7C) , indicating that IgG1 was dependent upon CD4+ T cell activation alone, whereas IgG2c was dependent on both CD4+ and CD8+ T cell activation. Moreover, IFN-csecreting cells (representing the Th1 response) after treatment with mixed CD8+ T cell-epitope peptides [31] were not induced in G4 and G5 (Fig. 7D) . These data showed that the RNA adjuvant enhanced an antigen-specific Th1 response via the TLR7 signaling pathway and CD4+/CD8+ T cell activation.

In this study, we demonstrated the adjuvant effect of CrPV-IGR IRES-derived ssRNA [23] using protein-based vaccines. The RNA adjuvant induced the expression of genes related to the innate immune response (IFN and TLR-Myd88), T cell activation, and leukocyte chemotaxis similar to poly(I:C) in hPBMCs; however, the induction levels were lower than those by poly(I:C), and most genes returned to normal levels within 24 h (Fig. 1 ). This transient induction of genes by the RNA adjuvant may be evidence of vaccine safety. Similarly, a previous report demonstrated the rapid disappearance of the RNA platform within 24 h after injection [23] . We did not observe any adverse effects, such as autoimmune symptoms, changes in liver function, or organ damage, in the RNA adjuvant-immunized mice, even at a higher dose (200 mg/-mouse) (data not shown). Although poly(I:C) can be a superior adjuvant, safety concerns, such as an excessive immune response, represents a significant drawback [15] . However, our RNA adjuvant appeared to overcome this safety issue without decreasing the immune-stimulation effect.

Among the key factors induced in RNA adjuvant-treated hPBMCs, intercellular adhesion molecule-1, IL-6, IL-15, CD80, and STAT1/3 were induced (Fig. 1) . In particular, IL-6, which is highly expressed in CD4+ T cells, is partially responsible for Th2 cytokine production in Th1-differentiated cells via protein kinase C theta and NF-jB signaling pathways [32] . Additionally, IL-15 stimulates T and natural killer cell immunity and promotes the generation of long-lived memory T cells [33] . Furthermore, STAT1 functions as a central regulator of innate and adaptive immunity, and STAT3 is essential for DC, Tfh cell, and memory CD8+ T cell development [34, 35] . Given that the RNA adjuvant activated the Th1 response and induced long-term immunity (Figs. 4-6) , induction of these genes might have played an important role in this immune response. Fig. 6 . RNA adjuvant-HPV-formulated VLP vaccine induces antigen-specific long-term immunity. BALB/c mice were intramuscularly immunized at 2-week intervals with three doses of 10-value VLP-HPV-L1 with alum and with/without protamine-formulated RNA adjuvant-HPV. (A) Overall study design. (B) HPV-L1-specific IgG1 and IgG2a levels were measured by ELISA from 2-weeks to 27-weeks after the first immunization. Data represent the mean ± standard deviation (n = 5 mice). (C) HPV-L1 NAb levels against different types of VLPs in the immunized mice were determined by pseudovirus PRNT at 27 weeks after the first immunization. Serum from the same group (n = 5) was pooled. (D) Splenocytes were harvested at 27 weeks after the first immunization and re-stimulated for 2 days with/without type 16 or 18 HPV-L1 protein or 10 types of VLP-HPV-L1. The frequency of HPV-specific IFN-c-producing cells was determined by ELISPOT. Data represent the mean ± standard deviation (n = 3) (triplicate results acquired from one pooled sample per group). The data were statistically analyzed by ANOVA for (B) and (D). The significance of differences between groups is indicated by different letters of the alphabet. p < 0.05. ELISPOT: enzyme-linked immunospot, IFN: interferon, VLP: virus-like particle, HPV: human papillomavirus, ELISA: enzyme-linked immunosorbent assay, PRNT: plaque-reduction neutralization test, NAb: neutralizing antibody.

Interestingly, we found that the expression of genes associated with the innate immune response, especially type 1 IFN-related genes, was similar between the RNA adjuvant-treated hPBMCs and PBMCs from a live-attenuated YF17D-immunized patient (Fig. 2) . Moreover, it is also important that IRF7 expression was induced in RNA adjuvant-treated hPBMCs similar to that observed following immunization with YF17D because IRF7 is an essential molecule for antigen-specific IFN production promoting T and B cell activation [36] . These data assumed that the RNA adjuvant may make inactivated and protein-based vaccines function like a live attenuated vaccine. However, further studies are required to confirm this hypothesis.

Furthermore, Tlr7-KO mice and CD4+/CD8+ T cell-depleted mice did not show any RNA adjuvant effects (Fig. 7) , consistent with previous results [22, 37, 38] . These data indicated that this RNA adjuvant might be efficacious for inducing a broad immune response mediated through TLR7 signaling and T cell activation. However, the PRR for the RNA adjuvant cannot be confined to TLR7 alone, as the RNA adjuvant could induce RIG-I-like receptor pathwayrelated molecules similar to R848 and poly I:C ( Supplementary  Fig. 4 ). This indicated that the RNA adjuvant is recognized by RIG-I-like receptor in addition to TLR7 to activate an immune response. These data are accordance with a previous study, which showed that RNA based adjuvants triggered TLR-and RIG-I-like helicase-dependent effect [22] . Therefore, the boosting with RNA adjuvant may partially induce an IgG1 (Th2) response in Tlr7-KO mice through the RIG-I-like pathway. However, we do not know why this pathway cannot induce IgG2c (Th1). Further studies are required to understand the detailed underlying mechanism.

In addition, although the RNA adjuvant mediated its effects through the TLR7 signaling pathway, we did not use TLR agonists, such as imiquimod (TLR7 agonist) and resiquimod (TLR7/8 agonist), which are chemical adjuvants [39] , as a controls in this study, as this study focused on RNA-based adjuvants, such as poly I:C, but not chemical agonist-adjuvants. Poly I:C is a dsRNA that shows strong adjuvanticity in inactivated and protein-based vaccines [12] [13] [14] ; therefore, it was used as a control in this study.

Upon immunization with the MERS S protein vaccine together with alum and the RNA adjuvant, MERS S-specific NAb production increased synergistically (Fig. 4B) . Our data suggest that T cell activation of G5 may also show a synergistic increase, however, more data are needed to confirm this (Fig. 4D) . Additionally, given the strong immune-stimulating effect of the RNA adjuvant, it was unsurprising that the RNA adjuvant-HPV group was able to maintain HPV-specific NAb levels and the Th1 response, even at 27weeks after first immunization (Fig. 6) . These results indicated that the RNA adjuvant triggered long-term immunity, including humoral and cellular immune responses. This is an important fea- Fig. 7 . The RNA adjuvant induces a Th1 response through TLR7 and CD4+/CD8+ T cells. Tlr7-KO mice and C57BL/6 (WT) mice were intramuscularly immunized at 2-week intervals with two doses of 1 mg MERS S soluble protein vaccine, 20 mg formulated RNA adjuvant or alum. (A) MERS S protein-specific IgG1 and IgG2c levels were measured by ELISA at 2 weeks after the first and second immunizations. Data represent the mean ± standard deviation (n = 3-5 mice). (B) Overall study design and mouse schedule for immunization and anti-CD4 or CD8 treatment. (C) MERS S protein-specific IgG1 and IgG2c levels were measured by ELISA at 2 weeks after the second immunization. Data represent the mean ± standard deviation (n = 3-5 mice). (D) The frequency of MERS S protein-specific IFN-c-producing cells was determined by ELISPOT after treatment with a CD8+ T cell-epitope peptide mixture (7 peptides for C57BL/6, as described previously [30] ) or MERS S protein. Data represent the mean ± standard deviation (n = 3) (triplicate results acquired from one pooled sample per group). The data were statistically analyzed by ANOVA for (A), (C), and (D). The significance of differences between groups is indicated by different letters of the alphabet. p < 0.05. Th: T helper cell, ELISPOT: enzyme-linked immunospot, S: spike, IFN: interferon, ELISA: enzyme-linked immunosorbent assay, PRNT: plaque-reduction neutralization test, MERS: Middle East Respiratory Virus, TLR: Toll-like receptor, KO: knockout, WT: wild-type.

ture, given that HPV should have a long-term protective effect after vaccination.

Here, we used one RNA molecule derived from the CrPV-IGR IRES in the absence of target genes as the adjuvant and demonstrated its adjuvanticity for the MERS S protein vaccine. The RNA adjuvant-containing target gene HPV-L1 also showed an immune-stimulating effect. Therefore, although we previously showed that an RNA platform could express a coding gene [23] , RNA adjuvant including target gene is not critical to ensure adjuvanticity. However, should this RNA platform be used as an RNA vaccine, it needs to encode target genes. Further, improvements in the delivery strategy to increase the expression efficiency in the inoculated animal muscle are warranted. Interestingly, we compared the adjuvanticity of CrPV-IGR IRES-derived RNA with those of other RNA virus-derived adjuvants, including coxsackievirus B3 (CVB3) and encephalomyocarditis virus (EMCV) IRESsderived RNAs. CrPV and CVB3 IRES-derived RNAs showed higher adjuvant effects compared to that of the EMCV IRES-derived RNA (data not shown). Studies are underway on the effect of the viral IRES type on the immune effect.

Use of the RNA adjuvant as an immune stimulator, regardless of the presence of antigen-encoding genes, has many advantages, including its applicability for protein-based vaccines, simple manufacturing procedures, and easy standardization. Moreover, it can potentially overcome challenges associated with protein-subunit vaccines, such as weak T cell activation. The associated immune responses consequently promote high NAb levels and balanced Th1 and Th2 responses, resulting in long-term immunity. Therefore, since the RNA adjuvant exhibited immune-enhancing capacity, it might be valuable not only as a first-line vaccine in an epidemic emergency but also to improve protein-based vaccine efficacy.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Vaccine production, distribution, access, and uptake

Vaccine adjuvants: why and how

Live attenuated vaccines: historical successes and current challenges

Correlates of adjuvanticity: a review on adjuvants in licensed vaccines

Mechanisms of action of adjuvants

Beyond antigens and adjuvants: formulating future vaccines

Towards an understanding of the adjuvant action of aluminium

Old and new adjuvants

Advances in aluminum hydroxide-based adjuvant research and its mechanism

Use of defined TLR ligands as adjuvants within human vaccines

Type I interferons as vaccine adjuvants against infectious diseases and cancer

Enhanced resistance of mice to virulent Japanese B encephalitis virus following inactivated vaccine and poly(I:C)

Vaccine adjuvant uses of poly-IC and derivatives

Synthetic double-stranded RNA induces innate immune responses similar to a live viral vaccine in humans

Particulate formulations for the delivery of poly(I:C) as vaccine adjuvant

Type I interferon in the pathogenesis of lupus

Intranasal Poly-IC treatment exacerbates tuberculosis in mice through the pulmonary recruitment of a pathogen-permissive monocyte/macrophage population

TLR-based immune adjuvants

A Sendai virus-derived RNA agonist of RIG-I as a virus vaccine adjuvant

Nucleoside-modified mRNA vaccines induce potent T follicular helper and germinal center B cell responses

Messenger RNA-based vaccines with dual activity induce balanced TLR-7 dependent adaptive immune responses and provide antitumor activity

A new RNA-based adjuvant enhances virus-specific vaccine responses by locally triggering TLR-and RLH-dependent effects

Development of an RNA expression platform controlled by viral internal ribosome entry sites

Measurement of neutralizing serum antibodies of patients vaccinated with human papillomavirus L1 or L2-based immunogens using furin-cleaved HPV Pseudovirions

Live attenuated yellow fever 17D infects human DCs and allows for presentation of endogenous and recombinant T cell epitopes

Yellow fever vaccine induces integrated multilineage and polyfunctional immune responses

Systems biology approach predicts immunogenicity of the yellow fever vaccine in humans

Generation of a transgenic mouse model of Middle East respiratory syndrome coronavirus infection and disease

Peptides used in the delivery of small noncoding RNA

RIG-I and other RNA sensors in antiviral immunity

Rapid generation of a mouse model for Middle East respiratory syndrome

Elevated IL-6 expression in CD4 T cells via PKCtheta and NF-kappaB induces Th2 cytokine production

IL15 and T-cell stemness in T-cell-based cancer immunotherapy

IFNs and STATs in innate immunity to microorganisms

STAT3 signaling in immunity

A cGAS-independent STING/IRF7 pathway mediates the immunogenicity of DNA vaccines

Both CD4+ and CD8+ T cells can mediate vaccineinduced protection against Coccidioides immitis infection in mice

Transient CD4+ T cell depletion results in delayed development of functional vaccine-elicited antibody responses

The use of TLR7 and TLR8 ligands for the enhancement of cancer immunotherapy

The authors thank Drs Jung Joo Hong and Hanseul Oh for technical assistance, Hae-Yung Ahn for assistance with animal care, and Dr. Jin-Seul Kim for assistance with HPV vaccine. 

There are no conflicts of interests to declare.

Supplementary data to this article can be found online at https://doi.org/10.1016/j.vaccine.2019.07.070.