Synthetic Abortive HIV-1 RNAs Induce Potent Antiviral Immunity

Strong innate and adaptive immune responses are paramount in combating viral infections. Dendritic cells (DCs) detect viral infections via cytosolic RIG-I like receptors (RLRs) RIG-I and MDA5 leading to MAVS-induced immunity. The DEAD-box RNA helicase DDX3 senses abortive human immunodeficiency virus 1 (HIV-1) transcripts and induces MAVS-dependent type I interferon (IFN) responses, suggesting that abortive HIV-1 RNA transcripts induce antiviral immunity. Little is known about the induction of antiviral immunity by DDX3-ligand abortive HIV-1 RNA. Here we synthesized a 58 nucleotide-long capped RNA (HIV-1 Cap-RNA58) that mimics abortive HIV-1 RNA transcripts. HIV-1 Cap-RNA58 induced potent type I IFN responses in monocyte-derived DCs, monocytes, macrophages and primary CD1c+ DCs. Compared with RLR agonist poly-I:C, HIV-1 Cap-RNA58 induced comparable levels of type I IFN responses, identifying HIV-1 Cap-RNA58 as a potent trigger of antiviral immunity. In monocyte-derived DCs, HIV-1 Cap-RNA58 activated the transcription factors IRF3 and NF-κB. Moreover, HIV-1 Cap-RNA58 induced DC maturation and the expression of pro-inflammatory cytokines. HIV-1 Cap-RNA58-stimulated DCs induced proliferation of CD4+ and CD8+ T cells and differentiated naïve T helper (TH) cells toward a TH2 phenotype. Importantly, treatment of DCs with HIV-1 Cap-RNA58 resulted in an efficient antiviral innate immune response that reduced ongoing HIV-1 replication in DCs. Our data strongly suggest that HIV-1 Cap-RNA58 induces potent innate and adaptive immune responses, making it an interesting addition in vaccine design strategies.


INTRODUCTION
Evoking potent and tailored antiviral responses by the host is paramount in combating viral infections (1). Dendritic cells (DCs) induce antiviral immune responses by recognizing invading viruses via pattern recognition receptors (PRRs). PRR triggering by viral pathogen-associated molecular patterns (PAMPs) induces DC maturation and activation as well as differentiation of naïve T cells (2)(3)(4). Certain PRRs such as the RIG-I-like receptors (RLRs) induce strong antiviral innate immune responses initiated by expression of type I interferon (IFN) responses. RLRs are cytosolic PRRs and sense virus infection by detection of specific viral RNA structures (5,6). RIG-I (DDX58) and MDA5 are two well-described RLRs and important in antiviral immunity to e.g., Influenza viruses, Dengue virus, and West Nile virus (7)(8)(9). RIG-I recognizes both uncapped 5 ′ ppp single stranded (ss) RNA and short double stranded (ds) RNA, whereas MDA5 senses long dsRNA (7,10,11). Upon activation, RIG-I and MDA5 engage with mitochondrial antiviral protein MAVS, leading to MAVS multimerization and subsequent recruitment of adaptor molecule TRAF3. TRAF3 mediates recruitment of serine/threonine-protein TANK-binding kinase 1 (TBK1) and IκB kinase ε (IKKε), resulting in activation of transcription factors IRF3 and NF-κB, leading to type I IFN and cytokine transcription (12,13). Autocrine and paracrine ligation of IFNβ to the heterodimeric transmembrane IFN receptor (IFNAR) on the cell surface ultimately results in transcription of a broad spectrum of interferon-stimulated genes (ISGs), of which many exhibit strong antiviral activity (14)(15)(16)(17)(18). NF-κB activation regulates transcriptional activation of a plethora of cytokine genes, including IL-1β and TNF that are important in innate immunity (19). In addition, IFNβ is crucial in driving IL-27 synthesis and thereby CD8 + T cell-dependent adaptive immune responses (20). Moreover, both IRF3 and NF-κB are involved in the induction of cytokines important in T helper (T H ) 1 differentiation (21)(22)(23).
Recently, the cytosolic DEAD-box RNA helicase 3 (DDX3) that resembles cytosolic DEAD box helicase RIG-I and MDA5, was shown to function as a PRR for human immunodeficiency virus 1 (HIV-1) (24)(25)(26)(27). DDX3 acts as a host factor for various viruses including hepatitis B and C virus and West Nile virus, facilitating virus replication (8,28,29). Thus, DDX3 is an important host factor for viruses and its function as a PRR might prevent escape of viruses from DDX3.
DDX3 is a well-known host factor required for HIV-1 propagation due to its role in transport of viral Tat mRNA and subsequent formation of translation initiation complexes (30)(31)(32). Upon initiation of HIV-1 infection, deficient transcription elongation leads to formation of prematurely aborted RNAs (33,34). Interestingly, DDX3 senses these abortive HIV-1 RNA transcripts, leading to MAVS-dependent type I IFN responses, indicating that DDX3 is a viral PRR for HIV-1 (27).
Abortive HIV-1 RNAs are generated during the early steps of HIV-1 transcription, consisting of an HIV-1-specific complex secondary RNA hairpin-like structure (TAR loop) and a 5 ′ cap, but lacking a poly A tail (33,34). The complex structure of the TAR loop, together with the 5 ′ cap is required for binding to DDX3, while the absence of the poly A tail prevents engagement of DDX3 with the cellular translational machinery (30). Gringhuis (27,35). Due to this viral inhibition mechanism, the breadth and potency of abortive HIV-1 RNAs in inducing antiviral immunity remains elusive.
To characterize the role of abortive HIV-1 RNA in establishing innate and adaptive immune responses without interference due to DC-SIGN inhibition, we developed a synthetic 5 ′ capped HIV-1 RNA of 58 nucleotides that mimics the naturally occurring abortive HIV-1 RNA (HIV-1 Cap-RNA 58 ). Our data strongly suggest that HIV-1 Cap-RNA 58 induces potent type I IFN responses in monocyte-derived DCs, macrophages as well as primary CD1c + DCs. Furthermore, HIV-1 Cap-RNA 58 is a potent stimulus that induces both innate and adaptive immune responses in monocyte-derived DCs. HIV-1 Cap-RNA 58 -dependent induction of DC maturation and cytokine secretion leads to T H 2 differentiation. Notably, HIV-1 Cap-RNA 58 responses inhibited ongoing HIV-1 infection. Our data further define the importance of sensing abortive HIV-1 transcripts to evoke strong antiviral immunity and provide a rationale for using HIV-1 Cap-RNA 58 in vaccine design strategies.

RNA Constructs
The synthetic abortive HIV-1 RNAs were designed based on the HIV-1 genome. HIV-1 Cap-RNA 58 consists of nucleotides 1-58 from the HIV-1 genome, including a 5 ′ m 7 GTP cap but lacking the poly A tail. HIV-1 Cap-RNA 630 consists of nucleotides 1-630 and also contains the 5 ′ m 7 GTP cap and is lacking the poly A tail. The 5 ′ m 7 GTP cap was incorporated using co-capping of 5 ′ m 7 GTP during in vitro transcription (IVT) (Biosynthesis, Table S1) as previously described (27). As a control RNA, 1-58 nucleotides were synthesized lacking the 5 ′ cap (HIV-1 control RNA 58 ). HIV-1 Cap-RNA 58 and HIV-1 control RNA 58 structures were predicted with the MFOLD program.

RNA Interference
RNA interference was performed using the Neon Transfection System according to manufacturer's protocol (Thermo Fisher). On day 4 of monocyte-derived DC cultures, cells were washed with PBS, resuspended in buffer R (Thermo Fisher) and divided according to the different short interfering (si) SMARTpool RNAs (all from Dharmacon). siDDX3 (M-006874-01), siMAVS (M-024237-02) or siNon-Target as a control (D-001206-13) were added to DC-buffer R mixtures and transfection of DCs with the siRNAs was achieved by subjecting them to 1,500V for 20 ms. Transfected cells were seeded in 24-wells plates in RPMI 1640 with 10% FCS (Invitrogen) and 2 mM L-glutamine (Lonza), without antibiotics. After 48 h, viable cells were harvested, washed and seeded in a 96-wells round bottom plate and incubated overnight at 37 • C, 5% CO 2 . Seventy-two hours after transfection, silencing of expression of target proteins in DCs was confirmed by quantitative real-time PCR and flow cytometry ( Figure S1) and cells were stimulated as previously described.
Quantitative Real-Time PCR mRNA was extracted using an mRNA capture kit (Roche) and was reverse transcribed to cDNA using a reverse transcriptase kit (Promega). Quantitative real-time PCR was performed on an ABI 7500 Fast Real-Time PCR detection system (Applied Biosystems) using SYBR Green (Thermo Fisher), with primers that were designed using Primer Express 2.0 (Applied Biosystems, Table S2). Expression of genes of interest were normalized to expression of household gene GAPDH, according to the formula Nt = 2 Ct(GAPDH)−Ct(target) . For each donor, expression levels induced upon stimulation with HIV-1 Cap-RNA 58 were set as 1.

p65 and IRF3 Translocation
DCs were stimulated for 4 h, fixed with 4% pFA and permeabilized with 0.2% Triton X-100 in PBS. p65 was stained with anti-p65 (1:50, 8242S, Cell Signaling) and IRF3 with anti-IRF3 (1:50, 4302S, Cell Signaling), followed by a secondary donkey anti-rabbit labeled with Alexa-546 (1:400, Invitrogen) and 1 µg/mL Hoechst (Invitrogen) and cellular localization was visualized with a 100x magnification, using Leica DM6 B upright microscope. Analysis was performed with LAS X Navigator software. Nuclear extracts (NE) were prepared 4 h after DC stimulation, using the NucBuster protein extraction kit (Novagen). Twenty micrograms of NE per sample was used to detect nuclear p65 or IRF3 using the TransAM NF-κB-p65 and IRF3 kits (Active Motif). OD450 nm values were measured using BioTek Synergy HT.
Elisa DC supernatants were harvest 24 or 48 h after stimulation and secretion of TNF, IL-6, and IL-12p70 protein (eBiosciences) was measured by ELISA as described by manufacturer. OD450 nm values were measured using BioTek Synergy HT.

Virus and Infection
DCs were infected with R5 HIV-1 strain NL4.3 BaL. NL4.3 BaL was produced as described previously (34,37). For DC infection, a multiplicity of infection (MOI) of 0.1-0.2 was used, depending on the virus batch. DCs were infected for 24 h, washed extensively and left in the presence of HIV-1 control RNA or HIV-1 Cap-RNA 58 for 5 days, after which intracellular p24 levels were measured using flow cytometry to determine infection.

Statistical Analysis
Statistics were performed using Student's t-test for paired (BAY 11-7082 inhibitor, IL-6 and TNF ELISAs, proliferation assay) and unpaired observations (all other experiments) using GraphPad version 8. Statistical significance was set at P < 0.05.

HIV-1 Cap-RNA 58 Induces Type I IFN Responses in Various Myeloid Cells
We used a synthetic RNA that mimics abortive HIV-1 RNAs: this synthetic HIV-1 Cap-RNA 58 consists of the first 58 nucleotides common to all HIV-1 transcripts and contains a 5 ′ cap while lacking a poly A tail (Table S1, Figure 1A). We investigated whether HIV-1 Cap-RNA 58 induced type I IFN responses in monocyte-derived DCs (DCs) by treating DCs with HIV-1 Cap-RNA 58 complexed with transfection reagent lyovec (vehicle control) to facilitate cytoplasmic delivery (27). HIV-1 Cap-RNA 58 induced strong IFNB transcription in DCs after 10 h of stimulation ( Figure 1B). Expression levels of interferonstimulated gene (ISG) Myxovirus resistance protein 1 (MxA) transcripts were significantly induced by HIV-1 Cap-RNA 58 in DCs as well when compared to untreated DCs ( Figure 1B). HIV-1 Cap-RNA 58 -induced type I IFN levels were compared to those observed with RLR agonist poly-I:C. After equalizing the amount of molecules of both HIV-1 Cap-RNA 58 and poly-I:C (Figure S1), HIV-1 Cap-RNA 58 -induced type I IFN were significantly higher than poly-I:C-induced levels after 10 h of stimulation ( Figure 1B). We next investigated whether HIV-1 Cap-RNA 58 activated responses in primary myeloid cells. Notably, treatment of CD14 + monocytes and monocytederived macrophages with HIV-1 Cap-RNA 58 induced type I IFN responses 10 h after stimulation, whereas primary blood CD1c + DCs showed type I IFN responses 8 h after stimulation. HIV-1 Cap-RNA 58 -induced type I IFN levels were higher than or comparable to poly-I:C-induced type I IFN responses ( Figure 1C). To assess the specificity of HIV-1 Cap-RNA 58 for DDX3 and MAVS-dependent signaling, both DDX3 and MAVS were silenced in DCs by RNA interference (RNAi) (Figure S2). Silencing of either DDX3 or MAVS expression completely abrogated IFNB induced by HIV-1 Cap-RNA 58 , 2 h after stimulation ( Figure 1D). Both vehicle control and HIV-1 control RNA did not induce IFNB expression (Figures 1A,D). These data indicate that HIV-1 Cap-RNA 58 triggers type I IFN in a variety of myeloid cells via DDX3 and MAVS, and that the 5 ′ cap is required for sensing by DDX3.

HIV-1 Cap-RNA 58 Induces DC Maturation
To determine whether HIV-1 Cap-RNA 58 induces adaptive immunity, we first examined the expression levels of costimulatory molecules CD80, CD83, and CD86 after stimulation with HIV-1 Cap-RNA 58 , by flow cytometry. HIV-1 Cap-RNA 58 induced expression of CD80, CD83, and CD86 compared to unstimulated DCs, albeit to a lesser extent than LPS (Figure 2A). HIV-1 Cap-RNA 58 induced expression levels of activation marker HLA-DR to a similar extent as LPS, whereas the expression levels of costimulatory molecule CD40 remained unaffected in contrast to LPS stimulation (Figure 2A). To investigate the role of IFNβ signaling in DC maturation, we neutralized IFNβ signaling by treatment with blocking IFNα/βR antibodies. CD86 induction by HIV-1 Cap-RNA 58 was partially but significantly blocked by blocking IFNα/βR antibodies (Figures 2B,C), suggesting that type I IFN induction increases CD86. Besides type I IFN-dependent CD86 expression, expression of costimulatory molecules can also be induced by the transcription factor NF-κB (38,39). To test whether CD86 expression is also NF-κB dependent, we blocked NF-κB activation using BAY 11-7082, a small molecule inhibitor for IκBα, as release of the inhibitory protein IκBα from the NF-κB dimer within the cytoplasm is mandatory for NF-κB activation (40). BAY 11-7082 treatment significantly reduced HIV-1 Cap-RNA 58 -induced CD86 expression (Figure 2C). Our results imply that HIV-1 Cap-RNA 58 treatment leads to IFNβ-and NF-κB-dependent DC maturation.
HIV-1 Cap-RNA 58 Activates IRF3 and NF-κB p65 We next investigated whether HIV-1 Cap-RNA 58 activated transcription factors IRF3 and NF-κB, known to be involved in transcriptional regulation of a plethora of cytokine and other genes required for the orchestration of innate and adaptive immune responses by DCs (14,15,19). DCs were treated with HIV-1 Cap-RNA 58 or LPS and translocation of p65, one of the primary active subunits within the dimeric NF-κB transcription factor family, was analyzed by immunofluorescence microscopy. Similarly, as was observed for IRF3, HIV-1 Cap-RNA 58 induced p65 translocation to the nucleus, albeit to a lesser extent than observed for LPS-treated DCs (Figures 3A,B). We next quantified nuclear translocation of p65 and IRF3 using a transcription factor binding assay using nuclear extracts from HIV-1 Cap-RNA 58 -activated DCs. HIV-1 Cap-RNA 58 induced significant translocation of the p65 unit, 4 h after stimulation, in contrast to HIV-1 control RNA (Figure 3C). Similarly, we quantified the nuclear translocation of IRF3, which was also detected in HIV-1 Cap-RNA 58 -activated but not HIV-1 control RNA-treated DCs ( Figure 3D). Thus, these data strongly indicate that HIV-1 Cap-RNA 58 activates both IRF3 and NF-κB, implying that it can play a significant role in establishing innate and adaptive immune responses.

HIV-1 Cap-RNA 58 Induces Expression of Pro-Inflammatory Cytokines
We next assessed whether treatment of HIV-1 Cap-RNA 58 triggered cytokine expression in DCs. HIV-1 Cap-RNA 58 significantly induced expression of the pro-inflammatory cytokines IL-6 and TNF at both mRNA and protein level, compared to vehicle control and HIV-1 control RNA-treated DCs (Figures 4A-C). Interestingly, HIV-1 Cap-RNA 58 treatment of DCs did not induce mRNA expression of IL1B, IL8, IL10, and IL23A (data not shown). mRNA expression levels of IL6 and TNF peaked at 10 h, while the ISG IL27A, which encodes IL-27p28, a subunit of IL-27, peaked at 8 h (Figure 4A). At peak level, the  IL-12p70 protein could be detected in the supernatant of DCs after HIV-1 Cap-RNA 58 treatment ( Figure 4C). Thus, HIV-1 Cap-RNA 58 induces a specific cytokine program primarily directed at pro-inflammatory conditions.

Different Viral HIV-1 RNAs Induce Similar Levels of Type I IFN and Pro-Inflammatory Cytokines
To assess whether the ability of HIV-1 Cap-RNA 58 to induce type I IFN and cytokine responses is due to the short RNA construct length, we examined the ability of a longer viral HIV-1 RNA in inducing immune activation. DCs treated with a 5 ′ capped 630 nucleotides long RNA corresponding to the Tat transcript, which includes the same 1-58 sequence as the HIV-1 Cap-RNA 58 at its start, but lacking a poly A tail (HIV-1 Cap-RNA 630 ) induced antiviral responses comparable to those induced by HIV-1 Cap-RNA 58 . Similar to HIV-1 Cap-RNA 58 , HIV-1 Cap-RNA 630 induced IFNB 2 h after stimulation reaching a peak level after 10 h (Figure 5A). Similarly, HIV-1 Cap-RNA 630 induced MXA and also another ISG A3G (encoding for APOBEC3G protein)  (Figures 4A,B). At 6,  8 and 10 h after stimulation with the two HIV-1 Cap-RNAs, we observed that both constructs induced similar levels of IL6, TNF, IL12A, and IL27A mRNA (Figure 5D). Although not significant, stimulation with HIV-1 Cap-RNA 630 showed a trend toward increased IL12A expression compared to HIV-1 Cap-RNA 58 after 6, 8, and 10 h of stimulation, and to a trend of increased IL27A expression after 10 h (Figure 5D). These data suggest that the length of viral RNA constructs does not affect the strength of the type I IFN and pro-inflammatory cytokine responses as long as it contains a 5 ′ cap and the first 58 nucleotides of the HIV-1 genome that form the TAR loop. Thus, synthetic HIV-1 RNAs induce antiviral innate immune responses independent of length.

HIV-1 Cap-RNA 58 -Activated DCs Induce T Cell Proliferation and Differentiation
We next investigated the ability of HIV-1 Cap-RNA 58 -treated DCs to activate T cells. First we analyzed proliferation of CellTrace Violet-labeled peripheral blood cells (PBLs) induced by coculture for 5 days with DCs that were treated with vehicle control, HIV-1 control RNA, HIV-1 Cap-RNA 58 or LPS. Flow cytometry analysis showed that HIV-1 Cap-RNA 58 -activated DCs enhanced proliferation of both CD4 + and CD8 + T cells compared to untreated DCs and DCs treated with vehicle control or HIV-1 control RNA (Figures 6A,B and Figure S3). We next analyzed T H 1 and T H 2 differentiation after DCnaïve CD4 + T cell cocultures by intracellular IFNγ and IL-4 expression, respectively. HIV-1 Cap-RNA 58 -activated DCs showed significant skewing toward T H 2 differentiation compared to the DCs treated with vehicle control or HIV-1 control RNA (Figures 6C,D). As expected, positive controls LPS and PGE 2 or LPS and IFNγ induced T H 2 and T H 1 responses, respectively, whereas LPS gave a mixed T H 1/T H 2 response (Figures 6C,D). Thus, our data show that HIV-1 Cap-RNA 58 -treated DCs induce CD4 + and CD8 + T cell activation and skew adaptive immune response toward a T H 2 phenotype.

HIV-1 Cap-RNA 58 Inhibits HIV-1 Infection of DCs
To investigate whether treatment with HIV-1 Cap-RNA 58 would block HIV-1 replication in DCs after infection, DCs were infected by R5-tropic HIV-1 NL4.3 BaL. After 24 h of ongoing HIV-1 infection the DCs were treated with HIV-1 Cap-RNA 58 or LPS After 5 days, infection levels were assessed by measuring intracellular p24 + cells by flow cytometry. Although infection levels differed per donor, HIV-1 Cap-RNA 58 decreased the percentage of HIV-1 p24 + DCs compared to HIV-1 control RNA-treated DCs in four different donors, albeit to a lesser extent than LPS (Figures 7A,B). These results imply that HIV-1 Cap-RNA 58 induces a functional antiviral response that limits HIV-1 infection.

DISCUSSION
RNA helicase DDX3 is important for the transport of HIV-1 Tat mRNA as well as the formation of translation initiation complexes required for HIV-1 translation (30)(31)(32). Besides its role as a host factor for HIV-1, DDX3 also functions as a viral sensor (27). Here we investigated the potency and breadth of DDX3 ligand HIV-1 Cap-RNA 58 , a synthetic mimic of the naturally occurring abortive HIV-1 RNAs. We observed that HIV-1 Cap-RNA 58 resulted in type I IFN responses in various myeloid cells in a DDX3-and MAVS-dependent manner. Our data further showed that HIV-1 Cap-RNA 58 inhibited ongoing HIV-1 replication in DCs, most likely via the induced innate antiviral type I IFN responses. Moreover, HIV-1 Cap-RNA 58 induced DC maturation and cytokine responses that led to adaptive T cell activation as well as differentiation. Thus, our data suggest that DDX3 is a PRR and shows that synthetic abortive HIV-1 RNA 58 is immunostimulatory.
The potency of abortive HIV-1 RNA to exert an antiviral role in response to viral sensing by its natural ligand DDX3 is not well-understood. Therefore, we aimed to assess the role of abortive HIV-1 RNA in induction of innate and adaptive immune responses by using a synthetic mimic of abortive HIV-1 RNA (HIV-1 Cap-RNA 58 ). Our data strongly suggest that synthetic HIV-1 Cap-RNA 58 encapsulated in lyovec triggered DDX3 and MAVS-mediated IFNβ responses. It has been described that DDX3 has been found in complexes with RIG-I and MDA5 and might therefore induce IFNβ responses via a RIG-I or MDA5-MAVS-dependent way (41). We have previously shown in 293T cells treated with CRISPR-cas9 that depletion of RIG-I and MDA5 did not affect HIV-1 Cap-RNA 58 -induced type I IFN responses, indicating that the HIV-1 Cap-RNA 58 -induced type I IFN responses described here are generated in a DDX3-MAVS-TBK1-IRF3-dependent manner (27). Previous studies have shown that knockdown of DDX3 or the generation of phosphorylation-deficient DDX3 mutants in cell lines resulted in TBK1/IKKε-dependent decrease of the IFNB promotor activity, providing evidence that DDX3 is involved in the induction of IFNB transcription (24,25). In line with the previous reported data obtained in cell lines, we have shown that DDX3 induces type I IFN responses in monocyte-derived DCs as well as primary monocytes, monocyte-derived macrophages and primary human CD1c + DCs.
Synthetic abortive HIV-1 Cap-RNA 58 contains a 5 ′ cap and a secondary TAR loop structure and is a ligand for DDX3, which leads to the induction of type I IFN responses (27). The HIV-1 Cap-RNA 58 -induced type I IFN responses were similar if not stronger than those observed by poly-I:C, which triggers RIG-I and MDA5. RIG-I and MDA5 distinguish the recognition of their viral ligands based on RNA structure and length. RIG-I recognizes ssRNAs and short dsRNAs, whereas MDA5 recognizes longer dsRNAs (42). Soto-Rifo et al. have described that DDX3 recognizes viral RNA constructs due to the presence of a 5 ′ cap in close proximity to a complex secondary structure (30). Whether DDX3 is able to distinguish between different viral RNA lengths and adapts subsequent immune activation is unknown. Here we aimed to assess whether differences in HIV-1 Cap-RNA construct length would lead to varying immune activation levels. We observed that the length of the synthetic viral ligand did not affect the strength of the antiviral type I IFN responses as both HIV-1 Cap-RNA 58 and HIV-1 Cap-RNA 630  Besides type I IFN responses, HIV-1 Cap-RNA 58 resulted in increased upregulation of costimulatory molecules CD80, CD83, and CD86. Furthermore, HIV-1 Cap-RNA 58 enhanced expression of HLA-DR but not CD40. Our data further indicate that DC maturation is dependent on both type I IFN and NF-κB activation. It remains to be established whether the effect of NF-κB is mediated via IFNβ or that other cytokines activated by NF-κB further affect DC maturation in combination with IFNβ. It has previously been described for various cell lines (mean ± s.d.). *P < 0.05, ***P < 0.001, Student's t-test. NS, not significant. that DDX3 expression knockdown results in decreased NF-κB p65 phosphorylation and cytokine responses suggesting that DDX3 plays a stimulatory role in NF-κB signaling (43). In addition, Ku et al. (44) described that in THP-1-differentiated macrophages DDX3 is important for TNF, IL-1β, CCL2, and CCL5 expression as knockdown of DDX3 expression impaired cytokine and chemokine expression in response to LPS and poly-I:C stimulations. Besides affecting pro-inflammatory cytokine responses, DDX3 knockdown also led to impaired migration and phagocytic capacities of THP-1-differentiated macrophages (44). Although it is unclear whether these functions can be induced by viral ligands, these data potentially imply that DDX3 could be involved in orchestrating various important functions in DCs. Our data underscore the importance of DDX3 as a viral sensor important for the induction of antiviral immunity.
We observed that HIV-1 Cap-RNA 58 induced expression of pro-inflammatory cytokines IL-6 and TNF, which are important for both innate and adaptive immune responses. Furthermore, HIV-1 Cap-RNA 58 induced IL12A but not IL12B mRNA expression which resulted in the absence of heterodimeric IL-12p70 protein that is crucial for the induction of T H 1 differentiation. The observed lack of IL-12p70 upon stimulation with HIV-1 Cap-RNA 58 might explain skewing of T helper differentiation toward a T H 2 phenotype by HIV-1 Cap-RNA 58treated DCs. Furthermore, we observed HIV-1 Cap-RNA 58dependent induction of IL27A, encoding for one of the subunits of heterodimeric IL-27 protein, important in the induction of follicular T helper (T FH ) cells (45). Both T H 2 and T FH cells are important for the induction of antibody responses against invading pathogens including viruses (46). T FH cells are important for the formation and maintenance of germinal centers (GCs) and subsequent differentiation of B cells in GCs (46). Once a B cell exits the GC, T H 2-induced IL-4 production can direct class switching from immunoglobulin G (IgG) to IgE antibodies (47,48). Recent studies show that T FH responses are required to induce broadly neutralizing antibodies against HIV-1 (49)(50)(51). Although it is unclear yet whether HIV-1 Cap-RNA 58 induces a T FH phenotype, the cytokine responses and T H 2 differentiation suggest that HIV-1 Cap-RNA 58 can be useful in vaccines to induce neutralizing antibodies against HIV-1 or other viruses.
Microbial LPS is increased in serum of HIV-1 infected individuals due to intestinal damage upon CD4 + T cell depletion. LPS as a potent immunostimulatory compound could be involved in inflammatory responses during chronic phase of infection (52,53). Interestingly, several studies suggest that HIV-1 replication in latent infected cells produces short abortive RNAs such as the DDX3 ligand HIV-1 Cap-RNA 58 (54,55). Our study suggest that these abortive RNAs can induce inflammatory responses. Thus, besides increased translocation of microbial LPS, the production of HIV-1 Cap-RNA 58 in latent infected cells can also contribute to immune activation observed in HIV-1 infected individuals during chronic phase of infection.
In conclusion, DDX3 is a highly versatile protein involved in a multitude of cellular processes. During HIV-1 infection, the dual role of DDX3 in exerting both proviral and antiviral capacities provides insight in its complexity and to the various roles DDX3 might play in establishing immunity. Here we have identified the antiviral role of DDX3 upon sensing of a viral-derived RNA and how its ligands can be used as adjuvants. Our data strongly indicate that HIV-1 Cap-RNA 58 induces potent antiviral innate and adaptive immune responses in human DCs or directed by human DCs. Our data shows that DDX3 is a pattern recognition receptor and its synthetic ligands can be used as adjuvants to induce potent immune responses.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Amsterdam University Medical Centers, location AMC Medical Ethics Committee according to the Declaration of Helsinki. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
MS designed, performed, and interpreted most experiments and prepared the manuscript. JS helped with study design. JH performed HIV-1 infections and subsequent FACS analyses. TK and EZ-W performed nuclear extract isolations. SG helped with study design and interpretation of data. TG supervised all aspects of this study.

FUNDING
This work was supported by Aidsfonds (P-9906) and the European Research Council (Advanced grant 670424).