Research ArticleInnate Immunity

TLR7 and TLR8 activate distinct pathways in monocytes during RNA virus infection

See allHide authors and affiliations

Science Signaling  29 Oct 2019:
Vol. 12, Issue 605, eaaw1347
DOI: 10.1126/scisignal.aaw1347

TLRing the antiviral response

Monocytes are critical orchestrators of immunity that both sense pathogens and produce cytokines. When comparing a range of RNA viruses, de Marcken et al. found that the human monocyte response to infection was virus specific and influenced by Toll-like receptor 7 (TLR7) and TLR8, which sense single-stranded RNA. Whereas TLR8 signaling stimulated monocytes to express type I interferon and cytokines involved in CD4+ T helper 1 (TH1) cell differentiation, TLR7 signaling promoted production of cytokines involved in TH17 cell differentiation. Only TLR7 activated monocyte Ca2+ flux, which suppressed antiviral interferon production. Thus, these receptors activate distinct pathways in monocytes, which can tailor the immune response to viral infection.

Abstract

Human blood CD14+ monocytes are bone marrow–derived white blood cells that sense and respond to pathogens. Although innate immune activation by RNA viruses preferentially occurs through intracellular RIG-I–like receptors, other nucleic acid recognition receptors, such as Toll-like receptors (TLRs), play a role in finely programming the final outcome of virus infection. Here, we dissected how human monocytes respond to infection with either Coxsackie (CV), encephalomyocarditis (EMCV), influenza A (IAV), measles (MV), Sendai (SV), or vesicular stomatitis (VSV) virus. We found that in monocytes, type I interferon (IFN) and cytokine responses to infection were RNA virus specific and differentially involved TLR7 and TLR8, which sense single-stranded RNA. These TLRs activated distinct signaling cascades in monocytes, which correlated with differences in the production of cytokines involved in the polarization of CD4+ T helper cells. Furthermore, we found that TLR7 signaling specifically increased expression of the transcription factor FOSL1, which reduced IL-27 and TNFα production by monocytes. TLR7, but not TLR8, activation of monocytes also stimulated Ca2+ flux that prevented type I IFN responses. Our work demonstrates that in human monocytes, TLR7 and TLR8 triggered different signaling pathways that contribute to distinct phenotypes during RNA virus infection. In addition, we defined individual targets within these pathways that promoted specific T helper and antiviral responses.

INTRODUCTION

Human monocytes are bone marrow–derived leukocytes from the innate immune system with important functions in pathogen sensing and innate immunity during bacterial, viral, and fungal infections. Monocytes are broadly classified on the basis of the expression of CD14 and CD16 into classical (CD14+), intermediate (CD14+CD16+), and nonclassical (CD14lowCD16+) monocytes (1, 2). However, single-cell transcriptomic data suggests that monocytes are likely heterogeneous (3). CD14+ classical monocytes are the most abundant population, representing 80 to 90% of circulating monocytes (4), and produce proinflammatory cytokines, phagocytose, and secrete reactive oxygen species after pathogen stimulation (2).

During the course of a viral infection, circulating monocytes rapidly leave the bloodstream and migrate to tissues where, after pathogen sensing and/or other stimuli, they differentiate into macrophages or dendritic cells (DCs) (4). This recruitment is essential for effective control, and ultimately, clearance of the infection. However, in addition to this function, monocytes can also exert direct antimicrobial responses and promote the adaptive immune response. Monocytes are equipped with pattern recognition and phagocytic receptors necessary for pathogen sensing and destruction, and they also produce proinflammatory cytokines, which may help initiate the adaptive immune response (4).

Although peripheral blood monocytes are a major population targeted by viruses during host infection (59), little is known about the early events that are triggered when RNA viruses interact with human monocytes and how these interactions modulate the phenotype and effector functions of the latter. Among the pattern recognition receptors expressed by monocytes, nucleic acid–sensing receptors are major components triggering cell activation during RNA virus infection (10). Retinoic acid–inducible gene I (RIG-I)–like receptors (RLRs), such as melanoma differentiation-associated protein 5 (MDA5), RIG-I, and DExH-box helicase 58 (DHX58 or LPG2), are cytoplasmic sensors of RNA (11), whereas Toll-like receptor 7 (TLR7) and TLR8 are intracellular sensors located in endosomes that recognize single-stranded RNA. Both types of receptors induce the expression of proinflammatory cytokines and type I interferon (IFN) response upon RNA virus sensing (11, 12). Monocytes from healthy individuals express RLRs (13) and both TLR7 (14) and TLR8 (15), and their expression can vary in patients with infectious diseases as compared to healthy individuals (16, 17).

In this study, we examined the signaling pathways downstream TLR7 and TLR8 that occur after monocyte infection with a panel of RNA viruses ex vivo and their capacity to drive specific cytokine production leading to the regulation of the adaptive immune response. We demonstrated that RNA virus infection induced a variety of phenotypes and distinct antiviral responses in human monocytes as a result of the simultaneous activation of various pattern recognition receptors. We focused on the role of TLR7 and TLR8 signaling, and we demonstrated that although both receptors recognize the same generic ligand, there were important differences between their activation. TLR7 preferentially promoted the expression of CD4+ T helper 17 (TH17) cell polarizing cytokines after virus infection, whereas TH1-type cytokine production and type I IFN response were dependent on TLR8 signaling. The differences in proinflammatory cytokine expression induced by TLR7 or TLR8 signaling were due to the TLR7-stimulated expression of Fos family member FOSL1, which inhibited type 1 cytokine expression. Furthermore, we uncovered a role for TLR7-dependent Ca2+ flux in inhibiting type I IFN responses. Thus, TLR7 and TLR8 activation triggers different signaling pathways in human monocytes that contribute to their distinct functions during RNA virus infection.

RESULTS

RNA virus infection of monocytes induces virus-specific cytokine production patterns

Circulating monocytes display distinct phenotypes in patients with different RNA virus infections, and although monocytes are a major population infected by viruses in peripheral blood (18), the early events that occur after interactions of viruses with monocytes are poorly understood. To examine this, ex vivo isolated CD14+ classical monocytes from healthy individuals were infected with several RNA viruses for 16 hours. The chosen viruses, i.e., Coxsackie virus (CV) (19, 20), encephalomyocarditis virus (EMCV) (21, 22), influenza A virus (IAV) (23, 24), measles virus (MV) (25), Sendai virus (SV) (26, 27), and vesicular stomatitis virus (VSV) (28, 29) are internalized by and actively infect monocytes in vivo and/or in vitro. As a readout, we examined two main monocyte functions: T cell polarization capacity and expression of type I IFN and IFN-stimulated genes (ISGs; Fig. 1), which constitute the antiviral response. For TH1-type cytokines, we assessed the expression of IL12B, IL18, IL27, and TNF, and for TH17-polarizing cytokines, we examined IL1B, IL6, IL23A, and TGFB1. As expected, human monocytes were rapidly activated in the presence of RNA viruses and produced proinflammatory and antiviral cytokines in a virus-specific fashion (Fig. 1). Some viruses preferentially induced the secretion of TH1-polarizing cytokines, such as EMCV, whereas others, such as MV and VSV, preferentially increased the expression of TH17-polarizing cytokines. Other viruses, such as IAV, induced the expression of both types of cytokines (Fig. 1A). These distinct gene expression patterns were confirmed at protein level by measuring cytokine release in the culture supernatant (Fig. 1B).

Fig. 1 RNA virus infections induce virus-specific cytokine and type I IFN responses by CD14+ monocytes.

(A) Quantitative real-time PCR (qRT-PCR) analysis of TH1-type (left) and TH17-type (right) proinflammatory cytokine mRNA expression in human CD14+ monocytes infected ex vivo with CV, EMCV, IAV, MV, SV, and VSV at an MOI of 5 for 16 hours. (B) ELISA analysis of proinflammatory cytokine secretion by virus-infected monocytes. (C) qRT-PCR analysis of type I IFN response mRNA expression in virus-infected monocytes. (D) ELISA measurement of IFNα and IFNβ production by virus-infected monocytes. Data are heat maps (A and C) or means ± SEM (B and D) of six donors from four independent experiments. *P < 0.05, **P < 0.005, and ***P < 0.001 by one-way analysis of variance (ANOVA) with Dunnett’s post hoc test to correct for multiple comparisons.

With regard to the type I IFN response, there was a notable variability in IFNA and IFNB expression after virus infection. Thus, whereas some viruses, such as EMCV, VSV, or IAV, significantly increased the expression of both antiviral cytokines, some other viruses, such as MV, CV, or SV, did not (Fig. 1C). IFNA and IFNB expression correlated with expression of ISGs, and thus, EMCV, VSV, and IAV infections promoted expression of ISGs, whereas CV barely induced the expression of any ISGs and MV and SV infection stimulated low but detectable amounts of ISG mRNAs (Fig. 1C). The results were confirmed at protein level by measuring IFNα and IFNβ in the culture supernatant of virus-infected monocytes (Fig. 1D). These data suggest that RNA viruses trigger the expression of proinflammatory cytokines and antiviral response in a virus-specific fashion in human CD14+ monocytes.

Single-stranded RNA-sensing TLRs contribute to virus-specific monocyte functions

During infection of antigen-presenting cells (APCs), RNA viruses are first sensed by host cells through several pattern recognition receptors (30), with a major contribution of those sensors that recognize viral nucleic acids (31), and in particular, RLRs. Subsequently, nucleic acid–specific TLRs are also involved in fine tuning of the immune response after RLR innate activation (11). We decided to examine the contribution of single-stranded RNA-sensing TLR (TLR7 and TLR8) to the cytokine patterns observed after infection of monocytes. TLR7 and TLR8 are both expressed by CD14+ monocytes ex vivo (fig. S1), and their abundance is increased after virus encounter (14, 32, 33). To test their involvement in monocyte activation, we blocked each of them individually with inhibitory oligodeoxynucleotides specific for TLR7 (IRS661) (34) and TLR8 (IRS957) (35, 36) in human monocytes before infection with RNA viruses. When we examined monocyte proinflammatory cytokine gene expression by real-time polymerase chain reaction (PCR) and protein secretion by enzyme-linked immunosorbent assay (ELISA), we found that stimulation of monocytes with IRS661 or IRS957 alone did not have any effect on cytokine expression (fig. S2). Both TLR7 and TLR8 blockade had virus- and cytokine-specific effects after infection (Fig. 2). TLR8 blockade did not inhibit the expression of IL1B nor IL23A in monocytes during virus infection. In contrast, in those virus infections that induced TH17-polarizing cytokines (IAV, MV, and VSV), TLR7 blockade statistically significantly decreased IL1B and IL23A expression while having no effect on EMCV, SV, and CV infections. The regulation of IL6 expression by TLR7 and TLR8 was virus specific, and although TLR7 blockade had no effect on IL6 expression upon CV, EMCV, SV, or VSV infections, it decreased IL6 expression in MV and IAV infections. Similarly, TLR8 blockade had no effect on IL6 expression triggered by all viruses except for EMCV, where it led to decreased gene expression (Fig. 2A). Blockade of both TLR7 and TLR8 by the simultaneous use of IRS661 and IRS957 did not completely abrogate cytokine expression, supporting the involvement of other pattern recognition receptors such as RLR, in innate immune cell activation after RNA infection (fig. S3).

Fig. 2 RNA virus infections induce different TLR7- and TLR8-specific cytokine expression patterns.

(A) qRT-PCR analysis of TH17-type (left) and TH1-type (right) proinflammatory cytokine mRNA expression in human CD14+ monocytes preincubated with IRS661 or IRS957 before infection with CV, EMCV, IAV, MV, SV, and VSV at an MOI = 5 for 16 hours. (B) ELISA analysis of cytokine secretion measured by virus-infected monocytes. All data are means ± SEM of five independent experiments with two donors in each. *P < 0.05, **P < 0.005, and ***P < 0.001 by one-way ANOVA with Dunnett’s post hoc test.

TLR7 and TLR8 had different effects on TH1-polarizing cytokine expression. Whereas TLR7 blockade had no effect on the expression of IL12B, IL27, or TNF triggered by any virus, TLR8 blockade significantly decreased the expression of IL12B and IL27 in CV, EMCV, MV, and SV infections. TNF was also inhibited in EMCV and IAV infections in the presence of the TLR8 antagonist IRS957 (Fig. 2A). TLR7 and TLR8 silencing in monocytes and subsequent virus infection confirmed the results obtained with IRS661 and IRS957 antagonist oligonucleotides (fig. S4). Cytokine secretion data assessed by ELISA supported the gene expression results and suggested that TLR7 was involved in the expression of TH17-polarizing cytokines [interleukin-1β (IL-1β), IL-6, and IL-23] after viral infection, whereas TLR8 promoted the expression of TH1-type cytokines (IL-27 and TNFα) and IL-6 (Fig. 2B).

TLR7 does not contribute to monocyte type I IFN response triggered by RNA viruses

We went on to examine the involvement of TLR7 and TLR8 signaling on the generation of a type I IFN response during virus infection. Monocytes were preincubated with IRS661 and IRS957 and infected with RNA viruses as above, and we determined the expression of IFNA1, IFNA2, and IFNB1 by real-time PCR (Fig. 3A). Whereas TLR8 blockade statistically significantly inhibited the expression of these genes in most virus infections, TLR7 blockade had no effect. In some cases, such as EMCV infection, TLR7 blockade statistically significantly increased IFNA1, IFNA2, or IFNB1 expression. These results were confirmed at protein level by measuring IFNα and IFNβ cytokines in the culture supernatant (Fig. 3B). Furthermore, the expression of the ISGs EIF2AK2, IFITM1, ISG15, MOV10, and TRIM5 was decreased by TLR8 blockade, whereas there was no effect after TLR7 blockade, with the exception of a slight reduction in IFITM1 and ISG15 expression after MV infection and EIF2AK2 and ISG15 after SV infection (Fig. 3C). TLR7 and TLR8 silencing further confirmed the absence of involvement of TLR7 in the type I IFN response after RNA virus infection and the role of TLR8 in triggering it (fig. S5).

Fig. 3 RNA virus infections induce different TLR7- and TLR8-specific type I IFN gene expression patterns.

(A) qRT-PCR analysis of IFNA and IFNB mRNA expression in human CD14+ monocytes preincubated with IRS661 or IRS957 before infection with CV, EMCV, IAV, MV, SV, and VSV at an MOI = 5 for 16 hours. (B) ELISA analysis of IFNα and IFN production by virus-infected monocytes. (C) qRT-PCR analysis of ISG mRNA expression in virus-infected monocytes. All data are means ± SEM of five independent experiments with two donors in each N.D. (not detected). *P < 0.05, **P < 0.005, and ***P < 0.001 by one-way ANOVA with Dunnett’s post hoc test.

Monocyte TLR7 triggering promotes CD4+ TH17 polarization, but TLR8 stimulation promotes TH1 differentiation

The different effects of TLR7 and TLR8 inhibition on proinflammatory cytokine expression and type I IFN responses led us hypothesize that TLR7 and TLR8 stimulation induce different functional phenotypes on CD14+ monocytes. To test this hypothesis, we chose imiquimod (IMQ) as a human TLR7-specific ligand and ssRNA40-LyoVec (ssRNA40) as a human TLR8-specific ligand, and we stimulated ex vivo isolated CD14+ monocytes with them to examine the expression of proinflammatory cytokines (Fig. 4). In agreement with the inhibition experiments, IMQ treatment induced the expression of IL1B, IL6, IL23A, and TGFB1 but failed to increase the expression of IL12B, IL18, IL27, or TNF. On the contrary, TH1-type cytokine gene expression was statistically significantly increased after ssRNA40 stimulation. TLR8 triggering also increased the expression of IL6 and IL1B, albeit at lower levels as compared to TLR7 stimulation (Fig. 4A). These differences were not due to a kinetics issue, because IMQ did not induce IL12B, IL27, or TNF expression at any time point examined during the 24 hours of stimulation (fig. S6A). Moreover, the differences in protein expression were not due to differences in cell numbers, as confirmed by flow cytometry (fig. S6, B and C). To determine whether the distinct cytokine expression patterns by IMQ and ssRNA40 were due to the specific concentrations of IMQ or ssRNA40 used, we stimulated monocytes with various concentrations of each TLR ligand covering the concentration ranges recommended by the manufacturer. Supernatant from the cultures were used to determine IL-1β, IL-6, IL-12, and TNFα secretion by ELISA (Fig. 4B). Whereas TLR8 stimulation induced the secretion of all cytokines examined, IMQ-stimulated monocytes did not promote secretion of IL-12 at any concentration tested, and it only induced a detectable amount of TNFα with the highest concentration used. The different cytokine production patterns were TLR7 and TLR8 specific, as TLR7 blockade with IRS661 inhibited IMQ-dependent secretion of IL-1β and IL-6, but not TNFα, and had no effect on cytokines released by ssRNA40-stimulated monocytes. In contrast, TLR8 blockade with IRS957 significantly decreased the secretion of IL-6 and TNFα but not IL-1β release upon ssRNA40 stimulation, while not affecting the production of IL-1β nor TNFα, and increasing IL-6 secretion after IMQ stimulation (Fig. 4C).

Fig. 4 TLR7-stimulated monocytes do not acquire a TH1-polarizing phenotype.

(A) qRT-PCR analysis of TH1-type and TH17-type cytokine mRNA expression by ex vivo isolated monocytes stimulated with vehicle, IMQ, or ssRNA40 for 16 hours. (B) ELISA analysis of cytokine secretion by monocytes stimulated with increasing concentrations of IMQ or ssRNA40 for 16 hours. (C) ELISA analysis of cytokine secretion by human CD14+ monocytes preincubated with IRS661 (gray), IRS957 (white), or vehicle (black) before stimulation with IMQ, ssRNA40, or vehicle for 16 hours. (D) qRT-PCR analysis of TH17-type (upper row) and TH1-type (lower row) cytokine mRNA expression by monocytes stimulated with vehicle, the TLR7-specific ligands Gardiquimod (GDQ) and Loxoribine (Loxo), or the TLR8-specific ligands ssRNA-DR and PolyU. (E) qRT-PCR analysis of chemokines and cell surface receptor mRNA expression by monocytes stimulated vehicle, IMQ, or ssRNA40 for 16 hours. (F) Flow cytometry analysis of ICOSL and CD40 abundance on monocytes stimulated with IMQ (black line), ssRNA40 (black line), or vehicle (gray line) for 6 hours in comparison to isotype control (gray curve) staining. Histograms with geometric mean fluorescence intensity (gMFI) for vehicle (gray) versus TLR treatment (black) are representative of five independent experiments performed with two donors in each. All quantified data (A to E) are means ± SEM of all experiments. *P < 0.05, **P < 0.005, and ***P < 0.001 by one-way ANOVA with Tukey’s post hoc test.

To further confirm the specificity of our results, we stimulated monocytes with other human TLR7- and TLR8-specific ligands, including gardiquimod (GDQ) and loxoribine (Loxo) as TLR7 ligands, and single-stranded RNA double-right (ssRNA-DR) and Poly(U) as TLR8-specific ligands. Gene expression analysis of TH1– and TH17-type cytokines confirmed that TLR7 ligands preferentially induce IL1B, IL6, and IL23A, whereas TLR8 ligands increase the expression of IL12B, IL27, and TNF (Fig. 4D), and this was confirmed at protein level in the culture supernatant (fig. S6D). Furthermore, we examined the expression of surface receptors and chemokines that have been associated to type 1 and 17 responses in monocytes. We found that TLR8 stimulation did not increase the expression of inducible T cell costimulator ligand (ICOSL), which is important for human TH17 differentiation (37), but did increase the expression of the TH1-related receptor CD40 (38). In contrast, TLR7 signaling slightly increased the expression of ICOSL but did not induce CD40 expression at both the RNA and protein levels (Fig. 4, D to F). Moreover, IMQ-treated monocytes statistically significantly increased the expression of the chemokine CCL20, the ligand for CCR6, which is preferentially expressed by TH17 cells (39). In contrast, ssRNA40-stimulated monocytes increased CXCL10 expression, the ligand for the chemokine receptor CXCR3, which is preferentially expressed by TH1 cells (40).

To functionally confirm the capacity of TLR7- and TLR8-stimulated monocytes to induce TH17 and TH1 cells, respectively, we stimulated monocytes with IMQ, ssRNA40, or vehicle and subsequently cocultured them with CD4+ T cells isolated from the same donors in the presence of only anti-CD3, to allow monocytes to provide the costimulatory signals under a similar polyclonal TCR stimulation (Fig. 5). Subsequently, CD4+ T cells were sorted by flow cytometry to examine the expression of TH1-related (IFNG and TBX21), TH2-related (IL4 and GATA3), TH17-related (IL17A and RORC), and regulatory T cell (Treg)-related (FOXP3 and IL10) cytokines and transcription factors. CD4+ T cells that were cocultured with ssRNA40-treated monocytes statistically significantly increased expression of IFNG and TBX21 and decreased expression of GATA3 and RORC when compared to CD4+ T cells cocultured with vehicle-treated monocytes. However, CD4+ T cells cocultured with IMQ-treated monocytes failed to induce IFNG or TBX21 but statistically significantly increased the expression of IL17A and RORC. No differences were observed in the expression of FOXP3 or IL10 by CD4+ T cells in any coculture (Fig. 5A). Cytokine release in the cocultures measured by ELISA confirmed that IFNγ was only detected in the cocultures with ssRNA40-treated monocytes, and IL-17 was preferentially secreted in cocultures with IMQ-treated monocytes (Fig. 5B). The lack of TH1 cell induction by IMQ-treated monocytes was also confirmed at early and late time points during the coculture period (fig. S6E).

Fig. 5 TLR7 and TLR8 differentially polarize CD4+ T cells.

(A) qRT-PCR analysis of TH1, TH2, TH17, and Treg cytokines and transcription factor mRNA expression in sorted CD4+ T cells after coculture for 24 hours with monocytes (Monoc) stimulated with vehicle, IMQ, or ssRNA40. (B) ELISA analysis of IFNγ and IL-17 production from cocultures of CD4+ T cells with monocytes stimulated with vehicle, IMQ, or ssRNA40 at the indicated monocyte:CD4 ratio for 72 hours. Data are means ± SEM of five independent experiments with two donors in each *P < 0.05, **P < 0.005, and ***P < 0.001 by one-way ANOVA with Tukey’s test (A) or two-way ANOVA with Tukey’s post hoc test (B).

The B cell adapter for phosphatidylinositol 3-kinase (PI3K)–dependent activation of PI3K (BCAP) signaling inhibits proinflammatory cytokine secretion after TLR stimulation in DCs (41). To examine whether differential activation of the PI3K pathway during TLR7 or TLR8 signaling was responsible, at least in part, for the phenotypes that we observed in monocytes, we stimulated monocytes ex vivo with IMQ or ssRNA40 in the presence of two widely used PI3K inhibitors (LY294002 and wortmannin), and we assessed IL-12 and IL-6 secretion. PI3K inhibition did not alter the secretion of IL-6 after IMQ or ssRNA40 stimulation but did reduce IL-12 production by TLR8-stimulated monocytes. Inhibition of PI3K after IMQ stimulation did not increase IL-12 expression, which suggests that the PI3K pathway is not involved in inhibition of proinflammatory cytokines in TLR7-stimulated monocytes (fig. S7).

TLR7 preferentially signals through mitogen-activated protein kinase and displays defective nuclear factor κB activation

A common feature of most TLR signaling pathways is the activation of three major transcription factors, i.e., AP-1 and nuclear factor κB (NFκB), responsible for proinflammatory cytokine release, and interferon regulatory factor 3 (IRF3), involved in the antiviral type I IFN response (42). We examined whether the differences in proinflammatory cytokine secretion patterns elicited by TLR7 and TLR8 stimulation were due to differences in their capacity to activate AP-1 and NFκB. For this, we stimulated human CD14+ monocytes with IMQ and ssRNA40, and we measured the expression of phosphorylated MKK3/6, MKK4/7, and MEK1/2 (Fig. 6). Mitogen-activated protein kinase (MAPK) kinases (MKKs) and MAPK kinases (MEKs) belong to the MAPK family and are involved in the activation of p38, c-Jun N-terminal kinase (JNK), and extracellular signal–regulated kinase (ERK), whose phosphorylation leads to AP-1 activation and secretion of proinflammatory cytokines (43). MKK3/6 and MKK4/7 were rapidly and comparably activated upon IMQ or ssRNA40 stimulation. However, there was a small but statistically significant increase in the phosphorylation of MEK1/2 after IMQ stimulation at 5 and 15 min that was not observed after ssRNA40 stimulation (Fig. 6, A and B). Subsequently, we assessed the expression of phosphorylated p38 and ERK1/2 after IMQ or ssRNA40 activation. TLR7 induced an increase in pp38 at all time points examined except for 60 min, whereas TLR8 stimulation only induced a modest increase at 5 and 30 min. The differences in p38 phosphorylation between IMQ- and ssRNA40-stimulated monocytes were statistically significant at all time points examined except for 60 min. Furthermore, whereas TLR8 failed to increase pERK1/2 expression when compared to vehicle-treated cells, TLR7-stimulated monocytes displayed a sharp increase in pERK1/2 expression 5 min after stimulation (Fig. 6, C and D). These data suggest that TLR7 signaling induces MAPK activation more effectively than TLR8 signaling, in agreement with the increased IL-1β and IL-23 production observed, which are in part activated by AP-1 (44, 45).

Fig. 6 TLR7 and TLR8 stimulation differentially activate MAPK and NFκB signaling.

(A and B) Flow cytometric analysis of pMKK3/6, pMKK4/7, and pMEK1/2 abundance in monocytes stimulated with IMQ (gray thick line) or ssRNA40 (black fine line) for 15 min in comparison to isotype control (gray curve) staining. Histograms (A) are representative of seven donors from five independent experiments. Normalized gMFI over time (B) are means ± SEM of all experiments. (C and D) Flow cytometric analysis of pp38 and pERK1/2 abundance in monocytes stimulated with IMQ or ssRNA40 for 30 and 5 min, respectively. Histograms (C) are representative of seven donors from five independent experiments. Normalized gMFI over time (D) are means ± SEM of all experiments. (E and F) Flow cytometric analysis of pNFκB p65 and IκBα abundance in monocytes stimulated with IMQ or ssRNA40 for the indicated time. Histograms (E) are representative of five donors from three independent experiments. Normalized gMFI over time (F) are means ± SEM of all experiments. *P < 0.05, **, ##P < 0.005, and ***, ###P < 0.001 by two-way ANOVA with correction for multiple comparisons with Tukey’s post hoc test.

We went on to examine NFκB activation after TLR7 and TLR8 stimulation. NFκB is a homo- or heterodimer of NFκB or Rel proteins. In resting cells, NFκB remains in the cytoplasm in an inactive state, complexed with the inhibitory protein inhibitor of nuclear factor κBα (IκBα). TLR triggering activates NFκB by both phosphorylation of its subunits and ubiquitination-mediated proteasomal degradation of IκBα, which leads to NFκB translocation to the nucleus (46, 47). We assessed NFκB activation by p65 subunit phosphorylation and IκBα degradation (Fig. 6, E and F). p65 phosphorylation showed a statistically significantly different kinetics on TLR7-stimulated monocytes as compared to TLR8-stimulated monocytes. Whereas ssRNA40 induced a robust activation of p65 maintained during 60 min of activation, IMQ-stimulated monocytes displayed a modest p65 activation at 15 and 30 min that rapidly vanished. Significant differences in phospho-p65 were observed at 45 and 60 min after activation when comparing IMQ- versus ssRNA40-treated monocytes (Fig. 6F). IκBα expression correlated with the differences in p65 phosphorylation. Thus, although ssRNA40 induced a statistically significant decrease in the expression of IκBα by 30 min of activation, this was almost absent in IMQ-treated monocytes, and there was only a statistically significant decrease in its expression at 60 min, supporting the observation that TLR7 stimulation fails to induce robust NFκB activation, and consequently, the expression of the type 1 cytokines that are dependent on NFκB function, such as TNFα (48, 49), IL-12 (50), and IL-27 (51, 52).

FOSL1 inhibits type 1 cytokines after TLR7 stimulation

AP-1 transcription factors are a family of heterodimers predominantly composed of one member of each of two major families of proteins: FOS (c-FOS, FOSL1, FOSL2, and FOSB) and JUN (c-JUN, JUNB, and JUND). In addition, members of the activating transcription factor (ATF) and MAF families can replace in some instances one of the FOS or JUN proteins in the dimer (53). AP-1 regulation occurs both transcriptionally and posttranslationally after exposure to several different stimuli (54), including TLR ligation (55). We examined the expression of AP-1 subunits elicited by ssRNA40 and IMQ as compared to vehicle-treated monocytes. The expression of all JUN family genes was similarly increased by either IMQ or ssRNA40 when compared to control (fig. S8A); however, there were statistically significant differences in the expression of ATF1, ATF2, FOSL1, and FOSL2. Monocytes stimulated with TLR7 increased expression of all four transcription factors as compared to vehicle-treated monocytes, whereas TLR8-stimulated cells decreased ATF1 and ATF2 and slightly increased FOSL2 expression when compared to vehicle-stimulated monocytes. ssRNA40 had no effect on FOSL1 expression (Fig. 7A).

Fig. 7 TLR7-dependent FOSL1 expression inhibits TH1-type cytokines.

(A) qRT-PCR analysis of ATF1, ATF2, FOSL1, and FOSL2 mRNA expression by monocytes stimulated with IMQ (black) or ssRNA40 (white) normalized to vehicle-treated monocytes. Normalized data are means ± SEM of two independent experiments performed with two donors in each. (B and C) qRT-PCR (B) and ELISA (C) analysis of TH1-type cytokine mRNA gene expression and production by ATF1-, ATF2-, FOSL1-, and FOSL2-silenced monocytes stimulated with IMQ (black) or vehicle (white) as compared to nontarget (NT)–silenced control. Data are means ± SEM of six donors from four independent experiments. (D and E) qRT-PCR analysis of FOSL1 (D) or TNF and IL27 (E) mRNA expression by monocytes stimulated with the indicated RNA viruses for 16 hours. Data are means ± SEM of five donors from four independent experiments. *, #P < 0.05, **P < 0.005, and ***, ###P < 0.001 by one-way ANOVA with Tukey’s post hoc test (D) or two-way ANOVA with Tukey’s post hoc test (A, B, C, and E) for comparison to vehicle treatment (A) or to NT small interfering RNA (siRNA) treated similarly (B, C, and E).

Unlike the rest of the FOS family, FOSL1 and FOSL2 proteins lack transactivation domains and display weak transactivation potentials. Under certain circumstances, they even act as inhibitors of AP-1 activity by competing for binding to AP-1 sites or by dimerizing with JUN and forming “inactive” heterodimers (54, 56, 57). On the basis of these observations and the fact that type 1 cytokines, which require AP-1 activation (51, 5864), were not produced after TLR7 signaling, we hypothesized that increased expression of FOSL1 and/or FOSL2 inhibited type 1 cytokines. To test this, we silenced ATF1, ATF2, FOSL1, and FOSL2 on CD14+ monocytes (fig. S8B), and we subsequently stimulated them with either vehicle or IMQ. Gene expression analysis demonstrated that ATF1, ATF2, and FOSL2 silencing significantly decreased the expression of IL1B and IL6 after TLR7 stimulation, which suggested that these transcription factors have redundant roles in the expression of type 17 cytokines (fig. S8C). IL23A expression was decreased in ATF2-silenced monocytes as compared to controls after IMQ treatment, which suggested that ATF2 promoted IL23A transcription. Genetic deletion of FOSL1 had no effect on any of the type 17 cytokines measured. However, when we examined the expression of type 1 cytokines (Fig. 7B), FOSL1 silencing significantly increased the expression of IL12B, IL27, and TNF after IMQ treatment, as compared to nontarget (NT)–transfected cells. ATF1, ATF2, and FOSL2 silencing had no effect on type 1 cytokine expression after IMQ stimulation. We confirmed the inhibitory effects of FOSL1 on type 1 cytokine secretion by measuring IL-27 and TNFα secretion in the culture supernatant from FOSL1-silenced monocytes (Fig. 7C).

Because TLR7 and TLR8 displayed nonoverlapping roles in the secretion of specific cytokines after RNA virus infections (Fig. 1), we examined the expression of FOSL1 in monocytes stimulated with RNA viruses. We observed that FOSL1 expression was significantly increased after IAV, MV, and VSV infections as compared to vehicle-treated monocytes (Fig. 7D), which displayed a more TH17-polarizing capacity that was partly TLR7 dependent (Fig. 2). To examine whether virus-induced FOSL1 regulates the expression of type 1 cytokines, we infected FOSL1- or NT-silenced monocytes with IAV, MV, and VSV and examined the expression of IL-27 and TNF. FOSL1-silenced cells increased TNF expression after IAV infection. Moreover, IL27 was increased in FOSL1-silenced monocytes after IAV and MV infection. These data suggest that RNA viruses that signal through TLR7 increase expression of FOSL1 to restrain type 1 proinflammatory cytokine production (Fig. 7E).

TLR7 signaling does not induce a type I IFN response in human CD14+ monocytes

An important function of myeloid cells is their capacity to promote an antiviral state after RNA virus infection by secreting type I IFN and expression of ISGs (65). Monocytes infected with several RNA viruses secreted type I IFN in a TLR7-independent manner (Fig. 3). To examine whether direct stimulation through TLR7 induced a type I IFN response, CD14+ monocytes were stimulated ex vivo with IMQ or ssRNA40 for 36 hours, and the expression of IFNA1, IFNA2, and IFNB1 was examined at various time points (Fig. 8A). Although IMQ did not increase expression any of the three cytokines at any time point, ssRNA40 induced all three with different expression kinetics. These data were confirmed at protein level by measuring IFNα and IFNβ in the culture supernatant (Fig. 8B). The lack of IFN response after TLR7 stimulation was not due to the specific concentration of IMQ used, because none of the various concentrations tested within the recommended range induced significant expression of any of the IFN cytokines (Fig. 8C), whereas increasing amounts of ssRNA40 increased the expression of IFNA1, IFNA2, and IFNB1 and secretion of IFNα and IFNβ proteins (Fig. 8D). Accordingly, the expression of ISGs was virtually absent in IMQ-stimulated monocytes, but TLR8 activation promoted the expression of all six ISG tested (Fig. 8E). To confirm that the defect in IFN expression after IMQ stimulation was due to a specific inability of TLR7 triggering to induce IFN in monocytes, we stimulated monocytes with GDQ and Loxo as TLR7 ligands, and with ssRNADR and PolyU as TLR8-specific ligands, and IFN gene expression was examined 16 hours later (Fig. 8F). Although neither GDQ nor Loxo induced IFN expression, both ssRNADR and PolyU statistically significantly increased the expression of IFNA1, IFNA2, and IFNB1. Pretreatment of monocytes with IRS661 and IRS957 confirmed the TLR8 specificity of the results obtained (Fig. 8G).

Fig. 8 TLR7 does not induce a type I IFN response in human CD14+ monocytes.

(A and B) qRT-PCR (A) and ELISA (B) analysis of IFNA and IFNB mRNA expression and production by human CD14+ monocytes stimulated with vehicle (white square), IMQ (black circle), or ssRNA40 (white circle) for 36 hours. (C and D) qRT-PCR (C) and ELISA (D) analysis of IFNA and IFNB mRNA expression and secretion by monocytes stimulated with different concentrations of IMQ or ssRNA40 for 16 hours compared to vehicle (dotted line). (E) qRT-PCR analysis of ISGs mRNA expression by monocytes stimulated with vehicle (white square), IMQ (black circle), or ssRNA40 (white circle) for 36 hours. (F) qRT-PCR analysis of IFNA and IFNB gene expression by monocytes stimulated with vehicle (white bar), the TLR7-specific ligands (black bars) GDQ and Loxo, or the TLR8-specific ligands (gray bars) ssRNA-DR and PolyU for 16 hours. (G) qRT-PCR analysis of IFNA and IFNB gene expression by monocytes preincubated with control (black), IRS661 (gray), or IRS957 (white) oligos before stimulation with vehicle, IMQ, or ssRNA40 for 16 hours. (H and I) Flow cytometry analysis of pTBK1 and pIRF3 abundance in monocytes stimulated with vehicle (gray curve), IMQ (thick gray line), or ssRNA40 (black thin line) for 15 and 30 min, respectively. Histograms (H) with gMFI for IMQ (gray) and ssRNA40 treatment (black) are representative of five donors from three independent experiments. Normalized gMFI (I) data are means ± SEM of all experiments. All other data (A to G) are means ± SEM of two independent experiments performed with two donors in each. *, #P < 0.05, **P < 0.005, and ***, ###P < 0.001 by two-way ANOVA with Tukey’s test for comparison to vehicle treatment (A, B, and E), to vehicle-treated cells with the same pretreatment (G), or to 0 min (I).

Induction of a type I IFN response in myeloid cells after nucleic acid–sensing TLR stimulation can involve the activation of TANK-binding kinase 1 (TBK1) by nucleic acid–sensing TLRs (66). Subsequently, IRF3 and IRF7 are activated (67, 68), and both are important for the expression of type I IFN genes (65). We examined the activation of the TBK1-IRF3/7 pathway in monocytes stimulated with IMQ or ssRNA40 as compared to vehicle-treated cells (Fig. 8, H and I). The activation of TBK1 was increased by ssRNA40 during the first 30 min of stimulation. In contrast, there was only a slight increase in the phosphorylation of TBK1 after IMQ stimulation at 15 min when compared to vehicle-treated monocytes. We then examined the phosphorylation of IRF3 by TBK1 at Ser386, and although both TLR7 and TLR8 stimulation induced IRF3 activation, phosphorylation of IRF3 by ssRNA40 was significantly stronger at each time point than that of IMQ-stimulated cells (Fig. 8I). RNA viruses promoted phosphorylation of IRF3, and those that predominantly signaled through TLR7 instead of TLR8 (VSV, MV, and IAV) activated less pIRF3 in the presence of the TLR7 inhibitor IRS661. In contrast, EMCV-dependent pIRF3, which preferentially activates TLR8, was only reduced in the presence of the TLR8 inhibitor IRS957 but not IRS661. Furthermore, blocking both TLR7 and TLR8 did not completely abrogate pIRF3, in agreement with the involvement of other nucleic acid–sensing mechanisms during infection with the RNA viruses tested (fig. S9).

Although recent data suggest that unlike TLR8, TLR7 signaling specifically induces activation of IRF5 over IRF3 (69), we did not observe preferential activation of IRF5 by TLR7 as compared to TLR8. These data are in agreement with the observation that TLR8 can also lead to the activation of IRF5 in human primary monocytes and macrophages (70). These data suggest that differential activation of IRF5 is not responsible for the differences in type I IFN responses observed between TLR7 and TLR8 stimulation (fig. S10).

TLR7-dependent Ca2+ signaling inhibits type I IFN response

The increased IRF3 activation by IMQ did not correlate with the lack of IFN gene expression, suggesting that other mechanisms could be inhibiting the expression of type I IFN genes. We hypothesized that a TLR7-specific pathway triggered by IMQ would inhibit IFN expression. An important difference that we had previously observed between TLR7 and TLR8 signaling on CD4+ T cells was an increase in intracellular Ca2+ triggered by TLR7 stimulation that did not occur after ssRNA40 stimulation (71). Because Ca2+ signaling alters the functions of myeloid cells, and it is an important modulator of cell responses during viral infections (72, 73), we examined how TLR7 stimulation may alter intracellular Ca2+ concentration in monocytes (Fig. 9A). Unlike ssRNA40, IMQ stimulation increased intracellular Ca2+ concentration in a dose-dependent manner (Fig. 9A). The effect observed was TLR7 dependent, because preincubation with IRS661 almost completely inhibited Ca2+ flux as compared to IRS control-treated monocytes (Fig. 9B). To determine the origin of the increase in Ca2+, we stimulated monocytes with IMQ in the presence of EGTA (extracellular Ca2+ chelator) and xestopongin C, an inhibitor of inositol 1,4,5-triphosphate–dependent Ca2+ release. Whereas chelation of extracellular Ca2+ did not affect IMQ-induced increase in Ca2+, xestopongin C inhibited it almost completely, suggesting that TLR7-driven increased in intracellular Ca2+ occurs through depletion of internal stores located in the endoplasmic reticulum (Fig. 9C).

Fig. 9 TLR7-dependent Ca2+ signaling inhibits type I IFN response.

(A to C) Flow cytometry analysis of Ca2+ flux over time in monocytes stimulated (downward arrows) with the indicated concentrations of IMQ, ssRNA40, or ionomycin (A), monocytes preincubated with control (black), or IRS661 (gray) oligos before stimulation with IMQ (B), or monocytes preincubated with vehicle (black), EGTA (left), or xestopongin C (right) before stimulation with IMQ (C). All histograms are a representative of five independent experiments. (D to G) qRT-PCR analysis of IFNA and IFNB (D), ISGs (E), and IRF3 and IRF7 (G) mRNA gene expression and IFNα secretion (F) by monocytes preincubated with vehicle or Quin-2 AM and stimulated with IMQ as indicated for 16 hours. Data are means ± SEM of four independent experiments (G) performed with two donors in each (D to F). *P < 0.05, **P < 0.005, and ***P < 0.001 by one-way ANOVA with Tukey’s post hoc test for comparison to IMQ treatment alone.

We hypothesized that TLR7-dependent Ca2+ flux may inhibit the expression of type I IFN. To test this, we used Quin-2 AM as a chelating agent to block free intracellular Ca2+ (71, 74), and we stimulated monocytes with IMQ to examine the expression of type I IFN genes (Fig. 9, D and E). Although Quin-2 AM alone had no statistically significant effect on the expression of any of the genes tested, stimulation of monocytes with IMQ in the presence of increasing concentrations of Quin-2 AM statistically significantly augmented the expression of IFNA1 and IFNA2. Accordingly, the expression of ISGs was also increased by IMQ when intracellular Ca2+ flux was blocked (Fig. 9D). The increased IFNA gene expression was confirmed at protein level by IFNα in the culture supernatant (Fig. 9F). When we examined whether Ca2+ blockade had an effect on IRF3 and IRF7 expression (Fig. 9G), we found that monocyte stimulation with IMQ in the presence of Quin-2 AM significantly increased the expression of IRF7 at 24 hours, but IRF3 expression was significantly decreased when compared to IMQ only–treated cells. These results were similar to the pattern of IRF3 and IRF7 expression observed after TLR8 stimulation as compared to vehicle-treated cells at similar time points (fig. S11). Because increases in intracellular Ca2+ concentration can activate nuclear factor of activated T cells (NFAT) (71), we preincubated monocytes with VIVIT peptide (75), which inhibits NFAT activation by preventing its nuclear translocation, before IMQ stimulation. We found that the TLR7-dependent antagonistic effects on proinflammatory cytokine and type I IFN gene expression were NFAT independent, because VIVIT peptide did not restore the expression of IL12, IL27, and TNF nor the expression of type I IFN genes (fig. S12). Together, these results suggest that TLR7-dependent Ca2+ signaling acts as an inhibitor of type I IFN response in human CD14+ monocytes, independently of IRF3 and IRF7 expression, as well as NFAT activation.

TLR7-dependent FOSL1 expression is Ca2+ and ERK dependent

To further explore how TLR7 stimulation increased the expression of FOSL1 (Fig. 7), we pharmacologically inhibited pathways previously identified that were activated after TLR7 activation, including p38, ERK, JNK (Fig. 6), and Ca2+ (Fig. 8). We stimulated monocytes with IMQ in the presence of inhibitors of p38 (SB203580), ERK1/2 (SCH772984), JNK (SP600125), and the Ca2+ intracellular chelator Quin-2 AM, and we measured FOSL1 expression (Fig. 10). Although p38 and JNK inhibition did not abrogate IMQ-stimulated FOSL1 expression, both ERK1/2 inhibition and Ca2+ chelation with Quin-2 AM statistically significantly inhibited the expression of FOSL1 after TLR7 stimulation. Because FOSL1 can inhibit IFN signaling in mice and in human cell lines (76), we examined whether TLR7-induced FOSL1 was required for the inhibition of type I IFN responses after IMQ stimulation (Fig. 10E). FOSL1 silencing in monocytes did not restore IFNA expression after IMQ activation nor increase IFNA gene expression after ssRNA40 stimulation. Together, these data suggest that both TLR7-dependent ERK1/2 activity and Ca2+ flux promote FOSL1 expression in human monocytes, but FOSL1 is not involved in the inhibition of type I IFN response after TLR7 stimulation.

Fig. 10 FOSL1 expression requires Ca2+ and ERK signaling.

(A to D) qRT-PCR analysis of FOSL1 mRNA gene expression by CD14+ monocytes preincubated with the p38 inhibitor SB203580 (A), the JNK inhibitor SP600125 (B), the ERK1/2 inhibitor SCH772984 (C), or the Ca2+ chelation agent Quin-2 AM (D) before stimulation with vehicle (white bars) or IMQ (gray bars) for 18 hours. Data are means ± SEM of four independent experiments. (E) qRT-PCR analysis of IFNA1 and IFNA2 mRNA expression by NT control or FOSL1-silenced monocytes stimulated with vehicle, IMQ, or ssRNA40. Data are means ± SEM of five donors from three independent experiments. *P < 0.05, **P < 0.005, and ***P < 0.001 by two-way ANOVA with Tukey’s test (A to C and E) for comparison between NT- and FOSL1-silenced monocytes (E), or one-way ANOVA with Tukey’s post hoc test (D).

DISCUSSION

The innate immune system is armed with receptors that sense pathogens to mount effective immune responses during infection. Monocytes are a major target of many viruses, and infection not only induces monocyte differentiation into other cell populations but it also activates them to perform effector functions (4). In this work, we dissected the TLR7- and TLR8-dependent events that tune the activation of human monocytes upon RNA virus infection, and we defined TLR7- and TLR8-specific signaling pathways in this process (fig. S13). Furthermore, by dissecting the signaling pathway downstream TLR7, we found that FOSL1 inhibits the expression of type 1 cytokines upon TLR7 signaling. Last, we described the TLR7-Ca2+ axis limited type I IFN responses in CD14+ monocytes.

While APCs in general use a variety of sensors to detect pathogens, initial virus sensing is carried out by RLRs leading to local IFN production (11). Subsequently, both RLR and TLR act in concert to specifically tune the adaptive immune response (31). Thus, whereas RLR signaling is essential for the initial triggering of antiviral responses, the TLR serve a secondary role to drive specific cytokine production and type I IFN responses to regulate and shape the type of adaptive immune response and programming of cell-mediated immunity (77, 78). In this regard, we observed that monocyte infection with a variety of RNA viruses led to various phenotypes that were dependent on TLR7 and/or TLR8 signaling. However, and in agreement with the involvement of RLR in innate immune activation, TLR7 or TLR8 blockade upon virus infection did not completely abrogate the expression of cytokines in our experimental system (79, 80). The concerted actions of these pathways to human monocyte function has not been addressed, and it will be interesting for future studies to examine the cross-talk between TLR7 and TLR8 signaling and other nucleic acid–sensing pathways in human monocytes during virus infection.

Our data demonstrated that TLR7 and TLR8 signaling in CD14+ monocytes are significantly different, and although TLR7 preferentially increased expression of TH17-polarizing cytokines through the activation of MAPK cascades that lead to AP-1 activation (Fig. 7 and Fig. 10), TLR8 predominantly induced type 1 cytokines, which are NFκB dependent (46). TLR7 and TLR8 signaling have differential roles in other immune cell populations, such as DCs and neutrophils (70, 8184), and TLR8 promotes preferential type 1 cytokine secretion when compared to TLR7 stimulation in natural killer cells (81). Furthermore, in other disease settings, TLR7 inhibits type 1 cytokine secretion (85), and TLR7 engagement on DCs leads to an activated phenotype with increase ability to promote TH17 responses in an in vivo model of uveitis (86) and in human DC in vitro (87). FOSL1 is expressed primarily in TH17 cells as compared to other T helper subsets, and IL-17 is directly controlled by FOSL1 (88). Moreover, FOSL1 expression is correlated with high psoriasis (89) and increased susceptibility to collagen-induced arthritis (88).

AP-1 family of transcription factors activate expression of genes involved in proliferation, activation, and other cell functions. The observation that FOSL1 inhibits type 1 proinflammatory cytokines underscores the complexity of the AP-1 family, which is generally overlooked, and raises the questions of whether individual stimuli activate specific AP-1 dimers, what the transcriptional targets for each dimer are, and what the distribution of AP-1 dimers across cells types or over time in a specific cell type are (54). Furthermore, in agreement with our results, other AP-1 members inhibit TLR signaling (90, 91). For example, loss of cFOS in mice results in increased NFκB signaling in response to lipopolysaccharide (LPS) (92, 93), and overexpression of FOSL1 in macrophages inhibits the expression of proinflammatory cytokines after LPS stimulation (94, 95). Consistent with ERK activation by Ca2+ in other cell systems (96, 97), we observed that the TLR7-dependent increase in FOSL1 expression was dependent on both increased Ca2+ concentrations and ERK activation. Therefore, it is plausible that these two events may occur sequentially after TLR7 activation to induce the expression of FOSL1, with TLR7-induced Ca2+ increase activating ERK activity, which in turn is required for the up-regulation of FOSL1.

The observation that TLR7 ligation inhibits type I IFN response by augmenting intracellular Ca2+ release in monocytes is unexpected, and further investigations are warranted to dissect the mechanisms by which calcium modulates type I IFN responses. However, several works have suggested a negative relationship between calcium signaling and IFN activity (98100), and it is well known that viruses use diverse strategies to inhibit type I IFN responses (101). It is thus tempting to speculate that RNA viruses would use TLR7 signaling when infecting CD14+ monocytes to avoid type I IFN responses and expand in the host, be transported to other tissues, etc. In support of our data, several signaling molecules that are involved in TLR7 signaling in human monocytes inhibit type I IFN responses in other cell types, such as MEK1/2-ERK (102) and FOSL1 (76). Moreover, in human macrophages after hepatitis C virus infection (14) and in human DCs after influenza infection or small-molecule activation (103), TLR7 stimulation limits type I IFN expression.

The involvement of nucleic acid–sensing mechanisms in the immune response against infections and even in autoimmune diseases (104) makes these pathways interesting targets for drug design. TLR agonists are being tested in clinical trials in cancer and infectious diseases, because they provide enhanced immune responses in these settings (10, 105). In this regard, our results have important consequences on the selection of these agonists as adjuvants in therapeutic immunizations against cancer or prophylactic vaccines against pathogens and demonstrate that the fundamental differences in signaling between these two related receptors could potentially be harnessed to tailor specific immune responses in various disease settings.

MATERIALS AND METHODS

Study subjects

Peripheral blood was drawn from healthy participants after informed consent and approval by the Institutional Review Board at Yale University and Imperial College London. All experiments were performed conformed to the principles set out in the World Medical Association Declaration of Helsinki and the Department of Health and Human Services Belmont Report.

Cell culture reagents

Cells were cultured in RPMI 1640 media supplemented with 2 nM l-glutamine, 5 mM Hepes, penicillin/streptomycin (100 U/μg per milliliter) (BioWhittaker), 0.5 mM sodium pyruvate, 0.05 mM nonessential amino acids (Life Technologies), and 5% human AB serum (Gemini Bio-Products). Monocytes were cultured in 96-well polypropylene plates to avoid nonspecific activation by adhesion to polystyrene.

The TLR7-specific inhibitory sequence IRS661, the TLR8-specific sequence IRS957, and an NT sequence as a control were synthesized by Sigma-Aldrich on a phosphorothioate backbone and used at 2 to 5 μM. None of these inhibitory sequences induced statistically significant cell death at the concentrations described above. The p38 inhibitor SB203580 (Selleckchem) (106, 107), the ERK1/2 inhibitor SCH772984 (Selleckchem) (108, 109), and the JNK inhibitor SP600125 (InvivoGen) (110, 111) were used at 2 and 10 μM. The Ca2+ chelator Quin 2-AM (Thermo Fisher Scientific) was used at 0.5 and 2.5 μM (71).

Cell isolation and activation with TLR ligands

Peripheral blood mononuclear cells were isolated from healthy donors by Ficoll-Paque PLUS gradient centrifugation (GE Healthcare). Classical CD14+ monocytes were isolated by positive selection using the EasySep Human CD14-Positive Selection Kit (STEMCELL Technologies). Total CD4+ T cells were isolated by negative selection using the EasySep Human CD4+ T cell Enrichment Kit (STEMCELL Technologies). B cell isolation was performed with the EasySep Human CD19 Positive Selection Kit II (STEMCELL Technologies). The following TLR ligands were used to stimulate CD14+ monocytes: IMQ, in concentrations ranging from 0.1 to 10 μg/ml, ssRNA40-LyoVec (InvivoGen), in concentrations ranging from 0.1 to 5 μg/ml, 0.5 mM Loxo, GDQ (2.5 μg/ml), ssRNA-DR (2.5 μg/ml), and PolyU (2.5 μg/ml). All TLR ligands (<0.001 EU/μg; EndoFit) were obtained from InvivoGen and resuspended in endotoxin-free water according to the manufacturer's recommendations.

Monocyte-RNA virus cocultures

CD14+ monocytes were stimulated with RNA viruses at a multiplicity of infection (MOI) of 5 for 16 hours at 37°C. The viruses used in this manuscript are CV B1 (Conn-5 strain), MV (Edmonton strain), SV (Sendai/52 strain), EMCV, all obtained from the American Type Culture Collection. VSV (Indiana strain) was a gift from A. Iwasaki’s laboratory (Department of Immunobiology, Yale School of Medicine), and IAV (PR8 strain) was a gift from A. Garcia-Sastre’s laboratory (Icahn School of Medicine, Mount Sinai, NY). Culture supernatant was collected for cytokine release measurement by ELISA, and cells were lysed in RLT Plus buffer (QIAGEN) for gene expression assays.

CD14+:CD4+ T cell cocultures

Freshly isolated CD14+ monocytes were resuspended at 106 cells/ml in complete medium and stimulated with IMQ (2.5 μg/ml), ssRNA40 (2.5 μg/ml), or vehicle for 16 hours at 37°C. Subsequently, monocytes were cocultured with isolated CD4+ T cells from the same donor in the presence of anti-CD3 (1 μg/ml; UCHT1 clone, BD Biosciences) at different T cell:monocyte ratios. Supernatant was collected after 24, 72, and 120 hours for IFNγ and IL-17 measurement by ELISA, and the cells were stained with LIVE/DEAD fixable violet dye (Thermo Fisher Scientific), CD14 and CD3. CD3+CD14 T cells were sorted on a FACSAria (BD Biosciences) and lysed with RLT Plus buffer (QIAGEN) to examine gene expression by real-time PCR.

Enzyme-linked immunosorbent assay

IL-1β, IL-6, IL-12, IL-27, and TNFα were measured from supernatants of stimulated monocyte cultures, and IL-17 and IFNγ from supernatants of monocyte:CD4+ T cell coculture experiments with Human DuoSet ELISA Kits from R&D Systems, according to the manufacturer’s recommendations. IFNα and IFNβ were measured from stimulated monocyte cultures using the VeriKine Human ELISA kits for IFNα and IFNβ, respectively (PBL Assay Science), using the manufacturer’s recommendations.

Quantification of mRNA expression levels by real-time PCR

RNA was isolated using the QIAGEN RNeasy Micro Kit (QIAGEN) or the ZR-96 Quick RNA (Zymo Research, Irvine, CA) following manufacturer′s guidelines and converted to complementary DNA by reverse transcription (RT) with random hexamers and Multiscribe RT (TaqMan Reverse Transcription Reagents; Thermo Fisher Scientific). For mRNA gene expression assays, probes were obtained from Thermo Fisher Scientific, and the reactions were set up following the manufacturer's guidelines and run on a StepOnePlus Real-Time PCR System (Thermo Fisher Scientific). Values are represented as the difference in cycle threshold (Ct) values normalized to β2-microglobulin for each sample as per the following formula: Relative RNA expression = (2—ΔCt) × 1000. Gene expression data were arranged in heat maps in Fig. 1 to identify virus-specific cytokine expression patterns. Colors are assigned using linear conditional formatting based on relative expression values of each gene for each donor and virus infection as compared to the expression value of that gene for all other donors and virus infections. Red denotes high expression and blue low expression.

For mRNA expression assays, the following probes were used: ATF1 Hs00909673_m1, ATF2 Hs01095345_m1, ATF3 Hs00231069_m1, ATF4 Hs00909569_g1, ATF5 Hs01119208_m1, ATF6 Hs00232586_m1, ATF7 Hs00232499_m1, B2M Hs00187842_m1, BST2 Hs00171632_m1, CCL20 Hs00171125_m1, CD40 Hs01002913_g1, CD74 Hs00269961_m1, CFOS Hs04194186_s1, CXCL10 Hs00171042_m1, CXCL8 Hs99999034_m1, EBI3 Hs01057148_m1, EIF2AK2 Hs00169345_m1, FOS Hs04194186_s1, FosB Hs00171851_m1, FOSL1 Hs04187685_m1, FOSL2 Hs01050117_m1, FOXP3 Hs01085834_m1, GATA3 Hs00231122_m1, GZMB Hs01554355_m1, ICOSLG Hs00323621_m1, IFITM1 Hs00705137_s1, IFITM2 Hs00829485_sH, IFNA1 Hs00855471_g1, IFNA2 Hs00265051_s1, IFNB1 Hs01077958_s1, IFNG Hs00989291_m1, IL10 Hs001174086_m1, IL12B Hs01011519_m1, IL1B Hs99999029_m1, IL17A Hs00174383_m1, IL21 Hs00222327_m1, IL23A Hs00372324_m1, IL27 Hs00377366_m1, IL4 Hs00174122_m1, IL6 Hs00174131_m1, IRF1 Hs009171965_m1, IRF3 Hs01547283_m1, IRF5 Hs00158114_m1, IRF7 Hs00185375_m1, IRF9 Hs00196051_m1, ISG15 Hs00192713_m1, JDP2 Hs00185689_m1, JUN Hs01103582_s1, JUNB Hs00357891_s1, JUND Hs04187679_s1, MOV10 Hs00253093_m1, MX2 Hs01550808_m1, RORC Hs01076122_m1, TBX21 Hs00203436_m1, TGFB1 Hs00998133_m1, TLR1 Hs00413978_m1, TLR2 Hs01014511_m1, TLR3 Hs01551078_m1, TLR4 Hs00152939_m1, TLR5 Hs00152825_m1, TLR6 Hs00271977_m1, TLR7 Hs00152971_m1, TLR8 Hs00152972_m1, TLR9 Hs00152973_m1, TLR10 Hs01675179_m1, TNF Hs00174128_m1, TREX1 Hs03989617_s1, TRIM5 Hs01552559_m1.

Flow cytometry staining

For cell surface staining, monocytes were stimulated with IMQ (5 μg/ml) or ssRNA40 (5 μg/ml) for 12 hours and CD40 and ICOSL were detected with anti-human CD40 (Clone 5C3) and anti-human ICOSL (clone 2D3), from BioLegend. For intracellular staining, monocytes were stimulated as above in the presence of GolgiStop (BD Biosciences) for 8 hours at 37°C. For both surface and intracellular staining, cells were exposed to LIVE/DEAD viability dye before antibody staining, and a subsequent Fc receptor (FcR) blocking (FcR blocking reagent, human, Miltenyi Biotec) was performed before fixation and the addition of antibodies. Cells were fixed using the Foxp3 staining buffers (Thermo Fisher Scientific) following the manufacturer’s recommendations. For intracellular cytokine staining, the following anti-human antibodies were used: IL-1β (clone AS10), IL-12 (clone C11.5), and TNFα (clone Mab11) from BD Biosciences and IL-23 p19 (clone 23dcdp) from eBioscience.

For flow cytometry phosphorylation assays, isolated monocytes were left at 37°C overnight in complete medium to rest. Cells were stimulated with IMQ (5 μg/ml) or ssRNA40 (5 μg/ml) and fixed at various time points with Fixation buffer (BD Biosciences), permeabilized with Perm Buffer III (BD Biosciences) according to the manufacturer’s instructions, and stained with anti-human phospho-p38 (Thr180/Tyr182), phospho-MEK1/2 (Ser218/Ser222), phospho-ERK1/2 (Thr202/Tyr204), phospho-p65 (Ser529), and phospho-TBK1 (Ser172) from BD Biosciences and anti-human phospho-MKK3/6 (Ser218), phospho-MKK4/7 (Ser271/Thr275), and phospho-IRF3 (Ser386) from Bioss Antibodies. Samples were run on a BD Fortessa cytometer, and data were analyzed using FlowJo software (TreeStar).

Gene silencing by small interfering RNA

Gene silencing on CD14+ monocytes was performed as previously published (112) with some modifications. Briefly, monocytes were resuspended at 1.5 × 106 cells/ml in complete medium and transfected with 200 mM small interfering RNA (siGENOME SMARTpool, Dharmacon) for TLR7, TLR8, ATF1, ATF2, FOSL1, or FOSL2 using HiPerFect (QIAGEN). 36 hours after transfection, some cells were lysed to confirm target gene silencing, and the rest of monocytes were harvested, plated at 5 × 104 cells per well in 96-well plates and stimulated with IMQ or ssRNA40 at 5 μg/ml for 16 hours. Supernatant was collected for ELISA determination of cytokines, and cells were lysed with RLT Plus buffer (QIAGEN) for gene expression analysis.

Measurement of intracellular calcium

Monocytes were labeled ex vivo with 5 μM Indo-1 AM (Thermo Fisher Scientific) in phosphate-buffered saline (PBS) for 45 min at 37°C. Cells were washed to remove excess of Indo-1 AM, resuspended at 106 cells/ml in PBS, and rested for 30 min at 37°C. When necessary, IRS661 or IRS control was added at 2.5 μM during the 30 min of resting. Xestopongin C (Sigma-Aldrich) and EGTA were added during the 30 min of resting at 1 μM (113, 114) and 2 mM (115), respectively. The samples were acquired on a BD Fortessa for approximately 2 min initially to determine basal levels of Ca2+. Subsequently, IMQ, ssRNA40, or ionomycin was added to the cells at the appropriate concentrations. Samples were acquired for approximately eight more minutes to determine the increase in intracellular Ca2+ concentration. Data were analyzed with FlowJo software.

Statistical analysis

GraphPad Prism software was used for statistical analysis.

SUPPLEMENTARY MATERIALS

stke.sciencemag.org/cgi/content/full/12/605/eaaw1347/DC1

Fig. S1. Expression of TLR7 and TLR8 by human CD14+ monocytes.

Fig. S2. TLR7 and TLR8 antagonists do not alter monocyte cytokine expression.

Fig. S3. TLR7 and TLR8 blockade alters monocyte cytokine expression after viral infection.

Fig. S4. TLR7 and TLR8 knockdown alters monocyte cytokine expression after viral infection.

Fig. S5. TLR8 promotes monocyte type I IFN response after viral infection.

Fig. S6. TLR7 activation of human monocytes promotes TH17 polarization of CD4+ T cells.

Fig. S7. Cytokine expression by TLR7-stimulated monocytes is PI3K independent.

Fig. S8. Redundant roles of ATF1, ATF2, and FOSL2 in inducing TH17-type cytokines after TLR7 stimulation.

Fig. S9. RNA viruses induce phosphorylation of IRF3 through TLR7 and TLR8 stimulation.

Fig. S10. TLR7 and TLR8 stimulation activate IRF5.

Fig. S11. Kinetics of IRF3 and IRF7 expression after TLR7 and TLR8 stimulation.

Fig. S12. TLR7-dependent inhibition of cytokine expression and type I IFN response is NFAT independent.

Fig. S13. Differences between TLR7 and TLR8 signaling in human CD14+ monocytes.

REFERENCES AND NOTES

Acknowledgments: We would like to thank C. Cotsapas for critical review of the statistical methods used in this manuscript and A. Iwasaki’s and A. Garcia-Sastre’s laboratories for providing reagents. Funding: This work was supported by funding from Yale and Imperial College (to M.D.-V.). Author contributions: M.d.M., K.D., A.C.D., A.S.G., and M.D.-V. performed the experiments. M.d.M., A.C.D., A.S.G., and M.D.-V. analyzed the data. M.d.M. and M.D.-V. designed the experiments and wrote the manuscript. M.D.-V. designed and supervised the work. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the manuscript or the Supplementary Materials.
View Abstract

Stay Connected to Science Signaling

Navigate This Article