Macromolecular Synthesis Shutoff Resistance by Myeloid Cells Is Critical to IRF7-Dependent Systemic Interferon Alpha/Beta Induction after Alphavirus Infection

Most previous research exploring the interaction of alphaviruses with host cell antiviral responses has been conducted using fibroblast lineage cell lines. Previous studies have led to the discovery of virus-mediated activities that antagonize host cell antiviral defense pathways, such as host cell translation and transcription inhibition and suppression of STAT1 signaling. However, their relevance and impact upon myeloid lineage cell types, which are key responders during the initial stages of alphavirus infection in vivo, have not been well studied. Here, we demonstrate the different abilities of myeloid cells to resist VEEV infection compared to nonmyeloid cell types and begin to elucidate the mechanisms by which host antiviral responses are upregulated in myeloid cells despite the actions of virus-encoded antagonists.

ABSTRACT

Alphavirus infection of fibroblastic cell types in vitro inhibits host cell translation and transcription, leading to suppression of interferon alpha/beta (IFN-α/β) production. However, the effect of infection upon myeloid cells, which are often the first cells encountered by alphaviruses in vivo, is unclear. Previous studies demonstrated an association of systemic IFN-α/β production with myeloid cell infection efficiency. Murine infection with wild-type Venezuelan equine encephalitis virus (VEEV), a highly myeloid-cell-tropic alphavirus, results in secretion of very high systemic levels of IFN-α/β, suggesting that stress responses in responding cells are active. Here, we infected myeloid cell cultures with VEEV to identify the cellular source of IFN-α/β, the timing and extent of translation and/or transcription inhibition in infected cells, and the transcription factors responsible for IFN-α/β induction. In contrast to fibroblast infection, myeloid cell cultures infected with VEEV secreted IFN-α/β that increased until cell death was observed. VEEV inhibited translation in most cells early after infection (<6 h postinfection [p.i.]), while transcription inhibition occurred later (>6 h p.i.). Furthermore, the interferon regulatory factor 7 (IRF7), but not IRF3, transcription factor was critical for IFN-α/β induction in vitro and in sera of mice. We identified a subset of infected Raw 264.7 myeloid cells that resisted VEEV-induced translation inhibition and secreted IFN-α/β despite virus infection. However, in the absence of IFN receptor signaling, the size of this cell population was diminished. These results indicate that IFN-α/β induction in vivo is IRF7 dependent and arises in part from a subset of myeloid cells that are resistant, in an IFN-α/β-dependent manner, to VEEV-induced macromolecular synthesis inhibition.

IMPORTANCE Most previous research exploring the interaction of alphaviruses with host cell antiviral responses has been conducted using fibroblast lineage cell lines. Previous studies have led to the discovery of virus-mediated activities that antagonize host cell antiviral defense pathways, such as host cell translation and transcription inhibition and suppression of STAT1 signaling. However, their relevance and impact upon myeloid lineage cell types, which are key responders during the initial stages of alphavirus infection in vivo, have not been well studied. Here, we demonstrate the different abilities of myeloid cells to resist VEEV infection compared to nonmyeloid cell types and begin to elucidate the mechanisms by which host antiviral responses are upregulated in myeloid cells despite the actions of virus-encoded antagonists.

INTRODUCTION

Alphaviruses are mosquito-vectored, positive-sense, single-stranded RNA viruses in the family Togaviridae and are broadly classified as either arthritogenic Old World alphaviruses (e.g., Sindbis virus [SINV], Ross River virus [RRV], and chikungunya virus [CHIKV]) or encephalitic New World alphaviruses (e.g., eastern equine encephalitis virus [EEEV] and Venezuelan equine encephalitis virus [VEEV]). Arthritogenic alphavirus infection causes a febrile illness leading to arthralgia/arthritis that can potentially last for months to years after primary infection (1), whereas infection with encephalitic alphaviruses can progress to fatal encephalitis in a significant number of cases ranging from 0.1 to 1% with VEEV to 30 to 70% with EEEV (2, 3).

During infection of humans and rodent models with alphaviruses, as with many arboviruses, subcutaneous deposition of virions can lead to infection of skin-resident and infiltrating myeloid-lineage cells, such as dendritic cells, macrophages, and Langerhans cells, which facilitate virus spread to regional draining lymph nodes, where a primary initial site of viral infection is established (4, 5). The course of arbovirus infection is significantly shaped by the interactions with myeloid cells, and a particular virus’ ability to exploit this interaction partly explains the virulences of different arboviruses in vivo (2). For example, the translation and replication of EEEV genomes in myeloid cells is suppressed by binding of the hematopoietic-cell-specific microRNA miR142-3p to specific sites in the EEEV 3′ untranslated region. This prevents the induction of systemic innate antiviral immune responses (including interferon alpha/beta [IFN-α/β]), allowing the virus to seed sites of replication apart from the inoculation site, and results in severe encephalitis in murine models and humans (6). Studies using EEEV mutants have demonstrated a strong association between levels of myeloid cell infection and systemic IFN-α/β production in vivo (6, 7). In contrast, very high levels of systemic IFN-α/β and other proinflammatory cytokines, such as interleukin 12 (IL-12), tumor necrosis factor alpha (TNF-α), MIG, and monocyte chemoattractant protein 1 (MCP-1) (8), are secreted by myeloid cells following VEEV infection of lymphoid tissue draining the infection site. The production of systemic IFN-α/β upregulates the expression of antiviral proteins and primes distant tissues against viral replication (2, 6, 7, 9–11), possibly limiting the severity of VEEV infection in humans, for example, in comparison with EEEV. These results suggest a direct association between myeloid cell infection efficiency in vivo and systemic serum IFN-α/β and proinflammatory cytokine levels. However, production of IFN-α/β by uninfected cells in lymphoid tissue has also been proposed (12, 13).

Studies with arthritogenic alphaviruses indicate that IFN-α/β produced by the activation of interferon regulatory factor 3 (IRF3) and the similarly acting but inducible IRF7 transcription factor and, specifically, systemic IFN-α/β production by monocytes and other myeloid cells can control virus replication and protect mice from mortality (14–18). As IRF7 can be constitutively expressed in myeloid lineage cells, such as macrophages and plasmacytoid dendritic cells (pDCs) (19–22), it is likely that this transcription factor plays a critical role in inducing IFN-α/β responses in these cells in vivo and in vitro following alphavirus infection. However, the role of IRF3 or IRF7 in IFN-α/β induction from myeloid cells or mediating protection during encephalitic alphavirus infection has not been explored.

In fibroblasts and other nonmyeloid cells, alphaviruses block IFN-α/β induction by efficiently inhibiting host macromolecular synthesis (specifically, translation and transcription) to the point where little to no IFN-α/β protein is detected in infected cell supernatants (23–28). SINV infection of fibroblast lineage cells activates the dimerization and nuclear translocation of IRF3, which subsequently fails to elicit transcription of IFN-α/β or antiviral effector genes due to virus nonstructural protein 2 (nsP2)-mediated inhibition of cellular transcription (26–28). VEEV-infected fibroblasts and neurons produce IFN-α/β mRNA but do not secrete IFN-α/β protein due to VEEV nsP2-mediated inhibition of translation (26–28). VEEV infection of myeloid cells in vitro, however, has shown production of high levels of IFN-α/β (7). Arthritogenic alphaviruses such as SINV and RRV also stimulate IFN-α/β production from myeloid cell cultures (29, 30). However, the impact of alphavirus macromolecular synthesis inhibition on IFN-α/β production by myeloid cells, which is likely critical to systemic IFN-α/β production and protection from disseminated disease, is unknown.

Here, we have studied the interaction of VEEV with myeloid cells to determine the importance of the IRF3 and IRF7 transcription factors for IFN-α/β induction in vitro and in vivo and to examine the characteristics of macromolecular synthesis inhibition and their effects on IFN-α/β production after infection. VEEV was chosen due to the robust tropism of the virus for these cell types in vitro and in vivo and its potential to provide a strong stimulus for myeloid cell IFN-α/β responses (2).

Infection of IRF3^−/− or IRF7^−/− mice with VEEV revealed a critical role for IRF7, but not IRF3, in the upregulation and secretion of systemic IFN-α/β. Similarly, IRF7 was critical for IFN-α/β production in primary myeloid cell cultures. Using cell sorting to group infected Raw 264.7 monocytes/macrophages based on the relative expression of virus proteins, a distinct population of cells was identified that was partially resistant to VEEV replication and translation inhibition and also capable of secreting high levels of IFN-α/β. Furthermore, we identified an infected cell population with higher virus replication, greater macromolecular synthesis inhibition, and significantly muted IFN-α/β responses. Partially resistant cells expressed elevated levels of interferon-stimulated genes (ISGs) and pattern-recognition receptors (PRRs) and/or induced these genes faster than more virus-sensitive cells following VEEV infection, potentially explaining their virus-resistant phenotype. Additionally, VEEV infection of primary bone marrow macrophages (BMMϕ) generated from wild-type (WT) or interferon α/β receptor^−/− (IFNAR^−/−) mice indicated that IFN-α/β signaling is involved in resistance of some cell subsets to macromolecular synthesis inhibition. From these data, we conclude that a subset of infected myeloid cells exhibits IFN signaling-dependent suppression of virus replication and macromolecular synthesis inhibition, allowing IRF7-dependent IFN-α/β secretion and systemic IFN production following infection.

RESULTS

IRF7 is critical for systemic IFN induction in vivo.

To determine the antiviral roles of various IFN-α/β pathway factors against VEEV infection, we infected mice lacking IFNAR1 (AB6), IRF3, or IRF7 with WT VEEV (Fig. 1A and B). Similar to previously published studies (31, 32), IFNAR^−/− mice were most susceptible to VEEV infection (average survival time [AST], 1.5 versus 5.8 days for the WT; P < 0.001), highlighting the essential role played by the IFN-α/β system in controlling VEEV infection in vivo. The AST for infected IRF7^−/− mice was decreased significantly, by ∼3 days (P ≤ 0.001) compared to the WT, whereas the reduction in AST in IRF3^−/− mice was more modest (∼1 day; P ≤ 0.001), which suggests IRF7 plays an important role in suppressing VEEV infection. Similarly, viral titers in tissues of IRF7^−/− mice were universally and significantly higher (P < 0.001) than titers in WT mice, whereas VEEV replication in IRF3^−/− mice was not significantly different in the majority of tissues from that in WT mice (Fig. 1B). The role of IRF7 in controlling VEEV infection in vivo was further highlighted when levels of systemic IFN-α/β were measured. Induction of systemic IFN-α/β as measured in serum was significantly dependent on IRF7, but not IRF3 (Fig. 2A). IFN-α/β induction in WT mice following VEEV infection was very robust (>10,000 IU/ml at 24 h postinfection [p.i.]), consistent with previous reports (2, 9), whereas IRF7^−/− mice secreted little to no serum IFN-α/β (50 to 200 IU/ml at 24 h p.i.). Serum IFN-α/β levels in IRF3^−/− mice were not significantly different from levels in WT mice through 24 h p.i., although detectable IFN-α/β was significantly lower later in infection as serum levels waned (data not shown). Taken together, these data suggest that IRF7 is important for control of VEEV replication in vivo, potentially through a critical activity of IRF7 in induction of systemic IFN-α/β.

FIG 1.

IRF7 protects mice from VEEV infection. (A) BL6, AB6, IRF3^−/−, and IRF7^−/− mice were infected subcutaneously in the hind leg footpad with 1,000 PFU of WT VEEV, and the AST was determined (4 mice per group). *, P < 0.001 using a Mantel-Cox log rank test. The data are representative of two independent experiments. d, days. (B) BL6, IRF3^−/−, and IRF7^−/− mice were infected subcutaneously in the hind leg footpad with 1,000 PFU of WT VEEV, and the indicated tissues were harvested at 24 h p.i. (6 mice per group). Viral titers were determined using plaque assays as described in Materials and Methods. ****, P < 0.0001; **, P < 0.001; *, P < 0.01; ns, not significant using two-way ANOVA. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

FIG 2.

IRF7 is important for IFN induction following VEEV infection both in vivo and in vitro. (A) BL6, AB6, IRF3^−/−, and IRF7^−/− mice were infected subcutaneously in the hind leg footpad with 1,000 PFU of WT VEEV (4 mice per group). Serum was collected at the indicated times p.i. for IFN bioassay as described in Materials and Methods. ****, P < 0.001; ns, not significant using two-way ANOVA. The error bars indicate standard deviations. The data are representative of the results of two independent experiments. (B) Primary 3-day macrophages and conventional dendritic cells were infected in triplicate with WT VEEV (MOI = 1), and supernatants were collected for IFN bioassay at the indicated times p.i. as described in Materials and Methods. ****, P < 0.001; *, P < 0.04; ns, not significant by t test. The error bars indicate standard deviations. The data are representative of the results of two independent experiments. (C) Primary 3-day macrophages and conventional dendritic cells were infected in triplicate with WT VEEV (MOI = 1), and the lysates were collected at the indicated times p.i. RT-PCR for positive-strand levels was performed as described in Materials and Methods. L.O.D., limit of detection.

IRF7 is important for IFN-α/β induction in primary macrophages and dendritic cells.

To further evaluate the role of IRF7 in the induction of IFN-α/β, we separately cultured primary macrophages and conventional dendritic cells (cDCs) from WT, IRF3^−/−, and IRF7^−/− mice and infected them with VEEV (Fig. 2B). These cells were used because similar cells are initially infected by VEEV in vivo during the initial rounds of infection (4, 5). Similar to mouse serum data, levels of secreted IFN-α/β from IRF7^−/− macrophages (P < 0.0001) and dendritic cells (P < 0.04) were significantly lower than from WT cells, highlighting the importance of the transcription factor in IFN-α/β induction by VEEV. In contrast, IFN-α/β levels from macrophages and dendritic cells lacking IRF3 were higher than those from WT cells, possibly due to enhanced virus replication (Fig. 2C) as a consequence of limited direct IRF3-mediated antiviral gene induction (33–36). Therefore, primary myeloid cells produce IFN-α/β following VEEV infection primarily through an IRF7-mediated signaling pathway.

Induction of IFN-α/β occurs despite rapid translation inhibition by VEEV.

Alphaviruses have been shown to efficiently replicate and inhibit host translation and transcription in nonmyeloid cell lines (murine embryonic fibroblast [MEF], BHK, etc.) (26, 28), activities that successfully block IFN-α/β induction following alphavirus infection in those cell types (27). We used Raw 264.7 cells (a mouse monocyte/macrophage line) to model VEEV infection in myeloid cells, as they are more susceptible to infection than primary macrophages, support efficient VEEV replication following infection (7), and are similarly capable of secreting IFN-α/β following infection with various alphaviruses (12). To study the interplay between VEEV-mediated host translation and transcription inhibition and IFN-α/β induction in myeloid cells, we measured levels of IFN-β mRNA and biologically active IFN-α/β produced following infection of Raw 264.7 cells (Fig. 3) with VEEV at a high multiplicity of infection (MOI) (MOI = 10), and tested the ability of VEEV to inhibit host translation and transcription in the cells (Fig. 4).

FIG 3.

Raw 264.7 cells, but not MEFs, successfully secrete IFN following VEEV infection. Raw 264.7 cells were infected in triplicate with WT VEEV (MOI = 10). The lysates and supernatant were collected at the indicated times p.i. RT-PCR (A) and IFN bioassays (B) were performed as described in Materials and Methods to determine IFN-β mRNA and secreted IFN levels. Infection at each time point was performed in duplicate. ****, P < 0.0001; **, P < 0.01 using two-way ANOVA. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

FIG 4.

Host translation, but not transcription, is inhibited early in Raw 264.7 cells following WT VEEV infection. Raw 264.7 cells were infected with WT VEEV (MOI = 10). (A) Lysates were collected in triplicate, and RNA was harvested for RT-PCR for gamma actin intron 3, performed as described in Materials and Methods. ****, P < 0.0001; ns, not significant using two-way ANOVA. The data are averages of two replicates. The error bars indicate standard deviations. (B) Cells were labeled in duplicate with 100 μCi/ml of [³⁵S]Cys/Met for 1 h at the indicated times p.i. The lysates were collected, resolved on SDS-PAGE gels, and visualized as described in Materials and Methods. (C) Densitometry was performed on the gels. ****, P < 0.0001; ns, not significant using two-way ANOVA. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

Levels of IFN-β mRNA in both MEFs and Raw 264.7 cells increased significantly (P < 0.0001) versus uninfected controls throughout the course of infection (Fig. 3A), possibly reflecting the relatively slow (e.g., versus SINV) induction of host transcription inhibition previously observed during VEEV infection (26, 28). In contrast, and consistent with previous reports (27), IFN-α/β secretion from MEFs was completely blocked. However, secretion of large amounts of IFN-α/β into the cell supernatant from infected Raw 264.7 cells was observed, which increased during the observation period of 18 h (Fig. 3B).

We further evaluated VEEV-mediated host transcription inhibition in Raw 264.7 cells using a previously described reverse transcription (RT)-PCR assay that measures levels of gamma actin intron 3 as an indicator of RNA polymerase (Pol) II transcriptional activity (28). VEEV successfully inhibited general host cell transcription in MEFs by 6 to 9 h p.i. (P < 0.0001), whereas transcription was not significantly blocked (P < 0.0001) in Raw 264.7 cells until 12 to 15 h p.i. (Fig. 4A). In contrast, use of ³⁵S pulse-chase analysis to measure levels of newly synthesized proteins in Raw 264.7 cells (Fig. 4B and C) revealed the ability of VEEV to rapidly and significantly (P < 0.0001) inhibit translation in most cells by 3 to 6 h p.i. Translation inhibition in MEFs was slightly delayed to 6 to 9 h p.i. (P < 0.0001). These data suggest that VEEV-mediated translation inhibition is more prominent early after infection of myeloid cells than transcription inhibition. However, in the infected cell population measured as a whole, neither transcription inhibition nor translation inhibition led to blockade of IFN-α/β secretion. Notably, translation inhibition was not absolute in the cultures, suggesting that IFN-α/β was either produced by poorly or uninfected cells or as part of low-level production of host proteins in infected cells in which host translation was efficiently inhibited but unknown mechanisms resulted in IFN-α/β production.

Most IFN-α/β is secreted by cells infected with actively replicating virus.

To identify the cellular source of IFN-α/β protein during VEEV infection, we attempted to define infected versus uninfected cell populations through use of green fluorescent protein (GFP)-expressing VEEV replicons (VREP GFP), which are not capable of spreading to neighboring cells following an initial round of infection (5). In addition, the replicons lack the capsid protein, which with New World alphaviruses has been implicated in host cell transcription inhibition (37), but do express the nsP2 protein, which inhibits translation (28). Raw 264.7 cells were infected with VREP GFP at an MOI (<1), at which ∼50% of the cells were positively infected, based on flow-cytometric analysis (Fig. 5). Cells were sorted into GFP-negative/low (−/low) and GFP-positive (+) populations (Fig. 5A), following which IFN-α/β mRNA and protein production and transcription and translation inhibition were measured (Fig. 5C and D). These two populations represented roughly equal percentages of the total culture (57% and 42%, respectively) (Fig. 5A). Upregulation of both IFN-β and IFN-α4 mRNAs were highest in VREP GFP⁺ cells (∼50-fold over mock-infected cells versus ∼5-fold for VREP GFP^−/low cells; P < 0.01). Additionally, the large majority of measurable IFN-α/β protein produced by sorted cells between 0 and 2.5 h postsorting was found in cells with actively replicating replicon genomes that produced clearly detectable GFP (a 4-fold difference between GFP^−/low and GFP⁺ populations; P < 0.0001). Due to the limit of detection of the cytometer, we were unable to completely separate GFP^low cells and truly GFP⁻ cells; therefore, the small amount of IFN-α/β observed in this population may have originated from GFP^low and not uninfected cells. Indeed, quantitative-PCR analysis for positive-sense RNA in the GFP⁺ cells revealed the presence of VEEV genomes, indicating either low levels of genome replication (Fig. 5B) or remnants of the original inoculum. These results suggest that monocytes/macrophages, at least in culture, primarily secrete IFN-α/β upon detection of an active infection. Furthermore, as expected, we observed no reduction in transcription levels (Fig. 5D) in VREP GFP-infected cells due to the absence of the viral capsid protein, whereas VREP GFP efficiently, but not absolutely, inhibited translation in GFP⁺ cells compared to GFP^−/low cells (P < 0.0001). This indicates that in the absence of the transcription-inhibiting capsid protein, significant translation suppression occurs in myeloid cells. These cells are capable of detectable IFN production even in the face of translation inhibition by nsP2; however, the suppressive effect of translation inhibition on IFN-α/β secretion (e.g., the impact of nsP2 expression) is not measured with this experimental design.

FIG 5.

IFN secretion from Raw 264.7 cells infected with VEEV replicon occurs primarily from positively infected cells. Raw 264.7 cells were infected with VEEV replicon expressing GFP at an MOI at which ∼50% of the cells were positively infected for 8 h. (A) Infected cells were sorted by level of expressed GFP into GFP^−/low and GFP⁺ populations as described in Materials and Methods. (B) RT-PCR for positive-strand levels was performed as described in Materials and Methods. ****, P < 0.0001; ***, P < 0.0008 using a t test. (C) RNA and supernatants were collected in triplicate for RT-PCR and IFN bioassay as described in Materials and methods. ****, P < 0.0001; **, P < 0.01; *, P < 0.035 using a t test. (D) RT-PCR was performed on sorted cells. The sorted cells were labeled with 100 μCi/ml of [³⁵S]Cys/Met for 1 h. The lysates were collected, resolved on SDS-PAGE gels, and visualized as described in Materials and Methods. Densitometry was performed on the gels. ****, P < 0.0001 using a t test. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

Infection at a high MOI reveals a subset of cells resistant to translation inhibition that secrete IFN-α/β following VEEV infection.

The above-mentioned data suggested that a population of infected cells with actively replicating viral genomes was capable of secreting IFN-α/β despite induction of virus-mediated host translation inhibition. However, the relationship between the extent of virus-mediated translation inhibition and the amount of IFN-α/β secreted from a given cell was not measured directly. As the intensity of GFP signal in the VREP GFP⁺ population was observed as a continuum of increasing expression levels (Fig. 5A), the IFN-α/β produced by the population may have been secreted exclusively by cells with lower GFP expression, which presumably resisted virus-mediated translation inhibition to a higher degree than cells with higher GFP signal intensity. Thus, we hypothesized that IFN-α/β production from a given infected cell would inversely correlate with the level of viral replication and, therefore, with the extent of virus-mediated translation inhibition. It is likely that the majority of VREP^−/low cells were not infected with a replicon particle due to the low MOI used (MOI < 1). Presumably, infecting the cells with a replicating WT VEEV at an MOI at which all the cells would be exposed to infecting virus particles would replicate the VREP GFP⁺ population and allow analysis of the range of myeloid cell responses to virus exposure.

To this end, we infected Raw 264.7 cells with WT VEEV expressing the MeDF reporter protein at a BHK cell MOI of 4, designed to expose all the cells to infecting virions (∼98% by Poisson distribution at an MOI of 4 [38]), and used cell sorting to separate the virus-exposed cells into three populations: MeDF^−/low, MeDF^medium, and MeDF^high (Fig. 6A). We sorted the cells into three populations to represent points along a continuum of virus gene expression in order to demonstrate the differential resistance of individual virus-exposed myeloid cells in a given population to VEEV-mediated macromolecular synthesis inhibition and to measure its effect on IFN-α/β mRNA and protein production. The MeDF reporter is expressed as a fusion with the viral structural proteins; therefore, the MeDF signal intensity directly reflects viral protein expression levels (39). At 8 h p.i., which was chosen to minimize exposure of cells to newly replicated virus, −/low cells were 42.6% of the total, medium cells were 22.2% of the total, and high cells were 26.2% of the total (Fig. 6A). To confirm that secondary infection by newly produced virions did not affect our analyses, additional infections were performed in the presence or absence of an anti-VEEV antibody (Fig. 7). Only a small difference, similar to interexperiment differences (Fig. 6 and 7), was observed in the percentage of cells in each population in the presence or absence of antibody, suggesting that secondary infection did not play a major role in delineating different populations. Upregulation of IFN-β and IFN-α4 mRNAs (Fig. 6C) was positively associated with increasing viral replication between −/low and medium populations (∼2-fold increase; P < 0.003) but was decreased in cells with the highest level of MeDF reporter expression (∼3-fold reduction between MeDF^medium and MeDF^high; P < 0.0001), possibly due to effects of infection on transcription efficiency. However, IFN mRNA levels were elevated in all populations compared to mock-infected cells (P < 0.0001). Equivalent secretion levels of IFN-α/β between 0 and 2.5 h postsorting were observed from MeDF^−/low and MeDF^medium populations (∼500 IU/ml), whereas IFN-α/β production was dramatically reduced (∼50 to 60 IU/ml; P < 0.0001 compared to MeDF^−/low and MeDF^medium) in MeDF^high cells (Fig. 6C). Quantitative-PCR analyses of these samples for virus genomes again revealed the presence of positive-sense genome RNA in all the populations (Fig. 6B). Positive-sense genome measurements reflected MeDF expression, with greater genome abundance in high versus medium populations (a trend was present that was not significant [P = 0.1065]) and significantly greater genome abundance in medium versus low (P < 0.001) populations. In all cases, the −/low population yielded significantly more positive-sense genomes than mock-infected cells (P < 0.0001).

FIG 6.

A subset of Raw 264.7 cells resistant to WT VEEV-induced translation inhibition secrete IFN. Raw 264.7 cells were infected with WT VEEV expressing MeDF blue reporter protein (MOI = 4) for 8 h. (A) Infected cells were sorted by level of expressed reporter protein into low, medium, and highly infected populations as described in Materials and Methods. (B) RT-PCR for positive-strand levels was performed as described in Materials and Methods. ****, P < 0.0001; ***, P < 0.001 using a t test. (C) RNA and supernatants were collected in triplicate for RT-PCR and IFN bioassay as described in Materials and Methods. ****, P < 0.0001; ***, P < 0.0005 using one-way ANOVA and a t test. (D) RT-PCR was performed on sorted cells. The sorted cells were labeled with 100 μCi/ml of [³⁵S]Cys/Met for 1 h. The lysates were collected, resolved on SDS-PAGE gels, and visualized as described in Materials and Methods. Densitometry was performed on the gels. ****, P < 0.0001; ns, not significant using one-way ANOVA. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

FIG 7.

Secondary infection does not substantially contribute to virus spread at early times p.i. Raw 264.7 cells were infected with WT VEEV expressing MeDF blue reporter protein (MOI = 7) for 8 h in the presence or absence of anti-VEEV antibody (Ab). Infected cells were sorted by level of expressed reporter protein into low, medium, and highly infected populations as described in Materials and Methods. The percentage of cells in each population is the average of those in three independent infections.

Measurement of gamma actin intron transcription levels did not demonstrate a significant (<50% of mock-infected cells) reduction in any population (Fig. 6D), consistent with previous data (Fig. 4A), where gamma actin intron transcription inhibition was first observed 12 to 15 h p.i. Furthermore, the extent of translation inhibition increased with greater viral replication, with the greatest magnitude of inhibition observed in MeDF^high cells (P < 0.0001 compared to mock-infected cells), whereas no significant inhibition was observed in the MeDF^−/low population (Fig. 6D). Thus, translation inhibition in myeloid cells was proportional to virus gene expression. Taken together, the data suggest that myeloid cells are capable of producing IFN-α/β until a certain level of VEEV replication and host translation inhibition is reached, following which IFN-α/β production is curtailed.

To confirm that the resistance of the MeDF^−/low population to VEEV infection was not due to lower numbers of virus particles entering the cells, we repeated the experiment with a high-MOI infection (MOI = 10) (Fig. 8). This reduced the −/low population to 20.1% of total cells and increased the medium group to 39.6%, while the high percentage remained relatively unchanged at 29.3% (Fig. 8A). Similar to previous observations, levels of IFN-β and IFN-α4 mRNAs (Fig. 8C) increased between MeDF^−/low and MeDF^medium populations (∼2.5 fold) but decreased between MeDF^medium and MeDF^high populations (∼3 fold; P < 0.009). Similarly, MeDF^−/low and MeDF^medium cells secreted IFN-α/β (∼200 IU/ml) after infection, which was significantly reduced (P < 0.015 compared to MeDF^−/low and MeDF^medium) in MeDF^high cells (Fig. 8C). Also, as with the infection at an MOI of 4, positive-sense genomes (Fig. 8B) were detected in all infected cell populations, with medium group levels significantly higher than −/low group levels (P < 0.02) and high group levels significantly above the medium group levels (P < 0.02). In addition, levels of transcription were not significantly (<50% of mock-infected cells) reduced in any population, whereas, unlike the MeDF^high population, MeDF^−/low and MeDF^medium cells resisted VEEV-mediated translation inhibition (Fig. 8D). Together, these data suggest that a higher MOI increased the number of cells expressing medium levels of virus proteins, possibly as a result of multiple-particle infection of somewhat susceptible cells in the −/low population. This suggests that resistance to translation inhibition and the IFN-α/β secretion capabilities of medium and −/low cells are largely intrinsic limitations in virus replication and suppression of macromolecular synthesis versus the high population rather than a simple difference in particle entry efficiency.

FIG 8.

Resistance of a subset of Raw 264.7 cells to WT VEEV is independent of the MOI used for infection. Raw 264.7 cells were infected with WT VEEV expressing MeDF blue reporter protein at a high MOI (MOI = 10) for 8 h. (A) Infected cells were sorted by level of expressed reporter protein into low, medium, and highly infected populations as described in Materials and Methods. (B) RT-PCR for positive-strand levels was performed as described in Materials and Methods. ***, P < 0.0008; **, P < 0.004; *, P < 0.02 using a t test. (C) RNA and supernatants were collected in triplicate for RT-PCR and IFN bioassay as described in Materials and Methods. ****, P < 0.0001; ***, P < 0.001; **, P < 0.009; *, P < 0.015; ns, not significant using one-way ANOVA and a t test. (D) RT-PCR was performed on sorted cells. The sorted cells were labeled with 100 μCi/ml of [³⁵S]Cys/Met for 1 h. The lysates were collected, resolved on SDS-PAGE gels, and visualized as described in Materials and Methods. Densitometry was performed on the gels. ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; ns, not significant using one-way ANOVA. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

Virus replication and translation inhibition increase in resistant cells over time.

We sought to determine whether resistance of the MeDF^−/low population was maintained for long incubation periods following infection or if IFN-α/β secretion was eventually blocked by VEEV-mediated translation inhibition in these cells. Measurement of IFN-α/β secretion from Raw 264.7 cells infected with VEEV at a high MOI (MOI = 10) (Fig. 3B) had previously demonstrated the capacity of at least some cells to secrete IFN-α/β for long periods following infection. We sorted cells infected with VEEV (MOI = 4) after at least 16 h of infection and measured IFN-α/β secretion and transcription and translation inhibition (Fig. 9A to C). The MeDF^−/low population decreased as a percentage of total cells at 16 h (∼12%) compared to 8 h (∼43%), suggesting that VEEV can gradually overcome the resistance in some of these cells either by continuing replication in previously infected cells, by the effects of secondary infection by newly produced virions, or by virion entry into naive cells, the latter an unlikely event considering the multiplicity of infection. Indeed, infecting cells with VEEV in the presence or absence of an anti-VEEV antibody for 16 h (Fig. 10) did not greatly affect the number of cells in each population, suggesting that the spread of newly produced virions did not play a major role in overcoming the resistance of myeloid cells to VEEV infection. Levels of IFN-β and IFN-α4 mRNAs (Fig. 9B) were dramatically reduced compared to 8 h p.i. in all populations (∼4- to 8-fold over mock-infected cells versus ∼20- to 60-fold at 8 h p.i.), suggesting that, similar to other data (Fig. 4A), infection downregulates production of IFN-α/β and other mRNAs late during infection. It is also possible that negative-feedback mechanisms reduce the transcription of these genes. Surprisingly, cells in the MeDF^−/low population continued to secrete IFN-α/β protein (∼200 IU/ml) at 16 h p.i., while IFN-α/β production declined significantly in the MeDF^medium (P < 0.02 compared to MeDF^−/low) and MeDF^high (P < 0.009 compared to MeDF^−/low) populations (Fig. 9B). The MeDF^high population exhibited the greatest degree of virus-mediated translation inhibition and secreted little to no IFN-α/β at either early or late times p.i.

FIG 9.

A subset of Raw 264.7 cells is resistant to WT VEEV-induced translation inhibition at late times p.i. Raw 264.7 cells were infected with WT VEEV expressing MeDF blue reporter protein (MOI = 4) for 16 h. (A) Infected cells were sorted by level of expressed reporter protein into low, medium, and highly infected populations as described in Materials and Methods. (B) RNA and supernatants were collected in triplicate for RT-PCR and IFN bioassay as described in Materials and Methods. ***, P < 0.0008; **, P < 0.009; *, P < 0.02; ns, not significant using one-way ANOVA and a t test. (C) RT-PCR was performed on sorted cells. The sorted cells were labeled with 100 μCi/ml of [³⁵S]Cys/Met for 1 h. The lysates were collected, resolved on SDS-PAGE gels, and visualized as described in Materials and Methods. Densitometry was performed on the gels. ****, P < 0.0001; ***, P < 0.001; *, P < 0.03; ns, not significant using one-way ANOVA and a t test. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

FIG 10.

Secondary infection does not substantially contribute to virus spread at late times p.i. Raw 264.7 cells were infected with WT VEEV expressing MeDF blue reporter protein (MOI = 7) for 16 h in the presence or absence of anti-VEEV antibody. The infected cells were sorted by level of expressed reporter protein into low, medium, and highly infected populations as described in Materials and Methods. The percentage of cells in each population is the average of those in three independent infections.

Overall transcription measured by gamma actin intron levels was elevated in all populations compared to mock-infected cells (Fig. 9C); however, a trend of reduced transcription was observed in MeDF^medium (P < 0.03) and MeDF^high (P < 0.515) populations relative to MeDF^−/low cells. At this time, production of IFN-α/β corresponded to relative translation inhibition, with no significant inhibition observed in MeDF^−/low cells while host translation was reduced to about half that observed at 8 h p.i. in MeDF^medium (P < 0.001) and MeDF^high (P < 0.0001) populations (Fig. 9C). These data suggest that, while initial rates of replication and host macromolecular synthesis inhibition may vary between populations with different viral gene expression levels, as infection proceeds, macromolecular synthesis inhibition increases in most infected cells. Furthermore, the MeDF^−/low population, which continuously produces IFN-α/β throughout infection, appeared to represent a subset of Raw 264.7 cells that resists VEEV-mediated translation inhibition and upregulates antiviral pathways despite virus-encoded antagonists of such responses.

IFN-α/β is important for control of VEEV infection in myeloid cells.

The above-mentioned data suggest that cells capable of producing IFN-α/β are initially partially or highly resistant to VEEV replication and macromolecular synthesis shutoff but that at least some of them may become more susceptible over time. Therefore, we assessed the contribution of IFN-α/β receptor signaling to this resistance. Secreted IFN-α/β from infected myeloid cells is capable of signaling through the IFNAR complex in an autocrine or paracrine manner to upregulate antiviral ISGs and resist or suppress infection (40). We infected BMMϕ from wild-type (B6) and IFNAR1^−/− (AB6) mice with VEEV (MOI = 4), sorted the cells into MeDF expression-based populations as described above (Fig. 11), and tested their abilities to secrete IFN-α/β and resist macromolecular synthesis inhibition. We hypothesized that AB6 BMMϕ would be significantly more sensitive to VEEV infection and macromolecular synthesis inhibition. Preliminary studies (not shown) indicated that a large percentage of B6 BMMϕ (>50%) expressed less GFP than the −/low population of infected Raw 264.7 cells even when using a high initial MOI (e.g., MOI = 10). Therefore, AB6 and B6 BMMϕ were sorted into only positive and negative populations (Fig. 11).

FIG 11.

Primary macrophages lacking IFNAR are more sensitive to WT VEEV infection. BMMϕ grown from B6 and AB6 mice were infected with WT VEEV expressing MeDF blue reporter protein (MOI = 4) for 8 h. (A) Infected cells were sorted by level of expressed reporter protein into MeDF-negative and -positive infected populations as described in Materials and Methods. (B) RT-PCR for positive-strand levels was performed as described in Materials and Methods. ****, P < 0.0001; **, P < 0.002 using a t test. (C) RNA and supernatants were collected in triplicate for RT-PCR and IFN bioassay as described in Materials and Methods. ****, P < 0.0001; ***, P < 0.0005; **, P < 0.005; *, P < 0.015; ns, not significant using a t test. (D) RT-PCR was performed on sorted cells. The sorted cells were labeled with 100 μCi/ml of [³⁵S]Cys/Met for 1 h. The lysates were collected, resolved on SDS-PAGE gels, and visualized as described in Materials and Methods. Densitometry was performed on the gels. ****, P < 0.0001; *, P < 0.03 using a t test. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

Compared with B6 cells, we observed a greater number of infected (positive) AB6 cells (41% AB6 versus 26% B6) (Fig. 11A), suggesting that IFN-α/β signaling mediates at least part of the initial resistance to VEEV infection detectable by flow cytometry. IFN-β and IFN-α4 mRNAs were upregulated to generally similar degrees in B6 (∼5- to 10-fold over mock-infected cells; P < 0.005) and AB6⁺ (∼10- to 15-fold over mock-infected cells; P < 0.0005) cells compared to negative populations. However, the positive population of B6 cells secreted significantly greater amounts of IFN-α/β protein 0 to 2.5 h postsorting (∼60 IU/ml; P < 0.0001) compared to the positive population of AB6 cells, from which little to no IFN-α/β (∼5 IU/ml) was released (Fig. 11C). Similarly, IFN-α/β secretion was observed only in B6⁺ cells (∼30 IU/ml; P < 0.015) when sorted later at 16 h p.i. As with Raw264.7 cells, quantitative RT (qRT)-PCR analysis of BMMϕ RNA revealed positive-sense genomic RNA in both MeDF⁻ and MeDF^high expression populations (Fig. 11B). However, levels of positive-sense RNA were approximately 10-fold lower in the B6 BMMϕ than in the middle MeDF-expressing Raw 264.7 cell population and 20-fold less than the high Raw264.7 cell population. Levels of positive-sense RNA were significantly higher in the AB6 versus B6 sorted populations (B6⁻ versus AB6⁻, P < 0.0001; B6⁺ versus AB6⁺, P < 0.002), suggesting greater replication, since the initial virus inocula would have been equal.

Transcription was inhibited (Fig. 11D) to a greater extent in AB6⁺ cells (∼75% reduced versus AB6⁻ cells; P < 0.0001), whereas B6⁺ cells resisted this viral activity to a greater degree (∼50% reduced versus AB6⁻ cells; P < 0.03). Similarly, VEEV blocked translation to a significantly greater extent (P < 0.0001) in infected AB6 cells than in infected B6 cells (Fig. 11D). Finally, BMMϕ generated from IRF 3X5X7^−/− mice, which do not secrete IFN-α/β following VEEV infection, showed phenotypes similar to those of AB6 cells after VEEV infection, exhibiting higher levels of viral gene expression and translation inhibition than B6 cells in the positive population (data not shown). Taken together, these data suggest that IFN-α/β production and/or signaling partially protects WT primary BMMϕ (and likely the −/low population of Raw 264.7 cells) from VEEV infection and macromolecular synthesis inhibition, allowing continued secretion of IFN-α/β.

ISGs are expressed at higher levels in myeloid cells resistant to VEEV infection.

Thus far, we have identified a subset of myeloid cells capable of resisting VEEV-mediated translation inhibition in an IFN-α/β-dependent manner and that secrete IFN-α/β following infection. However, the specific mechanisms underlying IFN-α/β-mediated resistance of some myeloid cells to VEEV infection are unknown. We speculated that this population of resistant cells may express innate immune pattern recognition receptors, transcription factors, and/or ISGs at higher levels when in a resting state than more virus-sensitive myeloid cells, thereby allowing these cells both to detect incoming virions more quickly and to induce a more rapid and robust antiviral response. It is also possible that resistant cells may upregulate ISGs faster than sensitive cells following autocrine or paracrine IFN-α/β secretion and subsequent receptor signaling that may occur after exposure to virions capable of infection, which leads to greater inhibition of viral replication. Due to the difficulty of sorting cells a priori into populations based upon IFN response pathway activity for analysis of preinfection mRNA levels, we utilized relative rates of virus replication at 8 h p.i., as performed in previous experiments, to segregate Raw 264.7 cells (Fig. 12A) and primary BMMϕ (Fig. 12B), followed by immediate quantitative assessment of IFN response pathway component mRNAs.

FIG 12.

ISGs are expressed at higher levels in myeloid cells resistant to VEEV infection. (A) Raw 264.7 cells were infected with WT VEEV expressing MeDF blue reporter protein (MOI = 4) for 8 h, and the infected cells were sorted by level of expressed reporter protein into low, medium, and highly infected populations. The lysates were collected in triplicate for RT-PCR as described in Materials and Methods. ****, P < 0.0001; ***, P < 0.001; **, P < 0.002 using a t test. (B) B6 and AB6 BMMϕ were infected with WT VEEV expressing MeDF blue reporter protein (MOI = 4) for 8 h, and the infected cells were sorted by level of expressed reporter protein into MeDF-negative and -positive infected populations. The lysates were collected in triplicate for RT-PCR as described in Materials and Methods. ****, P < 0.0001; ***, P < 0.0005; **, P < 0.01; ns, not significant using a t test. The error bars indicate standard deviations. The data are representative of the results of two independent experiments.

We measured mRNA levels of ISGs at 8 h p.i. as a proxy for the level of resistance to VEEV in a given myeloid cell prior to infection. We hypothesized that ISG mRNA levels at 8 h p.i. were at least partly dependent on whether a given population of myeloid cells was resistant or sensitive to VEEV infection, with the level of sensitivity determining the extent of translation and transcription inhibition established by VEEV following infection, which would subsequently determine the level of ISG upregulation by 8 h p.i. We assumed that the levels of ISG mRNAs at 8 h p.i. would be a reasonable indicator of the overall resistance of a given cell to inhibition of both virus-induced translation and transcription, since these mRNAs are upregulated by IFN-α/β signaling (41).

Analysis of the mRNA levels of four ISGs in sorted Raw 264.7 populations (Fig. 12A) indicated that the cells most resistant to VEEV infection (MeDF^−/low cells) expressed/upregulated ISGs to a greater extent than MeDF^medium (P < 0.002) or MeDF^high cells. Similarly, ISG mRNAs in MeDF^medium cells were upregulated to a significantly higher level than in MeDF^high cells (P < 0.002). Finally, ISG mRNA levels in primary BMMϕ (Fig. 12B) were uniformly higher in WT B6 BMMϕ than in BMMϕ lacking IFNAR (∼2- to 10-fold higher; P < 0.0005). B6-negative cells expressed the highest levels of ISG mRNAs, which suggested greater resistance to VEEV infection, whereas lower levels of ISG mRNAs in B6⁺ cells likely reflected greater viral replication and virus-mediated macromolecular synthesis inhibition in these cells. In contrast, lower ISG mRNA levels in AB6 cells (P < 0.0001) were consistent with their reduced ability to control viral replication due to the absence of IFN-α/β signaling. Overall, our data indicate that lower levels of viral replication are observed in myeloid cell populations expressing higher ISG mRNA levels at 8 h p.i., depending on IFN-α/β signaling, suggesting that possibly either elevated expression of ISGs prior to infection or faster ISG induction may contribute to the resistance of a subset of myeloid cells in a given population to VEEV infection and its capacity for interferon secretion.

DISCUSSION

Role of IRF7 during VEEV infection.

Infection of murine models with WT VEEV is usually fatal (42, 43) despite robust secretion of IFN-α/β in serum that is uniquely high among the alphaviruses (2, 7). In these studies, we demonstrate the importance of IRF7 for mediating resistance against VEEV in vivo. IRF7^−/− mice succumbed to infection significantly faster than WT mice when infected with VEEV (an ∼3-day reduction in AST), with higher viral loads in all examined tissues. As IFN-α/β induction from nonmyeloid cells is efficiently blocked by VEEV during infection (26, 27) and IRF7 is constitutively expressed only in myeloid lineage cells, such as macrophages and plasmacytoid dendritic cells (21, 22), we surmise that IRF7 provides an alternative signaling pathway for IFN-α/β induction following sensing of incoming virions by PRRs. Consistent with this hypothesis, we observed a dramatic reduction in serum IFN-α/β in VEEV-infected IRF7^−/− mice compared to WT mice and also in primary macrophages and dendritic cells generated from IRF7^−/− mice compared to cells from WT mice. Our data suggest that IRF7 plays a critical role in mediating protection both in vivo and in vitro following VEEV infection by regulating IFN-α/β induction from infected cells.

In contrast, IRF3 does not appear to play a major role in control of VEEV infection early after infection but may be important in viral clearance. IRF3^−/− mice were slightly more sensitive to VEEV infection than WT mice (∼1 day AST reduction; P < 0.001), with similar levels of tissue viral loads and serum IFN-α/β observed early after infection. This suggests that the effects of IRF-3 activation, at least on viral loads, are manifest during the later stages of infection. Interestingly, IFN-α/β production was higher in IRF3^−/− primary macrophages and dendritic cells, reflecting enhanced VEEV replication in those cells, as well as potentially greater compensatory signaling by IRF7. The greater replication and higher IFN induction in IRF3^−/− cells suggest a role for IRF3 in the control of VEEV infection, potentially via direct IRF3-mediated antiviral gene induction (33–36), a role not described for IRF7, as well as the presence of IRF7, which results in greater IFN secretion due to higher levels of viral PAMPs. In contrast, in IRF7^−/− cells, IRF3 can directly upregulate ISGs, which may potentially control VEEV replication without significant IFN induction. In vivo, a systemic IRF7-mediated IFN-induced antiviral state likely results in a lower viral burden in IRF3^−/− mice, whereas IRF7^−/− mice sustain higher levels of viral replication due to the lack of a systemic type I IFN response. The difference between in vitro and in vivo results likely reflects the presence of a systemic IFN response and its effect on various uninfected cells in vivo versus intracellular and local effects of IFN on only virus-exposed myeloid cells in vitro.

Our data are in agreement with observations from CHIKV-infected IRF3- and IRF7-deficient mice. Deletion of both IRF3 and IRF7 turns an avirulent murine CHIKV infection lethal (17), although deletion of only IRF3 or IRF7, in contrast to our observations with VEEV, does not result in a significant difference in mortality or disease phenotype compared to WT mice. Additionally, combined deletion of IRF3 and IRF7 dramatically reduced serum IFN-α/β levels in CHIKV-infected mice, suggesting that these host factors mediate the control of multiple alphaviruses primarily through their upregulation of IFN-α/β (17). Similarly, pDCs require IRF7 for IFN-α/β secretion to control CHIKV infection (13), and IRF7, but not IRF3, has been shown to mediate systemic IFN-α/β induction during West Nile virus infection of mice (44, 45).

During SINV infection of fibroblast lineage cells, IRF3 is phosphorylated, dimerizes, and successfully translocates into the nuclei of infected cells. However, no IFN-α/β or ISG mRNA is transcribed, nor is protein produced, whereas IFN-α/β mRNA is produced, at least initially, after VEEV infection, but no IFN-α/β protein is detected (26, 27). IRF3 translocation occurs before infecting alphaviruses can inhibit translation and/or transcription, though these viral activities play a major role in preventing subsequent IFN-α/β secretion. Since primary myeloid cells expressing IRF7, but not IRF3, successfully secrete IFN-α/β following VEEV infection, we speculate that IRF7-mediated IFN-α/β and ISG induction is either resistant to host cell translation and transcription inhibition or occurs at a faster kinetic pace than similar IRF3-mediated induction, or IRF-3-mediated activation by alphaviruses is defective in these cells. Macrophages have been shown to preload RNA Pol II complexes on the promoters of many rapid-response genes, including ISGs and IFN-β and NF-κB genes (46), which may help explain the disparity in IFN-α/β secretion from myeloid and nonmyeloid cells. Future experiments will explore the molecular mechanisms underpinning the difference between IRF7- and IRF3-mediated IFN-α/β induction in different cell types.

IFN-α/β induction from infected myeloid cells.

We observed transcription of IFN-α/β mRNA but no secretion of IFN-α/β protein following VEEV infection of MEF cultures in which most, if not all, cells were productively infected (consistent with previously published data [26]). Levels of IFN-α/β mRNAs in MEFs cumulatively increased throughout the course of infection despite VEEV capsid-mediated transcription inhibition (37), yet translation was suppressed, apparently preventing IFN-α/β protein secretion, which is similar to observations made using primary neurons infected with VEEV replicons lacking capsid (26).

In contrast to MEFs, primary macrophages, dendritic cells, and the Raw 264.7 monocyte/macrophage cultures successfully secrete IFN-α/β protein following infection with VEEV. Our data demonstrate that IFN-α/β protein production continues and cumulatively increases throughout the course of infection. Still, VEEV is able to rapidly, but not completely, inhibit translation when all the cells are assayed together in an infected Raw 264.7 cell culture, while transcription inhibition is presumably delayed due to the temporally separated synthesis of VEEV nonstructural protein 2 (translation inhibition without transcription inhibition) versus capsid (nuclear pore blockade), consistent with previous reports for fibroblasts (26–28). Incomplete transcription and/or translation inhibition allows some myeloid cells to respond to VEEV infection by secreting IFN-α/β.

Segregation of cells by VEEV relative replication efficiency identified a population of Raw 264.7 cells that resisted VEEV-mediated macromolecular synthesis inhibition and produced IFN-α/β following infection. This included cells negative for GFP expression by flow cytometry and those exhibiting an intermediate level of VEEV replication. However, qRT-PCR analysis revealed that all the cell populations contained virus positive-sense genomic RNA, which was found at different levels in the sorted populations, suggestive of relative virus replication. In contrast, infected cell groups, similar to infected unsegregated cultures, did not exhibit transcription shutoff. The resistance phenotype of the IFN-α/β-producing cells (−/low and medium populations) appeared to be largely independent of the initial MOI used, suggesting that an intrinsic property of these cells mediates initial resistance to VEEV. In contrast, when cells were incubated for longer periods before sorting, fewer cells were observed in the −/low population, while the medium population increased in number and the high population was unchanged. At the later times, when multiple rounds of virion release and infection and continuously increasing replication in initially infected cells would have occurred, translation shutoff resistance and IFN-α/β production were maintained in the −/low population, in contrast to more sensitive cells, in which VEEV gradually increased the magnitude of translation inhibition over time, with a corresponding reduction in IFN-α/β secretion from those cells. Thus, sensitivity of myeloid cells to VEEV infection occurs along a spectrum, with the magnitude and duration of IFN-α/β secretion reflecting the relative resistance to replication and macromolecular synthesis inhibition and some cells permitting increased replication and translation shutoff and others maintaining a resistant phenotype.

One possible mechanism for differential cell sensitivity to VEEV replication is the levels of ISGs and antiviral proteins expressed. Consistent with this, our data indicate higher expression of tested ISGs in cells resistant to VEEV than in more sensitive cells. Higher levels of antiviral proteins, such as PRRs and transcription factors, probably allow faster detection of incoming virus and/or stronger blockade of viral genome replication, which diminishes the impact of viral activities countering the upregulation of systemic antiviral responses in resistant myeloid cells. Such cells are thus more capable of activating systemic antiviral pathways, such as type I IFN. Consistent with our data, a previous study demonstrated the importance of a specific population of inflammatory Ly6C^high CCR2⁺ monocytes in controlling RRV infection in mice (15). Following in vivo RRV infection, these cells expressed elevated levels of IRF7 and were capable of inducing IFN-α/β, which is critical for control of acute RRV infection.

A second mechanism may act through the effect of IFN-α/β secreted from resistant myeloid cells by signaling in an autocrine manner to reduce VEEV replication, and this effect may be more rapid or robust in some cells than others, thereby accounting for the range of sensitivity to replication and macromolecular synthesis inhibition. We show enhanced susceptibility of macrophages lacking IFNAR to VEEV infection in comparison with WT cells associated with dramatically downregulated IFN-α/β production (and greater macromolecular synthesis shutoff) in those cells. IFN-α/β potentially mediates protection by upregulating levels of ISGs and antiviral pathway proteins, which enhance resistance of myeloid cells to VEEV-encoded activities targeting downregulation of such responses. Consistent with this, we observed a greater increase in mRNA levels of antiviral genes after infection of B6 macrophages than after infection of AB6 macrophages. Overall, we demonstrate the ability of a subset of myeloid cells to resist VEEV-mediated activities targeting the activation of antiviral pathways following infection, with IFN-α/β protein expression, in particular, reflecting translation inhibition resistance. Finally, as even some myeloid cells without IFNAR signaling remained resistant to virus replication and macromolecular synthesis inhibition, additional factors that affect replication efficiency, such as virus receptor abundance, availability of host factors required for replication, and/or other IFN-α/β-independent restrictive mechanisms, may determine replication efficiency in some myeloid cells.

A previous study concluded that IFN-α/β was secreted in two waves following infection of myeloid cells in VEEV-infected lymph nodes and in vitro cultures (12). Infected cells produced IFN-α/β in the first phase, with uninfected cells, upon detection of viral replication in neighboring cells, releasing an equivalent amount of IFN-α/β in a second wave. Our data using replicons competent for a single round of replication and incapable of spreading to neighboring cells suggest that the majority (∼80%) of IFN-α/β is produced from infected cells, at least in RAW cell and primary macrophage cultures. We were unable to completely separate uninfected cells from cells infected at very low levels due to the sensitivity of the instruments and assays used, and it is possible that the IFN-α/β secreted by virus-exposed GFP⁻ cells was being made by cells with low levels of viral replication. Supporting this idea, qRT-PCR analysis detected virus genomic RNA in all infected cell populations. However, production of IFN-α/β mRNA and protein was very low/undetectable in VREP and BMMϕ^−/low populations. Limited IFN-α/β upregulation in these populations was associated with positive-sense genome levels at or below 1 × 10⁶ genome equivalents (GE), whereas all the populations that induced an IFN response were above 1 × 10⁶ GE. However, it is also possible that other factors, such as expression levels of infection- or RNA-sensing receptors in uninfected cells, may play a role in the requirement for virus replication. The data presented in the previous report (12) may also have been confounded by the method used to differentiate uninfected cells from infected cells, with the possibility of significant cross-contamination between infected and uninfected cell contents, whereas our study utilized intact-cell sorting to more definitely separate different populations.

Similarly, the authors of a study exploring the role of plasmacytoid dendritic cells in controlling CHIKV infection concluded that pDCs can indirectly sense infection through cell-cell contact with CHIKV-infected cells and subsequently secrete IFN-α/β (13). IFN-α/β was secreted in an IRF7-dependent manner in the absence of mitochondrial antiviral signaling protein (MAVS), but not Toll-like receptor 7 (TLR7), suggesting that detection of viral single-stranded RNA (ssRNA), possibly from phagocytosed virions or extracellular vesicles and/or apoptotic blebs secreted by infected cells (47), drives IFN-α/β production in pDCs. Clearly, all the cells in our cultures were exposed to virions containing viral positive-sense genomic RNA during infection, and this form was detected in all the populations after infection. However, in the other study, pDCs were infected with CHIKV, although virus replication was not assayed and IFN-α/β protein production was not observed. In contrast with VEEV, CHIKV is poorly infectious for murine myeloid cells (48–50), which may affect both the responses of pDCs and more productively infected macrophages. Depletion of pDCs in vivo followed by infection with a myeloid-cell-tropic virus such as VEEV will be required to definitively assess their contribution to serum IFN-α/β production. While the possibility of uninfected myeloid cells communicating with infected cells and upregulating antiviral cytokine responses cannot be discounted, the current study, along with previous work showing that replication in myeloid cells in vivo is required for EEEV production of serum IFN (6), suggests that active entry of virions and subsequent viral replication in cells partially resistant to macromolecular synthesis inhibition contribute to induction of the robust IFN-α/β response observed systemically after VEEV infection.

MATERIALS AND METHODS

Generation of primary cells.

Primary 3-day cDCs and macrophages (BMMϕ) were generated as previously described (7). The cDCs were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 200 mM l-glutamine (l-Glut) (Sigma), 10,000 U/ml penicillin (Sigma), 10 mg/ml streptomycin (Sigma), 1 mM sodium pyruvate (Sigma), 1% nonessential amino acids (Sigma), 5 mM HEPES buffer (Sigma), 50 μM β-mercaptoethanol, 10 ng/ml granulocyte-macrophage colony-stimulating factor (GM-CSF) (Peprotech), and 10 ng/ml interleukin 4 (Peprotech). BMMϕ were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, 200 mM l-Glut, 10,000 U/ml penicillin (Sigma), 10 mg/ml streptomycin (Sigma), and 20% L929-conditioned cell supernatant. To harvest cells, bone marrow from femurs and tibias from the indicated mouse strains was plated for 1 h (37°C and 5% CO₂) in RPMI medium without supplements to allow attachment of macrophages. The nonadherent dendritic cells were removed and cultured separately in appropriate medium. Adherent macrophages were then cultured in appropriate media. After 3 days, nonadherent cDCs were pelleted, counted, and seeded for further use. Adherent macrophages were scraped off, counted, and seeded for experiments.

Cell culture.

Raw 264.7 cells (ATCC) were cultured in DMEM supplemented with 10% FBS, 200 mM l-Glut, 10,000 U/ml penicillin, and 10 mg/ml streptomycin. Tetracycline-inducible MEFs (Clontech) were grown in the above-mentioned medium supplemented with 50 mg/ml G418. BHK-21 and L929 cells (ATCC) were maintained in RPMI medium supplemented with 10% donor bovine serum (DBS), 10% tryptose-phosphate broth (TPB), 10,000 U/ml penicillin, and 10 mg/ml streptomycin. All the cells were grown at 37°C with 5% CO₂.

Viruses and virus infections.

The construction of cDNA clones of wild-type VEEV strains ZPC738 (43) and Trinidad Donkey (V3000) (51) has been previously described. Capsid fusion reporter viruses expressing enhanced GFP (eGFP) or MeDF blue timer protein (a gift from Peter Drain, University of Pittsburgh) were constructed as previously described (39) by inserting the eGFP or MeDF timer gene followed by the Thosea asigna 2A-like protease in frame between capsid and E3. GFP- or MeDF-expressing viruses were used interchangeably depending on the titer of a particular stock. The construction and packaging of a VEEV replicon expressing GFP (VREP GFP) (a gift from Robert Johnston, University of North Carolina—Chapel Hill) has been previously described (5). Virus stocks were created by electroporation of in vitro-transcribed capped viral RNA into BHK-21 cells (51). Virus titers were determined using a standard BHK-21 plaque assay. For virus infections, cells were seeded on plates overnight. The viruses were diluted to the indicated MOI in phosphate-buffered saline (PBS) with calcium/magnesium supplemented with 1% DBS. The growth medium was removed, and the cells were infected for 1 h at 37°C and 5% CO₂, following which the growth medium was replaced.

IFN bioassay.

The concentrations of biologically active mouse IFN-α/β in sera and cell culture supernatants were determined using a previously described IFN bioassay (52). Briefly, 3 × 10⁴ L929 cells/well were seeded in a 96-well plate; 200 μl of each sample was acidified to pH 2.0 using 1 M HCl for 24 h at 4°C. Samples were neutralized to pH 7.0 using 1 M NaOH, and 100 μl was added in duplicate to the cells. Samples were then diluted 2-fold across the plate and left for 24 h at 37°C. Encephalomyocarditis virus (EMCV) (4 × 10³ PFU/well) was used to infect the cells for 24 h. The cells were then fixed and stained with 1% crystal violet in 15% methanol. IFN concentrations were determined as the dilution of sample required to protect 50% of the cells in a well from the cytopathic effect (CPE) caused by EMCV compared with protection from an IFN standard at a known concentration.

Metabolic labeling.

To measure host translation, cells were labeled with [³⁵S]cysteine/methionine (MP Biochemical). Thirty minutes prior to labeling, the cells were washed twice with starvation medium (DMEM without Cys/Met [Cellgro] with 1% FBS, 200 mM l-Glut, and penicillin/streptomycin). The cells were then incubated in starvation medium for 30 min at 37°C. Next, the cells were labeled with 100 μCi/ml [³⁵S]Cys/Met in starvation medium for 1 to 2 h at 37°C. The cells were washed in PBS without calcium/magnesium, and the lysates were collected in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 1 mM EDTA, and 1 mM EGTA supplemented with protease inhibitors [1 mM phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, and 1 μg/ml pepstatin], and a phosphatase inhibitor cocktail [Sigma]). Equal volumes of lysates were resolved on an SDS-10% PAGE gel. The gels were fixed, dried, and exposed to photographic film (GE Healthcare) for 7 days at −80°C. Densitometry was performed using Image J software. Equal volumes of lysates were also resolved and stained for actin as a loading control.

RT-PCR.

Lysates were collected from infected cells in TRIzol, and total RNA was extracted using the manufacturer’s protocols (Ambion). For each sample, cDNA was made by transcribing 100 ng RNA using random hexamers and Moloney murine leukemia virus (MMLV) reverse transcriptase (Fisher; no. 28025013) according to the manufacturer’s instructions. Levels of IFN-α4 (sense, 5′-CTGCTGGCTGTGAGGAAATA-3′; antisense, 5′-GAAGACAGGGCTCTCCAGAC-3′), IFN-β (sense, 5′-GAACATTCGGAAATGTCAGG-3′; antisense, 5′-ACTGTCTGCTGGTGGAGTTC-3′), MDA5 (sense, 5′-CCCAACATTATGAGCAGTGG-3′; antisense, 5′-CGTGTCGTTGATTTGTAGGG-3′), IFIT1 (sense, 5′-GTGGCTCACATAGAGCAGGA-3′; antisense, 5′-AGTTTCCTCCAAGCAAAGGA-3′), ISG20 (sense, 5′-GGCACTGAGACAGGGCTT-3′; antisense, 5′-CCATGGATGTTCACAATGCT-3′), and ISG15 (sense, 5′-TCCATGACGGTGTCAGAACT-3′; antisense, 5′-GACCCAGACTGGAAAGGGTA-3′) mRNAs were measured using the indicated primers (in parentheses). These genes were selected based on the results of a previous study (53) in which they were upregulated by virus infection or IFN treatment of bone marrow-derived dendritic cells from wild-type mice, but not in cells derived from IFN receptor-null mice. To measure the rate of transcription inhibition, levels of mouse gamma actin intron 3 were measured using the following primers (sense, 5′-ACAGAACGCAAGCAGAAACG-3′; antisense, 5′-TGGCATTTCCTCCCTGAAGC-3′). 18S RNA levels were measured as a loading control (primers: sense, 5′-CGCCGCTAGAGGTGAATTTCT-3′; antisense, 5′-CGAACCTCCGACTTTCGTTCT-3′). Values were normalized to 18S RNA levels using the ΔΔC_T method (54), and reactions excluding reverse transcriptase were performed to rule out DNA contamination.

To measure levels of positive-strand viral RNA, reverse transcription was performed with a standard amount of total RNA (100 or 150 ng/sample) using MMLV reverse transcriptase (Fisher; no. 28025013). Specific-strand cDNA synthesis of VEEV TrD nsP2 was facilitated using a T7-tagged antisense primer, T7_TrD_R_Pos (5′-GCGTAATACGACTCACTATACAGGTACTAGGTTTATGCGC-3′), and 18S cDNA synthesis was performed with an antisense 18S primer (5′-CGAACCTCCGACTTTCGTTCT-3′). The following RT conditions were used. Primer, deoxynucleoside triphosphates (dNTPs), and RNA were added together, heated to 65°C for 5 min, and then transferred to ice for 2 min to enhance primer binding before the addition of enzymatic reagents. Extension was carried out at 37°C for 50 min, followed by reverse transcriptase inactivation at 70°C for 15 min. All data were obtained with an Applied Biosystems Quant Studio 6 Flex real-time PCR system. Positive-sense viral cDNA was amplified using TaqMan Fast Universal PCR master mix (2×; Fisher; no. 4352042) with T7 primer (5′-GCGTAATACGACTCACTATA-3′), VEEV_TrD_3374_Probe (5′-6-carboxyfluorescein [FAM]-TAGACTCTTCCAGTGGCAACTGCC- black hole quencher [BHQ]_1-3′), and TrD_F_Neg (5′-TCCGTCAGCTCTCTCGCAGG-3′). The PCR conditions included a 95°C denaturation and activation step for 20 s, followed by 45 cycles of denaturation at 95°C for 3 s, and then extension at 60°C for 20 s. Fluorescence intensity was measured at the end of each 20-s extension.

Genome equivalents were determined by interpolating measured nsP2 threshold cycle (C_T) values with those of a generated standard curve. The positive-sense standard curve was based on 10-fold dilutions of VEEV TrD replicon in vitro-transcribed RNA. The mean expression of 18S C_T values for each experimental group was used to correct loading errors within the group. The corrected C_T values were then interpolated with the correct standard curve.

Cell sorting.

For sorting experiments, cells were infected for the indicated times as described previously. For experiments where secondary infection was blocked, cells were infected for 1 h and washed three times in PBS without calcium and magnesium, and growth medium containing anti-VEEV antibody (ATCC anti-VEEV ascites fluid) diluted 1:100 (final concentration, 10 to 20 50% plaque reduction/neutralization titers [PRNT₅₀]) was added on top of the cells. Prior to sorting, the cells were washed twice in PBS without calcium and magnesium, scraped off, and collected into sorting buffer (PBS without calcium and magnesium supplemented with 10% FBS). The cells were sorted into the indicated populations depending on the expression levels of reporter proteins (GFP or MeDF timer protein) using a BD FACS Aria 1 SORP (special-order research product) cytometer. The default yield sort mask settings were used. GFP was excited using a 488-mm laser (100 mW), whereas MeDF timer protein was excited using a 405-mm laser (20 mW). The duration of each sort was generally 2 to 4 h. Following completion of sorting, equal numbers of cells for each group were concentrated and seeded into 24-well plates (100,000 to 200,000/well). The cells were plated for 30 min before being used to collect samples for metabolic labeling (2 h postsort), IFN bioassay (2.5 h postsort), or RT-PCR (2.5 h postsort).

In vivo experiments.

The indicated mouse strains (C57BL/6 [Charles River], AB6 [IFNAR^−/−; a gift from Sujan Shrestha]), and IRF3^−/− and IRF7^−/− strains [both gifts from Mike Diamond]) were used at 6 to 8 weeks of age and were infected with inocula (10 μl) containing 1,000 PFU of virus in the hind leg footpad using a gas-tight Hamilton syringe and a 27-gauge needle. The mice were observed every 24 h, and the AST and percent mortality were calculated. Sera and tissues were collected at the indicated times p.i.

Statistics.

GraphPad Prism software was used to perform the following statistical tests: Student’s t test, one-way analysis of variance (ANOVA), and two-way ANOVA.

Ethics statement.

All animal studies were performed in accordance with procedures from the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Animal procedures were conducted according to protocols (no. 16119340 and 17091503) approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh. Mice were anesthetized with isoflurane for viral infections. When meeting euthanasia criteria, mice were euthanized by overdosing with isoflurane and cervical dislocation. When collecting sera, mice were overdosed with isoflurane and exsanguinated by cardiac puncture.