Venezuelan Equine Encephalitis Virus Antagonizes the cGAS-STING Pathway

Highlights

What are the main findings?

• VEEV induces interferon-stimulated gene expression during late viral infection independent of STING phosphorylation at residue Ser366.
• VEEV suppresses agonist-induced phosphorylation of STING (Ser366).
• Priming the STING pathway with dsDNA suppresses alphavirus replication.

What are the implications of the main findings?

• Elucidating the mechanism by which VEEV suppresses phosphorylation of STING (Ser366) may reveal viral–host interactions with relevance for future therapeutic exploration.

Abstract

Venezuelan equine encephalitis virus (VEEV) is a mosquito-borne pathogen causing low mortality but high morbidity in humans, with 4–14% cases exhibiting neurological complications. While the cyclic GMP-AMP synthase–stimulator of interferon genes (cGAS–STING) pathway is canonically associated with double-stranded DNA (dsDNA) detection, it has been shown to respond to RNA viruses and subsequently limit viral pathogenesis. Several viruses antagonize this signaling cascade, underscoring the importance that cGAS–STING plays in host immunity. Previous studies regarding single-stranded RNA viruses revealed that cGAS–STING limits viral replication in Old World alphavirus chikungunya virus infections, but little is known about New World alphaviruses such as VEEV. Here, we investigate the impact that STING activation has on VEEV infection as a potential prophylactic and therapeutic intervention. VEEV infection alone did not induce STING phosphorylation at Ser366, but interferon-stimulated genes (ISGs) were upregulated during the late phase of infection. Loss of STING through siRNA showed a partial dependency on STING for ISG transcription, suggesting that STING activation may occur through a noncanonical process. Priming of the STING pathway prior to infection was found to be critical in limiting viral replication; however, targeting STING activation post-infection abrogated the antiviral effects that dsDNA had on VEEV. VEEV suppressed STING phosphorylation in a multiplicity of infection (MOI)-dependent manner with the most robust pSTING (Ser366) inhibition observed at an MOI of 10. Collectively, our results suggest that VEEV antagonizes canonical STING activation.

1. Introduction

Venezuelan equine encephalitis virus (VEEV) is a zoonotic mosquito-borne pathogen that can induce encephalitis in both humans and equines [1]. This single-stranded, positive-sense RNA virus is listed as an overlap select agent by the Centers for Disease Control and Prevention and the United States Department of Agriculture (USDA) due to ease of dissemination, high morbidity rates, and low mortality rates (<1%) in humans [2]. VEEV-induced neurological sequelae are observed in 4–14% of cases, with symptoms ranging from confusion to convulsions and cognitive impairment [3,4]. There is a critical need for therapeutic development against VEEV infection due to the lack of U.S. Food and Drug Administration-approved vaccines or treatments.

During infection, VEEV elicits a Type I interferon (IFN) response that ultimately aids in controlling pathogenesis; however, VEEV has simultaneously evolved ways to antagonize this pathway in which nonstructural protein 2 (nsP2) and capsid are largely responsible for dampening the host immune response [5,6,7,8]. Despite this ongoing “arms race” at the crux of viral–host interactions, the Type I IFN response can be exploited for therapeutic development, but the timing of intervention is paramount [4,9]. Multiple studies investigating the use of IFN as a pretreatment option for VEEV infection have shown promising results through reduction in viral titers in cortical neurons from CD-1 mice [10], increased survival and decreased morbidity in mice challenged intranasally with VEEV TC-83 [11], and subcutaneous and aerosol challenge with virulent strain Trinidad Donkey (TrD) [12]. One study evaluated the effects of IFN-α treatment administered concomitantly or post-infection in C57BL/6J mice intranasally infected with VEEV strain ZPC-738, upon which treatment delayed viral replication and associated morbidity but still resulted in 100% mortality [13], underscoring the importance of therapeutic timing.

Type I IFN can be stimulated through several cytosolic pattern recognition receptors (PRRs), including the retinoic acid-induced gene I (RIG-I) and melanoma differentiation-associated gene 5 (MDA5) to detect single-stranded RNA (ssRNA)/double-stranded RNA (dsRNA), and cyclic GMP-AMP synthase (cGAS) to detect cytosolic dsDNA [14,15]. RIG-I and MDA5 have been shown to respond to VEEV infection, leading to the activation of the Type I IFN response and release of interferon-stimulated genes (ISGs) [16]; however, little is known about cGAS detection of dsDNA during VEEV infection. cGAS produces a secondary messenger molecule, cyclic GMP-AMP (cGAMP), upon recognition of dsDNA or DNA:RNA hybrids in the cytosol, which binds to and activates endoplasmic reticulum (ER)-bound protein stimulator of interferon genes (STING) [17]. STING undergoes a conformational change and translocation to the ER–Golgi intermediate compartment, where recruited protein TANK-binding kinase 1 (TBK1) phosphorylates transcription factor interferon regulatory factor 3 (IRF3), initiating downstream signaling of IFN production and subsequent ISG production [18].

Despite cGAS canonically recognizing dsDNA, the cGAS-STING pathway has been shown to be activated in response to ssRNA viruses through mechanisms such as the endogenous release of mitochondrial DNA (mtDNA) into the cytosol during infection [19]. Research on alphaviruses and STING has predominantly been limited to an arthritogenic alphavirus, chikungunya virus (CHIKV). STING is essential in limiting CHIKV pathogenesis, and CHIKV can circumvent this pathway through degradation of cGAS [20,21]. Small-molecule STING agonists, G10 and C11, have been shown to inhibit CHIKV and VEEV viral replication in vitro by inducing the Type I IFN response in a STING-dependent manner [22,23]. Previous studies found that VEEV induces mitochondrial stress through accumulation of reactive oxygen species (ROS) and decreased mitochondrial membrane permeability (MMP), which can lead to the release of mtDNA into the cytosol [24,25,26]. These studies suggest that STING may be induced in VEEV-infected cells.

In this study, we sought to characterize the dynamic of STING signaling during VEEV infection. VEEV was found to elicit a Type I IFN response in vitro independent of STING (Ser366) phosphorylation. Pretreatment of cells with dsDNA, a known activator of the cGAS-STING pathway, reduced VEEV infectious virus, but decreased viral replication was not observed when cells were treated with dsDNA post-infection. Pretreatment of cells with dsDNA resulted in decreased replication of multiple alphaviruses, including Eastern equine encephalitis virus (EEEV), CHIKV, and Sindbis virus (SINV). Surprisingly, VEEV-infected cells that underwent dsDNA post-treatment led to the inhibition of STING (Ser366) phosphorylation in a multiplicity of infection (MOI)-dependent manner. Collectively, the data reveals VEEV’s ability to antagonize canonical STING activation.

2. Materials and Methods

2.1. Cell Culture

Human microglial cells (HMC3) (ATCC, CRL-3304, Manassas, VA, USA) were maintained in Eagle’s Minimum Essential Medium with L-glutamine (EMEM) (Quality Biological, 112-018-101, Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum (FBS) (Avantor, 97068-085, Radnor, PA, USA) and 1% penicillin/streptomycin (Pen-Strep) (Quality Biological, 120-095-721, Gaithersburg, MD, USA). Mouse embryonic fibroblast (MEF) cells (ATCC, CRL-2991, Manassas, VA, USA) and African green monkey kidney epithelial (Vero) cells (ATCC, CCL-81, Manassas, VA, USA) were maintained in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco, 11960-044, Grand Island, NY, USA) supplemented with 10% FBS, 1% L-glutamine (Corning, 25-005-Cl, Manassas, VA, USA) and 1% Pen-Strep.

2.2. VEEV TC-83 V5-C Design

A plasmid containing the V5-capsid gene block (V5-C gene block) was ordered from Integrated DNA Technologies and transformed into Escherichia coli using the One Shot™ TOP10 kit (Invitrogen, C404010, Carlsbad, CA, USA). Large-scale plasmid purification was performed with the Qiagen Plasmid Maxi kit (Qiagen, 12165, Germantown, MD, USA). A VEEV TC-83 plasmid with an ampicillin resistance marker (kind gift from Dr. Ilya Frolov, University of Alabama at Birmingham, Birmingham, AL, USA) and the V5-C gene block were digested with the restriction enzymes PspOMI (New England Biolabs, R0653, Ipswich, MA, USA) and AflII (New England Biolabs, R0520, Ipswich, MA, USA). The digested DNA samples were separated on a 1% agarose gel, and the bands of the appropriate sizes were excised, extracted, and purified using the QIAquick Gel Extraction kit (Qiagen, 28704, Germantown, MD, USA). The VEEV TC-83 plasmid and the V5-C gene block were then ligated using T4 DNA Ligase (Invitrogen, Cat #15224-041, Carlsbad, CA, USA), inserting the V5 tag at the N-terminus of the capsid, immediately after the methionine start codon. The ligated plasmid was transformed into E. coli using the One Shot™ TOP10 kit and plated on yeast extract peptone dextrose (YPD) agar containing 100 μg/mL ampicillin. Selected colonies were cultured in liquid YPD medium with 200 μg/mL ampicillin and incubated overnight at 37 °C in a shaking incubator. Plasmid DNA was extracted using the Qiagen QIAprep Spin Miniprep kit (Qiagen, 27106, Germantown, MD, USA), and successful insertion of the V5 tag was confirmed by gel electrophoresis and Sanger sequencing. This newly constructed plasmid was designated as the VEEV TC-83 V5-C plasmid. VEEV TC-83 V5-C was rescued and titered as described [27].

2.3. Viral Stocks and Infections

Venezuelan equine encephalitis virus strain Trinidad Donkey (VEEV TrD), Eastern equine encephalitis virus strain GA97 (EEEV GA97), and Sindbis virus (SINV) (EgAr 339) (BEI Resources, NR-15695, Manassas, VA, USA) stocks were generated as previously described [28]. Chikungunya virus (CHIKV) Vaccine Strain 181/Clone 25 was obtained from Dr. Naomi Forrester (University of Texas Medical Branch, Galveston, TX, USA) [29]. Experiments utilizing VEEV TrD and EEEV GA97 were performed at BSL3 in accordance with the Federal Select Agent Program. Experiments using VEEV TC-83 V5-C, SINV (EgAr 339), and CHIKV (181/25) were performed at BSL2.

Viral infections were conducted as previously described [30], with cells plated in 24-well plates (Greiner Bio-One, 662165, Monroe, NC, USA) at a seeding density of 7.5 × 10⁴ cells/well for HMC3 cells and 6 × 10⁴ cells/well for MEFs. At time of collection, supernatants were collected, cells washed twice with phosphate-buffered saline (PBS) (Corning, 21-040-CV, Manassas, VA, USA), and cells lysed in either blue lysis buffer or TRIzol as detailed below. Supernatants and lysates were stored at −80 °C until samples could be processed for downstream analysis.

2.4. Western Blot and Antibodies

Cells were collected in blue lysis buffer [30], 35 μL for HMC3 viral infections or 55 μL for MEF samples, and processed via Western blot as previously described [30] with modifications. Gel electrophoresis was conducted using NuPAGE 4–12% Bis-Tris gels with cell lysates totaling 20 μL loaded onto 12-well gels (Invitrogen, NP0322BOX, Carlsbad, CA, USA) or 25 μL for 10-well gels (Invitrogen, NP0321BOX, Carlsbad, CA, USA).

Antibodies specific to pSTING (Ser366) (50907), STING (13647), pTBK1/NAK (Ser172) (5483), TBK1/NAK (3504), and cGAS (79978) were obtained from Cell Signaling Technology (Danvers, MA, USA) and diluted 1:1000 in 5% BSA in Tris-Buffered Saline with Tween 20 (TBS-T). Goat anti-Rabbit IgG (H+L) Secondary Antibody HRP (Invitrogen, 32460, Carlsbad, CA, USA) was diluted 1:2000 in 5% milk in TBS-T. HRP-conjugated primary antibodies specific to β-actin (Abcam, ab49900, Waltham, MA, USA) and V5 tag (Bio-Rad, MCA1360P, Hercules, CA, USA) were diluted in 5% milk in TBS-T at 1:20,000 and 1:10,000, respectively.

MagicMark XP Western Protein Standard (Thermo Fisher Scientific, LC5602) was used for protein size determination. Precision Plus Protein Kaleidoscope Prestained Standards (Bio-Rad, 1610375, Hercules, CA, USA) were used to verify protein transfer and, when applicable, to determine where to cut the membrane prior to blocking. The MagicMark ladder was overexposed when imaging for pSTING (Ser366), STING, pTBK1/NAK (Ser172) and TBK1/NAK, in which case the ladder was covered and the membrane imaged for better band visualization. Membranes incubated with HRP-conjugated primary antibodies for β-actin and V5 tag resulted in high-intensity bands that rendered the ladder invisible. Western blot images for all proteins, including instances where the ladder has been covered, can be found in the Supplementary Materials Section.

Western blot membranes probed for multiple proteins were stripped of bound antibodies utilizing a stripping buffer made of 1.5% Glycine (Sigma-Aldrich, G7126, Saint Louis, MO, USA), 0.1% SDS (Sigma-Aldrich, L3771, Saint Louis, MO, USA), and 0.01% Tween 20 (Promega, H5151, Madison, WI, USA), pH adjusted to 2.2 with Hydrochloric acid (Avantor, JT5618, Radnor, PA, USA), in 1000 mL of autoclaved Millipore water. Membranes were incubated via rocking at room temperature in stripping buffer for 10 min, followed by washing with TBS-T for 10 min three times, and then blocked in 5% milk in TBS-T as normal.

2.5. Western Blot Quantification

Relative expression of pSTING (Ser366) and STING was quantified using ImageJ 1.54i software. Band intensities for pSTING, STING, and β-actin were measured based on a set region of interest. The background was subtracted prior to normalization of pSTING (Ser366) or STING to the loading control (β-actin), followed by normalization to a mock sample.

2.6. RNA Extraction and RT-qPCR

Cells were washed twice with PBS prior to being lysed with TRIzol (Invitrogen, 15596018, Carlsbad, CA, USA). Total RNA was purified using a Direct-zol RNA Miniprep kit (Zymo Research, R2052, Irvine, CA, USA) per the manufacturer’s protocol, including DNase I treatment, and RNA eluted in molecular-grade water (Corning, 46-000-Cl, Manassas, VA, USA). Samples were quantified on a NanoDrop Spectrophotometer (Thermo Fisher Scientific, ND-ONEC-W, Carlsbad, CA, USA) followed by normalization of samples to 10 ng/μL.

Viral RNA was quantified using 5 μL normalized RNA with the RNA UltraSense One-Step Quantitative RT-PCR System (Thermo Fisher Scientific, 11732927, Carlsbad, CA, USA), primer pairs specific to VEEV TC-83 nucleotides 7931 to 8005 (Forward 5′-TCTGACAAGACGTTCCCAATCA-3′ and Reverse 5′- GAATAACTTCCCTCCGACCACA-3′), and a TaqMan probe (5′-6-carboxyfluorescein-TGTTGGAAGGGAAGATAAACGGCTACGC-6-carboxy-N,N,N′,N′-tetramethylrhodamine-3′) on a StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific, 4376600, Carlsbad, CA, USA) as previously described [11,31]. Absolute quantification of viral RNA was determined by StepOnePlus software v2.3 based on standard curves generated from serial dilutions of VEEV TC-83 RNA.

Relative quantification of host gene expression was determined via the delta–delta Ct (ΔΔCt) method [32] utilizing the Taqman RNA-to-C_T 1-Step Kit (Thermo Fisher Scientific, 4392938, Carlsbad, CA, USA) as previously described [30] except with GAPDH used as the endogenous control. TaqMan Gene Expression Assays (FAM-MGB) specific to IFIT2 (Hs01584837_s1), IFIT3 (Hs01922752_s1), ISG15 (Hs01921425_s1), ISG20 (Hs00158122_m1), CXCL10 (Hs00171042_m1), STING/TMEM173 (Hs00736955_g1) and GAPDH (Hs02786624_g1) were obtained from Thermo Fisher Scientific (Carlsbad, CA, USA).

2.7. Crystal Violet Plaque Assay

Vero cells were plated in 12-well plates (VWR, 10062-894, Radnor, PA, USA) at a seeding density of 1.5 × 10⁵ cells/well. The following day, plaque assays were conducted using an established agarose overlay protocol [33].

2.8. dsDNA Treatment

Poly (dA:dt)/LyoVec (Invivogen, tlrl-patc, San Diego, CA, USA) was reconstituted per the manufacturer’s protocol to 50 μg/mL using the provided sterile endotoxin-free water and stored at 4 °C for up to one week.

HMC3 cells were infected with VEEV TC-83 V5-C and challenged, either pre- or post-infection, as indicated in the corresponding figures, with 500 μL/well of dsDNA (1 μg/mL) in supplemented EMEM or no treatment (media spiked with equal amounts of diluent).

2.9. Cell Viability and Drug Treatment

5,6-dimethylxanthenone-4-acetic acid (DMXAA) (Sigma-Aldrich, D5817-5MG, Saint Louis, MO, USA) was reconstituted to 10 mg/mL with dimethyl sulfoxide (DMSO) (ATCC, 4-X, Manassas, VA, USA) and stored in single-use aliquots at −20 °C.

For cytotoxicity assays, MEFs were plated at a seeding density of 4 × 10⁴ cells/well in a 96-well plate (Corning, 3903, Manassas, VA, USA). DMXAA (or DMSO) was serially diluted in supplemented DMEM and 200 μL/well added to cells prior to being incubated at 37 °C with 5% CO₂. ATP production was measured 48 h post-treatment using CellTiter-Glo (Promega, G7570, Madison, WI, USA), in accordance with the manufacturer’s protocol, and luminescence quantified on a Promega GloMax Discover (Promega, GM3000, Madison, WI, USA) to assess cell viability.

For viral titer analysis and generating protein lysates for Western blot analysis, MEFs were plated in 24-well plates at a seeding density of 6 × 10⁴ cells/well. The following day, DMXAA was serially diluted in supplemented DMEM, with the highest % DMSO serving as the untreated control. Cells were treated with DMXAA or DMSO for 21 h. At time of infection, cells from extra wells were trypsinized and quantified to determine accurate cell density for MOI calculation. Cells were infected with VEEV TC-83 V5-C at MOI 0.1 for one hour at 37 °C, 5% CO₂, with plates rocked every 15 min. Infection was removed, cells washed twice with PBS, fresh media added, and plates incubated until time of collection. Supernatants and protein lysates were collected 24 h post-infection as described above.

2.10. siRNA Transfection

HMC3 cells were plated in a 6-well plate at 1 × 10⁵ cells/well. The following day, cells were transfected with 50 nM ON-TARGETplus Human STING1 siRNA SmartPool (Horizon Discovery [Dharmacon], L-024333-00-0005, Lafayette, CO, USA) or ON-TARGETplus Non-targeting Control Pool (Horizon Discovery [Dharmacon], D-001810-10-05, Lafayette, CO, USA) using DharmaFECT 1 siRNA Transfection Reagent (Horizon Discovery [Dharmacon], T-2001-02, Lafayette, CO, USA). The transfection was done in accordance with the manufacturer’s protocol utilizing DharmaFECT reagent at 2.4 μL/well and Gibco Opti-MEM I Reduced Serum Medium (Gibco, 31985-062, Grand Island, NY, USA). Transfections were added to the cells and plates incubated at 37 °C, 5% CO₂, for 24 h. Spent media were replenished with complete EMEM and plates incubated for an additional 24 h. At 48 h post-transfection, cells were either mock-infected (media only) or infected with VEEV TC-83 at MOI 1 as detailed above. At 8 and 16 h post-infection, the supernatants were collected, cells washed twice with PBS, and lysates collected in either 300 μL TRIzol or 300 μL blue lysis buffer.

2.11. Plasmid Construction

VEEV TC-83 nonstructural protein expression vectors (nsP1, nsP2, nsP3, nsP4) were designed as an insertion of the corresponding individual nonstructural genes into the pcDNA3.1(+) plasmid under the control of the double (cytomegalovirus) CMV and T7 promoter between KpnI and XbaI restriction sites, except for the nsP4 protein, which was inserted between NheI and Xba restriction sites. Each nonstructural protein was additionally modified to contain an in-frame-fused V5 tag at the N-terminus and TAG used as a stop codon for each nsP. To ensure efficient translation, each nonstructural gene was flanked with the Kozak sequence upstream from the gene. The nsP gene synthesis and cloning into pcDNA3.1(+) were ordered and performed by GenScript Biotech Corporation (Piscataway, NJ, USA).

A VEEV TC-83 structural polyprotein expression vector (E3-E2-6K/TF-E1) that expresses all but the capsid’s structural genes was developed as an insertion of the structural polyprotein-encoding sequence into the pcDNA3.1(+) plasmid between KpnI and EcoRV restriction sites. The V5 tag was in-frame-used at the C-terminus of the polyprotein. The Kozak sequence was placed upstream from the polyprotein-encoding gene. ATG and TGA were used as a start and stop codon, respectively. The polyprotein gene synthesis and cloning into pcDNA3.1(+) were ordered at GenScript Biotech Corporation.

Construction of the plasmid expressing VEEV TC-83 capsid protein with a C-terminal V5 tag has been detailed elsewhere [34].

The detailed maps for all developed constructs are available upon request.

2.12. Plasmid Transfection

Individual plasmids expressing VEEV TC-83 nsP1, nsP2, nsP3, nsP4, capsid, E3-E2-6K/TF-E1, or pcDNA3.1(+) were transfected into HMC3 cells (1 × 10⁵ cells/well in 12-well plates) with TransIT^®-LT1 Transfection Reagent (Mirus, MIR 2300, Madison, WI, USA) per the manufacturer’s protocol, using the recommended starting conditions of 1 µg DNA and 3 µL Transfection Reagent. The transfection:plasmid master mixes were incubated at room temperature for 15 min. During the incubation period, old media were discarded and 1 mL fresh supplemented EMEM added to cells. Transfections were added to the cells and plates incubated at 37 °C, 5% CO₂, for 24 h.

The following day, the plasmid transfection was removed, cells washed once with PBS, and 1 mL/well of freshly reconstituted dsDNA diluted in EMEM to 1 µg/mL added to the cells and plates incubated at 37 °C, 5% CO₂. Cell lysates were collected at 17 h post-treatment with dsDNA in 80 µL blue lysis buffer.

2.13. Statistical Analysis

GraphPad Prism (San Diego, CA, USA) Version 10.6.1 was used for statistical analyses, with specific tests described within the figure legends. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001.

3. Results

3.1. VEEV Upregulates Type I IFN Genes During Late Infection in the Absence of STING Phosphorylation at Ser366

To test VEEV’s impact on the cGAS-STING pathway and Type I IFN response through downstream ISG production, host transcriptional and translational changes were assessed subsequent to viral infection by Western blot and gene expression analysis. Human microglial (HMC3) cells were infected with VEEV TC-83 V5-C with a V5 tag inserted at the capsid’s N-terminus at a low and high MOI (0.1 and 10, respectively) for up to 24 h. Phosphorylation of STING residue Ser366 serves as a common marker of STING activation as it relates to the STING-TBK1-IRF3 arm of the Type I IFN response [35,36]. Relative protein expression over time revealed a lack of STING phosphorylation (pSTING) following infection alone, with low levels of pSTING observed in all samples, including the mock-infected group, during both early and late viral infection (Figure 1A,B).

Elevated levels of pSTING were present only in the dsDNA-treated uninfected control and the corresponding total STING levels were decreased, as expected, due to a negative feedback loop [37]. V5-tagged capsid (V5-C) protein appeared as early as 4 h post-infection (hpi) when infected at MOI 0.1 and 2 hpi at MOI 10 (Figure 1A). VEEV capsid expression was most prominent during late infection at the high MOI (Figure 1B), in which replication kinetics for infectious virus measured by a plaque assay reflected this increase in capsid production at 8 hpi for MOI 10 compared to MOI 0.1 (Figure 1C). cGAS levels were largely unchanged following VEEV infection (Figure 1A,B). Transcriptional changes in ISGs were measured by RT-qPCR using the ∆∆Ct method for genes (IFIT2, IFIT3, ISG15, ISG20, and CXCL10) normalized to the endogenous control, GAPDH. These genes were chosen due to their regulation of the Type I IFN response with varying importance of antiviral mechanisms against VEEV. IFIT3 has been shown to be a key mediator of the IFN response restricting VEEV replication [38]. IFIT2 restricts VEEV via interaction with the 3’ UTR region of the genome, but its effectiveness is largely strain-dependent due to the genome variability in that region [39]. ISG20 exonuclease activity mediates VEEV translation through the upregulation of ISGs during infection but is not one of the three key regulators of the host antiviral mechanism against VEEV [38,40]. CXCL10 serves to recruit immune cells to the site of infection, and has been shown to be upregulated in mice post-infection [41]. In contrast, ISG15 does not have as defined a role in direct antagonism of VEEV as the aforementioned ISGs and chemokine but is consistently shown to be upregulated during infection, such as in the case of VEEV TrD-infected astrocytoma cells [42] and VEEV TC-83-infected primary human umbilical vein endothelial cells [43]. In our RT-qPCR analysis, GAPDH served as a proper endogenous control as determined by nonsignificant differences in Ct values regardless of infection and timepoint (Figure 1E). ISG production occurred as early as 16 hpi at high MOI, 10, and 24 hpi at low MOI, 0.1 (Figure 1D), with a greater fold change in mRNA levels 24 hpi with MOI 10. These data suggest that the Type I IFN response as indicated by ISG production happened during late VEEV infection independent of STING signaling via phosphorylation at residue Ser366.

STING siRNA was utilized as an orthogonal approach to determine whether ISG production during VEEV infection occurs independent of STING. HMC3 cells were transfected with siRNA targeting STING (siSTING) or a negative control (siNeg) prior to infection with VEEV TC-83 (MOI 1) for 8 and 16 h. Western blot and RT-qPCR analysis confirmed that STING expression had been successfully reduced (Figure S1A,C). VEEV replication was not altered by the reduction in STING (Figure S1B). Changes in CXCL10 gene expression, but not IFIT2, were observed in the VEEV-infected STING knockdown cells compared to the siNeg control (Figure S1C). These data indicate that STING partially contributes to ISG production during late VEEV infection but does so in the absence of canonical STING Ser366 phosphorylation.

3.2. STING Activation Inhibits VEEV Replication

Given the success of previous studies that primed the IFN pathway to inhibit viral replication, we sought to determine whether priming human microglial cells through STING signaling could reduce VEEV infectious virus. HMC3 cells underwent dsDNA treatment (0.5 μg) or were left untreated (No Tx; diluent in media) for six h followed by mock infection or infection with VEEV TC-83 V5-C at increasing MOIs (0.1, 1, 5, and 10) for 17 h (Figure 2A).

Protein lysates underwent immunoblotting to measure relative expression for STING phosphorylation (Ser366), as an indicator of cGAS-STING activation. Cells that underwent no infection and no treatment (mock) or dsDNA pretreatment only (dsDNA only) served as the negative and positive controls, respectively. HMC3 cells pretreated with dsDNA exhibited STING phosphorylation at all MOIs (Figure 2B,C), whereas cells that underwent infection alone had little to undetectable STING phosphorylation. Relative total STING was significantly reduced in cells that underwent dsDNA treatment and higher MOIs (5 and 10) compared to the mock (Figure 2D). Viral reduction, as determined by immunoblotting for V5-C, revealed decreased capsid protein levels for cells pretreated with dsDNA compared to the untreated group (Figure 2B), with the greatest decrease observed at MOI 0.1.

The impact of dsDNA on VEEV was also assessed by examining viral titers via plaque assays and viral RNA by RT-qPCR. Reduction in infectious virus due to dsDNA pretreatment occurred in an MOI-dependent manner, with the greatest reduction in viral titers occurring at the lowest MOI 0.1 (2.12 log₁₀-fold change, p < 0.001) compared to the untreated group (Figure 2E). Likewise, the same MOI-dependent trend was observed when assessing intracellular viral RNA levels, in which the greatest reduction in VEEV RNA was observed in the dsDNA-pretreated group at MOI 0.1 (1.87 log₁₀-fold change, p = 0.0062) (Figure 2F). These data indicate that priming microglial cells with dsDNA can activate STING via phosphorylation at Ser366 and is able to reduce VEEV infectious virus.

Additionally, we expanded our testing to include DMXAA, a mouse-specific STING agonist [44], to treat mouse embryonic fibroblasts (MEFs). Using non-toxic concentrations of DMXAA (Figure 3A), we observed a concentration-dependent impact on VEEV viral titers (Figure 3B) and confirmed this reduction through immunoblotting for V5-C as a means to measure capsid production (Figure 3C). MEFs treated with 100 μM DMXAA had the greatest effect on infectious virus (3.4 log₁₀-fold change, p < 0.0001) (Figure 3B) and a lack of a perceivable band for V5-C on Western blot (Figure 3C). These results provide further evidence that priming STING signaling reduces VEEV replication.

3.3. STING Activation Inhibits Viral Replication of Other Alphaviruses

Given that STING activation inhibited VEEV TC-83 V5-C viral replication and has been previously reported to reduce CHIKV replication [22,23,45], the impact of STING activation prior to infection by other alphaviruses was explored. We confirmed the impact of dsDNA treatment on VEEV infectious virus through measurement of viral titers via a plaque assay using VEEV virulent strain TrD, and expanded testing to include a New World alphavirus, EEEV, and Old World alphaviruses, CHIKV and SINV. The same MOI-dependent reduction in infectious virus was observed in cells that underwent dsDNA pretreatment, with the greatest decrease in viral titers observed at the lowest MOI (0.1) for TrD (1.76 log₁₀-fold change, p < 0.0001) (Figure 4A), EEEV GA Fatal (2.54 log₁₀-fold change, p = 0.0004) (Figure 4B), and CHIKV (2.15 log₁₀-fold change, p = 0.0226) (Figure 4C). Microglial cells infected with SINV had a similar trend in viral inhibition, but the lowest MOI that was statistically significant was MOI 1 (1.4 log₁₀-fold change, p = 0.009) (Figure 4D). These results indicate that STING activation can broadly impact multiple New and Old World alphaviruses.

3.4. VEEV Suppresses STING Ser366 Phosphorylation

To determine whether the timing of dsDNA treatment in human microglia was crucial to enable viral inhibition, we infected cells with VEEV TC-83 V5-C at increasing MOIs (0.1, 1, 5, and 10) or mock-infected them for 6 h followed by post-treatment with dsDNA (0.5 μg total) or diluent in media (no treatment) for 17 h, and supernatants and lysates were collected at 23 hpi (Figure 5A).

When administered post-infection, dsDNA treatment did not have an impact on infectious virus, as shown by the lack of change in viral titers between the untreated and dsDNA treatment groups (Figure 5B). These results were recapitulated through immunoblotting for V5-C, in which both groups had relatively equal band intensity, signifying no difference in capsid production (Figure 5C). TBK1 recruitment and phosphorylation play a critical role in the activation of the STING-TBK1-IRF pathway by phosphorylating STING at Ser366 [36,46]. We probed for pTBK1 (Ser172) as a measure of Type I IFN induction and precursor to STING phosphorylation, in which we observed increased pTBK1 for cells post-treated with dsDNA at all MOIs compared to the untreated group (Figure 5C). In contrast, we observed a reduction in pSTING in the dsDNA-treated group in an MOI-dependent manner, with the greatest decrease in pSTING levels observed at higher MOIs (5 and 10) (Figure 5C,D). Quantitation of relative pSTING (Ser366) and STING revealed that cells infected at MOI 0.1 had similar relative phosphorylation and total protein levels as the dsDNA control (Figure 5D,E). These results indicate that VEEV inhibits STING phosphorylation, but not TBK1 phosphorylation, in the presence of dsDNA post-infection.

3.5. VEEV Nonstructural Proteins, Capsid, and Structural Polyprotein Without Capsid Alone Do Not Inhibit STING (Ser366) Phosphorylation

To assess the VEEV protein(s) responsible for inhibiting STING phosphorylation (Ser366), we transfected HMC3 cells with individual nonstructural proteins (nsP1, nsP2, nsP3, or nsP4), the capsid protein, or the structural polyprotein without capsid (E3-E2-6K/TF-E1) for 24 h followed by dsDNA (1 μg) treatment for 17 h. Protein lysates were assayed for pSTING via Western blot to measure changes in Ser366 phosphorylation levels due to VEEV proteins. HMC3 cells that underwent no transfection (mock) or transfection with the pcDNA3.1 backbone served as the negative controls and cells transfected with backbone and challenged with dsDNA served as the positive control for pSTING (Ser366) expression. Protein lysates from the dsDNA post-treatment challenge (Figure 5C) for VEEV TC-83 V5-C (MOI 10)-infected HMC3 cells were used as a positive control for pSTING inhibition. Western blot analysis revealed unchanged pSTING levels when cells were transfected with VEEV nonstructural proteins, capsid, or E3-E2-6K/TF-E1 (Figure 6). These results indicate that expression of these proteins individually is not capable of suppressing STING (Ser366) phosphorylation.

4. Discussion

This study sought to elucidate the dynamic viral–host interplay between VEEV and STING, having utilized human microglial cells as our model for characterization, due to ongoing interest in the neurological complications associated with VEEV infection. Microglial cells have been shown to be the primary activators of the cGAS-STING signaling pathway [47]. Previous studies have successfully explored prophylactic and therapeutic options for VEEV viral inhibition by inducing a Type I IFN response in a STING-dependent manner [22,23]; however, the role that STING plays during VEEV infection remains unclear. The data presented here confirm that STING-pathway activation through pretreatment of cells with dsDNA leads to viral inhibition during VEEV infection (Figure 2 and Figure 3A), and the timing of therapeutic intervention is paramount, as evident in the lack of impact that dsDNA had on viral titers when administered 6 h post-infection (Figure 5B). Two synthetic small molecules, G10 [22] and C11 [23], were previously identified via high-throughput screening and shown to indirectly activate STING. VEEV replication was significantly inhibited in human fibroblasts when G10 or C11 was given prophylactically [22,23]. C11 exhibited therapeutic potential through reduction in VEEV infectious virus when cells were treated up to 12 hpi; however, antiviral activity was abrogated when C11 was administered 24 hpi [23]. Our studies coincide with the previous literature regarding the importance that timing of intervention plays in the drug’s antiviral activity. Irrespective of how drug mechanisms of action can differ among even similar compounds, the timing of addition is also dependent on the virus’s ability to establish productive infection prior to eliciting an innate immune response.

It is imperative to note that while we opted to use microglia as our model of infection, VEEV’s primary targets in the CNS are neurons [48,49]. Previous studies have indicated astrocytes to also be infected in vivo [50]; however, microglia have been noted as having sustained activation, thus driving neuroinflammation [51]. Cells within the CNS have low basal levels of STING expression, and it is often proposed that microglia are the predominant expressors of STING during neurological illness and disease [52]. The impact of STING signaling on recovery from post-traumatic brain injury in mice with a system-wide microglia STING knockout has been examined. Mice lacking functional STING fared better than mice with intact signaling through decreased cortical tissue loss and improved motor function, suggesting a protective role of STING [53]. There are conflicting reports in the literature as to whether microglia are the primary driver of STING activation, and in fact, some papers suggest endothelial cells and neurons to be targets in neurodegenerative diseases [54]. It would be beneficial to expand testing on the role that STING plays during VEEV infection to other cell types, including neurons and astrocytes. Additionally, studying STING signaling in vivo during VEEV infection should provide further insights into the importance of this pathway for neuropathogenesis and neurological sequelae.

Even though previous research has identified STING agonists capable of inhibiting VEEV replication, mechanistic insights into STING signaling and alphaviruses have been derived from studies with CHIKV. VEEV has been shown to induce mitochondrial dysregulation during early infection [24,25], and therefore, we suspected the cGAS-STING pathway to be activated. Our studies showed upregulation of multiple ISGs in the absence of STING Ser366 phosphorylation at the timepoints analyzed (Figure 1). It is possible that STING phosphorylation occurs during these timepoints, but limitations due to assay sensitivity may render STING phosphorylation below detectable levels, and activation may be happening earlier during infection. Timepoints prior to 2 hpi would need to be included in future studies to ascertain whether STING is being activated during early infection. In addition, loss of STING through siRNA demonstrated a partial dependence of CXCL10, but not IFIT2, expression on STING (Figure S1C), indicating that STING partially contributes to ISG induction. As STING phosphorylation is only one marker of STING activation, it is possible that STING is induced in VEEV-infected cells independent of STING Ser366 phosphorylation. Indeed, noncanonical activation of STING has been shown to be independent of cGAS and Ser366 phosphorylation [55]. Noncanonical activation of STING can occur upon detection of DNA damage, activating the DNA sensing proteins ataxia telangiectasia mutated (ATM), poly-ADP-ribose polymerase 1 (PARP1), and interferon-inducible protein 16 (IFI16). These proteins facilitate the assembly of a STING complex containing STING, p53, IFI16, and the E3 ubiquitin ligase tumor necrosis factor receptor-associated factor 6 (TRAF6) at the ER. TRAF6 mediates the poly-ubiquitination of STING, ultimately leading to the activation of NF-κB and stimulation of NF-κB-dependent gene expression. Future studies should explore noncanonical activation of STING in VEEV-infected cells.

CHIKV infection alone is sufficient to stimulate STING (pSTING Ser366) [45]. A study by Geng et al. (2020) discovered that macrophages harvested from Sting^Gt mice, lacking functional STING protein [56], were susceptible to CHIKV infection, resulting in increased viral titers. Interestingly, these macrophages were still able to mount a Type I IFN response in the absence of STING [20]. Another study, conducted by Akhrymuk et al. (2016), investigated pattern recognition receptors in mouse fibroblasts responsible for Type I IFN signaling after VEEV and SINV infection as measured by IFN-β production [16]. It was determined that both cytosolic receptors RIG-I and MDA5 are responsible for Type I IFN activation, likely through the detection of viral ssRNA and dsRNA, respectively; however, other PRRs responsible for this signaling could not be definitively ruled out due to limitations with the cell type used as the infection model [16]. The findings in our study related to ISG production corroborate that Type I IFN signaling in VEEV-infected microglial cells is likely due to activation of other immune pathways, such as RIG-I and MDA5 (Figure 7A). Confirmatory studies need to be conducted, such as evaluating ISG production in single and double knockouts for RIG-I and MDA5 in microglia cells and other cell types (neurons, astrocytes).

Our study is not without limitations, as most of our experiments were performed with VEEV TC-83. VEEV TC-83 has two distinct attenuating mutations, one in the 5′ UTR region (G3A) and the other in the E2 glycoprotein [57]. The G3A mutation in the 5′ UTR region leads to an increased IFN sensitivity [57,58]. The 5′ UTR region in TC-83 allows the virus to be recognized by host ISGs, such as IFIT1, making it IFN-sensitive, whereas the intact hairpin structure near the 5′ cap in the TrD strain affords the virus immune evasion capabilities by preventing interaction with IFIT1 [59]. HMC3 cells pretreated with dsDNA and infected with the virulent VEEV TrD strain (Figure 4A) exhibited similar patterns of reduction in viral titers to what was observed in VEEV TC-83-infected cells (Figure 2E), but with a less pronounced difference at higher MOIs. Despite pretreatment of cells with dsDNA successfully dampening viral replication, STING activation and/or viral antagonism may differ between the strains due to differences in response to IFNs and overall immune evasion. Viral antagonism of the cGAS-STING pathway by VEEV TC-83 has been demonstrated here, but future studies should include delineation of whether VEEV TrD exhibits a similar phenotype.

Another limitation of our study is the use of dsDNA as a means to activate the cGAS-STING pathway, as multiple other cytosolic sensors (i.e., DDX41, IFI16) can elicit a Type I IFN response to dsDNA including indirect activation of RIG-I [60,61]. Future studies will need to be conducted to determine the role that STING activation plays in reducing VEEV infectious virus independent of these other pathways.

The bulk of our assays looked at downstream regulation of the signaling cascade with a predominant focus on STING phosphorylation (Ser366) and ISG production; however, it is possible that viral antagonism can be occurring earlier in the pathway such as prior to the cell’s detection of mtDNA released within the cytosol, cGAS activation, and subsequent production of the secondary messenger molecule (Figure 7B). Studies targeting each step of the cGAS-STING pathway, starting with confirmation of mtDNA release into the cytosol, would provide crucial insight into which step(s) are undergoing viral antagonism. By deciphering whether mtDNA is being released into the cytosol post-VEEV-infection, cell lysates can undergo fractionation followed by qPCR analysis to quantitate mtDNA in the different fractions [62,63]. Additional experiments include looking at early signs of apoptosis through Annexin V staining [63], as well as evaluating whether cGAMP is being released through assays such as ELISAs or liquid chromatography–mass spectroscopy (LC-MS) [64]. The LC-MS methodology for measuring cGAMP, as described by Skeldon et al. (2025) [64], could be used in conjunction with the dsDNA challenge to determine whether the messenger molecule is able to interact with STING. Due to this assay using a His-tagged STING to pull down cGAMP, if viral antagonism is preventing this interaction, then we would see a decrease in cGAMP levels in lysates from VEEV-infected cells in an MOI-dependent manner when challenged with dsDNA. Studies centered around this activation and messenger molecule production are needed to determine whether VEEV regulates STING signaling upstream of STING activation via phosphorylation.

Additionally, STING is activated through phosphorylation by TBK1 after translocation of STING from the ER to the ER–Golgi intermediate complex (ERGIC). cGAMP binding to STING is a crucial step for the protein to exit the ER, but this transition from the ER to the ERGIC is mediated by the interaction with STING ER exit protein (STEEP) [65]. It is possible that VEEV inhibits translocation of STING from the ER, but experiments addressing this question will need to be conducted (Figure 7C). Confocal microscopy and co-immunoprecipitation would be useful tools to evaluate whether STING is being translocated from the ER to the ERGIC via STEEP, for direct interaction between STING and STEEP and STING and TBK1, and whether viral proteins may be interacting with any of these proteins.

To our knowledge, this study is the first to demonstrate VEEV’s ability to antagonize STING signaling through inhibition of STING (Ser366) phosphorylation. Microglial cells infected with VEEV TC-83 V5-C had an MOI-dependent decrease in protein expression for pSTING when challenged 6 h post-infection with dsDNA (Figure 5C). Earlier research revealed that CHIKV inhibits the cGAS-STING pathway through degradation of cGAS during early infection of human foreskin fibroblasts [21]. cGAS degradation occurs via an ATG7-dependent autophagy mechanism [21], with the induction of autophagy facilitating CHIKV replication [66]. Degradation of cGAS was not observed in VEEV-infected cells. The use of divergent strategies by alphaviruses to suppress STING signaling is not surprising, given that Old World and New World alphaviruses are known to utilize different mechanisms to evade immune responses to facilitate viral replication [67]. For example, alphaviruses inhibit RNA polymerase II transcription via unique processes to reduce the ability of host cells to respond to infection [68,69]. Old World alphavirus SINV and CHIKV nsP2 induces degradation of RNA polymerase II [70], whereas VEEV capsid blocks nucleocytoplasmic transport to suppress cellular transcription [71].

CHIKV nsP1 was found to directly interact with STING, serving a dual function in signaling inhibition and increasing nsP1 protein levels within the cell [21]. While our data does not show a similar pattern of cGAS reduction (Figure 1A,B), our findings of pSTING (Ser366) inhibition led us to assess which, if any, VEEV proteins led to this inhibition. Microglial cells transfected with VEEV capsid, structural polyprotein without capsid (E3-E2-6K/TF-E1), or nonstructural proteins (nsP1, nsP2, nsP3, or nsP4) did not reduce STING phosphorylation after challenge with dsDNA (Figure 6), and therefore is a key limitation to this study. It is possible that cGAS antagonism can happen during VEEV infection but through other modes of action outside of protein degradation, such as protein sequestration, preventing the downstream signaling cascade from taking place. Alternatively, degradation may be an even earlier event outside of the timepoints we test, such as prior to 2 h post-infection. A combination of structural and nonstructural proteins might be needed to decrease pSTING (Ser366), but future experiments need to be conducted to elucidate the mechanisms behind VEEV-induced inhibition of STING phosphorylation.