Diminished Innate Antiviral Response to Adenovirus Vectors in cGAS/STING-Deficient Mice Minimally Impacts Adaptive Immunity

Ad5CiG and Ad5βGal were previously described (52, 53) and were grown on a large scale in HEK-293 cells according to standard protocols, purified through two rounds of CsCl gradient ultracentrifugation, and stored at −80°C in storage buffer (10 mM Tris, 2 mM MgCl₂, 4.0% sucrose [pH 7.5]). Virus particle numbers were quantified by spectrophotometric detection of intact virions at an optical density at 260 nm (OD₂₆₀) (10¹² particles/OD₂₆₀ unit).

Eight- to twelve-week-old male and female C57BL/6 and C57BL/6J-Tmem173^gt/J (STING KO) mice were obtained from Jackson Laboratories, cGAS^−/− C57BL/6 mice were generously provided by H. Virgin, and IRF3^−/− mice were generously provided by T. Taniguchi through Riken BRC.

Bone marrow cells were extracted from the femurs and tibiae of mice. For bone marrow-derived macrophages (BMMOs), red cells were removed with a lysis solution (0.15 M NH₄Cl, 1 mM KHCO₃, 0.1 mM EDTA). Bone marrow cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 20% fetal bovine serum (FBS) and 25% supernatant derived from confluent L929 cells. On day 7, immature macrophages were collected. Bone marrow-derived dendritic cells (BMDCs) were differentiated in RPMI medium containing 10% FBS, 1% l-glutamine, and beta-mercaptoethanol (BME) in the presence of 20 ng/ml granulocyte-macrophage colony-stimulating factor (GM-CSF), as previously described (54). Briefly, bone marrow was harvested and plated into petri dishes at a density of 4 × 10⁵ cells/ml, 10 ml/dish. At day 3, 10 ml of medium was added, and at days 6 and 8, 10 ml of medium was replenished. Immature, nonadherent BMDCs were gently harvested at day 10. The typical purity of each population was >90%, as determined by fluorescence-activated cell sorter (FACS) analysis of CD11b and CD11c surface markers, respectively.

BMMOs were harvested at day 7, BMDCs were harvested at day 10, and cells were plated at densities of 5 × 10⁵ cells/ml in 12-well plates (Western blot and flow cytometry analyses) (5 × 10⁵ cells/well) and 10⁶ cells/ml in 6-well plates (for RNA extraction) (2.5 × 10⁶ cells/well). Twenty-four hours after plating, cells were infected with Ad5CiG at 20,000 virus particles (vp)/cell and diluted in Opti-MEM. For Western blot analysis, cells were harvested at 2, 4, and 6 h p.i. Flow cytometry was performed at 24 h p.i., and RNA extraction was performed at 6 h p.i.

Mice were infected by intraorbital intravenous (i.v.) injections with 10¹¹ vp/mouse of Ad5βGal diluted in DMEM. Mice were euthanized at 5 h, 24 h, and 28 days postinfection for liver and blood serum harvest. All experiments were performed in accordance with Weill Cornell Medical College IACUC protocols.

Antibodies were purchased from BioLegend (San Diego, CA). BMMOs and BMDCs were prepared from bone marrow as described above. For phenotypic analysis, cells were surface stained with the following monoclonal antibodies (MAbs): anti-CD11b allophycocyanin (APC) (catalog number 101212), anti-CD11c Pac Blue (catalog number 117322), anti-CD11c APC (catalog number 117310), anti-CD86 phycoerythrin (PE) (catalog number 105008), and anti-major histocompatibility complex class II (MHC-II) peridinin chlorophyll protein (PerCP)-Cy5.5 (catalog number 116416). After staining, washing, and fixation in 1% paraformaldehyde, cells were analyzed by using a FACSCanto flow cytometer (BD Biosciences). Data analysis was performed by using FlowJo software, and 12,000 events per sample were analyzed.

Whole-cell extracts were prepared by washing cells twice with ice-cold phosphate-buffered saline (PBS) and incubating them in lysis buffer (50 mM Tris [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1% NP-40) with the addition of Phosphatase Inhibitor Cocktails 1 and 2 (catalog numbers P2850 and P5726; Sigma) and protease inhibitors (30 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride [PMSF], 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 μg/ml pepstatin, 1 mM benzamidine) for 30 min at 4°C on a rocking platform before scraping and transferring cells to tubes. The lysates were cleared by centrifugation at 13,000 × g for 20 min at 4°C, and protein quantification was performed with the DC protein assay kit (Bio-Rad Laboratories).

For Western blot analysis, 20 μg total protein was separated by using standard 10% SDS-polyacrylamide gels and transferred onto polyvinylidene difluoride (PVDF) membranes (Immobilon P; Millipore). All blots were blocked in 5% skim milk in Tris-buffered saline (TBS) plus Tween (0.1%) at room temperature for 1 h. Phospho-IRF3 (Ser396) (catalog number 4947), phospho-STAT1 (58D6) (Tyr701) (catalog number 9167), beta-actin (catalog number 4967), total IRF3 (catalog number 4302), STING (catalog number 3337), TBK1 (catalog number 3503), phospho-NF-κB p65 (pNF-κB p65) (Ser536) (catalog number 3033), pTBK1 (Ser172) (catalog number 5483), green fluorescent protein (GFP) (catalog number 2956), and horseradish peroxidase (HRP)-linked anti-rabbit IgG (catalog number 7074) antibodies were obtained from Cell Signaling. Primary antibodies were used at a dilution of 1:2,000 to 1:3,000 in 5% bovine serum albumin (BSA)–TBS. The HRP-linked secondary antibody was diluted 1:4,000 in 5% milk–Tween–TBS. Detection was done with the Luminator Crescendo Western HRP substrate (Millipore).

Total cellular mRNA was isolated by using RNAzol reverse transcriptase (RT) (Molecular Research Center) according to the manufacturer's instructions. For RT quantitative PCR (RT-qPCR), a two-step protocol was employed: first, cDNA was synthesized from 2 μg total RNA in a volume of 20 μl by using random hexamer primers with the Maxima first-strand cDNA synthesis kit (Fermentas), and second, amplifications were carried out with a total volume of 15 μl by using the Maxima SYBR green/ROX qPCR master mix (Fermentas) in an Applied Biosystems Prism 7900H sequence detection system with SDS 2.1 software. Cycles consisted of an initial incubation step at 95°C for 10 min; 40 cycles at 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s; and a melting-curve analysis cycle. Data acquisition was performed during the extension step. All determinations were performed in technical triplicate. Nontemplate and no-RT controls were run with every assay and had cycle threshold (C_T) values that were significantly higher than those of experimental samples or were not detected. The relative abundance of each mRNA was calculated by the ΔΔC_T method (55, 56), normalizing to Tata-binding protein (TBP) expression with standardization to one reference sample, as indicated. For comparisons between cell lines, or in late-stage virus infections, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a normalization standard. Sequences of primers are available upon request.

Liver fragments were Dounce homogenized in lysis buffer containing 10 mM NaCl, 10 mM Tris (pH 8.0), 10 mM EDTA (pH 8.0), and 0.5% SDS, followed by proteinase K digestion. Crude homogenates were cleared by double centrifugation. Total DNA was purified by using NaCl salting-out and ethanol precipitation protocols (57). Following DNA purification and resuspension in 100 μl H₂O, OD₂₆₀ DNA concentrations were determined by using a NanoDrop instrument. Five nanograms of sample DNA in a final reaction mixture volume of 15 μl (Maxima SYBR green/ROX qPCR master mix [Fermentas] system) was used for each assay. Ad5 hexon and cellular control primers (IFN-β) were used to characterize each DNA sample. Normalization to the cellular genomic standard was used to derive the relative viral DNA yield/host genome. Assays were carried out with an Applied Biosystems Prism 7900H sequence detection system with SDS 2.1 software. Cycles consisted of an initial incubation step at 95°C for 10 min; 40 cycles at 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s; and a melting-curve analysis cycle. Data acquisition was performed during the extension step. Data points represent the averages of results from biological and technical triplicates.

The Tropix Galacto-Star chemiluminescent reporter gene assay (catalog number T1012; Applied Biosystems) was used for β-galactosidase (β-Gal) detection in the livers of infected mice at 24 h and 28 days p.i., according to the manufacturer's protocol. Liver fragments were sonicated on ice for 20 s in 1 ml lysis buffer with PMSF. Homogenates were spun down twice at 13,000 rpm for 15 min. The supernatant was heat inactivated at 48°C for 1 h to neutralize endogenous β-galactosidase activity and spun down at 13,000 rpm for 15 min. The protein concentration was determined by a Bradford assay (58). Ten microliters of the supernatant per well was loaded into 96-well Immulon assay plates and incubated with 100 μl reaction buffer for 30 min at room temperature. The luminescence in each well was measured for 1 s. Each sample was loaded in triplicates, and data were averaged. The background given by the lysis buffer was subtracted, and results were expressed per microgram of liver protein.

Mouse blood samples isolated at 28 days postinfection were left at room temperature for 15 min to allow clot formation and then centrifuged at 13,000 rpm for 10 min for serum separation. Mouse serum was immediately flash frozen and thawed when used for biological assays. Briefly, serum neutralization assays were carried out as follows. HeLa cells were plated out at 2.5 × 10⁴ cells/well in 96-well plates 24 h prior to infection. On the following day, serum was thawed, and 1:4 serial dilutions in DMEM were generated for each serum sample. Ad5βGal, diluted in DMEM for a final delivery of 10³ particles/cell, was incubated with a serum dilution (technical triplicates for each sample) for 1 h at 37°C. Following virus/serum incubation, the mixture was incubated with HeLa cells for 30 min at 37°C. Medium was removed, and cells were washed with fresh medium and incubated in 100 μl fresh medium for 24 h. At 24 h p.i., plates were spun down at 2,000 rpm for 5 min, and medium was removed. Cell lysates were harvested in a volume of 100 μl lysis buffer. Ten microliters of the supernatant was applied for a Tropix Galacto-Star activity assay as described above. The background generated by the lysis buffer was subtracted from the result for each assay point, and results were expressed as luminescence units.

Mouse serum derived from blood samples as described above was harvested at the indicated times and stored at −80°C until it was evaluated in cytokine assays. Serum was centrifuged for 15 min at 13,000 rpm to remove any residual cells or debris and then evaluated for cytokine secretion with Meso Scale Discovery (MSD) (Rockville, MD) technology using a mouse proinflammatory 7-plex tissue culture kit (catalog number K15012B-1). Kits were run according to the manufacturer's instructions. Samples were run in duplicate, at the Weill Cornell Medical School CTSC facility, by using Meso Scale Discovery Sector Imager 2400 (SI2400).

Mouse IFN-β was measured by using a mouse IFN-β (PBL Interferon Source) enzyme-linked immunosorbent assay (ELISA) kit (catalog number 42400; R&D Systems). Serum from mice was obtained as described above and stored at −80°C until it was evaluated for cytokine secretion. Kits were run according to the manufacturer's instructions. The absorbance at 450 nm was read within 5 min after the addition of the stop solution. Each sample was assayed as a technical duplicate.

Data were expressed as means ± standard errors of the means. Statistical analysis was performed with Student's t test. A P value of <0.05 was considered significant.

Short hairpin RNA (shRNA) knockdown assays targeting cGAS, STING, and TBK1 were previously used to identify these proteins as components of the antiadenovirus DNA recognition response in both human and murine cell lines (23, 37). To establish the essentiality of STING to adenovirus-induced recognition responses in primary murine macrophages, bone marrow-derived CD11b⁺ macrophages were isolated from WT C57BL/6 and Goldenticket mice containing a null STING allele (59). As a negative control, BMMOs were also isolated from C57BL/6 IRF3 knockout (IRF3^−/−) mice. BMMOs from each strain were infected with first-generation (E1⁻/E3⁻) replication-defective Ad5CiG (cytomegalovirus [CMV]-driven chloramphenicol acetyltransferase internal ribosome entry site [IRES] GFP [CATiresGFP] reporter gene cassette) (2 × 10⁴ vp/cell). Cell protein lysates were harvested over a 6-h time course. Western blot analysis of lysates from WT BMMOs revealed a time-dependent increase in the phosphorylation of TBK1 and IRF3 (Fig. 1A). TBK1 and IRF3 are primary response markers, which undergo activation as a direct consequence of viral DNA recognition by the cGAS/STING PRR machinery. In contrast to WT BMMOs, infection of STING^−/− BMMOs did not result in the phosphorylation of TBK1 or IRF3. Lysates generated by infection of BMMOs from negative-control IRF3^−/− mice indicated that the primary DNA recognition cascade was intact (enhanced phosphorylation of TBK1), which is consistent with IRF3 being a downstream substrate for TBK1.

IRF3 phosphorylation, dimerization, and translocation to the nucleus are required for transcriptional activation of IFN-β. IFN-β expression and secretion contribute to the secondary antiviral signaling response. Activation of type I IFN receptor (IFNRI) triggers the phosphorylation of STAT1; therefore, the induction of ^pTyr701STAT1 provides a marker for the secondary antiviral signaling response. Lysates harvested from infected STING^−/− or IRF3^−/− BMMOs were severely restricted in the phosphorylation of STAT1 compared to BMMOs from WT mice. Our data indicate that both primary and secondary signaling responses to rAdV were compromised in STING-deficient and IRF3^−/− BMMOs.

As indicated above, the combination of primary and secondary antiviral signaling leads to transcriptional activation of proinflammatory genes and ISGs. To determine the influence of the STING deficiency on antiviral gene induction, total RNA harvested at 6 h postinfection was characterized by a two-step RT-qPCR assay (Fig. 1B). Levels in all samples were normalized to the level of TBP by using the ΔΔC_T method, where the value for mock WT sample 1 was set as an arbitrary unit of 100. As a quantitative reference, the unnormalized average C_T for each mock WT primer pair is indicated above the expression bar (Δ1 log ∼3 C_T, and C_Ts of >30 indicate relatively low-abundance transcripts). Transcripts corresponding to IFN-β, ISG15, ISG54, VIP, CXCL9, TNF-α, IL-1β, and IL-6 were strongly induced following infection of BMMOs from WT mice. In contrast, induction of these transcripts was not observed following infection of STING^−/− or IRF3^−/− BMMOs. We also found that basal levels of several transcripts (ISG15, ISG54, viperin (VIP), TNF-α, and IL-1β) were significantly reduced in BMMOs from both knockout strains. Macrophage maturation following virus uptake results in the upregulation of the costimulatory protein CD86 at 24 to 48 h postinfection (Fig. 1C and D). CD86 surface expression levels following infection of STING^−/− or IRF3^−/− BMMOs were significantly lower than levels found with WT BMMOs. Based on these criteria, primary and secondary signaling markers, induction of antiviral gene expression, and upregulation of costimulatory CD86, BMMOs from STING-deficient mice were essentially nonresponsive to rAdV infection.

We next obtained BMMOs from cGAS (MB21D1) knockout mice (60), allowing us to confirm the role of cGAS in the antiviral response to rAdV. cGAS^−/− BMMOs infected with rAd5CiG were harvested over a 6-h time course for protein lysates or at 24 h postinfection to assess CD86 surface expression. Western blot analysis of cGAS BMMO-derived protein lysates revealed a complete lack of primary and secondary activation events (pTBK1, pIRF3, and pSTAT1) in cGAS KO compared to WT infections (Fig. 2A). Using CD86 upregulation as a maturation marker, cGAS BMMOs failed to mature in response to rAdV infection (Fig. 2B). These experiments demonstrate that the cGAS/STING DNA sensor complex functions as the dominant PRR pathway for rAdV infection of murine BMMOs.

While macrophages contribute to virus clearance through phagocytic elimination of virus and secretion of inflammatory chemokines, virus uptake by dendritic cells (DCs) results in the maturation of DCs into highly efficient professional antigen-presenting cells (APCs). Bone marrow progenitor cells grown in the presence of GM-CSF differentiate into a CD11c⁺ immature dendritic cell type (BMDCs). Using cultures of GM-CSF-differentiated BMDCs isolated from WT, STING^−/−, IRF3^−/−, and cGAS^−/− mice, we determined how these mutations influenced the BMDC antiviral response. Day 10 BMDCs were infected with 2 × 10⁴ Ad5CiG vp/cell, and protein lysates were harvested at the indicated times over a 6-h time course (Fig. 3A). Western analysis of WT BMDC protein lysates indicated that IRF3, TBK1, and STAT1 are phosphorylated in a time-dependent manner following exposure to rAd5CiG. In contrast, analysis of lysates from infected cGAS^−/− and STING^−/− BMDCs revealed levels of pIRF3, pTBK1, and pSTAT1 that were similar to those in uninfected mock samples. Consistent with observations made in BMMOs, infection of IRF3^−/− BMDCs resulted in the production of pTBK1 in a time-dependent manner. Additionally, low levels of pSTAT1 were present in infected IRF3^−/− lysates. These data indicate that primary and secondary antiviral signaling responses were compromised in cGAS^−/− and STING^−/− BMDCs.

We next harvested total RNA from mock- or rAdV-infected BMDCs at 6 h postinfection. Two-step RT-qPCR was carried on all RNA samples as described above. The first observation taken from the RT-qPCR assay (Fig. 3B) was that transcript induction in each of the knockout strains was significantly compromised compared to that in infected WT BMDCs. Transcripts corresponding to IFN-β, ISGs, and proinflammatory transcripts were all impacted by these mutations. The IL-1β transcript was not appreciably induced by rAd5CiG infection of WT BMDCs. A second observation taken from these assays was the existence of strain- and transcript-specific induction trends. For each strain, in comparisons between mock- and rAdV-treated samples, there were examples of low-level but significant transcript induction (notably with cGAS ISGs). The final observation taken from the RT-qPCR assay of BMDCs was that we did not find significant differences in basal levels of any ISG. Of note, low-level induction of IFN-β mRNA following infection of IRF3^−/− BMDCs was observed, which is consistent with the low level of pSTAT1 observed in the BMDC Western analysis of protein lysates (Fig. 3A).

Upregulation of costimulatory markers is an important functional measurement of BMDC maturation in response to pathogen activation. BMDCs at 24 h postinfection were analyzed by flow cytometry for increased CD86 surface expression levels for each knockout strain (Fig. 3C and D). Consistent with compromised primary and secondary signaling, levels of CD86 surface expression from cGAS- or STING-deficient BMDCs exposed to rAd5CiG were at mock-infected levels. Therefore, at the levels of primary and secondary signaling, induction of antiviral transcripts, and upregulation of costimulatory molecules, we have established that the cGAS/STING DNA-sensing cascade is a critical component of the antiviral recognition response in primary murine BMDCs. We note that in BMMOs, these mutations resulted in an essentially nonresponsive anti-rAdV phenotype, while in BMDCs, the antiviral responses were severely compromised but not completely eliminated.

In the murine systemic infection model, administration of rAdV exposes virus to serum factors, including native antibody and factor X, which bind virus. The vast majority of virus localizes to the liver, where a heterogeneous population of cells, including hepatocytes, endothelial cells, macrophages (Kupffer cells), and dendritic cells, is exposed to rAdV. To determine the role of cGAS and STING in the in vivo antiviral recognition response, intravenous administration of 10¹¹ rAd5βGal particles per mouse was carried out for each mouse strain. Mice were euthanized at 5 h postinfection for tissue and serum harvest (n = 6 to 9 mice/strain). We first characterized liver protein lysates for the activation of the IRF3 primary response marker and found that levels of pIRF3 from WT lysates were below detectable limits (Fig. 4A). From our experience, murine hepatocyte cell lines are nonresponsive to rAdV due to a lack of cGAS/STING expression (37). Hepatocytes make up the majority of the liver cell mass; therefore, the nonresponsive pool of cells may dilute and obscure the signal derived from responsive cells. When screening for the IFN secondary response marker pSTAT1, we found detectable induction of pSTAT1 in liver lysates from infected WT mice, whereas lysates from STING^−/− mice had lower basal and induced levels of pSTAT1. The data from Western analysis imply that a low level of interferon expression occurred following infection of WT mice, which was compromised in knockout mice. An IFN-β ELISA was employed to directly quantitate serum IFN-β levels in all strains at 5 h postinfection (n = 6 mice/strain) (Fig. 4B). Serum IFN-β levels from mock-treated mice (WT and each knockout strain) were uniformly below the limit of detection (LOD) for the ELISA. A significant level of IFN-β was detected in serum from infected WT mice. Serum levels of IFN-β from all infected knockout mice were found to be statistically indistinguishable from those of mock-treated serum samples.

To assess the impact of cGAS/STING cascade deficiencies on rAdV-induced secretion of proinflammatory chemokines and cytokines, a mouse proinflammatory multiplex strip serum assay was carried out (which detects TNF-α, IL-6, CXCL1/KC, IL-10, IL-1β, IFN-γ, and IL-12p70). Data are presented for both mock- and rAd5βGal-treated mice (n = 6 mice/strain) at 5 h p.i.(Fig. 4C). Infection of WT mice resulted in elevated serum levels of TNF-α, IL-6, KC, and IL-10 compared to those in mock-infected mice. Significant changes in IL-1β and IL-12p70 levels were not detected. rAd5βGal infection of knockout strains generated significantly lower levels of serum TNF-α, IL-6, and IL-10 than did rAd5βGal infection of WT mice. Levels of both KC and IFN-γ were trending down, but differences were not statistically significant. Furthermore, we found diminished basal levels of several cytokines in untreated IRF3^−/− and cGAS^−/− mice. Diminished basal level of cytokines contributed to the identification of strain-specific instances of cytokine induction following rAd5βGal infection of knockout mice. Overall, these data indicate that cytokine levels are reduced in knockout strains, but a residual low-level antiviral cytokine induction response persists.

Data indicate that antiviral signaling cascades are compromised in cGAS, STING, and IRF3 knockout mice. To determine how these deficiencies influence the transcriptional induction of antiviral response transcripts, total liver RNA was isolated from mock- or rAd5βGal-infected mice at 5 and 24 h postinfection and used in a two-step RT-qPCR assay with the indicated primer pairs (Fig. 5). Exposure of WT mice to rAd5βGal resulted in ∼2-log increases in IFN-β transcript and interferon-stimulated gene transcript levels corresponding to ISG15, ISG54, VIP, and CXCL9 at 5 h postinfection. Proinflammatory transcripts corresponding to TNF-α and IL-6 were significantly induced, whereas IL-1β was not. Levels of each transcript had declined by 24 h postinfection, with CXCL9 being the sole exception.

RT-qPCR analysis of total liver RNA from knockout strains at 5 h postinfection revealed significantly lower levels of the IFN-β transcript, which were similar to those in mock-treated mice (Fig. 5). For all strains, levels of IFN-β in mock-treated mice were extremely low and not significantly different from those for a no-RT control. Consistent with diminished levels of IFN-β and IFN-β transcripts (Fig. 4B and 5, respectively), rAdV-induced ISG and proinflammatory transcript levels from rAd5βGal-infected knockout strains were significantly lower than the levels observed for their WT counterparts, with the exception of CXCL9 (which was trending downward). Several additional observations were made from these assays. Basal levels of ISG15, ISG54, and VIP were diminished in IRF3 knockout samples but not in cGAS and STING knockout mice (in comparison to levels in mock-treated WT mice). It is apparent that in spite of the diminished levels of induced transcripts, there is persistent low-level induction of ISG15 and CXCL9 following Ad infection of knockout strains. Furthermore, ISG15 and CXCL9 were the only inducible transcripts whose levels were largely maintained above background levels at 24 h postinfection for all strains. At 24 h postinfection, levels of IFN-β, ISG54, VIG1, IL-6, and TNF-α were trending toward levels with mock treatment. Overall, these data strongly indicate that the cGAS/STING DNA-sensing cascade plays a major role in rAdV activation of innate antiviral responses in vivo.

Following systemic rAdV infection, the vast majority of virus is cleared through phagocytic uptake by macrophages and neutrophils. Hepatocyte uptake of rAdV and ensuing vector transgene expression serve as a marker for the persistence of virus-infected cells in immunocompetent or compromised mouse models. Based on our observations that cGAS/STING DNA sensing of rAdV makes a major contribution to the early innate antiviral response in vitro and in vivo, we asked if mice deficient in the cGAS/STING cascade demonstrate a compromised ability to clear virus-infected cells. Administration of 10¹¹ rAd5βGal vp/mouse for each strain (n = 6 to 9/strain/harvest time point, as described above) was followed by tissue harvest on day 1 or day 28 postinfection.

As described in Materials and Methods, liver segments were homogenized and clarified by centrifugation for use in a Tropix Galacto-Star chemiluminescent β-Gal reporter gene expression assay and compared to strain-specific mock-infected lysates (Fig. 6A). At 1 day postinfection, levels of β-Gal expression from WT and mutant strains (STING^−/−, IRF3^−/−, and cGAS^−/−) were equally elevated (∼3 logs) compared to those in mock-treated lysates. Under these experimental conditions, cGAS DNA-sensing cascade deficiencies did not result in significant differences in Ad-mediated liver transgene expression. In comparison to the levels of β-Gal expression at 1 day postinfection, liver homogenates from infected mice on day 28 postinfection revealed a significant loss of β-Gal expression for all mouse strains. We found no statistical difference between WT mice and each of the knockout strains. These data argue that clearance of virus-infected cells was not influenced by deficiencies of cGAS/STING DNA-sensing cascade components.

Liver DNA was harvested from mock-infected mice and infected mice at 1 day and 28 days postinfection for quantitative PCR of viral DNA. qPCR of persistent viral DNA provides a reporter gene-independent assay for quantifying rAd5βGal and vDNA persistence. Five nanograms of total DNA from each liver sample (n = 6 mice/strain/time point) was used in a standard qPCR assay (see Materials and Methods), using an Ad5 hexon-specific primer pair with normalization to the value for a cellular IFN-β genomic control using the ΔΔC_T method. An infected WT sample from day 28 was arbitrarily assigned a relative vDNA expression value of 1 as a unit standard for all samples. Mock-infected samples from each strain resulted in no detection of rAd5βGal, as expected (Fig. 6B). For each sample derived from a 24-h total DNA harvest, we found no statistically significant differences in rAd5 hexon PCR products between each mouse strain. These data are in agreement with the observations made in transgene expression assays. Levels of viral DNA and transgene expression were similar in WT and cGAS/STING-deficient mice at 24 h postinfection.

In agreement with data from the β-Gal expression assays, levels of vDNA isolated from livers at 28 days postinfection were equally reduced for all strains. In comparison to mock-treated samples, low-level residual vDNA was detected at 4 weeks postinfection in all infected strains. Based on data from the β-Gal reporter assay, residual vDNA present at these later time points was not yielding detectable reporter gene expression. Clearance of the rAd5 vector from the liver is partly attributed to the development of antigen-specific cytotoxic T cells (49, 51, 61). Based on the data presented, impairment of the cGAS/STING antiviral recognition response does not compromise virus clearance from the liver.

Adenovirus infections generate a potent neutralizing humoral immune response in the immunocompetent host. To assess the influence of cGAS/STING cascade mutants on the development of neutralizing antibody (NAb) against rAd5βGal, serum was harvested from mock- or rAd5βGal-infected mice at 28 days postinfection (n = 6/strain). A HeLa cell rAd5βGal transduction-blocking assay was used to assess NAb in serum samples from infected WT and knockout mice (as described in Materials and Methods) (Fig. 6C). Controls consisting of incubation of rAd5βGal with medium (no serum) or a 1:40 dilution of mock-treated WT serum resulted in similarly high levels of β-Gal expression. Sera from the infected WT strain and each cGAS/STING cascade-deficient strain at day 28 postinfection effectively blocked rAd5βGal transduction in a concentration-dependent manner. Although each strain developed NAbs against rAd5, higher luminescence levels in infected cGAS^−/− serum suggest that the NAb titer produced by infection of cGAS^−/− mice was slightly but significantly lower than the NAb titer generated from WT mice.