C57BL/6J (WT), STING ^-/- , and cGAS ^-/- mice were bred and maintained in-house at the University of Washington. Sting^fl/fl mice were a generous gift from Dr. Mohamed Oukka (University of Washington, Seattle, Washington). STING^fl/fl mice were made by a commercial service (Biocytogen, Wakefield MA) using a Cas9/sgRNA plasmid construct. In brief, a targeting vector was designed with a Neo cassette, flanked by Frt sites, and STING Exon 6 flanked by LoxP sites, was introduced into a B6-derived embryonic stem cells (ES). Targeted ES cells were introduced into host embryos, and cell embryos were surgically transferred into pseudo-pregnant (surrogate) mothers resulting in F0 heterozygous floxed mice on the B6 background. Chimeric mice were crossed to B6J (JAX #000664) for 6 generations. STING^fl/fl mice were bred to zbtb46 ^Cre mice (JAX #028538) to generate zbtb46^Cre x STING^f/fl/ mice (cDC STING cKO). The zbtb46^Cre x STING^f/fl genotyping was carried out using tail snips followed by PCR ( Supplementary Table 1 ). Zbtb46 is expressed by cDCs, but not plasmacytoid DCs (15), and as such cDC STING cKO mice should have STING selectively absent in cDCs. We verified that cDCs in cDC STING cKO mice were missing STING via qPCR for the detection of STING ( Supplementary Figure 1 ). Three to five STING^fl/fl, zbtb46^Cre and cDC STING cKO mice were sacrificed, spleens were collected, pooled, and made into single cell suspensions by crushing spleens through a 0.70μm filter followed by a 20ml wash with full RPMI media (RMPI 1640 supplemented with 10% FBS, non-essential amino acids, sodium pyruvate, pen/strep and β-mercaptoethanol). Red blood cells were lysed using an RBC lysis buffer (Thermofisher cat. no. 00-4300-54). Splenocytes were stained with Live/Dead fixable aqua dead (ThermoFisher Scientic cat. no. L34957), PerCP-conjugated anti-mouse CD11c (Biolegend cat. no. 117326), BV605-conjugated anti-mouse CD19 (Biolegend cat. no. 115540), BV605-conjugated anti-mouse CD3 (Biolegend cat. no. 100237), APC-conjugated anti-mouse Ly-6G (Biolegend cat. no. 127614), PE-conjugated anti-mouse NK1.1 (Biolegend cat. no. 108708), APC-Cy7-conjugated anti-mouse CD11c, and Pacific blue-conjugated anti-mouse MHC-II (Biolegend cat. no. 107620). A BD Biosciences FACS ARIA III cell sorter was used to isolate cDC (CD11c⁺MHCII^hi), and B and T cells (CD19⁺CD3⁺). RNA was isolated (Qiagen RNeasy kit cat. no. 74004) and then converted to cDNA (ThermoFisher cat. no. 4368814). STING expression in each cell type was measured via qPCR using primer F: GGGAGCCGAAGACTGTACAT; primer R: CGCTGTTGGAAAAACCCGA. All mice were maintained by the department of animal welfare (OAW) according to institutional Animal Care and Use Committee (IACUC)- approved protocols.

A DNA vaccine that encodes a codon-optimized full-length nucleoprotein from A/Puerto Rico/8/1934 (pNP) was used to immunize mice. Construction of the plasmid, pNP, employed for these studies is previously described (16). Briefly, the influenza NP gene is under control of the human cytomegalovirus (CMV) immediate early promoter. This plasmid also includes the following additional elements to optimize antigen expression: the hepatitis B virus (HBV) pre-S2 5’ untranslated region (UTR), rabbit beta globin poly A, rat insulin intron A, the HBV env enhancer and the CMV exon 1 and 2. Mice were injected with 10 μg of plasmid nucleoprotein (pNP) in 50μl PBS split between both tibialis anterior, and then were electroporated using a BTX agilepulse waveform electroporation system (cat. no. 47-0400N). BTX’s agilepulse voltage and pulse length settings used for intramuscular (IM) vaccinations were as follows: Group 1; 450 V pulse amplitude, 0.05ms pulse width, 300ms pulse interval, 500ms group intervals, 2 pulses, and Group 2; 110V pulse amplitude, 10ms pulse width, 300ms pulse interval, 500ms group interval, 8 pulses. Gene gun (GG) DNA vaccination was also used to vaccinate mice with pNP. Mice were vaccinated with 1μg pNP using helium at 400psi ( Supplementary Figure 2 ).

Sera were collected at 14-, 21- and 28-days post-vaccination to measure NP specific IgG, IgG₁, and IgG_2C antibody titers, and sera samples were stored at -80°C until the time of assay. Microtiter plates were coated at 1μg/ml of recombinant NP (Sino Biological inc, cat. no 11675-V08B) overnight at 4°C. Anti-NP IgG, IgG1 and IgG_2C concentrations (μg/ml) in sera were calculated using purified mouse IgG (Southern Biotech cat. no. 0107-01), IgG₁ (Southern Biotech cat. no. 0102-01), or IgG_2C (Southern Biotech cat. no. 1078-01) to produce a standard curve. Purified mouse antibodies were detected using 1μg/ml of goat anti-mouse IgG-HRP (Southern Biotech cat. no. 1030-05), IgG₁-HRP (Southern Biotech cat. no. 1070-05) or IgG_2C–HRP (Sothern Biotech cat. no. 1078-05), respectively. O-phenylenediamine dihydrochloride (OPD) substrate tablets (ThermoFisher cat. no. 34006) were used to observe HRP induced color changes. The HRP substrate reaction was terminated using 2N sulfuric acid after 2 minutes. Absorbance was measured at 450nm and concentrations of antibody were calculated using each plate’s internal standard curve. Sera dilutions were optimized for each experiment and timepoint to ensure absorbance reading fell within each plates internal standard curve. Data are representative of at least three independently preformed studies (n=3-9 mice/genotype/study).

Antigen (Ag)-specific CD8⁺ ( Supplementary Figure 3 ) and CD4⁺ T cells ( Supplementary Figure 5 ) were measured by flow cytometry using PE-and/or APC-conjugated-NP_366-374 (H-2k(b)) and PE-conjugated- NP_311-325 (I-A(b)) tetramers, respectively. Tetramers were obtained from the NIH tetramer core facility (Atlanta, GA USA). Mice were sacrificed, spleens collected and single cell suspensions were generated for each mouse as outlined above. A total of 1x10⁶ splenocytes were stained and analyzed via flow cytometry (BD bioscience LSRII). To measure the frequencies of Ag-specific CD4⁺ T cells, splenocytes were stained with the following antibodies: Live/Dead fixable aqua dead (ThermoFisher Scientic cat.no. L34957), BUV395-conjugated anti-mouse CD4 (BD cat. no. 565975), PerCP5.5-conjugated anti-mouse CD3 (Biolegend cat. no. 1002180), BV711-conjugated anti-mouse CD8 (Biolegend cat. no. 100748), PE-Cy7-conjugated anti-mouse CD44 (BD cat. no. 560569) and PE-conjugated- NP_311-325. To measure the frequencies of Ag-specific CD8⁺ cells, splenocytes were stained as follows: Live/Dead fixable aqua dead (ThermoFisher Scientic cat. no. L34957), PerCP5.5-conjugated anti-mouse CD3 (Biolegend cat. no. 1002180), BUV395-conjugated anti-mouse CD4 (BD cat. no. 565975), BV711-conjugated anti-mouse CD8 (Biolegend cat. no. 100748) and PE-Cy7-conjugated anti-mouse CD44 (BD cat. no. 560569).

Single cell suspensions from spleens were collected 21 days post-vaccination as outlined above then 1x10⁶ cells were plated in duplicate wells in a 96 well plate and rested for 24 hours (hrs.) in full RPMI media (details above) at 37°C. To evaluate CD8⁺ T cell polyfunctionality cells were either stimulated with 1μg/ml NP_366-374, or media only (unstimulated). After one hour, 1x brefeldin A (Biolegend cat. no. 420601) and FITC-conjugated CD107a/GranzymeB (Biolegned cat. no. 121606) were added to the wells and incubated overnight (~16hrs). The following day, cells were stained for IL-2 (Biolegend cat. no. 503808), IFN-γ (Biolegend cat. no. 505818), FITC-conjugated anti-human/mouse Granzyme B (Biolegend cat. no. 515403) and TNF-α (Biolegend cat.no 506308) using BD’s fixation/permeabilization solution kit (cat.no. 554714) and the manufacture’s protocol. T cells were first gated using BV605-conjugated anti-mouse CD3 (Biolegend cat. no. 100237). CD8⁺ T cells were gated using BV711-conjugated anti-mouse CD8 (Biolegend cat. no. 100748). The frequencies of CD8⁺ T cells expressing IL-2, TNF-α, IFN-γ and/or cytolytic markers CD107a/Granzyme B were then measured by flow cytometry (BD Biosciences LSRII) and Boolean gating ( Supplementary Figure 4 ). The frequencies of mouse CD8⁺ T cell in media were subtracted from their frequencies following NP peptide stimulation. These data were then used to calculate a polyfunctional index as described (17). CD8⁺ T cell polyfunctionality is defined as the frequency of CD8⁺ T cells expressing any three or more of the cytokines IFN-γ, IL-2, TNF-α and/or co-expressing the cytolytic marker CD107a/GranzymeB after stimulation with the immunodominant NP peptide, NP_366-374, specific for C57BL/6 mice.

To measure Th1 and Th2 cytokine expression, 1x10⁶ cells were plated in duplicate, rested for 24hrs in full media at 37°C and then one well was stimulated overnight (~16hrs) with 1μg/ml NP_311-325. Supernatants were collected and analyzed for IFN-γ, TNF-α, IL-2, IL-4, IL-6, IL-10 and IL-17a via flow cytometry (BD Bioscience LSRII) using BD’s mouse Th1/Th2/Th17 Cytometric Bead Array kit (cat. no. 560485) using the manufacture’s protocol.

To measure IFN-γ expressing CD4⁺ T cells mouse splenocytes were isolated as outlined above. A total of 1x10⁶ cells were pated in duplicate in a 96 well plate and rested for 24hrs. The cells were then stimulated with 1μg/ml NP_311-325, the immunodominant NP peptide for C57BL/6 CD4⁺ T cells or remained in media only. After one hour, 1x brefeldin A (Biolegend cat. no. 420601) was added to wells and the cells were incubated overnight (~16hrs). The following day splenocytes were stained with PerCP5.5-conjugated anti-mouse CD3 (Biolegend cat. no. 1002180), BUV395-conjugated anti-mouse CD4 (BD cat. no. 565975) and IFN-γ (Biolegend cat. no. 505818) using BD’s fixation/permeabilization solution kit (cat. no. 557414) and the manufacture’s protocol. The frequencies of CD4⁺ T cells expressing IFN-γ was measured by flow cytometry (DB Bioscience LRSII).

Mouse splenocytes were used to enumerate NP-specific antibody secreting cells 21 days post-vaccination. ELISPOT plates (Millipore Sigma cat. no. S2EM004M99) were coated with 10μg/ml recombinant NP (Sino Biological inc, cat. no 11675-V08B) in PBS and then incubated overnight at 4°C. Mouse splenocytes were processed as outlined above and then 3x10⁶ splenocytes in 100μl of media were plated in NP-coated plates in duplicate and then incubated overnight (~16hrs) at 37°C. The following day, the cells were discarded and the wells were washed 5 times with PBS. Following the final wash, 100μl of anti-mouse IgG-HRP, IgG₁-HRP, or IgG_2C–HRP diluted at 1:2000 in PBS were added to each well. Spots were developed using BD ELISPOT AEC substrate set (BD cat. no. 551951) and then visualized using a CTL imager (Cellular Technologies) and were enumerated using an ImageJ cell counter.

Sera were collected from mice prior to vaccination (baseline), and at 6hrs and 24hrs post-vaccination. Sera from WT, zbtb46^Cre x STING^fl/fl (cDC STING cKO) and cDC STING LitC mice were diluted 1:5 while sera from STING^-/- mice were diluted 1:2. Levels of IFN-γ, TNF, IL-2, IL-4, IL-6, IL-10 and IL-17a in diluted sera were then measured by flow cytometry (BD Bioscience LSRII) using BS’s mouse Th1/Th2/Th17 Cytometric Bead Array kit (cat. no. 560485) following the manufacture’s protocol. IFN beta (IFN-β) was measured in sera diluted 1:2 prior to vaccination and 6hrs post-vaccination using mouse IFN-β ELISA kit (PBL Assays cat. no. 42410-1).

All data are represented as the mean of individual mice ± SD. Statistical analyses were performed using one-way ANOVA followed by a turkey post-test and Student’s t test using Graphpad Prism 7.

To determine if cGAS and/or STING play a role in the ability of DNA vaccines to induce antibody (Ab) responses, WT, cGAS^-/- and STING^-/- mice were vaccinated by intramuscular delivery and electroporation (IM/EP) with 10μg of a codon optimized DNA vaccine expressing influenza A nucleoprotein (pNP). Sera were collected 14-, 21- and 28-days post-vaccination to analyze induction of NP-specific IgG Ab responses. STING^-/- mice generated significantly lower anti-NP IgG Ab concentrations at 21- and 28-days post-vaccination compared to WT and cGAS^-/- mice ( Figure 1A ), a result that confirms previous studies that STING, but not cGAS, plays a role in DNA vaccine immunogenicity (9). The lower IgG titers in STING^-/- mice led us to investigate if both Th1 (IgG_2C) and Th2 (IgG₁) Ab responses were impacted by STING or cGAS. Analysis of IgG_2C and IgG₁ titers, and IgG_2C/IgG₁ ratios in vaccinated WT, cGAS^-/- and STING^-/- mice showed that STING^-/- mice developed significantly lower Th1 associated anti-NP IgG_2C ( Figure 1B ) than WT and cGAS^-/- mice, but comparable levels of Th2 associated NP-specific IgG1 Ab titers ( Figure 1C ). A similar outcome was observed in cGAS^-/- , STING^-/- and WT mice immunized with the same DNA vaccine by gene gun indicating that these results are not dependent on the route of DNA vaccine delivery ( Supplementary Figure 2 ). Taken together, these data are consistent with previous findings showing that STING is required to induce Ab responses, but further shows that STING is necessary to generate Th1 IgG_2C antibody, but not Th2 IgG₁ responses.

A hallmark of DNA vaccines is their ability to induce Th1 responses that mediate the induction of CD8⁺ T cell responses including cytotoxic T lymphocytes (2–5, 18). To determine if STING or cGAS play a role in DNA vaccine induction of CD8⁺ T cell responses, WT, cGAS^-/- and STING^-/- mice were vaccinated with pNP by IM/EP vaccine delivery. Mice were sacrificed 21 days post-vaccination and the frequencies of antigen specific CD8⁺ T cells were measured using tetramers for NP_366-374, an immunodominant NP-specific CD8⁺ T cell epitope via flow cytometry ( Supplementary Figures 3A, B ). The frequencies of NP_366-374 tetramer binding CD8⁺ T cells were not statistically different between WT, cGAS^-/- and STING^-/- mice ( Figure 2A ) indicating that cGAS and STING are not required for DNA vaccine induction of antigen-specific CD8⁺ T cells. To determine if STING impacts effector functions of CD8⁺ T cells, we next measured the frequencies of polyfunctional CD8⁺ T cells expressing one or more cytokines IL-2, TNF-α, IFN-γ and the cytolytic markers CD107a/Granzyme B following stimulation with NP_366-374 by intracellular cytokine staining (ICS) then flow cytometry ( Supplementary Figure 4 ). To compare CD8⁺ T cell effector functions, a polyfunctionality index score was calculated for each mouse. The polyfunctional index enumerates cellular polyfunctionality as a one-dimensional value where greater value is given to CD8⁺ T cells that express more functions (17). CD8⁺ T cells in STING^-/- mice exhibited significantly lower polyfunctional index scores when compared to WT and cGAS^-/- mice ( Figures 2A, B ). These data suggest that STING and cGAS are not required for the generation of antigen-specific CD8⁺ T cells, but STING is required for the induction of polyfunctional CD8⁺ T cell responses.

Conventional dendritic cells (cDCs) are important for the generation of DNA vaccine induced Ab and CD8⁺ T cell responses (10–14), therefore, we hypothesized that STING within cDCs may be required to mediate DNA vaccine induction of Th1 and CD8⁺ T cell responses. To test this, we generated mice that lacked STING in cDCs (cDC STING cKO), but not in plasmacytoid DCs (19). WT, STING^-/- , cDC STING cKO and cDC STING littermate control (cDC STING LitC) mice (to control for the potential deleterious effects on STING due to the introduction of flox sites) were vaccinated by IM/EP delivery of 10μg of pNP. Mice were sacrificed 21 days post-vaccination and splenocytes were isolated to measure Ag-specific CD8⁺ T cell responses by NP_366-374 tetramer staining ( Supplementary Figures 3A, C ) and polyfunctional CD8⁺ T cell responses by ICS and flow cytometry, as described above. WT, STING^-/- , cDC STING cKO and cDC STING LitC mice had similar frequencies of tetramer binding CD8⁺ T cells ( Figure 3A ). In addition, WT, cDC STING cKO and cDC STING LitC mice developed similar levels of polyfunctional CD8⁺ T cell responses ( Figures 3B, C ). Together, these results indicate that while STING is required for induction of polyfunctional CD8⁺ T cell responses ( Figure 2 ), it is not required within cDCs ( Figure 3 ).

The production of pro-inflammatory cytokines influences the generation of Th1 or Th2 responses and CD4⁺ and CD8⁺ T cell effector functions (20–23). To determine if STING within cDCs influenced the induction of pro-inflammatory cytokines, WT, STING^-/- , cDC STING cKO and cDC STING LitC mice were vaccinated with pNP via IM/EP delivery. Serum concentrations of pro-inflammatory cytokines were measured prior to vaccination, at 6 hours and at 24 hours post-vaccination. A control group of mice were mock treated by IM injection of saline followed by IM/EP to control for the induction of pro-inflammatory cytokines induced by the vaccination method. Consistent with previous studies (9), we found that sera from STING^-/- mice has significantly lower levels of TNF-α ( Figure 4A ), IL-6 ( Figure 4B ) and IFN-β ( Figure 4D ) when compared to sera from WT, cDC STING cKO and cDC STING LitC mice. STING^-/- mice also had lower levels of IL-2 ( Figure 4D ) compared to sera from WT, cDC STING cKO and cDC STING LitC mice, but these differences did not reach statistical significance. These data suggest that STING mediates the induction of pro-inflammatory cytokines after DNA vaccination, but not within cDCs.

Since our data indicated that STING is required to mediate optimal IgG response and specifically, Th1-associated IgG_2C Abs, we next investigated if STING within cDCs was necessary to generate these responses. WT, STING^-/- , cDC STING cKO and cDC STING LitC mice were vaccinated with pNP via IM/EP delivery. IgG, IgG₁ and IgG_2C Ab responses were measured by ELISA prior to vaccination (D0), and at 14-, 21- and 28- days post-vaccination. STING^-/- and cDC STING cKO mice developed significantly lower NP-specific IgG ( Figure 5A ) and IgG_2C ( Figure 5B ) 21- and 28-days post-vaccination, but comparable IgG₁ ( Figure 5C ) responses when compared to WT and cDC STING LitC mice. Consequently, STING^-/- and cDC STING cKO mice also had lower Th1/Th2 ratios (IgG_2C/IgG₁) than the control mice ( Figure 5D ) indicating a lower Th1 response. Together, these results show that STING within cDCs is required to induce Th1 (IgG_2C) type Ab responses. To determine if the lower IgG_2C Ab responses observed in STING^-/- and cDC STING cKO mice was due to a reduction in the number of Ab-secreting B cells (ASCs), the number of NP-specific IgG ( Figure 5E ), IgG_2C ( Figure 5F ) and IgG₁ ( Figure 5G ) ASCs were quantified 21 days post-vaccination via a B cell ELIspot assay. STING^-/- and cDC STING cKO mice developed comparable number of IgG₁ ASCs, but significantly fewer IgG and IgG_2C ASCs than the control mice, further supporting our findings that STING is required within cDCs to induce IgG_2C secreting B cells following DNA vaccination.

To determine the role of STING within cDCs for DNA vaccine induction of CD4⁺ T cell responses WT, STING^-/- , cDC STING cKO and cDC STING LitC mice were IM/EP-vaccinated with pNP and the frequencies of NP-specific CD4⁺ T cells binding the NP-specific CD4⁺ T cell epitope, NP_311-325, were measured by tetramer staining and flow cytometry ( Supplementary Figure 5 ) 21 days post-vaccination. Both STING^-/- and cDC STING cKO mice exhibited a trend toward lower frequencies of Ag-specific CD4⁺ T cells, but these differences were not significantly different from control groups ( Figure 6A ). To determine the impact of STING within cDCs on the induction of Th1 CD4⁺ T cell responses, we evaluated cytokines produced by splenocytes stimulated with a C57Bl/6 CD4⁺ T cell NP immunodominant peptide NP_311-325. Splenocytes were isolated 21 days post-vaccination, stimulated overnight with 1μg/ml NP_311-325 and then concentrations of Th1 (IFN-γ, TNF-α, IL-2 and IL-6) and Th2 (IL-4) cytokines were analyzed in the supernatants using a flow-based cytokine detection kit (24). Overall, we observed no significant differences in the production of TNF-α, IL-2, IL-4 and IL-6 ( Supplementary Figure 6 ). We also measured IL-10 [a marker of Tregs (25)] and IL-17a cytokines but observed no differences between the groups suggesting that STING may have limited impact on the induction of Tregs or Th17 T cells following DNA vaccination ( Supplementary Figures 6D, F ) but more studies are required to determine if STING impacts these and other T cell subsets. However, STING^-/- and cDC STING cKO splenocytes produced significantly less IFN-γ when compared to WT and cDC STING LitC mice ( Figure 6B ), suggesting a role for STING within cDCs in the induction of Th1/IFN-γ-producing CD4⁺ T cell responses. To further investigate this, CD4⁺ T cells from STING^-/- and cDC STING cKO mice were stimulated in vitro with the NP CD4⁺ T cell epitope, NP_311-325, stained for IFN-γ by ICS and then analyzed by flow cytometry ( Figure 6C ). STING^-/- and cDC STING cKO mice developed significantly lower frequencies of IFN-γ⁺CD4⁺ T cells when compared to WT and cDC STING LitC mice ( Figure 6D ), further indicating that STING is required within cDCs for DNA vaccine induction of Th1 CD4⁺ T cell responses.