The common Sting1 HAQ, AQ alleles rescue CD4 T cellpenia, restore T-regs, and prevent SAVI (N153S) inflammatory disease in mice

The significance of STING1 gene in tissue inflammation and cancer immunotherapy has been increasingly recognized. Intriguingly, common human STING1 alleles R71H-G230A-R293Q (HAQ) and G230A-R293Q (AQ) are carried by ~60% of East Asians and ~40% of Africans, respectively. Here, we examine the modulatory effects of HAQ, AQ alleles on STING-associated vasculopathy with onset in infancy (SAVI), an autosomal dominant, fatal inflammatory disease caused by gain-of-function human STING1 mutations. CD4 T cellpenia is evident in SAVI patients and mouse models. Using Sting1 knock-in mice expressing common human STING1 alleles HAQ, AQ, and Q293, we found that HAQ, AQ, and Q293 splenocytes resist STING1-mediated cell death ex vivo, establishing a critical role of STING1 residue 293 in cell death. The HAQ/SAVI(N153S) and AQ/SAVI(N153S) mice did not have CD4 T cellpenia. The HAQ/SAVI(N153S), AQ/SAVI(N153S) mice have more (~10-fold, ~20-fold, respectively) T-regs than WT/SAVI(N153S) mice. Remarkably, while they have comparable TBK1, IRF3, and NFκB activation as the WT/SAVI, the AQ/SAVI mice have no tissue inflammation, regular body weight, and normal lifespan. We propose that STING1 activation promotes tissue inflammation by depleting T-regs cells in vivo. Billions of modern humans have the dominant HAQ, AQ alleles. STING1 research and STING1-targeting immunotherapy should consider STING1 heterogeneity in humans.

In this study, we discovered, surprisingly, that the HAQ, AQ splenocytes are resistant to STING1mediated cell death.We generated HAQ/SAVI(N153S) and AQ/SAVI(N153S) mice and found that the HAQ, AQ alleles prevent CD4 T cellpenia, increasing/restoring T-regs and alleviating/stopping tissue inflammation in SAVI mice, thus providing evidence for the in vivo significance of type I IFNsindependent, STING1-mediated cell death and potential AQ-based curative therapy for SAVI patients.

STING1 activation kills mouse spleen CD4, CD8 T, and CD19 B cells ex vivo
We first used the synthetic non-CDNs STING1 agonist diABZI (Ramanjulu et al., 2018) to induce lymphocyte death because diABZI induces cell death without the need for lipid transfection or detergent for cell permeabilization (Kabelitz et al., 2022;Messaoud-Nacer et al., 2022) and diABZI is in clinical trials (NCT05514717).Splenocytes from C57BL/6 N mice were treated with diABZI in culture, and cell death was determined by Annexin V and Propidium Iodide stain.Splenocyte cell death could be detected at 5 hr post diABZI treatment (Figure 1-figure supplement 1A).Dosage responses showed that ~25 ng/ml diABZI could kill 70% of splenocytes (Figure 1-figure supplement 1B).Similarly, STING1 agonists DMXAA and synthetic CDNs RpRpss-Cyclic di-AMP killed mouse spleen CD4, CD8 T cells, and CD19 B cells (Figure 1A).Thus, STING1 activation readily induces mouse lymphocyte death ex vivo.
HAQ, AQ, Q293 STING1 knock-in mouse splenocytes are resistant to STING1-mediated cell death ex vivo HAQ and AQ are common human STING1 alleles (Jin et al., 2011;Patel et al., 2017a;Mansouri et al., 2022).Previously, we reported that HAQ knock-in mice are defective in CDNs-induced immune responses, while CDNs responses in AQ knock-in mice are similar to WT mice (Mansouri et al., 2022).We treated splenocytes from HAQ and AQ mice with diABZI ex vivo and found, surprisingly, that both HAQ and AQ splenocytes were resistant to diABZI-induced cell death (Figure 1C and D).In comparison, IFNAR1 -/-splenocytes were killed by diABZI, confirming that STING1-mediated lymphocytes death are type I IFNs-independent (Figure 1C and D;Kuhl et al., 2023;Gulen et al., 2017;Murthy et al., 2020).
HAQ and AQ share the common A230 and Q293 residues changes.We thus generated a Q293 Sting1 knock-in mouse.Notably, the Q293 splenocytes were resistant to STING1 agonists 2'3'-cGAMP, RpRpss-Cyclic di-AMP, and diABZI-induced cell death (Figure 1E and F).Thus, the residue 293 of STING1 is critical for its cell death function.WT/HAQ, WT/AQ mouse splenocytes are partially resistant to STING1mediated cell death ex vivo WT/HAQ (34.3%) is the most common human STING1 genotype in East Asians, while WT/AQ (28.2%) is the 2 nd most common STING1 genotype in Africans (Patel et al., 2017a).We generated WT/ HAQ, WT/AQ mice and treated their splenocytes with mouse STING1 agonist DMXAA.WT/HAQ and WT/AQ splenocytes were protected from 25 µg/ml DMXAA-induced cell death (Figure 1G).A total of 100 µg/ml DMXAA could kill WT/HAQ and WT/AQ splenocytes, albeit less than WT/WT cells (Figure 1G).Thus, the HAQ and AQ alleles are dominant and likely impact STING1 activation even in heterozygosity.
STING1 activation kills primary human CD4 T cells but not CD8 T or CD19 B cells STING1 agonists-based clinical trials in humans have been disappointing (NCT02675439, NCT03010176, NCT05514717; Meric-Bernstam et al., 2023;Meric-Bernstam et al., 2022).We showed that the human STING1 gene might undergo natural selection during the out-of-Africa migration (Mansouri et al., 2022) sensitive to evolutionary pressure.Thus, we investigated STING1mediated death in primary human lymphocytes.
Human explant lung cells from the WT(R232)/WT(R232) donors were treated with STING1 agonists 2'3 c GAMP, RpRpss-Cyclic di-AMP, diABZI for 24 hr in culture.Lymphocyte cell death was determined by Propidium Iodide staining.Different from mouse lymphocytes, diABZI and RpRpss-Cyclic di-AMP killed human CD4 T but not CD8 T or CD19 B cells (Figure 2A and B).Human CD8 T and CD19 B cells are resistant to 500 ng/ml diABZI-induced cell death (Figure 2-figure supplement 1A).

WT/HAQ human CD4 T cells are resistant to low doses of diABZIinduced cell death
WT/HAQ mouse splenocytes are resistant to low-dose diABZI-induced cell death (Figure 1G).To extend our observation into primary human T cells, we obtained lung explants from WT/WT and WT/ HAQ individuals (Figure 2-figure supplement 1B) and treated them with diABZI in culture.25 ng/ ml diABZI killed WT/WT, but not WT/HAQ, human lung CD4 T cells (Figure 2C).

HAQ and AQ alleles rescue the lymphopenia and suppress myeloid cell expansion in SAVI(N153S) mice
The in vivo significance of the STING1/MPYS-cell death is unclear.Furthermore, multiple cell death pathways, that is apoptosis, necroptosis, pyroptosis, ferroptosis, and PANoptosis, are proposed (Murthy et al., 2020;Li et al., 2021;Song et al., 2022).The uncertainty likely results from studies using different cell types (primary cells vs cancer cell lines); species (human vs mouse); STING1 agonists (cGAMP, which requires cell permeabilization by detergents or lipid transfection, vs diABZi, DMXAA that can directly cross the membrane; Larkin et al., 2017;Cerboni et al., 2017;Gulen et al., 2017;Kabelitz et al., 2022;Wu et al., 2019).Critically, which mechanism is relevant in vivo, causing T  SAVI is an autosomal dominant, inflammatory disease caused by one copy of a gain-of-function STING1 mutant (WT/SAVI) Liu et al., 2014).CD4 T cellpenia was found in SAVI patients and SAVI mouse models (Liu et al., 2014;Luksch et al., 2019.STING1 activation in SAVI mice is independent of ligands and happens in vivo.We thus generated HAQ/SAVI(N153S) and AQ/SAVI(N153S) mice aiming to establish the in vivo significance and mechanism of STING1-cell death.

Discussion
This study, using the HAQ, AQ, SAVI(N153S) Sting1 knock-in mice, reveals the in vivo significance and mechanism of STING1-mediated CD4 T cell death.HAQ, AQ alleles prevent CD4 T cellpenia, and increase/restore CD4 T-regs in SAVI mice.The results are consistent with previous finding that the impaired CD4 T cell proliferation by the SAVI(V155M) mutant could be rescued by the addition of the HAQ allele in vitro (Cerboni et al., 2017).STING1 has been increasingly implicated in inflammatory diseases such as nonalcoholic fatty liver disease, nonalcoholic steatohepatitis, cardiomyopathy, obesity, diabetes, neurodegenerative diseases, aging, and kidney injury, many of which are independent of type I IFNs (Skopelja-Gardner et al., 2022;Bai and Liu, 2021;Gao et al., 2023a).It is tempting to suggest that STING1 activation in CD4 T cells leads to CD4 T-regs depletion that break tissue tolerance and exacerbates tissue inflammation.
Human immunodeficiency virus (HIV) primarily infects CD4 T cells and might activate the STING1 pathway in CD4 T cells (Monroe et al., 2014;Doitsh et al., 2014;Jakobsen et al., 2015;Silvin and Manel, 2015;Altfeld and Gale, 2015;Krapp et al., 2018).The loss of CD4 T cells is the hallmark of untreated HIV infection Gubser et al., 2019;Morou et al., 2019, and the measurement of CD4 T cell count is a central part of HIV care.We found that HAQ and AQ CD4 T cells are resistant to STING1mediated cell death.Mogensen and colleagues reported that HAQ/HAQ was enriched in HIV-infected long-term nonprogressors (Nissen et al., 2018).These HAQ/HAQ individuals had reduced inhibition of CD4 T cell proliferation and a reduced immune response to DNA and HIV (Nissen et al., 2018).It is likely that HIV infection activates STING1-cell death pathway in CD4 T cells.In HAQ/SAVI and AQ/ SAVI mice, one copy of HAQ, AQ allele suppressed CD4 T cell death.HAQ, AQ carriers might have fewer HIV-induced CD4 T cell death, thus being long-term nonprogressors in HIV infection-induced Acquired immunodeficiency syndrome (AIDS; Nissen et al., 2018).Targeting STING1 to prevent CD4 T cell death might be a valid therapy for AIDS.Activating the STING1 pathway is a promising strategy for cancer immunotherapy (Hines et al., 2023;Samson and Ablasser, 2022;Liu et al., 2021;Zheng et al., 2020;Corrales et al., 2015;Fu et al., 2015;Barber, 2011).Multiple STING1 agonists are in clinical trials (Meric-Bernstam et al., 2023;Meric-Bernstam et al., 2022).Recently, the safety issue emerged in some STING1 agonist trials (Meric-Bernstam et al., 2023;Meric-Bernstam et al., 2022).For example, in the STING1 agonist, ADU-S100, clinical trial, Grade 3/4 treatment-related adverse events were reported in 12.2% of 41 pretreated patients (NCT02675439) (Meric-Bernstam et al., 2023;Meric-Bernstam et al., 2022).The National Institutes of Health defines grade 3 as 'incapacitating; unable to perform usual activities; requires absenteeism or bed rest.'In a clinical trial using STING1 antibody-drug-conjugate (ADC) that conjugates diABZI to anti-HER2 Ab, a Grade 5 (fatal) serious adverse event was recorded and deemed related to the STING1-ADC (NCT05514717).SAVI disease, driven by overreacting STING1, is often fatal Liu et al., 2014.AQ, to a less degree, HAQ, suppress mortality in SAVI mice.Future STING1 clinical trials should be based on human STING1 genotype to achieve safe and effective responses.
Mechanistically, apoptosis, pyroptosis, ferroptosis, necroptosis, and PANoptosis have all been reported in STING1-mediated cell death (Kuhl et al., 2023;Jin et al., 2008;Gulen et al., 2017;Kabelitz et al., 2022;Murthy et al., 2020;Li et al., 2021;Song et al., 2022;Messaoud-Nacer et al., 2022;Tang et al., 2016).Different cell types and STING1 agonists used likely contributed to the inconsistency and complexity.Here, we focused on lymphopenia in the SAVI mice that avoids ligand-dependent, non-physiological dosage in STING1-mediated cell death.HAQ and with 100 ng/ml diABZi in culture for 24 hr.Cells were lysed and run on a 4~20% Mini-PROTEAN TGX Stain-Free Precast Gel.The blot was probed for STING1 antibody (Proteintech, 19851-1-AP).(B, D).IFNβ and TNF were determined by ELISA in the cell supernatant from E. (F).BMDM from WT/ WT, WT/SAVI, HAQ/SAVI and AQ/SAVI were treated with 400 µM cleavable chemical crosslinker DSP (Pierce, cat no: PG82081) in PBS for 1 hr at 4 °C.Cells were washed with PBS and lysed in RIPA buffer.Whole cell lysate was mixed with 4 x Laemmli Sample Buffer (BioRad, cat no 1610747) containing 5% 2-mercaptoethanol, heated at 95 °C for 10 min and, run on a 4~20% Mini-PROTEAN TGX Stain-Free Precast Gel.The blot was probed for STING1 antibody (Proteintech, 19851-  AQ alleles could prevent CD4 T cellpenia in the SAVI mice strongly indicating that residue A230 or Q293 prevent STING1-mediated CD4 T cell death in vivo.Splenocyte from Q293 mice were resistant to STING1 agonists-induced cell death ex vivo.Thus, it is likely that the Q293 residue is critical for STING1-mediate lymphopenia.Notably, Q293 is outside the C-terminal tail (CTT) (residues 341-379 of human STING1) critical for TBK1 recruitment and IRF3 phosphorylation (Liu et al., 2015) or miniCTT domain (aa343-354) (Cerboni et al., 2017), or the UPR motif (aa322-343) (Wu et al., 2019) important for T cells death in vitro.Further studies are needed to understand how the aa293 of STING1 mediates cell death in vivo.Noteworthy, AQ/SAVI cells had similar TBK1-IRF3, NFκB activation and STING1 degradation as the WT/SAVI cells.Yet, AQ/SAVI mice did not have CD4 T cellpenia as WT/SAVI mice suggeSting1 that the canonical STING1-TBK1-IRF3/ NFκB pathway, likely STING1 oligomerization, is not sufficient for the induction of cell death at the physiological condition.
We used the WT/N153S knock-in SAVI mouse model that spontaneously develop lung inflammation, T cell cytopenia, and early mortality, mimicking pathological findings in human SAVI patients (Warner et al., 2017).Using the WT/N153S SAVI mouse model and human Jurkat T cell line, it was proposed that STING1 activation causes chronic ER stress and unfolded protein response, leading to T cell death by apoptosis (Wu et al., 2019).Furthermore, the study showed that crossing WT/N153S mice to the OT-I mice reduced ER stress and restored CD8 + , but not CD4 + , T cells (Wu et al., 2019).The restoration of CD8 + T cells reduces inflammation and lung disease (Wu et al., 2019).However, human WT/N154S SAVI patients have normal CD8 + T cells numbers (Liu et al., 2014), and primary human CD8 + T cells are largely resistant to STING1-agonists-induced cell death ex vivo (Figure 2A; Kuhl et al., 2023).Thus, it is puzzling how restoring CD8 + T cells can rescue SAVI phenotypes since the SAVI patients already have normal CD8 + T cells numbers.
Finally, it is unexpected that both HAQ and AQ alleles are resistant to cell death.Our previous studies showed that the HAQ and AQ alleles have opposite functions (Mansouri et al., 2022).AQ-STING1, not HAQ-STING1, responds to CDNs (Jin et al., 2011;Patel et al., 2017a;Mansouri et al., 2022;Sebastian et al., 2020;Yi et al., 2013;Patel et al., 2017b;Nissen et al., 2018;Ruiz-Moreno et al., 2018;Movert et al., 2023).AQ mice are lean while HAQ mice are fat (Mansouri et al., 2022).Most importantly, HAQ was positively selected, while AQ was negatively selected, in modern humans outside Africans (Mansouri et al., 2022).Thus, the death pathway of STING1 is also distinct from the STING1 function that was naturally selected.It is worth noting that the AQ allele does better than the HAQ allele in suppressing SAVI disease.Thus, besides preventing cell death, additional mechanism by the AQ allele, for example fatty acid metabolism (Mansouri et al., 2022;Vila et al., 2022), is involved in curing SAVI.

The limitations of the study
The poor transferability of mouse to humans is a major issue in STING1 research (Meric-Bernstam et al., 2023;Meric-Bernstam et al., 2022).The present study used AQ/SAVI and HAQ/SAVI mice.Confirmation is needed in humans with the identification and evaluation of people who are AQ/SAVI, HAQ/SAVI.

Experimental design
The study was designed to reveal (i) the in vivo significance of the type I IFNs-independent, STING1dependent cell death function; (ii) the interplay between common STING1 alleles HAQ, AQ and the rare, gain-of-function SAVI STING1 mutation; (iii) the driver for the inflammatory SAVI disease.Mouse splenocytes, primary human lung cells, human THP-1 cells and HAQ, AQ, SAVI knock-in mice were used to establish the in vivo significance and human relevance.All the repeats were biological replications that involve the same experimental procedures on different mice.Where possible, treatments were assigned blindly to the experimenter by another individual in the lab.When comparing samples from different groups, samples from each group were analyzed in concert, thereby preventing any biases that might arise from analyzing individual treatments on different days.All experiments were repeated at least twice.

Mice
WT/SAVI(N153S) mice were purchased from The Jackson Laboratory.HAQ, AQ mice were previously generated in the lab (Patel et al., 2017a;Mansouri et al., 2022).The Q293 mice were generated by Cyagen Biosciences.Briefly, the linearized targeting vector was transfected into JM8A3.N1 C57BL/6 N embryonic stem cells.A positive embryonic stem clone was subjected to the generation of chimera mice by injection using C57BL/6 J blastocysts as the host.Successful germline transmission was confirmed by PCR sequencing.The heterozygous mice were bred to Actin-flpase mice [The Jackson Laboratory, B6.Cg-Tg (ACTFLPe)9205Dym/J] to remove the neo gene and make the Q293 knock-in mouse.Age-and gender-matched mice (2-6 month old, both male and female) were used for indicated experiments.WT/SAVI (male) x WT/HAQ (female), WT/SAVI (male) x WT/AQ (female) breeders were set up to generate HAQ/SAVI, AQ/SAVI mice.Mice were housed at 22 °C under a 12hr light-dark cycle with ad libitum access to water and a chow diet (3.1 kcal/g, Teklad 2018, Envigo, Sommerset, NJ) and bred under pathogen-free conditions in the Animal Research Facility at the University of Florida.Littermates of the same sex were randomly assigned to experimental groups.All mouse experiments were performed by the regulations and approval of the Institutional Animal Care and Use Committee at the University of Florida, IACUC202200000058.

Generation of THP-1 KO-STING1 cells stably expressing human STING1 alleles
THP1-Dual KO-STING1 cells were purchased from Invivogen (thpd-kostg).These cells are guaranteed mycoplasma-free and authenticated by the vendor.THP1-Dual KO-STING1 Cells in six-well plate were transfected with 1 µg STING1 plasmid (in pcDNA 3.1 vector) with Lipofectamine LTX and Plus Reagent (Invitrogen, cat no: A12621) according to the manufacturer's instructions.Transfecting Plasmid DNA into THP-1 Cells Using Lipofectamine LTX Reagent | Thermo Fisher Scientific -US.Forty-eight hours after the transfection, the cell medium was changed.G418 (1 mg/ml) was added to the culture to select STING1 expressing THP-1 cells.The G418-resistant cells were established and expanded.

Histology
Lungs and livers were fixed in 10% formalin, paraffin-embedded, and cut into 4 µm sections.Lung, liver sections were then stained for hematoxylin-eosin.All staining procedures were performed by the histology core at the University of Florida.Briefly, tissue sectins were immersed Harris Hematoxylin for 10 s, then washed with tap water.Cleard sections were re-immersed in EOSIN stain for ~30 s.The sections were washed with tap water until clear, then dehydrate in ascending alcohol solutions (50%, 70%, 80%, 95% x 2, 100% x 2).Afterwards, the sections werer cleared with xylene (3-4 x).The sections were mounted on glass slide with permount organic mounting medium for visulization.

Lung function
Pulmonary function was evaluated using an isolated, buffer-perfused mouse lung apparatus (Hugo Sachs Elektronik, March-Huggstetten, Germany), as previously described (Cai et al., 2020).Briefly, mice were anesthetized with ketamine and xylazine and a tracheostomy was performed, and animals were ventilated with room air at 100 breaths/min at a tidal volume of 7 μl/g body weight with a positive end-expiratory pressure of 2 cm H 2 O using a pressure-controlled ventilator (Hugo Sachs Elektronik, March-Huggstetten, Germany).

Isolation of lung cells
Cells were isolated from the lung as previously described (Mansouri et al., 2020).The lungs were perfused with ice-cold PBS and removed.Lungs were digested in DMEM containing 200 μg/ml DNase I (Roche, 10104159001), 25 μg/ml Liberase TM (Roche, 05401119001) at 37 °C for 2 hr.Red blood cells were then lysed and a single-cell suspension was prepared by filtering through a 70 µm cell strainer.

Flow cytometry
Single-cell suspensions were stained with fluorescent-dye-conjugated antibodies in PBS containing 2% FBS and 1 mM EDTA.Surface stains were performed at 4 °C for 20 min.For intracellular cytokine or transcription factor staining of murine and human cells, cells were fixed and permeabilized with the Foxp3 staining buffer set (eBioscience, cat no 00-5523-00).Cells were washed and stained with surface markers.Cells were then fixed and permeabilized (eBioscience, cat no.00-5523-00) for intracellular cytokine stain.Data were acquired on a BD LSRFortessa and analyzed using the FlowJo software package (FlowJo, LLC).Cell sorting was performed on the BD FACSAriaIII Flow Cytometer and Cell Sorter.

Human lung explants
Human lung explants were procured at the Lung Transplant Center, Division of Pulmonary, Critical Care and Sleep Medicine, Department of Medicine, University of Florida.Donor and patients consent was obtained for a research protocol (UF IRB201902955-Treatment with IFNβ Induces Tolerogenic Lung Dendritic Cells in Human advanced lung disease).Healthy donor lungs were surgically removed postmortem, perfused, small pieces were cut from the right middle and lower lobes for research purpose, and stored in cold Perfadex at 4 °C for no more than 12 hr before processing.Ex planted lungs from emphysema lung transplant patients were stored in cold Perfadex at 4 °C for no more than 12 hr before the process.No lung explants were procured from prisoners.

Statistical analysis
To gain statistical power, we employ three ~four mice/groups to characterize lung immunity.Ten mice/ group to monitor animal health.The statistical justification for group size was calculated using the SAS program to calculate the animal numbers.The analysis was carried out using a standard error of 0.5 for immunological assays, and a power of 0.9.All data are expressed as means ± SEM.Statistical significance was evaluated using Prism 9.0 software.Comparisons between two groups were analyzed by performing an unpaired Student's t test.Comparisons between more than two groups were analyzed by performing a one-way analysis of variance (ANOVA) with Tukey's multiple comparisons test.

Materials availability statement
The Q293 mouse is available upon request via a Material Transfer Agreement.

Figure 2
Figure 2 continued on next page Figure supplement 1. STING1 activation in primary human cells and THP-1 cells reconstituted with WT human STING1.

Figure 4 .
Figure 4. HAQ and AQ alleles prevent SAVI(N153S) disease in mice.(A, B, G) The size and body weight of HAQ/SAVI, AQ/SAVI, WT/SAVI and their littermates WT/HAQ, WT/AQ mice.(D, E, H, I).Airway resistance, and pulmonary artery pressure were determined as described in Materials and methods.(C, F).HAQ/SAVI, AQ/SAVI, WT/SAVI (10 mice/group), were monitored for survival by Kaplan-Meier.(J, K).Representative hematoxylin and eosin (H&E) staining of lung, liver sections from indicated mice.n=3-5 mice/group.Data are representative of three independent experiments.Graphs represent the mean with error bars indication s.e.m. p values are determined by one-way ANOVA Tukey's multiple comparison test.* p<0.05, **p<0.01.n.s.: not significant.(WBC): white blood cells; H: hepatocytes; K: Kupper cells; PV: portal vein; CV: central vein.The online version of this article includes the following figure supplement(s) for figure 4:

Figure 5 continued
Figure 5 continued