STING-mediated disruption of calcium homeostasis chronically activates ER stress and primes T cell death

Abstract

STING gain-of-function mutations cause lung disease and T cell cytopenia through unknown mechanisms. Here, we found that these mutants induce chronic activation of ER stress and unfolded protein response (UPR), leading to T cell death by apoptosis in the Sting^N153S/+ mouse and in human T cells. Mechanistically, STING-N154S disrupts calcium homeostasis in T cells, thus intrinsically primes T cells to become hyperresponsive to T cell receptor signaling-induced ER stress and the UPR, leading to cell death. This intrinsic priming effect is mediated through a novel region of STING that we name "the UPR motif," which is distinct from known domains required for type I IFN signaling. Pharmacological inhibition of ER stress prevented Sting^N153S/+ T cell death in vivo. By crossing Sting^N153S/+ to the OT-1 mouse, we fully restored CD8⁺ T cells and drastically ameliorated STING-associated lung disease. Together, our data uncover a critical IFN-independent function of STING that regulates calcium homeostasis, ER stress, and T cell survival.

Graphical abstract

Figure 1.

Sting^N153S/+ mouse T cells undergo spontaneous activation and apoptotic cell death. (A) FACS analysis of T cell death. Sting^N153S/+ or WT littermate mouse splenocytes were stained with T cell markers (α-CD3, CD4, or CD8 antibodies) and cell death markers (annexin V or CaspACE FITC-VAD-FMK). Frequencies of cell death of CD3⁺, CD4⁺, or CD8⁺ T cells were analyzed by FACS, and accumulated results from triplicate experiments are shown as bar graphs. Anx V, annexin V; Casp, caspase. (B) Immunoblots and qRT-PCR analysis of apoptosis markers in WT and Sting^N153S/+ mouse T cells. Pan T cells were isolated from spleens of Sting^N153S/+ or WT littermate mice. Indicated apoptotic protein and mRNA expression was analyzed by immunoblots (left) or qRT-PCR (right). Data are representative of at least three independent experiments. (C–E) FACS analysis of WT and Sting^N153S/+ mouse T cell activation and cell death. Fresh isolated splenocytes from Sting^N153S/+ or WT littermate mice were stained with CD4 and CD8 antibodies and T cell activation markers CD69 (C), activation marker CD44 and CD62L (D), or activation marker CD44 and cell death marker caspase (CaspACE FITC-VAD-FMK; E), then analyzed by FACS. (F and G) FACS analysis of mouse T cell proliferation and cell death. Sting^N153S/+ or WT littermate mouse splenocytes were stained with CFSE first, then either mock treated or stimulated with CD3/CD28 antibodies for 3 d. The cell proliferation of CD8⁺ T cells was analyzed by CFSE dilution (F). In separate experiments, splenocytes were stimulated with or without CD3/CD28 antibodies for 1 d, then CD4⁺ or CD8⁺ T cells were analyzed by annexin V staining (G). (H and I) FACS analysis of T cell proliferation and cell death after STING agonist DMAXX treatment. WT splenocytes were stained with CFSE first, then either mock treated or stimulated with DMXAA in the presence or absence of CD3/CD28 antibodies for 3 d. CD8⁺ T cell proliferation was analyzed by CFSE dilution as in F. Annexin V staining of CD4⁺ or CD8⁺ T cells was also done similar to G. Error bars, SEM. **, P < 0.01; ***, P < 0.001; ns, not significant. Student’s t test. Data in all panels are representative of at least three independent experiments.

Figure 2.

Chronic activation of ER stress and the UPR induces T cell death in the Sting^N153S/+ mouse. (A) Immunoblot analysis of the UPR and apoptosis in WT and Sting^N153S/+ immune cells. T cells, B cells, and macrophages were isolated from littermate WT and Sting^N153S/+ mice; and lysates from fresh isolated cells were used for immunoblot analysis of indicated proteins. (B) Immunoblot analysis of the UPR in primary MEFs or human fibroblasts (HFB). The primary MEF cells were isolated from WT and Sting^N153S/+ E13.5 embryos. Human skin fibroblasts were isolated from a healthy donor control (HC) or two independent SAVI patients (pt3 and pt6; Liu et al., 2014). (C) Immunoblot analysis of the UPR and apoptosis in WT and Sting^N153S/+ T cells. Pan T cells from littermate WT and Sting^N153S/+ mice were either mock treated (Vehicle) or stimulated with CD3/CD28 antibodies for 3 h. Lysates from treated cells were used for immunoblot analysis of indicated proteins. (D) FACS analysis of cell proliferation and cell death. Pan T cells were isolated from WT and Sting^N153S/+ mice, then either untreated (UT) or mock treated (DMSO) or treated with TUDCA or QVD-OPH in the presence or absence of CD3/CD28 antibodies. The cells were then stained with CaspACE FITC-VAD-FMK In Situ Marker and analyzed by FACS. (E) In vivo treatment of Sting^N153S/+ mice. Top: schematic diagram of the experimental design. Bottom: FACS analysis of CD4⁺ and CD8⁺ T cells in mouse peripheral blood (PB). Sting^N153S/+ mice were i.p. injected with PBS, TUDCA (250 mg/kg, three injections per wk), or QVD-OPH (10 mg/kg, three injections per wk). T cells from the peripheral blood were analyzed after 14 d. n = 8–10. Error bars: SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant. Student’s t test. Data shown are from a representative dataset from two repeat treatment experiments.

Figure 3.

Human STING-N154S primes T cell death through ER stress. (A) Kinetics of cell death measured by IncuCyte. Reconstituted Jurkat T cells were mock or Dox treated for 1 d, then stimulated with or without PMA+ionomycin (PMA+Iono) for another day. Cell death kinetics was monitored using IncuCyte with a live-cell compatible caspase 3/7 dye (CellEvent Caspase-3/7; see Materials and methods). (B and C) qRT-PCR analysis of UPR gene expression in reconstituted Jurkat T cells. Experiments in B are Jurkat T cells reconstituted with Vector, WT STING, or N154S mutant (indicated as different color bars) that were either mock treated or treated with Dox alone, PMA+ionomycin alone, or Dox and PMA+ionomycin (indicated on bottom). Experiments in C are all N154S cells, treated with indicated conditions on the right. UPR gene expression was measured by qRT-PCR. (D) Kinetics of cell death measured by IncuCyte in N154S cells. N154S cells were treated with Dox, PMA+ionomycin, and increasing doses of TUDCA (Tud; 50 µM, 100 µM, and 200 µM). Cell death kinetics was monitored by IncuCyte as in A. (E) Kinetics of cell death measured by IncuCyte in N154S/S366A cells. N154S/S366A cells were treated with Dox and PMA+ionomycin with and without TUDCA (100 µM). (F) IncuCyte analysis of reconstituted Jurkat T cells treated with Tg. Jurkat T cells reconstituted with indicated STING construct (top) were mock treated or treated with Tg (as indicated on the right). (G and H) T cell death analysis by annexin V staining and FACS analysis. Mouse primary T cells were stimulated for 1 d (G) or 2 d or 3 d (H) with mock or the indicated concentration of DMXAA in the presence or absence of CD3/CD28 antibodies. Both high-dose (10 µM) and low-dose (1.25 µM) DMXAA were analyzed at 1 d. Only low-dose DMXAA conditions were analyzed at 2 and 3 d. SKO, Sting^−/−. (I) qRT-PCR analysis of UPR gene expression in mouse primary T cells treated with DMXAA in the presence or absence of CD3/CD28 antibodies for 24 h. Error bars: SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant. Student’s t test. Data shown in all panels are representative of at least three independent experiments.

Figure 4.

A novel STING motif that regulates ER stress and the UPR. (A) qRT-PCR analysis of UPR gene expression in HEK293T cells. HEK293T cells were transiently transfected with indicated cGAS or STING or both expressing plasmids. The IFNB- and UPR-related gene expression was measured by qRT-PCR 36 h after transfection. (B) Schematic diagrams of STING WT and truncation constructs and data summary. (C, E, and F) Structural and functional domain-mapping studies. HEK293T cells were transiently transfected with indicated plasmids (x axis). GADD34, ATF3, and IFIT1 mRNA expressions were measured by qRT-PCR 36 h after transfection. (C) Truncation study of known STING domains. (E) Further fine mapping to define the UPR domain and critical residues in this region. FL, full length. (F) Comparison of the UPR motif versus the miniCTT motif by internal deletions. In the bottom panel of F, HEK293T cells were transfected with NF-κB luciferase reporter and indicated expressing plasmid, and the luciferase activity was assayed 24 h after transfection. The P value was calculated by comparing the indicated group with the vector control group. (D) The crystal structure of STING C terminus (4EF5) with key functional residues highlighted. Alignment of the UPR motif and CTT of STING homologs across different specifies are shown below. Error bars: SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001. Student’s t test. Data shown are representative of at least three independent experiments.

Figure 5.

RRDD mutation abolishes STING-N154S–mediated ER-to-Golgi translocation, UPR, and T cell death. (A) qRT-PCR analysis of UPR gene expression in reconstituted Jurkat T cells. Jurkat T cells reconstituted with vector, STING, N154S, or N154S/RRDD were induced with Dox (0.5 µg/ml) for 1 d, then stimulated with or without PMA (10 nM) + ionomycin (1 µM; PMA+Iono) for another day. UPR gene expression was measured by qRT-PCR. (B) IncuCyte measurement of cell death kinetics in N154S/RRDD Jurkat cells. N154S/RRDD cells were treated with conditions indicated on the right, and cell death kinetics was measured by IncuCyte. (C) Fluorescent microscopy images of STING localization. Sting^−/− MEFs were stably transduced with a retroviral vector expressing WT STING or indicated mutant STING (N154S or N154S/RRDD). STING localization was detected by immunofluorescence staining of STING. IRE1 was costained as an ER marker and GM130 was costained as a Golgi marker. Bars, 10 µm. (D) STIM1 (aa1–249) ectopic expression reduces ER stress and N154S Jurkat T cell death. N154S Jurkat T cells were reconstituted with empty vector or STIM1 (aa1–249). The cells were then mock treated or stimulated with PMA+ionomycin in the presence or absence of Dox. Cell death (top panel) was analyzed by annexin V staining. Expression of UPR genes (middle and bottom panels) was measured by qRT-PCR. STIM1 (aa1–249) expression is shown on the right. Error bars: SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant. Student’s t test. Data shown are representative of at least three independent experiments.

Figure 6.

STING-N154S disrupts ER calcium homeostasis. (A) Relative changes in cytosolic Ca²⁺ concentration monitored by Fura-2 ratios in Jurkat T cells expressing indicated STING protein. Cells were stimulated with 1 µM Tg and 3 mM EGTA; 2 mM Ca²⁺ were added back 6 min after stimulation. Fura-2 intensities at 340-/380-nm excitation were recorded. Intensity ratios for Fura-2 are plotted. Curve graph (left) and bar graphs (right): mean with SEM (number of cells measured in each experiment was as follows: Vector, 580; STING, 589; STING-N154S, 633; STING-N154S/RRDD, 933). (B) Relative changes in cytosolic Ca²⁺ concentration in primary mouse T cells from two littermate pairs of WT and Sting^N153S/+ mice. Cells were loaded with Fura2 and subjected to the same stimulation as in A. Fura-2 intensities at 340-/380-nm excitation were recorded. Intensity ratios for Fura-2 are plotted. Curve graphs: mean with SEM (number of cells measured in each experiment was as follows: WT-pair 1, 263; STING-N153S-pair 1, 194; WT-pair 2, 242; STING-N153S-pair 2, 271). Data are from a representative trace from two independent experiments. (C) qRT-PCR analysis of UPR gene expression in HEK293T cells treated with Tg or Tm for 16 h. (D) NFAT-Luc reporter assay in HEK293T cells treated with Tg or Tm. HEK293T cells were transfected with NFAT-Luc and TK-Renila Luc (control) plasmids for 12 h, then treated with Tg or Tm with and without cyclosporine A (CsA). Dual-luciferase activity was measured 16 h after treatment. (E and F) UPR gene expression and cell death analysis of N154S Jurkat T cells treated with Tg or increasing concentrations of Tm. N154S Jurkat T cells were stimulated with fixed amount Tg (20 nM) and increasing concentrations of Tm (range, 0.0625–1 μg/ml). UPR gene expression was measured by qRT-PCR (E). Cell death was measured 48 h after treatment (F). Error bars: SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005. Student’s t test. Data shown are representative of at least three independent experiments.

Figure 7.

Altered TCR in Sting^N153S/+ mice eliminates ER stress and T cell death. (A) Immunoblot of UPR proteins in fresh isolated T cells. The total T cells were isolated from the spleens of WT, Sting^N153S, OT-1, or Sting^N153SOT-1 mice. UPR protein Bip and Chop were detected by Western blot. (B) FACS cell death analysis of CD8⁺ T cells from spleens of indicated mouse strains (bottom). Total splenocytes were stained for CD8 and cell death markers as in Fig. 1 A. (C) T cell population analysis of indicated mouse strains. Total splenocytes isolated from the indicated mouse strains were stained for CD4 and CD8. Representative FACS plots are shown on the left, and CD8⁺ T cell percentages and cell numbers are summarized on the right. (D) T cell activation analysis of indicated mouse strains. Total splenocytes isolated from indicated mouse strains were stained for CD8, CD44, and CD62L. Representative FACS plots are shown on the left, and CD44^hi (antigen experienced or activated) in the CD8⁺ population are summarized on the right. (E) T cell proliferation analysis of indicated mouse strains. Total splenocytes were isolated from indicated mouse strains and stained with the CFSE dye. Then, cells were stimulated with CD3/CD28 antibodies for 3 d. CFSE dilution in the CD8⁺ T cell population was analyzed by FACS. A representative set of CFSE dilution plots is shown on the top, and summarized data from n = 4 mice per genotype are shown on the bottom. (F) T cell death analysis of indicated mouse strains. Total splenocytes from indicated mouse strains were stimulated with CD3/CD28 antibodies ex vivo for 3 d. Cells were then stained for CD8 and cell death markers and analyzed by FACS. A representative set of FACS plots is shown on the left, and summarized data from n = 4 mice per genotype are shown on the right. (G) Immunoblot of UPR proteins in indicated mouse T cells after CD3/CD28 antibody stimulation. Total T cells were isolated from spleens of indicated mouse stains, then stimulated with CD3/CD28 antibodies for 1 d. UPR proteins were detected by Western blot. Error bars: SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001. Student’s t test. Anx V, annexin V; Casp, caspase.

Figure 8.

Restoration of CD8⁺ T cells relieves inflammation and lung disease. (A) Spleen size and total splenocyte number of indicated mouse strains. (B) Histopathological analysis of the lung from indicated mouse strains. H&E images of paraffin-embedded lung sections from 3-mo and sex-matched WT littermate, Sting^N153S, OT-1, and Sting^N153S OT-1 mice. n = 4. Bars, 500 µm. (C and D) Multiplex cytokine analysis of serum from indicated mouse strains. The overall cytokine comparison is shown in the heat map (C), and the representative cytokine concentrations are shown on the bar graphs (D). n = 4. Error bars: SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant. Student’s t test.