cGAS activation converges with intracellular acidification to promote STING aggregation and pyroptosis in tumor models

Abstract

The cyclic GMP-AMP synthase (cGAS)/stimulator of IFN genes (STING) pathway is intimately associated with antitumoral immunity; however, the direct involvement of this pathway in tumor cell demise remains elusive. Here, we identified a compound, dodecyl 6-hydroxy-2-naphthoate (DHN), that induces pyroptosis in melanoma cells by activating noncanonical cGAS/STING signaling. DHN targets mitochondrial protein cyclophilin D (CypD) to induce the release of mitochondrial DNA, leading to cGAS activation and cyclic GMP-AMP (cGAMP) generation. Meanwhile, DHN-caused intracellular acidification induces protein kinase R-like endoplasmic reticulum kinase (PERK) activation, which promotes STING phosphorylation and polymerization in the presence of cGAMP, thereby facilitating the aggregation of STING in the ER, which serves as a platform to recruit Fas-associated via death domain (FADD) and caspase-8, leading to caspase-8 activation and subsequent gasdermin E cleavage, which ultimately results in pyroptosis of tumor cells and tumor regression in mouse models. The occurrence of this noncanonical cGAS/STING pathway-associated pyroptosis is also observed when both cGAS is activated and intracellular pH declines. Collectively, our findings reveal a pathway that links noncanonical cGAS/STING signaling to gasdermin E-mediated pyroptosis, thereby offering valuable insights for tumor therapy.

Keywords: Cancer; Cell biology; Mitochondria; Oncology.

Figure 1. DHN induces pyroptosis by caspase-8–mediated cleavage of GSDME.

Melanoma A375 cells were treated with DHN (15 μM) for 20 hours to assess pyroptotic features (including characteristic morphology, GSDME cleavage, and LDH release), unless specifically defined. (A) Compound library screening and chemical structure of DHN. (B) DHN-induced pyroptosis at different time points and cells with characteristic pyroptotic morphology indicated by red arrows. Cleavage of GSDME was detected by Western blot, and cell death was evaluated by accessing LDH release. (C–F) GSDME (C) or caspase-8 (F) were separately knocked down in cells or cells were cotreated with Z-VAD (D, 20 μM) or Z-IETD (E, 10 μM), followed by detection of pyroptosis. (G) GSDME^WT, GSDME^D229A, GSDME^D251A, GSDME^D256A, GSDME^D267A, GSDME^D270A, or GSDME^D279A were separately transfected into GSDME-knockdown cells, then the cleavage level of GSDME was detected. (H) GSDME^WT or GSDME^D270A were separately transfected into GSDME-knockdown cells, then pyroptosis was detected. Tubulin was used to determine the amount of loading protein. Data are presented as mean ± SEM of 3 independent experiments. Statistical analyses were determined by 1-way ANOVA with Tukey’s multiple-comparison test (B) and 2-way ANOVA with Tukey’s multiple-comparison test (C–F and H). P values are indicated in figures. Scale bars: 100 μm. All Western blots were repeated at least twice, and 1 of them is shown. IB, immunoblot; LE, long exposure; SE, short exposure.

Figure 2. DHN promotes the opening of mPTP by targeting mitochondrial protein CypD.

Melanoma A375 cells were treated with DHN (15 μM) for 20 hours to assess pyroptotic features (including characteristic morphology, caspase8/GSDME cleavage, and LDH release), unless specifically defined. (A) Chemical structure of DHN probe (DHN-P, left) and workflow of click chemistry for DHN-P (right). (B) Cells were treated with DHN-P (150 μM) for 2 hours. Azide-rhodamine was conjugated with DHN-P, and the localization of DHN-P is shown (Tom20, mitochondria marker; CALR, ER marker; GM130, Golgi marker; LAMP2, lysosomal marker). Scale bar: 20 μm. (C and D) Cells were treated with DHN in the presence of CsA (C, 5 μM) or in CypD-knockdown cells (D), followed by the detection of pyroptosis. Scale bars: 100 μm. (E) Cells were treated with DHN-P (150 μM) for 2 hours; azide-biotin was added to conjugate with DHN-P. DHN-P–targeted CypD was assayed by streptavidin beads. (F) The binding affinity between DHN and CypD was determined by surface plasmon resonance. (G) Cellular thermal shift assay. The proteins of CypD^WT or CypD^R97A/Q105A were immunoprecipitated from cells, followed by treatment with DHN and subsequent differential temperature incubation for 15 minutes. Resulting lysates were subjected to Western blot analysis. (H) CypD^WT or CypD^R97A/Q105A were separately transfected into CypD-knockdown cells, followed by detection of pyroptosis. Scale bar: 100 μm. (I) CypD was knocked down in cells, or cells were cotreated with CsA (5 μM) for 12 hours, followed by detection of mPTP opening. Tubulin was used to determine the amount of loading protein. DAPI was used to indicate nucleus in confocal microscopy. Data are presented as mean ± SEM of 3 independent experiments. Statistical analyses were determined by 2-way ANOVA with Tukey’s multiple-comparison test (C, D, G, H, and I). P values are indicated in figures. All Western blots were repeated at least twice, and 1 of them is shown.

Figure 3. DHN-induced mtDNA release activates the cytosolic cGAS.

Melanoma A375 cells were treated with DHN (15 μM) for 12 hours to show the puncta of cGAS and detect the release of mtDNA; and for 20 hours to assess pyroptotic features (including characteristic morphology, caspase8/GSDME cleavage, and LDH release), unless specifically defined. (A) Cells were cotreated with CsA (top, 5 μM), and CypD (middle) or ANT1 (bottom) was knocked down in cells, followed by detection of mtDNA release. (B and C) Cells were cotreated with CsA (B, 5 μM), and CypD or ANT1 (C) was knocked down in cells, then stained with anti-cGAS antibody. cGAS puncta were observed under confocal microscope. Scale bars: 20 μm. The percentage of cells with cGAS puncta was quantified (right, mean ± SEM, n = 3 repeats). The quantification and counting of 100 cells were performed 3 times in a single experiment, and the average value obtained from the 3 statistical measurements was recorded as 1 repetition. (D–F) Cells were cotreated with G140 (E, 30 μM), cGAS was knocked out (D), or STING was knocked down (F) in cells, followed by the detection of pyroptosis. Scale bars: 100 μm. Tubulin was used to determine the amount of loading protein. Data are presented as mean ± SEM of 3 independent experiments. Statistical analyses were determined by 2-way ANOVA with Tukey’s multiple-comparison test (A–F). P values are indicated. All Western blots were repeated at least twice, and 1 of them is shown.

Figure 4. DHN induces the formation of STING aggregates to recruit caspase-8 and GSDME.

Melanoma A375 cells were treated with DHN (15 μM) for 12 hours to show the puncta of STING in the ER and detect various proteins in the Triton X-100–insoluble (TI) fractions, unless specifically defined. (A) cGAS was knocked out in cells, or cells were cotreated with G140 (30 μM), and stained with STING antibody. STING puncta were observed under confocal microscope (left). Scale bars: 20 μm. The percentage of cells with STING puncta was quantified (right, mean ± SEM, n = 3 repeats). (B) Living cells were treated with DHN; puncta of STING and ER shown. Scale bars: 10 μm. (C and D) Observation of STING-associated ER structure using electron microscopy. A375 cells (C) or STING/APEX-expressing A375 cells (D) were treated with DHN for 12 hours; the ER morphology and the location of STING in ER was observed. Scale bars: 1 μm and 5 μm. (E) Indication of STING-associated organelles. Cells were transfected with STING-HA, the STING-associated organelles were immunoprecipitated, and then indicated by various antibodies (CALR, ER marker; GM130, Golgi marker; Tom20, mitochondria marker; LDHA, cytosol marker). (F) Cells were treated with DHN; puncta of STING and GSDME or caspase-8 shown. Scale bars: 5 μm and 10 μm. (G) The CypD-knockdown (left), STING-knockdown (right), or cGAS-knockout (middle) cells were treated with DHN. The localization of STING, cleaved CASP8, and GSDME in the TI is indicated. (H) Cells were transfected with GFP-V5-turboID or STING-V5-turboID and then labeled with biotin (100 μM) for 10 minutes; the biotin-labeled proteins were isolated and indicated by corresponding antibodies. Data are presented as mean ± SEM of 3 independent experiments. Statistical analyses were determined by 2-way ANOVA with Tukey’s multiple-comparison test (A). P values are indicated. All Western blots were repeated at least twice, and 1 of them is shown.

Figure 5. DHN-caused ER acid environment promotes the formation of STING aggregates.

Melanoma A375 cells were treated with DHN (15 μM) for 12 hours to show the puncta of STING in the ER and to detect STING polymer and location of various proteins in TI and for 20 hours to assess pyroptotic features, unless specifically defined. (A) CypD was knocked down in cells (A, left), or cells were cotreated with G140 (A, right, 30 μM), and then polymer of STING was indicated. (B and C) STING^WT, STING^C206S, and STING^C148S were separately transfected into STING-knockdown cells, then polymer of STING was indicated (B). STING puncta were observed under confocal microscope (C, left); scale bar: 20 μm. The percentage of cells with STING puncta was quantified (C, right). (D) Cells were treated with DHN at different concentrations (left), or were cotreated with NH₄Cl (right, 5 mM), followed by measurement of cytosolic pH values. (E–I) Cells were cotreated with NH₄Cl (5 mM), followed by detection of polymer of STING (E); ER morphology using electron microscope (F), scale bars: 1 μm and 2 μm (zoom); STING and ER puncta using confocal microscope (G), scale bar: 20 μm; the localization of STING, cleaved-CASP8, and GSDME in the TI (H); and pyroptosis, scale bar: 100 μm (I). Tubulin was used to determine the amount of loading protein. Data are presented as mean ± SEM of 3 independent experiments. Statistical analyses were determined by 1-way ANOVA with Tukey’s multiple-comparison test (D, left) and 2-way ANOVA with Tukey’s multiple-comparison test (C and D, right; and I). P values are indicated. All Western blots were repeated at least twice, and 1 of them is shown.

Figure 6. cGAS activation converges with intracellular acidification to induce pyroptosis.

Melanoma A375 cells were treated with different stimulants for 20 hours to assess pyroptotic features (including characteristic morphology, caspase8/GSDME cleavage, and LDH release), unless specifically defined. (A) Cells were cotreated with lactic acid (20 mM) and 2,3′-GAMP (10 μg/mL) or diABZI (10 μM), followed by the detection of pyroptosis. (B) Cells were infected with HSV1 (10 MOI) in the presence of lactic acid (20 mM) or HCl (20 mM), followed by detection of pyroptosis. Tubulin was used to determine the amount of loading protein. Data are presented as mean ± SEM of 3 independent experiments. Statistical analyses were determined by 2-way ANOVA with Tukey’s multiple-comparison test. Scale bars: 100 μm (A and B). P values are indicated. All Western blots were repeated at least twice, and 1 of them is shown.

Figure 7. DHN-induced phosphorylation of STING by PERK facilitates the polymerization of STING.

Melanoma A375 cells were treated with DHN (15 μM) for 12 hours to detect STING phosphorylation, the puncta of STING in the ER, and STING polymer and location of various proteins in TI; and for 20 hours to assess pyroptotic features, unless specifically defined. (A and B) Control (A, top) or PERK-knockdown A375 cells (B, bottom) were treated with DHN in the presence of NH₄Cl (A, bottom, 5 mM) or GSK2656157 (B, top, 10 μM). Cell lysates were incubated with calf intestinal alkaline phosphatase (CIAP) (A, middle). STING phosphorylation was analyzed using Phos-tag assays. (C) Cell lysates were incubated with CIAP (top). Cells were cotreated with GSK2656157 (middle, 10 μM), and PERK^WT or PERK^T982A were separately transfected into PERK-knockdown cells (bottom). (D) Cells were cotreated with GSK2656157 (10 μM); the interaction between STING and PERK was determined. (E) Cells were cotreated with NH₄Cl (top, 5 mM) and DHN, or treated with lactic acid (middle and bottom, 20 mM), followed by the detection of PERK and eIF2α phosphorylation. (F–H) Cells were cotreated with GSK2656157 (10 μM) or subjected to PERK knockdown, followed by detection of STING polymerization (F); STING puncta (G), scale bar: 20 μm; and pyroptosis, scale bar: 100 μm (H). Tubulin was used to determine the amount of loading protein. Data are presented as mean ± SEM of 3 independent experiments. Statistical analyses were determined by 2-way ANOVA with Tukey’s multiple-comparison test (G and H). P values are indicated. All Western blots were repeated at least twice, and 1 of them is shown.

Figure 8. Phosphorylation of STING at Ser345 and Ser358 by PERK is critical for DHN-induced pyroptosis.

Melanoma A375 cells were treated with DHN (15 μM) for 12 hours to detect STING phosphorylation and the puncta of STING in the ER; and for 20 hours to assess pyroptotic features, unless specifically defined. (A) STING^WT and STING^S345A/S358A were transfected into STING-knockdown cells, followed by the detection of STING phosphorylation (top). STING^WT and STING^S345A/S358A was incubated with PERK in vitro (bottom). (B–D) STING^WT and STING^S345A/S358A were transfected into STING-knockdown cells, followed by detection of STING polymerization (B); STING puncta (C), scale bar: 20 μm; and pyroptosis, scale bar: 100 μm (D). Tubulin was used to determine the amount of loading protein. Data are presented as mean ± SEM of 3 independent experiments. Statistical analyses were determined by 2-way ANOVA with Tukey’s multiple-comparison test (C and D). P values are indicated. All Western blots were repeated at least twice, and 1 of them is shown.

Figure 9. DHN inhibits tumor growth by inducing pyroptosis in mice.

A375 cells (2 × 10⁶) were injected subcutaneously into the posterior flanks of nude mice. After 4 days, DHN was intraperitoneally administered to the mice every other day for 2 weeks. The tumor volume and weight were recorded at the indicated times. (A–D) A375 cells were injected into BALB/c-nu mice to form subcutaneous xenografts (A, n = 6, scale bar: 1 cm). The expression of Ki67 is shown (B; n = 9 fields from 3 independent tumor tissues; scale bar: 100 μm). Tumors were collected for detection of GSDME (C). STING puncta are indicated by white arrows (D, left, scale bar: 20 μm), and the percentage of cells with STING puncta was quantified (D, right; n = 9 fields from 3 independent tumor tissues). (E–H) A375 cells with or without knockdown of CypD (E and F). Scale bar: 1 cm (E). STING or GSDME (G and H) was injected into BALB/c-nu mice to form subcutaneous xenografts (n = 8). Tumors were collected for detection of GSDME and monomers, dimers, and oligomers of STING. Scale bar: 1 cm (G). Tubulin was used to determine the amount of loading protein. Statistical analyses were determined by unpaired 2-tailed Student’s t test (A, B, and D) and 2-way ANOVA with Tukey’s multiple-comparison test (E and G). P values are indicated.

Figure 10. DHN induces antitumor immune responses in a mouse tumor model.

A375 cells (2 × 10⁶) were injected subcutaneously into the posterior flanks of nude mice. After 4 days, DHN was intraperitoneally administered to the mice every other day for 2 weeks. The tumor volume and weight were recorded at the indicated times. (A–C) A375 STING-knockdown cells with expression of STING^WT, STING^C206S, or STING^S345A/358A were injected into BALB/c-nu mice to form subcutaneous xenografts (I, n = 7). Tumors were collected for detection of GSDME (A), scale bar: 1 cm; monomers, dimers, and oligomers of STING (B); and STING phosphorylation (C). (D) B16 (top, n = 10) or Hepa1-6 (bottom, n = 8) cells were injected into C57BL/6 mice to form xenografts. DHN was intraperitoneally administered to the mice. Scale bar: 1 cm. (E) B16 cell–derived xenograft tumors were collected 24 hours after DHN (10 mg/kg) administration and then analyzed using flow cytometry to determine the proportion and activation status of immune cells within the tumor microenvironment (n = 5). Tubulin was used to determine the amount of loading protein. Statistical analyses were determined by unpaired 2-tailed Student’s t test (D and E) and 2-way ANOVA with Tukey’s multiple-comparison test (A). P values are indicated.