5,6-dimethylxanthenone-4-acetic acid (DMXAA), a partial STING agonist, competes for human STING activation

Abstract

5,6-dimethylxanthenone-4-acetic acid (DMXAA) is a mouse-selective stimulator of interferon gene (STING) agonist exerting STING-dependent anti-tumor activity. Although DMXAA cannot fully activate human STING, DMXAA reached phase III in lung cancer clinical trials. How DMXAA is effective against human lung cancer is completely unknown. Here, we show that DMXAA is a partial STING agonist interfering with agonistic STING activation, which may explain its partial anti-tumor effect observed in humans, as STING was reported to be pro-tumorigenic for lung cancer cells with low antigenicity. Furthermore, we developed a DMXAA derivative-3-hydroxy-5-(4-hydroxybenzyl)-4-methyl-9H-xanthen-9-one (HHMX)-that can potently antagonize STING-mediated immune responses both in humans and mice. Notably, HHMX suppressed aberrant responses induced by STING gain-of-function mutations causing STING-associated vasculopathy with onset in infancy (SAVI) in in vitro experiments. Furthermore, HHMX treatment suppressed aberrant STING pathway activity in peripheral blood mononuclear cells from SAVI patients. Lastly, HHMX showed a potent therapeutic effect in SAVI mouse model by mitigating disease progression. Thus, HHMX offers therapeutic potential for STING-associated autoinflammatory diseases.

Keywords: DMXAA derivative; HHMX; SAVI; STING; partial agonist.

Figure 1

DMXAA is a partial agonist for human STING interfering with agonist-mediated STING activation. (A) Fresh human PBMCs from three healthy donors (n=3) were stimulated with the indicated CDNs (10 µg/ml) for 24 h after 90 minutes of DMXAA (100 µg/ml) pretreatment. IFNβ, IFNγ, and IL-6 production was measured using ELISA. Data from each individual from two independent experiments are shown as bar graphs. (B, C) PMA-differentiated THP1 dual reporter cells were pretreated with DMXAA (100 µg/ml) for 90 min and stimulated with the indicated CDNs (10 µg/ml) for 12 or 24 h in three independent experiments. After RNA isolation and cDNA synthesis, the mRNA expression levels of IFNβ (B) and IL-6 (C) were measured using RT-PCR. 18S ribosomal RNA was used as an internal control. (B) IRF activity was determined by measuring secreted luciferase activity in the supernatants. (C) NF-κB activity was determined by measuring secreted embryonic alkaline phosphatase (SEAP) activity in the supernatant. Representative data from three independent experiments are shown as the mean ± SD of triplicates (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, Student’s t test). (D) PMA-differentiated THP1 dual reporter cells were pretreated with DMXAA (100 µg/ml) for 90 min and stimulated with the indicated CDNs (10 µg/ml) for 3 h in three independent experiments. P-TBK1, TBK1, P-IRF3, IRF3, NF-κB and P-NF-κB levels in the cell lysates were detected using western blotting and the blots from a representative experiment were shown. B-actin was used as loading control. Fold induction levels for P-TBK1, P-IRF3 and NF-κB relative to total TBK1, IRF3 and NF-κB from three independent experiments were plotted as mean ± SD (n=3, biological replicates) (*p < 0.05, **p < 0.01, Student’s t test).

Figure 2

Kinetics of the suppressive effect of DMXAA in THP1 dual reporter cells. (A) PMA-differentiated THP1 dual reporter cells were stimulated with various concentrations of DMXAA (1, 3, 10, 30, or 100 µg/ml) together with increasing concentrations of c-di-AMP (3, 10, or 30 µg/ml) for 24 h in three independent experiments. IRF activity was determined by measuring secreted luciferase activity in the supernatants. Representative data from three independent experiments are shown as the mean ± SD of triplicates (**p < 0.01, ****p < 0.0001, One-way ANOVA with Sidak’s multiple comparison test). (B) IFNβ production was measured using ELISA. Representative data from three independent experiments are shown as the mean ± SD of triplicates (**p < 0.01, ***p < 0.001****, p < 0.0001, One-way ANOVA with Sidak’s multiple comparison test). (C) PMA-differentiated THP1 dual reporter cells were stimulated with DMXAA (100 µg/ml) at the indicated time points (−3, 0, or 8 h) together with c-di-AMP (30 µg/ml) for 24 h in three independent experiments. IRF activity was determined by measuring secreted luciferase activity in the supernatants. IFNβ production was measured using ELISA. Representative data from three independent experiments are shown as the mean ± SD of triplicates (*p < 0.05, **p < 0.01, Student’s t test). (D) NF-κB activity was determined by measuring secreted embryonic alkaline phosphatase (SEAP) activity in the supernatant. Representative data from three independent experiments are shown as the mean ± SD of triplicates (*p < 0.05, **p < 0.01, ***p < 0.01, Student’s t test).

Figure 3

DMXAA derivative HHMX is a potent suppressor of STING-induced immune responses not only in human but also in mice. (A) Chemical structures of DMXAA and its derivative HHMX. (B) Fresh human PBMCs from healthy donors (n=10) were stimulated with the indicated CDNs (10 µg/ml) or transfected with PolydA:dT (1 µg/ml) for 24 h after 90 min of DMXAA (100 µg/ml) pretreatment. IFNγ and IFNβ productions were measured using ELISA (*p < 0.05, **p < 0.01, ****p < 0.0001, Student’s t test). (C) PMA-differentiated THP1 dual reporter cells were stimulated with HHMX (30 µg/ml) together with the indicated CDNs (10 µg/ml) for 6 or 24 h in three independent experiments. After the RNA isolation and cDNA synthesis, the mRNA expression levels of IFNβ and IL-6 were measured using RT-PCR. 18S ribosomal RNA was used as an internal control. Representative data from three independent experiments are shown as the mean ± SD of triplicates (**p < 0.01, ***p < 0.001, ****p < 0.0001, Student’s t test). (D) IRF activity was determined by measuring secreted luciferase activity in the supernatants, whereas NF-κB activity was determined by measuring secreted embryonic alkaline phosphatase (SEAP) activity in the supernatant. Representative data from three independent experiments are shown as the mean ± SD of triplicates (***p < 0.001, ****p < 0.0001, Student’s t test). (E) Cell death was measured using lactate dehydrogenase (LDH) release assay (from five independent experiments, n=5, biological replicates) according to manufacturer’s instructions. (F) J774 dual reporter cells were stimulated with increasing concentrations of HHMX together with the indicated CDNs (10 µg/ml), PolydA:dT, or Poly I:C (1 µg/ml) for 24 h in three independent experiments. IRF activity was determined by measuring secreted luciferase activity in the supernatants. NF-κB activity was determined by measuring secreted embryonic alkaline phosphatase (SEAP) activity in the supernatant. Data are shown as the mean ± SD from three independent experiments (n=3, biological replicates) (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, One-way ANOVA with Sidak’s multiple comparison test). (G) IFNβ and IL-6 concentrations in the supernatants were determined using ELISA. Data are shown as the mean ± SD from three independent experiments (n=3, biological replicates) (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, One-way ANOVA with Sidak’s multiple comparison test). (H) J774 dual reporter cells were stimulated with HHMX (30 µg/ml) together with the indicated c-di-AMP (10 µg/ml), PolydA:dT, or Poly I:C (1 µg/ml) for 3 h in three independent experiments. P-TBK1, TBK1 and β-tubulin in the cell lysates were detected using western blotting and the blots from a representative experiment were shown. B-tubulin was used as loading control. Fold induction levels for P-TBK1 relative to total TBK1 from three independent experiments were plotted as mean± SD. Data are representative of three independent experiments (n=3, biological replicates) (**p < 0.01, One-way ANOVA with Sidak’s multiple comparison test).

Figure 4

Kinetics of the suppressive effect of HMMX in THP1 dual reporter cells. (A) PMA-differentiated THP1 dual reporter cells were stimulated with various concentrations of HHMX (1, 3, 10, or 30 µg/ml) together with increasing concentrations of c-di-AMP (3, 10, or 30 µg/ml) for 24 h in five independent experiments. IRF activity was determined by measuring secreted luciferase activity in the supernatants Data are shown as the mean ± SD of five independent experiments (n=5, biological replicates). IFNβ concentration in the supernatants was determined using ELISA. Data are shown as the mean ± SD from three independent experiments (n=3, biological replicates) (*p < 0.05, **p < 0.01, ****p < 0.0001, One-way ANOVA with Sidak’s multiple comparison test). (B) PMA-differentiated THP1 dual reporter cells were stimulated with HHMX (30 µg/ml) at indicated the time points (−3, −1.5, 0, 1.5, or 3 h) together with c-di-AMP (30 µg/ml) for 24 h in three independent experiments. IRF activity was determined by measuring secreted luciferase activity in the supernatants. Representative data from three independent experiments are shown as the mean ± SD of triplicates from one experiment (n=3, technical replicates). IFNβ concentration in the supernatants was determined using ELISA. Representative data from three independent experiments are shown as the mean ± SD of triplicates from one experiment (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, Student’s t test.) (C) NF-κB activity was determined by measuring secreted embryonic alkaline phosphatase (SEAP) activity in the supernatant. Data are shown as the mean ± SD from three independent experiments (n=3, biological replicates) (*p < 0.05, **p < 0.01, Student’s t test.).

Figure 5

HHMX shows therapeutic potential against the STING-mediated autoinflammatory disease SAVI. (A) J774 dual reporter cells were transfected with the plasmids expressing WT STING or STING containing V146L, N153S, V154M, C205Y, R280Q, or R283G mutations together with HHMX (10 µg/ml) co-stimulation for 24 h in four independent experiments. IRF activity was determined by measuring secreted luciferase activity in the supernatants. NF-κB activity was determined by measuring secreted embryonic alkaline phosphatase (SEAP) activity in the supernatant. Data are shown as the mean ± SD of four independent experiments (n=4, biological replicates) (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, Student’s t test.). (B) The concentrations of IFNβ and IL-6 in the supernatants were measured using ELISA. Data are shown as the mean ± SD of four independent experiments (n=4, biological replicates) (**p < 0.01, ***p < 0.001, ****p < 0.0001, Student’s t test.). (C) Fresh human PBMCs isolated from a healthy donor or SAVI patient bearing heterozygous STING N154S mutation were treated with HHMX (30 µg/ml) for 24 h and after protein isolation; the phosphorylation of TBK1 and STAT1 was investigated using western blotting. Graphs showing the band quantifications for P-TBK1 and P-STAT1 after normalization to GAPDH were plotted. (D) Eight–to-thirteen-week-old N153S^+/- SAVI mice (n=4-6) and their WT littermate (n=3-6) controls were orally administered 500 µg of HHMX or control, respectively, for 3 weeks every weekday (5 times/week). (E) Body weight and symptoms of mice were monitored until the day of sacrifice. Body weight percentage on day 18 are shown as dot plots with mean values (*p < 0.05, Student’s t test). (F, G) Splenocytes isolated from the control or HHMX treatment mice were stimulated with cGAMP and c-di-AMP for 24 h and the supernatant (F) IL-6 and (G) IP-10 levels were measured using ELISA. Data is the representative of two independent experiments and individual data from each mouse are shown (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, Student’s t test). (H) The serum IL-6 levels of the HHMX-treated or control mice were measured using ELISA with the serum samples taken on the day of sacrifice. Data is the representative of two independent experiments and individual data from each mouse are shown (*p < 0.05, Student’s t test).

Figure 6

Mode of action of DMXAA and its derivative HHMX in human and mice. Mouse-specific STING agonist DMXAA is a partial STING agonist interfering with agonistic STING activation in humans. Based on this data, we developed a novel DMXAA derivative, 3-hydroxy-5-(4-hydroxybenzyl)-4-methyl-9H-xhanthen-9one (HHMX), which could antagonize STING signaling pathway both in human and mice to become potential therapeutic agent for STING-associated autoinflammatory diseases, like SAVI.