Karen L. Leach^‡ and Justin D. Hall^§

^‡MolPharm Consulting, West Haven, CT; ^§Pfizer, Worldwide Medicinal Chemistry, Groton, CT

Innate immunity is a key pathway activated in response to bacterial and viral infections. The invasion of foreign nucleic acids results in the production of interferons and cytokines that comprise the host defense. More recently, roles for innate immunity in tumorigenesis, autoimmune disease, and senescence also have been elucidated. In these diseases, it is ‘self’ nucleic acids, RNA and DNA (including mitochondrial DNA), which escape into the cytosol and trigger immune responses. Two proteins, cyclic GMP-AMP synthase (cGAS) and stimulator of interferon genes (STING), are the key sensors of nucleic acids in the immunity pathway. Emerging data for the regulation of cGAS and STING suggests that under some conditions, such as autoimmune disease, inhibition of the pathway is desirable, while in the case of tumor immunity, stimulation of cytokine production is beneficial. This suggests cGAS and STING are potential therapeutic targets, though the details of activation or inhibition remain to be elucidated.^1-4

Nucleic acids activate a number of cytosolic sensors, including retinoic acid-inducible gene I (RIG-I) and melanoma differentiation-associated protein 5 (MDA5) for RNA and absent in melanoma 2 (AIM2), DNA-dependent activator of interferon-regulatory factors (DAI), and interferon-γ-inducible protein 16 (IFI16) for DNA.⁵ An unresolved issue in the immunity field, however, was that none of the DNA sensors completely accounted for interferon (IFN) production. This conundrum was solved in 2013 with the discovery of a new cytosolic DNA sensor, cGAS, and accumulating evidence now suggests cGAS is the primary sensor in innate immune activation.^6,7

Cytosolic cGAS binds double-stranded DNA and catalyzes the production of the novel second messenger 2ʹ-3ʹ-cyclic AMP-GMP (2ʹ3ʹ-cGAMP) from ATP and GTP. cGAMP binds to the ER-resident protein STING. The resulting conformational change in STING leads to the recruitment of the kinase TANK-binding kinase 1, IFN-inducible gene activation, and IFN production via IRF3 phosphorylation and nuclear translocation (Figure 1).

Figure 1. Cytosolic DNA and cyclic dinucleotides trigger the innate immune system through IFN production. Double-stranded DNA-bound cGAS produces the secondary messenger cGAMP that binds and activates STING resulting in the production of interferon (IFN).

Studies comparing cyclic dinucleotide binding and STING activation identified genetic variants of human STING and important differences between human and mouse STING. The five human haplotypes are denoted WT (R232), REF (R232H), HAQ (R71H, G230A, R293Q), AQ (G230A, R293Q), and Q (R293Q).^8,9 Metazoan 2ʹ3ʹ-cGAMP contains G(2ʹ5ʹ)-pA(3ʹ5ʹ) phosphodiester linkages in contrast to bacterial cGAMP (3ʹ3ʹ-cGAMP), which contains G(3ʹ5ʹ)pA(3ʹ5ʹ) linkages. The R232H allele, which occurs in 14% of the population, responds to 2ʹ3ʹ-cGAMP, but responds weakly to bacterial cyclic dinucleotides. In contrast, STING-HAQ responds to both metazoan and bacterial cGAMPs and is found in 20% of the population. DMXAA (5,6-dimethylxanthenone-4-acetic acid, Vadimezan) is a small molecule that exhibits immune modulatory activities via induction of cytokines and shows efficacy in mouse tumor models.¹⁰ This compound was taken into clinical trials in combination with paclitaxel and carboplatin but failed in the phase III trials.¹¹ Although mouse and human STING share high sequence identity, it was shown subsequently that DMXAA activates mouse but not human STING. Mutation at a single cyclic-dinucleotide binding site of human STING (S162A) allows DMXAA binding and restores sensitivity.¹² 

Activation of cGAS and STING is important in host defense against pathogens, but uncontrolled activation of this pathway has been implicated in autoinflammatory disease including type I interferonopathies such as Aicardi-Goutières syndrome (AGS), a severe inflammatory disease, and systemic lupus erythematosus (SLE). Self-DNA that escapes into the cytosol is normally degraded by the primary mammalian exonuclease TREX1. TREX1 is one of seven human genes whose mutation causes AGS, and a small percentage of SLE patients have TREX1 mutations.^13,14 TREX1 knockout mice have elevated levels of dsDNA, elevated levels of cGAMP, and display multi-organ inflammation (especially myocarditis) leading to morbidity.^15,16 The double TREX1/cGAS knockout rescues the TREX1 phenotype, demonstrating a key role for cGAS stimulation in autoinflammation.¹⁷ Elevated levels of cGAMP have been reported in a subset of SLE patients with a more severe disease phenotype (as shown by higher SLEDAI scores) compared to SLE patients in whom no cGAMP was detected.¹⁸ In the case of STING, gain-of-function mutations result in the autoinflammatory disease SAVI (STING-associated vasculopathy with onset in infancy), characterized by interferonopathy, which causes skin lesions, interstitial lung disease, and systemic inflammation.¹⁹

In contrast to autoimmunity, in tumorigenesis the cGAS/STING pathway can be both stimulatory as well as inhibitory. Tumor-derived DNA is taken up by dendritic cells (DC) and activates cGAS/STING, resulting in type 1 IFN production and DC maturation. This innate immunity activation induces an adaptive immune response by stimulating CD8⁺ T cell priming, resulting in a tumor antigen-specific T cell response to kill cancer cells. In a mouse melanoma model, PD-L1 antibody treatment resulted in increased levels of tumor-specific CD8⁺ T cells in wild-type (WT) mice but not in cGAS knockout or mice which do not express STING (“golden ticket mice”). Likewise, tumor volume was decreased and survival was increased only in the WT mice, demonstrating the dependence on cGAS/STING. Direct intramuscular injection of cGAMP reduced tumor size in this and other mouse tumor models. The efficacy of DMXAA treatment in mouse tumor models has led to the discovery of stable cyclic dinucleotide STING agonists that activate human STING and show antitumor efficacy in colon, breast, and melanoma models.^20-22 The cancer vaccine STINGVAX that combines stable cyclic dinucleotide STING agonists with granulocyte-macrophage colony-stimulating factor (GM-CSF) is effective in multiple tumor models.²³ Clinical trials are ongoing to test the effect of STING agonists in patients with advanced/metastatic solid tumors or lymphomas (NCT02675439).

The complexity of the role of the cGAS/STING pathway in tumor immunity is increased by the observation that under some circumstances activation of this pathway promotes tumorigenesis. Mutagens such as 7,12-dimethylbenz(a)anthracene (DMBA) as well as cisplatin and etoposide induce nuclear DNA leakage, activation of cGAS and STING, and the production of proinflammatory cytokines such as IL-1 and TNF-α. These cytokines stimulate phagocyte infiltration, resulting in an increased inflammatory response and tumor development. STING-deficient mice are resistant to DMBA-induced skin cancer, demonstrating the requirement for cGAS/STING activation.²⁴ In contrast, STING-deficient mice developed colonic tumors at an enhanced frequency compared to WT mice.²⁵ STING agonists induced indoleamine 2,3-dioxygenase (IDO) activity, an immune checkpoint that activates regulatory T cells and suppresses immunity, and promoted tumor growth in a Lewis lung carcinoma model.^26,27 These examples illustrate that there are many details yet to be uncovered in how cGAS/STING are controlled in the context of DNA sensing. It appears that the ability of innate immune pathways to modify tumorigenesis depends on multiple factors including acute or chronic DNA exposure, level of STING activation, tumor cell types and location, and the tumor microenvironment.

As the biological roles of cGAS and STING unfold, assessing their potential as therapeutic targets is a critical next step. The identification of activators and inhibitors is key to that process. As outlined above, considerable progress has been made in identifying STING agonists, including testing them clinically. In the case of cGAS, we at Pfizer and others have established cGAS assays utilizing purified cGAS suitable for high-throughput screening.^28,29 Riboswitch aptamers have been used to measure bacterial dinucleotides, and Bose, et al. engineered an aptamer to measure 2ʹ3ʹ-cGAMP with high specificity.²⁸ This aptamer was used to measure cGAMP in a biochemical assay as well as cGAMP levels in DNA-stimulated L929 cells overexpressing cGAS.²⁸ Our group at Pfizer developed a specific monoclonal antibody that recognizes cGAMP and used it to establish a fluorescence polarization assay. Using this assay as well as structural and biophysical studies, we identified a low affinity fragment hit that was chemically optimized to bind cGAS with high affinity (PF-06928215, K_d = 200 nM) and inhibit enzymatic activity (Figure 2). This compound did not inhibit DNA-stimulated IFN-β production in THP-1 cells, which could be a result of poor cell permeability or lack of potency. It will be important to develop sensitive tools suitable to measure cGAMP levels in human biological samples as well as potent inhibitors that can be used to clinically test whether modulation of cGAS activity affects disease outcomes.

Figure 2. The small molecule inhibitor PF-06928215 causes a concentration-dependent inhibition of cGAS activity. Activity was monitored through displacement of a Cy5-cGAMP probe from the anti-cGAMP mAb 80-2, as described in Hall, et al. Image used under CC BY 4.0. 

Flip through the pages of the Cayman Currents issue on Innate Immune Signaling: cGAS/STING to learn more about the cGAMP ligands, 2'3'-cGAMP ELISA Kit, recombinant STING variants, and additional key proteins and antibodies Cayman Chemical offers to help in this endeavor.

References

1. Corrales, L., McWhirter, S.M., Dubensky, T.W., Jr., et al. J. Clin. Invest. 126(7), 2404-2411 (2016).

2. Chen, Q., Sun, L., and Chen, Z.J. Nat. Immunol. 17(10), 1142-1149 (2016).

3. Tao, J., Zhou, X., and Jiang, Z. IUBMB Life 68(11), 858-870 (2016).

4. Bose, D. Int. J. Mol. Sci. 18(11), 2456-2466 (2017).

5. Chow, J., Franz, K.M., and Kagan J.C. Virology 479-480, 104-109 (2015).

6. Sun, L., Wu, J., Du, F., et al. Science 339(6121), 786-791 (2013).

7. Vance, R.E. Immunity 45(2), 227-228 (2016).

8. Yi, G., Brendel, Y.P., Shu, C., et al. PLoS ONE 8(10), e77846 (2013).

9. Diner, E.J., Burdette, D.L., Wilson, S.C., et al. Cell Rep.3(5), 1355-1361 (2013).

10. Zhao, L., Ching, L-M., Kestell, P., et al. Br. J. Cancer 87(4), 465-470 (2002).

11. Lara, P.N., Jr., Douillard, J.Y., Nakagawa, K., et al. J. Clin. Oncol. 29(22), 2965-2971 (2011).

12. Gao, P., Ascano, M., Wu, Y., et al. Cell 153(5), 1094-1107 (2013).

13. Crow, Y.J., Hayward, B.E., Parmar, R., et al. Nature Genet. 38(8), 917-920 (2006).

14. Lee-Kirsch, M.A., Gong, M., Chowdhury, D., et al. Nature Genet. 39(9), 1065-1067 (2007).

15. Morita, M., Stamp, G., Robins, P., et al. Mol. Cell Biol. 24(15), 6719-6727 (2004).

16. Stetson, D.B., Ko, J.S., Heidmann, T., et al. Cell 134(4), 587-598 (2008).

17. Gray, E.E., Treuting, P.M., Woodward, J.J., et al. J. Immunol.195(5), 1939-1943 (2015).

18. An, J., Durcan, L., Karr, R.M., et al. Arthritis Rheumatol. 69(4), 800-807 (2017).

19. Liu, Y., Jesus, A.A., Marrero, B., et al. N. Engl. J. Med. 371(6), 507-518 (2014).

20. Wang, H., Hu, S., Chen, X., et al. Proc. Natl. Acad. Sci. USA114(7), 1637-1642 (2017).

21. Demaria, O., De Gassart, A., Coso, S., et al. Proc. Natl. Acad. Sci. USA112(50), 15408-15413 (2015).

22. Corrales, L., Glickman, L.H., McWhirter, S.M., et al. Cell Rep.11(7), 1018-1030 (2015).

23. Fu, J., Kanne, D.B., Leong, M., et al. Sci. Transl. Med. 7(283), 283ra52 (2015).

24. Ahn, J., Xia, T., Konno, H., et al. Nat. Commun. 5:5166 (2014).

25. Ahn, J., Konno, H., and Barber, G.N. Oncogene 34(41), 5302-5308 (2015).

26. Lemos, H., Huang, L., McGaha, T.L., et al. Eur. J. Immunol. 44(10), 2847-2853 (2014).

27. Lemos, H., Mohamed, E., Huang, L., et al. Cancer Res.76(8), 2076-2081 (2016).

28. Bose, D., Su, Y., Marcus, A., et al. Cell Chem. Biol. 23(12), 1539-1549 (2016).

29. Hall, J., Brault, A., Vincent, F., et al. PLoS ONE 12(9), e0184843 (2017).