cGAS-STING Activation in the Tumor Microenvironment and Its Role in Cancer Immunity
Stimulator of interferon (IFN) genes (STING) is a key mediator in the immune response to cytoplasmic DNA sensed by cyclic GMP-AMP (cGAMP) synthase (cGAS). After synthesis by cGAS, cGAMP acts as a second messenger activating STING in the cell harboring cytoplasmic DNA but also in adjacent cells through gap junction transfer. While the role of the cGAS-STING pathway in pathogen detection is now well established, its importance in cancer immunity has only recently started to emerge. Nonetheless, STING appears to be an essential component in the recruitment of immune cells to the tumor microenvironment, which is paramount to immune clearance of the tumor. This review presents an overview of the growing literature around the role of the cGAS-STING pathway in the tumor microenvironment, with a specific focus on the role that cancer cells may play in the direct activation of this pathway, and its amplification through cell-cell transfer of cGAMP.
KeywordscGAS cGAMP STING Interferon Connexin Cancer
Detection of pathogen-associated molecular patterns (PAMPs) by the immune system is the foundation of innate immunity and is one of the first lines of defense against infections. Viral nucleic acids and bacterial cell wall components (such as lipopolysaccharides or flagellar proteins) are PAMPs selectively detected by different types of innate immune receptors. These include membrane-bound Toll-like receptors (TLRs)—located at the surface of the cells or in the endosomes (e.g., TLR3, TLR4, and TLR7/8/9)—or cytoplasmic sensors such as retinoic acid-inducible gene I (RIG-I)-like helicases, which detect foreign RNAs [1, 2]. Activation of such innate immune sensors results in the production of antiviral and antibacterial proteins, including cytokines such as TNF-α or type I interferons (IFNs). Type I IFNs exert a strong antiviral effect through transcriptional induction of more than 2000 genes .
8.1.1 Stimulator of Interferon Genes: STING
Stimulator of interferon genes (STING, also known as MITA, ERIS, MPYS, or TMEM173) was independently discovered by four different groups in 2008 in an attempt to characterize mechanisms of DNA recognition resulting in the production of type I interferon (IFN), independently of TLR9 [4, 5, 6, 7]. STING is an ER-localized protein containing N-terminal transmembrane helices and a large C-terminal cytosolic domain. Its activation promotes signal transduction through the TBK1-IRF3 axis and production of type I IFNs [8, 9]. STING is directly involved in intracellular bacterial detection through sensing of cyclic dinucleotides (CDNs). During infection, CDNs are produced by bacteria such as Listeria monocytogenes and Mycobacterium tuberculosis [10, 11], which activate STING to promote strong immune responses to these pathogens.
8.1.2 Cyclic GMP-AMP Synthase: cGAS
Despite the initial demonstration of STING’s involvement in intracellular DNA sensing , a lack of evidence for a direct dsDNA-STING interaction prompted the community to suggest the existence of an upstream DNA sensor. Several potential cytoplasmic DNA sensors had been described (e.g., DAI, IFI16, DDX41, DNA-PK, MRE11, Sox2, and PQBP1, reviewed in ), before the discovery of cyclic GMP-AMP synthase (cGAS) by the Chen laboratory in 2013 . Following cytoplasmic DNA detection, cGAS produces an endogenous second messenger, 2′3′-cyclic GMP-AMP (2′3′-cGAMP), that binds directly to STING to promote its activation [15, 16, 17, 18, 19]. It is now clear that cGAS is an important immune receptor for DNA viruses and retroviruses. As such, its capacity to sense cytoplasmic DNA is not restricted to standard Watson-Crick DNA, and it can also sense DNA-RNA hybrids . The DNA cGAS senses can originate from various sites; in addition to bacterial and viral DNA, there is now good evidence that it senses self-DNA leaked from the nucleus  or mitochondria [2, 22].
By activating STING, DNA detection by cGAS culminates in the production of pro-inflammatory cytokines and the recruitment of the IFN response. Soon after the discovery of this pathway, the role of cGAS-STING in sterile inflammation (i.e., in the absence of any pathogen) was investigated. Mutations promoting a gain-of-function of STING have been shown to result in chilblain lupus , a rare type of cutaneous auto-inflammation. In addition, cGas and/or Sting genetic deletion was found to rescue animals from lupus-like diseases [24, 25, 26].
8.1.3 cGAS-STING in Tumor Development
Beyond its critical role in innate immunity, the cGAS-STING pathway is rapidly emerging as a critical player in the control of tumor development. Clearance of tumorigenic cells by the immune system initially relies on type I IFN production by dendritic cells (DCs) and the recruitment of CD8+ T cells, which promote the targeted death of such aberrant cells. Accordingly, chemically induced tumors develop better in mice lacking type I IFN signaling compared to their wild-type (WT) counterparts; similarly, a deficiency in type I IFN signaling results in poorer rejection of transplanted immunogenic tumors [27, 28].
Recruitment of CD8+ T cells in the tumor microenvironment is the key step in antitumor immunity and is directly dependent on type I IFN production by DCs [29, 30]. Critically, STING appears to play a unique, non-redundant role in the recruitment of CD8+ T cells to the tumor microenvironment [31, 32]. Indeed, genetic loss of a number of the most important signaling molecules known to date in innate immunity did not have any impact on CD8+ T cell infiltration in the tumor microenvironment in a syngeneic melanoma model [31, 32]. Conversely, the loss of STING abolished the spontaneous infiltration of CD8+ T cells in this melanoma model [32, 33].
8.1.4 STING Agonists as Antitumoral Adjuvants
The use of pathogen-associated molecular pattern (PAMP)-like molecules to potentiate DC activation upon radiation therapy and help tumor clearance is well established, with several clinical trials currently underway . Specifically, there has been a great deal of enthusiasm around the concept that synthetic STING ligands could be designed to facilitate tumor clearance [35, 36]. One STING ligand that showed early promise was DMXAA, a flavonoid compound comprising two phenyl rings and a heterocyclic ring. DMXAA was found to exert strong antitumoral activities in mice and to act as a direct ligand for mouse STING . However, DMXAA failed in human clinical trials  and was found not to be a direct ligand for human STING . Other research has focused on cGAMP, the second messenger that activates STING. Although cGAMP can be synthesized in vitro, its application as an adjuvant is partially limited by the fact that it is not membrane permeable. Nonetheless, cGAMP injection has been found to synergize with radiation therapy to control local and distant tumors in pancreatic cancer  and to decrease chemotherapy toxicity in colon cancer .
To remedy the limitations of natural cGAMP, various types of synthetic STING ligand based on the structures of CDNs have already been synthesized and tested in various cancer models (reviewed in ). Most of these ligands have been designed to target human STING and have the potential to be used in clinical trials, as exemplified by compound MIW815 (ADU-S100). In line with a role for STING in CD8+ T cell recruitment and considering prior evidence with DMXAA, intratumoral injection of these CDN molecules into different types of cancer (melanoma, colon, glioma, and breast carcinomas) caused rapid tumor regression and mediated lasting and systemic antigen-specific T cell immunity [37, 35, 39]. The action of STING agonists may not be limited to antitumor immunity, and they could also promote direct apoptosis of cancer cells, as suggested by observations in malignant B cell leukemia .
In addition to designing synthetic ligands, loading of otherwise non-permeable cyclic di-GMP into liposomes was found to stimulate clearance of murine melanoma . This suggests that experimentation with packaging and delivery of STING agonists may help enhance the therapeutic potential of these molecules.
8.1.5 STING and Immunotherapy
The tumor microenvironment plays a critical role in the control of CD8+ T cells to kill tumor cells. The specificity of CD8+ T cells relies on the surface expression of different receptors and ligands, creating an immune checkpoint to limit aberrant killing by CD8+ T cells. Programmed cell death 1 (PD-1) expression on CD8+ T cells inhibits their activation upon binding of its ligand, PD-L1. Expression of PD-L1 in the tumor microenvironment promotes inhibition of CD8+ T cell clearance. Consequently, strategies to block PD-L1 or its binding to CD8+ T, ultimately aimed at reigniting immune clearance of tumor cells, have shown great therapeutic potential in several types of cancers . Critically, the cGAS-STING axis appears to be very important in the capacity of PD-L1 blocking strategies to reactivate CD8+ T cells, as evidenced by cGAS- and STING-deficient animal models , and in studies using co-administration of PD-L1 with STING ligands [45, 46].
Conversely, there has been evidence that STING activation can dampen immune activity through induction of the immune checkpoint indoleamine 2,3-dioxygenase (IDO) [47, 48, 49]. Further, in a subtype of head and neck cancer where the tumor was in a repressed immune state, STING activation could not mediate recruitment of CD8+ T cells .
With these lines of evidence taken together, the role of STING in immunotherapy is context dependent, and further studies are needed to better understand the signaling pathways initiating the different outcomes of its activity. How STING is activated within tumor cells is not fully understood either and is a subject for further study, although the role of STING in immune cells during tumor immunity and tumor clearance is well documented. In this review, we will focus on the role of cGAS and STING within tumor cells. We will also address the role of cell-cell communication in STING activation within the tumor microenvironment.
8.2 Mechanisms Underlying IFN Production in Immune Cells Within the Tumor Microenvironment
8.2.1 STING Activation by Tumor DNA
The current model of IFN production and CD8+ T cell priming by tumor cells indicates a possible role for the activation of antigen-presenting cells (APCs) by DNA released from the tumor as a result of apoptosis and taken up by APC via phagocytosis. This is supported by the demonstration that DNA from the tumor could be observed in APCs where STING was activated and type I IFN produced [32, 31, 50]. In this model, sensing of apoptotic cell-derived nuclear DNA by DCs recruits the STING-IRF3 axis to promote type I IFN production, thereby enhancing the functionality of DCs in an autocrine loop. Activated DCs further mediate the activation and the clonal expansion of CD8+ T cells favoring tumor clearance .
However appealing, this model does not account for how the tumor DNA would be released from the endosome/phagolysosome of APCs to reach the cytoplasm, where cGAS is located. In fact, previous work indicates that phagocytosis of apoptotic cells does not result in IFN-I activation, due to sequestration of DNA in the phagosome . Similarly, despite tumor DNA being found in APC cells in vivo, dead tumor cells incubated with APCs did not engage an IFN response in vitro . Another limit of the phagocytosis model is that it restricts the role of tumor cells in initiating an immune response to only dying/apoptotic cells.
There is also emerging evidence that the growth and spread of tumors, such as melanoma tumors, which display very little cell death, are regulated by STING. Indeed, mouse melanoma cells transplanted in mice lacking STING were found to develop significantly more lung metastases than WT mice . Collectively, these examples indicate that the essential role of STING in tumor restriction is probably not limited to the activation of cGAS-STING in APCs upon phagocytosis of dead/apoptotic tumor cells.
8.2.2 Cytosolic DNA Activation of cGAS-STING
STING is indirectly activated by cytosolic DNA, through upstream cGAS engagement. Following cytoplasmic DNA recognition, cGAS produces cGAMP which acts as a second messenger between cGAS and STING. Critically, once synthesized, cGAMP has the capacity to be transferred between cells and mediate STING activation in adjacent cells , provided they form gap junctions with the cell making cGAMP. In homeostasis, physical restriction of DNA in the nucleus or phagosome (when the DNA is phagocytosed) ensures that self-DNA is not accessible to cytosolic cGAS. However, it is now apparent that self-DNA can escape from its original localization and activate the cGAS-STING pathway in certain contexts. To prevent aberrant cGAS recruitment, several key nucleases are at play to degrade DNA molecules escaping nuclear retention. Accordingly, defects in such DNases—e.g., TREX1 , DNase2a , or the nuclease-like SamHD1 —lead to cytoplasmic DNA accumulation and cGAS-STING activation by self-DNA. In addition, certain chemical modifications of DNA, such as 8-OHG, protect DNA from DNase degradation (such as by TREX1) thereby favoring cGAS activation [56, 25]. Decreases in genome stability or faults in DNA repair pathways can also lead to STING activation. Loss of ATM, which results in an impaired DNA damage response, increases genomic instability and causes spontaneous type I IFN production through STING activation . In line with this, a lack of RPA and/or RAD51, which normally bind to damaged DNA and sequester it in the nucleus, results in cGAS detection of leaked cytoplasmic DNA . Therefore, the cGAS-STING pathway is indirectly involved in the response activated upon DNA damage.
Genome stability is directly dependent on the capacity of the cell to stop and repair its DNA. During each cell division, multiple mutations are made but are also constantly being repaired. Oncogenes and tumor suppressors regulate cell cycle and proliferation. In cancer, these genes are often mutated to cause sustained proliferation. This constant pressure to proliferate leads to genomic instability by forcing the cells to divide even in the presence of damaged DNA . Accordingly, cytosolic DNA can be observed in B cell lymphomas, but not in normal B cells . Despite being predominantly produced by infiltrating immune cells, some growing tumors can also produce type I IFNs . Such type I IFN production can directly relate to cGAS-STING activation as recently suggested using a model of breast cancer with genomic instability . Although a detailed understanding about how nuclear DNA can be leaked to the cytoplasm remains elusive, there is data to suggest that such leakage results from an active process. Overexpression of RNaseH1 (that degrades the R-loops of DNA-RNA hybrids) reduces the level of cytosolic DNA and type I IFN production and hampers the rejection of the lymphoma tumors . Conversely, in prostate cancer cells, the MUS81 endonuclease cleaves DNA at stalled replication forks, promoting export of DNA products into the cytoplasm . This results in type I IFN production through STING activation by the cancer cells, which consequently enhances the rejection of the tumor in vivo . All together, these data strongly suggest that the cGAS-STING can be activated in select tumor cells, independently of APC phagocytosis.
8.2.3 cGAS-STING Activation in Tumor Cells
To date, most reports addressing the in vivo role of the cGAS-STING pathway in the tumor environment have relied on the use of Sting −/− tumor-bearing mice compared to WT mice [33, 32, 35]. These models, although supporting a critical role for STING in the tumor microenvironment, are unable to define the source of STING activation in the tumor itself. Given that about 50% of tumor cell lines express cGAS (e.g., 5/11 colorectal cancer cell lines  and 7/11 melanoma cell lines ), cGAMP production by tumors is likely to be a frequent occurrence, assuming they also have defective DNA repair capacity. Whether such tumor-derived cGAMP activates type I IFN production by cancer cells is dependent on the presence of STING [64, 65], but this cGAMP can also modify the tumor microenvironment through horizontal transfer of the second messenger to adjacent cells .
In support of the origin of non-hematopoietic-derived cGAMP, it should be noted that Trex1 −/− mice initially produce type I IFN in their non-hematopoietic compartment. This primary induction mediates recruitment of inflammatory cells that produce even more type–I IFN . Importantly, cGAMP may also potentiate local type I IFN production through nonimmune adjacent tissues. As such, in an in vivo model of melanoma engraftment, cGAMP passage from the tumor to the vasculature may explain the observation that the primary source of IFN was endothelial cells .
It is tempting to speculate that cGAS activation in tumor cells could explain why some patients develop spontaneous leukocyte infiltration and antitumor T cell responses. Such spontaneous T cell tumor infiltration has, for instance, been reported in melanoma  and ovarian , breast [69, 61], and colorectal cancers . Whether such infiltration depends on cGAS activation in the tumor and how such activation would take place are not currently defined. Future studies investigating the correlation between tumor-derived cGAMP and leukocyte infiltration may help refine disease prognosis—leukocyte infiltration is already a powerful prognostic factor in colorectal cancer patients —while being informative about the best therapeutic approach to be selected.
8.2.4 Loss of cGAS-STING Expression in Tumor Cells
Many bacteria and viruses have evolved to block innate immune pathways, thereby facilitating their intracellular survival. Similar selection pressure constantly operates on tumor cells, which attempt to evade clearance by the immune system. Type I IFN is crucial to immune cell recruitment into the tumor microenvironment, and STING is a key factor in such type I IFN production, as discussed previously. One could speculate that selective pressure operates to block the cGAS-STING pathway in tumor cells to facilitate immune evasion. Accordingly, cGAS-STING inhibition has been recently described in human cancers.
The first extensive study on the loss of the cGAS-STING pathway comes from the field of colorectal adenocarcinoma (CA). In that research, cGAS-STING activity was decreased in the vast majority of CA cell lines, which lacked the capacity to produce type-I IFN in response to cytoplasmic DNA . In some cell lines such as HT29 cells, the pathway was altered but still functional . In support of this in vitro data, one third of 48 clinical samples of adenocarcinoma analyzed showed a loss of cGAS expression . Interestingly, upregulation of cGAS expression in the early stages of cancer and disruption of the STING pathway in advanced stages were reported in a similar study . These data indicate that engagement of the pathway during early stages of tumor development, but inhibition of the pathway later in disease, favored tumor growth. Similarly to CA, melanoma cells displayed recurrent loss of cGAS-STING expression, ultimately inhibiting type I IFN production . Epigenetic repression was found to relate to the cGAS and STING inhibition proposed in these studies [64, 65], but other factors are most likely at play. In another study, ovarian cancer cells (serous, clear cell, and endometriosis) had lost responsiveness to DNA transfection via STING-IRF3 activation . These reports collectively suggest that loss of expression or activity of the cGAS-STING pathway may favor the development of the tumor cells.
Therapeutically, such a loss of the cGAS-STING response can be harnessed to obtain clinical benefit. Indeed, cGAS-STING-deficient tumors are more susceptible to viral oncolysis—where modified DNA viruses like HSV-1 have been used to target and kill cancer cells [65, 64].
8.2.5 cGAS-STING in Pathogen-Driven Carcinogenesis
Whether the cGAS-STING pathway is involved in pathogen-driven carcinogenesis is a question that remains elusive. Human papillomavirus (HPV) is the causative agent of cervical cancer and other types of cancer. HPV is a DNA virus, making it a potential target of cGAS sensing . Accordingly, and although not detailed in the current literature, the cGAS-STING pathway may play a role in HPV infection and cancer development. In line with this hypothesis, a single nucleotide polymorphism (SNP) in cGAS (rs311678) has been recently associated with a reduced risk of cervical precancerous lesions. This SNP was found to modulate cGAS expression in vitro, and higher cGAS was associated with a reduced risk of HPV infection . As described for other viruses, HPV can also counteract the activity of the innate immune system. In vitro expression of the HPV E2 protein in human primary keratinocytes downregulates the expression of STING and several other innate immune genes . In addition, STING repression has been shown in HPV+ low-grade squamous intraepithelial lesions, when compared with HPV− controls. Furthermore, viral oncogenes such as HPV E7 and E1A of human adenovirus A5 were proposed to bind to the N-terminal region of STING to reduce its downstream signaling .
In line with a role for the cGAS-STING pathway in reducing the risk of early stages of pathogen-driven cancer, STING expression is significantly decreased in gastric cancers when compared to non-tumor tissues . Helicobacter pylori, the main causative agent of gastric cancer, can activate STING and promote inflammation . These data collectively suggest a role for cGAS and STING in pathogen-mediated carcinogenesis.
8.3 Connexin Expression in Tumor Cells and Its Impact on Tumor Development
Beyond the inhibition of tumor-derived type I IFN production by cGAS, one could argue that loss of cGAS expression by a significant number of tumor cells is also important for stopping the propagation of cGAMP within the tumor microenvironment. As mentioned previously, cGAMP transfers horizontally through gap junctions to activate STING in adjacent cells . With this in mind, it is tempting to revisit previous works on the role of gap junctions in tumorigenesis.
8.3.1 Connexins and Gap Junctions
Gap junctions are formed through the interaction of connexins from both interacting cells to promote the intercellular circulation of ions and small molecules such as cGAMP. Among the family of connexins, connexin (CX) 43 and CX45 were found to be essential in human embryonic kidney 293T (HEK293T) cells for cGAMP horizontal transfer . Critically, transfer of cGAMP to HEK293T CX43/45-deficient cells could be restored through the expression of human CX26, CX31, CX32, CX40, CX43, and CX62 and mouse CX43 and CX45, suggesting that most connexins are able to transfer cGAMP (noting that human CX50 overexpression did not restore transfer) .
8.3.2 Loss of Intercellular Cell Communication in Early-Stage Tumors
Loss of intercellular communication by cancer cells was first described over half a century ago . Since their discovery, connexins have been shown to exert both pro- and antitumoral activities, making it a controversial field of research . The overall view is that retention of connexin expression benefits antitumoral activities at early stages but can later favor metastasis. Arguably, loss of connexin expression early in tumor development would be expected to reduce cGAMP horizontal transfer to adjacent cells and inhibit type I IFN induction—thereby favoring tumor initiation and immune evasion. Accordingly, CX43 expression is decreased in prostate cancer patients compared to controls. In line with what is observed with the loss of cGAS, the reduction of CX43 expression was found to correlate with advanced stages of cancer . Similar trends were observed in breast cancers and head and neck squamous cell carcinomas (HNSCC), where low expression of CX43 correlated with a negative prognosis . Given that HPV is also often detected in HNSCC tumors , it is of interest to note that HPV-E6 protein expression was associated with a reduction of gap junction formation .
Further supporting an antitumoral effect of connexins in early tumor growth, overexpression of CX43 was found to reduce melanoma tumor growth in vivo , while its downregulation stimulated the growth of prostate cancer cells . Re-expression of connexins in breast cancer cell lines implanted in vivo also reduced tumor growth . Qigesan, a molecule that increases the expression of connexins, reduced cell migration and invasion in esophageal cancer cells . Critically, in a model of chemically induced mammary tumors, enforced expression of connexin decreased the incidence of tumor formation, while no difference was observed on the tumor growth . These findings suggest that connexins may play a greater role in tumor initiation than tumor development . It should, however, be noted that the latter tumor model relied on tumoral expression of the Cre recombinase—which we linked recently to cGAS-STING engagement —opening the possibility for a direct role for the cGAS-STING pathway in the initiation of tumors in this model.
8.3.3 The Role of Connexin in Metastasis
Conversely, there are instances where increased expression of connexins was found to enhance metastasis. Indeed, in a mouse melanoma model, the metastatic capacity of the cancer cells was found to be dependent on CX26 expression . Furthermore, CX43 expression is induced in CA cell lines that display greater metastatic potential, while CX43 levels are almost absent in other tumor cell lines .
Critically, direct evidence for the role of cGAMP and gap junction in metastasis was recently reported. When breast and lung cancer metastatic cells migrated to the brain, they increased connexin expression, allowing passage of tumor cell-derived cGAMP to adjacent astrocytes . cGAMP transfer to astrocytes promoted astrocyte activation and the subsequent secretion of pro-inflammatory cytokines favoring tumor growth and chemotherapeutic resistance .
8.4 cGAS-cGAMP: Connexins-STING in Chemotherapy
The role played by intercellular cell communication during chemotherapy is also controversial: there are reports of both positive and negative roles for connexins on the outcome of chemotherapy. For example, increased expression of connexins enhanced the sensitivity of RKO colon cancer cells to diverse chemotherapeutic agents such as fluorouracil, oxaliplatin, and irinotecan . Critically, increased sensitivity was observed in vitro when the cells were more confluent or when they were treated with retinoic acid, which induces the expression of connexins . In addition, low expression of connexins correlated with reduced sensitivity of hepatocarcinoma cells to oxaliplatin . On the other hand, inhibition of connexins was reported to sensitize glioblastoma cells previously shown to be resistant to chemotherapy .
Chemotherapy, which mostly relies on the greater sensitivity of tumor cells to DNA damage, often induces type–I IFN production . We recently discovered that the DNA intercalating agent acriflavine could promote cGAMP synthesis in SV40T immortalized mouse embryonic fibroblasts (MEFs), in association with increased cytoplasmic DNA levels . Despite previous evidence of cGAS and STING activation following DNA damage and cytoplasmic DNA leakage [ 56, 57, 98], there had been no prior demonstration of the direct engagement of cGAS and cGAMP production in these contexts. Critically, the capacity of the topoisomerase I inhibitor, topotecan (TPT), to restrict breast cancer cell proliferation in vivo was abrogated in mice lacking STING . This, however, does not tease out whether cGAMP was generated by the tumor after chemotherapy treatment or by phagocytes. Similar results were found using irradiation of tumors . Interestingly, single-stranded DNA (ssDNA) leaking in the cytoplasm of the tumor cells following chemotherapy was shown to be a by-product of BLOOM syndrome helicase (BLM) and the exonuclease-RNase EXO1 . Along with the discovery that RPA and RAD51 normally work to retain ssDNA generated by BLM and EXO1 during DNA repair , these findings suggest that cytoplasmic DNA leakage is likely related to a saturation of the cell’s capacity to retain it in the nucleus . Although ssDNA is not supposed to activate cGAS, it can bind to it weakly, and it is likely that its modification upon DNA damage (such as by 8-OHG) somewhat favors cGAS activation [56, 101]. Further work is clearly warranted to better define the modalities of cytoplasmic DNA leakage and cGAS engagement upon DNA damage by chemotherapy.
8.5 DNA Damage Engages cGAS Activity and Horizontal STING Amplification
We have previously demonstrated that DNA damage can mediate cGAMP production [90, 21]. When using inducible Cre recombinase-mediated DNA damage, we observed that only a proportion of cells displayed the hallmarks of DNA damage (through γH2AX staining) . Critically, we demonstrated that the capacity of damaged cells to generate a widespread type I IFN production in the cell monolayer was strongly dependent on cell density—which we attribute to a connexin-dependent transfer of cGAMP . This illustrates the capacity of healthy adjacent cells to amplify the signal of selected damaged cells and suggests that a similar feedback loop could operate in the tumor microenvironment upon induction of DNA damage by chemotherapy. Surprisingly, cGAS-depleted bone marrow-derived DCs can be activated after co-incubation with irradiated tumor cells, albeit modestly, while STING depletion completely thwarted the effect. This residual activation in cGAS-deficient cells suggests that engagement of cGAS in phagocytes is not essential for STING activation by irradiated tumor cells . From this point of view, the capacity of tumors cells to generate cGAMP and transfer it to adjacent cells may play an important role in the outcome of chemotherapy.
In summary, the cGAS-STING axis is crucial to cancer immunity. There is accumulating evidence for a role for STING in tumor DCs and cross priming of CD8+ T cells. These findings clearly suggest that STING ligands have strong therapeutic potential. Nonetheless, therapeutic STING activation may also contribute to the tumor expressing interferon-stimulated genes previously linked with chemoresistance [97, 101, 102]. As such, critical questions regarding the modalities of activation of the pathway in the tumor microenvironment and its impact on chemotherapies are still to be answered. For instance, while DCs have a central role in mediating the recruitment of an antitumor immune response through STING, how other cell types like macrophages and neutrophils contribute to this pathway should also be addressed. Critically, the source of cGAMP activating STING in APCs and its putative modalities of transfer to APCs are not defined. In light of the current literature reviewed herein and our own experiments, we propose a model, illustrated in Fig. 8.1, in which cGAMP can be synthesized by tumor cells to play a role in immune cell activation through immune gap junctions. While this is supported by the demonstration that metastatic breast cancer cells could transfer cGAMP to astrocytes in the brain , direct evidence of cGAMP production by tumor cells and its transfer to APCs remains to be found. Defining if and how cGAMP can be made by tumors has the potential to help understand why select tumors are devoid of infiltrating immune cells. Given that the efficacy of preferred chemotherapies has been associated with STING signaling in vivo (e.g., irinotecan), these may be particularly effective when used in patients with active cGAS and functional connexin tumors.
We thank Rebecca Smith, Jonathan Ferrand, and Lise Boursinhac for the assistance in the preparation of the manuscript. This work was funded in part by the Australian NHMRC (1062683 and 1081167 to M.P.G.), the Australian Research Council (140100594 Future Fellowship to M.P.G.), and the Victorian Government’s Operational Infrastructure Support Program. G.P. is a fellow from Fonds de Recherche du Québec – Santé (FRQS), Canada (35071).
- 5.Sun W, Li Y, Chen L, Chen H, You F, Zhou X, Zhou Y, Zhai Z, Chen D, Jiang Z (2009) ERIS, an endoplasmic reticulum IFN stimulator, activates innate immune signaling through dimerization. Proc Natl Acad Sci U S A 106(21):8653–8658. doi: 10.1073/pnas.0900850106 CrossRefPubMedPubMedCentralGoogle Scholar
- 7.Jin L, Waterman PM, Jonscher KR, Short CM, Reisdorph NA, Cambier JC (2008) MPYS, a novel membrane tetraspanner, is associated with major histocompatibility complex class II and mediates transduction of apoptotic signals. Mol Cell Biol 28(16):5014–5026. doi: 10.1128/MCB.00640-08 CrossRefPubMedPubMedCentralGoogle Scholar
- 16.Gao P, Ascano M, Wu Y, Barchet W, Gaffney BL, Zillinger T, Serganov AA, Liu Y, Jones RA, Hartmann G, Tuschl T, Patel DJ (2013) Cyclic [G(2′,5′)pA(3′,5′)p] is the metazoan second messenger produced by DNA-activated cyclic GMP-AMP synthase. Cell 153(5):1094–1107. doi: 10.1016/j.cell.2013.04.046 CrossRefPubMedPubMedCentralGoogle Scholar
- 18.Diner EJ, Burdette DL, Wilson SC, Monroe KM, Kellenberger CA, Hyodo M, Hayakawa Y, Hammond MC, Vance RE (2013) The innate immune DNA sensor cGAS produces a noncanonical cyclic dinucleotide that activates human STING. Cell Rep 3(5):1355–1361. doi: 10.1016/j.celrep.2013.05.009 CrossRefPubMedPubMedCentralGoogle Scholar
- 22.White MJ, McArthur K, Metcalf D, Lane RM, Cambier JC, Herold MJ, van Delft MF, Bedoui S, Lessene G, Ritchie ME, Huang DC, Kile BT (2014) Apoptotic caspases suppress mtDNA-induced STING-mediated type I IFN production. Cell 159(7):1549–1562. doi: 10.1016/j.cell.2014.11.036 CrossRefPubMedPubMedCentralGoogle Scholar
- 23.Konig N, Fiehn C, Wolf C, Schuster M, Cura Costa E, Tungler V, Alvarez HA, Chara O, Engel K, Goldbach-Mansky R, Gunther C, Lee-Kirsch MA (2017) Familial chilblain lupus due to a gain-of-function mutation in STING. Ann Rheum Dis 76(2):468–472. doi: 10.1136/annrheumdis-2016-209841 CrossRefPubMedGoogle Scholar
- 24.Mackenzie KJ, Carroll P, Lettice L, Tarnauskaite Z, Reddy K, Dix F, Revuelta A, Abbondati E, Rigby RE, Rabe B, Kilanowski F, Grimes G, Fluteau A, Devenney PS, Hill RE, Reijns MA, Jackson AP (2016) Ribonuclease H2 mutations induce a cGAS/STING-dependent innate immune response. EMBO J 35(8):831–844. doi: 10.15252/embj.201593339 CrossRefPubMedPubMedCentralGoogle Scholar
- 29.Diamond MS, Kinder M, Matsushita H, Mashayekhi M, Dunn GP, Archambault JM, Lee H, Arthur CD, White JM, Kalinke U, Murphy KM, Schreiber RD (2011) Type I interferon is selectively required by dendritic cells for immune rejection of tumors. J Exp Med 208(10):1989–2003. doi: 10.1084/jem.20101158 CrossRefPubMedPubMedCentralGoogle Scholar
- 32.Woo SR, Fuertes MB, Corrales L, Spranger S, Furdyna MJ, Leung MY, Duggan R, Wang Y, Barber GN, Fitzgerald KA, Alegre ML, Gajewski TF (2014) STING-dependent cytosolic DNA sensing mediates innate immune recognition of immunogenic tumors. Immunity 41(5):830–842. doi: 10.1016/j.immuni.2014.10.017 CrossRefPubMedPubMedCentralGoogle Scholar
- 33.Demaria O, De Gassart A, Coso S, Gestermann N, Di Domizio J, Flatz L, Gaide O, Michielin O, Hwu P, Petrova TV, Martinon F, Modlin RL, Speiser DE, Gilliet M (2015) STING activation of tumor endothelial cells initiates spontaneous and therapeutic antitumor immunity. Proc Natl Acad Sci U S A 112(50):15408–15413. doi: 10.1073/pnas.1512832112 CrossRefPubMedPubMedCentralGoogle Scholar
- 35.Ohkuri T, Ghosh A, Kosaka A, Zhu J, Ikeura M, David M, Watkins SC, Sarkar SN, Okada H (2014) STING contributes to antiglioma immunity via triggering type I IFN signals in the tumor microenvironment. Cancer Immunol Res 2(12):1199–1208. doi: 10.1158/2326-6066.CIR-14-0099 CrossRefPubMedPubMedCentralGoogle Scholar
- 37.Corrales L, Glickman LH, McWhirter SM, Kanne DB, Sivick KE, Katibah GE, Woo SR, Lemmens E, Banda T, Leong JJ, Metchette K, Dubensky TW Jr, Gajewski TF (2015) Direct activation of STING in the tumor microenvironment leads to potent and systemic tumor regression and immunity. Cell Rep 11(7):1018–1030. doi: 10.1016/j.celrep.2015.04.031 CrossRefPubMedPubMedCentralGoogle Scholar
- 38.Lara PN Jr, Douillard JY, Nakagawa K, von Pawel J, McKeage MJ, Albert I, Losonczy G, Reck M, Heo DS, Fan X, Fandi A, Scagliotti G (2011) Randomized phase III placebo-controlled trial of carboplatin and paclitaxel with or without the vascular disrupting agent vadimezan (ASA404) in advanced non-small-cell lung cancer. J Clin Oncol 29(22):2965–2971. doi: 10.1200/JCO.2011.35.0660. Epub 2011 Jun 27.CrossRefPubMedGoogle Scholar
- 40.Baird JR, Friedman D, Cottam B, Dubensky TW Jr, Kanne DB, Bambina S, Bahjat K, Crittenden MR, Gough MJ (2016) Radiotherapy combined with novel STING-targeting oligonucleotides results in regression of established tumors. Cancer Res 76(1):50–61 https://doi.org/10.1158/0008-5472.CAN-14-3619 CrossRefPubMedGoogle Scholar
- 45.Moore E, Clavijo PE, Davis R, Cash H, Van Waes C, Kim Y, Allen C (2016) Established T cell-inflamed tumors rejected after adaptive resistance was reversed by combination STING activation and PD-1 pathway blockade. Cancer Immunol Res 4(12):1061–1071. doi: 10.1158/2326-6066.CIR-16-0104 CrossRefPubMedPubMedCentralGoogle Scholar
- 46.Fu J, Kanne DB, Leong M, Glickman LH, McWhirter SM, Lemmens E, Mechette K, Leong JJ, Lauer P, Liu W, Sivick KE, Zeng Q, Soares KC, Zheng L, Portnoy DA, Woodward JJ, Pardoll DM, Dubensky TW Jr, Kim Y (2015) STING agonist formulated cancer vaccines can cure established tumors resistant to PD-1 blockade. Sci Transl Med 7(283):283ra252. doi: 10.1126/scitranslmed.aaa4306 CrossRefGoogle Scholar
- 48.Huang L, Li L, Lemos H, Chandler PR, Pacholczyk G, Baban B, Barber GN, Hayakawa Y, McGaha TL, Ravishankar B, Munn DH, Mellor AL (2013) Cutting edge: DNA sensing via the STING adaptor in myeloid dendritic cells induces potent tolerogenic responses. J Immunol 191(7):3509–3513. doi: 10.4049/jimmunol.1301419 CrossRefPubMedPubMedCentralGoogle Scholar
- 49.Lemos H, Huang L, Chandler PR, Mohamed E, Souza GR, Li L, Pacholczyk G, Barber GN, Hayakawa Y, Munn DH, Mellor AL (2014) Activation of the STING adaptor attenuates experimental autoimmune encephalitis. J Immunol 192(12):5571–5578. doi: 10.4049/jimmunol.1303258 CrossRefPubMedPubMedCentralGoogle Scholar
- 55.Kretschmer S, Wolf C, Konig N, Staroske W, Guck J, Hausler M, Luksch H, Nguyen LA, Kim B, Alexopoulou D, Dahl A, Rapp A, Cardoso MC, Shevchenko A, Lee-Kirsch MA (2015) SAMHD1 prevents autoimmunity by maintaining genome stability. Ann Rheum Dis 74(3):e17. doi: 10.1136/annrheumdis-2013-204845 CrossRefPubMedGoogle Scholar
- 56.Gehrke N, Mertens C, Zillinger T, Wenzel J, Bald T, Zahn S, Tuting T, Hartmann G, Barchet W (2013) Oxidative damage of DNA confers resistance to cytosolic nuclease TREX1 degradation and potentiates STING-dependent immune sensing. Immunity 39(3):482–495. doi: 10.1016/j.immuni.2013.08.004 CrossRefPubMedGoogle Scholar
- 57.Hartlova A, Erttmann SF, Raffi FA, Schmalz AM, Resch U, Anugula S, Lienenklaus S, Nilsson LM, Kroger A, Nilsson JA, Ek T, Weiss S, Gekara NO (2015) DNA damage primes the type I interferon system via the cytosolic DNA sensor STING to promote anti-microbial innate immunity. Immunity 42(2):332–343. doi: 10.1016/j.immuni.2015.01.012 CrossRefPubMedGoogle Scholar
- 58.Wolf C, Rapp A, Berndt N, Staroske W, Schuster M, Dobrick-Mattheuer M, Kretschmer S, Konig N, Kurth T, Wieczorek D, Kast K, Cardoso MC, Gunther C, Lee-Kirsch MA (2016) RPA and Rad51 constitute a cell intrinsic mechanism to protect the cytosol from self DNA. Nat Commun 7:11752. doi: 10.1038/ncomms11752 CrossRefPubMedPubMedCentralGoogle Scholar
- 60.Lam AR, Le Bert N, Ho SS, Shen YJ, Tang ML, Xiong GM, Croxford JL, Koo CX, Ishii KJ, Akira S, Raulet DH, Gasser S (2014) RAE1 ligands for the NKG2D receptor are regulated by STING-dependent DNA sensor pathways in lymphoma. Cancer Res 74(8):2193–2203. doi: 10.1158/0008-5472.CAN-13-1703 CrossRefPubMedPubMedCentralGoogle Scholar
- 61.Parkes EE, Walker SM, Taggart LE, McCabe N, Knight LA, Wilkinson R, McCloskey KD, Buckley NE, Savage KI, Salto-Tellez M, McQuaid S, Harte MT, Mullan PB, Harkin DP, Kennedy RD (2017) Activation of STING-dependent innate immune signaling By S-phase-specific DNA damage in breast cancer. J Natl Cancer Inst 109(1). doi: 10.1093/jnci/djw199 CrossRefPubMedPubMedCentralGoogle Scholar
- 63.Ho SS, Zhang WY, Tan NY, Khatoo M, Suter MA, Tripathi S, Cheung FS, Lim WK, Tan PH, Ngeow J, Gasser S (2016) The DNA structure-specific endonuclease MUS81 mediates DNA sensor STING-dependent host rejection of prostate cancer cells. Immunity 44(5):1177–1189. doi: 10.1016/j.immuni.2016.04.010 CrossRefPubMedGoogle Scholar
- 66.Gall A, Treuting P, Elkon KB, Loo YM, Gale M Jr, Barber GN, Stetson DB (2012) Autoimmunity initiates in nonhematopoietic cells and progresses via lymphocytes in an interferon-dependent autoimmune disease. Immunity 36(1):120–131. doi: 10.1016/j.immuni.2011.11.018 CrossRefPubMedPubMedCentralGoogle Scholar
- 70.Galon J, Costes A, Sanchez-Cabo F, Kirilovsky A, Mlecnik B, Lagorce-Pages C, Tosolini M, Camus M, Berger A, Wind P, Zinzindohoue F, Bruneval P, Cugnenc PH, Trajanoski Z, Fridman WH, Pages F (2006) Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science 313(5795):1960–1964. doi: 10.1126/science.1129139 CrossRefPubMedGoogle Scholar
- 71.Pages F, Kirilovsky A, Mlecnik B, Asslaber M, Tosolini M, Bindea G, Lagorce C, Wind P, Marliot F, Bruneval P, Zatloukal K, Trajanoski Z, Berger A, Fridman WH, Galon J (2009) In situ cytotoxic and memory T cells predict outcome in patients with early-stage colorectal cancer. J Clin Oncol 27(35):5944–5951. doi: 10.1200/JCO.2008.19.6147 CrossRefPubMedGoogle Scholar
- 75.Xiao D, Huang W, Ou M, Guo C, Ye X, Liu Y, Wang M, Zhang B, Zhang N, Huang S, Zang J, Zhou Z, Wen Z, Zeng C, Wu C, Huang C, Wei X, Yang G, Jing C (2016) Interaction between susceptibility loci in cGAS-STING pathway, MHC gene and HPV infection on the risk of cervical precancerous lesions in Chinese population. Oncotarget. doi: 10.18632/oncotarget.12399
- 76.Sunthamala N, Thierry F, Teissier S, Pientong C, Kongyingyoes B, Tangsiriwatthana T, Sangkomkamhang U, Ekalaksananan T (2014) E2 proteins of high risk human papillomaviruses down-modulate STING and IFN-kappa transcription in keratinocytes. PLoS One 9(3):e91473. doi: 10.1371/journal.pone.0091473 CrossRefPubMedPubMedCentralGoogle Scholar
- 78.Song S, Peng P, Tang Z, Zhao J, Wu W, Li H, Shao M, Li L, Yang C, Duan F, Zhang M, Zhang J, Wu H, Li C, Wang X, Wang H, Ruan Y, Gu J (2017) Decreased expression of STING predicts poor prognosis in patients with gastric cancer. Sci Rep 7:39858. doi: 10.1038/srep39858 CrossRefPubMedPubMedCentralGoogle Scholar
- 81.Xu N, Chen HJ, Chen SH, Xue XY, Chen H, Zheng QS, Wei Y, Li XD, Huang JB, Cai H, Sun XL (2016) Reduced Connexin 43 expression is associated with tumor malignant behaviors and biochemical recurrence-free survival of prostate cancer. Oncotarget 7(41):67476–67484. doi: 10.18632/oncotarget.11231 CrossRefPubMedPubMedCentralGoogle Scholar
- 82.Danos K, Brauswetter D, Birtalan E, Pato A, Bencsik G, Krenacs T, Petak I, Tamas L (2016) The potential prognostic value of Connexin 43 expression in head and neck squamous cell carcinomas. Appl Immunohistochem Mol Morphol 24(7):476–481. doi: 10.1097/PAI.0000000000000212 CrossRefPubMedGoogle Scholar
- 85.Tittarelli A, Guerrero I, Tempio F, Gleisner MA, Avalos I, Sabanegh S, Ortiz C, Michea L, Lopez MN, Mendoza-Naranjo A, Salazar-Onfray F (2016) Overexpression of Connexin 43 reduces melanoma proliferative and metastatic capacity. Br J Cancer 115(9):e14. doi: 10.1038/bjc.2016.296 CrossRefPubMedPubMedCentralGoogle Scholar
- 87.Qin H, Shao Q, Curtis H, Galipeau J, Belliveau DJ, Wang T, Alaoui-Jamali MA, Laird DW (2002) Retroviral delivery of Connexin genes to human breast tumor cells inhibits in vivo tumor growth by a mechanism that is independent of significant gap junctional intercellular communication. J Biol Chem 277(32):29132–29138. doi: 10.1074/jbc.M200797200 CrossRefPubMedGoogle Scholar
- 90.Pepin G, Ferrand J, Honing K, Jayasekara WS, Cain JE, Behlke MA, Gough DJ, Williams BRG, Hornung V, Gantier MP (2016) Cre-dependent DNA recombination activates a STING-dependent innate immune response. Nucleic Acids Res 44(11):5356–5364. doi: 10.1093/nar/gkw405 CrossRefPubMedPubMedCentralGoogle Scholar
- 91.Ito A, Katoh F, Kataoka TR, Okada M, Tsubota N, Asada H, Yoshikawa K, Maeda S, Kitamura Y, Yamasaki H, Nojima H (2000) A role for heterologous gap junctions between melanoma and endothelial cells in metastasis. J Clin Invest 105(9):1189–1197. doi: 10.1172/JCI8257 CrossRefPubMedPubMedCentralGoogle Scholar
- 92.Zhang A, Hitomi M, Bar-Shain N, Dalimov Z, Ellis L, Velpula KK, Fraizer GC, Gourdie RG, Lathia JD (2015) Connexin 43 expression is associated with increased malignancy in prostate cancer cell lines and functions to promote migration. Oncotarget 6(13):11640–11651. doi: 10.18632/oncotarget.3449 CrossRefPubMedPubMedCentralGoogle Scholar
- 96.Murphy SF, Varghese RT, Lamouille S, Guo S, Pridham KJ, Kanabur P, Osimani AM, Sharma S, Jourdan J, Rodgers CM, Simonds GR, Gourdie RG, Sheng Z (2016) Connexin 43 inhibition sensitizes chemoresistant glioblastoma cells to temozolomide. Cancer Res 76(1):139–149. doi: 10.1158/0008-5472.CAN-15-1286 CrossRefPubMedGoogle Scholar
- 99.Kitai Y, Kawasaki T, Sueyoshi T, Kobiyama K, Ishii KJ, Zou J, Akira S, Matsuda T, Kawai T (2017) DNA-containing exosomes derived from cancer cells treated with topotecan activate a STING-dependent pathway and reinforce antitumor immunity. J Immunol 198(4):1649–1659. doi: 10.4049/jimmunol.1601694 CrossRefPubMedGoogle Scholar
- 100.Deng L, Liang H, Xu M, Yang X, Burnette B, Arina A, Li XD, Mauceri H, Beckett M, Darga T, Huang X, Gajewski TF, Chen ZJ, Fu YX, Weichselbaum RR (2014) STING-dependent cytosolic DNA sensing promotes radiation-induced type I interferon-dependent antitumor immunity in immunogenic tumors. Immunity 41(5):843–852. doi: 10.1016/j.immuni.2014.10.019 CrossRefPubMedPubMedCentralGoogle Scholar
- 102.Gaston J, Cheradame L, Yvonnet V, Deas O, Poupon MF, Judde JG, Cairo S, Goffin V (2016) Intracellular STING inactivation sensitizes breast cancer cells to genotoxic agents. Oncotarget. doi: 10.18632/oncotarget.12858