Human DNA-PK activates a STING-independent DNA sensing pathway

Katelyn Burleigh, Joanna H. Maltbaek, Stephanie Cambier, Richard Green, Michael Gale Jr., Richard C. James, Daniel B. Stetson

Preprint posted on 23 March 2019

Article now published in Science Immunology at

Taking the STING out of sensing cytosolic DNA: A new pathway in humans mediated by the protein DNA-PK

Selected by Connor Rosen

Categories: immunology, microbiology


The presence of cytosolic DNA is a key danger signal in immunity, triggering a global anti-viral state through the activation of type I interferons. The canonical DNA-sensing pathway in mammalian cells is the cGAS-STING pathway. Recognition of DNA by cGAS leads to production of the second messenger cGAMP, which activates STING, leading to subsequent activation of the transcription factor IRF3 and type I interferon production. While multiple regulators of the process or alternate sensors have been proposed, all proposed DNA-sensing pathways to date converge on STING. This preprint by Burleigh et al uncovers a STING-Independent DNA Sensing Pathway (SIDSP) that exists in human cells.


Key Findings:

  • A STING-independent DNA sensing pathway exists in human, but not mouse, cells.

The authors began by examining the activity of the adenovirus E1a viral oncogene, which has STING-antagonist activity. However, the authors demonstrate that E1a also suppresses responses to DNA in STING-deficient cells, suggesting a second mechanism of DNA sensing also antagonized by E1a. They further show that STING-deficient human, but not mouse, cells respond to DNA with robust type I interferon induction, in an IRF3-dependent manner.

  • The SIDSP requires DNA-PK activity.

The previous experiments established that damaged DNA (i.e. with ends exposed) triggered the SIDSP. This suggested a potential role for DNA-PK, a sensor of DNA double-stranded breaks involved in DNA repair. Pharmacological and genetic blockade of DNA-PK established that DNA-PK is required for the SIDSP. This enabled further studies of the gene program induced by the SIDSP, using antagonists of DNA-PK.

  • DNA-PK phosphorylates HSPA8.

The authors identified that the SIDSP, but not STING or RIG-I signaling, also results in the phosphorylation of the heat shock protein HSPA8. The C-terminus of HSPA8 has a sequence matching the IRF3 phosphorylation site, and seems to be phosphorylated at the corresponding residues. The phosphorylation of HSPA8 by DNA-PK also matched the species-specificity of the SIDSP, as mouse DNA-PK was unable to phosphorylate either mouse or human HSPA8 while human DNA-PK could phosphorylate either HSPA8. This suggests that HSPA8 phosphorylation may be a component of the SIDSP, or minimally indicative of conditions under which the SIDSP is functional.



This study convincingly demonstrates the existence of the SIDSP in human cells and sets multiple paths for future investigation, through the delineation of an effector, target, and inhibitors of the pathway. Additionally, both agonists and antagonists of STING are under preclinical or clinical evaluation for tumor immunotherapy and treatment of inflammatory diseases, respectively, and the discovery of complementary pathways existing in humans but not mice may be critical for fully effective targeting of DNA sensing.


Moving Forward:

  • This work opens up a wide field of investigation, including many questions the authors bring up. Some particularly interesting avenues will be the regulation of self/non-self discrimination by DNA-PK, the intermediate proteins between DNA-PK and IRF3 phosphorylation (if any), and the importance of HSPA8 phosphorylation both in the SIDSP and more broadly in cellular physiology.
  • The SIDSP has slower kinetics than the STING-dependent pathway, as shown globally through the mRNA-seq experiment. This may suggest possible signaling steps for the SIDSP – changes such as degradation of inhibitory checkpoint, new translation (potentially of HSPA8 client proteins), or cellular redistribution of proteins (e.g. nuclear-cytoplasmic shuttling of DNA-PK) all would require longer kinetics than the phosphorylation cascade of cGAS-STING signaling. It will be interesting to know if there are anti-viral advantages to delayed kinetics, or if this simply reflects the signaling process.
  • It is intriguing that the SIDSP operates through critical components of fundamental cell biological pathways – DNA-PK (DNA-repair) and HSPA8 (essential chaperone activity). This is in contrast to cGAS and STING, which are generally described as non-essential and without well-defined roles outside immune regulation. One might imagine, then, that viral antagonism of the SIDSP will induce distinct cell stress phenotypes through loss of critical cell biological functions that may serve as signals to target an infected cell for destruction – a “backup plan” within the same pathway. As a specific example of this, does ICP0 antagonism (the non-oncogenic viral antagonist described) cause sufficient cell stress changes to render the infected cell targetable by the immune system? That is, does a cell with viral antagonism of DNA-PK look, for example, “transformed” enough to be killed through immunosurveillance mechanisms?


Posted on: 19 April 2019


Read preprint (1 votes)

Have your say

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Sign up to customise the site to your preferences and to receive alerts

Register here