During our investigation of STING-mediated metabolic changes, we found that adding STING agonist to cells caused a metabolic change not at early time points (1 hour or 4 hours after stimulation), but at later timepoints (20 hours post stimulation). The changes that we observed were an active mitochondrial metabolism characterised by increased oxygen consumption and a mild increase in glycolysis, accompanied by a strong decrease in mitochondrial membrane potential and mitochondrial fission. The loss of mitochondrial membrane potential and fission is not typically linked to increased mitochondrial function, and for us this was a very interesting and unique response to stimulation.
Since the effect was only observed at later timepoints, we suspected that it was not due to intrinsic STING signalling but due to a second-wave factor produced upon activation. The first potential candidate for this effect upon STING activation was IFN-I. To test this, we tested if IFN-I alone would have the same effect, and if blocking IFN-I would block this effect. Indeed, addition of recombinant IFN-I to macrophages showed the same metabolic effect at early timepoints. And the effect upon STING agonist or IFN-I could be blocked by blocking the IFN-I receptor. We then dedicated our study to determining the metabolic changes upon IFN-I stimulation rather than the use of direct STING agonists.
As we observed more mitochondrial fission upon STING/IFN, we used a genetic model to determine if fission was required for this phenotype. However, we found that the metabolic changes, excluding fission, still occurred in this model. Based on literature about downstream effects of IFN-I, we then tested some IFN stimulated genes and identified one that mediates the effects of STING/IFN stimulation.
To then study the effects of these changes, we tested primary cytokine production in both the fission-defective model and cells deficient in the target IFN-stimulated gene to differentiate between effects of decreased mitochondrial fission vs the other metabolic phenotypes observed. We found that in our genetic models, primary cytokine production upon STING agonist stimulation was not affected. We then tested the response to IFN-I stimulation and found an increase in IFN stimulated genes in fission-deficient cells. This indicated that the induction of mitochondrial fission by IFN-I acts as a negative feedback loop.
We then tested the effect on phagocytosis, focussing on the uptake of dead cells, efferocytosis. We found that IFN-I stimulation increased the efferocytosis rate of macrophages and this increase could be inhibited in cells deficient in the identified target IFN-stimulated gene. IFN-I can be inflammatory and is important for viral clearance, but it also serves a homeostatic function in the resolution of inflammation. Part of the resolution of inflammation is the clearance of dead cells and debris by macrophages, and we found that the metabolic changes induced by IFN-I were important for this resolutive function.
