In the life-threatening condition known as sepsis, the body responds to infection by inducing widespread biochemical changes that make the situation worse, some of which can lead to a severe decline in blood pressure. Several molecular factors that alter the constriction of blood vessels are involved in this decline, including nitric oxide, prostaglandins and oxidants such as hydrogen peroxide. In 2010, kynurenine — a metabolic product of the amino acid tryptophan — was identified1 as another factor that causes blood vessels to widen during sepsis. Writing in Nature, Stanley et al.2 (who work in the same laboratory as the researchers who identified kynurenine) now say that they got the wrong culprit.
Stanley and colleagues’ work begins as a classic whodunnit. The authors found that carefully purified kynurenine often did not cause blood-vessel widening (vasodilation) of isolated blood vessels in the context of inflammation, despite the previous report1. However, they consistently observed vasodilation using a mixture of tryptophan and either the enzyme indoleamine 2,3-dioxygenase 1 (IDO1) or singlet oxygen, a reactive oxygen species that is generated by IDO1. The expression of IDO1 is normally low in cell types other than immune cells, but can be upregulated by inflammatory proteins known as cytokines and by redox stress3,4 — which means that IDO1 is often expressed in the presence of oxidants.
Both IDO1 and singlet oxygen are involved in the production of kynurenine from tryptophan (Fig. 1a), but also in the making of other metabolites. The authors’ findings therefore suggested that another vasodilator was being formed. They hunted it down, and found it to be a compound that they call cis-WOOH (Fig. 1b), which is formed by IDO1 in a reaction involving tryptophan and singlet oxygen in the presence of hydrogen peroxide.
IDO1 activity is conventionally thought to be stimulated by chemical reducing agents and inhibited by hydrogen peroxide5. By contrast, Stanley et al. found that reducing agents did not generate cis-WOOH in their system, whereas exposure to hydrogen peroxide did. In a series of clever chemistry experiments, the authors showed that: in the presence of hydrogen peroxide, IDO1 generates singlet oxygen, followed by cis-WOOH; the oxygen atoms that are added to tryptophan to form cis-WOOH are derived from singlet oxygen rather than hydrogen peroxide; and both IDO1 activity and singlet oxygen are required for tryptophan to elicit vasodilation.
In the previous work from the same group1, kynurenine was reported to dilate blood vessels by causing the signalling molecule cyclic guanosine monophosphate (cGMP) to activate an enzyme called protein kinase G1α (PKG1α, which has a dimeric structure assembled from two identical protein monomers). Stanley and co-workers observed that vasodilation mediated by cis-WOOH needs much less cGMP than does blood-vessel relaxation mediated by kynurenine, but still requires activation of PKG1α.
The authors determined that cis-WOOH activates PKG1α by oxidizing a specific cysteine amino-acid residue (Cys 42) in the enzyme. This causes a disulfide bond to form6 between the Cys42 residues in the monomers of PKG1α. This oxidation has also been reported to stimulate the enzyme’s activity in a certain type of artery (resistance arteries) independently of cGMP levels6, and to contribute to the maintenance of normal blood pressure and to the ability of resistance arteries to dilate when exposed to hydrogen peroxide7.
The authors found that incubation of PKG1α with cis-WOOH or with tryptophan in the presence of IDO1 induces the formation of the disulfide bond. Moreover, vasodilation of blood vessels isolated from a mouse model of sepsis — endotoxaemic mice, in which the animals are exposed to bacteria-derived toxins — was suppressed when the animals expressed a mutant version of PKG1α that could not dimerize at Cys 42 on exposure to oxidants. The authors then observed that upregulation and activation of IDO1 in a mouse model of atherosclerosis (an inflammatory disease that is characterized by narrowing of blood vessels) contributes to tryptophan-mediated vasodilation and blood-pressure control, which suggests that IDO1 could be a target in efforts to develop treatments for inflammatory diseases.
The discovery adds cis-WOOH to a sizeable list of vasodilators that are involved in inflammatory responses, and reveals a specific role for singlet oxygen in the physiology of diseased mammalian cells (Fig. 1). Levels of IDO1 expression and activity must be high to form the singlet oxygen that, in turn, generates cis-WOOH; such conditions are common in inflamed tissues, but not healthy ones. The conversion of tryptophan to cis-WOOH therefore occurs under conditions of oxidative and inflammatory stress. Although hydrogen peroxide can cause vasodilation in the absence of IDO1 by inducing disulfide-bond formation between the Cys-42 residues in PKG1α, the authors found that some of the relaxation of artery walls that was induced by hydrogen peroxide under normal conditions depends on the presence of IDO1.
The idea that kynurenine-induced vasodilation is caused by the activation of PKG1α by cGMP1 now looks doubtful. Stanley et al. suggest that the vasodilation reported in the previous study was caused by contamination of kynurenine with cis-WOOH. However, the link between cis-WOOH and vasodilation does not depend on cGMP, but rather on PKG1α oxidation, so the authors’ proposal cannot explain the earlier findings.
Could selective inhibition of IDO1 be used therapeutically to reduce the large drop in blood pressure that is experienced by people who have sepsis? Perhaps, although several other approaches are also possible, given that many vasodilating factors are involved in systemic inflammation. Moreover, IDO1 has many crucial roles in immune cells, so its inhibition is likely to have several effects. For example, IDO1 inhibitors have been developed for use in cancer therapy8 — many tumours express IDO1, which helps them to evade the immune system by suppressing the functioning of certain immune cells and enhancing that of others9. Clinical trials of such compounds have been disappointing, in part because of the complexity of the pathway being targeted9. Nevertheless, it would be interesting to see whether those trials indicate that IDO1 inhibition can raise blood pressure, particularly in people who have inflammatory syndromes and elevated levels of kynurenine.
Another consideration with respect to the therapeutic use of IDO1 inhibitors is that IDO1 upregulation during sepsis might protect the body from a hyperactive immune system by increasing tolerance to endotoxaemia10. And although kynurenine and its metabolites are highly relevant to cancer and immune modulation3,5,11, the extent to which tumours and the immune system are affected by tryptophan signalling through the pathway now identified by Stanley and co-workers remains to be seen.
Vasodilation by cis-WOOH ultimately derives from the oxidation of Cys 42 in PKG1α. However, Cys 42 might not be the only cysteine residue that can be oxidized to activate the enzyme; Cys 117 seems to be another12. The effects of the Cys 42 mutation that Stanley et al. used to prevent PKG1α dimerization are also controversial. Some studies12,13 have found that it might depress cGMP-stimulated PKG1α activity. Furthermore, PKG1α dimerization triggered by Cys 42 oxidation occurs in the hearts of mice stressed by high blood pressure (pressure overload), yet its prevention in mice with the Cys 42 mutation enhances PKG1α-mediated protection against such stress14 — not by altering PKG1α activation, but by changing the intracellular localization of that enzyme and therefore its interactions with other proteins. Stanley and colleagues found that levels of IDO1 in heart-muscle tissue in mice were undetectable during pressure overload, but rose markedly in response to administration of interferon-γ (a protein that is involved in many inflammatory responses), which suggests a role for IDO1-generated cis-WOOH in inflammatory conditions of the heart. These details will be interesting to unpick, as the plot of this vasodilation whodunnit thickens.
Nature 566, 462-464 (2019)