Activation of NAD(P)H oxidase by tryptophan-derived 3-hydroxykynurenine accelerates endothelial apoptosis and dysfunction in vivo

Abstract

Rationale: The kynurenine (Kyn) pathway is the major route for tryptophan (Trp) metabolism in mammals. The Trp-Kyn pathway is reported to regulate several fundamental biological processes, including cell death.

Objective: The aim of this study was to elucidate the contributions and molecular mechanism of Trp-Kyn pathway to endothelial cell death.

Methods and results: Endogenous reactive oxygen species, endothelial cell apoptosis, and endothelium-dependent and endothelium-independent vasorelaxation were measured in aortas of wild-type mice or mice deficient for nicotinamide adenine dinucleotide phosphate [NAD(P)H] oxidase subunits (p47(phox) or gp91(phox)) or indoleamine-pyrrole 2,3-dioxygenase 1 with or without angiotensin (Ang) II infusion. As expected, AngII increased plasma levels of Kyn- and 3-hydroxykynurenine-modified proteins in endothelial cells in vivo. Consistent with this, AngII markedly increased the expression of indoleamine-pyrrole 2,3-dioxygenase in parallel with increased expression of interferon-γ. Furthermore, in wild-type mice, AngII significantly increased oxidative stress, endothelial cell apoptosis, and endothelial dysfunction. These effects of AngII infusion were significantly suppressed in mice deficient for p47(phox), gp91(phox), or indoleamine-pyrrole 2,3-dioxygenase 1, suggesting that AngII-induced enhancement of Kynurenines via NAD(P)H oxidase-derived oxidants causes endothelial cell apoptosis and dysfunction in vivo. Furthermore, interferon-γ neutralization eliminates AngII-increased superoxide products and endothelial apoptosis by inhibiting AngII-induced Kynurenines generation, suggesting that AngII-activated Kyn pathway is interferon-γ-dependent. Mechanistically, we found that AngII-enhanced 3-hydroxykynurenine promoted the generation of NAD(P)H oxidase-mediated superoxide anions by increasing the translocation and membrane assembly of NAD(P)H oxidase subunits in endothelial cells, resulting in accelerated apoptosis and consequent endothelial dysfunction.

Conclusions: Kyn pathway activation accelerates apoptosis and dysfunction of the endothelium by upregulating NAD(P)H-derived superoxide.

Keywords: 3-hydroxykynurenine; NAD(P)H oxidase; apoptosis; indoleamine-pyrrole 2,3,-dioxygenase; kynurenine; oxidative stress; tryptophan.

Figure 1. P47^phox or gp91^phox deletion attenuates AngII infusion-mediated endothelial apoptosis and dysfunction, in vivo

WT, P47^phox−/− and gp91^phox−/− mice were infused with saline or AngII (1000ng/min*Kg) for 2 weeks. A and B, Deletion of p47^phox and gp91^phox ablates AngII-induced superoxide production in vivo. AngII-stimulated superoxide anions were detected by using DHE staining, as described in Materials and Methods. C, Deletion of p47^phox and gp91^phox abolishes AngII-induced endothelial cell apoptosis. Apoptotic cells were detected by TUNEL-positive cells in ECs (n=4–6 per group, n=4 sections for each sample). D and E, p47^phox or gp91^phox deficiency rescues endothelial dysfunction induced by AngII. Endothelial function was assayed by Ach induced endothelium-dependent relaxation in WT, p47^phox−/− and gp91^phox−/− mice aortas. F and G, SNP induced endothelium-independent relaxation in WT, p47^phox−/− and gp91^phox−/− mice aortas are not affected by AngII. *, p<0.05 versus WT-saline; **, p<0.01 versus WT-saline; #, p<0.05 versus WT- AngII, n=4 to 10 per group.

Figure 2. IDO deletion alleviates endothelial apoptosis and dysfunction via superoxide anions elimination, in vivo

WT and IDO^−/− mice were administrated with saline or AngII (1000ng/min*kg) for 4 weeks. A, AngII increases Kyn. The serum concentrations of Kyn were detected by HPLC. B, AngII increases IDO expression in vivo. C, AngII increases the detection of 3-OHkyn modified proteins in WT but not in IDO^−/− mice. B and C, Representative immunohistochemical staining (original magnification, X400) and quantifications for IDO (B) and 3-OHkyn modified proteins (C) in aortas. D, AngII increases superoxide anions in WT but not in IDO^−/− mice. Superoxide anions were detected by using DHE staining. E and F, IDO deletion abrogates AngII-mediated endothelial apoptosis. Triple immunofluorescence photographs show apoptotic ECs (green) in WT and IDO^−/− mice aortic endothelium (CD 31, red); nuclei (Hoechst, blue). TUNEL positive cells colocalized with CD31 positive cells and CD31 positive cells were counted in all aortic sections (original magnification, 25×). Histogram shows the mean number of TUNEL-CD31 positive nuclei/CD31 positive nuclei. Images are representatives of a study conducted in a separate set of animals (n=6 per group, n=4 sections for each sample). G and H, The endothelium-dependent relaxation and -independent relaxation in mouse aortas of WT and IDO^−/− with or without AngII. G, IDO deficiency improves the impairment of endothelial function caused by AngII. H, SNP induced endothelium-independent relaxation in mice aortas of WT and IDO^−/− are not affected by AngII. *, p<0.01 versus WT-saline; #, p<0.05 versus WT-AngII, n=4 to 10 per group.

Figure 3. Both AngII-increased levels of Kyn and 3-OHkyn-modified proteins and AngII-induced superoxide anions and endothelial apoptosis are IFN-γ-mediated in mouse aortas in vivo

A, WT and IDO^−/− mice were administrated with saline or AngII (1000ng/min*kg) for 4 weeks. AngII increases the detection of IFN-γ in both of WT and IDO^−/− mice. B–F, WT mice were intraperitoneally injected with anti-IFN-γ or control rat IgG1 antibodies for 4 weeks beginning on the day of AngII infusion. B, IFN-γ neutralization eliminates Kyn. The serum concentrations of Kyn were detected by HPLC. C, IFN-γ neutralization inhibits IDO expression. D, IFN-γ neutralization ameliorates 3-OHkyn modified protein levels. E, IFN-γ neutralization abrogates AngII-induced superoxide anions in vivo. F, IFN-γ neutralization abolishes AngII-induced active caspase-3 staining. A, C, D and F, Representative immunohistochemical staining (original magnification, X400) and quantifications for the levels of IFN-γ (A), IDO (C), 3-OHkyn modified proteins (D) and active caspase-3 (F) in aortas, WT mice infused with saline for 4 weeks were employed as control (C, D and F). *, p<0.01 versus WT saline or WT AngII-anti-IgG, n=4 to 6 per group.

Figure 4. The inhibition of IDO or KMO abrogates ROS generation and cell apoptosis induced by IFN-γ

A, IFN-γ-induced endothelial cell apoptosis is dose-dependent. HAECs were incubated for 48 hours with various concentration of IFN-γ. B, IFN-γ-induced endothelial cell apoptosis is time-dependent. HAECs were treated with 100ng/ml IFN-γ for different time. C–F, HAECs were pretreated for 1 hour with medium alone or medium containing MT (200µM) or Ro61-8048 (50 µM) before incubation with or without IFN-γ (100ng/ml) for 48 hours. G–J, MLECs isolated from WT and IDO^−/− mice were challenged with or without 100ng/ml IFN-γ for 48 hours. C and G, Pretreatment with IDO inhibitor (1-MT) or KMO inhibitor (Ro61-8048) (C), or IDO deficiency (G) inhibits IFN-γ-induced protein cleavage of parp, caspase-3 and caspase-7 in ECs in vitro. D and H, Inhibition of IDO (D & H) or KMO (D) blocks IFN-γ-induced caspase-3 activation in ECs. E & I, Inhibition of IDO (E & I) or KMO (E) eliminates IFN-γ-induced cell apoptosis. F & J, Inhibition of IDO (F and J) or KMO (F) attenuates IFN-γ-induced ROS formation in ECs. A–C and G, Cell lysates were subjected to immunoblot analysis to detect the protein cleavage of PARP, caspase-7 and caspase-3. GAPDH and β-actin were used as loading controls. D and H, Capase-3 activity was measured in a fluorescent microplate reader. The results were the mean ± S.E. of three independent measurements. E and I, Nuclear DNA fragment as a sign of apoptosis was detected using a TUNEL assay. Data were reported as %TUNEL-positive in a microscopic field ±S.E. F and J, Superoxide production was assessed by DHE/HPLC. *, p<0.05 versus control; **, p<0.01 versus control; #, p<0.05 versus IFN- γ (A–F) or WT IFN-γ (G–J); ##, p<0.01 versus IFN- γ (A–F) or WT IFN-γ (G–J).

Figure 5. IFN-γ induces IDO and KMO protein expressions and endogenous kynurenine formation

A–C, HAECs were incubated for 24 hours with various concentration of IFN-γ. D–F, HAECs were treated with 100ng/ml IFN-γ for different time. A and D, IFN-γ induces the expressions of IDO and KMO in a dose- and time- dependent manner. Cell lysates were subjected to immunoblot analysis to detect IDO, KMO and β-actin. B and E, IFN-γ increases Kyn but decreases Trp in a dose- and time- dependent manner. C and F, IFN-γ increases IDO activity in a dose- and time- dependent manner. G–I, Inhibition of IDO reduces IFN-γ induced Kyn formation and Trp consumption. G, HAECs were pretreated for 1 hour with medium alone or medium containing 200µM 1-MT before incubation with or without IFN-γ (100ng/ml) for 48 hours. H and I, MLECs were stimulated with or without 100ng/ml IFN-γ for 48 hours. B, E and G–I, The levels of Trp and kyn in the culture medium were determined by HPLC. *, p<0.05 versus control; **, p<0.01 versus control; #, p<0.05 versus IFN-γ (G) or WT IFN-γ (H & J); ##, p<0.01 versus WT IFN-γ.

Figure 6. IFN-γ induced 3-OHkyn is both IDO- and KMO-dependent

A and B, Exogenous 3-OHkyn reacts with intracellular proteins. Haecs were incubated with or without 60µM 3-OHkyn for 24 hours. C–H, Inhibition of IDO or KMO (C–E), or IDO deficiency (F–H) eliminates 3-OHkyn modification of intracellular proteins induced by IFN-γ. C–E, Haecs were pretreated for 1 hour with medium alone or medium containing 200µM 1-MT or 50 µM Ro61-8048 before incubation with or without 100ng/ml IFN-γ for 48 hours. F–H, MLECs were challenged with or without 100ng/ml IFN-γ for 48 hours. A, C and F, Immunocytochemistry was performed to localize 3-OHkyn modifications by mouse anti-3-OHkyn monoclonal primary antibody and Alexa594 red-conjugated goat anti-mouse IgG secondary antibody. Cells were counterstained with a nuclear stain, DAPI (blue) (From up: 3-OHkyn modifications, DAPI and merge). E and H, Cell lysates were subjected to immunoblot analysis to detect 3-OHkyn modified proteins. GAPDH and β-actin were used as loading controls. *, p<0.01 versus control; #, p<0.01 versus IFN- γ.

Figure 7. 3-OHkyn is apoptosis-inducing kynurenine

A–C, 3-OHkyn induces cell apoptosis. HAECs were treated with complete serum medium or serum free medium alone or containing Kyn (60µM) or 3-OHkyn for 24 hours. D–F, 3-OHkyn induces cell apoptosis in a dose-dependent manner. HAECs were incubated for 24 hours with various concentration of 3-OHkyn. G–I, 3-OHkyn induces cell apoptosis in a time-dependent manner. HAECs were treated with 60µM 3-OHkyn for different time. A, D and G, Cell lysates were subjected to immunoblot analysis to detect the protein cleavage of parp, caspase-7 and caspase-3. B, E and H, The results of capase-3 activity were the mean ± S.E. of three independent measurements. C, F and I, Data were reported as %TUNEL-positive in a microscopic field ±S.E. *, p<0.05 versus SFM or control; **, p<0.01 versus control.

Figure 8. 3-OHkyn reduces cell viability but induces mitochondrial cytochrome c release, and ROS formation in a NAD(P)H oxidase-dependent manner

A–C, HAECs were incubated for 24 hours with indicated concentration of 3-OHkyn. D–J, HAECs were incubated with or without 3-OHkyn (60 µM) for 24 hours, which were infected with ad-GFP or ad-SOD1 (G)/2 (D), or pretreated with medium alone or medium containing tempol (10µM) or apocynin (100µM) for 1 hour (E & F). A, 3-OHkyn reduces endothelial cell viability in a dose-dependent manner. Cell viability was assessed by using MTT reduction assays. B and E, 3-OHkyn induces intracellular ROS generation, which is ablated by NAD(P)H oxidase inhibitor. Superoxide production was assessed by DHE/HPLC. C, 3-OHkyn accelerates mitochondrial cytochrome c release. Cytochrome c was detected in cytosolic fractions using a monoclonal antibody. D, F and G, Overexpress of SOD or pretreatment with NAD(P)H oxidase inhibitor attenuates 3-OHkyn-triggered apoptosis. Cell lysates were subjected to immunoblot analysis to detect SOD1 (G) or 2 (D) and the protein cleavage of PARP, caspase-7 and caspase-3. H, 3-OHkyn promotes translocation and membrane assembly of p47^phox and p67^phox. Immunoblot analysis of NAD(P)H oxidase subunits (p47^phox and p67^phox) were applied in membrane proteins and whole cell lysates by immunoblot. I, 3-OHkyn increases NAD(P)H activity. NAD(P)H activity was measured by a low concentration based lucigenin (5µM) assay. The results were the mean ± S.E. of three independent measurements. J, 3-OHkyn reacts with and chemically modifies p47^phox and p67^phox. Cell lysates were subjected to immunoprecipitation (IP)/immunoblot analysis to detect p47^phox and p67^phox modified by 3-OHkyn. Na+/K+ ATPase and β-actin were used as loading controls. *, p<0.05 versus control; **, p<0.01 versus control; #, p<0.05 versus 3-OHkyn only.