Transcriptional Regulation of Cystathionine-γ-Lyase in Endothelial Cells by NADPH Oxidase 4-Dependent Signaling

Abstract

The gasotransmitter, hydrogen sulfide (H2S) is recognized as an important mediator of endothelial cell homeostasis and function that impacts upon vascular tone and blood pressure. Cystathionine-γ-lyase (CSE) is the predominant endothelial generator of H2S, and recent evidence suggests that its transcriptional expression is regulated by the reactive oxygen species, H2O2. However, the cellular source of H2O2 and the redox-dependent molecular signaling pathway that modulates this is not known. We aimed to investigate the role of Nox4, an endothelial generator of H2O2, in the regulation of CSE in endothelial cells. Both gain- and loss-of-function experiments in human endothelial cells in vitro demonstrated Nox4 to be a positive regulator of CSE transcription and protein expression. We demonstrate that this is dependent upon a heme-regulated inhibitor kinase/eIF2α/activating transcription factor 4 (ATF4) signaling module. ATF4 was further demonstrated to bind directly to cis-regulatory sequences within the first intron of CSE to activate transcription. Furthermore, CSE expression was also increased in cardiac microvascular endothelial cells, isolated from endothelial-specific Nox4 transgenic mice, compared with wild-type littermate controls. Using wire myography we demonstrate that endothelial-specific Nox4 transgenic mice exhibit a hypo-contractile phenotype in response to phenylephrine that was abolished when vessels were incubated with a CSE inhibitor, propargylglycine. We, therefore, conclude that Nox4 is a positive transcriptional regulator of CSE in endothelial cells and propose that it may in turn contribute to the regulation of vascular tone via the modulation of H2S production.

Keywords: ATF4; CSE; NADPH oxidase 4; endothelial cell; hydrogen sulfide; reactive oxygen species (ROS); redox signaling; transcription regulation; transgenic mice; vascular tone.

FIGURE 1.

Nox4 regulated CSE expression in endothelial cells. A, representative Western blot and quantitative histograms indicating Nox4 and CSE protein expression in HUVECs after 48-h Nox4 or β-gal (B Gal) overexpression. B and C, QPCR analyses of Nox4 and CSE mRNA expression after Nox4 or β-gal overexpression in HUVEC for times and multiplicity of infection doses as indicated. D, QPCR analyses of CSE mRNA expression in HUVEC after a 10 mm H₂O₂ or control H₂O treatment for 1 h. E, QPCR analyses of Nox4 and CSE mRNA expression in HUVECs after 24 h of treatment with Nox4-targeted siRNA (siNox4) or control siRNA (siScram) and quantitative histogram and representative Western blot depicting CSE protein expression in HUVEC after 48 h of treatment with siNox4 or control siScram. F, QPCR analyses of relative expression of Nox1, Nox2, and Nox4 mRNA in HUVECs. G, representative Western blot and quantitative histogram indicating p22^phox protein expression and corresponding QPCR analyses of CSE mRNA expression in HUVEC after 24 h of treatment with siRNA targeted to p22^phox (sip22^phox) or control siRNA (siScram). All data normalized to β-actin protein or mRNA expression. n = 3 in all cases. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001 A.U., absorbance units.

FIGURE 2.

Nox4 regulated CSE expression in CMECs leading to reduced PE-induced aortic constriction. A and B, QPCR analyses of Nox4 and CSE mRNA expression in CMECs isolated from WT and eNox4 Tg mice. C, representative Western blot and corresponding densitometric analysis of CSE protein expression in CMECs from WT and eNox4 Tg mice. For A, B, and C, data are representative of 3 separate isolations, where n = 2/3/4 in each group in each isolation. A.U., absorbance units. D, PE-induced constriction of aortae isolated from WT or eNox4 Tg mice, preincubated with or without 20 mm PPG for 30 min. All data normalized to β-actin mRNA and protein expression. *, p < 0.05; **, p < 0.01; ****, p < 0.0001. Myography data: WT n = 8 (12 rings), eNox4 Tg n = 6 (9 rings), WT + PPG n = 8 (13 rings), eNox4 Tg + PPG n = 9 (16 rings). *, p < 0.05; **, p < 0.01 (* compares WT with eNox4 Tg); +++, p < 0.001; ++++, p < 0.0001 (+ compares eNox4 Tg with eNox4 Tg + PPG).

FIGURE 3.

Nox4 regulated CSE expression via ATF4. A, representative Western blot and quantitative densitometric analyses of ATF4 protein expression and the corresponding QPCR analyses of CSE mRNA expression in HUVECs after 24-h ATF4 or control pCDNA3.1 overexpression. A.U., absorbance units. B, representative Western blot and quantitative densitometric analyses of ATF4 protein expression in HUVEC after 48 h of treatment with ATF4-targeted siRNA (siATF4) or control siRNA (siScram) together with 24 h Nox4 or β-gal (B Gal) overexpression as indicated. C, QPCR analyses of ATF4 and CSE mRNA expression in HUVECs after treatments as in B. D, representative Western blot and corresponding densitometric analyses of ATF4 protein expression in CMECs isolated from WT and eNox4 Tg mice. All data are normalized to β-actin mRNA and protein expression. n = 3; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

FIGURE 4.

Nox4 elicited a cellular stress response. A, representative Western blots and quantitative densitometric analysis of ATF4, total eIF2α and phosphorylated eIF2α protein (Pi-eIF2α) in HUVECs after 24-h Nox4 or β-gal (B Gal) overexpression. A.U., absorbance units. B, QPCR analyses of Nox4 and CSE mRNA expression in HUVEC after 1 h of incubation under atmospheric (Control (21%)) or hypoxic (Hypoxia (1% O₂) conditions. C, HUVEC proliferation assessed after treatment with siRNA targeted to Nox4 (siNox4) or control siRNA (siScram) for 24 h. Cells were then seeded at equal densities onto E-plates (ACEA), and respective cell index was subsequently measured on an xCELLigence real time cell analyzer for 24 h. Cells were cultured in normal media or in media supplemented with 1 mm dimethyloxalylglycine (DMOG), administered 5 h after plating. The cell indices were normalized at the time point of compound administration in all cases (5 h). D, QPCR analyses of PERK and CSE mRNA expression in HUVECs after 48 h of treatment with siRNA targeted to PERK (siPERK) or control siRNA (siScram) together with 24 h Nox4 or β-gal overexpression as indicated. E, QPCR analyses demonstrating relative PHD4 mRNA expression in (rat) neonatal cardiomyocytes and HUVEC. F, PCR analysis of spliced XBP1 mRNA in HUVEC splicing after 24-h Nox4 or β-gal overexpression as indicated. G, representative Western blot and corresponding densitometric analysis for cleaved and total ATF6 protein expression in HUVECs after 24-h Nox4 or β-gal overexpression as indicated. HUVECs treated with tunicamycin (TM; 2 μg/ml) for 2 h compared with control (vehicle; DMSO) served as a positive control. All data were normalized to β-actin mRNA or protein expression apart from protein that was normalized to total eIF2α protein. n = 3 in all cases, *, p < 0.05; **, p < 0.01; ***; p < 0.001; ****, p < 0.0001.

FIGURE 5.

Nox4 regulated CSE transcription through the HRI/eIF2α/ATF4 signaling module. A, QPCR analyses of HRI and CSE mRNA expression in HUVEC after 48 h of treatment with siRNA targeted to HRI (siHRI) or control siRNA (siScram) together with 24-h Nox4 or β-gal overexpression as indicated. A.U., absorbance units. B, representative Western blot analyses of Nox4, total eIF2α, and phosphorylated eIF2α protein (Pi-eIF2α) and quantitative densitometric analyses of Pi-eIF2α, normalized to total eIF2α phosphorylation, after treatments as in A. C, representative Western blot analyses and quantitative densitometric analyses of ATF4 protein expression after treatments as in A. D and E, QPCR analyses of Nox4 and CSE mRNA expression in HUVEC after inhibition of heme biosynthesis with 1 mm 4,6-dioxoheptanoic acid (DA) (D) or inhibition of heme oxygenase-1 with 10 μm tin protoporphyrin (SnPP) (E) for 24 h. All data were normalized to β-actin mRNA and protein expression apart from B. n = 3 for mRNA analyses, n = 5 for protein analyses. *, p < 0.05; **, p < 0.01; ***; p < 0.001; ****, p < 0.0001.

FIGURE 6.

ATF4 induced CSE transcription via direct binding to a cis regulatory intronic site. A and B, luciferase activity resulting from HEK cells transfected with constructs as indicated together with (empty vector control plasmid) pCDNA3.1 or overexpressed ATF4. RLU, relative light units. n = 4. *, p < 0.05; ****, p < 0.0001. n/s, not significant. C, schematic representation of alignment of putative ATF4-binding sites within intron 1 of mouse and human CSE gene loci. The genomic fragments contained in each luciferase construct are indicated. D and E, formaldehyde cross-linked chromatin prepared from HEK cells transfected with pCDNA3.1 or ATF4 incubated with normal rabbit IgG (negative control), anti-acetyl-histone H3 (positive control), or anti-ATF4 as indicated. Aliquots of chromatin before immunoprecipitation served as a positive control (input). Purified DNA was analyzed using primers specific for site A or site B as indicated in panel C or exon 3 of Human RLP30. The results presented are representative of three separate experiments.

FIGURE 7.

A schematic illustration of the mechanism underscoring the Nox4-induced increase in CSE expression in endothelial cells. Hypoxia potentially acts as an upstream signal to induce Nox4 expression. Enhanced Nox4 activity leads to HRI activation and subsequent phosphorylation of eIF2α on serine 51. eIF2α phosphorylation attenuates global translation but permits ATF4 protein expression. ATF4 then binds to an intronic enhancer element in intron 1 of the CSE gene and subsequently promotes CSE expression.