Heme oxygenase-1 induction and organic nitrate therapy: beneficial effects on endothelial dysfunction, nitrate tolerance, and vascular oxidative stress

Abstract

Organic nitrates are a group of very effective anti-ischemic drugs. They are used for the treatment of patients with stable angina, acute myocardial infarction, and chronic congestive heart failure. A major therapeutic limitation inherent to organic nitrates is the development of tolerance, which occurs during chronic treatment with these agents, and this phenomenon is largely based on induction of oxidative stress with subsequent endothelial dysfunction. We therefore speculated that induction of heme oxygenase-1 (HO-1) could be an efficient strategy to overcome nitrate tolerance and the associated side effects. Indeed, we found that hemin cotreatment prevented the development of nitrate tolerance and vascular oxidative stress in response to chronic nitroglycerin therapy. Vice versa, pentaerithrityl tetranitrate (PETN), a nitrate that was previously reported to be devoid of adverse side effects, displayed tolerance and oxidative stress when the HO-1 pathway was blocked pharmacologically or genetically by using HO-1(+/-) mice. Recently, we identified activation of Nrf2 and HuR as a principle mechanism of HO-1 induction by PETN. With the present paper, we present and discuss our recent and previous findings on the role of HO-1 for the prevention of nitroglycerin-induced nitrate tolerance and for the beneficial effects of PETN therapy.

Figure 1

Effects of the HO-1 inducer hemin on GTN-induced tolerance and effects of the HO-1 suppressor apigenin on PETN side effects. Hemin (25 mg/kg) was administrated by single i.p. injection on day 3 of GTN treatment (6.6 μg/kg/min for 4 days via s.c. infusion) and markedly improved vascular GTN responsiveness (see area between curves) as demonstrated by isometric tension studies (a), a significant decrease in mitochondrial ROS formation (b), and improvement of mitochondrial ALDH-2 activity (c). Effects of bolus bilirubin on ROS formation in isolated heart mitochondria from GTN in vivo treated rats were determined by L-012 (100 μM) ECL in the presence of 2.5 mM succinate and bilirubin (0–25 μM) (d). Apigenin (10 mg/kg/d) was coinfused over 4d together with PETN (10.5 μg/kg/min for 4 days via s.c. infusion). Apigenin cotreatment decreased PETN vasodilator potency (see area between curves) and induced a tolerance-like right shift in the PETN concentration-relaxation-curve (E). This observation was accompanied by increased mitochondrial ROS formation (F). Data are mean ± SEM of n  =  8–12 (a), 40 (b), 6–18 (c), 4–6 (d), 9–11 (e), and 28–43 (f) independent experiments. *P < 0.05 versus GTN or PETN treatment. Modified from [31].

Figure 2

Effects of HO-1 deficiency versus HO-1 induction on vascular improvement by pentaerithrityl tetranitrate (PETN). (a, b) PETN treatment (75 mg/kg/d for 4d) had no effect on PETN potency (PETN-induced relaxation) in aorta from control mice (HO-1^+/+) but caused nitrate tolerance (see area between curves) in aorta from mice with partial HO-1 deficiency (HO-1^+/−). In accordance, cardiac mitochondrial ROS formation (L-012 ECL) was increased in PETN-treated HO-1^+/− mice. P < 0.05: *versus HO-1^+/+/DMSO; ^#versus HO-1^+/−/DMSO. S: solvent; P: PETN-treated. (c, d) Hemin (25 mg/kg i.p.)-triggered HO-1 induction improved high-dose AT-II (1 mg/kg/d for 7d)-induced endothelial dysfunction (ACh-response) in aorta (see area between curves) and NADPH oxidase activity in heart (lucigenin ECL) from control mice (HO-1^+/+). P < 0.05: *versus AT-II-treated HO-1^+/+/DMSO. (e,  f) PETN (75 mg/kg/d for 7d) failed to prevent endothelial dysfunction (ACh-response) induced by low-dose AT-II (0.1 mg/kg/d for 7d) in aorta from HO-1^+/− mice (see area between curves). In accordance, PETN did not improve NADPH oxidase activity (2-HE formation by HPLC analysis) in cardiac samples from AT-II-treated HO-1^+/− mice. P < 0.05: *versus AT-II-treated HO-1^+/−/DMSO. All data are mean ± SEM of aortic rings and hearts from 4-5 animals/group. Modified from [32].

Figure 3

Scheme illustrating the mechanisms underlying the oxidative stress concept of nitrate tolerance in response to GTN treatment and the mechanisms underlying the beneficial vascular effects in response to PETN. PETN and GTN are bioactivated by mitochondrial ALDH (ALDH-2) yielding 1,2-glyceryl dinitrate and PETriN, respectively, as well as a yet undefined nitrogen species (NO_x, probably nitrite) that undergoes further reduction by the mitochondrial respiratory chain or acidic disproportionation to form an activator of sGC (probably nitrico xide). GTN treatment induces mitochondrial reactive oxygen and nitrogen species formation (ROS/RNS). These ROS/RNS in turn inhibit the GTN bioactivation process by inactivation of ALDH-2 or by inhibiting the repair system of the ALDH-2, which includes lipoic acid, as well as a reductase system depending on the NADH or NADPH (lipoicacid reductase (LAR), thioredoxin/thioredoxin reductase (Trx/TrxR) or glutathione/glutathione reductase (GSH/GR). In contrast to GTN, PETN provides potent antioxidative effects by inducing HO-1 and ferritin, which in turn decrease ROS levels and therefore protect the ALDH-2 from ROS mediated inactivation. Adapted from [31].

Figure 4

Comparison of the HO-1 promoter sequences of different species. (a) Comparison of the 5′-flanking sequences (10 kb) of the rat, mouse, rhesus macaque, chimpanzee, and human (arrow) HO-1 gene using the ECR-Browser software (http://ecrbrowser.dcode.org/). The search area was 10 bp, and the minimal homology was 80%. The height of the curves (50% < X < 100%) indicates the homology (red: intergenic regions, green: single repeats, yellow: untranslated regions of the RNA (UTR), blue: exon, salmon: intron, pink ECR, above). (b) Map of transcription factor binding sites (TFBS) in the human HO-1 promoter. TFBS labeled in red have been verified experimentally in different cells systems (see also [33]).

Figure 5

Effects of organic nitrates on the human HO-1 promoter activity. DLD1-HO-11kb-Prom cells were treated for 8 h with isosorbide dinitrate ISDN, PETN, isosorbide-5-mononitrate ISMN, or GTN at a concentration of 50 μM or the respective solvent (DMSO, H₂O, EtOH). Extracts were prepared, and luciferase activity and protein content were measured, and luciferase activity was normalized to the protein content. Shown (mean ± SEM; n = 8–10) are the normalized luciferase activity values. The normalized luciferase activity of the cell treated with solvent was set to 100%. (***= P < 0.001, ns: not significant different from solvent-treated cells.) These data were partly published in [34].

Figure 6

PETN-induced enhancement of the human HO-1 promoter activity depends on NRF2. DLD1-HO-11kb-Prom cells were transfected with a specific anti-NRF2 siRNA (siNRF2) or a nonrelated control siRNA (siCon). After 48 h, the cells were incubated with PETN or the solvent DMSO. Extracts were prepared and luciferase activity and protein content were measured, and luciferase activity was normalized to the protein content. Shown (mean ± SEM; n = 5-6) are the normalized luciferase activity values. The normalized luciferase activity of the cell treated with solvent and siCon was set to 100%. (**= P < 0.01, *= P < 0.05, ns: not significant different to DMSO and siCon-treated cells. Modified from [34].

Figure 7

Comparison of the 3′-UTR of the HO-1 mRNA of different species. (a) The 3′-UTR sequences of the HO-1 mRNA from rat, mouse, rhesus macaque, chimpanzee and human was compared using the ECR-Browser software. The search area was 10 bp and the minimal homology was 80%. The height of the curves (50% < X < 100%) indicates the homology (red = intergenic regions, yellow = untranslated regions of the RNA (UTR), blue = exon, pink ECR, above). (b) Using the software RNALogo (http://rnalogo.mbc.nctu.edu.tw/createlogo.html) a consensus sequence of all 5 3′-UTR sequences was generated. The height of the letters indicate the frequency of the appearance of this base. AREs are marked by a red box. A putative HuR binding site is indicated.

Figure 8

Post-transcriptional regulation of the human HO-1 expression. Human endothelial EA.hy 926 cells were transiently transfected with pGL3-Control-HO-1-3-UTR and pRenilla (normalization of transfection efficiency). After 24h the cells were treated with 50 μM nitroglycerin (GTN) or PETN (or the solvents ethanol [EtOH] or DMSO) for 6h. Extracts were prepared and luciferase and Renilla activity were determined. The luciferase activity was normalized to the renilla activity. Shown (mean ± SEM; n = 6–8) are the normalized luciferase activity values. The normalized luciferase activity of the cell treated with solvent were set to 100% (*= P < 0.05; ns = not significant versus EtOH- or. DMSO treated cells).

Figure 9

Molecular mechanisms of PETN-mediated enhancement of HO-1 expression. The high amounts of bioactive NO generated from PETN (but not GTN) activate the transcription factor NRF2 and thereby enhance the HO-1 promoter activity. In addition, the interaction of the stabilizing RNA binding protein HuR with the 3′-UTR of the HO-1 mRNA is enhanced. Both effects result in an enhancement of HO-1 expression (E: exon, I: intron).