|Home | About | Journals | Submit | Contact Us | Français|
Heme oxygenase-1 (HO-1), via its enzymatic degradation products, exhibits cell and tissue protective effects in models of vascular injury and disease. The migration of vascular smooth muscle cells (VSMC) from the medial to the intimal layer of blood vessels plays an integral role in the development of a neointima in these models. Despite this, there are no studies addressing the effect of increased HO-1 expression on VSMC migration.
The effects of increased HO-1 expression as well as biliverdin, bilirubin, and carbon monoxide (CO), were studied in in vitro models of VSMC migration. Induction of HO-1 or CO, but not biliverdin or bilirubin, inhibited VSMC migration. This effect was mediated by the inhibition of Nox1 as determined by a range of approaches including detection of intracellular superoxide, NADPH oxidase activity measurements, and siRNA experiments. Furthermore, CO decreased PDGF-stimulated, redox-sensitive signaling pathways.
Herein we demonstrate that increased HO-1 expression and CO decreases PDGF-stimulated VSMC migration via inhibition of Nox1 enzymatic activity. These studies reveal a novel mechanism by which HO-1 and CO may mediate their beneficial effects in arterial inflammation and injury.
Heme oxygenase (HO)-1 is an inducible stress protein that has cellular and tissue protective effects in vascular injury and disease1, 2. The tissue protection of HO-1 likely relates to the production of its enzymatic products, biliverdin (BV)/bilirubin (BR) and carbon monoxide (CO)2, 3. BV and BR are antioxidants that can provide protection against oxidative stress in cell culture and in vivo3–5. Recent evidence also demonstrates anti-inflammatory and anti-proliferative properties of these pigments6. Although toxic at high concentrations, low concentrations of CO confer anti-inflammatory, anti-apoptotic, anti-proliferative, and vasodilatory effects7. Indeed, both CO and BV/BR have been shown to inhibit vascular smooth muscle cell (VSMC) proliferation in vitro and neointima formation in response to vascular injury6, 8.
In animal models of vascular injury, intimal and medial thickening is thought to be attributable not only to VSMC proliferation but also migration of VSMC from the media to the intima 9, 10. The NADPH oxidase isoform Nox1 and Nox4 were recently shown to play important roles in the migration of VSMC 11–14. The NADPH oxidase family of enzymes catalyzes the one electron reduction of molecular oxygen to superoxide (O2−). Superoxide may be dismutated to hydrogen peroxide by various isoforms of superoxide dismutase (SOD). These reactive oxygen species (ROS) promote the migration of VSMC via activation of redox-sensitive kinases and/or inhibition of phosphatases15.
Prior studies suggest that CO and/or BV/BR may inhibit NADPH oxidase activity16–18. Additionally, BV/BR may scavenge NADPH oxidase-derived ROS by due to their antioxidant properties. Thus, in this study we hypothesized that increased expression of HO-1 and heme degradation products exhibit anti-migratory properties in addition to their anti-proliferative effects. We further hypothesized that the inhibition of NADPH oxidase activity or scavenging of ROS by these HO-1-derived products mediates this anti-migratory effect.
Tricarbonyldichlororuthenium (II) dimer (CORM-2), ruthenium (III) chloride hydrate (RuCl3), peg-SOD, platelet-derived growth factor BB (PDGF-BB), dihydroethidium (DHE), Triton X-100, dimethyl sulfoxide, and reduced nicotinamide adenine dinucleotide phosphate (NADPH) were from Sigma (St. Louis, MO). Diphenylene iodonium (DPI) was from Calbiochem (San Diego, CA). Lucigenin was from Alexis (San Diego, CA). Antibodies against total and phospho-ERK, p42/p44, JNK and p38 antibodies were from Cell Signaling Technology Inc. (Danvers, MA). Co(III) Protoporphyrin IX chloride (CoPP) was from Frontier Scientific (Logan, UT).
Rat aortic smooth muscle cells (RASMC) were isolated by collagenase/elastase digestion and maintained in DMEM with 10% FBS, 2mM L-glutamine,100 units/ml penicillin, and 100μg/ml streptomycin. RASMC were used at passages 3–8.
Adenovirus expressing recombinant β-galactosidase (AdLacZ) and recombinant rat HO-1 (Ad HO-1) have been described elsewhere19.
The activity of heme oxygenase enzymes was determined in cell extracts by measuring conversion of hemin to bilirubin as previously described 20.
PDGF-induced migration of RASMC was measured by wounding a monolayer of cells and monitoring the decrease in area after 18h.
PDGF-induced Intracellular O2− production was evaluated by measuring the conversion of DHE to hydroxyethidium in a fluorimeter (TECAN infinite M200).
Membrane fractions were prepared as previously described21.
NADPH oxidase activity was measured by monitoring lucigenin (5μM) chemiluminescence or cytochrome c reduction21.
The medial layers of aorta were isolated, membrane fractions prepared and NADPH oxidase activity measured by lucigenin (5μM) chemiluminescence21.
Statistics were performed using Graphpad Prizm software. Data were analyzed by one-way ANOVA followed by Tukey’s test for multiple comparisons. For grouped analysis data were analyzed by two-way ANOVA followed by Bonferonni’s post hoc-test.
Increased expression of HO-1 has been shown to modulate VSMC proliferation8, but not VSMC migration. We, therefore, examined PDGF-induced migration of RASMC with or without prior treatment with HO-1 adenovirus (AdHO-1), or control β-galactosidase adenovirus (AdLacZ). In all cases cells were treated with adenovirus 24h prior to stimulation with PDGF. As seen in Fig 1B, overexpression of HO-1 mediated by infection of cells with AdHO-1 resulted in decreased RASMC migration. The inhibition of VSMC migration correlated with both increased HO-1 protein expression and activity (Fig 1A). Similar results were obtained using the chemical inducer of HO-1, CoPP (supp. Fig 1).
HO-1 catalyzes the breakdown of heme into CO, free iron and biliverdin (and subsequently bilirubin). To determine which of these products are involved in the regulation of PDGF-induced RASMC migration by HO-1 we treated cells with CO gas (250ppm), biliverdin (30μM), or bilirubin (10μM). Only the addition of CO to the cells decreased PDGF-induced migration (Fig 1C) providing evidence that CO is responsible for the anti-migratory effects seen with increased HO-1 expression. As an alternative to CO gas, we also tested whether a CO-releasing molecule, CORM-2, could inhibit RAMSC migration. CORM-2 dose-dependently inhibited RASMC migration (Fig 1D). Importantly, RuCl3 did not affect PDGF-induced migration demonstrating that the effect of CORM-2 was due to CO and not the presence of the ruthenium base compound. Similar results were obtained using the Boyden chamber assay (supp. Fig 2).
We hypothesized that the anti-migratory effects of CO might be mediated via inhibition of NADPH oxidase-derived ROS. Using qRT-PCR, we confirmed the expression of Nox1 and Nox4 but not Nox2 in our RAMSC cultures (supp. Fig 3). Our first approach to elucidate a possible role for ROS and NADPH oxidase in the CO-dependent inhibition of VSMC migration, was to evaluate levels of O2− in RASMC in response to PDGF. We observed that PDGF induced a rapid and sustained increase in O2− levels (supp. Fig 4) as measured by DHE fluorescence. To determine if the ROS production was NADPH oxidase-dependent, we pre-incubated the cells with the non-specific flavo-protein inhibitor DPI, which is often used as a preliminary indicator of NADPH oxidase activity. DPI abolished PDGF-induced increases in ROS levels suggesting a possible role for NADPH oxidase in PDGF-induced O2− production (Fig 2A). Peg-SOD (10U/ml) was used to validate the specificity of the assay for O2−.
To determine whether CO gas (250ppm) or CORM-2 could inhibit PDGF-stimulated increases in O2− RASMC were treated with CO gas or CORM-2 and 30 min later stimulated with PDGF. Superoxide production was monitored by DHE as above. Both CO gas and CORM-2, but not RuCl3, decreased PDGF-induced O2− production (Fig 2B).
To more directly test the hypothesis that HO-1/CO inhibits PDGF-induced NADPH oxidase activity, NADPH-dependent O2− production was measured in 28,000xg membrane fractions from treated cells using lucigenin-enhanced chemiluminescence. Treatment of RASMC with PDGF caused a >2 fold increase in NADPH dependent O2− production that was inhibited by DPI and scavenged by peg-SOD (Fig 3A). Importantly, L-NAME or allopurinol had no effect on NADPH-dependent superoxide production providing evidence that the superoxide measured was not from uncoupled nitric oxide synthase or xanthine oxidase activity. The effect of DPI or peg-SOD on NADPH oxidase activity was further confirmed via measurement of O2− using cytochrome c reduction in similar experiments (Fig 3B). We next tested whether induction of HO-1 expression could inhibit PDGF-induced NADPH oxidase activity in RASMC. The increased HO-1 expression and activity due to AdHO-1 resulted in significant inhibition of PDGF-induced NADPH oxidase activity (Fig 3C).
To determine whether the effect of HO-1 on NADPH oxidase activity was due to CO, RASMC were pretreated with different concentrations of CO gas (100 or 250 ppm) or 100μM CORM-2 and stimulated with PDGF. CO gas as well as CORM-2 decreased PDGF-induced NADPH oxidase activity (Fig 4A). In ex vivo experiments strips of medial smooth muscle were preincubated with or without CORM-2 for 30 min followed by an additional 45 min incubation with PDGF. Membrane fractions from the treated medial strips were then prepared and NADPH oxidase activity measured. CORM-2, as well as peg-SOD, inhibited NADPH oxidase activity in the medial layer of the rat aorta (Fig 4B) providing evidence that CO could inhibit NADPH oxidase activity in an intact tissue.
We next sought to determine whether CO could directly inhibit NADPH oxidase activity. Fig 4C shows representative tracings of real-time lucigenin chemiluminescence from membrane fractions derived from control or PDGF-stimulated cells. Addition of NADPH to the membrane fraction from PDGF-stimulated cells results in a greater than 2-fold increase in lucigenin chemiluminescence compared to membrane fraction from control cells. Addition of CO saturated buffer (final conc. of approx. 9μM based on the solubility of CO in water) to the PDGF treated sample after NADPH addition caused a rapid decrease in lucigenin chemiluminescence.
To test whether the inhibition of migration by HO-1/CO is mediated by inhibition of NADPH oxidase we transiently transfected RASMC with siRNA against Nox1, Nox4 or a non-targeting (NT) siRNA. Transfection of RASMC with Nox1 or Nox4 siRNA decreased their mRNA levels by approximately 85% and 60% respectively after 72h (supp. Fig 5A). Nox4 but not Nox1 or NT siRNA significantly inhibited basal NADPH oxidase activity measured in membrane fractions from unstimulated cells (Fig 5A). Alternatively, Nox1 siRNA, but not NT or Nox4 siRNA, reduced PDGF-induced NADPH oxidase activity to basal levels as measured by lucigenin chemiluminescence in membrane fractions from PDGF-stimulated cells (Fig 5B). Importantly, CORM-2 inhibited PDGF-stimulated NADPH oxidase acitivity in Nox4 but not Nox1 siRNA treated cells. Both Nox1 and Nox4 siRNA, but not NT siRNA significantly reduced PDGF-induced migration, however, only cells treated with NT or Nox4 siRNA were still sensitive to inhibition of migration by CORM-2 (Fig 5C). Similar results were obtained using the Boyden chamber assay (supp. Fig 5B). These data strongly support the hypothesis that, while Nox1 and Nox4 are both important in mediating RASMC migration, HO-1/CO inhibits RASMC migration via inhibition of Nox1.
NADPH oxidase-derived ROS have been shown to activate pro-growth, pro-migratory pathways in RASMC22, 23. Serum starved RASMC were stimulated with PDGF with or without prior treatment with CORM-2. Western blot analysis revealed that treatment of RASMC with CORM-2 resulted in decreased phosphorylation of ERK1/2, p38, JNK, and AKT when compared to control samples (Fig 6A–F). Each of these pathways is known to be involved in RASMC migration as specific inhibition of these pathways results in decreased migration. Finally, RuCl3 had no effect on these redox sensitive signaling pathways (supp. Fig 6).
The protective role of heme oxygenase-1 has been studied in the context of various vascular diseases. In children lacking a functional HO-1 allele, atherosclerosis (hyperlipidemia, fatty streaks and plaques) is increased24. Moreover, HO-1 over-expression reduces lesional area in the aorta of ApoE−/− mice25. HO-1 is induced after balloon angioplasty in rats26 and neointimal hyperplasia is exacerbated in HO-1 null mice. Additionally, both CO 8, 27 and biliverdin 28 have been shown to inhibit neointima formation. Interestingly, to date, there exists no study exploring the effect of HO-1/CO on vascular smooth muscle cell migration, an integral process to the development of atherosclerosis and restenosis after angioplasty.
Herein, we provide evidence that induction of HO-1 expression in RASMC inhibits PDGF-induced migration. CO mediated the inhibition of migration by HO-1 since CO gas or the CO releasing molecule, CORM-2, but not biliverdin or bilirubin was able to inhibit RASMC migration. We therefore, focused on the mechanism by which CO inhibits PDGF-induced RAMSC migration.
Many of the pro-migratory signaling pathways stimulated by PDGF are mediated by ROS15. In VSMCs, antioxidants block migration in response to PDGF 29. In contrast, VSMCs extracted from Nox1 or p22phox-overexpressing mouse aortas exhibit an increase in PDGF-stimulated migration 30. These studies led us to hypothesize that CO inhibits RASMC migration via inhibition of NADPH oxidase activity. Indeed, we demonstrate that CO inhibits PDGF-induced O2− production in intact cells as well as NADPH oxidase activity in membrane fractions from PDGF-stimulated cells. By isolating the medial layer of rat aorta we were likewise able to demonstrate the inhibition of PDGF-stimulated NADPH oxidase activity in an ex vivo setting. Furthermore, direct addition of CO to the membrane fraction isolated from PDGF-treated cells rapidly decreased NADPH-stimulated O2− production. These data provide compelling evidence that CO inhibits NADPH oxidase activity in RASMC by directly interacting with the enzyme.
Aortic VSMC express both Nox1 and Nox4. Additionally, recent studies demonstrate an important role for either Nox1 or Nox4 in VSMC migration 11, 12. In particular, Lee et al. demonstrated that VSMC derived from Nox1 null mice exhibit decreased PDGF-induced migration whereas VSMC derived from smooth muscle specific Nox1 overexpressing mice exhibit enhanced migratory responses. In this study, the specific involvement of Nox1 in PDGF-stimulated NADPH oxidase activity was revealed by experiments demonstrating that this activity could be prevented by Nox1 siRNA but not Nox4 siRNA. Both Nox1 siRNA and Nox4 siRNA were able to prevent PDGF-induced migration to various extents, however, cells treated with Nox1 siRNA were resistant to further inhibition by CO whereas Nox4 treated cells were not. From these observations we draw the conclusion that CO inhibits Nox1-dependent ROS production leading to inhibition of VSMC migration.
Early studies on Nox2 (a.k.a. gp91phox, cytochrome b558) utilized CO as a tool to study the two heme moieties within the protein31, 32. These in vitro studies using partially purified enzyme generally concluded that CO binds the heme group in Nox2 poorly if at all. One inherent drawback of those studies, however, is that partially purified Nox2 was used which likely contained only the membrane components of the enzyme complex. At the time, it was not known that additional cytosolic subunits, such as p47 and p67, are required in the enzymatic complex. This leaves open the possibility that the conformation of Nox2 maybe altered when bound to these subunits, which could allow for interaction of CO with the heme group. Indeed, more recent studies examining the effect of CO on Nox2 have demonstrated changes in the heme absorbance spectra in response to CO16, 17. Despite this demonstration, these studies did not address whether CO could directly inhibit NADPH oxidase enzymatic activity. Rather these studies demonstrate decreased NADPH oxidase activity in intact cells treated with CO and thus could not distinguish whether CO was inhibiting signaling processes leading to the activation of Nox2 or directly inhibiting Nox2 activity. Indeed, our studies demonstrating a direct effect of CO on NADPH oxidase enzymatic activity are unique in this respect. Taken together, our studies along with those showing alteration of the heme spectra of Nox2 suggest that CO may indeed inhibit NADPH oxidase enzymatic activity via binding to one or both of the heme groups in the Nox subunit of the enzyme complex.
As discussed earlier many of the pro-migratory pathways stimulated by PDGF are redox sensitive15. One important mediator of growth factor responses in VSMC is Akt. ROS sensitivity of Akt is conferred by the phosphorylation of MAPKAPK-2 by p38 MAPK (mitogen activated protein kinase), a redox-sensitive kinase33. This leads to recruitment of MAPKAPK-2 to an Akt –p38MAPK complex and phosphorylation of Akt 34. Besides p38MAPK other MAPKs, are sensitive to ROS. c-Jun NH2 terminal kinase (JNK) activation in response to Ang II is blocked by antioxidants 35. ERK1/2 was the earliest discovered redox sensitive kinase having been shown to be activated by direct addition of hydrogen peroxide to cells 22. Additionally, Janus tyrosine kinases (JAKs) activate ERK1/2 in VSMCs and JAK2 activation in response to Ang II was shown to be attenuated by NADPH oxidase inhibitors36. In agreement with these findings we were able to demonstrate that, in RASMC, inhibition of Nox1 activity by CO correlates with decreased phosphorylation of AKT as well as the MAPKs, p38, ERK1/2, and JNK1.
The use of CORM-2 in this study raises some question as to whether the effects seen are due to direct effects of CO or secondary effects of CORM-2. In particular, CORM-2 may induce HO-1 and thus contribute to the effects of CORM-2 in the wound migration assay. However, our studies show an effect of CO gas in this model of migration assay. Furthermore, CORM-2 also inhibits RASMC migration in the Boyden chamber assay of migration. The time frame of this assay (4h) is such that one would not expect increased HO expression to be a factor. In terms of NADPH oxidase activity, these assays are all on a short time scale and thus induction of HO-1 by CORM-2 cannot be a factor. Concerns that the ruthenium metal center might be mediating the effects of CORM-2 were assuaged by the demonstration that ruthenium chloride had no effects in the models studied.
In conclusion, our studies demonstrate that in RASMC CO inhibits Nox1-dependent migration stimulated by PDGF. Furthermore, we show that CO inhibits Nox1 activity, likely via direct interaction with the enzyme complex. These studies reveal a novel mechanism by which increased HO-1 expression and activity and HO-1-derived CO may mediate their beneficial effects in arterial inflammation and injury.
Sources of Funding: This work was supported by NIH HL085134 to P.M.B.