Protection from a prolyl hydroxylase domain-containing enzyme (PHD) inhibitor, desferoxamine (DFO), was recently reported to be dependent on production of reactive oxygen species (ROS). Ischemic preconditioning triggers the protected state by stimulating nitric oxide (NO) production to open mitochondrial ATP-sensitive K+ (mitoKATP) channels, generating ROS required for protection. We tested whether DFO and a second PHD inhibitor, ethyl-3,4-dihydroxybenzoate (EDHB), might have similar mechanisms. EDHB and DFO increased ROS generation by 50–75% (P < 0.001) in isolated rabbit cardiomyocytes. This increase after EDHB exposure was blocked by Nω-nitro-l-arginine methyl ester (l-NAME), an NO synthase (NOS) inhibitor; ODQ, a guanylyl cyclase antagonist; and Rp-8-bromoguanosine-3′,5′-cyclic monophosphorothioate Rp isomer, a PKG blocker, thus implicating the NO pathway in EDHB's signaling. Glibenclamide, a nonselective KATP channel blocker, or 5-hydroxydecanoate, a selective mitoKATP channel antagonist, also prevented EDHB's ROS production, as did blockade of mitochondrial electron transport with myxothiazol. NOS is activated by Akt. However, neither wortmannin, an inhibitor of phosphatidylinositol-3-kinase, nor Akt inhibitor blocked EDHB-induced ROS generation, indicating that EDHB initiates signaling downstream of Akt. DFO also increased ROS production, and this effect was blocked by ODQ, 5-hydroxydecanoate, and N-(2-mercaptopropionyl)glycine, an ROS scavenger. DFO increased cardiomyocyte production of nitrite, a metabolite of NO, and this effect was blocked by an inhibitor of NOS. DFO also spared ischemic myocardium in intact hearts. This infarct-sparing effect was blocked by ODQ, l-NAME, and N-(2-mercaptopropionyl)glycine. Hence, DFO and EDHB stimulate NO-dependent activation of PKG to open mitoKATP channels and produce ROS, which act as second messengers to trigger entrance into the preconditioned state.
- nitric oxide synthase
- prolyl hydroxylase inhibitor
- reactive oxygen species
myocardial ischemia activates an array of adaptive reactions, which include changes in cell metabolism, new vessel formation, matrix production, and even myocyte replication (5, 43). Several transcription factors have been shown to mediate physiological adaptations to hypoxia, including hypoxia-inducible transcription factors (HIF), which have been identified as important regulators of cellular responses to oxygen deprivation (21, 37, 42, 47). Under normoxic conditions, prolyl-4-hydroxylase domain (PHD)-containing enzymes catalyze the hydroxylation of key proline residues in the HIF subunit, marking it for proteasome-mediated degradation. PHD-mediated hydroxylation cannot occur when oxygen tension is reduced. Hence, HIF accumulates in the nucleus of hypoxic cells, causing HIF-mediated gene transcription (13). Cai et al. (7) recently demonstrated that HIF is involved in hypoxia-induced cardioprotection.
Wright et al. (44) reported that PHD inhibition with ethyl-3,4-dihydroxybenzoate (EDHB) induced the glucose transporter GLUT-1, nitric oxide (NO) synthase (NOS)-2, and the antioxidant heme oxygenase in cardiomyocytes. More importantly, they found that these cardiomyocytes were protected against simulated ischemia. Nwogu et al. (26) showed that inhibition of PHD can improve cardiac remodeling after myocardial infarction. Recently, we were able to demonstrate that inhibition of PHD increases HIF in hearts and improves cardiac function (32).
The iron chelator desferoxamine (DFO), which is also known to inhibit PHDs (3) and, thereby, preserve HIF, was first reported to be cardioprotective in 1987 (1). In that study, it was assumed that DFO acted by preventing reactive oxygen species (ROS) formed by the Haber-Weiss reaction, which requires free iron. Most recently, Dendorfer et al. (9) reported that DFO reduced infarct size in rats pretreated just 2 h before the ischemic insult. As expected, DFO increased HIF-2α mRNA levels. Surprisingly, they were able to block the DFO-induced protection with the ROS scavenger N-(2-mercaptopropionyl)glycine (MPG), indicating that ROS production was actually required for DFO's protection. It is well known that certain receptor agonists, including bradykinin and acetylcholine, also can protect the heart against ischemia by an ROS-dependent mechanism (27). In that scheme, ROS act as diffusible second messengers to trigger entrance into the preconditioned state. We wondered whether the PHD inhibitors might somehow be activating the preconditioning mechanism directly through the ROS-dependent signal transduction pathway. Gi-coupled receptors, including those binding bradykinin and acetylcholine, trigger the production of ROS by a complex pathway in which phosphatidylinositol-3 (PI3)-kinase activates Akt through phosphoinositide-dependent kinase (PDK)-1 and PDK-2, which in turn activate endothelial NOS (11, 17, 18). The resulting NO stimulates guanylyl cyclase, which then activates PKG. PKG causes mitochondrial ATP-sensitive K+ (mitoKATP) channels to open (29). MitoKATP channel opening leads to production of ROS by the mitochondria. Those ROS then trigger the protective process (18, 29, 30). Accordingly, in the present study, we tested two known PHD blockers on isolated rabbit cardiomyocytes to determine whether they would trigger ROS production and, if so, by what pathway. We also tested whether DFO and EDHB could precondition the intact heart when given just 15 min before ischemia and whether that protection was dependent on NO and guanylyl cyclase.
This study was performed in accordance with The Guide for the Care and Use of Laboratory Animals (25) and approved by the Institutional Animal Care and Use Committee.
Adult Rabbit Myocytes
The method for isolating rabbit ventricular myocytes has been described in detail elsewhere (2, 29). Briefly, New Zealand White rabbits were anesthetized with pentobarbital sodium (30 mg/kg iv), anticoagulated with heparin (1,000 U/kg iv), and ventilated with 100% oxygen. Hearts were excised, quickly mounted on a Langendorff apparatus, and retrogradely perfused with modified Ca2+-free Krebs-Henseleit-HEPES buffer containing collagenase (type 2, Worthington, Lakewood, NJ; 200 U/ml) at 37°C. After 15–25 min, the softened heart was minced and passed through a 200- to 350-μm-diameter nylon mesh. Viable myocytes were separated by repetitive slow-speed centrifugation and made Ca2+ tolerant by stepwise restoration of Ca2+ to 1.25 mM in buffer containing 2% bovine serum albumin. At least 2 × 107 viable, Ca2+-tolerant cells were extracted per heart. Preparations were considered satisfactory if rod-shaped cells accounted for >65% of the cells.
Immediately after the isolation and separation procedure, cells were plated on poly-d-lysine-laminin-coated 24-well plates (Becton Dickinson, Bedford, MA) using medium 199 supplemented with creatine (5 mM), l-carnitine (2 mM), and taurine (5 mM; CCT medium 199) as described by Piper et al. (33) and Mitcheson et al. (22). Penicillin (100 U/ml) and streptomycin (100 μg/ml) were added as antibiotics. Cells were stored in an incubator at 37°C in air enriched with 5% CO2. The medium was changed after 3–4 h; then the cells were allowed to equilibrate for ≥18 h. Experiments were performed 1–4 days after myocyte isolation.
Each experiment started with a change of medium for 10 min. In those experiments in which the effect of a blocker was investigated, the medium contained the blocker. The medium was then removed and replaced with medium containing the drug or drug + blocker (according to protocol) and reduced MitoTracker Red (1 μM). This reduced form of the probe is nonfluorescent and becomes fluorescent when oxidized by ROS. The oxidized product is bound to thiol groups and proteins within mitochondria. After incubation with MitoTracker Red for 15 min, the cells were washed with fresh MitoTracker Red-free CCT medium 199, which removes the unbound and, thus, voltage-dependent pool of oxidized probe held in the cells and any unreacted MitoTracker Red to stop the reaction. After the washing process, all that remains is the protein-bound fluorescent product, which is stable for ≥30 min. Previous investigations have documented that the described protocol of a timed incubation followed by washing permits reliable measurement of ROS generation and is insensitive to changes in mitochondrial transmembrane potential (16, 28).
The effects of two PHD inhibitors, EDHB (1 mM) and DFO (1 mM), were examined. In experiments in which the effects of the blockers myxothiazol (Myxo, 200 nM), a mitochondrial electron transport blocker; MPG (1 mM), an ROS scavenger; glibenclamide (Gli, 50 μM), a nonspecific KATP channel antagonist; 5-hydroxydecanoate (5-HD, 200 μM), a specific mitoKATP channel antagonist; Nω-nitro-l-arginine methyl ester (l-NAME, 200 μM), an NOS antagonist; 1-H-[1,2,4]oxadiazole[4,3-a]quinoxalin-1-one (ODQ, 10 μM), a blocker of soluble guanylyl cyclase; 8-bromoguanosine-3′,5′-cyclic monophosphorothioate Rp isomer (Rp-8-Br-cGMPS, 50 μM), a PKG blocker; Akt inhibitor [1L-6-hydroxymethyl-chiro-inositol 2-(R)-2-O-methyl-3-O-octadecylcarbonate, 20 μM], an antagonist of Akt; and wortmannin (Wort, 100 nM), a blocker of PI3-kinase, were to be examined, each blocker was present in the medium during the 10-min period before the addition of reduced MitoTracker Red and agonist.
Measurement of ROS production.
Experiments were designed such that four different conditions were always simultaneously evaluated. In each field, rod-shaped cells were considered viable, whereas round cells were considered dead. Mitochondrial ROS generation was analyzed by measuring the fluorescence of ≥25–50 individual rod-shaped cells that were randomly selected within each well. The average fluorescence of 150–300 cells was computed and expressed as a percentage of the average single-cell fluorescence in the respective control well in the same chamber. Thus treated cells were compared only with contemporary untreated cells of the same age and isolation and stained with the same MitoTracker Red preparation. The fluorescence of single cells was quantified as described previously (28) using a Nikon TMS-F microscope with a ×20 objective, an XF filter set (Omega Optical, Brattleboro, VT), a xenon light source with an optical filter changer (excitation at 560 nm and emission at 610 nm, light source 0.1 s; model lambda 10-2, Sutter Instruments, Novato, CA), and software from Intracellular Imaging (Cincinnati, OH). Each set of experiments was repeated on different days with cells of different ages and from at least two different isolations.
Measurement of nitrite production as an indicator of NOS activity.
Activation of NOS results in the generation of NO. The latter is quickly metabolized to nitrite, a stable product that can be measured and is reflective of NOS activity (14, 20). Each experiment started with a change of medium for the cells for 15 min. In those experiments in which the effect of the NOS inhibitor S-ethyl-N-[4-(trifluoromethyl)phenyl]isothiourea (ETU, 100 μM) was investigated, the medium contained ETU. The medium was then removed and replaced with medium containing 1 mM DFO with or without ETU. After 30 min the medium was removed, and nitrite was measured by reductive gas-phase chemiluminescence (10, 34). Samples of medium (100 μl) were injected into a reaction chamber filled with an iodide-triiodide-containing mixture that reacts with nitrite to release NO into the gas phase. Released NO was carried by a helium gas stream to the chemiluminescence detector, where NO reacts with ozone to form nitrogen dioxide. A proportion of the latter arises in an electronically excited state, which, on decay to its ground state, emits light in the near-infrared region that can be quantified with a photomultiplier. Provided ozone is present in excess and the reaction conditions are kept constant, the intensity of the emitted light is directly proportional to the NO concentration. All determinations were made in triplicate and averaged.
Isolated Rabbit Heart Model
New Zealand White rabbits of either gender were anesthetized with pentobarbital sodium, intubated, and mechanically ventilated with 100% oxygen. As previously described (8), a 2-0 silk suture was passed around a branch of the left coronary artery to form a snare by passage of the ends of the thread through a small vinyl tube. Hearts were removed, mounted on a Langendorff apparatus, and perfused with modified Krebs-Henseleit bicarbonate buffer containing (in mM) 118.5 NaCl, 25 NaHCO3, 4.7 KCl, 1.2 MgSO4, 1.2 KH2PO4, 2.5 CaCl2, and 10.0 glucose. A fluid-filled latex balloon was inserted into the left ventricle and inflated to set an end-diastolic pressure of 5 mmHg. All hearts were allowed to equilibrate for 30 min before the protocols were started.
Five groups of hearts were studied. All hearts were subjected to 30 min of regional ischemia and 120 min of reperfusion. Control hearts received no treatment. EDHB strongly reduced coronary flow and left ventricular developed pressure. Because we could not exclude the possibility that EDHB may have ischemically preconditioned the hearts, we decided not to complete investigations of this agent in intact hearts. Fortunately, DFO did not have these complicating hemodynamic effects, so a second group of hearts was treated with 1 mM DFO for 10 min followed by 5 min of washout before the 30-min coronary occlusion. Two additional groups were also treated as described above with DFO, but, in addition, 2 μM ODQ or 300 μM MPG was added to the perfusate for 20 min starting 5 min before and ending 5 min after DFO treatment. These latter protocols allowed an additional 5 min of washout before coronary occlusion. Finally, in the fifth group, DFO was infused as described above. l-NAME (200 μM) was coinfused starting 5 min before DFO and continuing until the onset of the coronary occlusion. Thus, in this group, there was no washout period.
Infarct size measurement.
As previously reported (8), the coronary artery was reoccluded, and the risk zone was delineated as the tissue without fluorescence with 2- to 9-μm-diameter green fluorescent microspheres (Duke Scientific, Palo Alto, CA) injected into the perfusate. The hearts were weighed, frozen, and then cut into 2-mm-thick slices, which were incubated in 1% triphenyltetrazolium chloride in sodium phosphate buffer (pH 7.4) at 37°C for 8 min. Triphenyltetrazolium chloride stains noninfarcted myocardium brick red. The slices were immersed in 10% formalin to preserve the tissue. The areas of infarct and risk zone for each slice were determined by planimetry, and volumes were calculated by multiplying each area by slice thickness and summing them for each heart. Infarct size is expressed as a percentage of the risk zone.
All drugs required for cell isolation and culture were purchased from Sigma Chemical (St. Louis, MO). Reduced MitoTracker Red was purchased from Molecular Probes (Eugene, OR); EDHB, DFO, Gli, 5-HD, MPG, and Myxo from Sigma; l-NAME, ODQ, Rp-8-Br-cGMPS, and ETU from Alexis Biochemicals (San Diego, CA); and Akt inhibitor from Calbiochem (San Diego, CA). Distilled water, DMSO, or ethanol was used to dissolve the drugs and prepare stock solutions. The final DMSO and ethanol concentrations were <2%.
Fluorescence measurements provide data in arbitrary units. To remove the variability caused by different lots of MitoTracker Red, cell age, and environmental conditions, average cell fluorescence was calculated and compared with that of simultaneously studied control cells as described above. Therefore, fluorescence data for each experiment are provided as a percentage of the control (mean ± SE). To minimize further the possible influence of these variables on the data, ANOVA for repeated measures with Tukey's post hoc test was used to evaluate differences in mean fluorescence of the groups within the same experiment. Similarly, differences in nitrite production among groups within the same experiment were analyzed by ANOVA for repeated measures with Tukey's post hoc test. Baseline hemodynamic variables and risk zone and infarct size data among groups were compared by one-way ANOVA with Tukey's post hoc test. Changes in serial measurements of hemodynamics for any given group were analyzed by ANOVA for repeated measures with Tukey's post hoc test. P < 0.05 was considered significant.
Attenuation of ROS generation.
Exposing cardiomyocytes to DFO or EDHB led to a robust and highly significant increase in fluorescence that reflected enhanced ROS production (53 ± 12 and 71 ± 16%, respectively, P < 0.001 vs. control cells for both; Fig. 1). To confirm that the increase in MitoTracker Red fluorescence induced by DFO was caused by free radical production, cells were coincubated with the free radical scavenger MPG (Fig. 1A). MPG abolished the increase in fluorescence [3 ± 6%, P = not significant (NS)], whereas MPG alone had little effect. Blockade of the mitochondrial respiratory chain at site III with the selective blocker Myxo prevented EDHB's effect on ROS generation (4 ± 8%, P = NS), suggesting that the mitochondrial respiratory chain was the source of the free radicals (Fig. 1B). Myxo alone had no effect on ROS generation.
KATP channel blockade.
To test whether EDHB-induced ROS generation was dependent on KATP channel opening, two different KATP channel antagonists were used. As expected, EDHB increased ROS production by 50 ± 12% (P < 0.001). Gli, a nonselective closer of KATP channels, prevented the effect of EDHB on ROS generation (Fig. 2). Selective blockade of mitoKATP channels with 5-HD produced similar results. Neither Gli nor 5-HD had an independent influence on ROS generation. Exposure of cardiomyocytes to DFO caused a robust increase in ROS production (57 ± 9%, P < 0.001) that could again be blocked by coincubation with 5-HD (−1 ± 6%, P = NS; Fig. 3).
When cells were coincubated with EDHB and the NOS antagonist l-NAME, EDHB-triggered ROS production was abrogated: 50 ± 6% (P < 0.001) for EDHB vs. −5 ± 6% (P = NS) for EDHB + l-NAME (Fig. 4). l-NAME alone had no impact on ROS production.
Soluble guanylyl cyclase.
To determine whether the NO-sensitive soluble guanylyl cyclase was involved in the signaling cascade, we coincubated cells with ODQ. EDHB and DFO significantly increased ROS production (75 ± 19 and 53 ± 12%, respectively, P < 0.001 for both). Inhibition of the NO-sensitive guanylyl cyclase with ODQ completely abolished the increase in ROS production: 2 ± 13% (P = NS) for EDHB + ODQ and 0 ± 6% (P = NS) for DFO + ODQ (Fig. 5). As demonstrated previously (29), ODQ alone had no effect on ROS generation.
Inhibition of PKG by the cell-permeant PKG inhibitor Rp-8-Br-cGMPS blocked EDHB-induced ROS generation in cardiomyocytes (Fig. 6). The inhibitor alone had no influence on fluorescence.
PI3-kinase and Akt.
The preconditioning mimetics acetylcholine and bradykinin signal through NO to open mitoKATP channels and produce ROS, and PI3-kinase and Akt are upstream of NO in this pathway (18). Because the increased ROS generation stimulated by EDHB depends on NO signaling, we wanted to see if PI3-kinase or Akt might also be involved. The selective PI3-kinase blocker Wort had no effect on EDHB-induced ROS generation: 36 ± 4 and 39 ± 5% for EDHB and EDHB + Wort, respectively (P < 0.001 for both; Fig. 7A). Inhibition of Akt also did not block EDHB-induced ROS production: 71 ± 16 and 74 ± 13% for EDHB and EDHB + Akt inhibitor, respectively (P < 0.001 for both; Fig. 7B). Neither Akt inhibitor nor Wort alone had an effect on ROS production.
NO (nitrite) production.
DFO increased nitrite production by isolated cardiomyocytes by 30% compared with untreated cells: 141.7 ± 5.8 vs. 109.0 ± 4.5 nM (P < 0.0.001; Fig. 8). This effect was completely blocked by ETU, an NOS inhibitor. ETU alone had no effect. This was a little surprising, because we expected ETU to block basal NOS activity. However, the medium itself obviously contained nitrite (111.5 ± 8.0 nM), thus possibly impeding the detection of any basal NOS activity.
Isolated Rabbit Hearts
Only DFO could be thoroughly tested in isolated hearts (see methods). DFO caused a very minor, but significant, decrease in coronary flow and left ventricular developed pressure (LVDP; Table 1). It is unlikely that this would contribute to protection. Prior studies in our laboratory showed that such changes in hemodynamics do not result in cardioprotection (38). Coinfusion of ODQ further reduced coronary flow and LVDP. A reduction of coronary flow and LVDP could theoretically ischemically precondition the heart and decrease infarct size. As noted below, however, ODQ, which alone does not have an effect on infarct size, blocked DFO-induced protection. Thus the changes in hemodynamics did not precondition the heart.
There were no differences in risk zone volume in the five groups (Table 2). Infarction in control hearts averaged 32.8 ± 1.2% of the risk zone (Fig. 9). In the few hearts examined, EDHB salvaged ischemic myocardium, but the adverse hemodynamic profile of severely depressed coronary flow and LVDP complicated interpretation of the results, causing us to abandon the group. DFO reduced infarct size to 10.3 ± 2.4% (P < 0.001 vs. control). Coadministration of the guanylyl cyclase inhibitor ODQ or the NOS blocker l-NAME aborted protection (33.2 ± 5.5 and 30.5 ± 2.8% infarction, respectively). Similarly, the protective effect of DFO was blocked by MPG (35.9 ± 5.6% infarction, P = NS vs. control). We previously reported that neither ODQ alone (29), l-NAME (24), nor MPG (24, 46) had a significant effect on infarct size in this model. Absolute risk zone size can be an independent determinant of the percent infarction of the risk zone of the rabbit heart (45). To ensure that the differences in infarction were not influenced by different risk zone volumes among the groups, the size of the risk zone was plotted against infarct volume for all hearts. Whereas regression lines for the control, DFO + ODQ, DFO + l-NAME, and DFO + MPG groups were not different, covariate analysis revealed that the regression for hearts treated with DFO alone was significantly shifted toward smaller infarcts for any risk zone size (data not shown).
This study demonstrates for the first time that EDHB and DFO activate pathways in cardiomyocytes that are identical to those that are known to trigger the preconditioned state. As previously reported for bradykinin and acetylcholine, EDHB opened mitoKATP channels by activating the NOS-guanylyl cyclase-PKG cascade. MitoKATP channel opening directly leads to ROS formation by mitochondria, because it causes alkalinization of the matrix (12), and ROS are critical components of protection after preconditioning (4, 8, 39) as well as DFO treatment (9). This ROS production is all the more surprising because DFO has received much attention as an antioxidant and would have been expected to suppress, rather than promote, ROS generation. The actual mechanism by which EDHB and DFO activated NOS is unknown, but certainly DFO increased NO production by 30%, and this effect was blocked by ETU, an NOS inhibitor. As previously suggested (6), NO plays a key role in several forms of preconditioning and cardioprotection.
We were prompted to investigate these compounds after Dendorfer et al. (9) reported that DFO pretreatment could reduce infarct size by an ROS-dependent mechanism. Because EDHB caused a severe reduction of coronary flow, we believed that it would be impossible to separate any potential independent protective effect of EDHB from that of associated ischemia-related preconditioning. Despite the myocardial salvage seen with EDHB in several hearts, this conundrum forced us to abandon further investigation of this agent in intact hearts. It would appear that the coronary constriction caused by EDHB is unrelated to the agent's PHD inhibition, because DFO alone had no marked effect on coronary perfusion.
In the present study, DFO caused a reduction of infarct size comparable to that seen with ischemic preconditioning, thus confirming the report by Dendorfer et al. (9). After completing the cardiomyocyte studies, we predicted that DFO's protection would be dependent on NO and cGMP, and this prediction was confirmed when l-NAME and ODQ blocked DFO's anti-infarct effect. NO activates guanylyl cyclase to produce cGMP, which in turn stimulates PKG. Direct activation of PKG leads to an increase in ROS generation, an effect that is dependent on mitoKATP channel opening (29). Thus our observation implicates the involvement of the NO-PKG cascade in this cardioprotection and also argues against DFO's protection being the result simply of an antioxidant effect. This conclusion is further supported by the effective blockade of DFO's protective effect by MPG, an exogenous ROS scavenger. Because DFO is already available for human use, it is an attractive candidate for clinical trials. In a recent trial, DFO was infused before and during coronary artery bypass grafting (31). DFO significantly increased left ventricular ejection fraction and improved the wall motion score. It was assumed that the beneficial effect was caused by suppression of oxygen free radicals by DFO (31); however, the present results would suggest a very different mechanism.
In the present study and that of Dendorfer et al. (9), hearts were reperfused for only 2 h before infarct size was measured. Reddy et al. (35) also pretreated canine hearts with DFO and reported a limitation of infarct size. That protection could not be achieved when DFO was given at reperfusion, again supporting a preconditioning mechanism. In a subsequent study, however, these investigators extended the reperfusion time to 24 h and failed to see protection from DFO, suggesting that DFO-induced protection may not be sustained (36). On the other hand, protection from ischemic preconditioning is known to be long-lasting (23), and it would appear that DFO triggers the same pathway used by preconditioning. Perhaps discrepancies are related to the very different doses of DFO in the three studies. Whereas Reddy et al. treated their dogs with only 5 mg/kg infused intravenously over 30 min before ischemia, Dendorfer et al. administered an intraperitoneal bolus of 200 mg/kg to their rats. For our rabbits, the DFO concentration in the buffer was 1 mM (656 mg/l), again a much higher concentration than would have been realized by Reddy et al.
HIF is a transcription factor leading to induction of target genes such as vascular endothelial growth factor (15), but the resulting GLUT-1, heme oxygenase 1, and NOS proteins (44) appear no earlier than 1 h after pharmacological stimulation. The shortest interval between DFO treatment and the onset of myocardial ischemia studied by Dendorfer et al. (9) was 2 h, so it is not so easy to eliminate the possible involvement of a gene product in DFO's infarct-sparing effect that they documented, even though elevated message to any of HIF's known gene targets was not observed. Interestingly, there was an increase in HIF-2α mRNA 4 h after DFO treatment. In the present study, ROS generation in cardiomyocytes was seen just 15 min after application of the PHD inhibitors, making it very unlikely that HIF or any of its target gene products was involved in this process.
Although EDHB (44) and DFO (40) are PHD inhibitors, it was not possible to test whether the effect described here was specifically due to PHD inhibition. The mechanism could just as easily be related to their activity as iron chelators (41), although this property has been implicated as the cause of their PHD inhibition, because the oxygen-dependent hydroxylation by PHD requires iron as a cofactor (19).
Hence, it is possible that EDHB and DFO may protect ischemic hearts by a nonenzymatic mechanism(s) that is not dependent on PHD blockade. Although, originally, our intention was to study the effect of DFO, we subsequently selected EDHB for collateral studies, precisely because it too was a known PHD inhibitor. It is curious that two PHD inhibitors trigger similar intracellular signaling that is likely to be independent of HIF and that at least one of these PHD inhibitors has cardioprotective properties. Because we did not measure activation of HIF in this study, we cannot unequivocally exclude its participation. However, it is very possible that the PHD-blocking action of DFO and EDHB is merely an epiphenomenon with little bearing on the mechanism of cardioprotection.
In summary, two chemically dissimilar PHD inhibitors, EDHB and DFO, were found to activate a well-documented cardioprotective pathway in cardiomyocytes. These agents triggered the intracellular signaling pathway associated with preconditioning that involves production of NO and cGMP, activation of PKG, opening of mitoKATP channels, and ROS production. As predicted, DFO also showed a powerful anti-infarct effect, and that protection was dependent on NOS and guanylyl cyclase.
This study was supported in part by National Heart, Lung, and Blood Institute Grants HL-20648 and HL-50688. S. Philipp was supported by a grant from the Deutsches Institut für Bluthochdruckforschung (Heidelberg, Germany).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by the American Physiological Society