INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

HEME IS ENZYMATICALLY DEGRADED to carbon monoxide, bilirubin, and iron, three important molecules that have attracted great interest because of their possible role in modulating physiological functions. Heme oxygenase, existing in constitutive (HO-2) and inducible isoforms (HO-1), is the rate-limiting step in the oxidative cleavage of heme. Whereas HO-2 is discretely localized in the nervous and vascular tissue, HO-1 can be expressed in essentially every tissue upon appropriate stimulation (1, 11, 15). In fact, HO-1 is very sensitive to upregulation by a variety of stress mediators. Because of this sensitivity and because HO-1 induction has been repeatedly associated with protection against cellular injury (2, 19, 20, 23, 33), it is believed that HO-1 is a fundamental endogenous defensive system. One aspect of this defense seems to be strictly connected to the antioxidant properties of bilirubin. At micromolar concentrations, bilirubin can efficiently scavenge peroxyl radicals in vitro, and its antioxidant capacity is further enhanced at low oxygen tension (27). Other intracellular effects of bilirubin that may become relevant in pathophysiological situations include an inhibitory action on protein kinase C activity (26) and inhibition of superoxide production by activated NADPH oxidase (14). Few studies, however, have examined whether bilirubin is actually increased when HO-1 is upregulated and whether the functional consequences of enhanced bilirubin generation on cellular activities. We (7, 12) have recently reported that HO-1-derived bilirubin is cytoprotective against oxidative stress in vitro; in particular, maximal defense was obtained when bilirubin was actively produced by HO-1 at the time of oxidative challenge (7). Furthermore, our results have established an important role for exogenous and endogenous bilirubin in the amelioration of postischemic myocardial dysfunction in the isolated perfused rat heart (8). Because oxidative stress plays a pivotal part in the progression of cell damage during ischemia-reperfusion events, we wanted to extend our previous findings by analyzing more closely the mechanisms involved in the protection afforded by HO-1. In particular, we concentrated our studies on the effect of HO-1-derived bilirubin and on the critical role of heme availability in HO-1-mediated cellular defense. For this purpose, we utilized an in vitro system of rat cardiomyocytes subjected to hypoxia-reoxygenation.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reagents. Hemin and tin protoporphyrin IX (SnPP) were purchased from Porphyrin Products (Logan, UT). Dichlorofluorescein diacetate was from Molecular Probes (distributor: Cambridge Bioscience; Cambridge, UK). Bilirubin and all other reagents were obtained from Sigma.

In vitro model of hypoxia-reoxygenation. H9c2 rat cardiomyocytes were purchased from American Type Culture Collection (Manassas, VA), cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 4 mM L-glutamine, 100 U/ml penicillin, and 0.1 mg/ml streptomycin. Cardiomyocytes at confluence were transferred to an air-tight chamber and flushed with a mixture of 95% N₂-5% CO₂. The gas was infused continuously into the air-tight chamber at a flow rate of 5 l/min for the first 2 h and at 1 l/min for the following hours of incubation, as previously described (18). Within the hypoxia chamber, cells were maintained in a humidified atmosphere at 37°C. In some experiments, cells were exposed to 18 h of hypoxia and then transferred to normoxic conditions (5% CO₂ and air) for reoxygenation studies.

Heme oxygenase activity assay in cardiomyocytes. Cardiomyocytes were exposed to hypoxia for 6, 12, 18, and 24 h or reoxygenation for various periods of time after hypoxia (18 h). Heme oxygenase activity was also measured in cardiomyocytes exposed to 18 h of hypoxia and reoxygenation in the presence of 5 µM hemin, 0.5 µM bilirubin, or 3 µM SnPP (an inhibitor of heme oxygenase enzymatic activity). Heme oxygenase activity was determined at the end of each treatment as previously described by our group (10, 18, 20). Briefly, microsomes from harvested cells were added to a reaction mixture containing NADPH, rat liver cytosol as a source of biliverdin reductase, and the substrate hemin. The reaction was conducted at 37°C in the dark for 1 h and terminated by the addition of 1 ml of chloroform, and the extracted bilirubin was calculated by the difference in absorbance between 464 and 530 nm ( = 40 mM¹ cm¹).

Determination of bilirubin release in culture medium. Cells were exposed to 18 h of hypoxia alone or hypoxia in the presence of 5 µM hemin or 3 µM SnPP. In additional experiments, cardiomyocytes were subjected to hypoxia-reoxygenation in the absence or presence of 5 µM hemin, 0.5 µM bilirubin, or 3 µM SnPP. At the end of the incubation, heme oxygenase-derived bilirubin was determined in the culture medium as described recently by Turcanu et al. (28) and by our group (12).

Western blot analysis for HO-1. Samples of cardiomyocytes treated for the heme oxygenase activity assay were also analyzed by Western immunoblot technique as described previously (10). Briefly, an equal amount of proteins (30 µg) for each sample was separated by SDS-PAGE and transferred overnight to nitrocellulose membranes, and the nonspecific binding of antibodies was blocked with 3% nonfat dried milk in phosphate-buffered saline. Membranes were then probed with a polyclonal rabbit anti-HO-1 antibody (Stressgen; Victoria, Canada) (1:1,000 dilution in Tris-buffered saline, pH 7.4) for 2 h at room temperature. After three washes with phophate-buffered saline, blots were visualized using an amplified alkaline phosphatase kit from Sigma (Extra-3A), and the relative density of bands was analyzed by an imaging densitometer (model GS-700, Bio-Rad, Herts, UK). Blots shown are representative of three independent experiments.

RNA extraction and Northern blot analysis. Cardiomyocytes were subjected to hypoxia, hypoxia in the presence of 5 µM hemin, or hypoxia-reoxygenation, and total RNA was isolated by phenol-chloroform using the method described by Chomczynski and Sacchi (5). Total RNA was run on a 1.3% denaturing agarose gel containing 2.2 M formaldehyde and transferred onto a nylon membrane. The membrane was hybridized using [-³²P]dCTP-labeled cDNA probes to rat HO-1 gene as previously described (30, 31), and staining of the 18S rRNA band (or rat GAPDH gene) was used to confirm integrity and equal loading of RNA. The hybridized membrane was exposed to radiographic film, and bands were analyzed using an imaging densitometer. Blots shown are representative of three independent experiments.

Assessment of cell viability. Trypan blue uptake and a colorimetric assay kit from Serotec (Oxford, UK) were used to investigate cell viability as already described (7, 20). Cardiomyocytes were exposed to hypoxia, hypoxia in the presence of 5 µM hemin, or hypoxia-reoxygenation. Damage was also evaluated in cells subjected to hypoxia-reoxygenation in the presence of 0.5, 1, and 5 µM bilirubin or 3 µM SnPP.

Measurement of intracellular reactive oxygen species generation. The production of reactive oxygen species (ROS) was monitored by measuring changes in fluorescence resulting from oxidation of an intracellular probe. Dichlorofluorescein diacetate enters the cells, and the acetate group is cleaved by cellular esterases, trapping the dichlorofluorescein inside; the increase in fluorescence of dichlorofluorescein is indicative of ROS formation (4, 29). Cardiomyocytes were exposed to 18 h of hypoxia alone or 18 h of hypoxia in the presence of 5 µM hemin, 0.5 µM bilirubin, or 3 µM SnPP. Dichlorofluorescein diacetate (10 µM) was present for the entire incubation period in both control and treated cells. At the end of each treatment, cells were trypsinized and centrifuged (500 g for 5 min), and the pellet was resuspended in 300 µl of reading buffer (2% fetal calf serum, 0.02% sodium azide, and 1 mM EDTA in phosphate buffer). Fluorescence was measured using a FacsScan flow cytometer (excitation wavelength = 488 nm; Becton Dickinson). Because porphyrins demonstrate a high degree of electron resonance that contribute to their colorful and fluorescent properties, we confirmed in preliminary experiments that bilirubin, hemin, and SnPP do not affect the accuracy of the assay (data not shown).

Statistical analysis. Differences in the data among the groups were analyzed by using one-way analysis of variance combined with the Bonferroni test. Values were expressed as means ± SE, and differences between the groups were considered to be significant at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Time course of HO-1 expression and cell viability after hypoxia-reoxygenation in rat cardiomyocytes. We first examined whether changes in oxygen tension affected the pattern of HO-1 expression and heme oxygenase activity in rat cardiomyocytes. As shown in Fig. 1A, hypoxia caused a time-dependent increase in heme oxygenase activity, with maximal levels observed at 18-24 h. This effect was accompanied by augmented expression of HO-1 protein and upregulation of the HO-1 gene, as determined by Western and Northern blot technique, respectively (Fig. 1, B and C). It has been demonstrated that ROS released from mitochondria during brief hypoxia activate signaling pathways involved in protection of cardiomyocytes against subsequent ischemia-reperfusion injury (preconditioning phenomenon) (29). Heme oxygenase could be one of the systems involved in this defense, because our group (8) reported that HO-1 induction ameliorates postischemic myocardial dysfunction in the isolated perfused rat heart. Therefore, we performed experiments to determine whether mitochondrial-derived ROS are possible triggers of hypoxia-mediated HO-1 expression. We found that heme oxygenase activity was still elevated in cells exposed to low oxygen tension in the presence of the mitochondrial electron transport inhibitors myxothiazol (1 µM) and rotenone (7.6 µM), which have been shown to abolish ROS generation during hypoxia and prevent the preconditioning effect (29) (data not shown). These results indicate that ROS produced by the mitochondria are not likely to play a major role in the events leading to HO-1 induction in our experimental setting. We were interested to know whether high HO-1 expression could be maintained during the reoxygenation event, because previous studies (16) reported that induction of HO-1 is evident as early as 30 min after reperfusion in a model of isolated rat hearts. Interestingly, transfer of cardiomyocytes after 18 h of hypoxia to normoxic conditions (reoxygenation) resulted in a marked decrease in heme oxygenase activity and HO-1 protein over time, with a return to basal levels at 24 h reoxygenation (Fig. 2, A and B). Assessment of viability by means of a colorimetric assay indicated that no apparent damage was present in cardiomyocytes exposed to 18 h of hypoxia compared with control cardiomyocytes (Fig. 3). Elevated cell injury was, however, observed at 3 and 6 h of reoxygenation, in line with previous data showing that cell death is primarily manifested during reperfusion but not after ischemia in cultured cardiomyocytes (29). At 24 h reoxygenation there was a recovery in cell viability, possibly because by the time the measurements were taken the surviving cardiomyocytes had started to replicate. Similar results were obtained with trypan blue uptake [a method to determine viability usually associated with necrotic cell death (3)]. Although no significant increase in cell death was observed after a period of 18 h hypoxia, a considerable number of cardiomyocytes died at 3 and 6 h of reoxygenation (Table 1).

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Effect of severe hypoxia on heme oxygenase-1 (HO-1) expression in cultured cardiomyocytes. A: cells were exposed for 6, 12, 18, and 24 h to hypoxia, and heme oxygenase activity was measured at the end of each time point. The results were compared with untreated cells (0 h hypoxia); each bar is the mean ± SE of n = 6 experiments. *P < 0.05 vs. 0 h hypoxia. B and C: cardiomyocytes were incubated in hypoxic conditions as in A. Representative Western (B) and Northern blots (C) for HO-1 protein and HO-1 mRNA expression are shown. The 18 S band is shown to confirm integrity and equal loading of RNA.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Effect of reoxygenation on HO-1 expression in cardiomyocytes. After exposure to hypoxia for 18 h to allow strong upregulation of HO-1 (0 h reoxygenation), cells were transferred to normoxic conditions for reoxygenation studies. A: heme oxygenase activity was determined at various time points and each bar is the mean ± SE of n = 6 experiments. *P < 0.05 vs. 0 h reoxygenation. B: representative Western blot of HO-1 protein expression at various times of reoxygenation.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Viability of cardiomyocytes subjected to hypoxia-reoxygenation. Cells were exposed to hypoxia for 18 h, followed by different times of reoxygenation. Cell viability was assessed using a commercial kit (see MATERIALS AND METHODS) according to manufacturers instructions. Each bar is the mean ± SE of n = 6 experiments. *P < 0.05 vs. control (CON).

Bilirubin production is elevated in cardiomyocytes exposed to hypoxia in the presence of hemin. Because hypoxia stimulated the expression of HO-1 in cardiomyocytes, we wanted to establish whether increased heme oxygenase activity was accompanied by augmented levels of bilirubin, the end product of heme degradation by heme oxygenase. We indeed detected a small, but significant (P < 0.05), elevation in bilirubin in the medium of cells incubated under hypoxic conditions (Fig. 4D). It is noteworthy that heme oxygenase activity showed a 3.8-fold increase after 18 h of hypoxia (Fig. 1A and 4C), whereas bilirubin only exhibited a 1.4-fold increment. These data suggest that the availability of the substrate heme could be a limiting factor in the production of bilirubin by heme oxygenase. To verify this hypothesis, we incubated cardiomyocytes under hypoxia in the presence of a low concentration of hemin (5 µM) to continuously provide substrate for heme oxygenase activity. This treatment potentiated the upregulation of HO-1 gene and protein expression, as well as heme oxygenase activity compared with hypoxia alone (Fig. 4, A-C). Notably, bilirubin accumulated in the culture supernatant was also considerably augmented (Fig. 4D).

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effect of hemin on hypoxia-mediated HO-1 induction and bilirubin production in cardiomyocytes. Cells were incubated for 18 h in complete medium alone (CON), hypoxia (Hyp), normoxia in the presence of 5 µM hemin (Hemin), or hypoxia in the presence of 5 µM hemin (Hyp + Hemin). After the treatments, cardiomyocytes were analyzed for HO-1 protein (A) and HO-1 mRNA (B) expression. GAPDH is shown to confirm integrity and equal loading of RNA. C: cardiac heme oxygenase activity. D: bilirubin released in the culture medium determined at the end of the treatments described above. Each bar is the mean ± SE of n = 6 experiments. *P < 0.05 vs. CON; P < 0.05 vs. Hyp or Hemin alone.

HO-1 induction or exogenous bilirubin protect cardiomyocytes against the reoxygenation damage. To determine a possible involvement of HO-1 in protection against reoxygenation injury, untreated cells were transferred to normoxic conditions after 18 h of hypoxia, a time point when high HO-1 and heme oxygenase activity levels were detected. Despite induction of HO-1 during hypoxia, cardiomyocytes still displayed a pronounced damage upon reoxygenation (see Fig. 3). We postulated that greater protection could not be achieved because of the limited generation of heme catabolites, namely iron, bilirubin, and carbon monoxide, which may possess cytoprotective properties. If our hypotheses were correct, cardiomyocytes stimulated to produce higher levels of bilirubin (and therefore also the other products of heme degradation by heme oxygenase) during hypoxia should exhibit improved ability to counteract the reoxygenation injury. Indeed, we found that when cardiomyocytes exposed to hypoxic conditions in the presence of 5 µM hemin were subjected to the reoxygenation event, cell viability was almost completely preserved (Fig. 5 and Table 1). To analyze a potential contribution of the antioxidant bilirubin to this effect, cells were also incubated for 18 h at low oxygen tension with 0.5, 1, and 5 µM bilirubin. Interestingly, protection against reoxygenation damage was only observed with 0.5 µM (Fig. 5 and Table 1), but not with 1 or 5 µM bilirubin (data not shown). In addition, if bilirubin (0.5 µM) was added at the onset of reoxygenation, cardiomyocytes were still susceptible to reoxygenation-mediated injury. Specifically, at 3 h of reoxygenation, cell viability was 68% of control in untreated cells compared with 69% in bilirubin-treated cells. It has to be noted that the extent of protection afforded by 0.5 µM bilirubin was smaller compared with that obtained in hemin-treated cells (Fig. 5 and Table 1).

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Viability of cardiomyocytes subjected to hypoxia-reoxygenation in the presence of hemin or exogenous bilirubin. Cells were subjected to 18 h of hypoxia, followed by reoxygenation for different times in the presence of 5 µM hemin or 0.5 µM bilirubin. Untreated cells were exposed to hypoxia-reoxygenation in complete medium alone. Each bar is the mean ± SE of n = 6 experiments. *P < 0.05 vs. 0 h reoxygenation; ** and P < 0.05 vs. untreated.

Effect of hemin treatment on heme oxygenase activity and bilirubin production during reoxygenation. Having established that incubation of cardiomyocytes under hypoxic conditions in the presence of hemin or 0.5 µM bilirubin considerably reduced the reoxygenation damage, we wondered whether these treatments had any effect on the levels of heme oxygenase activity or bilirubin production during reoxygenation. This was an important point to take into consideration, because we had already observed that cells exposed to 18 h of hypoxia exhibited a rapid decrease in heme oxygenase activity and HO-1 expression during reoxygenation (see Fig. 2, A and B). Of interest, we found that elevated heme oxygenase activity could be maintained during reoxygenation if cardiomyocytes were subjected to hypoxia in the presence of 5 µM hemin (Fig. 6A). Likewise, bilirubin generation did not change at the time of reoxygenation in nontreated cells, but remained considerably higher in hemin-treated cells (Fig. 6B). At 24 h of reoxygenation, we detected an increase in bilirubin levels in both untreated and hemin-treated cells; this effect is most likely explained by the accumulation of the bile pigment into the culture medium over a long period of incubation. When exogenous bilirubin was applied, heme oxygenase activity during reoxygenation showed a decline similar to that of untreated cells (data not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Effect of hemin treatment on heme oxygenase activity and bilirubin production after hypoxia-reoxygenation. Cardiomyocytes were subjected to 18 h of hypoxia, followed by reoxygenation for different times in the presence of 5 µM hemin or complete medium alone. Control cells were exposed to normoxia for 18 h (CON). Heme oxygenase activity and bilirubin released into the culture medium are shown in A and B, respectively. Each bar is the mean ± SE of n = 6 experiments. *P < 0.05 vs. untreated; P < 0.05 vs. 0 h reoxygenation; **P < 0.05 vs. CON.

Enhancement of HO-1 expression or incubation with exogenous bilirubin reduces the generation of reactive oxygen species after hypoxia in cardiomyocytes. Oxidative stress is one of the stimuli leading to tissue HO-1 induction, and because heme oxygenase enzymic activity removes the prooxidant molecule heme and generates the free radical scavengers biliverdin and bilirubin, HO-1 is usually regarded as an antioxidant-inducible cellular defense. However, a direct demonstration that high HO-1 levels or exogenously applied bilirubin are associated with reduced production of ROS is still lacking in cardiac tissue. We investigated this possibility by employing a dye that fluoresces upon reaction with intracellular oxygen species (see MATERIALS AND METHODS). As shown in Fig. 7, cardiomyocytes exposed to 18 h of hypoxia exhibited increased fluorescence compared with control cells, indicating that there was augmented formation of ROS under low oxygen tension. Interestingly, treatment of cardiomyocytes with 5 µM hemin during hypoxia considerably diminished ROS production (Fig. 7), and a similar reduction was observed in the presence of 0.5 µM exogenous bilirubin. Thus increased HO-1-derived products via hemin stimulation or application of bilirubin are effective ways to lower oxidant stress in cells subjected to hypoxia-reoxygenation.

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Hemin and bilirubin decrease reactive oxygen species generation during hypoxia. Cardiomyocytes were subjected to 18 h hypoxia in the presence of complete medium alone (Hyp), 5 µM hemin (Hyp + hemin), 0.5 µM bilirubin (BR), or 3 µM tin protoporphyrin IX (Hyp + SnPP). Control cells were incubated in normoxic conditions for 18 h (CON). Fluorescent dye dichlorofluorescein diacetate was used to determine the production of reactive oxygen species and fluorescence intensity was measured by a FacsScan flow cytometer (excitation wavelength = 488 nm). Each bar is the mean ± SE of n = 3 independent experiments. *P < 0.05 vs. CON; P < 0.05 vs. Hyp.

Effect of SnPP on heme oxygenase activity, bilirubin production, and cell viability during hypoxia-reoxygenation. We conducted further experiments with the inhibitor of heme oxygenase activity SnPP. In preliminary studies, we determined that 3 µM SnPP was sufficient to inhibit the augmented heme oxygenase activity detected during hypoxia (Fig. 8A). Likewise, the slight increase in bilirubin production was abolished by this concentration of inhibitor (Fig. 8B). In contrast, when cardiomyocytes were incubated in hypoxic conditions in the presence of 5 µM hemin and 3 µM SnPP, the inhibitor was not able to completely block heme oxygenase activity or bilirubin generation (Fig. 8, A and B). These results indicate that SnPP acts efficiently as an inhibitor of heme oxygenase activity during hypoxia alone; its competitive action is, however, greatly diminished when cells are exposed to hypoxia in the presence of hemin. When cell viability and ROS formation were assessed in cardiomyocytes exposed to hypoxia-reoxygenation in the presence of 3 µM SnPP, we did not detect any difference compared with hypoxia-reoxygenation alone (data not shown and Fig. 7, respectively). The fact that SnPP did not exacerbate the reoxygenation-mediated damage or further increase ROS production, effects that might be expected following blockade of heme oxygenase activity, leads us to speculate that the contribution of HO-1 to protection during hypoxia-reoxygenation alone is minimal, possibly because the products of this pathway are not generated in sufficient amounts to counteract the stressful insult. An alternative explanation could derive from nonspecific properties of SnPP and other metalloporphyrins, which possess the ability to interact directly with oxidant molecules thereby preventing oxidative injury (12, 32). These antioxidant effects may readily mask the consequences of heme oxygenase activity inhibition in the context of certain experimental conditions.

View larger version (16K): [in this window] [in a new window]   Fig. 8.   Effect of SnPP on heme oxygenase activity and bilirubin production during hypoxia. Cardiomyocytes were exposed to 18 h hypoxia in the presence of the following: 1) complete medium alone (Hyp); 2) 3 µM SnPP (Hyp + SnPP); 3) 5 µM hemin; and 4) 5 µM hemin and 3 µM SnPP (Hyp + hemin + SnPP). Control cells were incubated in normoxic conditions for 18 h (CON). Heme oxygenase activity and bilirubin released into the culture medium are shown in A and B, respectively. Each bar is the mean ± SE of n = 6 experiments. *P < 0.05 vs. CON; P < 0.05 vs. Hyp + hemin + SnPP.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We show that rat cardiomyocytes stimulated to express HO-1 and produce high levels of the antioxidant bilirubin exhibit an increased resistance to reoxygenation damage following a period of sustained hypoxia. The mechanism(s) underlying this effect is partly explained by a considerable reduction in reactive oxygen species formed during hypoxia-reoxygenation. Our data reinforce the notion that HO-1 is a crucial inducible protective pathway and highlight the importance of at least one of the HO-1 products, bilirubin, as a strong contributor to this protection.

In a recent report, we (8) demonstrated that upregulation of HO-1 before ischemia ameliorates myocardial function and reduces infarct size on reperfusion of isolated rat hearts. This was associated with augmented endogenous bilirubin and could be mimicked by exogenous administration of very low concentrations of bilirubin (100 nmol/l). Interestingly, we also found that mitochondria, which are an important site of injury and constitute a major source of reactive oxygen species during reperfusion, were well preserved in both hearts overexpressing HO-1 and those treated with exogenous bilirubin. These results prompted us to examine how the hypoxia-reoxygenation event modulates HO-1 expression in cultured cardiomyocytes and the specific function of HO-1 and its product bilirubin in protection against reoxygenation damage. We first observed that severe hypoxia strongly upregulates cardiac HO-1 gene and protein, as well as heme oxygenase activity in a time-dependent manner. The pattern of expression is very similar to our previous studies using endothelial cells (18), although the extent of induction was smaller in cardiomyocytes. When cardiomyocytes were reoxygenated, HO-1 expression and heme oxygenase activity rapidly declined. This was possibly caused by the termination of the hypoxic stimulus; alternatively, because a pronounced damage and an increase in cell death were detected within the first 6 h of reoxygenation, HO-1 protein could be a target of the reoxygenation injury. Maulik et al. (16) have shown that HO-1 mRNA was not induced by ischemia, but its expression was considerably enhanced after 30 min of reperfusion in isolated rat hearts. The discrepancy between our and those results are likely explained by the different experimental models and the duration of the ischemic-hypoxic period.

HO-1 has been reputed as an essential enzyme for resistance to oxidative stress (25); however, the marked increase in HO-1 and heme oxygenase activity levels detected after hypoxia were not sufficient to prevent reoxygenation-mediated damage in our experiments. We reasoned that measuring bilirubin after hypoxia could give us a clearer indication of whether elevated HO-1 expression can be efficiently translated into enhanced HO-1-derived products. Interestingly, bilirubin released into the medium of cardiomyocytes subjected to hypoxia was only slightly increased compared with untreated cells. Therefore, HO-1 induction by low oxygen tension was not accompanied by complete protection against reoxygenation injury probably because of the limited production of bilirubin and other metabolites deriving from heme oxygenase enzymatic activity. These data also suggest that, although cardiomyocytes acquire the potential ability to produce bilirubin during hypoxia by virtue of high HO-1 and heme oxygenase activity, a limitation of the substrate heme could be responsible for the small amount of the bile pigment actually generated. This hypothesis was confirmed by the results obtained after exposing cells to hypoxia in the presence of low concentrations of hemin to sustain a continuous release of heme degradation products. Indeed, under these conditions, bilirubin was considerably elevated after hypoxia and during the reoxygenation period. Furthermore, and in contrast to untreated cells, the augmented heme oxygenase activity observed after hypoxia was maintained for the entire reoxygenation phase. When tested for damage, cardiomyocytes incubated in hypoxia in the presence of hemin exhibited a greater resistance to the reoxygenation insult both in terms of cell viability and number of dead cells. Thus, when HO-1-derived products are generated in sufficient amounts, cells can apparently defend themselves against injury caused by reoxygenation. We also acknowledge the possibility that the protection afforded by hemin is attributable to preservation of high HO-1 levels at reoxygenation.

Because bilirubin is an antioxidant, it is rational to consider that part of the protection is due to the increased production of the bile pigment in cardiomyocytes exposed to hypoxia in the presence of hemin. We consequently evaluated if exogenous bilirubin added during hypoxia could reproduce the same defensive effect. Three different concentrations (0.5, 1, and 5 µM) were used and only 0.5 µM bilirubin was able to protect. We speculate that concentrations of 1 or 5 µM bilirubin may be too high to be tolerated by cells for a long period of time (18 h hypoxia) and that the bile pigment exerts then its cytotoxic properties. It is interesting that the extent of protection obtained with exogenous bilirubin was inferior to that observed in cardiomyocytes subjected to hypoxia plus hemin; this indicates that other products of heme breakdown, namely carbon monoxide and presumably ferritin induction following iron release, may actively contribute to the increased resistance to damage. In line with this idea, recent evidence has demonstrated a role for exogenous carbon monoxide in attenuating hyperoxic lung injury (24) and HO-1 prevents cell death by controlling cellular iron efflux (9). The fact that 0.5 µM bilirubin added at the onset of reoxygenation did not improve cell viability suggests at least two plausible explanations. First, by adding it at reoxygenation, bilirubin did not have adequate time to reach the subcellular compartments (mitochondria and others) most targeted by the reoxygenation injury, because ROS generation is expected to occur within minutes (13) after transfer of cardiomyocytes from hypoxic to normoxic conditions (21% O₂). Second, the bile pigment confers protection not only because of its antioxidant capacities, but also because of other defensive mechanisms still to be identified but that require several hours to be activated. These findings confirm similar results we obtained in the isolated rat heart: exogenously applied bilirubin can ameliorate postischemic heart functions only when perfused through the system before the ischemic event, but not if administered at reperfusion (J. Clark and R. Motterlini, unpublished observations). From these evidences, we propose that bilirubin belongs to the category of antioxidants that need to be delivered before the reoxygenation-reperfusion events occur to exhibit beneficial effects.

ROS appear to be the main cause of cellular damage of reperfused tissue; therefore, various strategies aimed at reducing the reoxygenation damage have relied on increasing antioxidant defenses, either in exogenous or endogenous form. We hypothesized that HO-1 induction and exogenous bilirubin confer resistance to cardiomyocytes by diminishing the production of ROS typical of the reoxygenation event, thereby attenuating the initial and principal trigger of cytotoxicity. This point is sustained by the data collected using a fluorescent dye to detect ROS formation. We observed an enhanced ROS signal in cells subjected to hypoxia; this signal was, however, considerably decreased when cells were made hypoxic in the presence of hemin or bilirubin. The fact that hemin and exogenous bilirubin produced a similar reduction in the generation of ROS suggests that the antioxidant capacity of the HO-1 pathway relies solely on the free radical scavenging properties of bilirubin (and possibly biliverdin). On the other hand, the complete preservation of cell viability afforded by hemin against the reoxygenation damage substantiates the idea that other metabolites of the HO-1 pathway contribute to this defense via mechanisms independent from prevention of oxidative stress.

In summary, we have shown that hypoxia highly stimulates HO-1 expression in cardiomyocytes and that the HO-1 pathway can provide significant protection against reoxygenation-mediated cell damage. The reduction of oxidant stress by bilirubin during hypoxia-reoxygenation appears to be one of the mechanisms underlying this effect. Our present and previous (7) data also put a strong emphasis on the actual capability of cells to produce heme-derived products once HO-1 protein is upregulated. It is in fact evident that even following HO-1 induction and increased heme oxygenase activity, we should not always assume an enhancement of bilirubin and carbon monoxide levels, at least in a cell culture system where the substrate heme might be a limiting factor. With regard to hypoxia, decreased oxygen availability could also interfere with and hinder proper functioning of heme oxygenase, because the oxidation of heme is a complex reaction that consumes oxygen and requires several electrons (22). It is therefore intriguing that the enzyme has still a great capability to produce bilirubin when cells are exposed to very low oxygen tension for a long period of time in the presence of hemin, such as in the case of our study. Notably, biochemical studies have shown that hydrogen peroxide can substitute oxygen in the conversion of heme to verdoheme catalyzed by either HO-1 and HO-2 (21). Hence, the possibility that heme oxygenase activity could be supported by hydrogen peroxide during hypoxia cannot be ruled out, because this oxidant appears to be generated from superoxide metabolism by hypoxic cardiomyocytes (29). From these presented results, it is tempting to speculate that heme oxygenase might become one of the selected and favored pathways to operate at low oxygen tension, by virtue of the important roles its products may have in counteracting hypoxia-mediated cellular dysfunction (6, 8, 17).