Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Heme oxygenase (HO) controls the initial and rate-limiting step in heme catabolism. The enzyme cleaves heme to biliverdin, which is converted subsequently to bilirubin by biliverdin reductase. Iron is released when the heme ring is open and carbon monoxide is liberated. The heme molecule plays a central role in biological processes as the prosthetic moiety of hemoproteins involved in cell respiration, energy generation, oxidative biotransformation, growth-differentiation processes, and the generation of inflammatory mediators such as eicosanoids and nitric oxide. Three HO isozymes, the products of distinct genes, have been described (McCoubrey et al., 1997; Shibahara et al., 1993). HO-1, which is distributed ubiquitously in mammalian tissues, is induced strongly and rapidly by many compounds that elicit cell injury; the natural substrate of HO, heme, is itself a potent inducer of the enzyme (Yoshida et al., 1998). HO-2, which is suggested to be constitutively expressed, is present in high concentrations in such tissues as brain and testis and is thought to be noninducible (McCourbrey et al., 1992) and HO-3, which is not active in heme degradation. HO-1 activity is increased in whole animal tissues and in cultured cells following treatment with heme, metals, and hemodynamic forces, and in response to oxidative stress-inducing substances (Applegate et al., 1991; Otterbein et al., 1995; Wagner et al., 1997; Ishizaka et al., 1997). HO-1 also is induced by heat shock (Mitani et al., 1993) and the enzyme belongs to a class of macromolecules known as stress proteins, which are responsive to various types of acute cellular injuries (Shibahara et al., 1987; Dennery et al., 1996).

Induction of HO-1 is of considerable importance in the initiation of cellular protective mechanisms following exposure to hyperoxia (Dennery et al., 1996, 1997) and to various forms of cell-stressing stimuli (Stocker, 1990; Applegate et al., 1991; Vogt et al., 1995). This idea derives, in part, from the fact that increased HO activity enables the removal of heme, a lipid-soluble compound that is the transmissible form of the potent pro-oxidant iron, resulting in the generation of the heme metabolites bilirubin and biliverdin that have significant antioxidant and anticomplement properties (Stocker et al., 1987; Llesay and Tomato, 1994). Indeed, in vivo study showed that induction of HO-1 coupled to ferritin synthesis is a rapid, protective antioxidant response in rhabdomyolysis-induced kidney injury in the rat (Nath et al., 1992). Poss and Tonegawa (1997) also showed that mice deficient in the HO-1 gene were vulnerable to oxidative stress generated from heme and H₂O₂. Abraham et al. (1995) demonstrated that HO-1 gene overexpression via adenovirus-mediated HO-1 cDNA transfer protects coronary endothelial cells from oxidative injury produced by exposure to free heme/hemoglobin (Wagener et al., 1997).

Processes or agents that produce oxidative tissue injury and inflammation, such as heme and inflammatory cytokines, uniquely threaten the endothelial system (Stocker et al., 1987; Nath et al., 1992; Nakagami et al., 1993; Llesay and Tomato, 1994;Vogt et al., 1995; Poss and Tonegawa, 1997). This system is continuously exposed to circulating erythrocytes and to exogenous heme released by these cells after cellular damage. Heme, ubiquitous in nature, is an essential molecule for much biological function (for review, see Abraham et al., 1996). In addition, heme has multiple immunoregulatory actions. It stimulates cell proliferation and differentiation and it promotes angiogenesis (Lu and Broxmeyer, 1983; Novogrodsky et al., 1989; Abraham et al., 1996; Deramaudt et al., 1998), a phenomenon closely related to inflammatory reaction (Stolz et al., 1996). However, free heme may be deposited in tissues under pathological conditions and promote oxidative damage (Balla et al., 1993; Zager et al., 1995). In fact, pathological heme levels have been shown to occur in hemolytic diseases such as sickle cell anemia (Hebel et al., 1988). Moreover, heme may be released locally in extremely high concentrations after hemorrhage (>350 µM) as has been shown in experimentally induced subarachnoid hematoma in dogs (Letarte et al., 1993).

Wagener et al. (1997) demonstrated that heme stimulates the expression of HO activity, intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1, and endothelial leukocyte adhesion molecule (E-selection) in human endothelial cells. The in vivo significance of ICAM-1 expression on endothelium is supported by observations of their enhanced expression during atherogenesis, sickle cell disease, and vascular inflammatory disorders (Chernick et al., 1989; Wick et al., 1995; Moore et al., 1997). However, the link between HO isoforms and heme-mediated adhesion molecule expression in this model has not been established. This study was undertaken to examine the inducibility of HO-1 in endothelial cells and to assess the relation of this enzyme isoform to the heme degradation in human umbilical vein endothelial cells (HUVECs). The results of this study demonstrate that HUVECs express basal and inducible HO-1 and that an increase in HO-1 expression is associated with marked alleviation of heme-mediated adhesion molecule expression in endothelial cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents. The following inducers of HO were used: hemin and CoCl₂ (Shibahara et al., 1993; Letarte et al., 1993) (Sigma Chemical Co., St. Louis, MO). Tin (stannic) mesoporphyrin (SnMP) (Porphyrin Products, Inc., Logan, UT) was used as an inhibitor of HO activity (Abraham et al., 1995). All porphyrin solutions were freshly prepared as described in Chernick et al. (1989). CoCl₂ was dissolved in 1 M sodium citrate. N-acetylcysteine (NAC) and the glutathione monoethyl ester L--glutamylglycine ethyl ester (GSH-ET) were used as glutathione precursors.

Antibodies. For flow cytometry studies, REK-1, a monoclonal mouse antibody against human ICAM-1, which was a gift from Yvette van Kooyk (University Hospital, St. Radboud, Nijmegen, the Netherlands), was used. As a secondary antibody, a fluorescein isothiocyanate (FITC) fluorochrome conjugated to goat anti-mouse IgG (Amersham Corp., Arlington Heights, IL) was used. Human HO-1 antibodies were generated in our laboratory (Abraham et al., 1987), and HO-2 antibodies were obtained from StressGen Biotechnologies Corp. (Sidney, British Columbia, Canada).

Cell Culture and Incubations. HUVECs (second to eighth passage) (Clonetics Corp., San Diego, CA) were cultured in endothelial cell growth medium 2 containing 2% fetal bovine serum (EGM-2) (Clonetics Corp.) at 37°C and 5% CO₂ and used for the experiments described below.

Lactate Dehydrogenase (LDH) Activity Cytotoxicity Assay. Cell viability was analyzed with an in vitro LDH-based toxicology assay kit (KR430) as described by the manufacturer's protocols (Sigma Chemical Co., St. Louis, MO). Inhibition of endothelial cell growth, which resulted in a concomitant change in the amount of total cytoplasm LDH activity, indicated the degree of cytotoxicity. LDH activity-mediated formation of NADH was measured and analyzed with the conversion of a tetrazolium dye into a colored compound, which was measured with an enzyme-linked immunosorbent assay-plate reader at 490 and 690 nm.

Western Blot Analysis of HO-1 and HO-2. Cells were harvested with 1× reporter lysis buffer (Promega Biotec, Madison, WI) and supplemented with protease inhibitor cocktail (Sigma Chemical Co.) as described in Abraham et al. (1987). After 15 min at room temperature, the lysate was spun in a microcentrifuge at 10,000g for 5 min at 4°C. The supernatant was collected for Western blot or HO activity measurement. Protein levels were visualized by immunoblotting with polyclonal rabbit antiserum directed against human HO-1 or HO-2. Briefly, 50 µg of lysate supernatant was separated by SDS-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane (Amersham Pharmacia Biotech, Piscataway, NJ) with a semidry transfer apparatus (Bio-Rad Laboratories, Inc., Richmond, CA). The membranes were incubated with 5% milk in 10 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.05% Tris-buffered saline/Tween 20 buffer at 4°C overnight. After being washed with Tris-buffered/Tween 20 buffer, the membranes were incubated with a 1:2000 dilution of anti-HO-1 or anti-HO-2 antibodies for 1 h at room temperature with constant shaking. Then the filters were washed and subsequently probed with horseradish peroxidase-conjugated donkey anti-rabbit IgG (Amersham) at a dilution of 1:2000. Chemiluminescence detection was performed with the Amersham enhanced chemiluminescence detection kit according to the manufacturer's instructions.

HO Activity Assay. HO activity was assayed by preparing cell homogenate from HUVECs treated with heme, CoCl₂, SnMP, or a combination. Cell homogenate was incubated with 50 µM heme, 2 mg/ml rat liver cytosol (source of biliverdin reductase), 1 mM MgCl₂, 3 U of glucose 6-phosphate dehydrogenase, 1 mM glucose 6-phosphate, and 2 mM NADP⁺ in 0.5 ml of 0.1 M potassium phosphate buffer, (pH 7.4) for 30 min at 37°C. Placing the tubes on ice terminated the reaction, and bilirubin was extracted with chloroform as described in Chernick et al. (1989). The amount of bilirubin generated was determined by scanning spectrophotometer (Lambda 17 UV/VIS, Perkin-Elmer Cetus Instruments, Norwalk, CT) and was defined as the difference between 464 and 530 nm (extinction coefficient, 40 mM¹ · cm¹ for bilirubin) (Chernick et al., 1989). Results are expressed as picomoles of bilirubin per milligram of protein per hour.

Microsomal Heme Determinations. Microsomal heme was determined as the pyridine hemochromagen by the reduced minus oxidized difference spectrum between 400 and 500 nm with an absorption coefficient of 32.4 mM¹ · cm (Fuhrop et al., 1975).

HO-1 and HO-2 Sense and Antisense ODN Treatment. Cells at 50% confluence were washed three times with PBS and were cultured in medium containing 10% Nu-serum, which lacks nuclease activity (Collaborative Research Inc., Waltham, MA) for 5 h before addition of the ODNs. N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methylsulfate (DOTAP; Boehringer Mannheim Diagnostics, Indianapolis, IN) was used as a vehicle for transfecting cells with the ODNs. The proportions used were 2 µg ODN/1 µg DOTAP/ml of medium and the protocol used was as described by the manufacturer. The sense/antisense ODN for HO-1 and HO-2 were directed against the flanking translation initiation codon in the human HO-1 cDNA. The antisense were for HO-1 5'-CGCCTTCATGGTGCC-3' and for HO-2 5'-ATCATTCATCCTGCC-3', whereas the sense sequences for HO-1 5'-GGCACCATGAAGGCG-3' and for HO-2 sense 5'-ATCATTCATCCTGCC-3'. Each ODN was phosphorothioated on the first three bases on the 3' end and purified by HPLC. Cells were incubated for 24 h with the ODNs, and then the medium was replaced with fresh medium containing 10% fetal bovine serum and incubated for 24 h in the absence or presence of heme.

Flow Analysis for ICAM-1 Surface Expression on HUVECs. Confluent HUVEC monolayers were incubated with different agents. The cells were washed twice, trypsinized, transferred to an Eppendorf tube, and washed with cold PBS containing Ca²⁺/Mg²⁺/0.1% BSA/0.05% NaN₃, which is hereafter referred to as PBS-plus. The cells were incubated with ICAM-1 antibody in the appropriate dilution for 1 h at 4°C followed by washing with 1.5-ml of cold PBS-plus. The cells were then incubated with FITC-conjugated secondary antibodies in the recommended dilutions for 1 h at 4°C and washed twice, first with 1.5 ml of PBS-plus and then with 1.5 ml of PBS containing Ca²⁺/Mg²⁺/0.05% NaN₃. Finally, the cells were fixed in 200 µl of 1.5% paraformaldehyde in PBS and kept in the dark at 4°C until analyzed later that day. The cell preparations were analyzed with a Coulter Profile II (Coulter Immunology, Hialeah, FL) flow cytometer equipped with a 15-mW air-cooled argon laser emitting at 488 nm. During analysis, forward- and side-light scatter parameters were adjusted to enable gating of all cells. Five thousand cells were then analyzed for mean green (FITC) fluorescence intensity as a measure of ICAM-1 expression.

Statistical Analysis. Results are expressed as means ± S.E. Statistical analyses were performed with Student's two-tailed t test and ANOVA for the comparison of two treatments. P values < .05 were considered to be statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Flow Analysis of Time- and Concentration-Dependence of Heme-Induced Endothelial ICAM-1 Expression. Concentration and time dependence of heme-mediated ICAM-1 expression was analyzed by flow cytometry. As seen in Fig. 1, a concentration-dependent increase in ICAM-1 was found in HUVECs exposed to heme (10-100 µM) for all analyzed time points. Although ICAM-1 was constitutively expressed in HUVECs, a 5-fold increase was seen as early as 1 h after treatment with 100 µM heme. A gradual increase for ICAM-1 expression in time also was seen in cells exposed to lower concentrations of heme (10 and 50 µM). Twenty-four hours of incubation with increasing heme concentrations at 10, 50, and 100 µM resulted in an increase of ICAM-1 expression of 2.7-, 6.4-, and 12-fold, respectively, compared with untreated cells. In addition, high concentrations of heme (100 µM) were shown to cause cell injury and cell death as determined by LDH activity assay (48% ± 10% inhibition of LDH activity; unpublished data). Because 50 µM heme was shown to substantially induce ICAM-1 expression after 24 h and to cause no significant cell toxicity, this heme concentration was used in subsequent flow cytometry experiments.

View larger version (57K): [in this window] [in a new window]   Fig. 1.   Effects of heme on the induction of ICAM-1 in HUVECs as a function of heme concentration and time. Cells were treated with increasing concentrations of heme (0, 10, 50, and 100 µM) and analyzed for ICAM-1 expression at the indicated number of hours with flow cytometry as described in Materials and Methods. The y-axis indicates the mean fluorescence intensity of 5000 cells. Each bar represents the mean ± S.D. of three separate experiments. *P < .05 compared with control.

Modulation of HO Activity in HUVECs. To assess the effect of HO on heme-induced adhesion molecule expression, it was essential to evaluate the effect of inducers and inhibitors on HO activity. As seen in Fig. 2, incubation of HUVECs with CoCl₂ (50 µM) or heme (50 µM) resulted in a significant increase of cellular HO activity of 30 and 40%, respectively, compared with nontreated HUVEC (674.6 pmol of bilirubin/mg protein/h). SnMP, a known inhibitor of HO activity (Chernich et al., 1989), has been used clinically to control hyperbilirubinemia in humans (Kappas et al., 1995); therefore, its effect on HO activity in HUVECs was assessed. Cells treated with SnMP showed a 4-fold decrease in HO activity compared with nontreated cells. Moreover, cells treated with a combination of SnMP and heme demonstrated a 3.9-fold decrease in HO activity compared with cells treated with heme alone. These results show that CoCl₂ significantly induces HO activity, whereas SnMP acts as a potent inhibitor.

View larger version (52K): [in this window] [in a new window]   Fig. 2.   Heme oxygenase activity was measured after 24 h of incubation of HUVECs with inducers and inhibitors of HO. Results are described as picomoles of bilirubin formed per milligram of protein per hour. Control experiments were performed by exposing cells to the medium. Each bar represents the mean ± S.D. of three separate experiments. , significantly different from heme-treated cells (P < .05). *, significantly different from control cells (P < .05).

Heme-HO-Mediated ICAM-1 Expression. We analyzed the role of HO activity in modulating heme-induced ICAM-1 expression with an inducer (CoCl₂) or inhibitor (SnMP) of enzyme activity. HUVECs pretreated for 24 h with SnMP (20 µM) followed by heme treatment for 24 h resulted in a 2-fold enhancement in ICAM-1 expression compared with heme alone. In contrast, HUVECs pretreated with CoCl₂ (50 µM), followed by exposure of heme (50 µM) for 24 h showed a 33% decrease in heme-induced ICAM-1 expression. Interestingly, endothelial cells treated with SnMP showed a significant increase in ICAM-1 expression compared with control cells (Fig. 3), whereas exposure to CoCl_2, an inducer of HO activity, showed no significant increase in ICAM-1 expression. These results suggest that down-regulation of HO activity by SnMP inhibited the basal levels of HO activity and buildup of cellular heme with associated marked elevation of ICAM-1.

View larger version (57K): [in this window] [in a new window]   Fig. 3.   Heme-mediated endothelial ICAM-1 expression under HO-stimulating or -inhibiting conditions was measured with flow cytometry. Endothelial cells were pretreated with either CoCl2 or SnMP for 24 h followed by incubation with heme (50 µM) for 24 h. The y-axis indicates the mean fluorescence intensity of 5000 cells. Each bar represents the mean ± S.D. of three separate experiments *P < .05 compared with control; P < .05 compared with heme.

Contribution of HO-1 and HO-2 Isoforms on Total HO Activity and Heme. We next investigated the relative contribution of HO isoforms to elevation of HO activity after heme treatment. Because SnMP is an inhibitor of both HO-1 and HO-2 activity, we designed an antisense strategy to selectively inhibit HO-1 and HO-2 activity. Cells were treated with either HO-1 antisense ODNs or HO-2 antisense ODNs as described in Materials and Methods, and HO activity was assessed. The basal levels of HO activity in endothelial cells were 0.798 ± 0.16 nmol bilirubin/mg/h (Fig. 4), which was increased 3- to 4-fold after heme exposure. Preincubation of endothelial cells with HO-1 antisense ODNs at a concentration of 3, 6, and 9 µg/ml resulted in inhibition of HO activity in a dose-dependent manner. Treatment of HUVECs with HO-1 antisense (9 µg/ml) resulted in a 70% decrease of HO activity compared with the control (Fig. 4). In contrast, addition of HO-1 sense ODN did not affect HO activity. Treatment of endothelial cells with HO-2 antisense ODN (9 µg/ml) failed to inhibit HO activity at the same levels as HO-1 antisense ODN (Fig. 4). The inhibitory effect of HO-2 antisense ODNs on total activity was ~21%. These results suggested that elevation of HO activity in heme-treated endothelial cells is largely dependent on up-regulation of HO-1 gene expression.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   HO activity was measured in HUVECs treated with HO-1 or HO-2 sense or antisense ODNs in lipofection complex for 24 h; heme (50 µM) was added for an additional 24 h. And HO activity was measured as described in Materials and Methods.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Spectral measurement of heme in endothelial cells pretreated with antisense ODN or sense for HO-1 and HO-2. Cell cultures were treated with antisense ODNs, and heme (10 µM) was added to cell culture for 24 h. Cells were washed and heme levels were assessed on cell homogenate. Spectral measurement of reduced verses oxidized hemochromagen was performed as described in Materials and Methods.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Western blot analysis of HO-1 and HO-2 in endothelial cell homogenate, untreated cells (lanes 1 and 2), cells treated with heme (50 µM) (lanes 3 and 4), cells incubated with HO-1 antisense ODN (lanes 5 and 6), and cells preincubated with HO antisense ODN followed by heme (lanes 7 and 8).

Elevation of ICAM-1 by HO-1 Antisense ODN. We next investigated the effect of selective inhibition of HO-1 and HO-2 on heme-mediated adhesion molecules' ICAM-1 expression. Treatment of endothelial cells with heme causes elevation of adhesion molecules (Fig. 7; cyan histogram). Pretreatment of endothelial cells with HO-1 antisense ODN causes up-regulation of heme-mediated ICAM-1 expression (Fig. 7, red histogram). Inclusion of HO-1 sense ODN or HO-2 antisense ODN to endothelial cell culture failed to increase ICAM-1 expression (data not shown). Addition of either HO-1 or HO-2 antisense ODN to endothelial cells in the absence of heme did not result in modulation of the basal levels of ICAM-1, and the yield was similar to that of the untreated cells (Fig. 7; cyan histogram). These results suggest that inhibition of HO activity by HO-1 antisense ODN resulted in potentiation of heme-mediated ICAM-1 expression.

View larger version (57K): [in this window] [in a new window]   Fig. 7.   Effect of HO-1 antisense ODNs on heme-mediated endothelial ICAM-1 expression as measured by flow cytometry. The open histogram represents control staining of untreated cells with an FITC-conjugated secondary antibody. The cyan histogram represents ICAM-1 expression by cells incubated with heme (50 µM) for 24 h. The red histogram represents ICAM-1 expression by cells pretreated with HO-1 antisense ODNs for 24 h followed by incubation with heme (50 µM) for an additional 24 h.

Glutathione-Dependent Decrease in ICAM-1 Expression following Heme Exposure. Because heme iron is considered a pro-oxidant by generation of the hydroxyl radical (·OH) and an antioxidant such as glutathione (Aoki et al., 1996) or bilirubin (Stocker et al., 1987) has been shown to alleviate such toxicity, the effect of glutathione levels on heme-induced ICAM-1 expression was studied with flow cytometry. Glutathione is known to be one of the most important oxidant scavengers and to modulate adhesion molecule expression (Aoki et al., 1996). NAC and GSH-ET are known to increase intracellular levels of glutathione. HUVEC monolayers, preexposed to 5 mM NAC or GSH-ET for 24 h followed by exposure to heme for 24 h, were analyzed for ICAM-1 expression. NAC and GSH-ET (5 mM) decreased heme-induced ICAM-1 expression by 37 and 44%, respectively (Fig. 8). NAC and GSH-ET alone did not change ICAM-1 expression.

View larger version (49K): [in this window] [in a new window]   Fig. 8.   Flow cytometric analysis of heme-induced ICAM-1 expression on HUVECs pretreated with the glutathione precursor NAC (5 mM) or GSH ET (5 mM) for 24 h. The y-axis indicates the mean fluorescence intensity of 5000 cells. Exposing cells to medium alone performed control experiments. Each bar represents the mean ± S.D. of three separate experiments. *P < .05 compared with control; P < .05 compared with heme.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The results of our study indicated that heme, which is known to induce HO activity in renal tissue (da Silva et al., 1994; Agarwal et al., 1995) and other mammalian cells (Lutton et al., 1992), acts similarly in HUVECs. Heme is also a more potent inducer of HO-1 than HO-2 isoforms. Addition of heme in a concentration- and time-dependent manner to HUVECs resulted in a sustained increase of ICAM-1 expression and enhancements of inflammatory processes. The participant's role of heme in inflammation in vivo emphasized the observation that administration of heme to healthy volunteers induced thrombophlebitis (Simionatto et al., 1988; Green and Tsao, 1990). Furthermore, excess free heme released from damaged erythrocytes following hemorrhagic injury or in conditions such as hemolytic anemia and sickle cell disease has been linked to blood vessel occlusion and endothelial dysfunction. Thus, heme-induced endothelial cell activation may result in vascular inflammatory disorders, including acute renal failure, atherosclerosis, ischemia-reperfusion injury, and sickle cell disease; therefore, suppression of heme-induced endothelial ICAM-1 expression could contribute to the prevention of inflammatory complications.

Study of the effect of increased HO activity on heme-mediated adhesion molecule expression in endothelial cells was an important aspect of these experiments. Our data showed that preinduction of HO activity by CoCl₂ resulted in a marked decrease in heme-induced ICAM-1 expression, whereas suppression of HO activity by SnMP increased heme-induced ICAM-1 expression. With antisense ODNs, we were able to evaluate the contribution of HO-1 and HO-2 to the total HO activity, ICAM-1 expression, and degradation of heme. Addition of HO-1 antisense ODNs to cell cultures prevented heme degradation and elevation of cellular heme levels, which is associated with up-regulation of adhesion molecules, ICAM-1, and HO activity. HO-1 has been postulated to act as a defense protein, providing protection to endothelial cells that are in a diseased state, at least in part by degrading the excessive amount of exogenous heme to bilirubin and carbon monoxide (Marks et al., 1991; Abraham et al., 1995; Wagener et al., 1997).

Acute phase proteins and the acute phase responses play an important role in the inflammatory process (Heinrich et al., 1990). Some of the acute phase proteins are suggested to have anti-inflammatory activities associated with their antioxidant properties. HO-1 is among the acute phase proteins (Mitani et al., 1993) and is induced within minutes by heme and other oxidative stress-inducing agents (for review, see Abraham et al., 1996). Expression of this protein also has been associated with increased tolerance to various forms of stressful stimuli. This process involves the transcriptional activation of several regulatory sites in the HO-1 promoter region; activator protein 1, activator protein 2, and interleukin-6-responsive elements; nuclear factor B; and heme-responsive elements that were found in the promoter region of this gene (Alam and Zhining, 1992; Lavrovsky et al., 1994; Abraham et al., 1996; Lu et al., 1998). Up-regulation of HO-1 gene expression by oxidative stress is mainly through activation of nuclear factor B1 (Lavrovsky et al., 1993, 1994; Kurata et al., 1996).

Thus, up-regulation of HO-1 expression may be beneficial for the control of various types of inflammatory reactions. Several studies in animal models clearly affirm the role of HO-1 as a tissue-protective response to injury and inflammation in vivo (Heinrich et al., 1990; Willis et al., 1996). Preinduction of HO-1 inhibits the monocyte transmigration induced by low-density lipoprotein (Ishizaka et al., 1997). Hemodynamic forces up-regulate HO-1 gene expression and generation of carbon monoxide in vascular smooth muscle cells and concomitant inhibition of platelet aggregation (Wagner et al., 1997). Induction of HO-1 has been shown to be coupled to ferritin synthesis (Eisenstein et al., 1991) and has provided significant protection against cisplatin- and rhabdomyolysis-induced injury to the kidney; inhibition of HO activity produced deterioration of renal function in this model system (Nath et al., 1992; Agarwal et al., 1995). It has been shown that inhibition of HO-1 gene expression is associated with a decrease in ferritin synthesis and enhanced cell death (Vile et al., 1994). In this study, inhibition of HO-1 was associated with elevation of ICAM-1 and elevation of cellular heme content. It remains to be investigated whether heme iron or nonheme iron is involved in ICAM-1 expression.

HO-1 is a stress protein and the direct involvement of HO-1 was implicated in the suppression of inflammation (Willis et al., 1996) and modulation of hypertension (da Silva et al., 1994; Motterlini et al., 1998). Furthermore, Johnson et al. (1995) concluded that carbon monoxide arising from heme via metabolism by HO exerts a vasodilator effect and that inhibition of HO by SnMP causes vascular constriction (Johnson et al., 1996). In addition, da Silva et al. (1996) demonstrated that preinduction of HO-1 enhanced cell resistance to oxidative stress; however, long-term expression had a deleterious effect on cell viability. This hypothesis is suggested by the fact that elevated levels of HO activity were returned to normal levels after removal of the inducers (Lutton et al., 1992). Our finding suggests that HO-1 gene expression should be closely controlled and expressed at low levels in physiological conditions.

The importance of HO-1 in the protection of cells against the detrimental effects of free heme/hemoglobin is supported by the observation that humans and mice deficient in the HO-1 gene suffer from several pathological conditions, including endothelial cell detachment and dysfunction with subendothelial deposition of foreign agents and arterial occlusion (Yachie et al., 1999). The mechanisms by which HO activity provides endothelial cell protection against heme-induced ICAM-1 and oxidative injuries remain to be elucidated. The ability of HO to reduce heme, a pro-oxidant property, and generate bilirubin, an antioxidant property, and diminish cellular free iron could be a factor that contributes to the alleviation of the inflammatory response. However, we cannot rule out the possibility that HO down-regulation of adhesion molecule expression via the formation of CO, which like nitric oxide, may modulate adhesion molecule expression (De Caterina et al., 1995). This finding implicates heme HO isoforms analogous to that of nitric oxide synthase or cyclooxygenase, where HO-2 may participate in the physiological regulation of cellular function, and HO-1 may play an important role in modulating responses to injury and other insults under pathological conditions. Thus, HO-2 may be the active form in the normal endothelium, but in response to injury, the cell recruits additional protection by enhancing HO activity via induction of HO-1 (Abraham et al., 1995; Zakhary et al., 1996).

Our findings show that administration of CoCl₂ significantly inhibited heme-induced adhesion molecules. In contrast, HO-1 antisense ODNs potentiated heme-induced oxidative stress, suggesting that induction of HO-1 isoforms constitutes part of the mechanism of the anti-inflammatory effect of HO-1 induction and provides further evidence of a significant relationship between HO-1 and the genesis of tissue inflammation. Our findings also indicate that the relationship can be regulated by simple pharmacological means and raise the possibility that regulation of HO activity-inflammatory response relationships in this manner may be applicable to the clinical setting.