| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Review Article |
Department of Pharmacology, University of Saskatchewan, Saskatoon, Saskatchewan, Canada (L.W.); and Department of Biology, Lakehead University, Thunder Bay, Ontario, Canada (R.W.)
Abstract
Abstract I. Introduction II. Milestones for the Biological and Physiological Studies of Carbon Monoxide A. Endogenous Production of Carbon Monoxide B. Physiological Functions of Carbon Monoxide III. Heme-Dependent and -Independent Endogenous Production of Carbon Monoxide A. Primary Structure of Different Heme Oxygenase Isoforms B. Tissue-Specific Distribution of Heme Oxygenases C. Subcellular Localization of Heme Oxygenases D. Up-Regulation of the Expression and Activity of Heme Oxygenases E. Down-Regulation of the Expression and Activity of Heme Oxygenases F. Biological Functions of Heme Oxygenases 1. Production of Carbon Monoxide, Biliverdin/Bilirubin, and Ferrous Iron. 2. Heme Metabolism. 3. Heme-Containing Protein and Heme-Binding Protein. G. Nonheme Sources of Carbon Monoxide Production IV. Catabolism of Endogenous Carbon Monoxide A. Expiration B. Scavenging C. Oxidation V. Physiological Roles of Carbon Monoxide A. Carbon Monoxide and Circulatory System 1. Cardiac Function. 2. Vascular Contractility. 3. Platelet Aggregation and Monocyte Activation. B. Carbon Monoxide and Nervous System 1. Hypothalamic-Pituitary-Adrenal Axis. 2. Glia. 3. Circadian Rhythm Control. 4. Odor Response Adaptation. 5. Nociception and Chemoreception. 6. Thermal Regulation. 7. Learning, Memory, and Behavior. 8. Vision. 9. Hearing. 10. Peripheral Autonomic Nervous System. C. Carbon Monoxide and Respiratory System D. Carbon Monoxide and Reproductive System E. Carbon Monoxide and Gastrointestinal System F. Carbon Monoxide and Liver G. Carbon Monoxide and Kidney H. Carbon Monoxide and Pancreas VI. Pathophysiological Implications of Abnormal Heme Oxygenase/Carbon Monoxide System A. Neurodegenerations and Brain Disorders B. Hypertension 1. Hypertension Induced by Heme Oxygenase Inhibitors. 2. Spontaneously Hypertensive Rats. 3. Salt-Induced Hypertension. 4. Angiotensin II-Induced Hypertension. 5. One Kidney-One Clip Renovascular Hypertension. 6. Pulmonary Hypertension. 7. Portal Hypertension. C. Carbon Monoxide and Inflammation D. Cardiac Hypertrophy and Heart Failure E. Transplantation 1. Allograft Survival. 2. Xenograft Survival. F. Apoptosis and Cellular Proliferation 1. Vascular Smooth Muscle Cells. 2. Endothelial Cells. 3. Other Types of Cells. G. Oxidative Stress VII. Heme Protein-Dependent Cellular and Molecular Mechanisms for Carbon Monoxide Effects VIII. Interaction of Carbon Monoxide with Different Ion Channels A. Carbon Monoxide and KCa Channels B. Heme, Heme Oxygenase, 20-Hydroxyeicosatetraenoic Acid, and KCa Channels C. Carbon Monoxide and KATP Channels D. Carbon Monoxide and Calcium Channels E. Carbon Monoxide and Other Ion Channels 1. Na+ Channels. 2. Nonselective Cationic Channel. 3. Cyclic Nucleotide-Gated Ion Channels. IX. Interaction of Heme Oxygenase/Carbon Monoxide and Nitric-Oxide Synthase/Nitric Oxide Systems A. Influence of Carbon Monoxide on Nitric-Oxide Synthase/Nitric Oxide System 1. Carbon Monoxide Potentiates the Activity of Nitric-Oxide Synthase/Nitric Oxide System. 2. Carbon Monoxide Reduces the Activity of the Nitric-Oxide Synthase/Nitric Oxide System. 3. Carbon Monoxide Reduces the Expression of Nitric-Oxide Synthase. B. Influence of Nitric Oxide on Heme Oxygenase/Carbon Monoxide System 1. Nitric Oxide Potentiates the Activity of Heme Oxygenase/Carbon Monoxide System. 2. Nitric Oxide Reduces the Activity of Heme Oxygenase/Carbon Monoxide System. X. Therapeutic Applications of Carbon Monoxide A. Up-Regulating the Expression of Heme Oxygenase 1. Genetic Approaches. 2. Nongenetic Approaches. B. The More the Merrier? Down-Regulating the Expression or Activity of Heme Oxygenase 1. Genetic Approaches. 2. Nongenetic Approaches. C. Inhalation of Carbon Monoxide D. Use of Carbon Monoxide-Releasing Compounds E. Use of Prodrugs to Generate Carbon Monoxide XI. Conclusions and Perspectives
Over the last decade, studies have unraveled many aspects of endogenous production and physiological functions of carbon monoxide (CO). The majority of endogenous CO is produced in a reaction catalyzed by the enzyme heme oxygenase (HO). Inducible HO (HO-1) and constitutive HO (HO-2) are mostly recognized for their roles in the oxidation of heme and production of CO and biliverdin, whereas the biological function of the third HO isoform, HO-3, is still unclear. The tissue type-specific distribution of these HO isoforms is largely linked to the specific biological actions of CO on different systems. CO functions as a signaling molecule in the neuronal system, involving the regulation of neurotransmitters and neuropeptide release, learning and memory, and odor response adaptation and many other neuronal activities. The vasorelaxant property and cardiac protection effect of CO have been documented. A plethora of studies have also shown the importance of the roles of CO in the immune, respiratory, reproductive, gastrointestinal, kidney, and liver systems. Our understanding of the cellular and molecular mechanisms that regulate the production and mediate the physiological actions of CO has greatly advanced. Many diseases, including neurodegenerations, hypertension, heart failure, and inflammation, have been linked to the abnormality in CO metabolism and function. Enhancement of endogenous CO production and direct delivery of exogenous CO have found their applications in many health research fields and clinical settings. Future studies will further clarify the gasotransmitter role of CO, provide insight into the pathogenic mechanisms of many CO abnormality-related diseases, and pave the way for innovative preventive and therapeutic strategies based on the physiologic effects of CO.
Were scientists to remake the classic 1932 film Dr. Jekyll and Mr. Hyde, the starring role would go not to Frederic March but to an infamous gas molecule, namely carbon monoxide (CO1). The "evil side" of CO has been known for hundreds of years, far longer than Mr. Hyde's name. As a result, we are already accustomed to the idea that CO is nothing but a toxicant, waste, or pollutant. As an environmental toxicant, pathological levels of CO can induce both acute and chronic health hazards at societal and individual levels. On the other hand, CO has earned the image of Dr Jekyll in recent years. Mounting evidence now speaks loudly and clearly for the vital importance of CO, which is generated within our body, in regulating many biological functions. The sources of CO, its local concentrations, and its interaction with a specific environment integrally determine which side of this double-natured gas will dominate.
CO is the diatomic oxide of carbon. At temperatures above -190°C, CO is a colorless and odorless gas. The specific gravity of CO is 0.967 relative to air, and its density is 1.25 g/l at standard temperature and pressure. CO is a chemically stable molecule because of its formal triple bond. Chemical reduction of CO requires temperatures well above 100°C. The water solubility of CO is very low (354 ml/dl; 44.3 ppm by mass) at standard temperature and pressure (Allen, 1977
). CO cannot react with water without substantial energy input. Even for molecular oxygen, the reaction rate of CO is slow and needs a high activation energy (213 kJ/mol). Theoretically speaking, CO can be involved in redox reactions (Allen, 1977). Free CO does not readily react with reducing agents, including hydrogen. The coordinated CO has greater reactivity than the free gas, and the reduction of CO can be greatly facilitated by transition metals (Shriver, 1981). Once formed, metal carbonyls are relatively stable until CO is displaced, e.g., by molecular O2.
All types of incomplete combustion of carbon-containing fuels yield CO. Natural processes such as metabolism and production of CO by plants and oceans or wildfires release CO into the atmosphere. Oxidation of methane and nonmethane hydrocarbons by hydroxyl radicals and ozone, either natural or anthropogenic, is also a significant mode of CO production in the atmosphere. The most noticeable human activities which produce CO are internal combustion engines, appliances fueled with gas, oil, wood, or coal, and solid waste disposal. The use of smoking tobacco or inadequately vented stoves is examples of CO accumulation in a closed field. CO intoxication mainly results from exposure to environmentally generated CO at high concentrations and/or prolonged exposure periods. Being one of the most abundant air pollutants in North America, CO readily accumulates in the atmosphere and in our bodies, yet its physiochemical properties are hard to detect. CO intoxication can be fatal, the reason for its being known as a "silent killer". A simple increase in ambient CO levels will not necessarily lead to human intoxication, which is also influenced by the functional status of pulmonary ventilation, the endogenous buffering capacity, i.e., level of carbonmonoxy-hemoglobin A (COHb), and the partial pressures of CO and oxygen. Elevated CO concentration in the bloodstream accelerates binding of CO to normal adult hemoglobin (HbA), forming COHb. The formation of COHb impairs two functions of HbA. The oxygen storage function of HbA is significantly reduced since the affinity of CO to HbA is about 210 to 250 times greater than that of oxygen. The oxygen transportation function of COHb is also reduced, as the release of oxygen from COHb to the recipient tissue becomes more difficult. Both decreased arterial O2 content (poor O2 binding to hemoglobin) and decreased tissue PO2 (increased affinity of COHb for O2) (Stewart, 1975) cause hypoxia. The brain and heart are the organs most vulnerable to CO-induced acute hypoxia, due to their high demand for oxygen. Neurological injuries are manifested as headaches, dizziness, weakness, nausea, vomiting, disorientation, visual confusion, collapse, and coma. Without immediate treatment, the neurological injuries can be fatal. Severe or chronic CO intoxication of a developing fetus may increase the risk of retarded development of the central nervous system with abnormalities in visual perception, manual dexterity, learning, driving performance, and attention level. Chronic exposure of adult humans or experimental animals to CO has also been shown to induce cardiovascular disorders, including arteriosclerotic heart diseases and cardiac hypertrophy (Wang, 2004). The combination of CO with other heme-proteins, such as cytochrome P450, cytochrome c oxidase, catalase, and myoglobin may also in part account for the toxic effects of CO (Piantadosi, 2002). However, it has to be pointed out that living cells can tolerate CO in the concentration range of 0.01% (100 ppm) for several hours (Otterbein and Choi, 2000). Exposure to 500 ppm CO continuously for up to 2 years without deleterious effects has been reported in rodents (Stupfel and Bouley, 1970; Otterbein and Choi, 2000).
CO is a paradigm in itself. Advances in several frontiers in the last decade have given the evil image of CO a make-over. Endogenous CO production has been illustrated in great details, from its enzymatic catalyzation process to its variations in different tissues under different conditions. The physiological importance of endogenous CO to the homeostatic control of the human body has been reevaluated and realized in neuronal and cardiovascular systems and almost every other system and every tissue type. The numbers of publications and conferences dedicated to CO physiology and CO pharmacology have been increasing at a stunning speed with contributions from numerous research laboratories around the world. Although there is still reasonable skepticism on the physiological importance of CO, the current research advancements have lead to a significant acknowledgment that CO, joining with other endogenous gases including nitric oxide (NO) and hydrogen sulfide (H2S), is one member of a new class of physiologically important "gasotransmitters", a nomenclature composed of "gas" and "transmitters" (Wang, 2002).
II. Milestones for the Biological and Physiological Studies of Carbon Monoxide
A. Endogenous Production of Carbon Monoxide
As early as the 1850s, French physiologist Claude Bernard recognized the reversible binding of CO with hemoglobin as a potent chemical reaction that could cause asphyxia (Bernard, 1857). A later study in 1895 showed the antagonistic effect of high partial pressure of O2 on CO binding to hemoglobin (Haldane, 1895). The first indication for endogenous CO production was made by Saint-Martin and Nicloux in 1898. Warburg reported in 1930 that CO could inhibit respiration in yeast in a light-sensitive manner, extending the original discovery that COHb could be dissociated by exposure to light of appropriate wavelengths (Haldane and Smith, 1896). These pioneer studies initiated a century of investigations for the biological actions of CO. In the early 1950s, Sjöstrand for the first time provided actual experimental evidence for the existence of CO in our body (Sjöstrand, 1950 and 1952). He observed that decomposition of hemoglobin in vivo produced CO. Endogenous production of CO can be regulated. Increased heme levels after erythrocyte destruction increased endogenous CO production, reflected by the elevated COHb level (Coburn et al., 1965 and 1966). When heme metabolism is abnormally increased, such as in hemolysis, CO production rate in our body can increase tremendously (Coburn et al., 1966). In the late 1960s, Tenhunen and colleagues ascribed the driving force for endogenous source of CO production to heme oxygenase (HO) (Tenhunen et al., 1968, 1969, 1970; Landaw et al., 1970). The inducible HO isoform, HO-1, was identified in 1974 in two different laboratories (Maines and Kappas, 1974; Yoshida et al., 1974). Maines' laboratory in 1986 identified the constitutive HO isoform, HO-2, from rat liver microsomes (Maines et al., 1986; Trakshel et al., 1986). About 10 years later, the same laboratory identified the third HO isoform, HO-3 (McCoubrey et al., 1997b) (Fig. 1).
|
). The average physiological concentrations of CO in tissues are rather low, in the nanomolar range if based on normal levels of COHb of 1 to 2% (Piantadosi, 2002). However, the accuracy of this estimation is questionable, since CO generated in living cells would first be scavenged in the cytosol before being released into the bloodstream where COHb is formed. Previous studies have shown poor correlation of COHb levels with biological changes induced by CO (Stewart, 1975) and remnant effects of CO after COHb elimination (Halperin et al., 1959).
B. Physiological Functions of Carbon Monoxide
While investigation on endogenous production of CO and its regulation progressed continuously, especially in regards to the expression and biological implications of HO, the questions of how our body uses, or why it needs, endogenous CO had not been answered by conventional understanding of this gas as an endogenous waste or by-product of heme metabolism. It was not until little more than a decade ago that an appreciation of the biological and physiological functions of endogenous CO, and the role of CO in the beneficial effects of HO, was made. The break-through discovery of NO opened the way to further research on membrane/receptor-independent signaling gas molecules. In 1991, Marks and colleagues predicted that there was a metabolic reason and a physiological meaning for the production of CO in our bodies (Marks et al., 1991). This pioneering thinking stirred up the resurgence of CO as a physiological signaling molecule (Wang, 1998). Two years later, Snyder's research team provided the first comprehensive evidence for the role of CO as an endogenous neural messenger, based on the effect of HO inhibitors and the histological location of HO (Verma et al., 1993). Rattan and Chakder (1993), also using HO inhibitors, demonstrated that endogenous CO was involved in the relaxation of opossum internal anal sphincter (IAS) in response to nonadrenergic noncholinergic (NANC) nerve stimulation.
The relaxation of pulmonary vasculature induced by exogenous CO under nomoxic conditions was reported as early as the late 1970s (Sylvester and McGowan, 1978). However, cardiovascular researchers struggled for many years to find the evidence for the vasoactive effect of endogenous CO. By inhibiting HO activity with zinc protoporphyrin-IX (ZnPP), Suematsu et al. (1994) provided the evidence that ZnPP treatment reduced endogenous CO generation and increased vascular resistance in rat liver. Other laboratories at the same time argued that the vascular effect of ZnPP might not be related to inhibition of HO provoked (Ny et al., 1995; Zygmunt et al., 1994). Ensuing studies applying HO substrate and/or up-regulating HO-1 expression confirmed the vasorelaxant effect of endogenous CO (Wang 1996; Wang et al., 1997a).
The identification of the physiological role of endogenous CO was greatly facilitated by the use of gene knockout or gene overexpression techniques. The first HO-2 null mutant mouse was produced in 1995 (Poss et al., 1995). Poss and Tonegawa (1997) first generated mice deficient in HO-1 by targeted deletion of a 3.7-kb region including exons 3 and 4 and a portion of exon 5 of the mouse HO-1 gene. The direct relevance of the HO/CO pathway to human health was drawn by the reported first human case of HO-1 deficiency in Japan in 1999 (Yachie et al., 1999; Ohta et al., 2000). This HO-1 deficient patient died at age 6, showing growth retardation, anemia, thrombocytosis, hyperlipidemia, leukocytosis, elevated serum levels of ferritin and heme, and lower serum levels of bilirubin.
Clearly, the momentum in the study on CO biology in recent years has injected more and more enthusiasm into study on HO biology. Research on CO and HO is now closely integrated and coevolving. This HO/CO field is experiencing phenomenal growth, spurred on by scientists and health workers, from the laboratory bench to the hospital bed side, and by trainees, from graduate students to postdoctoral fellows.
III. Heme-Dependent and -Independent Endogenous Production of Carbon Monoxide
Upon the action of hydrogen peroxide or ascorbic acid, heme methylene bridges can be broken and CO released (Bonnett and McDonagh, 1973; Brown et al., 1978; Guengerich, 1978). Cytochrome P450 can be inactivated by free NADPH oxidation, NADPH-dependent monooxygenase reactions and lipid peroxidation. This self-inactivation also leads to the breakage of the bond between heme and apoenzyme and heme degradation (Karuzina et al., 1999).
Although nonenzymatic heme metabolism as mentioned above occurs in vivo, the majority of CO in our body is produced by enzymatic heme metabolism. This metabolism is catalyzed by HO, mainly occurring in the reticuloendothelial system of the spleen and liver (Maines, 1988). There are three isoforms of HO. HO-1 is the inducible isoform. An increased cellular stress level is one common denominator for most of the stimuli to up-regulate de novo transcription of HO-1 (Applegate et al., 1991; Tyrrell, 1999). HO-2 is constitutively expressed in many mammalian cells. HO-3 is also a constitutive isoform of HO (McCoubrey et al., 1997b). One theory suggests that HO-3 may be derived from retrotransposition of the HO-2 gene since the HO-3 gene does not contain introns (Scapagnini et al., 2002b).
A. Primary Structure of Different Heme Oxygenase Isoforms
Four mammalian HO-1 genes have been cloned and sequenced, including rat (Muller et al., 1987), mouse (Alam et al., 1994), human (Yoshida et al., 1988), and chicken (Lu et al., 1998). Their protein homology is about 80%. Human HO-1 gene is located on chromosome 22q12. HO-2 has been cloned from rat, mouse, and human and they share more than 90% protein homology. Human HO-2 is mapped to chromosome 16q13.3 (Kutty et al., 1994a; Abraham et al., 1996).
The HO-1 gene consists of 5 exons and 4 introns, spanning 
14 kb. A promoter sequence is located approximately 28 base pairs upstream from the starting site of transcription. Different transcriptional enhancer elements reside in the flanking 59 region (Fogg et al., 1999). Consensus binding sites for oxidative stress-responsive transcription factors, including nuclear factor-
B (NF-B), activator protein-1 (AP-1), AP-2, Sp1, upstream stimulatory factor, c-myc/max and interleukin-6 response elements have been reported in the promoter region of the human HO-1 gene (Deramaudt et al., 1999; Lavrovsky et al., 1994; Muraosa et al., 1996; Sato et al., 1990; Tyrrell et al., 1993), suggesting a potential role for these factors in modulating HO-1 induction. Positive regulatory regions of the human HO-1 promoter contain consensus binding sites for AP-1, STATx, c-Rel, hepatocyte nuclear factor-1 and factor-4, GATA-X, and cadmium response element (Takahashi et al., 1999b and 2002; Takeda et al., 1994). Negative regulatory regions of the human HO-1 promoter contain consensus binding sites for negative regulatory elements (NRE) boxes (Takahashi et al., 1999b) and a polymorphic GT repeat region (Yamada et al., 2000). It should be pointed out that some of these regulatory regions may function in one type of mammalian cells, but not necessarily in others (Takahashi et al., 1999b). The HO-1 gene from different species may have quite different responses to the given inducers, depending on variances in the promoter regions of the HO-1 gene. For example, in response to interferon-
or hypoxia, mouse HO-1 gene is activated but repression of human HO-1 gene occurs (Sikorski et al., 2004). Accordingly, hypoxia up-regulates the expression of HO-1 in rat, bovine, mouse, and monkey cells but is a repressor of HO-1 expression in human cells (Kitamuro et al., 2003, Lee et al., 1997). In another case, heat shock induced the expression of HO-1 in rodent cells (Shibahara et al., 1987), but HO-1 expression in human cells was not altered by heat shock stimulation (Okinaga et al., 1996; Sato et al., 1990; Shibahara et al., 2002). Finally, an internal enhancer region was identified in human HO-1 promoter (Hill-Kapturczak et al., 2003). This region is only responsive to heme and cadmium.
The HO-2 gene consists of 5 exons and 4 introns. As the products of different genes, HO-1 (32 kD) and HO-2 (36 kD) share roughly 40% amino acid homology (Maines, 1997). HO-1 and HO-2 share a common 22 amino acid domain (differing in just one residue), named "HO signature". This signature sequence may be responsible for the heme degradation capacity (Maines, 1997). HO-1 and HO-2 proteins are immunologically distinct. HO-3 (33 kD) shares about 90% amino acid homology with HO-2. There are no introns in HO-3 gene (Scapagnini et al., 2002a).
Both HO-2 and HO-3, but not HO-1, are endowed with 2 Cys-Pro residues as the core of the heme-responsive motif, a domain critical for heme binding but not for its catalysis (Hon et al., 2000; McCoubrey et al., 1997a). HO-3 is a poor heme catalyst and its role may be limited to heme binding and/or sensing, considering the presence of 2 heme-responsive motifs in its amino acid sequence.
B. Tissue-Specific Distribution of Heme Oxygenases
The spleen has the most abundant expression of HO-1. In fact, under physiological conditions, the spleen may be the only organ in which HO-1 overpowers HO-2. In many other mammalian tissues, HO-1 is ubiquitously induced. HO-2 is predominantly expressed in the brain and testes and also constitutively expressed in other tissues, including endothelium, distal nephron segments, liver, myenteric plexus of the gut, and in other tissues at low levels (Ewing and Maines, 1992). The profile of HO isoform expression can change under different conditions. Upon stimulation, HO-1 expression in the testes overpowers the expression of HO-2. The same profile shift holds true for the lung, brain, and other tissues. Although early studies did not find HO-1 proteins in the brain (Ewing and Maines, 1992), recent studies have detected the HO-1 mRNA in different regions of the brain, especially in the hippocampus and the cerebellum (Scapagnini et al., 2002a). The expression of HO-1 protein in rat hippocampus at different ages was also reported (Huang et al., 2004). HO-3 has only been found in rat tissues, including brain, liver, kidney, and spleen (McCoubrey et al., 1997b). Using RT-PCR method, Scapagnini et al. (2002a) found high levels of HO-3 transcript in rat cerebellum and the hippocampus. HO-3 transcript was also detected in primarily cultured rat type I astrocytes, but not cortical neurons. In situ hybridization using an HO-3 specific riboprobe showed the expression of HO-3 mRNA in hippocampus, cerebellum and cortex. The expression of functional HO-3 protein, even in rat tissues, is in doubt since no report is available to convincingly demonstrate the detection of HO-3 proteins in any tissues. Hayashia et al. (2004) recently re-examined the expression of HO-3 gene in rat tissues. They concluded that the reported HO-3 genes might be processed pseudogenes derived from HO-2 transcripts. Genomic PCR, RT-PCR, and Western blot studies all lead these authors to the same conclusion that HO-3-related protein(s) is unlikely expressed in rat tissues.
C. Subcellular Localization of Heme Oxygenases
HO-1 is traditionally viewed as a microsomal protein, primarily localized in endoplasmic reticulum (Maines, 1988). This protein can be translocated to the nucleus of differentiated astroglial cells (Li Volti et al., 2004), where it may participate in the regulation of heme metabolism. Since nuclear heme can activate a high molecular weight complex by altering the affinity of HapI to Hsp90, the translocation of HO-1 to the nucleus would be a factor in the regulation of different transcriptional factors (Li Volti et al., 2004). Subcellular localization of HO-1 has also been detected in cytoplasm, nuclear matrix, mitochondria, and peroxisomes of parenchymal and nonparenchymal liver cell populations (Immenschuh et al., 2003). By monitoring the formation of bilirubin as an indicator of HO activity, Srivastava and Pandey (1996) detected significant HO activity in mitochondria from liver, spleen, kidney and brain, but not cerebral mitochondria. Since these activities can only be detected after the animals were treated with cobalt and hemin, HO-1 localization in mitochondria under these conditions was suggested. HO-2 proteins are anchored to the endoplasmic reticulum by a hydrophobic sequence of amino acids at the carboxyl terminus of the protein (Wagener et al., 2003). The possibility has also been raised that HO-2 may be anchored to plasma membranes directly or indirectly through big-conductance KCa channels (Williams et al., 2004).
D. Up-Regulation of the Expression and Activity of Heme Oxygenases
HO-1 is also known as the stress protein HSP32 (Keyse and Tyrrell, 1989). A great array of endogenous and exogenous stimuli can induce the expression of HO-1. Among the known HO-1 inducers are heme and heme derivatives, heat shock, heavy metals, NO and NO donors, oxidized lipids, hyperoxia, lipopolysaccharides, phorbol ester, sodium arsenite (Sardana et al., 1981), radiation, ultraviolet, hydrogen peroxide, hypoxia, endotoxin, growth factors [platelet-derived growth factor (PDGF) and transforming growth factor 
(TGF-)], various electrophiles, okadaic acid, methylglyoxal (Wu, 2005), curcumin (Motterlini et al., 2000), oxidative stress (Nath et al., 2001), cytokines (interleukin-1, interleukin-6, interleukin-10, TNF-
, interferon-), shear stress (Wagner et al., 1997), intensive light, angiotensin II (Ang-II), glucose deprivation, and other injuries. HO-1 expression is up-regulated by exogenous CO (Ndisang and Wang, 2003a and 2003b; Carraway et al., 2002). Depending on the cell types and the nature of the stimuli, HO-1 induction may be mediated by different signaling pathways. These pathways may include cAMP-dependent mechanisms (Durante et al., 1997a; Immenschuh et al., 1998), protein kinase C (PKC), Ca2+-calmodulin-dependent protein kinase and the phosphoinositol pathway (Terry et al., 1999). Mitogen-activated protein kinases (ERK and P38) and tyrosine phosphorylation are also involved in HO-1 induction in some tissues (Elbirt et al., 1998; Shan et al., 1999; Alam et al., 2000; Chen and Maines, 2000). A role for PI3K activation was demonstrated for the effect of hepatocyte growth factor (Tacchini et al., 2001).
Being a constitutive isoform, HO-2 can also react to certain stimuli by changing its expression level or activity. A consensus sequence of the glucocorticoid response element is located in the promoter region of the HO-2 gene (Liu N et al., 2000). Correspondingly, the expression of HO-2 is up-regulated by adrenal glucocorticoids (Weber et al., 1994) and opiates (Li and Clark, 2000a). Induction of HO-2 by cortisteron has been shown in neonatal rat brain (Raju et al., 1997). In addition, endothelial cells treated with the NO synthase (NOS) inhibitor L-NAME and HO inhibitor zinc mesoporphyrin exhibited a significant up-regulation of HO-2 mRNA. Estrogen up-regulates HO-2 in endothelial cells (Tschugguel et al., 2001). Several mechanisms are responsible for the regulation of HO-2 activity. 1) HO-2 activity can be rapidly and transiently increased by calcium-calmodulin binding to HO-2 during neuronal activity (Boehning and Snyder, 2004). Boehning et al. (2004) demonstrated, with the aid of a yeast two-hybrid screen assay, the calcium-dependent high-affinity binding of calmodulin to HO-2. Consequently, the catalytic activity of HO-2 was increased significantly in stimulated neurons presumably within milliseconds. 2) PKC and phorbol esters phosphorylate HO-2 and increase the HO-2 activity (Dore et al., 1999). 3) Selective phosphorylation of HO-2 by CK2 (formerly known as casein kinase 2) has been shown, which markedly augments the catalytic activity of HO-2 (Boehning et al., 2003). CK2 is activated by PKC. This activation process is much slower than the calcium-calmodulin-dependent activation of HO-2 (Boehning et al., 2004). 4) Glutamate, via activation of the metabotrophic glutamate receptor (mGluR), stimulates HO-2 activity in neurons (Boehning et al., 2004). Nathanson et al. (1995) found that glutamate augmented endogenous CO production, which was abolished after inhibition of HO-2 and PKC. HO-2 is also activated by glutamate in cerebral vascular endothelium (Parfenova et al., 2001) and smooth muscle cells (Leffler et al., 2003), leading to increased CO production. This effect of glutamate, independent of cytosolic calcium, was abolished by inhibiting protein tyrosine kinase but potentiated by inhibiting protein tyrosine phosphatases (Leffler et al., 2003).
E. Down-Regulation of the Expression and Activity of Heme Oxygenases
HO-1 expression can be reduced by interferon- (Takahashi et al., 1999a) or hypoxia (Nakayama et al., 2000) in human glioblastoma, human umbilical vein endothelial cells, coronary artery endothelial cells, astrocytes, and many other human cell lines (Kitamuro et al., 2003). Bach1, a member of basic leucine-zipper factors, is a heme-regulated transcriptional repressor for the HO-1 gene (Kitamuro et al., 2003; Sun et al., 2002; Kitamuro et al., 2003). The reduced expression of HO-1 may help to preserve intracellular heme as an important substrate of certain heme proteins, and reduce energy expenditure for heme catabolism. It seems that the repression of HO-1 by hypoxia or heat shock is unique for many types of human cells, and the same hypoxia stimuli increase the expression of HO-1 in rodent cells (Shibahara et al., 2003). However, hypoxia does increase the expression of HO-1 in certain types of human cells, such as human retinal pigment epithelium (RPE) cells (Udono-Fujimori et al., 2004). Down-regulation of HO-2 expression has not been reported.
Whereas little is known about an endogenous HO activity inhibitor, an array of pharmacological blockers of total HO activity (HO-1 and HO-2) have been used as an invaluable tool in dissecting out the physiological role of endogenous CO. The most widely studied and used HO blockers are metalloporphyrins, including chromium mesoporphyrin (CrMP), manganese protoporphyrin, manganese mesoporphyrin, zinc protoporphyrin (ZnPP), tin mesoporphyrin, and tin protoporphyrin (SnPP) (Vreman et al., 1993). The majority of these metalloporphyrins are light sensitive. Therefore, by altering lighting conditions one can either optimize the specificity of action and preserve the integrity of most metalloporphyrins or purposely inactivate these agents as internal controls. Depending on the species of metal cations that is linked to the porphyrin ring as well as the substitutions in the ring, the potency of metalloporphyrins varies (Vreman et al., 1993). By carefully choosing the concentrations, these metalloporphyrins can specifically inhibit HO without interacting with other cellular components. This is specifically relevant to the reported inhibition of NOS and soluble guanylyl cyclase (sGC) activities by metalloporphyrins (Luo and Vincent, 1994; Grundemar and Ny, 1997). For example, SnPP is 10 times more potent in inhibiting HO-2 activity than NOS or sGC activities when used at 7.5 µM (Zakhary et al., 1996). Inhibition of muscle relaxation and suppression of cAMP and cGMP by the metalloporphyrins have also been reported (Ny et al., 1995), possibly due to the drug's interaction with membrane receptors or their downstream signal transduction pathways. The permeability of the blood-brain barrier to different metalloporphyrins varies. SnPP can easily pass the blood-brain barrier, whereas ZnPP cannot (Li and Clark, 2000b). This property will be important in determining the administration routes for applying metalloporphyrin in vivo for different purposes.
F. Biological Functions of Heme Oxygenases
1. Production of Carbon Monoxide, Biliverdin/Bilirubin, and Ferrous Iron.
HO forms a complex with NADPH-dependent flavoprotein reductase (cytochrome P450 reductase) and biliverdin reductase (a cytosolic enzyme) on the endoplasmic reticulum (Maines, 1988). In the presence of functional HO, the porphyrin ring of heme (ferroprotoporphyrin IX) is broken and oxidized at the -methene bridge, producing equimolar amounts of CO, ferrous iron, and biliverdin (Tenhunen et al., 1968; Maines, 1988; Ortiz de Montellano, 1998). The heme-CO metabolism pathway, as illustrated in Fig. 2, also requires the participation of NADPH and O2. HO has an apparent Km for O2 of 12 mM in liver. Therefore, the presence of O2 is required for HO activity, but even during severe hypoxia HO continues to produce CO. Cytochrome P450 reductase transfers electrons to the HO-heme complex. The process of endogenous CO production displays a wide color spectrum. Black heme breaks down to green biliverdin and colorless CO. Biliverdin complexes with iron until its final release. Yellowish bilirubin is generated from biliverdin, catalyzed by biliverdin reductase (Tenhunen et al., 1968).
|
; Stocker et al., 1987). Bilirubin protects cells from a 10,000-fold excess of hydrogen peroxide (Baranano et al., 2002; Dore et al., 1999; Llesuy and Tomaro, 1994; Stocker et al., 1987). Over production of bilirubin leads to jaundice, a clinical syndrome of hyperbilirubinemia. Beyond its antioxidant capability, biliverdin is also an endogenous inhibitor of sGC (Koglin and Behrends, 2002). At the micromolar range, biliverdin, but not bilirubin, inhibits the basal and NO-stimulated activity of recombinant (1)/(1) isoform of sGC without affecting the affinity of sGC for GTP.
Heme catabolism-generated ferrous iron, being an oxidant, stimulates the synthesis of ferritin (Eisenstein et al., 1991) through its regulatory protein binding and activation of iron response elements (Hanson et al., 1999; Pantopulos et al., 1997). Ferritin, an intracellular iron repository, allows safe sequestration of unbound iron liberated from heme degradation. In this way, ferritin possesses additional antioxidant capabilities (Balla et al., 1992; 1995). It has been estimated that each apo-ferritin molecule (450 kD) can sequester about 4500 iron atoms (Harrison and Arosio, 1996). Modulation of intracellular iron stores and increased efflux of free iron has recently been suggested as a mechanism for the cytoprotective effects of HO-1 (Ferris et al., 1999). In this activity HO-1 cooperates with a recently identified Fe-ATP pump (Baranano et al., 2000). Once HO-1 was knocked out, iron accumulated in the cells. The HO-1 knockout mice exhibit hepatosplenomegaly, lymphadenopathy and leukocytosis. Massive iron overload in the liver and kidneys in these mice leads to death at a young age (Poss and Tonegawa, 1997a; 1997b). Iron accumulation in the absence of HO-1 makes these animals much more susceptible to oxidative stress damage. Transfection of cells derived from HO-1 knockout animals with HO-1 cDNA restored the normal cellular iron levels in these cells (Ferris et al., 1999). Iron accumulation was also documented in the liver and kidney of the human HO-deficiency case (Yachie et al., 1999). Finally, a feedback mechanism between iron and HO-1 exists. Iron has been shown to regulate transcriptional expression of HO-1 and iNOS (Abraham et al., 1988; Baranano et al., 2002).
2. Heme Metabolism.
Regulating cellular heme level is another important function of HO. After completing their life cycle of 120 days, red blood cells release hemoglobin into circulation. Haptoglobin captures free hemoglobin and transports it to the reticuloendothelial system in the spleen, liver, and bone marrow. The rapid transformation of hemoglobin to methemoglobin also occurs, leading to the release of the incorporated heme. The free heme will then be carried by hemopexin or albumin to the reticuloendothelial system. In reticuloendothelial system, HO functions as the rate-limiting enzyme in further heme degradation (Abraham et al., 1988). HO activity in different types of cells is also largely responsible for the degradation of heme derived from denatured heme proteins other than hemoglobin.
Heme consists of ferrous iron complexed with the four porphyrin groups (Beri and Chandra, 1993). The organic synthesis of heme was realized as early as 1927 in Hans Fischer's laboratory (Watson, 1965). The full elucidation of all enzymes involved in heme synthesis in mammalian cells was not available until about 25 years later (Shemin and Wittenberg, 1951). Now we know that heme is synthesized in all human nucleated cells using glycine and succinyl CoA as the precursors. Involving 8 different enzymes, heme synthesis starts in the mitochondria, continues in the cytosol, and is completed in the mitochondria. The iron is eventually inserted by ferrochelatase. Heme can also be derived from the degradation of hemoproteins. This recycling process would be energy efficient without the need to start over again from glycine and succinyl-CoA.
The turnover of heme is rapid. For instance, cultured cerebellar granule cells consume 17% of the total heme pool to produce about 1 to 5 µmol CO in 5 h (Ingi et al., 1996b). The physiological level of free heme in normal cells is below 1 µM (Sassa, 2004). At this concentration range, free heme down-regulates 
-aminolevulinate synthase and reduces the expression of Bach1. The latter will lift inhibition of HO-1 gene expression (Sassa, 2004).
Extracellular heme is transported into cell via a heme transporter (Worthington et al., 2001). Due to its lipophilic nature, heme readily moves around among different organelles (Ingi et al., 1996b) and it interacts with many cellular membranes and organelles, including lipid bilayers, mitochondria, cytoskeleton, nuclei, and several intracellular enzymes (Nath et al., 1998; Ryter and Tyrrell, 2000). Free heme at high concentrations catalyzes oxidative reactions to generate reactive oxygen species (Jeney et al., 2002), mainly due to the catalytic effect of heme. This would explain increased expression of adhesion molecules such as ICAM-1, VCAM-1, and E-selectin on endothelial cells in the presence of high levels of heme (Wagener et al., 2003). The pro-oxidant effect of free heme is not due to the release of iron from heme molecule (Vincent et al., 1988) since heme induced more peroxidation of rat liver microsomal lipid in the presence of H2O2 than iron and that iron release is very low under the conditions employed. Increased oxidative stress affects a variety of substrates, including lipids, thiols, proteins, carbohydrates, and nucleic acids (Ryter and Tyrrell, 2000). Normal cellular functions would be disturbed and pathological cellular injuries exacerbated (Dennery et al., 1998; Ryter and Tyrrell, 2000). An abnormally high heme level is correlated with many diseases, such as nephrotoxin-induced renal injury (Agarwal et al., 1995). The elevated circulatory heme level results from either excessive filtration of heme proteins as would occur in rhabdomyolysis (Nath et al., 1992) or from the destabilization of intracellular heme proteins (e.g., cytochromes) in ischemia-reperfusion and nephrotoxin-induced renal injury (Balla et al., 1993; Agarwal et al., 1995; Shimizu et al., 2000). When free heme exceeds the physiological range, its cytotoxic role dominates its constitutive role in heme protein formation. Heme-responsive genes remain repressed when heme at low concentrations binds to Bach1 with MARE sequences (Ogawa et al., 2001). However, at higher concentrations heme inactivates the binding of Bach1, allowing access of transcription factors such as Nrf2 to interact with the MARE sequences (Ogawa et al., 2001). This in turn activates the heme-responsive gene. In this regard, intact HO activity is crucial for the removal of the prooxidant heme (Balla et al., 1991; Jeney et al., 2002). Cell lines derived from the HO-1 deficient human were strongly sensitive to the injury induced by hemin (Jeney et al., 2002).
3. Heme-Containing Protein and Heme-Binding Protein.
Most mammalian cells contain a "free" heme pool, i.e., nonprotein bound heme, providing heme for the synthesis of heme-containing proteins and for CO production (Ponka, 1999). In this article, we refer to heme-containing proteins as heme proteins. As the prosthetic moiety for heme proteins, the availability of heme influences the metabolism of hemoglobin, myoglobin, cytochromes, prostaglandin endoperoxide synthase, NOS, catalase, peroxidases, respiratory burst oxidase, guanylyl cyclase, tryptophan dioxygenase, pyrrolases, and many others (Table 1). These heme proteins play important roles in regulating cellular functions, from oxygen delivery and mitochondrial respiration to signal transduction (Platt and Nath, 1998). Among heme proteins also are cyclooxygenase isoforms that need the heme prosthetic group for their catalytic activity (Smith and Marnett, 1991).
|
The function of HO is also linked to another class of protein, i.e., heme-binding proteins. These proteins, without heme in their molecular structure, have vivid affinities for it. As such, they regulate the availability of heme for the catalytic activity of HO and CO production as well (Table 2). Examples include HBP23 (Iwahara et al., 1995), and the glutathione S-transferases. Hemopexin, which has the highest affinity for circulatory heme of all heme-binding proteins (Muller-Eberhard, 1988), completely inhibited heme-catalyzed lipid peroxidation at concentrations slightly higher than that of heme, suggesting a unique role for this acute phase protein in antioxidant defense mechanisms. The protein itself was not oxidized, presumably because the putative bis-histidyl heme-hemopexin complex cannot interact with H2O2. Rat and human albumin and rat glutathione S-transferases, proteins with moderate affinities for heme, decreased heme-catalyzed lipid peroxidation in a dose-dependent manner but were subject to oxidation. Bovine albumin and rat liver fatty acid-binding protein have lower affinities for heme. These proteins enhanced, instead of inhibiting, lipid peroxidation. In short, heme-binding proteins may enhance, decrease, or completely inhibit heme-catalyzed oxidations and in doing so the proteins themselves may be oxidized depending upon their relative affinities for heme, the nature of the amino acids in the vicinity of the bound catalyst, and the availability of a free coordination site on the iron.
|
In summary, the levels and activities of heme, HO, and CO are interrelated closely (Fig. 3). HO-2 may function as a physiological regulator of cellular functions via controlling the size of the free heme pool and providing physiologically important heme metabolites. The elevated heme level itself presents a pro-oxidant threat. The cell copes with this threat by up-regulating HO-1 expression. The latter degrades heme to increase the production of CO. The role of HO-3 in heme degradation is very limited. It may function as a heme sensing or heme binding protein (McCoubrey et al., 1997b). Thus, an appropriate heme level is reached and cellular homeostasis maintained.
|
G. Nonheme Sources of Carbon Monoxide Production
Heme-independent sources for endogenous CO production have been reported (Rodgers et al., 1994). In comparison with HO-catalyzed reactions, heme-independent CO production is minor and its physiological importance has not been properly evaluated. The CO thus formed diffuses into the blood, is carried via hemoglobin, and excreted in the lungs.
Phenobarbital and diphenylhydantoin enhanced CO production in humans (Coburn, 1970a) and so did progesterone (Delivoria-Papadopoulos et al., 1974). CO production during auto- and enzymatic oxidation of phenols was also reported (Miyahara and Takahashi, 1971). Photo-oxidation of organic compounds is another source of endogenous CO production (Rodgers et al., 1994). Iron-ascorbate-catalyzed lipid peroxidation of microsomal lipids and phospholipids results in endogenous CO production, a process usually linked to oxidative stress (Nishibayashi et al., 1968; Vreman et al., 1998; Wolff and Bidlak, 1976). Reduction of cytochrome b5 is accompanied by emergence of the absorption peak at 450 nm (Archakov et al., 1975). This likely reflects the yield of endogenous CO formation during lipid peroxidation. Archakov et al. (2002) further investigated the CO production from rabbit liver microsomes during iron-dependent lipid peroxidation, induced by phenobarbital (a specific inductor of cytochrome P450 2B4). This CO production was NADPH-dependent and required Fe2+. Whereas inhibition of HO activity with Zn-PP did not alter the CO production, addition of cytochrome P450 inhibitors SKF 525A and metyrapone reduced CO formation rate. Furthermore, there was no heme destruction in microsome preparation used in this study. Desferal and -tocopherol potently inhibit lipid peroxidation via acting on the protein bound Fe2+ in microsomes. After microsomes were treated with these two antioxidants, CO production decreased, accompanied by a decline in the formation rate of malonyldialdehyde (MDA), an end product of lipid peroxidation. These data convincingly demonstrated lipid peroxidation-induced, but not heme-dependent, CO production.
Endogenous production of CO from lipid peroxidation is ubiquitous to many types of cells, including brain, kidney, lung, spleen and blood (Vreman et al., 1998). Although no CO production from lipid peroxidation has been detectable in rat liver and heart (Vreman et al., 1998), CO thus generated was observed in rabbit (Archakov et al., 2002) and mouse (Usami et al., 1995) hepatic microsomes.
IV. Catabolism of Endogenous Carbon Monoxide
The majority of CO from systemic production is exhaled through the lung. CO rapidly diffuses across the alveolar –capillary membrane, a process influenced by alveolar gas volume, ventilation, and the concentration of hemoglobin in the pulmonary capillaries (Coburn and Forman, 1987). The local partial pressures of both CO and O2 determine the cellular concentrations of CO since these two gases compete for the same iron or copper binding sites. Exhaled CO from humans has been recently used as a novel noninvasive indication of heme metabolism and the related endogenous levels of CO and bilirubin. The end-tidal breath CO, corrected for ambient CO, has been clinically used for understanding the mechanisms of jaundice in healthy term infants in a variety of conditions (Stevenson et al., 1994; Okuyama et al., 2001). Seshadri et al. (2003) monitored exhaled CO in patients with advanced ischemic and nonischemic cardiomyopathy. It was found that exhaled CO was lower in patients with cardiomyopathy at rest and immediately after exercise than healthy people. A lower CO production in these patients was thus indicated.
Under normal conditions, most of our body CO is endogenously generated from HO-catalized heme metabolism. Also contributing to our body CO store are exogenously inhaled ambient CO gas or xenobiotics which is metabolized to CO by cytochrome P450 in the liver (Kubic and Anders, 1978). CO is bound to hemoglobin in the red blood cells as COHb (80%). Other intracellular heme proteins are responsible for the residual CO loading (Coburn, 1970b). The CO body stores are exchangeable. It was suggested that during hypoxia more CO would move from blood to tissue where CO binds to heme proteins (Coburn and Mayers, 1971). However, this redistribution of CO has not been demonstrated under physiological conditions.
Oxidation of CO in mammalian tissues under physiological conditions has not been shown. It has been believed, however, for a long time that a small metabolic endpoint of CO in living tissues is the oxidization to CO2 (Fenn and Cobb, 1932). The catalyzation of this process by reduced cytochrome c oxidase in mitochondria was reported in 1965 (Tzagoloff and Wharton, 1965) and later confirmed by many other laboratories (Young and Caughey, 1986; Young et al., 1979; Vijayasarathy et al., 1999). It seems that whether the interaction of CO with cytochrome c oxidase leads to ferrous carbonyl formation or the formation of CO2 depends on the CO/O2 ratios. The latter determined molecular configuration of the oxidase (Young and Caughey, 1990). The oxidation of CO in vivo is much slower than the rate of endogenous CO production. The rate of CO oxidation, however, increases in proportion to tissue CO store (Luomanmaki and Coburn, 1969). In microbes living on CO, CO is oxidized to CO2 in the presence of CO dehydrogenase (Ragsdale, 2004). The oxidation or oxygenation of CO to CO2 is also well appreciated in chemistry, e.g., the metal-catalyzed water gas shift reaction generates H2 and CO2 from CO and H2O, but this CO oxidation also requires temperatures beyond physiological tolerance of most organisms. CO in the atmosphere can react with hydroxyl radicals to yield HO2 (hydroperoxyl radical) and CO2 (Allen and Root, 1957).
V. Physiological Roles of Carbon Monoxide
A. Carbon Monoxide and Circulatory System
1. Cardiac Function.
The cardiac distribution of HO-1 has been largely detected in heart vascular wall, not in cardiomyocytes. The expression level of HO-1 proteins in normal myocardium often falls below the detectable level by Western blot (Nishikawa et al., 2004; Lakkisto et al., 2002). However, HO-1 protein expression is significantly up-regulated by hemin or other pathological stimuli, such as myocardial infarction (Lakkisto et al., 2002). Hypoxia-induced up-regulation of HO-1 in heart also increased CO production. The latter improves cardiac blood supply by relaxing vascular tone in heart, constituting a cardiac defense mechanism (Grilli et al., 2003). CO perfusion reduced the contractility of isolated rat papillary muscle (Liu et al., 2001). Clark et al. (2003), using a CO carrier, demonstrated that CO protected myocardial cells and isolated rat hearts against ischemia-reperfusion (I/R) injury as well as cardiac allograft rejection in mice. Exogenous CO also limited ischemia–reperfusion injury in vivo in a mouse model of myocardial infarction (Guo et al., 2004). Hearts from HO-1 knockout mice have greater susceptibility to I/R injury (Yoshida et al., 2001). Cardiac-specific overexpression of HO-1 leads to attenuated myocardial injury after I/R in transgenic mice (Yet et al., 2001).
Two HO-2 homologous transcripts (1.9 and 1.3 Kb) and HO-2 proteins are constitutively expressed in the atrium, ventricles and descending aorta (Ewing et al., 1994). HO-2 is also expressed in intracardiac neurons (Hassall and Hoyle, 1997). Therefore, cardiac CO production can be catalyzed by HO-2 from different types of cells, cardiomyocytes, vascular smooth muscle cells, and neurons.
2. Vascular Contractility.
The phenomenon of CO-induced vasorelaxation was originally unmasked in 1984 when McGrath and Smith showed the relaxation of rat coronary artery in response to exogenous CO. It was quickly demonstrated in the following years that the vasorelaxant effect of exogenous CO was not mediated by endothelium (Coceani et al., 1988; Graser et al., 1990; Vedernikov et al., 1989). The vasorelaxant effect of endogenous CO was indicated as inhibition of HO activity with metalloporphyrins either increased the perfusion pressure of isolated rat liver (Suematsu et al., 1994) or decreased the diameter of pressurized gracillis muscle arterioles (Kozma et al., 1999). Pretreatment of rat tail artery tissue with hemin to promote HO activity suppressed phenylephrine-induced vasoconstriction in a time- and concentration-dependent manner (Wang et al., 1997b). Moreover, hemin-dependent vasorelaxation was abolished after the tail artery tissues were coincubated with either oxyhemoglobin to scavenge CO or ZnPP to inhibit HO, suggesting that hemin increased endogenous production of CO (Wang et al., 1997b). Mounting evidence now demonstrates that the CO-induced vascular relaxation is ubiquitous (Wang, 1998). The long list of CO-relaxed vascular tissues includes rat tail artery (Wang, 1996; Wang et al., 1997a), lamb ductus arteriosus (Coceani et al., 1988), rat and rabbit thoracic aorta (Lin and McGrath, 1988; Furchgott and Jothianandan, 1991), canine carotid, coronary and femoral arteries (Vedernikov et al., 1989), rat coronary artery (McGrath and Smith, 1984), guinea pig coronary artery (Gagov et al., 2003), porcine coronary artery and vein (Graser et al., 1990), rat hepatic vein (Pannen and Bauer, 1998), piglet mesenteric artery (Villamor et al., 2000
