to red pulp fibrosis, atrophy and functional hyposplenism in older mice, recapitulating the asplenia of an HO-1-deficient patient which was previously reported in Japan [5]. During constitutive erythrophagocytosis, HO-1-deficient macrophages including Kupffer cells appeared to exhibit cell death by themselves. This study also postulated that failure of tissue macrophages to remove senescent erythrocytes led to intravascular hemolysis and increased expression of hemopexin and haptoglobin, the heme and hemoglobin scavenger proteins, in the liver. Lack of macrophages expressing the Hp receptor, CD163, diminished the ability of Hp to neutralize circulating Hb, and iron overload occurred in kidney proximal tubules, which were able to catabolize heme with HO-2. Thus, in HO-1(-/-) mammals, the reduced function and viability of erythrophagocytosing macrophages appear to be the main causes of tissue damage and iron redistribution.
CO Derived from HO-2 Targets Hepatic Stellate Cells to Relax Sinusoids
Many years ago, we first showed that CO serves as a vasorelaxing factor that reduces sinusoidal tone and is necessary for maintenance of the microvascular perfusion [6, 7]. This is based on the fact that zinc protoporphyrin IX, a potent inhibitor of HO, abolished CO generation and simultaneously increased the baseline vascular resistance in the perfused liver. The role of endogenous CO in sinusoidal relaxation was also shown in experiments using free oxyhemoglobin which captures both NO and CO, and methemoglobin that is able to capture NO but not CO [2]: only oxyhemoglobin was able to reproduce the vasoconstrictive effect of the HO inhibitor. Although the gas appears to diffuse freely across the cell membrane, our results collected from the liver did not support such a notion: when encapsulated with liposome, oxyhemoglobin lost its ability to constrict sinusoids because of limited accessibility to the space of Disse across the endothelial fenestration, suggesting that the locus of the gas action is extrasinusoidal, perhaps involving hepatic stellate (Ito) cells. Ito cells constitute the most abundant resource of soluble guanylate cyclase in the liver that serves as a receptor for CO-responsive modest upregulation of cyclic GMP through the gas binding to the prosthetic ferrous heme of proteins.
Although a lot of evidence suggests that CO generated from HO plays a protective role against organ dysfunction, few studies have provided evidence for quantitative information of endogenous CO generation: according to previous data, local concentrations of CO in the liver under physiologic conditions appeared to be around 1 μmol/l. Under pathologic conditions including ischemia-reperfusion, acetaminophen-induced liver injury [8], endotoxemia [9] or excessive heme overloading [10], the CO concentrations reached 2-4 μmol/l. Still unclear is the extent to which NO could interact or compete with CO in vivo, mainly because of the lack of quantitative information on functionally intact NO in situ. Although sinusoidal endothelium constitutes a major cellular component of NO synthase expression, NO released from the enzyme may be entrapped by circulating erythrocytes or cancelled by superoxide anion spontaneously released from Kupffer cells [11].
Since the potency of NO to activate the cyclase is extremely greater than that of CO [12], CO appears to be able to activate soluble guanylate cyclase in vivo, but only when local NO concentrations are low. Actually, when the liver is exposed to endotoxemia to induce inducible NO synthase [9], amounts of cyclic GMP are dictated by NO but not CO. Since cyclic nucleotides upregulate the transcriptional expression of HO-1, the endotoxemic liver overexpresses HO-1; this response downregulates inducible NO synthase through degradation of the prostethic heme for this enzyme to suppress NO, and subsequently upregulates CO to maintain sinusoidal relaxation for the blood supply. Under these circumstances, CO relaxes sinusoids through cyclic GMP-independent mechanisms [9]. In this relaxation mechanism, the ability of CO to inhibit cytochrome P450 epoxygenases was suggested to be involved: in other other words, the stress-inducible CO is necessary to maintain sinusoidal blood flow and subsequently to guarantee bile output under disease conditions.
CBS as a CO Sensor Mined by Metabolomic Analyses
Because of the nature of the gas to bind to the metal-centered prosthetic groups of macromolecules, it is not unreasonable to hypothesize that biological gases bind to enzymes in metabolic systems that constitute a major class of such proteins. Based on this assumption, we have recently applied metabolomic analyses assisted by capillary electrophoresis combined with mass spectrometry (CE-MS) in order to mine footprints of the gases of interest on alterations in small molecular metabolites in the metabolic systems [13, 14]. In recent studies, we examined the effects of CO on the metabolic systems using several different experimental models where the gas was upregulated significantly. Results suggested that CO has the ability to inhibit the transsulfuration pathway that is rate-limited by cystathionine β-synthase (CBS), a heme-containing enzyme [14].
Based on these data, we examined the roles of the enzyme for a CO-specific receptor candidate in vivo. Several lines of biochemical evidence support the concept that CBS acts as a CO sensor. First, studies using recombinant CBS have shown that CO inhibits CBS with a Ki value of approximately 5 μmol/l [15], being comparable to the concentration occurring in the liver. Second, murine hepatocytes express both CO-producing HO and H2S-producing CBS with additional HO-1 induced in both hepatocytes and Kupffer cells under stress conditions. The close proximity of the enzyme distributions taken together with measured CO concentrations in the models, and the kinetics of CBS activity led us to hypothesize that CBS is acting as a CO sensor in vivo. In vivo pulse-chase analyses suggest that CBS is the enzyme that actually modulates the metabolic flux of this pathway, as judged so far from the results collected by 15N-methionine flux analyses using CE-MS [14]. To note is that CBS is the enzyme that generates H2S. CO-overproducing livers showed a decrease in labile H2S amount, whereas the livers of heterozygous CBS knockout mice did not show any notable decrease in H2S in response to stress-inducible levels of CO, suggesting that the gas inhibits the activity of CBS in vivo. Such a stress-inducible suppression of H2S in the liver stimulates
We believe that the gas involving CO has multiple modes of biological actions through varied specific receptors. CBS appears to be one of such multiple receptors that regulate organ functions. In order to mine novel receptors that have not yet been identified, further investigation with different technological approaches should obviously be necessary.
Acknowledgments
This work was supported by JST, ERATO, Suematsu Gas Biology Project, Tokyo 160-8582. Establishment of metabolomic analysis was supported by a Global COE Project for Human
