- Email: [email protected]
Does carbon monoxide have a physiological function?
TiPS - May 1991 [Vol. 221
185
ists, due to the much smaller dose of the full agonist required for 50% effect in the high density cells. An advantage of this low dose might be a smaller risk of side&ects not related to an interaction with the G_4BA receptor complex’. The apparent difference in selectivity of the three agonists disappears almost fully, taking the peak effect ratio as an index (Fig. 3&ures 2F-H and Fig. 3 show that the preference for the effects in high density cells diminishes if the peak effect of the full agonist falls upon the highest part of the log concentration-effect curve, whereas it remains unchanged with the partlal agonist (at = 0.1). On the other hand, if the therapeutic aim is mediated by lower density cells (central muscle relaxation, sleep, anesthesia), this can only be achieved with a full agonist benxodiazepine. These simulations show that, indeed, pa&ally agonistic benzodianepines have a greater specificity for the effects induced via
the low density cells. This greater selectivity, however, only becomes obvious when doses almost approaching E, are administered. Such doses may be required because the anxidytic action does not last long enough. Aprt from choosing a partial agonist, solutions for this clinical problem might be found by lowering the dose and increasing the frequency of administration, but also by choosing a full agonist with another pharmacokinetic profile. Figure 4 shows that in this way the duration of action (~0% effect) can be prolonged from 6 to 24 units of time without a significant increase in the side-effects. By choosing such an agonist the rate of recovery (and hence the risk of rebound) is also decreased. It should be stressed that this choice cannot be made solely on the basis of terminal half-lives of the drugs but on the basis of the whole pharmaco kinetic profiles and the concentrationsffect relationships, since in these examples all half-lives of both
/
4
..I
r
:
,
,
.
’
ide have a physiological function? Gerald S. Marks, James F. Brien, Kanji Nakatsu and Brian E. McLaughlin ReccnNy mdothrlium-derived rclaxfng factor (EDEN) has been identified as nifric oxide. The source of the nitric oxfde is L-arginine, and the r-urginincnitric oxide pathwsy has been proposed to function as ~1 widespread transduction mechanism for the regulation of cell fmction and communication. Gerald Marks and colleagues suggest that carbon monoxide, which is formed en enousty from hemc catabolism snd which shares some of the chenricel an9 biofogfcelproperties of nitric oxide, may play a similar role. This would be achieued by carbon monoxide binding to the ironatomofthe hrme moiety of soluble guanylyl cydase and to the iron-sulfur centers of macrophage enzymes.
C. S. Marks, 1. F. B&II and K. N&&u are prafc*m ml It.E. MC&I&&I ir o Research ~~~~~~~~~~~~~ Unlucrsity,K;QMO~K7L 3N6, ‘Ontrrio, Canada.
References 1 Hwfely. W.,MarIin, 1.R. and Schach,P. (19SU) Trcnkc Phamad. Sri. II, 4Sz-456 2 !$pH&
1. (1968)BfopJJml. Drug Dfsp.
3 Sk, Y. J., Smcini, A. F., Scaf, A. H. j., Kmten, U. W. and Agoshm, S. (I%), An&h. AR@. 65,233-239
:
:.:!“y
Until recently, a pharmacologist/ toxicologist would have considered nitric oxide solely from the
drugs are the same (Fig. 1C caption). Comparing in oiao data from animals and humans with results from in niho experiments’ has been fruitful in improving insight into the mechanisms underlying the in vioo phenomena. It is difficult, however, to analyse the sig nificance of the model based on such comparisons. Simulations and graphical representation of the results may reveal the Emits within which conclusions may be drawn, for instance in the development of new drugs or the clinical use of both new and existing drugs. This contribution can therefore be considered as an extension of the article by Haefely et al.‘.
point of view of its toxic effects. Thus, in a recent edition of a highly respe&d toxicology text’, nitric oxide is listed as a vasculotoxic agent that causes vacuolation of arteriolar endothelial cells, edema, and a thickening of alveolar capillary membranes. In 1988, a surprised scientific world
learned that EDEF is the simple molecule nitric oxidti-‘, whose biosynthetic precursor is Larginine). The formation of nitric oxide from t-arginine was shown tooccurinavarietyofcelltyPes, and nitric oxide is now believed to have a variy of important physiological roles . It has been shown to relax arteries, arterioles and veins in several species, to inhibit platelet aggregation, and to cause platelet disaggregation’4. It appears likely that nitric oxide formation mediates the L-arginineinduced activation of macrophages to the micmbiocidal and tumoricidal states*“. Activated (cytotoxic) macrophages exhibit a variety of biochemical changes, including lntmc&dar iron loss, inhibition of mitochondrial respiration via inhibition of several enzymes of the mitochondrial electron transport chain, inhibition of the citric acid cycle enzyme aconitase, and inhibition of DNA synthesii. The possibiBty that nitric oxide, a toxic gas, might play an im ant physiological role woul 8”not have surp&d Pam&~ who propounded the well-known principle of toxicology in the 16th are century: ‘All substances - bI47J9Ii3O2.00 Ltd. WY) OlM
(0 199,. lOnevImsdsnn FuIdUum
TiPS - Mny 2992 [Vol. 221
186
a
histidine
poisons; there is none which is
not a poison; the right dose differentiates a poison and a remedy’ According to current ideas ‘, nitric oxide exerts its physiologi
b
histidine
which similarly both form sixcoordinate complexes with hemoproteins, produce the same movement of the iron atom into the plane of the porphyrin. Both carbon monoxide and nitric oxide appear to interact with Fe-S centers of a variety of enzymes to give enzyme inhibition. The interaction of nitric oxide with Fe-S centers results in degradation of the F&I centers, while the interaction of carbon monoxide with Fe4 centers dots not”. It is important to note that while the simple diatomic gases nitric oxide and carbon monoxide do share some properties there are some important differences. Llnlike carbon monoxide, nitric oxide has an additional electron that is readily lost to give the nitrosonium ionif. The nitrosonium ion in turn is involved in the formation of many nitric oxidenretal complexes. Nitric oxide undergoes both oxidative and reductive reactions. While carbou monoxide is stable to oxygen, nitric oxide reacts rapidly with oxygen to form nitrogen dioxide”. Nitric oxide has a higher affinity for hemoglobin than carbon monoxide, and also a greater rate of reaction with hemoglobin than carbon monoxide”. What is the evidence for a physiological role for carbon monoxide? Like nitric oxide, carbon monoxide has been shown to inhibit platelet aggmgation’5J6, apparently due to activation of guanylyl cyclase”. Recently it has been reported’s that carbon monoxide, although less potent than
nitric oxide, produces relaxation of dog femoral, carotid and coronary arterial preparations. The concentrations of carbon monoxide used in the organ bath were 10V6 tu 6 X io-” M. It has been proposed that endogenously produced carbon monoxide modulates blood vessel tone’s by activating guanylyl cyclase19. Eamos et nl.m have demonstrated that carbon monoxide relaxes rat coronary and aortic vascular smooth muscle preparations by an endothelium-independent mechanism. The concentrations of carbon monoxide employed were between 2.5% and 5% (Ref. 20). Carbon monoxide has been shown to cause a time-dependent increase in levels of ,cCMP in rat cultured aortic smooth muscle cells*‘. The elevation of cGMP could be correlated temporally with a decrease in Ca*+ levels and a relaxation of rat aortic smooth muscle preparations. The effects of carbon monoxide are reversible?. The lamb ductus venosus sphincter, like the ductus arteriosus, has been shown to rely upon an intramural cytochrome P-450 mechanism to develop its contractile tone=. Carbon monoxide reacts with the cytochrome P-450 system to relax completely the ductus. Although a proataglandin mechanism is believed to be responsible for the relaxation of the ductus venosus and ductus arteriosus, is it possible that endogenous carbon lmmoxide also plays a role in relaxation? Since carbon monoxide can inhibit FtS Centers of enzymes that are also inhibited by nitric oxide, is it possible that carbon monoxide might also exert microbiocldal and/or tumoricidal effects in macrophages? Endogenous sources If carbon monoxide exerts a physiological effect, it must have endogenous sources. At least two sources are available, one of which is lipid peroxidatlor?. Whether this carbon monoxide production is physiological and whether its production can be regulated by the cell is unknown. A second source is the metabolism of heme to biliverdin and carbon monoxide, catalysed by heme oxy genase (Fig. 2). The amount of carbon monoxide formed via this route is 0.4 ml h-’ (16.4 pmol h”)
TiPS - May 1991 [Vol. 12J . (Ref. 24). Heme oxygenase activity is high in organs rich in reticuloendothelial cells (i.e. spleen, bone marrow and liver), which is consistent with the role of the reticuloendothelial system in hemoglobin catabolism. Heme oxygenase is also present in brain, lung, kidney, small intestinal mucosa, chick embryo heart, and peritoneal and alveolar macrophages25 • Two distinct constitutive forms of heme oxygenase have recently been identified in a variety of rat tissues25 • One of the isozymes is inducible by heme and other agents, whereas the second form is not inducible. The inducible form of heme oxygenase is most abundant in tissues such as spleen and liver, in which the bulk of heme catabolism normally takes place. Recently, a physiological role has been postulated for one of the products of heme metabolism - biliverdin, the formation of which is catalysed by heme oxygenase. Biliverdin and bilirubin, formed by the reduction of biliverdin, have been shown to function as antioxidants26 . There are data that support the idea of a 'beneficial' role for bilirubin as a physiolo~ical, chain-breaking antioxidanf6. A portion of the carbon monoxide formed from heme degradation is exhaled and a portion is present as carboxyhemoglobin. In nonsmoking adults less than 1% of hemoglobin is present in the form of carboxyhemoglobin24 . As shown in Fig. 2, heme oxygenase catalyses the formation of carbon monoxide from heme metabolism; some of the carbon monoxide binds to hemoglobin, and we propose that some may bind to the heme moiety of cytosolic guanylyl cyclase and/or to Fe-S centers of other enzymes in order to exert physiological effects. Levels- of the inducible form of heme oxygenase are increased by its substrate heme as well as by hemoglobin2s. In addition, a variety of transition metals and heavy metals induce heme oxygenase activity. Moreover, a variety of hormones and chemicals, as well as endotoxin and interferon inducers, increase hepatic heme oxygenase activity2s. It is probable that heme functions as a physiological regulator of heme oxygenase, since maintenance of cel-
187 BILIRUBIN-IX~
NADP~ NADPH
~
L-
-'
BILIVERDIN-IX«
NADPH
0,
inhibition of enzymes C O - - -....~I with Fe-S
NADPH-cyWchrome c (f'-450) reductase
cenlres
co
GTP
cGMP
Fig. 2. Schematic illustration of the formation of blllverdln-lXa, iran (Fe) and carbon monoxide (CO) faliowing heme metaboNsm. catalysed by heme oxygenase acting in concert with NADPH-cy/OChrame c (P-450) reductase. Biliverdin-IXa is subsequently converted to blllrubln-IXIl! by biliverdin reductase. It is proposed that carbon monoxide binds to the iran atom of the heme moiety of cytosolic guanylyl cyclase in the cell in which CO is produced or in an adjacent cell. Activation ofguanylyl cyclase results in elevatedcGMP content, which leads via a series ofenzymic reactions to the physiological etfect, e.g. smooth muscle relaxation. It is proposed that in macraphages CO activates guanylyl cyclase and elevates cGMP content, thereby changing metabolism and ultimately facilitating micrabiocidal and/or tumaricidal effects. In addition, CO produced by macraphages binds to F~ centers of enzymes, causing inhibition, and thereby facilitates micrabiacidal and/or tumaricidal effects.
lular heme levels within narrow limits is important. Sn-protoporphyrin is the most potent of a variety of metal-protoporphyrin chelates that inhibit heme oxygenase 25 . Rat glioma cell heme oxygenase has been shown to be a heat shock protein27. It has recently been reported that the human 32 kDa stress protein induced by exposure to arsenite and cadmium ions is heme oxygenase28; however, the functional role of heme oxygenase under stress conditions remains to be elucidated. Since the metabolism of heme, catalysed by heme oxygenase, leads to metabolites that can provide antioxidant, antiplatelet and vasodilator activity (and conceivably microbiocidal activity), is it possible that this is the functional role of heme oxygenase under stress conditions? Moncada and associates8 ,9,29 have proposed 'that the L-arginine-nitric oxide pathway is a widespread transduction mechanism for the regulation of cell function and communication'.
Is it possible that the degradation of heme to biliverdin and carbon monoxide may also play a role in the regulation of cell function and communication? The following experimental investigations could be conducted to test the hypothesis that carbon monoxide has a physiological function: • Measurement of carbon monoxide production by various tissues. A sensitive gas-liquid chromatographic method for carbon monoxide is now available 30 • • Measurement of the tissue distribution of the two heme oxygenase isozymes by using sensitive immunological techniques. • Comparison of the physiological properties of carbon monoxide and nitric oxide. • Quantification of the amount of carbon monoxide formed in tissues, to determine if it is sufficient to produce physiological effects. • Comparison of carbon monoxide with nitric oxide as a mediator of cytotoxicity in activated macrophages.
TiPS -May 1991 [Vool.121
108 ?? Determination of the physiological effects of inhibitors of heme oxygenase.
A&Zowl*+Xi&s We thank Dr Laura A. Andersson, Oregon Graduate Institute, Beaverton, Orenon for information on the coiformation of the iron atom in hemoproteins following interaction with nitric oxide or carbon monoxide. This work was supported by the Heart and Stroke Foundation of Ontario (Grant T MB), Canada. References 1 Batazs, B., Hanig, J. P. and Herman, H. (1986) in C~wwtt ad
Doulf’s Toxicology. 7%~ Basic Science of Poisons (3rd cdn) (Kbassm, C. D., Amdur, M. 0. and Doutl, J., eds), pp. 387-411, Macmiltan 2 Fuw, R 1. (19BB)in Vasodilafior IVanhntrwP P_. ____, M.. l=c.\. I .-...__.._, ___,, ml TT_ @&4& Raven
/
3 Ignarro, L. J., Bytes, R. E. and Woods, K. S. (I%?& in Vosodilation(Vanhoutte, Raven P. M.,ed.). pp. 4274, 4 Palmer, R. M. J., Ferrig, A. C. and
Moncada, S. (1987) Nahrre 327,524-526 5 palmer, R M. J., Ashton, D. S. and Moncada, S. (1968) N&trt 333,664+X6 6 Madetta, hi. A. (1989) Trends Biochcm. Sci. 14, -92 7 Kadomski. M. W.. Palmer. R M. 1. and Muncada, S. (19B7) Lacet ii, lB57-1058 8 Moncada, S., Radomsk M. W. and Palmer, R M. J. (l!IBB)Biocbem.Pbamocol. 37,2495-2501 ^ .y Moncada, S., palm, R M J. Gd k$gsgsl~~19B9) Biocbem.Pburmacol. 10 Hibba, J. B., Jr, Taintor, R R, Varrin, 2. and Rachlin, E. M. (1988) Biocbem. Biopbys.Res.Cornman.157,B7-94 11 Ignarro, L. J. (19B9)Scmin. Hnnafol. 26, 63-76 12 Meyer, J. (1961) Arch. Biocbcm.Biopbys. 210; 246-256 13 Cotton, F. A. and Wilkinson, C. (1962) Advanced Inorganic Cbcmistty, pp. 225-260; 611-631, lnterscimce 14 Meyer, M., Schuster, K-D., Schulz, H., Mohr, M. and Ptiper, J. (1990) 1. A&. Pbysiol.68.2344-2357 15 Mansouri, A. and Peny, C. A. (1962) l’brvmb. Heemostesis 48,2B&2BB 16 Manscw!, b and Peq C. A. wB4) Experientia40,515-517 17 Briine, 8. and Ullrich, V. (1987) Noi. Pbarmacol.32.497-504 18 Vedemikov, Y. P., G&r, T. and Vanin, A. F. (19B9) Biomd. Biocbim. Acta 48,
601-603 19 G&er, 7.. Vedemikov, Y. P. and Li. D. S. (1990) Biomcd. Biocbim.Acta 4%
zo
Et% s.,Lin, H. ydM%the (198q) Biochem.
I. I.
3R: 1368-1370 21 Un, H. and McGrath. J, J. (19ee) tifi Sk 43.1813-1816
22 Adeagbo, A. S. 0. et al. (1990)J. Pbannn-
_A Lxp. e.I%-.u.3 aW! CYI. rnrr. &l&WJ917 23 Wolff, D. G. and Bidkk, W. R (1976) Biocbcm. Biopbp. Res. Commun. 73,
24 Smith, R P. (1966) in Gwerett end
Doull’s roxic&gy. The Basic science of Poisons (3rd edn) (Kiaawn, C. D..
25,349-369 26 Stocker, R., Yasnamc40,Y.. McDonagh, A. F., Glazer, A N. and Ames, 8. N. (19B7) 5cienc. 235,x&3-1046 27 Shib;ilhm, S., Mtilter, R M. and Taguchi. H. (19B7) J. Bid. Chem. 262,
l2689-l2w2 28 Taketmt. S.. Kohno. Ii., Ymhinua, 7.
md T&m, R @B9j FEBS Len: 245, 173-176 29 Moncada, S. (1990) Eur. /. Pkmttwccd.
163, N-159 30 Vreaun, J. H. and !3tevenwn, D. K. (1988) Awl. Biocbem.16B, 31G3tl
GlDWth~doKoIpIK
~iauuicrprtbrnyr
Signalling targets for anticancer drug &ve~opment Garth Powis Intracellularsignallingpathways mediatingthe effects of grvwth factors and oncogenes on ccl1growth and trunsformationpresent a challenging new class of target sitesfor anticancer drug devebpment. Several dnqs are already available that may act in this wuy, includingdrugs that act on prvtein serinel threonine kinases. protein fyrusine kinases and phospholfpaae C, as well as inhibitors of myo-inositol signalfing.As our undetisnding of the signding pathways involved in growth con&Z increases, new sites fo.=phunnacofogicaJ intervention will become apparent. Garth Powls reviews the evidence that this approach may eventually lead to new, more selective drugs for treating cancer. Traditionally, anticancer drugs have been targeted to inhibit DNA synthesis and function. Although drug treatment has given limited success against some rapidly growing cancers, the majority of human cancers remain refractory and there is an urgent &p$~&~$,~&~~~~~~~~f 200 First Shrti Soutbwcst, Rochester, 55% USA.
Mli
need for more effective anticancer drugs. A new approach is to use the signalling pathways that mediate the effects of growth factors and oncogeneson cell prolifnation as the molecular targets for anticancer drug development. The purpose of this review is to summarize briefly the rationale for this approach, to describe some drugs that may act in this way and to indicate potential new target sites for drug development.
G~wth hctont acton cell surfacereceptorstoactivateanumber of intracellular signal&j as&es that lead to cell pro&ration thatarelnthroughmechanlsms completely understood (see Pig. 1). More than tine growth factor is requhvd for normal cell proliferation, which may be related to a requirement for interactions between the signalling pathways. Proto-ontxgenes, genes that control normal cell growth, frequently encode proteins that are components of signall@ pathways. The 0verexpicLwion or mutation of aproto-oncogenetoglveanoncogene can lead to the conatitutive activation of a sign&lx~ pathway, whereby the transformed cell receives a signal for unrestrained growth. The inhibition of a slgnalling pathway that is activated by an oncOgene could lead to the reversal of ce.ll transformation. Since there are man pathways in growthfactorsl gnrl;ing , it might be possible to inhlblt a pae&w~y activated by an 0 leave intact other sign 2iYn ng pathways necessuy for mnmrl cell growth. Furthermore ,slncethere are signalling pathways apparently specifically associated with
185
ists, due to the much smaller dose of the full agonist required for 50% effect in the high density cells. An advantage of this low dose might be a smaller risk of side&ects not related to an interaction with the G_4BA receptor complex’. The apparent difference in selectivity of the three agonists disappears almost fully, taking the peak effect ratio as an index (Fig. 3&ures 2F-H and Fig. 3 show that the preference for the effects in high density cells diminishes if the peak effect of the full agonist falls upon the highest part of the log concentration-effect curve, whereas it remains unchanged with the partlal agonist (at = 0.1). On the other hand, if the therapeutic aim is mediated by lower density cells (central muscle relaxation, sleep, anesthesia), this can only be achieved with a full agonist benxodiazepine. These simulations show that, indeed, pa&ally agonistic benzodianepines have a greater specificity for the effects induced via
the low density cells. This greater selectivity, however, only becomes obvious when doses almost approaching E, are administered. Such doses may be required because the anxidytic action does not last long enough. Aprt from choosing a partial agonist, solutions for this clinical problem might be found by lowering the dose and increasing the frequency of administration, but also by choosing a full agonist with another pharmacokinetic profile. Figure 4 shows that in this way the duration of action (~0% effect) can be prolonged from 6 to 24 units of time without a significant increase in the side-effects. By choosing such an agonist the rate of recovery (and hence the risk of rebound) is also decreased. It should be stressed that this choice cannot be made solely on the basis of terminal half-lives of the drugs but on the basis of the whole pharmaco kinetic profiles and the concentrationsffect relationships, since in these examples all half-lives of both
/
4
..I
r
:
,
,
.
’
ide have a physiological function? Gerald S. Marks, James F. Brien, Kanji Nakatsu and Brian E. McLaughlin ReccnNy mdothrlium-derived rclaxfng factor (EDEN) has been identified as nifric oxide. The source of the nitric oxfde is L-arginine, and the r-urginincnitric oxide pathwsy has been proposed to function as ~1 widespread transduction mechanism for the regulation of cell fmction and communication. Gerald Marks and colleagues suggest that carbon monoxide, which is formed en enousty from hemc catabolism snd which shares some of the chenricel an9 biofogfcelproperties of nitric oxide, may play a similar role. This would be achieued by carbon monoxide binding to the ironatomofthe hrme moiety of soluble guanylyl cydase and to the iron-sulfur centers of macrophage enzymes.
C. S. Marks, 1. F. B&II and K. N&&u are prafc*m ml It.E. MC&I&&I ir o Research ~~~~~~~~~~~~~ Unlucrsity,K;QMO~K7L 3N6, ‘Ontrrio, Canada.
References 1 Hwfely. W.,MarIin, 1.R. and Schach,P. (19SU) Trcnkc Phamad. Sri. II, 4Sz-456 2 !$pH&
1. (1968)BfopJJml. Drug Dfsp.
3 Sk, Y. J., Smcini, A. F., Scaf, A. H. j., Kmten, U. W. and Agoshm, S. (I%), An&h. AR@. 65,233-239
:
:.:!“y
Until recently, a pharmacologist/ toxicologist would have considered nitric oxide solely from the
drugs are the same (Fig. 1C caption). Comparing in oiao data from animals and humans with results from in niho experiments’ has been fruitful in improving insight into the mechanisms underlying the in vioo phenomena. It is difficult, however, to analyse the sig nificance of the model based on such comparisons. Simulations and graphical representation of the results may reveal the Emits within which conclusions may be drawn, for instance in the development of new drugs or the clinical use of both new and existing drugs. This contribution can therefore be considered as an extension of the article by Haefely et al.‘.
point of view of its toxic effects. Thus, in a recent edition of a highly respe&d toxicology text’, nitric oxide is listed as a vasculotoxic agent that causes vacuolation of arteriolar endothelial cells, edema, and a thickening of alveolar capillary membranes. In 1988, a surprised scientific world
learned that EDEF is the simple molecule nitric oxidti-‘, whose biosynthetic precursor is Larginine). The formation of nitric oxide from t-arginine was shown tooccurinavarietyofcelltyPes, and nitric oxide is now believed to have a variy of important physiological roles . It has been shown to relax arteries, arterioles and veins in several species, to inhibit platelet aggregation, and to cause platelet disaggregation’4. It appears likely that nitric oxide formation mediates the L-arginineinduced activation of macrophages to the micmbiocidal and tumoricidal states*“. Activated (cytotoxic) macrophages exhibit a variety of biochemical changes, including lntmc&dar iron loss, inhibition of mitochondrial respiration via inhibition of several enzymes of the mitochondrial electron transport chain, inhibition of the citric acid cycle enzyme aconitase, and inhibition of DNA synthesii. The possibiBty that nitric oxide, a toxic gas, might play an im ant physiological role woul 8”not have surp&d Pam&~ who propounded the well-known principle of toxicology in the 16th are century: ‘All substances - bI47J9Ii3O2.00 Ltd. WY) OlM
(0 199,. lOnevImsdsnn FuIdUum
TiPS - Mny 2992 [Vol. 221
186
a
histidine
poisons; there is none which is
not a poison; the right dose differentiates a poison and a remedy’ According to current ideas ‘, nitric oxide exerts its physiologi
b
histidine
which similarly both form sixcoordinate complexes with hemoproteins, produce the same movement of the iron atom into the plane of the porphyrin. Both carbon monoxide and nitric oxide appear to interact with Fe-S centers of a variety of enzymes to give enzyme inhibition. The interaction of nitric oxide with Fe-S centers results in degradation of the F&I centers, while the interaction of carbon monoxide with Fe4 centers dots not”. It is important to note that while the simple diatomic gases nitric oxide and carbon monoxide do share some properties there are some important differences. Llnlike carbon monoxide, nitric oxide has an additional electron that is readily lost to give the nitrosonium ionif. The nitrosonium ion in turn is involved in the formation of many nitric oxidenretal complexes. Nitric oxide undergoes both oxidative and reductive reactions. While carbou monoxide is stable to oxygen, nitric oxide reacts rapidly with oxygen to form nitrogen dioxide”. Nitric oxide has a higher affinity for hemoglobin than carbon monoxide, and also a greater rate of reaction with hemoglobin than carbon monoxide”. What is the evidence for a physiological role for carbon monoxide? Like nitric oxide, carbon monoxide has been shown to inhibit platelet aggmgation’5J6, apparently due to activation of guanylyl cyclase”. Recently it has been reported’s that carbon monoxide, although less potent than
nitric oxide, produces relaxation of dog femoral, carotid and coronary arterial preparations. The concentrations of carbon monoxide used in the organ bath were 10V6 tu 6 X io-” M. It has been proposed that endogenously produced carbon monoxide modulates blood vessel tone’s by activating guanylyl cyclase19. Eamos et nl.m have demonstrated that carbon monoxide relaxes rat coronary and aortic vascular smooth muscle preparations by an endothelium-independent mechanism. The concentrations of carbon monoxide employed were between 2.5% and 5% (Ref. 20). Carbon monoxide has been shown to cause a time-dependent increase in levels of ,cCMP in rat cultured aortic smooth muscle cells*‘. The elevation of cGMP could be correlated temporally with a decrease in Ca*+ levels and a relaxation of rat aortic smooth muscle preparations. The effects of carbon monoxide are reversible?. The lamb ductus venosus sphincter, like the ductus arteriosus, has been shown to rely upon an intramural cytochrome P-450 mechanism to develop its contractile tone=. Carbon monoxide reacts with the cytochrome P-450 system to relax completely the ductus. Although a proataglandin mechanism is believed to be responsible for the relaxation of the ductus venosus and ductus arteriosus, is it possible that endogenous carbon lmmoxide also plays a role in relaxation? Since carbon monoxide can inhibit FtS Centers of enzymes that are also inhibited by nitric oxide, is it possible that carbon monoxide might also exert microbiocldal and/or tumoricidal effects in macrophages? Endogenous sources If carbon monoxide exerts a physiological effect, it must have endogenous sources. At least two sources are available, one of which is lipid peroxidatlor?. Whether this carbon monoxide production is physiological and whether its production can be regulated by the cell is unknown. A second source is the metabolism of heme to biliverdin and carbon monoxide, catalysed by heme oxy genase (Fig. 2). The amount of carbon monoxide formed via this route is 0.4 ml h-’ (16.4 pmol h”)
TiPS - May 1991 [Vol. 12J . (Ref. 24). Heme oxygenase activity is high in organs rich in reticuloendothelial cells (i.e. spleen, bone marrow and liver), which is consistent with the role of the reticuloendothelial system in hemoglobin catabolism. Heme oxygenase is also present in brain, lung, kidney, small intestinal mucosa, chick embryo heart, and peritoneal and alveolar macrophages25 • Two distinct constitutive forms of heme oxygenase have recently been identified in a variety of rat tissues25 • One of the isozymes is inducible by heme and other agents, whereas the second form is not inducible. The inducible form of heme oxygenase is most abundant in tissues such as spleen and liver, in which the bulk of heme catabolism normally takes place. Recently, a physiological role has been postulated for one of the products of heme metabolism - biliverdin, the formation of which is catalysed by heme oxygenase. Biliverdin and bilirubin, formed by the reduction of biliverdin, have been shown to function as antioxidants26 . There are data that support the idea of a 'beneficial' role for bilirubin as a physiolo~ical, chain-breaking antioxidanf6. A portion of the carbon monoxide formed from heme degradation is exhaled and a portion is present as carboxyhemoglobin. In nonsmoking adults less than 1% of hemoglobin is present in the form of carboxyhemoglobin24 . As shown in Fig. 2, heme oxygenase catalyses the formation of carbon monoxide from heme metabolism; some of the carbon monoxide binds to hemoglobin, and we propose that some may bind to the heme moiety of cytosolic guanylyl cyclase and/or to Fe-S centers of other enzymes in order to exert physiological effects. Levels- of the inducible form of heme oxygenase are increased by its substrate heme as well as by hemoglobin2s. In addition, a variety of transition metals and heavy metals induce heme oxygenase activity. Moreover, a variety of hormones and chemicals, as well as endotoxin and interferon inducers, increase hepatic heme oxygenase activity2s. It is probable that heme functions as a physiological regulator of heme oxygenase, since maintenance of cel-
187 BILIRUBIN-IX~
NADP~ NADPH
~
L-
-'
BILIVERDIN-IX«
NADPH
0,
inhibition of enzymes C O - - -....~I with Fe-S
NADPH-cyWchrome c (f'-450) reductase
cenlres
co
GTP
cGMP
Fig. 2. Schematic illustration of the formation of blllverdln-lXa, iran (Fe) and carbon monoxide (CO) faliowing heme metaboNsm. catalysed by heme oxygenase acting in concert with NADPH-cy/OChrame c (P-450) reductase. Biliverdin-IXa is subsequently converted to blllrubln-IXIl! by biliverdin reductase. It is proposed that carbon monoxide binds to the iran atom of the heme moiety of cytosolic guanylyl cyclase in the cell in which CO is produced or in an adjacent cell. Activation ofguanylyl cyclase results in elevatedcGMP content, which leads via a series ofenzymic reactions to the physiological etfect, e.g. smooth muscle relaxation. It is proposed that in macraphages CO activates guanylyl cyclase and elevates cGMP content, thereby changing metabolism and ultimately facilitating micrabiocidal and/or tumaricidal effects. In addition, CO produced by macraphages binds to F~ centers of enzymes, causing inhibition, and thereby facilitates micrabiacidal and/or tumaricidal effects.
lular heme levels within narrow limits is important. Sn-protoporphyrin is the most potent of a variety of metal-protoporphyrin chelates that inhibit heme oxygenase 25 . Rat glioma cell heme oxygenase has been shown to be a heat shock protein27. It has recently been reported that the human 32 kDa stress protein induced by exposure to arsenite and cadmium ions is heme oxygenase28; however, the functional role of heme oxygenase under stress conditions remains to be elucidated. Since the metabolism of heme, catalysed by heme oxygenase, leads to metabolites that can provide antioxidant, antiplatelet and vasodilator activity (and conceivably microbiocidal activity), is it possible that this is the functional role of heme oxygenase under stress conditions? Moncada and associates8 ,9,29 have proposed 'that the L-arginine-nitric oxide pathway is a widespread transduction mechanism for the regulation of cell function and communication'.
Is it possible that the degradation of heme to biliverdin and carbon monoxide may also play a role in the regulation of cell function and communication? The following experimental investigations could be conducted to test the hypothesis that carbon monoxide has a physiological function: • Measurement of carbon monoxide production by various tissues. A sensitive gas-liquid chromatographic method for carbon monoxide is now available 30 • • Measurement of the tissue distribution of the two heme oxygenase isozymes by using sensitive immunological techniques. • Comparison of the physiological properties of carbon monoxide and nitric oxide. • Quantification of the amount of carbon monoxide formed in tissues, to determine if it is sufficient to produce physiological effects. • Comparison of carbon monoxide with nitric oxide as a mediator of cytotoxicity in activated macrophages.
TiPS -May 1991 [Vool.121
108 ?? Determination of the physiological effects of inhibitors of heme oxygenase.
A&Zowl*+Xi&s We thank Dr Laura A. Andersson, Oregon Graduate Institute, Beaverton, Orenon for information on the coiformation of the iron atom in hemoproteins following interaction with nitric oxide or carbon monoxide. This work was supported by the Heart and Stroke Foundation of Ontario (Grant T MB), Canada. References 1 Batazs, B., Hanig, J. P. and Herman, H. (1986) in C~wwtt ad
Doulf’s Toxicology. 7%~ Basic Science of Poisons (3rd cdn) (Kbassm, C. D., Amdur, M. 0. and Doutl, J., eds), pp. 387-411, Macmiltan 2 Fuw, R 1. (19BB)in Vasodilafior IVanhntrwP P_. ____, M.. l=c.\. I .-...__.._, ___,, ml TT_ @&4& Raven
/
3 Ignarro, L. J., Bytes, R. E. and Woods, K. S. (I%?& in Vosodilation(Vanhoutte, Raven P. M.,ed.). pp. 4274, 4 Palmer, R. M. J., Ferrig, A. C. and
Moncada, S. (1987) Nahrre 327,524-526 5 palmer, R M. J., Ashton, D. S. and Moncada, S. (1968) N&trt 333,664+X6 6 Madetta, hi. A. (1989) Trends Biochcm. Sci. 14, -92 7 Kadomski. M. W.. Palmer. R M. 1. and Muncada, S. (19B7) Lacet ii, lB57-1058 8 Moncada, S., Radomsk M. W. and Palmer, R M. J. (l!IBB)Biocbem.Pbamocol. 37,2495-2501 ^ .y Moncada, S., palm, R M J. Gd k$gsgsl~~19B9) Biocbem.Pburmacol. 10 Hibba, J. B., Jr, Taintor, R R, Varrin, 2. and Rachlin, E. M. (1988) Biocbem. Biopbys.Res.Cornman.157,B7-94 11 Ignarro, L. J. (19B9)Scmin. Hnnafol. 26, 63-76 12 Meyer, J. (1961) Arch. Biocbcm.Biopbys. 210; 246-256 13 Cotton, F. A. and Wilkinson, C. (1962) Advanced Inorganic Cbcmistty, pp. 225-260; 611-631, lnterscimce 14 Meyer, M., Schuster, K-D., Schulz, H., Mohr, M. and Ptiper, J. (1990) 1. A&. Pbysiol.68.2344-2357 15 Mansouri, A. and Peny, C. A. (1962) l’brvmb. Heemostesis 48,2B&2BB 16 Manscw!, b and Peq C. A. wB4) Experientia40,515-517 17 Briine, 8. and Ullrich, V. (1987) Noi. Pbarmacol.32.497-504 18 Vedemikov, Y. P., G&r, T. and Vanin, A. F. (19B9) Biomd. Biocbim. Acta 48,
601-603 19 G&er, 7.. Vedemikov, Y. P. and Li. D. S. (1990) Biomcd. Biocbim.Acta 4%
zo
Et% s.,Lin, H. ydM%the (198q) Biochem.
I. I.
3R: 1368-1370 21 Un, H. and McGrath. J, J. (19ee) tifi Sk 43.1813-1816
22 Adeagbo, A. S. 0. et al. (1990)J. Pbannn-
_A Lxp. e.I%-.u.3 aW! CYI. rnrr. &l&WJ917 23 Wolff, D. G. and Bidkk, W. R (1976) Biocbcm. Biopbp. Res. Commun. 73,
24 Smith, R P. (1966) in Gwerett end
Doull’s roxic&gy. The Basic science of Poisons (3rd edn) (Kiaawn, C. D..
25,349-369 26 Stocker, R., Yasnamc40,Y.. McDonagh, A. F., Glazer, A N. and Ames, 8. N. (19B7) 5cienc. 235,x&3-1046 27 Shib;ilhm, S., Mtilter, R M. and Taguchi. H. (19B7) J. Bid. Chem. 262,
l2689-l2w2 28 Taketmt. S.. Kohno. Ii., Ymhinua, 7.
md T&m, R @B9j FEBS Len: 245, 173-176 29 Moncada, S. (1990) Eur. /. Pkmttwccd.
163, N-159 30 Vreaun, J. H. and !3tevenwn, D. K. (1988) Awl. Biocbem.16B, 31G3tl
GlDWth~doKoIpIK
~iauuicrprtbrnyr
Signalling targets for anticancer drug &ve~opment Garth Powis Intracellularsignallingpathways mediatingthe effects of grvwth factors and oncogenes on ccl1growth and trunsformationpresent a challenging new class of target sitesfor anticancer drug devebpment. Several dnqs are already available that may act in this wuy, includingdrugs that act on prvtein serinel threonine kinases. protein fyrusine kinases and phospholfpaae C, as well as inhibitors of myo-inositol signalfing.As our undetisnding of the signding pathways involved in growth con&Z increases, new sites fo.=phunnacofogicaJ intervention will become apparent. Garth Powls reviews the evidence that this approach may eventually lead to new, more selective drugs for treating cancer. Traditionally, anticancer drugs have been targeted to inhibit DNA synthesis and function. Although drug treatment has given limited success against some rapidly growing cancers, the majority of human cancers remain refractory and there is an urgent &p$~&~$,~&~~~~~~~~f 200 First Shrti Soutbwcst, Rochester, 55% USA.
Mli
need for more effective anticancer drugs. A new approach is to use the signalling pathways that mediate the effects of growth factors and oncogeneson cell prolifnation as the molecular targets for anticancer drug development. The purpose of this review is to summarize briefly the rationale for this approach, to describe some drugs that may act in this way and to indicate potential new target sites for drug development.
G~wth hctont acton cell surfacereceptorstoactivateanumber of intracellular signal&j as&es that lead to cell pro&ration thatarelnthroughmechanlsms completely understood (see Pig. 1). More than tine growth factor is requhvd for normal cell proliferation, which may be related to a requirement for interactions between the signalling pathways. Proto-ontxgenes, genes that control normal cell growth, frequently encode proteins that are components of signall@ pathways. The 0verexpicLwion or mutation of aproto-oncogenetoglveanoncogene can lead to the conatitutive activation of a sign&lx~ pathway, whereby the transformed cell receives a signal for unrestrained growth. The inhibition of a slgnalling pathway that is activated by an oncOgene could lead to the reversal of ce.ll transformation. Since there are man pathways in growthfactorsl gnrl;ing , it might be possible to inhlblt a pae&w~y activated by an 0 leave intact other sign 2iYn ng pathways necessuy for mnmrl cell growth. Furthermore ,slncethere are signalling pathways apparently specifically associated with