Invited Symposium: Quinones and Other Reactive Oxygen Species in Neurobiologic, Apoptotic, and Neurotoxic Processes
What We Know & What We Don't Know
A. NORMAL FUNCTIONS(1) What we know.
(i) Catecholamine ortho-quinones (CAQs) are necessary precursors in the biosynthesis of neuromelanin (NM) (Bindoli et al. 1992; Costa et al, 1992). Therefore the presence of NM in a neuron necessitates the presence of the relevant CAQ as well. NM occurs in abundance in the dopaminergic cells of the substantia nigra pars compacta (SNpc) and ventral tegmental area (VTA), in the noradrenergic cells of the locus coeruleus (LC) and is also found in the A1-A3 noradrenergic nuclei in the medulla (Bogerts, 1981; Saper & Petito, 1982). In the case of the C1- C3 adrenergic nuclei in the medulla Gai et al (1993) have presented evidence that very few of the adrenergic neurons are pigmented. However, adrenochrome may be formed but not metabolized as far as neuromelanin. Its chief metabolite is adrenolutin which does not form neuromelanin readily (Solano, 1998).
Neuromelanin is not found in these cells at birth but appears at around 6 months of age and then steadily accumulates during life. Therefore dopaminochrome, and noradrenochrome must (and adrenochrome may) occur in the brain together with related compounds on the NM biosynthetic pathway and must be continually produced during life.
(ii) The presence of 5-cysteinyl dopamine in brain (Carlsson et al. 1994) also indicates that DAQs are present too since it is a metabolite of dopamine o-quinone.
(iii) The enzymes DT-diaphorase and NADPH P450 cytochrome reductase that interconvert DAQs are present in brain (Segura-Aguilar, 1996).
(iv) There is now considerable evidence that dopamine neurotoxicity is mediated by DAQs acting not on DA receptors but on NMDA glutamate receptors (Michel and Hefti, 1990; Cadet and Kahler, 1994; Ben-Sacher et al. 1995). This suggests that DAQs are located so that they can reach NMDA receptors.
(v) NM is confined to cell bodies and large proximal dendrites.
(vi) The o-semiquinone free radical form of CAQs are potent neurotoxins, bind covalently to SH groups on proteins and inactivate enzymes such as COMT.
(vii) Very little is known about the basic pharmacology of CAQs derived from dopamine and norepinephrine. More is known about adrenochrome (the CAQ derived from adrenaline). It inhibits COMT (Urabe et al. 1994), promotes the synthesis of prostaglandins in brain tissue in vitro (White and Wu, 1975) promotes the secretion of nerve growth factor by L-M cells (Napolitano et al. 1993), inhibits hexokinase and succinic dehydrogenase (Grof, 1963), and acts as a powerful stimulant of guanylcyclase activity in cell free systems (Liang and Sacktor,1978). But it is not clear which, if any, of these have physiological significance. There is one report that 5,6-dihydroxyindole is cytotoxic (Urabe et al. 1994). It also inhibits certain lipoxygenases (Napolitano et al., 1993). More work done in this area may be expected to produce some data to work on.
(viii) Adrenochrome (or some metabolite such as adrenolutin—the 5,6-dihydroxyindole derivative of adrenaline) has been shown to be psychotomimetic in humans (Hoffer et al. 1954; Schwartz et al. 1956a; Taubman and Jantz, 1957; Grof, 1963) and to produce abnormal EEG patterns in humans and abnormal EEG changes and behaviors in animals similar to those produced by LSD (Schwartz et al, 1956b; Taborsky, 1968).
(ix) Outside the brain adrenochrome has been reported to be the principle metabolite of adrenaline by polymorphonuclear leucocytes (Matthews et al, 1985). Dhalla et al (1989) have reported that adrenolutin, but not adrenochrome, occurs in normal plasma. This claim certainly needs confirmation.
(x) Catecholamines in the brain are protected against auto-oxidation by cellular antioxidants, in particular ascorbate, glutathione (GSH) and possibly other antioxidants such as carnosine, uric acid, metatonin and others.
(2) What we don’t know.
(i) Where are CAQs formed?
Since NM is found only in the cell body, presumably the final stage of its synthesis—the polymerization of 5,6-dihydroxyindole—must occur in the cytoplasm of the neurone. There is no evidence that NM is transported anywhere. However, any of its soluble precursors may be formed elsewhere and transported to the cell body. Furthermore CAQs formed elsewhere—in the synapse for example—may be metabolized to non-toxic 0-methylated derivatives of the o-hydroquinone or of 5,6-dihydroxyindole, which are then excreted in the urine. Thus CAQs formed do not have to end up as neuromelanin. As we will see one important possible site of DAQ formation is inside the glutamate synapse. Catecholamines inside their synaptic vesicles are well protected against oxidation by high ascorbate and carnosine levels. But the cytoplasm and the synaptic cleft offer more oxidative environments. Dopamine is released from dendrites in the SN and VTA (Nirenberg 1997) and so could oxidize locally to form NM.
Enzymes capable of oxidizing catecholamines to o-quinones in vitro include xanthine oxidase (whose levels in brain are low), dopamine beta-monoxygenase, monoamine oxidase and prostaglandin H synthase (cyclo-oxygenase). The last is an important enzyme in the post-synaptic cascade of the NMDA receptor. So where is dopamine likely to come into contact with PGH synthase? One locus might be the post-synaptic cascade of the NMDA receptor in the synapses formed by glutamatergic axons, arising from the prefrontal cortex, on the dopaminergic axons in the striatum (Burns et al. 1994; Grace et al. 1998). Another locus could be glutamatergic synapses on dopaminergic neuronal bodies.
But where could dopamine come into contact with PGH synthase inside the glutamatergic neuron? A large proportion of dopamine receptors, when activated by dopamine, are internalized into the post-synaptic neuron by endocytosis (Dumartin et al. 1998) together with the bound neurotransmitter (Koenig and Edwardson 1997). This endocytosis was described by Dumartin et al. (1998) as “acute” (within ~4 minutes) and “dramatic”. So this might enable dopamine and PGHsynthase to interact. But it is uncertain if the small amount of dopamine that could be internalized by this mechanism could be effective. However, these redox cycling mechanisms act rather like enzymes and allow a small quantity of the recycling agent chemically to alter a large quantity of what is being recycled (Segura-Aguilar 1998). Polypeptide neuromodulators internalized by endocytosis are known to take part in the subsequent post-synaptic metabolic pathways (Koenig & Edwardson, 1997). However, which of these enzymes, if any, actually oxidizes catecholamines in vivo remains to be determined.
(ii) What is the normal function(s), if any, of CAQs?
In the past NM has been regarded as a mere ‘waste-product’ of catecholamine metabolism (Bogerts, 1981). However, its precursors may exert some normal function in the brain—if so, what? One role may have to do with synaptic plasticity. One has to ask oneself what normal function could potent neurotoxins like catecholamine o-semiquinones have if not deletion of unwanted tissue such as spines that are no longer required. Glutamate synapses are carried mainly on dendritic spines. Spines are highly dynamic structures and are continually being pruned and formed (Quartz and Sejnowski 1997). In the NMDA receptor post-synaptic cascade the activation of PGHs and NOs both lead to the release of AA as well as large amounts of ROS including H2O2. NO, AA and H2O2 are all freely diffusible molecules and can diffuse back into the synaptic cleft. NO exists mainly as the strongly oxidant neurotoxic nitric oxide radical NO.. Furthermore NO reacts with the superoxide anion to form the potent oxidant peroxynitrite. Any of these oxidant molecules could contribute to spine pruning.
Another key factor in this redox equation may be catecholamines, particularly dopamine. Many glutamate synapses have a dopaminergic bouton-en-passage closely attached to one side allowing diffusion of dopamine released into the glutamate synapse. Moreover it has recently been discovered (Sulzer et al, 1998) that many dopaminergic neurones have dopaminergic boutons-en-passage close to the target neuron but, contrary to Dale’s principle, have glutamatergic terminals that synapse with proximal dendrites of the target neuron via asymmetric synapses. Thus stimulation of these VTA neurons results in a fast glutamate mediated response in the target neuron and a slow dopamine-mediated response which acts partly presynaptically via DA receptors to inhibit glutamate release and may act partly by a redox mechanism on the glutamatergic synapse by the mechanism suggested in this paper. Catecholamines are potent antioxidants (Liu & Mori, 1993). Dopamine is released widely in the brain, in particular in the frontal cortex, following the receipt of positive reinforcement by the organism. Therefore, such release could tilt the redox balance inside the glutamate synapse towards the neuroprotective antioxidant side allowing synaptic growth promoted, for example, by metabotropic glutamate receptors. This would tend to promote those circuits active at the time of glutamate release. Whereas circuits active during periods of low dopamine release would lack this antioxidant growth-promoting effect and could even be deleted by an excess of oxidant molecules. This biochemical system may play a role in synaptic plasticity, neurocomputation and learning (see Smythies, 1997 for details). This mechanism is proposed as an addition not a replacement in addition to the many specific pre-and post-synaptic dopamine receptor-mediated mechanisms at the glutamate synapse.
However, in the case of most physiological antioxidants like vitamin E and ascorbate, the oxidized form that results from interaction with the free radical is not itself a highly toxic molecule, and, moreover, it is converted back into the active form by additional mechanisms involving a chain of antioxidants—e.g. glutathione, alpha-lipoic acid, caretenoids, and NADH, as well as vitamin E and ascorbate for each other (Chan 1993; Böhm et al. 1997). Dopamine at first sight might appear to be an unsatisfactory candidate for an antioxidant role, since dopamine oxidation leads eventually to the production of highly toxic free radical o-semiquinones, which cannot be reconverted into dopamine, as well as large amounts of new ROS. However, the first stage of dopamine oxidation is the conversion of dopamine to the uncyclized dopamine quinone. This stage is fully reversible by ambient antioxidants such as ascorbate and glutathione thus setting up a typical redox cycle. Dopamine quinone may then be further metabolized by three routes:—
Dopaminochrome may then be metabolized either of two routes. The first is by the enzyme DT-diaphorase to form the relatively non-toxic dopamine o-hydroquinone, which is then converted into O-methylated or sulfated products which are excreted. The second route is by the enzyme NADPH cytochrome P450 reductase to form the highly toxic free radical dopamine o-semiquinone. Both the hydroquinone and the semiquinone are then converted to 5,6-dihydroxyindole, which polymerizes to form neuromelanin. Thus, in an environment with sufficient ambient antioxidants dopamine could act as an antioxidant by reducing an ROS and being converted into dopamine o-quinone in the process. The ambient antioxidants then convert the dopamine o-quinone back to dopamine, thus providing the necessary redox recycling mechanism. Support for this hypothesis is provided by the observation that dopamine oxidizes to its o-quinone more rapidly, and the o-quinone cyclizes more slowly to its aminochrome, than do other catecholamines (Hawley et al. 1967; Graham, 1979). The production of the toxic o-semiquinone and additional ROS production is thus avoided. 5-cysteinylization and 5-glutathionylization may form the major routes of metabolism of dopamine quinone and ring closure, leading to the production of toxic o-semiquinones and further ROS terminating in neuromelanin formation, may occur only when supplies of cysteine and glutathione are exhausted (Carstam et al. 1991; Odh et al. 1994; Cheng et al. 1996). This redox cycling mechanism between dopamine and dopamine quinone would provide for modulation of the redox state of the glutamate synapse resulting in spine growth or deletion in synchrony with the reinforcement status of incoming stimuli as signaled by volume dopamine release. The other synaptic antioxidants acting by themselves could not respond in this manner.
It might also be argued that the dopamine content of the cortex is too low to support the purely chemical reactions suggested by this hypothesis. However the overall level is not what is important. What is important is the level at each glutamate synapse. Since many of these have a dopamine bouton-en-passage immediately adjacent (Kötter 1994), I would argue that this might provide the quantity of dopamine required. However, in a low antioxidant environment, dopamine can auto-oxidize, either spontaneously or under the influence of peroxinitrite, or other oxidant including free iron, to form neurotoxic o-quinones. If the level of dopamine released is insufficient to maintain a reductive redox potential in the glutamate synapse, yet sufficient to yield a toxic level of DAQs upon oxidation, then this mechanism may contribute to spine pruning.
This system can be further investigated by several routes:
(i) Urinary, CSF and brain tissue levels of stable intermediates in neuromelanin formation from the three catecholamines, such as the 5,6-dihydroxy indoles and their O-methylated products, can be examined in various situations.
(ii) The present work on 5-cysteinyl dopamine needs to be extended and complemented by work on the 5-cysteinyl derivatives of noradrenaline and adrenaline and on 5-glutathionyl derivatives of all three catecholamines and in additional brain areas besides the striatum. Furthermore 5-cysteinyl adrenaline should be looked for in the C1-C3 medullary group.
(iii) The status of neuromelanin in schizophrenia has been the subject of some preliminary investigations which showed abnormalities in some cases. This needs to be repeated with more subjects and extended to include the C1-C3 group of neuromelanin-containing neurons.
(iv) The claim made by Gai et al (1993) that the neuromelanin-containing cells in the medulla do not contain PNMT, and are thus not adrenergic, needs to be confirmed. It should be noted that the material they used came from severe cases of Parkinson’s disease who had lost up to 50% of their C1 and C3 pigmented neurons. This work needs to be repeated using material from normal subjects.
(v) The question as to the mode of psychotomimetic action of adrenochrome needs to be addressed. Clinically its effects appear to be more like those of cannabis than LSD and it has been classified as an ‘imagery-producing psychodysleptic’ (Smythies, 1999). Similar studies need to be carried out on the o-quinone derivatives of dopamine and norepinephrine. Since human work is hardly possible these days, animal models need to be employed. Adrenochrome is easier to work with as it is a more stable molecule. Pharmacological studies on its effects on serotonin, cannabinoid, NMDA and other relevant receptors might produce some interesting
B. WHAT IS THE ROLE OF CA-0-QUINONES IN DISEASE?(1) Parkinson’s disease.
Since the disease is due to the destruction of the neuromelanin-containing neurons of the SNpc and locus coeruleus (as well as the C1-C3 medullary group) CAQs may well be involved (Hirsch 1994). Neuromelanin is normally neuroprotective as it chelates toxic heavy metals and has antioxidant properties (Korytowski et al. 1995). But in excess it becomes neurotoxic (Enochs et al. 1994). Many workers have linked PD to excess heavy metals in the environment. (e.g. Enochs et al. 1994; Goldsmith et al. 1990). The substantia nigra in PD is the site of increased oxidative stress, increased pro-oxidant iron levels and reduced antioxidant defenses (Castellani et al, 1996; Drukarch et al, 1997; Toffa et al. 1997; Pearce 1997; Zhang and Dryhurst, 1994)). MAO activity is increased (Drukhardt et al. 1997) as are levels of pro-oxidant NO radicals (Nishibayashi et al. 1996) in the SNpc in PD.
The excess iron accumulates within the degenerating NM-containing neurons (Lereugle et al. 1996). Much of this is bound in the ferric form to the iron-binding protein lactotransferrin (Lereugle et al. 1996). Serum transferrin receptor concentration strongly correlates with mortality from in PD cases but not in controls (Marder et al. 1998). The Lewy bodies have increased activity of haem oxygenase which produces free iron (Schipper et al. 1998). In this connection it is of interest that Muthane et al. (1998) have reported from a study in India that Hindu’s have 40% lower levels of melanized neurons in the SNpc that do Europeans as well as a very low incidence of PD, whereas they report that Parsees in India have an incidence of PD equivalent to the high level in industrialized countries. Poor rural Muslims also have a low incidence of PD (Razdan et al, 1994). Hindus and poor rural Muslims have an almost exclusively vegetarian diet, whereas the wealthier Parsees eat meat. But perhaps more importantly Indians eat large amounts of curry spices that strongly inhibit iron adsorption (Fleming et al. 1998). Curiously vitamin C is a potent enhancer of iron adsorption (Fleming et al. 1998; Cook, 1998). This suggests that PD cases should eat plenty of curry but avoid vitamin C supplements. Mice given iron overload developed low levels of GSH, high GSSG and high levels of hydroxyl radicals but not increased levels of lipid peroxidation (Lan & Jiang. 1997). However, these mice were much more susceptible than normal to the neurotoxin MPTP. These authors concluded that high dietary levels of iron induce oxidative stress but not neuronal damage, which needs a second triggering factor. One such factor may be low activity of COMT (Kunugi, 1997): COMT protects against the formation of toxic catecholamine o-semiquinones by O-methylation of the hydroquinone.
The question of whether free iron occurs in the brain under normal circumstances is controversial (Koppenol 1998; Mumby et al. 1998) although it may be relevant to disease states. Free iron is a potent converter of catecholamines to CAQs. Furthermore, CAQs release free iron from ferritin (Tanaka, 1997). There is also evidence in PD of an increase in the oxidative pathway of DA metabolism— DAQ synthesis is raised in the SN (Mattammal et al. 1995) and the 5-cysteinyldopamine/HVA ratio is raised in CSF (Cheng et al. 1996) and the 5-cysteinyldopamine/DOPAC ratio is raised in the severely degenerated SNpc in PD (Fornstedt et al. 1989). Furthermore, although GSH levels in the SNpc are low there is no concomitant increase in GSSG levels (Shen et al. 1996). This suggested that the low GSH level is not due to its consumption as an antioxidant but to a high level of complex formation with excess levels of DA and NE o-quinones. Lewy bodies consist of neurofilaments thought to be cross-linked by DA o-quinones (Montine et al. 1995). In conclusion there is good evidence that CAQs (together with increased iron levels, increased oxidative stress and reduced antioxidant defenses) may be closely involved in the pathobiochemistry of PD but much more work is needed to determine the exact mechanism involved.
I have recently covered this topic extensively elsewhere (Smythies, 1997). Briefly the evidence that CAQs may be involved in schizophrenia is:—
(i) Adrenochrome is a psychotomimetic agent with effects more similar to those produced by cannabis than LSD (Hoffer et al, 1954; Schwartz et al. 1996a,b; Jantz and Taubman, 1998; Grof, 1963). No such tests have been carried out with dopaminochrome or noradrenochrome. Adrenochrome may occur in the C1-C3 adrenergic neurons in the medulla that project to key midline limbic thalamic nuclei. However more work is necessary to determine this.
(ii) Schizophrenics have increased brain levels of 5-cysteinyl dopamine.
(iii) Schizophrenics are weak in the enzymatic mechanisms that protect against CA o-semiquinone formation i.e. transmethylation reactions (weak MAT and SHMT) and antioxidant defenses.
(iv) There are preliminary reports that neuromelanin itself is abnormal in some cases of schizophrenia.
| Discussion Board | Previous Page | Your Symposium |
|Smythies, J; (1998). The Role of Catecholamine O-quinones in Health and Disease: What We Know and What We Don't Know. Presented at INABIS '98 - 5th Internet World Congress on Biomedical Sciences at McMaster University, Canada, Dec 7-16th. Invited Symposium. Available at URL http://www.mcmaster.ca/inabis98/kostrzewa/smythies0199/index.html|
|© 1998 Author(s) Hold Copyright|