The excitotoxin hypothesis: therapeutic considerations in amyotrophic lateral sclerosis Richard A. Smith, Kristen L. Vaiderhaug and Peter J. Wedlund Center for Neurologic Study, San Diego, CA, USA. School of Pharmacy, University of Kentucky, Lexington, KE, USA (PJW). Introduction The ALS landscape is littered with etiologic hypotheses. Some of these are highly speculative and until such time as they can be documented, they offer little promise for therapy. Others, such as the theories relating ALS to an excitotoxin, conceivably lead to an immediate opportunity for treatment. Whenever scientists or physicians think about applying results from the laboratory to the clinic, it is possible to elicit contradictory opinions as to the justification for going to a clinical trial. One might expect physicians to favour treatment and scientists to advise caution, but opinions do not easily sort along professional lines. The person whom this discussion is meant to benefit, namely the patient, is rarely a party to such discussion. This has led some commentators to wonder whether professionals in the form of doctors, hospitals and regulatory agencies are not overly paternalistic in their deliberations about experi- mental treatment (Feenberg, in press). As in the example of AIDS one can expect patient groups to become increasingly vocal in their support of treatment trials. However, the ambiguity which this subject musters is not confined to patients and physicians. Granting agencies often give treatment trials a very low priority, and in one instance refused to provide funding for the experimental treatment of ALS. The argument for treatment trials By way of review, what is the argument for initiating treatment trials to counter a putative environmental toxin? In this volume, evidence has been presented that the incidence of ALS has increased within the continental United States and decreased in the Pacific where the disease has been endemic following World War II (Whiting, 1988; Spencer et al., 1987). Data of this sort suggest the possibility of an increased environmental burden in some populations and a decrease in others. In Guam, Drs Whiting and Kurland and more recently Dr Spencer have implicated Cycad as the cause of ALS-parkinsonism-dementia complex. On Prince Edward Island, a neurotoxin has been traced to consumption of mussels (Teitelbaum et al., 1990). Thus, in two geographic locations specific compounds have been suggested as toxic to the nervous system. Further, there is evidence that one or more of these may be acting in the manner of an excitotoxin. These effects may be direct or occur indirectly through biotraiisforma- tion, as in the case of MPTP which is thought to injure dopaminergic neurons after oxidation by monoamine oxidase (Snyder & D'Anato, 1985). Although MPTP is not an environmental toxin, the naturally occurring and structurally related tetrahydro-@ carbolines are found in plants and have been suggested as a possible causative agent of Parkinson's disease in man (Fonne-Pfister et al., 1987). In humans, the hepatic microsomal enzyme system metabolizes many such potential toxins in order to enhance their water solubility and excretion by the body (Waring et al., 1989). One of the cytochrome P-450 enzymes (P-450dbi) involved in metabolism of toxins and capable of binding to both MPTP and tetta-beta-carbolines is found in both liver and central nervous system (Fonne-Pfister et al., 1987). Furthermore, the expression of the cytochrome P-450dbi enzyme is dependent upon genetic factors. About 6-7 percent of the Caucasian population is unable to express this particular enzyme (Guttendorf et al., in press). Thus, it n-tight be predicted that individuals lacking this enzyme might be uniquely susceptible to neurotoxins and the subsequent development of neurologic disease. To pursue this possibility, we have undertaken studies to phenotype the cytochrome P450 isoenzymes in ALS and control patients. Cytochrome P450 enzymes: phenotype studies The cytochrome P430 enzymes are believed to have evolved over millions of years as a means to detoxify naturally occurring toxins, principally produced by plants (Nelson & Strobel, 1987). Individual expression of the repertoire of cytochrome P450 enzymes is genetically determined (Funck-Brentano et al., 1989). Phenotypic variation of the cytochrome oxidase P430-dbl and P450-mp enzymes is associated with debrisoquine and mephenytoin 4'-hydroxylation, respectively (Inaba et al., 1985). To examine the expression of these enzymes subjects are challenged with a test drug and urine is collected over the ensuing 8 hours (Guttendorf et al., in press). The urine sample is then examined to characterize the ability of a particular isoetizyme of P-450 to metabolize the probe substrate. In the case of dextromethorphan, cytochrome P-450dbi is responsible for the 0-demethylation of this probe to dextrorphan (Fig. 1). The dextromethorphan and dextrorphan are then quantified in each sample by HPLC. From this quantification, a metabolic ratio (MR) is determined using the following formula: MR = moles of dextromethorphan in urine sample moles of cfextrorphan in urine sample In general, the population segregates into two groups based upon the value for the MR. Based on this subjects can be characterized as either efficient or slow metabolizers of dextromethorphan. A slow metabolizer is classified as having a metabolic ratio greater than 0.30 and these individuals are considered deficient in their expression of the cytochrome P-450dbl isozyme. In the case of mephenytoin, cytochrome P-450 mp is primarily responsible for the conversion of the S-enantiomer of this racemic drug to 4- ------------------------------------------------------------------------------------------------ Table 1. Cytochrome P450 phenotypes in ALS and control subjects. ALS Controls 1.0200 0.0038 0.0007 0.0019 0.0450 0.0017 0.0038 0.0066 0.0003 0.0092 0.0020 0.0134 0.0006 0.0018 0.0003 0.0366 0.0002 0.0049 0,0005 0.0010 0.0011 0.0033 0.0110 1.3677 0.0020 0.0058 0.0004 0.0238 0.0016 0,0017 0.0021 0.0168 0.0330 0.0011 7.8304 O@0520 0.0048 0.0045 0.0243 0.0032 0.0016 0.0114 0,0147 0.0094 0.0010 0.0023 0.0039 0.0036 0.0146 0,0028 0.0011 0.1954 6.4921 1.3365 0.0092 0.0026 0.0381 0.1502 0.0048 0.0012 7.1247 Slow metabolizers > 0.30. ---------------------------------------------------------------------------------------------------- hydroxymephenytoin. The stereospecificity in the metabolism of racemic mephenytoin and the marked difference between the rates at which the R- and S-enantiomers of mephenytoin are eliminated by the body, allow for the determination of cytochrome P- 450mp activity based solely upon the ratio of mephenytoin enantiomers in a urine sample. A slow metabolizer of mephenytoin is classified as someone with a ratio of S-/R-mephenytoin in urine of 0.90 or greater in'urine samples collected after ingestion of 100 mg of racemic mephenytoin (Guttendorf et al., in press). Among the 34 ALS patients examined so far, three have been classified as slow metabolizers of dextromethorphan (Table 1) and another three as slow metabolizers of mephenytoin. Among the 27 control subjects, three have been classified as slow metabolizers of dextromethorphan. None have been found to be slow metabolizers of mephenytoin thus far. Although these studies do not suggest that a deficiency in the cytochrome P-430 system is responsible for the susceptibility of ALS, this line of inquiry may still be fruitful in the future. Polymorphic expression of other isozymes within the cytochrome P-450 superfamily undoubtedly exist, and other biotransformation reac- tions involving toxins (eg conjugation with glucuronides) have yet to be studied in ALS patients. Another interesting twist to be considered is the possible relationship of dietary choice to phenotypic variation within the cytochrome P450 system (Britto et al., in press). Comparing food preferences in poor metabolizers of debrisoquine (n = 29) and mephenytoin (n = 18), with 134 controls, it was noticed that individuals with low levels of cytochrome P-450dbi were disinclined to eat cauliflower or coconut whereas those with diminished P-450mp activity avoided spinach and cabbage. With the majority of food stuffs there were no significant differences in preference. People's tastes for meat, dairy food, fats, sweets, starches, spices, etc. were similar in both groups. Considering the case for a dietary link with ALS it may be worthwhile to conduct similar studies in a large group of ALS patients. If the example of carcinogens is pertinent, this is likeiv to be rewarding because persons are exposed daily to a varietv of naturally occurring dietary hazards (Ames, Magaw & Bold, 1987) Therapeutic possibilities Turning to the opportunities for treatment, there seems ample justification, based on laboratory studies, to undertake a treatment trial of ALS with an NMDA blocking agent. Evidence in support of the excitotoxin hypothesis is indirect in the case of ALS, but compelling in the case'of ischemic or hypoglycemic insult (Collins, Dobkin & Choi, 1989). The mechanism for glutamate toxicity - glutamate being a prototypic excitotoxin, is well understood in general terms, although the specifics are still to be worked out (Choi, 1987). Following release of glutamate into the extracellular space, the sodium channel is activated, resulting in neuronal swelling, followed subsequently by an influx of extracellular calcium. Under conditions of ischemia, hippocampal neurons, notably in the CA, sector, are particularly vulnerable. In a seminal study, Simon and his colleagues demonstrated that hypoxic neuronal damage may be markedly reduced by the use of an NMDA antagonist in vivo (Simon et al., 1984). Similar results have been seen in vitro (Choi, Koh & Peters, 1988). Future therapy with NMDA antagonists will involve a number of considerations. Competitive NMDA antagonists do not readily cross the blood brain barrier (Albers, Gadberg & Choi, 1989). Fortunately, the noncompetitive NMDA antagonists do not have this limitation. Currently some of these could be employed in therapeutic trials: MK-801 (an anticonvulsant), dextromethorphan (a cough suppresant) and ketamine (an anesthetic). In the immediate future, dextromethorphan would seem to be the most likely treatment candidate since it is available in the market place, administered orally and well tolerated. Dextromethorphan (DM) is an over the counter antitussive analog of morphine. The drug has been shown to bind to both NMDA and sigma receptors (Mussachio. et al., 1989). The latter are particularly enriched in motor areas. Following absorption, the drug is oxidized by the cytochrome oxidase system to dextrorphan and conjugated to glucuronide (see Fig. 1). Both dextromethorphan and dextrorphan are potent NMDA antagonists, but once conjugated to glucuronide, dextrorphan would not be expected to cross the blood brain barrier. In'rats, there is considerable variability in brain plasma ratios for dextromethorphan (Walker & Hunt, 1989), brain levels ranging from 7-25 times that of plasma. There does not appear to be a good correlation between CSF and brain levels. As noted earlier, the degradation of dextromethorphan (DM) exhibits polymorphic features. In a study of 519 Americans, approximately 94.5 percent of the population were seen to be fast DM metabolizers and 6.7 percent slow metabolizers (Guttendorf et al., in press). This distribution may vary in different ethnic groups. Following the administration of 60 mg of DM to subjects who were slow or intermediate metabolizers, steady state concentrations of DM were reached in approximately four days, achieving levels of approximately 30 ng/ml in intermediate metabolizers and 300 ng/ml in slow metabolizers. In in-vitro studies, dextromethorphan exhibits NMDA blockade at levels as low as 250 ng/ml, although protection is not complete at this level (Choi et al., 1987). Intravenous Or intraperitoneal administration of DM has been protective at doses ranging from 10-35 mg/kg in animal models of hypoxic injury (Prince & Feeser, 1988). These modes of administration are impractical for longterm human trials, but they give hope that therapeutic levels of DM can be achieved in a treatment trial. Unfortunately, it is not known whether this is the case. For the majority of ALS patients, treatment with oral dextromethorphan would have to involve large doses of medication or concomitant administration of drugs which block the degradation of DM. Walker has reported administration of approximately one gram of DM by mouth to a small group of patients with Huntington's disease (Walker & Hunt, 1989). Out of a group of eleven patients, seven reported side effects including rash, clumsiness, dysarthria, drowsiness and increased rigidity. Several patients achieved serum levels above 250 ng/ml. In approximately half the patients, levels hovered at 50 ng/ml, in spite of being on doses of DM ranging from 200-960 mg/day. The median peak tolerated dose in this study was approximately 400 mg per day. Reflecting on this data, one could surmise that the patients who achieved high serum levels of DM were intermediate or slow metabolizers. At the minimum, this study reinforces the need to phenotype the cytochrome P-430 enzyme in experimental subjects and to monitor serum levels. This can not be overstated in the instance of therapy which is associated with individual variation in the metabolic handling of the test drug. In the case of treatment with DM, serum levels might vary 10-250 fold from one patient to another. Since blood levels would be expected to be minimal in the majority of patients (90 percent or more) - even when high dose therapy is employed, it may be useful to block the degradation of DM with a second drug. Using human liver microsomes a diverse group of compounds have been screened for ability to block the cytochrome oxidase system. Inhibition was demonstrated for a variety of classes of drugs (Table 2). -------------------------------------------------------------------------------------------------------- Table 2. Classes of drugs which inhibit cytochrome P450. 49 Anticonvulsants 0 Hypnotics and sedatives 0 Tricyclic antidepressants 0 Monoamine oxidase inhibitors 0 Antipsychotics 0 O-Adrenoceptorblocking drugs *Cardiovascular drugs 4D Antimalarials * Antihistamines 0 Alkaloids and analgesics -------------------------------------------------------------------------------------------------------- Among these, quinidine was the most potent, but it is not a substrate for the enzyme (Inaba et al., 1985). At steady state plasma levels of 2-7 1AM, quinidine almost totally blocks the oxidation of sparteine or debrisoquine in human subjects, suggesting that it may be an ideal drug to boost DM levels in an ALS treatment trial (Broden -et al., 1987). It is assumed that the side effects of NMDA blockade are reversible, but a recent report cautions that this may not be the case (Olney, Labruyere & Price, 1989). Following administration of MK 801 and ketamine to rats vacuolar degeneration was noted in the cingulate and retrospienial cortices. Both of these drugs, like DM, are NMDA antagonists. There are other reservations to initiating treatment of.ALS with NMDA antagonists. While it has been established that NMDA receptors decorate hippocampal neurons and presumably are critical to memory registration, analogous data are not well established for the motor system (Greenamyre, Young & Penney, 1984). Although there is ample evidence to incriminate excitotoxins as a mechanism for cell death, it remains to implicate them in ALS and it remains to be determined whether NMDA blockade can achieve therapeutic benefits without causing unacceptable consequences. Conclusions Returning to the general question of the appropriateness of treating ALS patients experimentally when the opportunity arises, it is worth recalling that there is always an element of the unknown. Further, experimental treatment involves risks, some of which can not be anticipated even when extensive background studies have been undertaken. However, the strength of well thought out treatment programs is that they provide the chance to test etiologic hypotheses. Until such a time as we know more about ALS, this may be one of the most rewarding means to advance knowledge of a poorly understood disease. Acknowledgement - This manuscript is dedicated to William and Arlene Doro, Arlene and George Hecht, John Stockton, and George and Joan Thagard. References Albers GW, Goldberg MP & Choi DW (1989): N-methyl-D-aspartate antagonists: Ready for clinical trial in brain ischemia? An?i. Netirol. 25, 398-403. Ames BN, Magaw R & Gold LS (1987): Ranking possible carcinogenic hazards. Scietice 236, 271- 280. Britto MR, McKean HE, Bruckner GG & Wedlund Pj (In press): Polymoiphisms in oxidative drug metabolism: Relationship to food preference. Br. f. Clitt. Pliar?nac. Brosen K, Gram LF, Haghfelt T & Bertilsson L (1987): Extensive metabolizers of debrisoquine become poor metabolizers during quinidine treatment. Pitartnac. Toxicol. 60, 312-314. Choi DW (1987): Ionic dependence of glutamate neurotoxicity. 1. Nezirosci. 7, 369-379. Choi D, Koh J & Peters S (1988): Pharmacology of giutamate neurotoxicity in cortical ceil culture: Attentuation by NMDA antagonists. 1. Netirosci. 8, 185-196. Choi DW, Peters S & Viseskul V (1987): Dextrorphan and levorphanoi selectively block N-methyl-D-asparate receptor-mediated neurotoxicity on cortical neurons. 1. Pharmac. Exp. Tlzer. 242, 713-720. Collins RC, Dobkin BH, Choi DW (1989): Selective vulnerability of the brain: New insights into the pathophysiology of stroke. Atiti. I)it. Med, 10, 992-1000. Feenberg A (In press): Experimental Treatment of Incurable Disease: Resolving the Ethical Dilemma. Handbook of Amyotrophic Lateral Sclerosis, editor R Smith, Marcel-Dekker, N.Y. Fonne-Pfister R, Bargetzi Mj & Meyer A (1987): MPTP, the neurotoxin inducing Parkinson's disease is a potent competitive inhibitor of human and rat cytochrome P450 isozymes (P450bufl, P45;dbi) catalyzing debrisoquine 4-Hydroxylation. Biocitem. Res. Communic. 148, 1144-1130. Funck-Brentano C, Turgeon J, Woosley RL & Roden DM (1989): Effect of low dose quinidine on encainide pharmocokinetics and pharmocodynamics: Influence of genetic polymorphism. Pliarmac. Exp. Tlter. 249, 134-142. Greenamyre, JT, Young, AB & Penney, JB (1984): Quantitative autoradiographic distribution of L-[3 HI glutamate-binding sites in rat central nervous system. 1. Neitrosci. 4, 2133-2144. Guttendorf Rj, Britto M, Wedlund Pj, Blouin RA, Fo'ster TS, John W & Pittman KA (In press): Rapid screening for polymorphisms in dextromethorphan and mephenytoin metabolism. Br. 1. Cliti. Pitartnac. Inaba T, Juxima M, Mahon WA & Kalow W (1985): In vitro studies of two isozymes of human liver cytochrome P-450: rvlephenytoin p-hydroxylase and sparteine monooxygenase. Drug Metab. Disp. 13, 443-,448. Mussachio JM, Klein M & Paturzo jj (1989): Effects of dextromethorphan site ligands and allosteric modifiers on the binding of (+) [HI-3-(-3-hydroxyphenyl)-N-(I-propyl)piperidine. Molec. Pliarmac. 35, 1-5. Nelson DR & Strobel HW (1987): Evolution of cytochrome P-450 proteins. Mol. Biol. Evol. 4, 372- 593. Olney JW, Labruyere J & Price MT (1989): Pathologic changes induced in cerebrocortical neurons by phencyclidine and related drugs. Scieiice 244, 1360-1362. Prince D & Feeser HR (1988): Dextromethorphan protects against cerebral infarction in a rat model of hypoxia-ischemia. Neitrosci. Lett. 85, 291-296. Simon RP, Swan JH, Griffiths R & Meldrum BS (1984): Blockade of N-methyl-D-aspartate receptors may protect against ischemic damage in the brain. Scieiice 226, 85@52. Spencer PS, Nunn PB, Hugon J, Ludolph AC, Ross SM, Roy DN & Robertson RC (1987): Guam amyotrophic lateral sclerosis-parkinsonism-dementia linked to a plant excitant neurotoxin. Sciepice 237, 517-57"7. Snyder SH & D'Amato Rj (1983): Predicting Parkinson's disease. Nature 317, 198-199. Teitelbaum JS, Zatorre Rj, Carpenter S, Gendron D, Evan AC, Giedde A & Cashman N (1990): Neurologic sequelae of domoic acid intoxication due to ingestion of muscles from Prince Edward Island. Nezu. Ettgl. 1. Med. 322, 1781-1787. Walker FO & Hunt VP (1989): An open label trial of dextromethorphan in Huntington's Disease. Clin. Netiropiiarn7ac, 12, 322-330. Waring RH, Sturman SW, Smith MCG, Steventon GB, Heafield MTE & Williams AC (1989): S- methylation in motoneuron disease and Parkinson's disease. La?tcet 2, 3@357. Whiting MG (1988): Toxicity of cvcads: implications for neurodegenerative diseases and cancer. Transcripts of Four Cycad Conferences. First Cycad Conference p. @, 1962. New York, Third World Medical Research Foundation.