Stereospecificity of the gliotoxic and anti-neurotoxic actions of alpha-aminoadipate

Neuroscience Letters, 19 (1980) 277-282 277 © Elsevier/North-Holland Scientific Publishers Ltd. STEREOSPECIFICITY OF T H E GLIOTOXIC AND ANTI-NEUROTOXIC ACTIONS OF ALPHA-AMINOADIPATE JOHN W. OLNEY*, TAISIJA de GUBAREFF and JAMES F. COLLINS Department of Psychiatry, Washington University School of Medicine, 4940 Audubon Ave., St. Louis, MO 63110 (U.S.A.) and (J.F.C.) Department of Chemistry, Sir John Cass School of Science and Technology, City of London Polytechnic, 31 Jewry Street, London, EC3N 3EY (U.K.) (Received May 23rd, 1980) (Revised version received June 26th, 1980) (Accepted July 2nd, 1980) SUMMARY The glutamate (Glu) analog, DL-alpha-aminoadipate (DL-czAA), and the separate D and L isomers of czAA, were administered subcutaneously to infant mice and histopathological effects on the arcuate hypothalamic (AH) nucleus were studied. L- ¢xAA induced striking gliotoxic and neurotoxic changes; D-aAA and DL-c~AA respectively induced mild and extreme gliotoxic but not neurotoxic changes. The neurotoxicity of L-czAA is of interest in view of its known neuroexcitatory potential. The non-neurotoxicity of DL-otA.A implies effective antagonism by o-czAA of the neurotoxicity of L-c~AA, which is of interest in that D-ceAA is recognized as an effective antagonist of amino acid excitants and is thought to block specifically at the excitatory receptor. Aspartic (Asp), glutamic (Glu) and cz-aminoadipic (aAA) acids are structurally analogous dicarboxylic amino acids which differ only in the length of their carbon chains. Curtis and Watkins [2, 3] have established that Asp, Glu and several structural analogs excite neuronal firing when administered by microelectrophoresis onto the surfaces of central neurons but DL-o/AA was considered inactive as a neuronal excitant. In a previous study [10], we explored the molecular specificity of the neurotoxic activities of Glu (its neuron-necrosing action in the retina and arcuate nucleus of the hypothalamus following subcutaneous administration) and found 278 that analogs with excitatory activity also exhibited Glu-type neurotoxic activity. Since the same order of potencies was demonstrable for the excitatory and neurotoxic activities of these compounds, we postulated that an excitatory mechanism underlies their neurotoxic activities and proposed referring to these agents as 'excitotoxins' (reviewed in ref. 7). We included DL-c~AA in our original tests [10] and observed that this compound induced distinctive cytotoxic changes in glial and ependymal cells in the same regions of the central nervous system where systemically administered Glu destroys neurons, but Dk-ceAA did not induce detectable toxic changes in neurons. Several laboratories [1, 5, 11] have demonstrated that c~AA antagonizes the depolarizing action of excitatory amino acids, the most specific antagonism being demonstrable when the D-isomer of c~AA was tested against the D-isomer of the potent Asp analog, N-methylaspartate (NMA) [11]. Recently, we reported the related finding that DL-oeAA, in doses lower than those required to induce a gliotoxic effect, blocks the in vivo neurotoxic action of NMA and Glu on arcuate hypothalamic (AH) neurons [9]. The neurotoxicity of these agents was not totally abolished, however, a fact that we suggested might relate to the observation by McLennan and Hall [5] that the L-isomer of o~AA is a weak neuroexcitant. Since only the oL-racemic mixture was available, we could not increase the dose of the antagonist (D-oeAA) without simultaneously introducing equal parts of an unwanted excitotoxin (L-ceAA). Recently, we synthesized the separate isomers of (~AA to permit the present in vivo experiments aimed at clarifying the stereospecificity of the cytotoxic and anti-cytotoxic activities of D~.-oeAA. DL-oeAA was obtained from Sigma (St. Louis, MO) and the separate D and ~ isomers were synthesized by one of us (J.F.C.) as follows: DL-o~AA (8.05 g) was dissolved in a warm solution of L-lysine (7.3 g) in water (15 ml) which was filtered and the filtrate diluted with hot methanol (20 ml). At room temperature, diethyl ether was added until the mixture became cloudy; it was then cooled to 4°C, allowed to stand, and the first crop of crystals removed. Successive crops of crystals were obtained by addition of more ether. The initial crop (c~D = +24 °) (6 N HC1) was predominately L-lysine 279 Fig. 1. a -d : light micrographic views of the arcuate hypothalamic nucleus (AH) of the 4-day-old mouse 4 h after subcutaneous administrat ion of 1 mg / g D-~AA (a), 2 mg /g D-c~AA (b), 2 mg /g L-c~AA (c) or 2 mg /g DL-~AA (d). The D-isomer at 1 m g / g (a) causes neither glial nor neuronal toxicity, but at double this dose (b) begins to induce detectable swelling of glial processes and of the ependymal cells lining the ventricle. The L-isomer at 1 m g / g (not shown) induces a moderate degree of glial and ependymal edema which is accompanied by necrosis of AH neurons. At 2 mg / g (c) both of these toxic manifestat ions of the L-isomer are more extreme. Note the bulls-eye profiles (dark pyknotic nuclei surrounded by clear halos of edematous cytoplasm) of acutely necrotic AH neurons (arrows). Edematous swelling affecting a portion of the AH neurons plus all o f the ependyma and glia gives an extremely rarified appearance to the AH region. In d the neurotoxic component of the reaction is absent and the glio-ependymal toxicity is maximized. Thus, the net effect of administering the D-isomer together with the L-isomer is additive with respect to gliotoxicity and subtractive with respect to neurotoxicity. 280 with an aqueous solution of NaCI at a dose equimolar to 2 mg/g c~AA. All animals were sacrificed by aldehyde perfusion fixation (1°70 paraformaldehyde and 1.5% glutaraldehyde in 0.1 M phosphate buffer) 4 h after treatment for combined light and electron microscopic examination of the hypothalamus by methods described previously [6]. No histopathological changes were detected in the hypothalami of NaCl-injected controls. The D-isomer of c~AA was without discernible effect at 1 mg /g and induced mild to moderate swelling of glial processes at 2 mg/g (Fig. la, b). Neurons remained entirely unaffected in the AH region of D-c~AA-treated mice. The L-isomer induced more dramatic gliotoxic changes and these were accompanied by an unequivocal display of neurotoxic changes as well. Both the gliotoxic and neurotoxic activities of the L-isomer were clearly evident at 1 mg/g and substantially more severe at 2 mg/g (Fig. lc). The degenerating neurons, examined by both light and electron microscopy (Fig. 2) had an identical appearance to those that have been described in the hypothalamus of Glu-treated animals in a number of our previous publications [6, 7, 10]. Changes in glia consisted of massive edematous swelling of the entire cell, including the nucleus, and conspicuous mitochondrial changes which have been described previously [10]. The De-isomer, at 2 mg/g , induced a reaction that clearly involved glial and ependymal elements with apparent total sparing of neurons (Fig. ld). Whereas glutamate and other excitotoxic amino acids typically induce marked swelling of post-synaptic dendritic and somal components of the neuron [6, 7, 10], evidence for this type of cytotoxic degeneration was not observed in the AH of animals treated with DL-o~AA. Ultrastructural evaluation of these hypothalami revealed that the Fig. 2. An electron micrograph depicting the toxic action of L-~AA (1 mg/g) on both neurons and glia. The necrotic neuron (NN) and swollen edematous glial cell (G) in the same field with a normal appearing neuron (N) is characteristic of the 'mixed cytopathology' induced by L-aAA in immature mouse AH. x 5000. 281 numerous massively swollen profiles were restricted to a non-neuronal compartment which received no synaptic contacts and which, by other criteria (appearance of the nucleus, mitochondria and the presence of fibrillary material in the cytoplasm), permitted identification of these elements as glia. Cells lining the ventricular wall, which were easily identified by their location as ependymal cells, were massively swollen. In addition, the numerous normal appearing neuronal cell bodies in the arcuate region attested to the non-neurotoxic character of the pathological process. The present findings corroborate our prior observation [10] that DL-aAA has gliotoxic but not neurotoxic activity. The new finding that L-o~AA, which was recently described as a neuronal excitant [5], has significant neurotoxicity adds new support to the excitotoxic hypothesis (that an excitatory mechanism underlies the neurotoxicity of amino acid excitants), More importantly, the demonstration that L- ~AA has significant neurotoxicity but D-c~AA has not, coupled with the observation that the DL racemic mixture is devoid of neurotoxic activity, warrants the inference that D-aAA is an effective antagonist of the neurotoxicity of the L-isomer. Blockade of the neurotoxic action by a molecule (D-~AA) known to antagonize the depolarizing action of amino acid excitants [1, 5, 11], an action which is thought to be mediated through excitatory receptors on the dendrosomal surfaces of the neuron [2, 3, 5], strongly implicates these receptors and an excitatory (depolarization) mechanism in the neurotoxic activities of the amino acid excitants. Recently we demonstrated that DL-oLAA effectively suppresses the neurotoxic action of either NMA or Glu on arcuate hypothalamic neurons [9]. This and the present findings suggest that the D-isomer of c~AA may be a relatively potent antagonist of the excitotoxic activity of these agents since the D-component of the racemic mixture totally abolishes the excitotoxic activity of its L counterpart while also counteracting a substantial amount of the excitotoxic activity of the additional excitant (either Glu or NMA) which we were administering [9]. The striking toxic effects of L-aAA on glial and ependymal elements and the relative impotence of D-aAA in producing such changes, makes it clear that the L- isomer is responsible for much of the glial-ependymal toxic activity seen in experiments with the DL racemic mixture. The fact that the D-isomer blocked the neurotoxic but not the gliotoxic manifestations of the L-isomer and, in fact, contributed modestly to the severity of the gliotoxic reaction, signifies that these are two separate and distinct toxic actions. Determining the nature of the gliotoxic mechanism and clarifying why it is relatively stereospecific for the L-isomer is a challenge for future research. Since t~-c~AA exhibited weak gliotoxicity in these experiments, one might ask whether there is any relationship between the action of this isomer on glia and its ability to antagonize amino acid-induced excitations. The answer must necessarily be no, since L-~AA which is an excitatory agonist rather than antagonist, is more potent than D-c~AA as a gliotoxin. If the gliotoxicity were linked with antagonism of excitatory activity, it would have to be the L-isomer that is the more effective antagonist. 282 It has been postulated [4, 8] that excitotoxins found naturally in brain (e.g. glutamate and aspartate) may play a role in certain human neurodegenerative disorders in that excessive stimulation of excitatory synaptic receptors by these agents might lead to excitotoxic degeneration of the post-synaptic neuron. If this postulation is correct, it follows that an agent capable of antagonizing excitatory activity at these receptors might provide a prophylactic approach to the management of such conditions. It is not likely that D-e~AA would be useful for this purpose because, like other straight chain dicarboxylic amino acids resembling Glu and Asp in structure, it probably does not pass from blood into brain except in circumventricular organ regions of brain. Our recent demonstration that DL-aAA antagonizes the neurotoxic activity of either NMA or Glu [9], plus the present evidence suggesting that this blocking action is attributable to the D-isomer, does serve as evidence, however, that Glu neurotoxicity can be effectively antagonized - apparently at the excitotoxic receptor - and that this principle, when the ideal molecule that distributes favorably across blood-brain barriers is found, could conceivably be of value in human pharmacoprophylaxis. A C K N O W L E D G E M E N T S Supported in part by U.S. Public Health Service Grants NS-09156, DA-00259, NIMH Research Career Scientist Award MH-38894 (.I.W.O.) and grants from the Huntington's Chorea and Wills Foundations. REFERENCES 1 Biscoe, T.J . , Davies, J., Dray, A., Evans, R.H., Francis, A.A. , Martin, M.R. and Watkins, J .C., Depression of synaptic excitation and of amino acid-induced excitatory responses of spinal neurons by D-c~-aminoadipate, c~,~-diaminopimelic acid and HA-966, Europ. J. Pharmacol . , 45 (19771 315-316. 2 Curtis, D.R. and Watkins, ] .C. , The excitation and depression of spinal neurones by structurally related amino acids, J. Neurochem., 6 (19601 117 141. 3 Curtis, D.R, and Watkins, J .C. , Acidic amino acids with strong excitatory actions on mammal ian neurons, 3. Physiol. (Lond.), 116 (19631 1 14. 4 McGeer, E.G. and McGeer, P.L., Duplication of biochemical changes of Hunt ington ' s chorea by intrastriatal injections of glutamic and kainic acids, Nature (Lond.), 263 (19761 517 518. 5 McLennan, H. and Hall, J .G., The action of D-¢~-aminoadipate on excitatory amino acid receptors of rat thalamic neurons, Brain Res., 149 (19781 541-545. 6 Olney, .I.W., Glutamate-induced neuronal necrosis in the infant mouse hypothalamus: an electron microscopic study, J. Neuropath. exp. Neurol., 30 (19711 75 90. 7 Olney, J .W., Neurotoxicity of excitatory amino acids. In E. McGeer, J. Olney and P. McGeer (Eds.I, Kainic Acid As a Tool in Neurobiology, Raven Press, New York, 1978, pp. 95-121. 8 Olney, J .W., Excitotoxic amino acids and Hunt ington ' s disease. In T.N. Chase, A. Wexler and A. Barbeau (Eds.), Advances in Neurology, Hunt ington 's Chorea, Raven Press, New York, 1979, pp. 609 -624. 9 Olney, 3.W., de Gubareff , T. and Labruyere, .I., ~-Aminoadipate blocks the neurotoxic action of N+ methylaspartate, Life Sci., 5 (1979) 537-540. 10 Olney, .I.W., Ho, O.k . and Rhee, V+, Cytotoxic effects of acidic and sulphur-containing amino acids on the infant mouse central nervous system, Exp. Brain Res., 14 (19711 61-76. 11 Watkins, ] .C. , Excitatory amino acids. In E. McGeer, 3. Olney and P. McGeer (Eds.), Kainic Acid As A Tool In Neurobiology, Raven Press, New York, 1978, pp. 37-- 70.