Glutamate is a cotransmitter in ventral midbrain dopamine neurons

Parkinsonism & Related Disorders Parkinsonism and Related Disorders 7 (2001) 261±264

www.elsevier.com/locate/parkreldis

Glutamate is a cotransmitter in ventral midbrain dopamine neurons S. Rayport a,b,* a

Departments of Psychiatry, Anatomy and Cell Biology, and Center for Neurobiology and Behavior, Columbia University, New York, NY 10032, USA b Department of Neuroscience, NYS Psychiatric Institute, New York, NY 10032, USA

Abstract Interactions between apparently separate dopaminergic and glutamatergic pathways ®gure prominently in the pathophysiology of Parkinson's Disease. So it is surprising that the ventral midbrain dopamine neurons, which give rise to the dopaminergic pathway, may themselves also be glutamatergic. We have addressed this idea in both rat and monkey brain and found that most ventral midbrain dopamine neurons exhibit glutamate immunoreactivity. We used postnatal cell culture to examine ventral midbrain dopamine neurons more closely. In vitro most dopamine neurons exhibit glutamate immunoreactivity, as well as immunoreactivity for phosphate-activated glutaminase, the enzyme principally responsible for the synthesis of neurotransmitter glutamate; inhibition of glutaminase reduces glutamate staining. In single cell microcultures, dopamine neurons make both dopaminergic and glutamatergic synaptic varicosities. Stimulation of individual dopamine neurons evokes a fast excitatory synaptic response mediated by glutamate; it also evokes dopamine release that inhibits the excitatory response via presynaptic D2 receptors. Thus, dopamine neurons appear to exert rapid synaptic actions via their glutamatergic synapses and slower modulatory actions via their dopaminergic synapses, including possibly inhibition of their own glutamatergic synapses. So, in the setting of dopamine neuron demise, there will be a loss of both dopaminergic and glutamatergic inputs to the striatum; furthermore, glutamate released by dopamine neurons may contribute to an excitotoxic cascade and the death of neighboring dopamine neurons. q 2001 Elsevier Science Ltd. All rights reserved. Keywords: Synaptic heterogeneity; Corelease; Cell culture; Glutamic acid; Immunocytochemistry; Excitotoxicity

Colocalization of neurotransmitters has been an enduring theme in neurobiology. The initial view was that classical neurotransmitters were colocalized with neuropeptides, and that the former mediated information ¯ow while the latter modulated it [1]. More recently, several studies have indicated that classical transmitters can be colocalized in spinal or peripheral neurons and this extends from recent demonstrations of corelease of two inhibitory transmitters in the case of GABA and glycine [2], to studies of corelease of transmitters with apparently opposing actions in the case of corelease of ATP and GABA [3] and of ACh and glycine [4]. In the brain, the three major CNS monoaminergic systems appear to corelease glutamate (GLU) [5±7]. Depending on its synaptic role, colocalization of GLU in ventral midbrain dopamine (DA) neurons carries signi®cant implications for the signal DA neurons convey to their striatal targets as well as for neurodegenerative disorders such as Parkinson's disease. * Columbia Ð Psychiatry/Neuroscience, 1051 Riverside Drive, NYSPI unit 62, New York, NY 10032-2695, USA. Tel.: 11-212-543-5641; fax: 11-212-504-3135. E-mail address: [email protected] (S. Rayport).

Many studies have examined DA neuron signaling. In the early studies, DA agonists were generally found to produce inhibition [8]. Subsequently, diverse modulatory actions of DA have been described in the analyses of DA actions at individual DA receptors, and in the coupling of these receptors to other membrane receptors [9] or to a number of second messenger systems [10], in turn modulating the phosphorylation state of receptors and channels, and ultimately gene expression [11]. Dopamine is released in part at synapses, in what has been termed wiring transmission, however, with signi®cant spill-over, and in part at varicosities in a paracrine mode, in what has been termed volume transmission [12±16]. Thus, DA action is probably more constrained by the distribution and coupling of postsynaptic receptors than by discrete connectivity. GLU, in contrast, is released in wiring transmission mode and acts principally on immediately adjacent postsynaptic receptors. The fact that DA acts principally via volume transmission accounts for the success of DA augmentation therapies, such as provision of l-DOPA or transplantation of DA releasing cells. However, such therapies are imperfect because they do not fully restore normal DAergic tone; they may also be

1353-8020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved. PII: S 1353-802 0(00)00068-7

262

S. Rayport / Parkinsonism and Related Disorders 7 (2001) 261±264

imperfect because they do not restore DA neuron synapses and thus other transmitters that DA neurons may release. In contrast to the inhibitory actions of DA agonists, the initial studies showed that stimulation of DA neurons produced excitation of striatal neurons [8]. Subsequently, Wilson and colleagues [17] stimulated the nigra in intact animals while recording intracellularly from striatal neurons and found that nigral stimulation evoked a large excitatory response. This was due to activation of long-loop pathways through the cortex and thalamus. However, after these pathways were transected, and time was allowed for degeneration of the cut axons, nigral stimulation still evoked a small, long-latency, putatively monosynaptic EPSP, which was attributed to DA neurons. More recent studies using extracellular recording have shown that activation of DA neurons either by stimulation of their axons in the medial forebrain bundle (MFB) [18] or by infusing NMDA onto their cell bodies [19] elicits a D1-mediated slow excitation. In addition, MFB stimulation produces fast excitatory responses in a subset of cells. While these responses could be due to antidromic activation of the target neurons, they could also be due to direct monosynaptic inputs from DA neurons. A pharmacological test of this idea has yet to be done. Support for fast monosynaptic actions of DA neurons comes also from studies in triple explant cultures of substantia nigra/striatum/cortex, in which stimulation of nigral neurons evokes striatal EPSPs [20]. Interestingly, perisynaptic clusters of mGluR5 receptors have been identi®ed recently at DA neuron synapses in the striatum [21]; these receptors could mediate fast excitation [22]. Several morphological studies support the idea that DA neurons release GLU (reviewed in Ref. [23]). Among these, the most compelling is the observation by Kaneko and colleagues [24] that DA neurons immunostain for phosphate-activated glutaminase (PAG), the enzyme responsible for the synthesis of most neurotransmitter GLU (viz. Ref. [25]). Further support came from the work of Hattori and colleagues [26], who described a population of DA neuron terminals, identi®ed by orthograde transport of tritiated leucine, that did not stain for TH and had asymmetric synaptic specializations, consistent with excitatory actions. More recently, Ingham and colleagues [27] showed that 6-OH-DA lesions of the MFB reduce the number of striatal synapses with asymmetric specializations by almost 20%, which could be due to the loss of excitatory DA neuron synapses. Finally, we have shown directly that ventral midbrain DA neurons in both the rat and primate brain are immunoreactive for GLU [7]. 1. Morphological studies in cell culture To examine GLU cotransmission more directly, we have placed postnatal rat DA neurons in culture [28]. We chose postnatal culture because the relevant brain areas are well de®ned by early postnatal life, so that appropriate

dissections can yield enriched cultures. With careful dissection, postnatal cultures can be obtained from rat ventral midbrain where the majority of neurons are DA neurons, as veri®ed by immunostaining for tyrosine hydroxylase (TH) [28,29]. The non-DA neurons are GABAergic (L. Lin and S. Rayport, unpublished observations). Similarly, striatal or nucleus accumbens cultures may be obtained where about 95% of the neurons are identi®ed as mediumspiny GABAergic neurons [30±32]. Furthermore, postnatal neurons have already undergone the in¯uence of synapsing with their expected targets [33], so that they are relatively mature as compared to DA neurons taken from embryos. We ®rst asked whether GLU immunoreactivity re¯ects neurotransmitter GLU [7]. We found that about 85% of DA neurons were GLUergic, similar to the colocalization in brain sections. Immunostaining of cultures from different brain regions with known cell populations revealed that both GLUergic and GABAergic neurons stained for GLU, so GLU immunostaining in culture revealed both GLU that is a precursor to GABA as well as neurotransmitter GLU. However, the presence of cells that were negative both for GLU and for GABA argues that the contribution of the metabolic pool of GLU is less signi®cant. Since ventral midbrain DA neurons are rarely GABAergic, we could assume that most GLU staining of DA neurons re¯ects the transmitter pool of GLU. In the intact brain, precursor GLU is less of a confound than in culture since most GABA neurons are synaptically active, and thus precursor levels do not build up [34,35]. If neurotransmitter GLU arises from the action of PAG, then inhibition of PAG should reduce GLU immunostaining. To test this, we used the suicide inhibitor of PAG 6diazo-5-oxo-l-norleucine (l-DON) and its inactive enantiomer d-DON. l-DON reduced the incidence of GLU colocalization in DA neurons and eliminated GLU staining of DA neuron processes, while cultures treated with d-DON were indistinguishable from controls. Thus, GLU staining of DA neurons appears to re¯ect neurotransmitter GLU. 2. DA neurons in single cell microcultures We then asked whether DA and GLU might be released from the same or different sites. To examine this, we placed DA neurons in single cell microcultures (for microculture methodology, see Ref. [36]). In such microcultures, all processes and varicosities of necessity derive from the single neuron in the microculture. Immunostaining for TH and for GLU revealed two partially overlapping sets of processes and varicosities, so that there were TH-only varicosities, GLUonly varicosities and varicosities that stained for both TH and GLU. The GLUergic varicosities were typically found near the cell body overlaying the proximal dendrites, while the doubly stained and TH-only varicosities tended to be more distributed towards the microculture perimeter, and were in most instances did not contact dendrites. At the ultrastructural

S. Rayport / Parkinsonism and Related Disorders 7 (2001) 261±264

level, microcultures stained just for TH (with diaminobenzidine) revealed TH staining to be quite heterogeneously distributed, with considerable variation of the intensity of staining of different processes, as well as within individual processes. Synaptic specializations seen under the electron microscope were mainly TH-negative, showed asymmetric specializations and were found near the cell body. Rare TH stained presynaptic varicosities were seen; these had synapses with symmetric specializations. Placing single neurons in microcultures favors the genesis of recurrent synapses or autapses [36], providing a single cell system to assay the transmitters a given neuron releases. We found that the majority of DA neurons in microcultures (identi®ed after recording by TH immunostaining) formed excitatory autaptic EPSPs. In some DA neurons, spikes were followed by reverberatory activity similar to the epileptiform-like activity described in single GLUergic hippocampal neurons in microcultures [37]. EPSPs were blocked removal of extracellular Ca 21, by kynurenate (in the higher concentration range), and by CNQX, indicating synaptic mediation involving AMPA receptors. In low Mg 21 saline, the later phase of the EPSP was attenuated by the NMDA antagonist APV, suggesting an NMDA receptor contribution. Of several other GLU agonist molecules that might be released, only GLU itself stimulates both NMDA and AMPA receptors, arguing that the autaptic EPSPs are mediated by release of GLU. In voltage clamp recordings, CNQX completely blocked autaptic EPSCs, revealing no DAergic component. However, application of the D2 antagonist sulpiride augmented autaptic EPSCs, arguing that the neurons did indeed release DA, which was inhibiting the EPSC. Consistent with this scenario, the D2 agonist quinpirole inhibited autaptic EPSCs. To identify the mode of DA action, we rested cells for several minutes and then compared the rested and subsequent responses. Sulpiride had no effect on the initial EPSC after the rest, arguing against mediation by ambient DA. Instead, sulpiride attenuated the decline in EPSCs with repeated stimulation, arguing for activitydependent release. Paired-pulse experiments showed that the locus of DA action was likely presynaptic. 3. Implications The present results suggest that where there is corelease in target areas, DA neurons may convey fast signals via their GLUergic synapses and slower modulatory signals via their DAergic synapses. Consistent with this, we have found recently that DA neurons make GLUergic synaptic connections with target neurons in micro-cocultures, and that these connections show D2 modulation [38]. However, the relative signals conveyed by the two sets of DA neuron synapses in the intact brain will of necessity be determined by their overlap (or lack thereof), whether the synapses impinge on the same or different neurons, and in areas of

263

overlap the microanatomic relationships. In the setting of Parkinson's disease and DA neuron demise, differential loss of the two sets of synapses may account in part for differences in the evolving motor de®cits. GLUergic synapses are inherently plastic, so differing patterns of DA neuron activity may engender changes in the ef®cacy of DA neuron GLUergic synapses. Furthermore, DA acting via D1-like receptors appears to be necessary for both induction and maintenance of the late phase of longterm potentiation [39,40] suggesting that coincident release of DA and GLU might enhance GLUergic transmission. De®cits in DA release would then secondarily reduce GLUergic transmission. However, to the extent that DA may also inhibit GLU release, the relationship will prove inherently more complex. In Parkinson's disease, excitotoxicity appears to contribute to the demise of DA neurons [41,42], which show a selective vulnerability to excitotoxicity [43]. While the dense GLUergic afferentation of the ventral midbrain Refs. [44,45] is without doubt the major source of GLU that might mediate excitotoxicity in the setting of Parkinson's disease, DA neurons when damaged would be expected to release colocalized GLU, and possibly could contribute both to their own demise as well as to the demise neighboring DA neurons. Acknowledgements Thanks to my colleagues for their contributions, and to the Burroughs Wellcome Fund, NIDA and NARSAD for support. References [1] HoÈkfelt T, Johansson O, Goldstein M. Chemical anatomy of the brain. Science 1984;225:1326±34. [2] Jonas P, Bischofberger J, Sandkuhler J. Corelease of two fast neurotransmitters at a central synapse. Science 1998;281:419±24. [3] Jo YH, Schlichter R. Synaptic corelease of ATP and GABA in cultured spinal neurons. Nature Neurosci 1999;2:241±5. [4] Tsen G, Williams B, Allaire P, Zhou YD, Ikonomov O, Kondova I, Jacob MH. Receptors with opposing functions are in postsynaptic microdomains under one presynaptic terminal. Nature Neurosci 2000;3:126±32. [5] Fung SJ, Chan JYH, Manzoni D, White SR, Lai YY, Strahlendorf HK, Zhuo H, Liu RH, Reddy VK, Barnes CD. Cotransmitter-mediated locus coeruleus action on motoneurons. Brain Res Bull 1994;35:423±32. [6] Johnson MD. Synaptic glutamate release by postnatal rat serotonergic neurons in microculture. Neuron 1994;12:433±42. [7] Sulzer D, Joyce MP, Lin L, Geldwert D, Haber SN, Hattori T, Rayport S. Dopamine neurons make glutamatergic synapses in vitro. J Neurosci 1998;18:4588±602. [8] Siggins GR. Electrophysiological role of dopamine in the striatum: excitatory or inhibitory?. In: Lipton MA, DiMascio A, Killam KF, editors. Psychopharmacology: a generation of progress, New York: Raven Press, 1978. p. 143±57. [9] Liu F, Wan Q, Pristupa ZB, Yu XM, Wang YT, Niznik HB. Direct protein±protein coupling enables cross-talk between dopamine D5 and g-aminobutyric acid A receptors. Nature 2000;403:274±80.

264

S. Rayport / Parkinsonism and Related Disorders 7 (2001) 261±264

[10] Huff RM. Signaling pathways modulated by dopamine receptors. In: Neve KA, Neve RL, editors. The dopamine receptors, Totowa, NJ: Humana Press, 1997. p. 167±92. [11] Adams MR, Ward RP, Dorsa DM. Dopamine receptor modulation of gene expression in the brain. In: Neve KA, Neve RL, editors. The dopamine receptors, Totowa, NJ: Humana Press, 1997. p. 305±42. [12] Garris PA, Ciolkowski EL, Pastore P, Wightman RM. Ef¯ux of dopamine from the synaptic cleft in the nucleus accumbens of the rat brain. J Neurosci 1994;14:6084±93. [13] Descarries L, Watkins KC, Garcia S, Bosler O, Doucet G. Dual character, asynaptic and synaptic, of the dopamine innervation in adult rat neostriatum: a quantitative autoradiographic and immunocytochemical analysis. J Comp Neurol 1996;375:167±86. [14] Greengard P, Allen PB, Nairn AC. Beyond the dopamine receptor: the DARPP-32/protein phosphatase-1 cascade. Neuron 1999;23:435±47. [15] Bunin MA, Wightman RM. Paracrine neurotransmission in the CNS: involvement of 5-HT. Trends Neurosci 1999;22:377±82. [16] Zoli M, Jansson A, Sykova E, Agnati LF, Fuxe K. Volume transmission in the CNS and its relevance for neuropsychopharmacology. Trends Pharmacol Sci 1999;20:142±50. [17] Wilson CJ, Chang HT, Kitai ST. Origins of postsynaptic potentials evoked in identi®ed rat neostriatal neurons by stimulation in substantia nigra. Exp Brain Res 1982;45:157±67. [18] Gonon F. Prolonged and extrasynaptic excitatory action of dopamine mediated by D1 receptors in the rat striatum in vivo. J Neurosci 1997;17:5972±8. [19] Gonon F, Bloch B. Kinetics and geometry of the excitatory dopaminergic transmission in the rat striatum in vivo. Adv Pharmacol 1998;42:140±4. [20] Plenz D, Kitai ST. Organotypic cortex-striatum-mesencephalon cultures: The nigrostriatal pathway. Neurosci Lett 1996;209:177±80. [21] Paquet M, Bradley SR, Conn J, Smith Y. Subcellular and subsynaptic distribution of metabotropic glutamate receptors in the monkey striatum. Soc Neurosci Abstr 1999;25:445. [22] Heuss C, Scanziani M, GaÈhwiler BH, Gerber U. G-protein-independent signaling mediated by metabotropic glutamate receptors. Nature Neurosci 1999;2:1070±7. [23] Hattori T. Conceptual history of the nigrostriatal dopamine system. Neurosci Res 1993;16:239±62. [24] Kaneko T, Akiyama H, Nagatsu I, Mizuno N. Immunohistochemical demonstration of glutaminase in catecholaminergic and serotoninergic neurons of rat brain. Brain Res 1990;507:151±4. [25] Laake JH, Takumi Y, Eidet J, Torgner IA, Roberg B, Kvamme E, Ottersen OP. Postembedding immunogold labelling reveals subcellular localization and pathway-speci®c enrichment of phosphate activated glutaminase in rat cerebellum. Neuroscience 1999;88:1137±51. [26] Hattori T, Takada M, Moriizumi T, Van der Kooy D. Single dopaminergic nigrostriatal neurons form two chemically distinct synaptic types: possible transmitter segregation within neurons. J Comp Neurol 1991;309:391±401. [27] Ingham CA, Hood SH, Taggart P, Arbuthnott GW. Plasticity of synapses in the rat neostriatum after unilateral lesion of the nigrostriatal dopaminergic pathway. J Neurosci 1998;18:4732±43. [28] Rayport S, Sulzer D, Shi W-X, Sawasdikosol S, Monaco J, Batson D, Rajendran G. Identi®ed postnatal mesolimbic dopamine neurons in

[29] [30] [31] [32]

[33] [34] [35] [36]

[37] [38] [39] [40] [41] [42] [43]

[44]

[45]

cell culture: morphology and electrophysiology. J Neurosci 1992;12:4264±80. Masuko S, Nakajima S, Nakajima Y. Dissociated high-purity dopaminergic neuron cultures from the substantia nigra and the ventral tegmental area of the postnatal rat. Neuroscience 1992;49:347±64. Shi W-X, Rayport S. GABA synapses formed in vitro by local axon collaterals of nucleus accumbens neurons. J Neurosci 1994;14:4548± 60. Shetreat ME, Lin L, Wong AC, Rayport S. Visualization of D1 dopamine receptors on living nucleus accumbens neurons and their colocalization with D2 receptors. J Neurochem 1996;66:1475±82. Wong AC, Shetreat ME, Clarke JO, Rayport S. D1- and D2-like dopamine receptors are colocalized on the presynaptic varicosities of striatal and nucleus accumbens neurons in vitro. Neuroscience 1999;89:221±33. Perrone-Capano C, Da Pozzo P, di Porzio U. Epigenetic cues in midbrain dopaminergic neuron development. Neurosci Biobehav Rev 2000;24:119±24. Ottersen OP, Storm-Mathisen J. Glutamate- and GABA-containing neurons in the mouse and rat brain, as demonstrated with a new immunocytochemical technique. J Comp Neurol 1984;229:374±92. Conti F, Rustioni A, Petrusz P, Towle AC. Glutamate-positive neurons in the somatic sensory cortex of rats and monkeys. J Neurosci 1987;7:1887±901. Segal MM, Baughman RW, Jones KA, Huettner JE. Mass cultures and microislands of neurons from postnatal rat brain. In: Banker G, Goslin K, editors. Culturing nerve cells, 2nd ed. Cambridge: MIT Press, 1998. p. 309±38. Segal MM. Epileptiform activity in microcultures containing one excitatory hippocampal neuron. J Neurophysiol 1991;65:761±70. Joyce MP, Rayport S. Mesoaccumbens dopamine neuron synapses reconstructed in vitro are glutamatergic. Neuroscience 2000;99:445±56. Otmakhova NA, Lisman JE. D1/D5 dopamine receptor activation increases the magnitude of early long-term potentiation at CA1 hippocampal synapses. J Neurosci 1996;16:7478±86. Frey U, Morris RGM. Synaptic tagging: implications for late maintenance of hippocampal long-term potentiation. Trends Neurosci 1998;21:181±8. Lange KW, Kornhuber J, Riederer P. Dopamine/glutamate interactions in Parkinson's disease. Neurosci Biobehav Rev 1997;21:393± 400. Sonsalla PK, Albers DS, Zeevalk GD. Role of glutamate in neurodegeneration of dopamine neurons in several animal models of parkinsonism. Amino Acids 1998;14:69±74. Marey-Semper I, Gelman M, LeÂvi-Strauss M. A selective toxicity toward cultured mesencephalic dopaminergic neurons is induced by the synergistic effects of energetic metabolism impairment and NMDA receptor activation. J Neurosci 1995;15:5912±8. Sesack SR, Pickel VM. Prefrontal cortical efferents in the rat synapse on unlabeled neuronal targets of catecholamine terminals in the nucleus accumbens septi and on dopamine neurons in the ventral tegmental area. J Comp Neurol 1992;320:145±60. Parent A, Parent M, Charara A. Glutamatergic inputs to midbrain dopaminergic neurons in primates. Parkinsonism Related Disord 1999;5:193±201.