[1] 6w Striatal plasticity and extrapyramidal motor dysfunction

[1] 160w Knowledge of molecular events contributing to motor dysfunction in Parkinson's disease has advanced rapidly during the past decade. Studies in animal models as well as in patients afflicted by this disorder suggest that the nonphysiologic stimulation of striatal dopamine receptors, first as a result of dopaminergic denervation and later as a consequence of the intermittent high-intensity stimulation produced by standard therapeutic regimens, leads to plastic changes in striatal medium spiny neurons. The clinical appearance of parkinsonism and subsequently of motor response complications is associated with the aberrant activation of signaling cascades within medium spiny neurons that modify the phosphorylation state of their ionotropic glutamatergic receptors. Resultant NMDA and AMPA receptor sensitization augments cortical excitatory input to these spiny efferent neurons, thus altering striatal output in ways that compromise motor function. These findings have already yielded new insight into mechanisms subserving motor memory and synaptic integration as well as accelerated development of novel approaches to the improved treatment of motor disability.

[1] 148w Plastic neuronal responses to various patterns and intensities of receptor stimulation have been extensively studied in various brain regions, especially in relation to long term potentiation and long term depression in the hippocampus [1]. More recent expansions of this research to the striatum have begun to show how activity patterns of mesencephalic dopaminergic systems can critically affect the regulation of synaptic function at certain stages of learning and memory [2 -4]. In this regard, particular attention has focused on how dopamine (DA) receptor mediated mechanisms modulate the sensitivity of ionotropic glutamatergic receptors on striatal medium-sized spiny neurons [5,6]. Since dopaminergic receptors are located in close proximity to N-methyl-D-aspartate (NMDA) and alpha-amino-3-hydroxy-5-methyl-4-isoxazole proprionic acid (AMPA) receptors on the dendritic spines of these neurons, these mechanisms could be crucial to understanding the persisting changes in motor behavior produced by the nonphysiologic stimulation of DA receptors occurring in Parkinson's disease (PD).

[2] 165w Iontropic glutamatergic receptors of the NMDA type function as ligand-gated ion channels, especially for calcium, while those of the AMPA type largely serve as channels for sodium and other monovalent ions [29]. The NMDA receptor complex is a tetramer, assembled from NR1 subunits and NR2 subunits, the latter consisting of four homologous isoforms of which the NR2B and less frequently the NR2A occur in striatum [30,31]. AMPA receptors are composed of tetrameric combinations of GluR1-4 subunits; all but the GluR-4 subunits appear to be expressed by medium spiny neurons [32]. The regulation of NMDA and AMPA receptor subunits occurs through protein phosphorylation, especially at sites along their intracellular carboxy tails [33,34]. Phosphorylation of tyrosine residues has been primarily linked to the modulation of channel characteristics including open state probability, while serine/threonine phosphorylation has been more closely related to the regulation of receptor anchoring to plasma [35,36]. A number of phosphorylation sites as well as protein kinases mediating phosphorylation at these sites have now been identified.

[3] 102w Alterations in striatal NMDA and AMPA receptor phosphorylation associated with dopaminergic denervation and subsequent intermittent stimulation result from the activation of intracellular signal transduction cascades linking co-expressed dopaminergic and glutamatergic receptors [20]. DA receptors are located relatively proximally along the necks of spiny neuron spines, while ionotropic glutamate receptors are situated within the postsynaptic density at the distal spine tips [37,38]. The distance between these receptors is small, thus facilitating bidirectional signaling between these and other neurotransmitter receptors [39]. The degree and pattern of NMDA and AMPA subunit phosphorylation can quickly change, reflecting the countervailing actions of relevant kinases and phosphatases [40].

[4] 157w The phosphorylation state of tyrosine residues on striatal ionotropic glutamatergic receptors undergoes characteristic alterations as a result of dopaminergic denervation and later as a consequence of intermittent dopaminomimetic drug treatment. NMDA NR2A and NR2B subunits, but not NR1 subunits, can be directly phosphorylated by src family protein tyrosine kinases including fyn [40]. In rats, the 6-OHDA induced destruction of the nigrostriatal pathway increases NR2B tyrosine phosphorylation in total striatal homogenates, without altering receptor protein expression [41,42]. For NMDA receptors on striatal membranes but not in the cytosolic compartment, however, NR2B subunit tyrosine phosphorylation and the expression of NR2B as well as NR1 protein are reduced by 6-OHDA, suggesting a redistribution of these receptors from the membrane to the cytoplasmic compartment [43,44]. Following intermittent levodopa treatment sufficient to induce motor response alterations, there is a substantial rise in tyrosine phosphorylation of both total and membrane associated NR2A and especially NR2B subunits, while the protein changes are normalized [42,43].

[5] 128w The relation between tyrosine phosphorylation of NMDA receptors and dopaminergic regulation of motor function has been explored using compounds that selectively inhibit tyrosine kinases. The unilateral injection of the protein tyrosine kinase inhibitor genestein into the striatum of rats whose ipsilateral nigrostriatal system has been interrupted by 6-OHDA produces modest antiparkinsonian effects as evidenced by the induction of contralateral rotation [42]. The intrastriatal administration of genistein to parkinsonian rats given intermittent levodopa treatment normalizes the motor response and attenuates the enhanced tyrosine phosphorylation of both the NR2A and NR2B subunits [42]. These results support the possibility that hyperphosphorylation of tyrosine residues on NMDA receptor subunits, due to kinase activation or phosphatase inhibition, contributes to the motor dysfunction induced by dopaminergic denervation as well as by subsequent dopaminomimetic treatment.

[6] 42w The phosphorylation state of tyrosine residues on AMPA receptors has received less investigative attention and striatal changes have not been reported in parkinsonian animals [32,33,45]. DA system destruction has, however, been observed to reduce AMPA GluR1 subunit expression in rat striatum [44,46].

[7] 204w Phosphorylation of serine/threonine residues also serves as a crucial regulatory mechanism for ionotropic glutamatergic receptors [2,3,30]. Multiple serine phosphorylation sites have been identified on NMDA receptor subunits: on NR1 subunits, for example, cyclic AMP-protein kinase A (PKA) can phosphorylate Ser897 and protein kinase C (PKC) phosphorylates Ser890 and Ser896 [47]; on NR2B subunits, PKC directly phosphorylates Ser1303 and Ser1323 and calcium/calmodulin-dependent protein kinase II (CaMKII) phosphorylates Ser1303 and/or the cognate site on the NR2A subunit [48,49]. Serine phosphorylation of striatal NMDA receptors is affected by dopaminergic denervation as well as by intermittent dopaminomimetic stimulation. In rats, destruction of rat nigrostriatal dopaminergic system with 6-OHDA increases NR2A subunit phosphorylation as well as NR2A protein; twice daily levodopa treatment that induces motor response changes leads to a further elevation in the serine phosphorylation of NR2A but not of NR2B subunits without associated changes in subunit protein levels [50]. More specifically, chronic intermittent D1 agonist treatment increases NR2A but decreases NR2B serine phosphorylation while chronic intermittent D2 agonist treatment has no effect on NR2A but increases NR2B phosphorylation. A 6-OHDA lesion decreases serine phosphorylation of membrane associated NR1 subunits at residues 890 and 896, while intermittent levodopa treatment increases NR1 phosphorylation at Ser890, Ser896 and Ser897 [43].

[8] 127w Serine phosphorylation also plays a crucial regulatory role for AMPA receptors. PKA, PKC and CaMKII PKC contribute to the regulation of AMPA channel function, in part, via phosphorylation of GluR1 and GluR2 subunits [51,52]. In nucleus accumbens cells, GluR1 phosphorylation at Ser845 increases with D1 DA agonist stimulation, a response that is attenuated by D2 agonist stimulation [53]. In rats, a 6-OHDA lesion of the DA system reduces AMPA GluR1 mRNA expression in striatal tissues [44], but has no affect on serine phosphorylation at the GluR1 Ser831 PKC site; levodopa treatment sufficient to induce motor response changes in parkinsonian rats, on the other hand, significantly increases Ser831 phosphorylation [54]. Similarly, striatal GluR1 Ser845 phosphorylation is unaffected by dopaminergic denervation, but rises with levodopa induced response changes [55].

[9] 285w Recent studies have begun to detail signaling pathways linking dopaminergic and glutamatergic receptors within CNS neurons [20,56]. All three of the major second messenger regulated protein kinases-PKA, PKC and CaMKII-have now been implicated in the altered serine phosphorylation of NMDA and AMPA receptors associated with the nonphysiologic stimulation of rat striatal dopaminergic receptors. PKA inhibition in medium spiny neurons by the intrastriatal injection of Rp-cAMPS has limited antiparkinsonian efficacy in 6-OHDA lesioned rats but potently reverses the motor response changes produced by intermittent levodopa therapy [57]. Since Rp-cAMPS normalizes the shortened response duration to D1 but not to D2 DA receptor agonist challenge, the participation of PKA in the pathogenesis of these response modifications may be limited to striatal spiny neurons that express functioning D1 DA receptors and project via the direct stiatonigral pathway. Inhibition of CaMKII by intrastriatally injected KN-93 has no antiparkinsonian activity, but reverses levodopa-induced response alterations [50]. In contrast to results with PKA inhibition, selective CaMKII inhibition affects D2 as well as D1 DA receptor mediated mechanisms. The acute intrastriatal injection of the CaMKII inhibitor KN93 not only normalizes the levodopa-induced motor response alterations but also attenuates the D1 and D2 receptormediated changes in serine phosphorylation of NR2A and NR2B subunits, respectively. Conversely, the protein serine phosphatase 2A inhibitor, okadaic acid, potentiates these response alterations. PKC augmentation as a consequence of the direct intrastriatal transfer of the catalytic domain of constitutively active PKC elevates Ser831 phosphorylation of GluR1 receptor subunits and hastens onset of the shortened motor response duration produced by levodopa therapy [54]. In PKC gene transferred animals, intrastriatal injection of the PKC inhibitor NPC-15437 attenuated both the increased GluR1 phosphorylation and the accelerated onset of the levodopa-induced response modifications.

[10] 94w The foregoing observations are consistent with the view that both denervation and intermittent stimulation of dopaminergic receptors on striatal spiny neurons activate internal signaling cascades capable of affecting the phosphorylation state of coexpressed ionotropic glutamatergic receptors. The nature of the signaling kinase/phosphatase alteration presumably influences the pattern of phosphorylation change and in turn the distinctive modifications in glutamate receptor function. Conceivably, resultant changes in cortical glutamatergic input to spiny neurons modify the output of these striatal efferents in ways that contribute to the appearance of parkinsonian signs as well as levodopa-associated response alterations [20].

[11] 75w If the nonphysiologic stimulation of striatal dopaminergic receptors alters the synaptic efficacy of adjacent ionotropic glutamatergic receptors as a consequence of changes in their phosphorylation state, then pharmacologic manipulation of these glutamate receptors might be expected to influence the associated motor dysfunction. Increasing evidence from studies in parkinsonian models and patients suggests that alterations in the sensitivity of NMDA and AMPA receptors do indeed occur with denervation-associated parkinsonism as well as with treatment-associated response modifications.

[12] 138w In parkinsonian rodents, NMDA receptor antagonists can potentiate the antiparkinsonian effects of dopaminomimetics and suppress levodopa-induced response alterations [58 -61]. Direct injection studies suggest that the effects of systemically administered NMDA antagonists are mediated in the striatum, presumably at receptors expressed on the dendrites of medium spiny neurons [59]. The motoric responses to both competitive and noncompetitive NMDA antagonists may reflect preferential activity at the D2 DA receptor mediated indirect striatopallidal pathway [61 -64]. In MPTP lesioned monkeys, the noncompetitive NMDA antagonist amantadine has rather limited antiparkinsonian activity but strongly suppresses levodopa-induced dyskinesias [65]. While not all noncompetitive or competitive antagonists have the ability to reduce levodopa-induced response fluctuations and dyskinesias in parkinsonian primates [66,67], animal model studies provide support for the view that dopaminergic denervation and especially intermittent DA receptor stimulation lead to striatal NMDA receptor sensitization.

[13] 87w Clinically, the ability of the NMDA receptor antagonist, amantadine, to palliate parkinsonian symptoms in early stage disease is well established [68]. More recently, this and other noncompetitive NMDA receptor antagonists, including dextrorphan and dextromethorphan, have been found to diminish levodopa-induced motor response fluctuations and peak dose dyskinesias [69 -71]. Subsequent studies have confirmed the safety and enduring efficacy of amantadine in treating the motor complication syndrome [72,73]. They further suggest that mechanisms underlying motor dysfunction in PD patients resemble those occurring in animal models of this disorder.

[14] 221w Observations in parkinsonian rodents and primates reveal differences between drugs that primarily block NMDA receptor NR2B subunits or NR2A subunits. Selective inhibition of rat NR2B subunits by the noncompetitive allosteric site antagonist Co 101244/PD 174494, but not of NR2A subunits by the competitive antagonist MDL 100,453 reverses levodopa-induced response alterations [42]. In parkinsonian monkeys, selective NR2B blockade with ifenprodil or CP101,606 has antiparkinsonian activity when given de novo [74] or with levodopa [75], while NR2B blockade with Co 101244/PD 174494 or CP 101,606 reduces levodopa-induced dyskinesias [75,76]. In contrast, selective NR2A antagonism with MDL 100,453 potentiates antiparkinsonian responses but exacerbates levodopainduced dyskinesias [76]. Taken together, these results support the possibility that the previously noted increase in tyrosine phosphorylation of NR2B subunits (possibly together with increased serine phosphorylation of NR2B subunits in D2 DA receptor bearing spiny neurons) may be particularly crucial to the NMDA receptor sensitization that favors the appearance of parkinsonism and especially to the altered dopaminergic responses (fluctuations and dyskinesias) produced by levodopa treatment. Since NR2B subunits predominate in mammalian striatum, while NR2A subunits are more highly concentrated in other brain regions, the foregoing possibilities may explain the relatively weak antiparkinsonian action of subunit nonselective NMDA antagonists. Furthermore, they suggest that NR2B subunit selective antagonists may have a better therapeutic index than nonsubunit selective NMDA antagonists in PD patients.

[15] 149w The functional implications of striatal AMPA receptor sensitization has also received investigative scrutiny in parkinsonian animals. Pharmacologic studies in 6-OHDA lesioned rats suggest that some AMPA antagonists can reduce parkinsonian signs when administered alone or with levodopa as well as prevent or reverse levodopa-induced motor response alterations [61,64,77]. Interestingly, the competitive AMPA antagonist NBQX has no effect on D2 DA agonist-induced rotations, but has been observed to potentiate the antiparkinsonian response to a D1 agonist [61]. In contrast, the competitive AMPA antagonist LY293558 reverses the response alterations elicited by a D-2 agonist but not those occurring with a D-1 agonist in levodopa-treated rats [78]. Thus in rodents the ability of AMPA antagonists to diminish denervation-induced parkinsonism may largely relate to their influence on the D1 direct striatonigral output pathway, while their ability to mitigate levodopa-induced response alterations appears more closely associated with effects on the D2 indirect striatopallidal pathway.

[16] 100w In parkinsonian primates, the competitive antagonist NBQX reportedly has inconsistent effects on parkinsonism [61,79,80]. However, a noncompetitive antagonist at the AMPA allosteric modulation site (LY 300164; talampanel) attenuates the dyskinesiogenic action of levodopa, while modestly augmenting its antiparkinsonian activity; in contrast, a positive modulator of AMPA receptors (CX 516) by itself has no antiparkinsonian activity, but exacerbates levodopa-induced dyskinesias [81]. Taken together with the efficacy of the putative glutamate release inhibitor riluzole [64,82], current evidence suggests that sensitization of striatal AMPA as well as NMDA receptors contributes to the motor dysfunction associated with dopaminergic denervation and dopaminomimetic induced response alterations.

[17] 95w Striatal spiny neuron dendrites are not only host to dopaminergic projections from substantia nigra and glutamatergic terminals descending from the cerebral cortex but also express receptors for numerous other transmitter systems. These include the serotonin 5HT2A, adenosine A2a, adrenergic alpha 2, and cannabinoid CB-1 receptors. By signaling through common phosphorylation -dephosphorylation cascades, activity at these receptors has the potential to modulate the effects of nonphysiologic DA receptor stimulation on glutamatergic receptor function. This possibility has profound implications both for understanding the pathogenesis of as well as for developing novel treatments for motor dysfunction in PD.

[18] 144w The A2a adenosine receptor is abundantly expressed on medium spiny neurons, especially those also having functional D2 DA receptors [95]. A2a receptors appear to signal in part through activation of PKA and PKC [96,97], which in turn can affect the regulation of ionotropic glutamatergic receptors [33,51]. Drugs that block A2a receptors in parkinsonian rodents and primates possess antiparkinsonian and especially antidyskinetic activity [98,99]. Specifically, studies in parkinsonian rats have revealed that the A2a receptor antagonist KW-6002 reverses both the levodopa-induced motor response alterations and the associated increase in Ser845 phosphorylation of striatal GluR1 subunits [55]. KW-6002 also reverses parkinsonian disability in MPTP-lesioned marmosets without inducing dyskinesias; even in parkinsonian primates that manifest these abnormal movements due to prior levodopa treatment, KW-6002 produces little if any dyskinesias [98]. Moreover, coadministration of KW-6002 daily apomorphine injections to MPTP-lesioned monkeys acts prophylactically to prevent dyskinesia onset [55].

[19] 123w Pilot clinical studies with KW-6002 (istradefylline) prompted by the foregoing animal model results suggest that at doses, which alone have no apparent effect on parkinsonian severity, the A2a antagonist potentiates the antiparkinsonian response to low dose levodopa with less dyskinesias than produced by optimal dose levodopa monotherapy [100]. KW-6002 also prolongs the efficacy half-time of levodopa and thus should reduce the severity of motor fluctuations. In a larger trial at lower doses, KW6002 was found to diminish 'off' time while increasing 'on' time with dyskinesia [101]. These findings are consistent with the view that drugs acting at transmitter receptors on spiny neurons to modify signaling kinases which affect the regulation of co-expressed ionotropic glutamatergic receptors can benefit extrapyramidal motor dysfunction in parkinsonian patients.

[20] 270w Serotonin 5HT2A receptors are also richly expressed on the dendrites of striatal spiny neurons [102] and their stimulation activates signaling pathways involving cAMP-and calcium-activated kinases such as PKA and PKC [103,104]. It is thus conceivable that drugs acting selectively at these receptors might also influence striatal mechanisms and thus motor function. Indeed, recent investigations in parkinsonian rats have shown that quetiapine, an atypical antipsychotic with 5HT2A/C as well as DA D2/D3 antagonist activity, normalizes the shortened levodopa response duration produced by chronic intermittent dopaminomimetic treatment, the rodent model of human motor fluctuations [105]. Moreover, quetiapine co-treatment acts prophyllactically to prevent onset of the levodopa-induced shortening in response duration. At a dose that was effective in reversing the curtailed motor response to levodopa, quetiapine diminished the enhanced Ser831 phosphorylation of spiny neuron GluR1 subunits produced by intermittent levodopa therapy [105], an effect consistent with the ability of 5HT2A receptor stimulation to increase serine phosphorylation at this AMPA subunit residue [106]. Similarly, in parkinsonian monkeys quetiapine co-administration attenuates levodopa-induced dystonic as well as choreiform dyskinesias at doses that do not interfere with the antiparkinsonian response to the DA precursor [105]. It is unlikely that DA receptor inhibition accounts for the ability of quetiapine to reduce levodopa-induced dyskinesia, since the same dose fails either to exacerbate parkinsonian signs when given alone or to attenuate the antiparkinsonian response to levodopa. Moreover, in parkinsonian rodents quetiapine prolonged rather than shortened the response to levodopa or DA agonists, as would be expected from drugs acting primarily via dopamine D2/D3 receptor blockade. These results can be expected to stimulate clinical explorations of their therapeutic applications to PD.

[1] 132w Increasing insight into striatal mechanisms contributing to symptom production in PD will undoubtedly accelerate the development of novel approaches to therapy. The clinical appearance of parkinsonism and later of motor response complications is associated with a complex array of functional alterations in medium spiny neurons. The nonphysiologic stimulation of DA receptors on these striatal neurons activates signaling cascades linking them to nearby ionotropic glutamatergic receptors. As a result, the phosphorylation state of NMDA and AMPA subunits changes in ways that profoundly influence their synaptic efficacy and thus excitatory input from cortex to striatum. Acccordingly, the output of these striatal efferents undergoes various changes that compromise motor function. The detailed elucidation of relevant molecular mechanisms will have important implications for our understanding of brain disorders involving the integration and retention of synaptic information.

[2] 92w More specifically, the possibility that medium spiny neuron dysregulation contributes to the motor disabilities attending dopaminergic denervation and intermittent stimulation suggests a new therapeutic paradigm for PD. Rather than continuing classical attempts to replace the deficient neurotransmitter, it may prove safer and more effective to pursue alternative pharmacologic approaches to the preventionor reversal of the reactive downstream changes produced by current therapy. Potential strategies include the pharmaceutical targeting of medium spiny neuron (1) protein kinases whose anomalous activation by the nonphysiologic stimulation of DA receptors modifies the phosphorylation state of glutamatergic receptors;

[3] 66w (2) ionotropic glutamatergic receptors whose sensitization corrupts striatal output; (3) cell surface receptors capable of modulating signal transduction cascades that link dopaminergic and glutamatergic receptors; and (4) transcription factors that promote retention of maladaptive responses to past stimuli. The vigorous exercise of these and related therapeutic options may prove helpful not only for those suffering from PD but also for those disabled by other neurodysfunctional disorders.

[1] 297w The nigrostriatal dopaminergic system mainly operates tonically. Nerve impulse activity in neurons comprising this pathway occurs rather constantly at a modal rate of 4 Hz, except when briefly interrupted by phasic bursts triggered in nonhuman primates by unexpected but potentially rewarding stimuli [7]. Since the quantity of transmitter release reflects the rate of neuronal firing, intrasynaptic DA receptors are generally exposed to stable transmitter concentrations. Most dopaminergic receptors, however, are located extrasynaptically [8,9], and are activated by DA diffusing from the synaptic cleft into the extracellular space [10]. Normally, striatal DA concentrations remain within a fairly narrow range [11,12]. But this situation changes dramatically in PD when dopaminergic receptors come in contact with nonphysiologic levels and patterns of stimulation [13,14]. First, it changes as dopaminergic neurons degenerate and DA levels decline. The abnormally low-intensity stimulation received by striatal dopaminergic receptors in untreated PD leads to the appearance of the cardinal signs of this disorder [15]. Later, it changes even more with the initiation of dopaminomimetic therapy. With few residual DA neurons, standard treatment regimens subject postsynaptic DA receptors to intermittent periods of high-intensity stimulation. Intermittency is the consequence of dopaminergic neuron loss, which curtails vesicular mechanisms for transmitter storage [16,17]. High-intensity is indicated by the excessive DA and DA metabolite levels found in striatal tissue and lumbar CSF with conventional therapy [18,19]. Episodic, high-intensity stimulation is no more physiologic than the subthreshold stimulation attending dopaminergic denervation [3]. Accordingly, striatal dopaminergic transmission remains nonphysiologic in those with advanced PD whether left untreated or given standard therapeutic regimens. Conceivably, either type of abnormal DA receptor stimulation may be able to alter spiny neuron function in ways that contribute to the increasing severity of wearing-off fluctuations as well as to the onset of unpredictable fluctuations and peak dose dyskinesias [20].

[2] 260w Both the structure [2] and function of striatal medium spiny neurons are affected by chronic exposure of their DA receptors to nonphysiologic stimulation. As observed in animal models of PD, functional effects include changes in DA receptor-linked signaling that affect the phosphorylation state and thus the synaptic efficacy of co-expressed ionotropic glutamatergic receptors. Unilateral injection of the catecholamine-specific neurotoxin 6-hydroxydopamine (6-OHDA) into rat substantia nigra destroys DA neurons projecting to the ipsilateral striatum and produces contralateral hypokinesia mimicking what occurs in human PD. The acute administration of levodopa produces contralateral rotations in these animals. With chronic levodopa treatment, given intermittently to simulate dosing schedules in parkinsonian patients, alterations in motor response appear within a few weeks that resemble human wearing-off and on -off fluctuations [13,21]. Animals receiving the same daily dose of levodopa by continuous round-the-clock infusion do not develop these motor response changes. With dopaminergic denervation and to an even greater extent with intermittent dopaminergic therapy, the sensitivity of NMDA and AMPA receptors on GABAergic spiny neurons rises. As a result, cortical glutamatergic excitation of these GABAergic efferent neurons increases and hence so does the activity of their GABA synthesizing enzyme, glutamic acid decarboxylase [22]. Similarly, mRNA and peptide levels for such spiny neuron co-transmitters as enkephalin, dynorphin and neurotensin undergo characteristic alterations in both these conditions, but not in parkinsonian rats given continuous levodopa treatment [22 -25]. Associated neurophysiologic and neuropharmacologic observations indicate that the transmitter modifications reflect increases in the rates and burstiness of medium spiny neuron firing that underlie the alterations in motor behavior [26 -28].

[3] 149w In parkinsonian animals and patients with a severe loss of DA neurons, it can take only a few days or weeks of intermittent dopaminomimetic treatment to produce motor response changes [21,83 -86]. Similarly, in both PD models and patients it appears to require a few weeks for these response alterations to significantly diminish when intermittent treatment is replaced by relatively continuous dopaminergic stimulation [87 -89]. These latter observations suggest that mechanisms responsible for the response modifications to dopaminomimetic therapy may linger long after removal of the inciting stimulus. Recent observations in parkinsonian rats suggest that the expression and maintenance of levodopa-induced motor response alterations are linked to the regulation of cAMP response element-binding protein (CREB), a transcription factor implicated in the formation and persistence of long-term memory [90,91]. The transcriptional activation of CREB depends on its phosphorylation at Ser133 either directly or indirectly by kinases such as PKA [92].

[4] 168w Unilateral nigrostriatal system destruction causes a small increase in the Ser133 phosphorylation of CREB in ipsilateral medium spiny neurons in response to acute levodopa challenge [93]. Intermittent levodopa treatment sufficient to induce motor response alterations in these hemiparkinsonian animals potently augments both total phospho-CREB expression and the number of phospho-CREB-positive neurons in the ipsilateral striatum; these changes persist for up to four weeks following withdrawal of the twice-daily levodopa therapy [93]. Agents, such as CREB antisense or the PKA inhibitor Rp-cAMPS, that disrupt the activity of this transcriptional factor attenuate both the change in motor response duration and the degree of striatal CREB phosphorylation. These results, together with the finding that chronic intermittent stimulation of D1 receptors, but not of D2 receptors, increases the magnitude of striatal phospho-CREB expression, indicate that the activation of D1/PKA-mediated phosphorylation of CREB family proteins contribute to the expression and maintenance of levodopa-induced motor response alterations [93,94]. Further elucidation of neuronal mechanisms regulating the persistence of motor memories could have important therapeutic implications.