|Home | About | Journals | Submit | Contact Us | Français|
Chronic L-DOPA treatment for Parkinson's disease often results in the development of abnormal involuntary movement, known as L-DOPA-induced dyskinesia (LIDs). Studies suggest that LIDs may be associated with aberrant cortico-striatal plasticity. Using in vivo extracellular recordings from identified type I and type II medium spiny striatal neurons, chronic L-DOPA treatment was found to produce abnormal cortico-striatal information processing. Specifically, following chronic L-DOPA treatment in dopamine-depleted rats, there was a transition from a cortically-evoked long-term depression (LTD) to a complementary but opposing form of plasticity, long-term potentiation, in type II `indirect' pathway neurons. In contrast, LTD could still be induced in type I neurons. Interestingly, the one parameter that correlated best with dyskinesias was the inability to de-depress established LTD in type I medium spiny striatal neurons. Taken as a whole, we propose that the induction of LIDs is due, at least in part, to an aberrant induction of plasticity within the type II indirect pathway neurons combined with an inability to de-depress established plastic responses in type I neurons. Such information is critical for understanding the cellular mechanisms underlying one of the major caveats to L-DOPA therapy.
Parkinson's disease (PD) is characterized by a progressive degeneration of nigrostriatal dopaminergic neurons1 and replacement therapy with the dopamine (DA) precursor L-3, 4-dihydroxyphenylalanine (L-DOPA) is the most effective therapy2. Unfortunately, it often induces adverse fluctuations in motor response such as the end-of-dose “wearing off”, and the development of involuntary movements, known as L-DOPA-induced dyskinesia (LID). Thus, patients may cycle between ON state complicated by LIDs and OFF state in which the motor symptoms re-emerge.
Several studies suggest that aberrant corticostriatal plasticity may underlie the development of LID3–5. Recently, Picconi and colleagues have described plastic changes in the striatum of 6-hydroxydopamine (6-OHDA) lesioned rats displaying LID6, representing a significant advance in understanding the mechanisms underlying this complication. One of the caveats of such studies is that the examination of neuroplasticity is performed in slices in which extracellular solutions are artificially modified to reduce extracellular Mg2+, which would favor long-term potentiation (LTP) development6, 7. This significantly alters the response consistently observed in the majority of studies examining neuroplasticity in the striatum using both in vitro and in vivo approaches8, 9. Indeed, the prominent form of neuroplasticity observed in the striatum of rats upon application of high frequency stimulation is not LTP but rather a complimentary process known as long-term depression (LTD).
The striatum has two distinct output projection pathways 10and a significant literature suggests that the motor symptoms of PD arise due to an imbalance within these pathways11–13. Thus, it is surprising that the majority of studies investigating synaptic plasticity in the striatum of PD and/or dyskinetic rats have failed to distinguish between these striatonigral and striatopallidal neurons.
Given evidence for altered synaptic plasticity associated with LID and evidence of alterations in direct vs indirect pathways in PD, we examined the regulation of striatal neuron plasticity in vivo in normal vs DA lesioned rats, and further subdividing L-DOPA-treated rats into those exhibiting dyskinesia and those that do not. We employed our previously established criteria14, further confirmed here, to distinguish striatonigral and striatopallidal neurons to determine whether synaptic plasticity is altered in the direct vs indirect pathways.
All experiments were performed in accordance with the guidelines outlined in the USPHS Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh.
Unilateral 6-OHDA lesions of the medial forebrain bundle (mfb) were performed as previously described15. Following three weeks recovery, rats were treated twice daily with L-DOPA (10 mg/kg i.p.) and benserazide (7.5 mg/kg i.p.) or vehicle (saline 1 ml/kg i.p.) for three weeks. This dose of L-DOPA was specifically selected to induce dyskinesia in approximately 50% of the animals6.
The expression of L-DOPA-induced dyskinesia has been extensively studied6, 16, 17. Thus, over the three-week period LIDs were examined using the standard abnormal involuntary movement (AIMs) scale and a test for limb asymmetry. Briefly, rats were placed individually in a glass cylinder 20 mins following the injection of L-DOPA/benserazide. AIMS were rated between 0–4 for orolingual, axial and limb over 5 mins. Limb asymmetry was determined by the number of wall contacts by the ipsilateral vs. contralateral forelimb over 5 mins.
Given that the effect of L-DOPA on striatal dopamine decreases after 6–7 hours18, 19, rats were anaesthetized with urethane (1.2g/kg i.p.) 12–24 hours following the last L-DOPA injection. Stimulation of the motor cortex (M1) and of the substantia nigra pars reticulate (SNr) was performed with bipolar concentric stimulating electrodes (NEX100). Extracellular glass microelectrodes (impedance 6–14MΩ) were lowered into the dorsal striatum and type I and II medium spiny neurons (MSN) were identified based on paired-pulse (inter-stimulus interval: 100 ms) cortical stimulation as previously reported14, 20. Antidromic activation of the SNr was used to further confirm the identity of striatonigral neurons. Once a stable putative MSN was identified, the stimulation intensity was adjusted to evoke an action potential approximately 50 % of the time and the baseline response to M1 stimulation (0.5 Hz pulses) was recorded for 20 mins. After the baseline period, high frequency stimulation (HFS) of the M1 pathway (3 × 2 s trains, 1 mA, 50 Hz, at a rate of 0.05 trains per second) was applied. The response to M1 stimulation was recorded for a further 40 mins before low frequency stimulation (LFS) of the M1 pathway (5 Hz, 500 pulses, 1 mA). M1 evoked responses were examined for an additional 40 mins.
For retrograde tracing experiments both control and unilateral 6-OHDA lesioned rats (250–300g) were anesthetized with ketamine/xylazine (80/12mg/kg i.p. respectively). Recombinant cholera toxin subunit B conjugated to Alexa Fluor 488 (1.0 μg in 0.5 μl dPBS) was microinjected into the external globus pallidus. Following one week recovery, rats were anaesthetized with urethane (1.2g/kg i.p.) and type I and II MSN's were identified by paired-pulse M1 stimulation. Extracellular glass microelectrodes (in situ impedance 40–80MΩ) filled with 2% Neurobiotin were connected to an Axoclamp-2A amplifier in bridge mode to enable simultaneous recording and current injection. Once a stable putative MSN was identified, current (+1 to +4 nA, 200 ms duration, 2 Hz) was passed and the cell entrained to depolarizing current steps.
At the cessation of experiments, the recording site was marked via electrophoretic ejection of Pontamine sky blue dye (−25 μA constant current: 20–30 mins) and iron deposited by passing constant current (0.3 mA, 10 s) across the stimulating electrode. Rats were perfused transcardially with paraformaldehyde, decapitated, and their brains removed and post-fixed for 24 hours. Brains were cryoprotected then sliced coronally (60 μm). Sections were Nissl stained for histochemical verification with reference to a stereotaxic atlas21. Additional sections of striatum were processed for tyrosine hydroxylase immunoreactivity as previously described22 (Supplemental Fig S1). Juxtacellularly labelled cells were processed with streptavidin-Alexa Fluor 594.
Electrophysiological analysis was performed using custom-designed software (Neuroscope). Densitometric analysis of each striatal section was performed with reference to the M1 contralateral to the lesion using the SCION imaging system (PC version of NIH image). All data are represented as the mean ± the standard error of the mean (S.E.M.). Data were analyzed using a one-way ANOVA with repeated measures followed by the Holm-Sidak test. Multiple comparisons were analyzed using a two-way ANOVA followed by the Holm-Sidak test (SigmaStat software program, Systat Software Inc).
Detailed methodology is described in the Supplementary Methods.
All 6-OHDA-lesioned rats used in this study demonstrated a substantial loss of striatal tyrosine-hydroxylase immunoreactivity in comparison with intact animals (−93.3 ± 0.9 % decrease, n=62).
As previously described, repeated L-DOPA induced gradual development of dyskinesia in approximately 50 % of 6-OHDA-lesioned rats 6, 16, 17. A significant increase in AIMS was observed for one group of rats (dyskinetic) after day 6 and increased gradually until day 17 (day 1: 0.26 ± 0.13 axial AIMS; 0.44 ± 0.17 limb AIMS; 0.07 ± 0.07 orolingual AIMS; day 17: 3.52 ± 0.23 axial AIMs; 3.74 ± 0.16 limb AIMS; 3.62 ± 0.22 orolingual AIMS; p<0.05; n=27) (Fig 1 A to C). A gradual increase in the limb-use asymmetry score was also observed (day 1: −5.7 ± 10.6 %; day 17: 64.8 ± 10.2 %, p<0.05, n=27) (Fig 1D). The non-dyskinetic group failed to show significant changes in AIMs (day 1: 0.08 ± 0.05 axial AIMs; 0.17 ± 0.12 limb AIMs; no orolingual AIMs; day 17: no axial AIMs; 0.29 ± 0.16 limb AIMs; 0.04 ± 0.04 orolingual AIMs; NS; n=24; Fig 1 A to C) or limb-use asymmetry score (day 1: −26.8 ± 10.3 %; day 17: −28.3 ± 7.6%, NS, n=24; Fig 1D). Thus, consistent with previous characterizations, the expression of LIDs was extremely robust with rats displaying either a full complement of axial, orofacial and limb AIMs or no dyskinesias.
Striatonigral (type I) and striatopallidal (type II) neurons were distinguished by their response to cortical stimulation, as described previously14 (Fig 2 and and3).3). Thus, striatonigral neurons responded to the paired cortical stimulation with a higher probability of spike generation in response to the first pulse, whereas striatopallidal neurons were more responsive to the second pulse (Fig 2 and and33)14. The identification of these distinct MSNs was confirmed by antidromic activation from the SNr as well as juxtacellular labeling combined with retrograde tracing from the GP. Specifically, identified type I neurons were consistently antidromically activated from the SNr whereas type II neurons were not (Fig 2; type I: 12/19 neurons; type II: 5/25; p<0.05, z=2.60, z-test with Yates correction). Furthermore, juxtacellularly-labeled type II neurons were consistently labeled retrogradely from the GP, whereas type I neurons were devoid of retrograde tracer (Fig 3; type I: 2/14 neurons; type II: 8/12; p<0.05, z=2.33 z-test with Yates correction). These data substantiate our previous characterizations of type I and type II neurons as MSNs compromising the direct and indirect pathways, respectively14, 20, 23.
Consistent with previous results24, 25, HFS of the corticostriatal pathway of intact rats induced LTD of cortically-evoked action potential generation in both types of neurons (striatonigral: p<0.05, n=8/8; striatopallidal: p<0.05, n=6/8; Fig 4A and B). In lesioned rats, the majority of striatopallidal neurons didn't exhibit LTD following HFS (ns; n=8/10; Fig 4D), confirming the block of corticostriatal plasticity after dopaminergic denervation6, 26.. Interestingly, striatonigral neurons in lesioned animals still exhibited LTD after HFS (p<0.05; n=7/7) that was not significantly different from the response recorded in intact animals (ns; Fig 4C). All average values are summarized in supplemental table 1.
LFS consistently reverses established plasticity in corticostriatal synapses4, 6, 9, 27, 28. Thus, LFS performed 40 minutes after HFS induced no change in striatopallidal neurons of intact and lesioned animals (ns; n=8 and n=7, respectively, Fig 4B and D). Comparison of evoked activity following HFS with that post LFS revealed a significant reversal of the LTD in striatonigral neurons of control animals (p<0.05; n=8, Fig 4A). Moreover, this return to baseline is lost after 6-OHDA lesions (p<0.05 in comparison with control; n=7) (Fig 4C).
LTD and LTP are long-term changes in synaptic efficacy that constitute one of the cellular models for information storage. In addition, depotentiation and de-depression represent another level of plasticity and can alter past synaptic information processing29. Previous studies have suggested that aberrant depotentiation may lead to the development of abnormal motor patterns6. Therefore, we assessed synaptic plasticity changes in both L-DOPA treated dyskinetic and non-dyskinetic rats. In striatonigral neurons, LTD was induced by cortical HFS in both non-dyskinetic (p<0.05; n=5/7; Fig 5A) and dyskinetic rats (p<0.05; n=7/9; Fig 5C). No significant differences were observed between the LTD in dyskinetic and non-dyskinetic rats (ns; Fig 5A, n=7 and Fig 5C, n= 5, respectively). In non-dyskinetic rats, LFS attenuated the HFS-evoked LTD (−22.4 ± 9.9%; p<0.05; n=5), whereas no effect on the LTD in dyskinetic rats was observed (ns; n=5).
Interestingly, HFS in striatopallidal neurons induced changes in synaptic plasticity in both dyskinetic and non-dyskinetic rats. Thus, in contrast to the LTD observed in untreated rats in vivo, HFS induced a complementary but opposite form of synaptic plasticity, i.e. LTP. As a result, an increase in the cortical-evoked spike probability of MSNs was observed following HFS in both non-dyskinetic (p<0.05; n=9/11; Fig 5B) and dyskinetic rats (p<0.05; n=9/12; Fig 5D). Moreover, LFS fully reversed the LTP in type II neurons from dyskinetic rats (p<0.05 in comparison with 20 minutes after HFS and ns in comparison with control; n=6), whereas non-dyskinetic rats failed to show reversal by LFS (ns compared to 20 minutes after HFS, p<0.05 compared to control; n=9).
The data presented here demonstrate major adaptive changes in corticostriatal information processing following chronic L-DOPA administration in the hemi-parkinsonian rat model. Specifically, we report that the primary deficit in synaptic information processing observed following 6-OHDA lesion is the inability to induce corticostriatal LTD in indirect pathway neurons. In addition we demonstrate that chronic L-DOPA treatment results in dramatic alterations in both the induction and normalization of synaptic plasticity in both direct and indirect pathway MSNs.
Based on intracellular staining and responses to dopamine agonists the characterization of type I and type II neurons and their identification as neurons corresponding to the `direct' and `indirect' pathways, respectively, has been extensively documented23. In the present study, this characterization has been confirmed using retrograde tracing (from the GP) combined with juxtacellular labeling as well as by antidromic activation from the SNr. It should be noted that antidromic activation of striatal neurons has been considered unreliable14, presumably due to current shunting by K+ channels14, 30. Indeed, in the present study, 35% of type I neurons failed to show an antidromic response. This could be the reason why a recent set of publications report the opposite characterization of MSN's based exclusively on antidromic activation from the SNr31.
HFS of the motor cortex has been demonstrated consistently to induce LTD of cortically-evoked action potentials in striatal MSNs in vitro9 as well as in vivo8, 32. In the current study, changes in plasticity, either potentiation or depression, were measured as changes in spike probability. HFS resulted in the induction of LTD in both striatonigral and striatopallidal neurons of control rats. Following 6-OHDA lesion, LTD could still be evoked in striatonigral neurons; in contrast, there was no significant effect of cortical HFS in striatopallidal neurons. These data are consistent with previous studies examining the role of dopamine D1 and D2 receptors in corticostriatal plasticity. Thus, corticostriatal LTD can be attenuated by sulpiride, and is absent in brain slices from both 6-OHDA lesioned rats7 and mice lacking D2 receptors33. Thus, the absence of LTD in striatopallidal neurons observed in the current study is consistent with previous data reporting the dependence of LTD on D2 receptor activation34. In contrast to the effect on striatopallidal neurons, there was no effect of unilateral DA depletion on the ability of cortical HFS to induce LTD in striatonigral neurons. This result is consistent with previous studies showing that corticostriatal LTD persists in mice lacking D1 receptors27, although D1 receptors antagonists will block corticostriatal LTD35. Taken together, our data suggest that a decrease in DA levels within the striatum would result in the loss of a component essential for the development of LTD predominantly in striatopallidal neurons, and a loss of control over motor inhibition.
Although corticostriatal plasticity following chronic treatment with L-DOPA has been investigated previously, such studies were limited in their interpretation due to several important factors: 1) these studies were performed in vitro in which afferent processes have been severed6;2) changes in LTP were biased by either depolarization of the neuronal membrane in vivo36 or omission of magnesium ions in vitro6, 7,; and 3) neurons of the direct and indirect pathway were not distinguished, confounding the interpretation of the functional outcome of the observed changes . Thus, we examined the effect of chronic L-DOPA treatment on corticostriatal plasticity in both striatonigral and striatopallidal neurons. Consistent with our observation that unilateral DA depletion only altered activity-dependent plasticity in striatopallidal neurons, the alterations in HFS-induced corticostriatal plasticity observed following chronic L-DOPA administration were restricted to the indirect projection neurons. Indeed, there was no disruption in cortically-evoked LTD in direct projecting neurons following DA denervation or after chronic L-DOPA treatment, suggesting that the induction of corticostriatal plasticity in striatopallidal neurons is not sensitive to tonic changes in dopaminergic transmission. In contrast, our most striking result is that chronic L-DOPA treatment resulted in a transition from cortically-evoked LTD to the complementary but opposite process of LTP in striatopallidal neurons. This is consistent with previous studies showing that HFS of cortical fibers from D2 receptor mutant mice robustly induces an NMDA independent LTP27. Such a significant change in corticostriatal information processing is likely to have a dramatic effect on motor output under conditions of high cortical activity, and interfere with the ability of the indirect pathway to effectively modulate direct pathway motor output.
Previous data suggested that aberrant de-potentiation of LTP may play a role in LIDs. Thus, low frequency stimulation of cortical afferents can reverse LTP evoked by HFS6. Similarly, in the present study LFS was able to attenuate LTD evoked by cortical HFS in striatonigral neurons recorded from intact rats. In contrast in hemiparkinsonian rats, LFS further depressed cortically driven responses suggesting a potential dopamine-dependent mechanism contributing to de-depression.
Interestingly, whereas no differences in HFS-evoked responses were observed between dyskinetic and non-dyskinetic rats, significantly different results were seen following LFS. Thus, chronic L-DOPA appeared to repair the deficit in LTD reversal observed in striatonigral neurons of lesioned rats, but only in the non-dyskinetic rats. In contrast, the deficit in LTD-reversal was retained in striatonigral neurons of dyskinetic rats. These data are in agreement, at least in part, with previous experiments reporting that rats displaying LIDs fail to reverse HFS-induced LTP in unidentified striatal neurons recorded in vitro6. Our data suggest a loss of regulation over LTD in dyskinetic rats causing the striatonigral neurons to retain abnormally large depression of responses. Such an imbalance may be further exacerbated by the indirect neurons, which were found to display LTP (rather than LTD) following chronic L-DOPA treatment; an effect that could be reversed by LFS only in dyskinetic rats but not in non-dyskinetic rats. Consequently, in dyskinetic rats the regulation of synaptic plasticity in direct and indirect pathways are out of balance; LTD in striatonigral neurons is abnormally persistent and fails to be down-modulated, whereas in striatopallidal neurons LTP is abnormally sensitive to reversal. As such, there would be a strong attenuation in the ability of the indirect pathway to regulate the direct pathway during low activity periods.
One potential caveat in the interpretation of the current results is that chronic L-DOPA treatment exhibited substantial effects in both dyskinetic and non-dyskinetic rats. However, it must be noted that all rats in the current study were recorded in the drug-free state while the expression of LIDs is most prominent during peak cerebrospinal fluid levels of L-DOPA37. Nonetheless, it has been suggested that the potential clinical actions of chronic L-DOPA administration arise from the induction of a priming effect that only manifests as LID during high levels of the drug38. Thus, we propose that the aberrant corticostriatal plasticity in indirect pathway neurons may be responsible, at least in part, for the so called “priming” of the system. In addition, a deficit in the ability to reverse established plastic changes in direct neurons is observed only in L-DOPA-treated rats that display dyskinesia in response to an acute L-DOPA injection. This inability to reverse established plasticity is suggested to result in a pathological storage of nonessential motor information. Thus the aberrant induction of striatal plasticity in striatopallidal neurons combined with the disparate response to LFS in striatonigral and striatopallidal neurons during OFF states could facilitate LIDs expression during ON states.
The authors thank Niki MacMurdo & Emily Mahar for their technical assistance, Brian Lowry for the production, development and support with the custom designed electrophysiology software (Neuroscope) and Kathryn Gill for critical reading and helpful discussions. This work was supported by the USPHS DA15408 and MH57440 (AAG), a Young Investigator Award from NARSAD - The Mental Health Research Association (DJL) and an award from Fondation pour la Recherche Medicale (PB).
This work was supported by the USPHS DA15408 and MH57440 (AAG), a Young Investigator Award from NARSAD - The Mental Health Research Association (DJL) and an award from Fondation pour la Recherche Medicale (PB).
Financial disclosure:Grants: Lundbeck; Boards: Puretech Ventures; Consultants: Johnson & Johnson, Taisho; Honoraria: Lilly
Author roles: Belujon: 1, 2, 3; Lodge: 1, 2, 3; Grace: 1, 2C, 3B