Presynaptic Mechanisms of Motor Complications in Parkinson Disease

The pathogenesis of levodopa-related motor complications (motor fluctuations and dyskinesias) is still far from clear, but there is converging evidence to suggest that presynaptic mechanisms play a major role. In fact, we have recently developed a mathematical probability model that can accommodate the whole spectrum of motor complications often seen in Parkinson disease (PD).¹ The model is based on dynamic studies using positron emission tomography^{2 - 3} and other sources of research evidence.¹

It is pertinent to emphasize that dopamine cell loss seems to be a requirement for levodopa-related motor fluctuations but not for dyskinesias.¹ In fact, recent animal experiments have shown that dyskinesia can occur in normal subjects treated with higher doses of levodopa.⁴ Although it may seem trivial, one should also bear in mind that the objective of dopaminergic treatment in PD is not simply to reach normal levels of dopamine in the striatum but rather to restore normal stimulation of postsynaptic dopamine receptors.¹ Our probabilistic model of motor complications in PD can be summarized as follows:

It is well known that exogenously administered levodopa is converted into dopamine and stored in the vesicles present in nigrostriatal nerve terminals. The ultimate fate of this exogenously derived dopamine mostly depends on the relationship between dopamine release and dopamine reuptake rates. Evidence suggests that a 100-mg tablet of levodopa (with carbidopa) may be sufficient to replenish the depleted nigrostriatal system with dopamine.¹ Higher doses of levodopa lead to increases in the quantal size (ie, greater number of dopamine molecules per releasable vesicle).¹

Naturally, the loss of dopamine terminals in PD compromises the reuptake capacity of the system; the molecules of released dopamine that are not taken back into the nerve terminal will be metabolized and lost. Therefore, the greater the loss of dopamine terminals, the shorter the duration of the clinical benefit obtained after each dose of levodopa, which explains the well-known association between PD severity and motor fluctuations.¹ Interestingly, however, the model shows that differences in compensatory mechanisms (particularly, dopamine release rate)^{1 ,5} can also explain why, for the same degree of PD severity, some patients develop fluctuations while others do not.¹ Thus, individuals who have a higher release rate will lose dopamine faster and, consequently, display a shorter duration of response to levodopa. The increased risk for motor fluctuations in younger patients with PD may be related to their relatively high dopamine release rate compared with older patients.²

Remarkably, our probabilistic model also explains the occasional occurrence of on-off oscillations after a dose of levodopa through a pure presynaptic mechanism. This “yo-yo-ing” phenomenon has long puzzled the scientific community¹ and was indeed one of the major reasons for invoking postsynaptic mechanisms in the pathogenesis of motor complications in PD. The model shows, however, that the yo-yo-ing phenomenon may simply obey random variations in the amount of dopamine of releasable vesicles. In support of this notion, recent work suggests that releasable vesicles are randomly dispersed throughout the vesicle cluster.⁶

One finding that is often used to buttress the theory that motor fluctuations are predominantly postsynaptic in nature is the observation that the response to apomorphine may be shorter in patients with motor fluctuations compared with stable responders.⁷ However, it is obvious that presynaptic mechanisms are involved in this differential response to apomorphine. First, the model shows that the distribution of vesicular levels of endogenous dopamine in the off state is particularly skewed in patients with motor fluctuations, the degree of skewness being inversely proportional to the amount of endogenous dopamine molecules per vesicle. That is, the higher vesicular release rate found in fluctuators is associated with a greater number of vesicles containing low levels of dopamine (and this principle holds true even in fluctuators with identical loss of dopamine nerve terminals to that observed in stable responders).¹ This scenario sets the stage for a reduced contribution of endogenous dopamine to the maintenance of postsynaptic dopamine receptor stimulation in the fluctuator group, which becomes particularly apparent as the effect of apomorphine fades. In addition, apomorphine (as well as other direct dopamine agonists) acts not only on postsynaptic dopamine receptors but also on D2 autoreceptors present on dopamine cells (soma, dendrites, and terminals).⁸ It is well known that the stimulation of these autoreceptors leads to decreased vesicular dopamine release.^{1 ,8} In fact, the regulatory role of D2 autoreceptors is a major mechanism of control of synaptic dopamine levels.⁸ Interestingly, there is evidence to suggest that fluctuators have autoreceptor dysfunction, showing lack of response to dopaminergic stimulation.¹ Hence, under the effect of apomorphine, stable responders (but not fluctuators) would tend to increase the vesicular levels of endogenous dopamine. Consequently, only stable responders can transitorily maintain optimal levels of postsynaptic dopamine receptor stimulation as the effect of apomorphine dissipates, thereby leading to longer duration of the clinical benefit obtained after apomorphine administration.

Patients with PD often develop levodopa-related dyskinesias, but these abnormal movements can also appear after the administration of direct dopamine agonists. In fact, this observation prompted the postsynaptic theory of dyskinesias. It was initially thought that the pathogenesis of dyskinesias was related to dopamine receptor up-regulation.⁹ However, in vivo positron emission tomography studies showed that patients with dyskinesias lack any specific change in either D1 or D2 receptors.^{3 ,10} The focus was then moved to postsynaptic mechanisms further downstream in the circuitry of the basal ganglia. In fact, different neurotransmitters, including glutamate, opioids, and adenosine (as well as the corresponding systems), have been implicated, but no specific factor to blame has yet been found.¹⁰ It is unclear whether the contribution of these neurotransmitters to the pathogenesis of dyskinesias is causal or secondary. In addition, the site where these substances exert their putative antidyskinetic effect also remains to be defined. Thus, for example, the antidyskinetic effects of glutamate antagonists¹¹ could simply reflect the presynaptic modulatory effects that these drugs have on dopamine release.^{8 ,12}

We proposed that dyskinesias are primarily related to large swings in the level of occupancy/stimulation of postsynaptic dopamine receptors.^{1 ,3} This simple mechanism of dyskinesias can be easily accommodated into the probabilistic model of motor fluctuations previously outlined, and explains a number of clinical and research observations. This integrative model predicts that the risk of dyskinesias increases as PD progresses.^{1 ,3} Thus, relative to the off state, increasing numbers of postsynaptic receptors need to be stimulated to reach an on state, which would lead to dyskinesias. On the other hand, for the same degree of loss of nigrostriatal terminals, the greater the release rate of dopamine, the greater the change in synaptic levels of dopamine after levodopa administration, and the higher the risk of levodopa-related dyskinesias. These predictions explain the well-known association between motor fluctuations and dyskinesias.^{1 ,10} Because levodopa produces a higher degree of change in synaptic dopamine levels in fluctuators, the model also explains why the dose of levodopa needed to evoke dyskinesias is lower in this group of patients compared with stable responders.¹³ Also in keeping with the model are recent postmortem observations, which have shown that dopamine is mostly metabolized in the extracellular compartment in patients with dyskinesias.¹⁴

Naturally, dyskinesias can be controlled by decreasing the quantal size (by lowering the dose of levodopa), but this often leads to worsening parkinsonian symptoms and more severe motor fluctuations. In contrast, dramatic increases in the quantal size induced by very large doses of levodopa can evoke dyskinesias even in normal subjects. As predicted by the model, this will occur whenever the induced change in the level of dopamine receptor occupancy is of enough magnitude. The model also predicts that any therapeutic maneuver aimed at reducing oscillations in the level of dopamine receptor stimulation would decrease the risk of dyskinesias and would also be effective in treating these abnormal movements. In keeping with these predictions, the risk of dyskinesias seems to be reduced in patients with PD on direct dopamine agonists,^{3 ,10} and there is evidence that maintaining quasi-constant levels of dopamine receptor stimulation helps control dyskinesias (and motor fluctuations).¹⁵ The latter observation highlights the notion that motor fluctuations and dyskinesias not only share similar mechanisms of pathogenesis but also benefit from the same therapeutic strategies.

One final lesson about the pathogenesis of motor complications in PD can be derived from clinical observations obtained from patients with dopa-responsive dystonia. These patients have the same degree of striatal dopamine depletion as patients with PD and who also receive long-term treatment with levodopa.¹ If postsynaptic mechanisms contributed in any major extent to the pathogenesis of motor fluctuations and dyskinesias in PD, these motor complications should also occur in dopa-responsive dystonia. However, as predicted by our model, significant levodopa-related motor fluctuations do not occur in dopa-responsive dystonia, and dyskinesias, when present, are usually mild and easily controlled by reducing the dose of levodopa.¹

In summary, while postsynaptic changes likely play a role in the pathogenesis of motor complications (particularly dyskinesias), the following key points should be adequately weighed: (1) the probabilistic model discussed here is sufficient to explain motor fluctuations based on altered presynaptic function; (2) our model also suggests that dyskinesias may not need the contribution of postsynaptic mechanisms; and (3) postsynaptic changes are most likely secondary to altered presynaptic function and could not otherwise occur.

Correspondence: Raúl de la Fuente-Fernández, MD, Division of Neurology, Hospital A. Marcide, 15405 Ferrol, Spain (rfuente@medynet.com).

Accepted for Publication: September 5, 2006.

Financial Disclosure: None reported.