Levodopa

For almost 30 years, levodopa has been administered to millions of patients with Parkinson's disease as the gold standard of treatment. In everyday practice, levodopa is the most effective symptomatic treatment available: patients live longer and their quality of life improves, often dramatically initially and thereafter remaining satisfactory. However, two types of problems which we are ill-equipped to handle can arise as the disease progresses. In end-stage patients, axial signs such as postural instability and difficulty walking, sphincter dysfunction, dysarthria, or mental deterioration can appear. These symptoms respond poorly to levodopa and its derivatives. Additionally, disabling motor side effects of levodopa(abnormal involuntary movements and fluctuations of performance) can occur in early stages, and require adjustment of the daily doses and treatment regimen.

In recent years, we have encountered a growing number of patients without such side effects who are not deriving the full benefit from antiparkinsonian treatment. The reason behind the appearance of this new category of patients seems evident: these patients are undertreated. They remain bradykinetic, depressed, and incapable of leading a normal existence because, after several months or years of illness, their levodopa treatment is insufficient or they receive only low doses of antiparkinsonian adjuvants without levodopa itself. Patients with more advanced disease can remain immobile for part of the day. Sometimes this immobility results from taking only two or three doses of levodopa daily, despite the drug's short half-life, but most often the individual doses are too small. A slight increase in each dose or increased fractionation of the daily doses can dramatically reduce motor instability and transform the existence of these patients.

When asked why they avoid levodopa or limit its administration, the patients, their families, and the prescribing physicians give different answers. Some believe that levodopa causes the irreversible development of dyskinesias and motor swings that can be avoided or delayed by retarding treatment. Others believe that long-term treatment with levodopa can accelerate the death of already fragile dopaminergic neurons. This opinion gains credence from the observation that levodopa can be a source of potentially deleterious free radicals. Therefore, it is commonly held that because "there is sufficient concern that levodopa may be toxic and thereby enhance the progression of Parkinson's disease … it seems prudent to delay the use of levodopa until this treatment becomes necessary."¹

If levodopa causes irreversible deleterious effects that accelerate the degeneration of dopaminergic neurons, prescribing it with caution and as late as possible would be justifiable. However, if there is only the question of reversible side effects, then levodopa therapy can be safely prescribed and adjusted to the particular needs of each patient, similar to any other substitutive treatment (e.g., thyroid extracts in patients with hypothyroidism). In this brief review, we consider this issue.

Does levodopa damage catecholaminergic neurons? In vitro and in vivo studies. Tissue culture evidence. Levodopa can be toxic to dopaminergic and other neurons in vitro.^2-6 Most of these results have been obtained by exposing neuroblastoma PC12, pheochromocytoma cells, or mesencephalic dopaminergic neurons to levodopa at concentrations ranging from 100 to 250 µM for 1 to 5 days. Although the mechanisms underlying these toxic effects are not fully understood, reactive oxygen radicals, hydrogen peroxide, and quinones generated by auto-oxidation of levodopa and its decarboxylated metabolite dopamine may be involved. Protection from the adverse effects of levodopa by antioxidants such as ascorbic acid or sulfhydryl compounds supports this theory.^3,5 When low concentrations of levodopa(50 µM) are used, however, trophic effects are unexpectedly observed and include an increase in the number and branching of dopamine processes.⁷ This indicates that the concentration of levodopa in vitro is a key factor.

Most of the in vitro experiments showing toxicity of levodopa were performed using neuronal cultures containing reduced numbers of glial cells, or in the absence of glia. Astrocytes take up levodopa and dopamine and possess the metabolic machinery required to inactivate both compounds.⁸ Levodopa is converted into dopamine by glial dopa-decarboxylase and further metabolized by monoamine oxidase (MAO) and catechol-O-methyltransferase (COMT). Hydrogen peroxide generated by catabolism via MAO-B can be efficiently detoxified by the enzymes catalase and glutathione peroxidase, which are present in relatively high concentrations in glial cells. Protection from the toxic effects of hydrogen peroxide by astrocytes has been observed in cultured striatal neurons, and it has been calculated that one astrocyte has the capacity to protect 20 neurons against the toxicity induced by the application of 100 µM hydrogen peroxide.⁹ The ability of astrocytes to participate in the enzymatic catabolism of levodopa and dopamine is only one aspect of the complex interactions that exist between glia and neurons. Glial cells are known to provide a nutritive and protective environment for neurons. For instance, glia-conditioned medium increases the survival of dopaminergic neurons (cultured in the absence of glia) and protects them from the toxic effect of high concentrations of levodopa (200µM).¹⁰ This finding suggests that glial cells release soluble factors able to promote survival of dopaminergic neurons and render them resistant to oxidative stress. Therefore, results obtained in vitro appear to be highly dependent on the conditions of culture used. Neurons cultured in the absence of glia lack the enzymatic protection and trophic support that these cells provide and are consequently highly susceptible to oxidative and other types of stress. Any modifications that render the condition of culture more similar to that present in vivo strengthen the capacity of cells to resist stress.

Recent studies have suggested an aspect of levodopa not previously appreciated: its neuroprotective and neurotrophic capacity. Exposure to levodopa of mixed mesencephalic neurons and glial cultures increases the cell concentrations of reduced glutathione (the substrate of glutathione peroxidase, an enzyme that detoxifies the radical superoxide) and protects dopaminergic neurons from the toxic effect of strong oxidants.¹¹ Not only can exposure to levodopa trigger a cellular response that results in enhanced resistance to oxidative stress, but the drug may have a neurotrophic effect on its own. In fact, addition of a low concentration of levodopa (50 µM) to cultures of rat fetal midbrain neurons and cortical astroglia increases survival and promotes neurite extension of dopaminergic neurons, an effect that is not dependent on dopamine synthesis because it is not blocked by the decarboxylase inhibitor carbidopa.⁷

In summary, high concentrations of levodopa can increase oxidative stress and produce cytotoxic effects on pure dopaminergic neuron cultures; levodopa, however, may also have protective effects, either at low concentrations (i.e., in the range of those measured in the plasma of patients with Parkinson's disease [5 to 50 µM])¹² or in the presence of glial cells, an in vitro experimental situation which reproduces more closely the in vivo conditions where dopaminergic neurons are embedded in a matrix of astrocytes.

Human and animal studies. Human studies have shown that administration of levodopa reduces the death rate in patients with Parkinson's disease.^13-15 This observation cannot reasonably be used as an argument to support a neuroprotective effect of levodopa, however, because this action is more likely to be due to the reduction in patients' disability. Long-term administration of levodopa to normal individuals fails to cause substantia nigra damage, as shown in pathologically confirmed cases.^16,17 Yet it may be argued that normal individuals have adequate protective mechanisms against oxidative stress. The neuropathologic study by Yahr et al.¹⁸ is therefore of major interest, because no histopathologic difference was found between the substantia nigra of patients with Parkinson's disease who had never been treated with levodopa and those who had. That study would have been pivotal in showing the absence of toxicity of levodopa in vivo if the investigators had been able to perform a precise cell count of melanized neurons in the substantia nigra. Nevertheless, to date, no study has shown that long-term administration of levodopa can destroy dopaminergic neurons in humans.

Several studies have shown the absence of toxicity of levodopa in experiments performed in vivo in rodents. No reduction in the number of dopaminergic cells was observed in the substantia nigra of unlesioned rats and mice treated with high doses of levodopa for up to 18 months.^19-21 Similar experiments have been performed in rodents with partial lesions of the dopaminergic nigrostriatal system (a model that mimics parkinsonism in humans). A modest reduction in the number of dopaminergic neurons was observed in the ventral tegmental area of rats previously lesioned with 6-hydroxydopamine and subsequently given long-term levodopa treatment.²² This finding, however, was not replicated in another recent study that showed no reduction in the number of dopaminergic neurons in the substantia nigra and ventral tegmental area of rats with moderate or severe lesions of the nigrostriatal dopaminergic system given long-term levodopa treatment (6 months).²³ Furthermore, before any conclusions drawn from the following two studies^24,25 can be used to support the claim that levodopa destroys dopaminergic cells in vivo, certain experimental factors should be taken into account. Although local damage was observed after direct injection of dopamine into the rat striatum in the first study,²⁴ the range of dopamine doses injected was high in a model that is far from reproducing physiologic conditions. Moreover, although cell survival and neuritic growth of fetal mesencephalic tissue were shown to be poor in grafted rats treated with levodopa,²⁵ these changes were reversible after prolonged withdrawal of treatment. In fact, no study has yet shown that levodopa can accelerate the degeneration of dopaminergic cells in vivo. Indeed, the contrary might be possible. The fact that oral treatment with levodopa for 6 months increased the density of dopaminergic fibers in the striatum of rats with partial lesions of the nigrostriatal pathway²⁶ suggests that levodopa could have beneficial effects in patients with Parkinson's disease.

In conclusion, the absence of levodopa toxicity reported in rodents and humans with or without dopaminergic loss is striking and suggests that the deleterious effects seen in in vitro studies in the absence of glia are not directly transferable to the in vivo situation.

Does levodopa result in irreversible dysfunction of the basal ganglia in patients with Parkinson's disease? The administration of levodopa, an amino-acid precursor of dopamine, reduces parkinsonian symptoms by allowing normal dopaminergic transmission to be re-established, making it an effective substitutive symptomatic treatment. After several years of levodopa treatment, however, parkinsonian disability in most patients increases despite adequate levodopa treatment, and patients start to experience both motor fluctuations and dyskinesia. The question is whether these new symptoms, which were not manifest at the onset of the disease in untreated patients, result from an irreversible neuronal dysfunction coupled with the natural worsening of brain lesions or from a transitory pharmacologic effect of levodopa on nerve cells.

Does long-term levodopa treatment contribute to the decreased response to treatment of parkinsonian symptoms? In most patients with parkinsonism, the response to levodopa decreases with time, and treatment becomes progressively less effective in reducing parkinsonian disability. Even in later stages of the disease, however, symptoms such as akinesia, rigidity, and tremor are still ameliorated by long-term levodopa treatment.²⁷ This improvement of the classic triad of parkinsonism results from the selective re-establishment of normal dopamine transmission in the striatum, essentially sparing the output systems of the basal ganglia. In contrast, the decreased response of motor disability to levodopa is accompanied by the progressive appearance of axial symptoms and signs such as dysarthria, gait disorders, postural instability, and cognitive disorders. These symptoms show little or no response to levodopa because they are caused by additional, nondopaminergic brain lesions that are not located downstream from nigrostriatal neurons; otherwise, they would antagonize the effects of levodopa. They are also distinct from the dopaminergic lesions(loss of noradrenergic, serotoninergic, subcorticocortical neurons; neurons containing peptides such as cholecystochinin-8, methionine [MET]- and leucine [LEU]-enkephalin; and cortical neurons showing Alzheimer-like neuropathologic changes).²⁸

Levodopa's ability to cause the permanent dysfunction of nondopaminergic neuronal systems seems unlikely. If levodopa does not aggravate the loss of vulnerable dopaminergic neurons, as discussed, why should it affect nonsusceptible neurons? Therefore, although this has yet to be proven, we conclude that the decreasing response to levodopa of parkinsonian symptoms during the course of the disease results not from neuronal loss provoked by the drug, but rather from the disease-related degeneration of additional nondopaminergic neuronal systems.

Does long-term levodopa administration result in irreversible levodopa-induced adverse reactions? As the duration of both disease and treatment increases, motor side effects become more frequent, severe, and complex. At early stages of the disease, motor performance fluctuations take the form of end-of-dose deterioration characterized by the loss of clinical benefit within hours of levodopa administration. Smooth fluctuations shift with time to "on/off" phenomena, which tend to become chaotic and not always obviously related to levodopa intake.²⁹ Levodopa-induced dyskinesia can appear as monophasic (peak-dose) dyskinesias, which occur during the period of maximal relief of parkinsonian symptoms, or biphasic(onset- and end-of-dose) dyskinesias, observed when parkinsonian disability is decreasing and increasing.³⁰

Why does levodopa, which should simply reverse the parkinsonian syndrome, lead to dyskinesias and sometimes unbearable "on/off" effects even when the most suitably adapted dose is used? The pathophysiology of levodopa-related motor complications is poorly understood. Such symptoms are not observed in untreated patients with Parkinson's disease, nor are they seen in normal individuals (mistakenly diagnosed as having Parkinson's disease) who received levodopa for long periods.^14,15 These motor complications are therefore due to the conjunction of two factors: the central dopaminergic lesions and long-term administration of levodopa. The severity of striatal dopaminergic denervation plays an essential role in the appearance of levodopa-related complications. Thus, the frequency and severity of levodopa-induced dyskinesias and motor fluctuations are correlated with the severity of the parkinsonian syndrome and the predominance of dopamine-dependent symptoms, such as akinesia and rigidity.²⁸ By analogy with experimental studies in animals,³¹ hyperkinesias triggered by levodopa may appear when the level of dopaminergic denervation in the striatum has reached 90%, which is the case in the dorsal part of the striatum in patients with Parkinson's disease.³²

Although we cannot prevent worsening of the neurodegenerative process, we can adapt the manner of levodopa prescription. To this end, certain factors must be considered. First, the higher the amount of levodopa absorbed at each dose, the more severe the side effects, so each individual dose during the day should preferably be reduced to a minimum. Second, the more intermittent the administration of levodopa, the earlier and more severe the motor side effects,³³ so the treatment should be given as continuously as possible to limit fluctuations in the levels of levodopa in the blood.³⁴ This can be achieved by increased fractionation of the daily dose, by using controlled-release preparations of levodopa, or by adding adjuvants, such as dopaminergic D2 agonists or COMT inhibitors to prolong the effect of the medication. Finally, adverse motor reactions are most frequent in patients who have been treated with levodopa earlier during the course of the disease,³⁵ an observation that has led many experts to delay prescribing levodopa to their patients. However, evidence shows that levodopa-induced motor complications appear at approximately the same time and with the same degree of severity, whether the medication is prescribed early or late in the course of the disease.^36,37 This suggests that the increased dyskinesia and response swings are most likely the result of more rapid deterioration due to the disease itself, because levodopa was given earlier to patients who needed it sooner. It is common practice to avoid early prescription of levodopa in patients for whom side effects will be more unpleasant than the discomfort caused by the disease (particularly in severe levodopa-responsive forms of the disease) and to avoid delaying prescription of levodopa treatment in patients who will benefit from a reduction in discomfort. Nonetheless, given that the treated parkinsonian patient presents a different clinical picture from that of the untreated patients, presumably the brain of a patient with Parkinson's disease who has received long-term treatment with levodopa is not identical to that of an untreated patient. Whether this is the case, the crux of the matter is the question of reversibility: if levodopa is withdrawn after long-term treatment with the drug, will the parkinsonian brain return to its "natural" state?

To answer this question, we must first consider the cellular and molecular changes that occur in the treated parkinsonian patient. When treatment is initiated, assuming that the dopaminergic lesions are still only moderate, enough dopamine remains at the level of the presynaptic dopaminergic terminals in the striatum to buffer fluctuations in dopamine levels by capturing the mediator when the concentration is too high and releasing it when the level is too low. During the period of end-of-dose deterioration, the buffering capacity of striatal dopaminergic nerve endings declines. The motor response becomes sinusoidal, with extracellular dopamine concentrations changing in parallel with fluctuations in blood levodopa levels. At this stage, the therapeutic fluctuations of levodopa can be largely explained by the pharmacokinetic factors of medication absorption.³⁸ At the "on/off" phenomena stage, pharmacodynamic factors gain the upper hand over pharmacokinetic factors, particularly at the postsynaptic level downstream from dopaminergic neurons.^39,40 The progressive degeneration of dopaminergic neurons of the nigrostriatal pathway is accompanied by a moderate hypersensitivity of dopaminergic receptors D1 and D2²⁸ and a dysfunction of the efferent pathways in the basal ganglia (in particular a hyperactivity of the internal pallidum),^41,42 which are normalized by the administration of levodopa.^28,42 The motor responses induced by the levodopa (dyskinesias, "on/off" effects) therefore imply a behavioral sensitization phenomenon, a process by which repeated stimulation of neurotransmitter receptors in the brain results in a progressive enhancement of responsiveness.

Several factors are involved in causing this sensitization, including changes in the activity of the residual dopaminergic neurons⁴³ and changes in postsynaptic receptor responsiveness. Sensitization does not involve a desensitization of the dopaminergic receptors, because central dopamine receptors remain available for stimulation if sufficient levodopa can be delivered to the brain.⁴⁴ Rather, sensitization is due to selective dopaminergic D2 receptor priming,⁴⁵ and perhaps even more to the reversible appearance of dopaminergic D3 receptors in the denervated striatum (a brain area from which this receptor subtype is normally absent) after repeated intermittent administration of levodopa.⁴⁶ The latter observation is probably decisive, even if striatal D3 receptor protein and messenger RNA were found to be normal or significantly elevated in a few patients with Parkinson's disease who were given levodopa.⁴⁷ Long-term treatment with levodopa is also reflected in physiologic changes in all basal ganglia downstream from the dopaminergic denervation. Prolonged dopaminergic denervation leads to loss in the specificity of electrophysiologic responses,⁴⁸ and even to changes in neuronal plasticity, as reported at the level of the substantia nigra and striatum in patients with Parkinson's disease.⁴⁹ If postsynaptic changes at the receptor level are indeed reversible after longterm levodopa treatment, no data yet indicate that this might also be the case for changes in plasticity. The most we can say is that, from clinical experience, motor complications induced by levodopa are reduced or abolished when the treatment is interrupted for several days (such drug holidays are no longer performed), which in some cases allows the levodopa treatment to be resumed at a lower daily dose and thus limits the severity of motor complications.⁵⁰ Far more conclusive is the observation of a few patients with Parkinson's disease in whom levodopa treatment has been withdrawn or markedly reduced for several months because the parkinsonian disability has been abolished by continuous high-frequency bilateral stimulation of the subthalamic nuclei. In these patients, resumption of levodopa treatment for a short period after the suspension of stimulation has shown that levodopa-induced dyskinesias are not immediately observed or else remain attenuated (Bejjani et al., unpublished data, 1998). Nevertheless, these results must be confirmed in more patients.

In conclusion, all of the experimental and clinical data available to date suggest that the molecular and cellular changes brought about by the long-term administration of levodopa in patients with Parkinson's disease are reversible.

Acknowledgments

The author thanks Drs. E. Borroni, P.P. Michel, and M. Ruberg for their help in the preparation of this manuscript.