Biochemical aspects of Parkinson's disease

The present era of biochemical research into Parkinson's disease (PD) started about 40 years ago. Before that time, only conjectures could be made about the biochemical abnormalities that underlie the clinical features of this brain disorder. The anticholinergic properties of the empirically used atropine-like antiparkinson remedies gave rise to the idea that PD was connected with a disturbance of cholinergic brain mechanisms.^1,2 In 1960, the dopamine (DA) deficit in the basal ganglia of patients dying with PD was discovered.³ The steps that led to this discovery have been described elsewhere.^4,5 This observation was the starting point for a series of biochemical investigations in human postmortem brain that provided a firm basis for understanding of the biochemical pathology, neurotransmitter pathophysiology, and rational pharmacotherapy of PD with levodopa.

Unlike much other progress in medicine, the DA era of PD was not an outgrowth of the accumulated knowledge gained in prior studies of the biochemistry of the human brain. Studies of chemically unstable neurotransmitters and enzymes in autopsied human brain had been rare, having long been regarded with great suspicion. The success of the DA substitution therapy with levodopa, which was without precedent in the history of research into chronic, degenerative, progressive human brain diseases, has definitely established the biochemical study of postmortem human brain as an important branch of present-day neuroscience research. The field of research in PD that has most recently sprung up, dealing with the biochemical causes of the DA neuron degeneration, is a logical outgrowth of the original work on DA in PD.

This article reviews the most relevant biochemical observations made in studies of the autopsied PD brain, discusses open questions, and considers some possibilities for future research.

The basic biochemical changes in PD brain. Among the neurotransmitter abnormalities found in the autopsied brains of patients with PD, the loss of DA in the striatal nuclei (i.e., caudate nucleus and putamen) stands out as the single most prominent biochemical change, closely connected with the motor disorder of PD.

In PD the entire nigrostriatal DA neuron is affected. Fully developed clinical PD is regularly accompanied by a severe DA deficit in the striatum.³ Also reduced are all other biochemical markers for the presynaptic striatal DA terminals, such as the levels of the major DA metabolite homovanillic acid (HVA),⁶ the DA synthetic enzymes tyrosine hydroxylase,^7,8 and L-dopa-decarboxylase,^7,8 and the DA transporter sites^9,10 (table 1).

In the advanced stages of PD, the striatal DA loss exceeds the 80% mark, with the DA levels in the putamen being consistently more reduced than in the caudate nucleus.^3,6,11-14 This putamen-caudate difference is due to the uneven pattern of loss of the melanin-containing DA perikarya in the compact zone of the substantia nigra. The nigral cell loss is more pronounced in the ventral cell groups that project to the putamen than in the dorsal cell groups that innervate the caudate nucleus.^15,16 Although the degree of the striatal DA loss correlates significantly with the degree of nigral cell loss,¹¹ latter is, on average, distinctly less than would be expected from the degree of the striatal DA loss¹¹(table 2).

The two specific patterns of striatal DA loss. In the idiopathic variety of PD, the loss of the striatal DA follows two regular, characteristic inter- and sub-regional patterns. First, in every patients,^17,18 the putamen loses considerably more DA than the caudate nucleus.^3,6,11-14 Second, within the putamen the caudal portions are more depleted of DA than the rostral portions, whereas in the caudate nucleus this rostrocaudal gradient goes in the opposite direction^12-14 (table 2).

The clinical threshold of the striatal DA loss. The striatal DA changes remain clinically silent until the threshold value of 60-80% DA loss is reached.¹¹ Within the symptomatic range of DA loss(<60% reduction), a correlation is seen¹¹ between the degree of DA loss and the severity of PD symptoms (table 2). In every clinical case of PD,¹⁸ the DA loss in the putamen, but not in the caudate nucleus, exceeds the critical threshold value of 60-80%. As a motor disorder, PD is essentially a disorder of the putamen("motor loop"^19,20) function.

Compensatory biochemical changes in PD striatum. The high threshold for the striatal DA loss to induce clinical manifestations of PD shows that lower degrees of DA loss are functionally compensated for by the remaining DA neurons.²¹ The mechanisms for the DAergic compensation are twofold (table 2). First, the remaining DA neurons increase their DA metabolism, i.e., DA release and synthesis, as evidenced by the shifting of the striatal HVA:DA ratio in favor of the metabolite (HVA).^6,21 This is an early compensatory mechanism, already operating during losses of less than 60% DA.²¹ Second, in the advanced, decompensated stages of DA loss (<90% loss), the number of the postsynaptic D₂ DA receptor sites increases in the PD striatum.²² This maximizes the therapeutic efficacy of the DA substitution treatment in the clinically overt stage of PD.

Levodopa as a DA substitution treatment for PD. The severe lack of DA in the PD striatum constitutes a firm biochemical basis for the use of the DA precursor levodopa as a rational and thus far most efficacious drug treatment for PD.^23-25 The clinical use of levodopa represents a DA substitution therapy. In the striatum of levodopa-treated patients, the mean concentrations of DA were (1) 9-15-fold higher than those in non-dopa-treated patients, were (2) related to the time before death of the last dose of levodopa, and were (3) greater in the striatum of patients clinically classified as "good responders" compared with"poor responders."⁷

Open questions. The severe striatal DA deficit is, in principle, sufficient to explain the main motor features of PD and the anti-PD efficacy of DA substitution with levodopa. However, the many other striatal and extrastriatal neurotransmitter changes found in PD brain raise questions that should be answered if we wish to obtain a greater understanding of what this disorder of basal ganglia function is as a brain disease in its own right.

In PD the striatal GABA levels are above normal. In PD striatum, the concentrations of GABA are elevated.^26,27 The most marked and statistically significant GABA changes were measured in the putamen (+16-72%), especially in those (caudal) subdivisions in which the DA loss was greatest. Milder GABA elevations (0 to+26%) were seen in the generally less DA-depleted caudate nucleus.²⁷ [GABA levels were also above normal (+30%) in the external and internal segments of the globus pallidus. This change and moderate elevations of glutamate in putamen (+28-42%) and caudate nucleus(+11-25%) were statistically not significant.] In the most affected caudal putamen, the negative correlation between the (increased) GABA levels and the(reduced) DA levels was statistically significant.²⁷

The GABAergic medium spiny (projection) neurons of the striatum are the primary target of the nigrostriatal DA neurons.²⁸ In PD, loss of the DAergic innervation appears to be directly responsible for the increased GABA levels, as evidenced by the significant negative correlation between the striatal DA loss and the GABA elevation. The functional significance for PD of this striatal GABA response to DAergic denervation remains to be determined.

Is extrastriatal DA involved in PD? In contrast to the obligatory striatal DA loss, the reduction of DA levels in other subcortical(and cortical) brain regions is less prominent and is not always present(e.g., hypothalamus²⁹). There is nevertheless no reason to doubt that, if present, such DA changes may be significantly involved with the overall PD disability and/or with levodopa's therapeutic or side effects.

The basal ganglia output nuclei. In addition to the input stations of the basal ganglia, i.e., the DArich caudate nucleus and putamen, all basal ganglia output nuclei, including the lateral and medial globus pallidus and the reticular zone of the substantia nigra, as well as the subthalamic nucleus, contain noticeable amounts of DA ^3,12,30,31 and DA receptor sites.^32-35 Therefore, DA downstream of the striatum is also likely to control the activity flow through the basal ganglia. This is supported by a recent electrophysiologic study indicating that DA attenuates the action of GABA in rat globus pallidus.³⁶

In PD, the DA levels in all output nuclei, especially the lateral globus pallidus, are greatly reduced and the ratio of HVA to DA is shifted, as in the striatum, in favor of HVA (table 3). The extrastriatal DA deficit most likely contributes to the motor disorder of PD.³⁷ Reversal of this DA deficit with, e.g., levodopa can be assumed to add to the overall antiparkinson efficacy of the DA substitution treatment. Local application of DAergic agents into the rat globus pallidus has been shown to elicit various motor behaviors.^38-40

These observations raise three important questions. First, are the levodopa-induced dyskinesias due to a DAergic overstimulation of the output nuclei rather than the striatum? If so, which output nucleus may be preferentially involved in this levodopa side effect? Finally, is the persistent beneficial effect of postero-medial pallidotomy perhaps due to direct surgical elimination of the morphologic substrates for the levodopa-induced dyskinesias?

The nucleus accumbens. In PD, a loss of neuronal perikarya less marked than in the substantia nigra is also seen in the ventral tegmental area (VTA) of the mesencephalon,⁴¹ the site of origin of much of the DAergic innervation of the nucleus accumbens (and the limbic cortex).⁴² As a consequence, in PD the concentrations of DA and HVA in the nucleus accumbens (and in several cortical areas) are significantly reduced. As in the striatum, in the nucleus accumbens the ratio of HVA to DA is shifted in favor of HVA.⁴³ On average, however, the DA loss in nucleus accumbens is distinctly less severe than in the striatum (table 3).

The accumbens DA is assumed to support both motor and limbic functions, especially those that are reward-related.⁴⁴ The DA loss in the nucleus accumbens in PD is therefore likely to play a part in the biochemical pathology of parkinsonian motor deficits, especially brady- and akinesia, as well as the frequently observed depressed mood and reduced ability to experience reward.

Two intriguing questions remain to be answered. First, is the degree of the accumbens DA reduction- about 50 to 60%-sufficient to produce overt clinical deficits? Second, to what extent are the various postsynaptic accumbens DA receptors, including the recently characterized D₃ receptors,⁴⁵ involved in levodopa's therapeutic and/or side effects, especially dysinesias? Measuring DA in accumbens "core" and"shell," the two functionally distinct accumbens subdivisions recently identified,⁴⁶ might help to answer these questions.

Is PD caused by a specific mechanism? Evidence from secondary parkinsonism. In addition to PD, many other degenerative brain disorders presenting with parkinsonian symptoms are accompanied by loss of striatal DA. These neurodegenerative conditions include postencephalitic parkinsonism,^3,11 progressive supranuclear palsy,⁴⁷ striatonigral degeneration (unpublished results), corticobasal ganglionic degeneration,⁴⁸ neuronal intranuclear inclusion body disorder,⁴⁹ dementia-parkinsonism-motor neuron disease,⁵⁰ Creutzfeldt-Jakob disease,⁵¹ Pick's disease,⁵² Rett syndrome,⁵³ and Hallervorden-Spatz disease.⁵⁴

Compared with PD, however, none of the secondary parkinsonian conditions thus far examined had inter- and subregional patterns of caudate and putamen DA loss similar to the DA patterns typical of PD (table 4). Does this difference mean that the striatal DA loss, as typically seen in PD, in addition to being sufficient to produce the cardinal motor signs of PD is also specific for the idiopathic condition? Moreover, does this difference imply that the loss of DA neurons in PD is caused by a single specific agent or mechanism different from the (possibly multifactorial) mechanisms involved in the other (secondary) parkinsonian conditions?

The non-DA changes in PD: The similarity to the changes in the MPTP primate. Narabayashi⁵⁵ has recently suggested that PD begins as a pure nigrostriatal DA disorder but that in the later course of the disease the pathologic process spreads to other neurotransmitter systems. Is there any biochemical evidence to support this possibility?

As judged from postmortem biochemical analyses at death many patients with PD exhibit changes in several indices of non-DA brain neurotransmitters, including noradrenaline,^{3,12,29,56,57} serotonin,^12,29,58 GABA,²⁷ glutamate,²⁷ met-enkephalin,⁵⁹ preproenkephalin,⁶⁰ cholecystokinin-8,⁵⁹ neurotensin,⁵⁹ substance P,⁵⁹ and others (figure 1). Although many of these changes, especially those involving above-normal levels(e.g., GABA; see above), are caused by the disturbed interaction with the severely affected nigrostriatal DA system,^21,61,62 some have a morphologic basis (e.g., cell loss in nucleus basalis,⁶³ locus ceruleus,¹⁵ raphe nuclei⁶⁴).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Some non-dopamine neurotransmitter/neuromodulator changes (shaded columns) in the brain of Parkinson's disease patients: comparison with the dopamine loss in the nigrostriatal neuron system (black column). ACC = nucleus accumbens; CCK-8 = cholecystokinin-8; CTX = cortex, DA = dopamine; GABA =γ-aminobutyric acid; Glu = glutamate; 5-HT = serotonin; Met-enk = met-enkephalin; NA = noradrenaline; NT = neurotensin; ppenk = preproenkephalin; PUT = putamen; SN = substantia nigra; SP = substance P; STR= striatum. (For references, see text.)

In contrast to the always present and severe loss of striatal DA, the losses in most non-DA systems, as exemplified by hypothalamic noradrenaline and serotonin levels,²⁹ vary among patients, ranging from no change to marked loss. This shows that the non-DA changes are not obligatory for PD; instead, they may be, as suggested by Narabayashi,⁵⁵ the result of the progressive spreading of the disease pathology, a process whose speed and extent may greatly differ among individual patients.

The spreading in PD of the degenerative process specifically to noradrenaline and serotonin neurons is reminiscent of the changes produced in primates by MPTP. Acute exposure to MPTP selectively affects the nigrostriatal DA system.⁶⁵ In contrast, repeated MPTP doses also affect the brain noradrenaline, and serotonin neurons.⁶⁶ Most significant, in PD as well as in the MPTP-treated primate, the DA, noradrenaline and serotonin changes show an identical regional preference. In both conditions the number of brain areas with significant DA loss was more prominent in the subcortex than in the cerebral cortex. In contrast, in both the MPTP primate and in PD, the most pronounced noradrenaline and serotonin changes occurred in cortical regions^66,67(figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Number of subcortical and cortical brain areas affected by significant dopamine (DA), noradrenaline (NA), and serotonin (5-HT) loss in the MPTP primate versus Parkinson's disease (PD), expressed as percent of the total number of subcortical (n = 12) and cortical areas (n = 12) analyzed. (For references, see text.)

This raises some crucial questions. Does the observation that MPTP mimicked the regional brain monoamine "disease susceptibility" to PD imply that, in principle, both conditions have a similar underlying disease process? Does this striking similarity therefore mean that a basically similar, MPTP-like biochemical mechanism is also involved in PD?

Is the degeneration of the nigrostriatal DA neurons a dying-back process? In PD, the degree of loss of striatal DA^11,56 and tyrosine hydroxylase activity⁷ exceeds the loss of nigral DA,³¹ tyrosine hydroxylase activity,⁷ (table 5) or cell bodies¹¹ (table 2). This mismatch suggests that in PD the degeneration of the nigrostriatal DA neurons may be a dying-back process that begins in the striatal terminals rather than in the nigral cell bodies. An analogous discrepancy between the striatal DA loss and nigral DA and cell loss had been found in some other striatal DA deficiency conditions, such as corticobasal ganglionic degeneration,⁴⁸ dementia-parkinsonism-motorneuron disorder,⁵⁰ Rett syndrome,⁵³ Lesch-Nyhan syndrome,⁶⁸ and in a group of patients with olivopontocerebellar atrophy.⁸ In sharp contrast, no such mismatch is seen in progressive supranuclear palsy.⁴⁷ In neuronal intranuclear inclusion body disorder, which primarily affects neuronal cell bodies, the DA loss in the substantia nigra was more pronounced than in the striatum.⁴⁹

Taken by itself, the striatal-nigral DA mismatch does not prove that in PD the loss of DA neurons starts in the striatum. However, the dying-back possibility is strongly supported by two crucial observations: In the MPTP-treated primate, there is a corresponding PD-like mismatch between the striatal and nigral DA loss^66,67(table 5). This correspondence is important because the mismatch in the MPTP primate is explained by the fact that the neurotoxin-induced condition does indeed start in the striatal DA terminals.⁶⁹ An analogous mechanism for the identical mismatch in PD therefore suggests itself. This conclusion is directly supported by a recent morphologic study of caudate biopsy specimens taken from a patient with PD. In this electron microscopic study, the presence of neurodystrophic changes indicative of a retrograde, dying-back degenerative process has been directly demonstrated.⁷⁰

The DA transporter and the cause of nigral cell death in PD: Synthesis of evidence and a testable hypothesis. In PD, the surviving melanin (DA)-containing neurons of the substantia nigra have been reported to have lower levels of DA transporter mRNA than nigral neurons of normal controls.⁷¹ One possible interpretation of this observation is that the level of DA transporter expression might determine the vulnerability of the nigral neurons to the parkinsonian insult. This is an intriguing possibility because, in the primate substantia nigra, the density pattern of the mRNA for the DA transporter was found to closely match the pattern of cell loss in PD. Neurons in the ventral tier cell groups, which in PD are most markedly affected, had markedly higher DA transporter density than neurons in the dorsal tier and the VTA cell groups.⁷²

DA transporter involvement may answer difficult questions about PD biochemistry. In addition to its physiologic role in the inactivation(reuptake) of synaptic DA, the DA transporter also allows DA-toxic compounds, such as MPP⁺ (formed from MPTP), to enter the DA neurons.⁷³ This raises a crucial question. Is a neurotoxic DA transporter substrate involved in the etiology of PD? The correspondence between the patterns of the DA transporter and the nigral cell loss appears to support this possibility. If this is the case, then several otherwise difficult to answer biochemical questions would find a simple answer. Both the specific biochemical pattern of striatal DA loss and the corresponding histologic pattern of nigral cell loss could, in fact, be attributed to a single intrinsic property of the DA neurons per se, i.e., the different DA transporter density within the affected (nigrostriatal) brain region. The pattern of cell damage produced by a DA-toxic agent passing through the "gate" of the DA transporter would of necessity reflect the(uneven) distribution pattern of the DA transporter. The possibility that the DA neuron death in PD is a dying-back degeneration would become quite realistic, considering that the DA transporter is particularly enriched on the striatal terminals rather than on substantia nigral cell bodies.⁷⁴ The spreading of the neurodegenerative process later in the course of PD would find a plausible explanation because higher(cumulative) levels of the toxic compound in question might be needed to enter and damage the extrastriatal DA neurons and the non-DA monoamine(noradrenaline, serotonin) neuron systems.

What could the toxic DA transporter substrate be chemically? The (endogenous) formation and/or accumulation of a DA-toxic compound in the PD brain might be due to a local or general, possibly genetic-based, metabolic disturbance. The compound in question could be either extracellular DA itself or one of its (oxidized) metabolites, or an isoquinoline orβ-carboline derivative, or any other metabolic product that is actively accumulated in the brain DA, (and to some degree also noradrenaline and serotonin) neurons. The toxicity of such a compound would crucially depend on two factors: its relative affinity for the DA transporter and other monoamine transporters, and the regional distribution of the respective transporter sites and their density per neuron. An additional factor might be an anomaly of the DA transporter itself. Recently, an association with PD of a polymorphism in the DA transporter gene has indeed been reported.⁷⁵

Oxidative stress, mitochondrial dysfunction, and the DA transporter. The hypothesis that in PD a the DA transporter substrate may cause death of the nigrostriatal neurons is of considerable practical value. It can without difficulty accomodate the great number of mechanisms proposed to lead to the "final common pathway" of oxidative stress and/or mitochondrial dysfunction of the DA neuron.^76,77 As best shown by the DA-toxic effect of MPTP, many of the biochemical changes observed in PD substantia nigra⁷⁶ can be produced by a DA transporter substrate (MPP⁺) with neurotoxic properties. In principle, this is a testable hypothesis. Fully effective DA transporter (reuptake) blocking compounds, alone or in combination with a neurotrophic factor, should favorably influence the development and course of PD.