Introduction
The functional capabilities of positron emission tomography (PET) and single photon emission tomography (SPECT) are among the most useful and revealing techniques to image the brain in Parkinson’s disease (PD). Both offer a unique opportunity to study the brain at work, with only minimal invasiveness. They involve the intravenous administration of a small dose of a compound or a drug, labeled with a short half-life isotope. The drug can be a selective marker of a variety of biological processes, for example, glucose metabolism, blood flow, or dopaminergic or serotoninergic function. The behavior of the chosen drug can be monitored in vivo by measuring, with specialized scanners and detectors, the distribution of the radioactive tracer over time.

In PD, multiple aspects of brain function can be studied to satisfy both clinicians and basic researchers, with not only the potential to provide insights in the pathophysiology of PD and its underlying neurochemical processes, but also to improve differential diagnoses, study the complications of treatment, and explore the mechanisms of drug action.

More importantly, it allows longitudinal follow-up of disease progression and the exploration of the compensatory neurochemical changes associated with long-term treatment or with surgical interventions. While there have been a number of radiotracers developed for SPECT, this article focuses on the use of PET tracers and only points out where similar studies have been performed using SPECT.

Tracers of the dopaminergic presynaptic system
Although a PET scanner is not needed to diagnose PD, the presence of a deficit in presynaptic activity, its pattern, and the conservation of dopamine receptor function can assist in the diagnosis. Currently, a deficiency in striatal dopamine can be detected by one of three methods. As dopamine does not cross the blood/ brain barrier, an analog of its precursor, 6-[18F]-fluoro-L-dopa (FDOPA), is routinely used. First reported in 1983,[1] FDOPA uptake and storage (as fluoro-dopamine) in the terminals of nigro-striatal dopamine neurons, provide an in vivo index of the integrity of the dopamine terminals.

That index, the uptake rate constant, obtained during short (90-minute) clinical scans, is a valid measure of dopaminergic integrity and correlates well with the number of surviving dopamine neurons in human subjects.[2,3] During longer scans (180 to 240 minutes), a more complex analysis method allows the estimation of an effective rate of dopamine turnover. This index of dopamine turnover is increased in animals with drug-induced parkinsonism and correlates with classical postmortem indices.[4]

FDOPA may help in the differential diagnosis between idiopathic PD and PD-like syndromes. FDOPA has demonstrated reduction of dopamine activity in PD patients, an often asymmetric decline and, in keeping with postmortem findings, with a rostro-caudal gradient with the caudate (the rostral part of the striatum) less affected and the more caudal putamen most severely affected Figure 1 [9].[5] Most often, Parkinson-plus syndromes such as multiple system atrophy and progressive supranuclear palsy show an equal reduction in FDOPA uptake in the caudate and putamen.[6]

Although there is an inverse correlation between FDOPA uptake and disease duration and severity, the symptoms of PD do not develop until there is an 80% reduction in striatal dopamine concentrations or a 50% loss of nigral dopamine neurons. FDOPA studies of idiopathic PD suggest that the threshold for development of symptoms lies around 55% of normal FDOPA values. FDOPA allows early detection of PD, even in asymptomatic patients; preclinical deficits have been demonstrated in neurologically normal patients exposed to 1-methyl-4-phenyl-1, 2,3,6 tetrahydropyridine (MPTP, a neurotoxin that selectively destroys the nigral dopamine neurons)[7] and in unaffected identical twins of patients with PD.[8]

Disease progression has been studied in the same subject populations over an extended period of time. The FDOPA striatal rate of decline is higher in the MPTP-exposed patients and the idiopathic PD patients as compared to normal subjects[8,9] and appears to be about 7% to 9% in PD patients, with a lower rate of decline in the caudate compared to the putamen.[10]

FDOPA has also been used in an attempt to unravel the mechanisms of motor fluctuations in response to treatment and dyskinesias. If fluctuations reflect a loss of buffering capacity due to loss of the dopamine terminals, patients with fluctuations should have a lower FDOPA uptake than patients with a stable response to treatment. Both our group and that of Leenders[5] have found evidence for reduced FDOPA in fluctuators but with considerable overlap, suggesting that additional factors must play a role.

FDOPA is increasingly used to monitor, over months and years, the effects of treatments such as transplanted dopaminergic-producing tissues. The use of PET has demonstrated a substantial increase in FDOPA in the putamen of patients following human fetal nigral grafts Figure 2 [10].[11] Another potential use is to monitor the effects of neuroprotective therapies or surgery such as deep brain stimulation of the subthalamic nucleus.

Another interesting aspect of dopamine transmission is the dopamine transporter. This protein lies on the membrane of the dopamine terminals and is an important factor in the inactivation of dopamine. Several cocaine analogues have been labeled to study the dopamine transporter. At UBC, we use [11C]-d-threo methylphenidate (MP), a non-cocaine analog with high selectivity, to measure binding potential to the dopamine transporter.

The dopamine transporter tracers have all yielded similar findings in PD, mainly a pattern and a rate of progression similar to that described above for FDOPA. Nevertheless, dopamine transporter tracers appear to be more sensitive than FDOPA in detecting changes in patients with very early, unilateral PD, and provide better early discrimination from normal controls.[12]

Another tracer, [11C]-dihydrotetrabenazine (DTBZ), which labels the central vesicular monoamine transporter, appears to be independent of this type of internal regulation and therefore may afford a more direct estimate of dopamine nerve terminal density. As the central vesicular monoamine transporter is present in all monoaminergic neurons, DTBZ is not a truly selective tracer. However, accumulation of DTBZ in the striatum mostly reflects the presence of dopamine terminals, which represent over 90% of striatal monoaminergic terminals.

Frey and colleagues[13] have demonstrated reduced DTBZ binding in PD, with asymmetric striatal activity and a rostrocaudal gradient similar to that seen with FDOPA and dopamine transporter ligands. Using DTBZ as a gold standard of the number of dopamine nigral neurons in a cohort of patients with idiopathic PD, we were able to confirm and demonstrate in vivo, by PET, the existence of compensatory mechanisms in the striatal dopamine pathway. When compared to DTBZ binding, the binding of MP to the dopamine transporter was decreased and the uptake of FDOPA increased to a greater extent than would be predicted, demonstrating downregulation of the dopamine transporter and upregulation of dopa decarboxilase (the enzyme that metabolizes FDOPA into [18F]fluoro-dopamine) in early PD patients Figure 3 [11].[14]

Tracers of the dopaminergic receptors
Several tracers of the dopamine receptors are currently available. The binding of several ligands of the dopamine D1- and D2-like receptors has been extensively studied, because of the possible involvement of dopamine receptors in the genesis and/or maintenance of motor complications associated with levodopa therapy. However, apart from a transient increase in the density of dopamine D2 receptors in the putamen of early, untreated PD,[15] most in vivo PET studies (as in postmortem studies) have not revealed definite differences in the dopamine receptors in patients with stable response, motor fluctuations, or with dyskinesias.[16]

Our own studies support this finding and agree with extensive research in animal models of PD, which suggest that most treatment complications stem from or reflect alterations downstream to dopaminergic receptors. Some Parkinson-plus syndromes such as progressive supranuclear palsy and multiple system atrophy are associated with a loss of striatal dopamine D2 receptors.

Non-dopaminergic tracers
Other transmitter systems can be studied with PET/SPECT that can potentially provide information on the physiology of complications associated with treatment. A reduction in the binding of [11C]diprenorphine, a non-selective tracer of the opiate receptors, has recently been reported in the striatum of patients with dyskinesias, while non-dyskinetic patients were the same as normal subjects.[17] Alternatively, a decrease in [11C]-cyclofoxy, a tracer of the opiate receptors, was reported in monkeys with long-term MPTP-induced parkinsonism who had never been treated with levodopa and had no dyskinesias.[18] Although detailed interpretation of these findings is not possible, they would be in keeping with the upregulation of dynorphin reported in postmortem studies of animal models of PD.

Although the studies of specific neurotransmitters may be the best way in which to uncover the pathophysiology of PD and the mechanisms of drug actions, studies of glucose metabolism can also yield useful information. In keeping with current hypotheses of basal ganglia function, fluorodeoxyglucose studies demonstrated an increase in glucose metabolism in the pallidum of PD patients.[19] More recently, the pattern of glucose metabolism is being investigated as a means to understand the consequences of PD to brain activity as a whole and of therapeutic strategies such as pallidotomy and deep brain stimulation.[20]

The future
Although the majority of PET and SPECT tracers used to study PD provide mainly descriptive information, there is increasing emphasis on the development of tracers aimed at studying disease pathophysiology and pharmacological mechanisms in vivo. For example, recent studies in parkinsonian patients using [11C] PK11195, a compound that binds to activated microglia, revealed increased binding throughout the basal ganglia,[21] suggesting the presence of an ongoing inflammatory process (in keeping with postmortem findings). While these results have not been duplicated at our centre, it remains an important area to pursue.

Other advances in PET are the increasing use of tracers as surrogate markers of physiological responses. Changes in raclopride (a relatively specific but low affinity tracer of dopamine D2 receptor) binding in response to changes in competition from endogenous dopamine (for instance, after administration of levodopa or amphetamine) are now being exploited to estimate changes in dopamine release. Methods like the Scatchard analyses, which were until now restricted to the laboratory, are being adapted for in vivo studies to yield the density and affinity of receptors.[15]

In parallel, significant improvement in technology in the form of higher sensitivity and resolution tomographs will soon allow the exploration of the various functions, processes, and mechanisms now restricted to the striatum in small extrastriatal nuclei such as the substantia nigra or the subthalamus.

Glossary [12]

Acknowledgments
The authors wish to acknowledge the support of the UBC/TRIUMF PET team, the Neurodegenerative Disorder Centre, the Medical Research Council of Canada, the Pacific Parkinson’s Research Institute and TRIUMF, and the National Institutes of Health.

Dr Doudet is an associate professor in the Department of Medicine, Division of Neurology at the University of British Columbia. Thomas Ruth is the director of the UBC/TRIUMF PET Program.