OUP user menu

Huntington's disease progression
PET and clinical observations

Thomasin C. Andrews, Robert A. Weeks, Nora Turjanski, Roger N. Gunn, Laura H. A. Watkins, Barbara Sahakian, John R. Hodges, Anne E. Rosser, Nicholas W. Wood, David J. Brooks
DOI: http://dx.doi.org/10.1093/brain/122.12.2353 2353-2363 First published online: 1 December 1999

Summary

Using serial [11C]SCH 23390- and [11C]raclopride-PET, we have measured the rate of loss of striatal dopamine D1 and D2 receptor binding over a mean of 40 months in nine asymptomatic adult Huntington's disease mutation carriers, four patients with symptomatic disease, seven mutation-negative controls and three subjects at risk for the disease. Eight of the nine asymptomatic Huntington's disease mutation carriers had serial [11C]raclopride-PET and showed a mean annual loss of striatal D2 binding of 4.0%. Only five of these eight, however, showed active progression, and they had a mean annual loss of D2 binding of 6.5%. All nine asymptomatic mutation carriers had serial [11C]SCH 23390-PET and showed a mean annual loss of striatal D1 binding of 2.0%. Four of these subjects demonstrated active progression and they had a mean annual loss of 4.5%. Our four symptomatic Huntington's disease patients showed a mean annual loss of D2 binding of 3.0% and of D1 binding of 5.0%. Loss of striatal D1 and D2 binding was significantly greater in the known mutation carriers than in the combined at-risk and gene-negative groups (P < 0.05). At follow-up PET all subjects were clinically assessed using the Unified Huntington's Disease Rating Scale. Scores for motor function and total functional capacity correlated with PET measures of striatal dopamine receptor binding both in the asymptomatic mutation carriers (D1, P < 0.01) and across the combined asymptomatic and clinically affected Huntington's disease mutation carrier group (D1 and D2, P < 0.001). We conclude that PET measures of striatal D1 and D2 dopamine binding can be used to identify asymptomatic Huntington's disease mutation carriers who are actively progressing and who would thus be suitable for putative neuroprotective therapies. Measures of disease progression rates in Huntington's disease patients and asymptomatic mutation carriers will be of critical importance in future trials of experimental restorative treatments.

  • Huntington's disease
  • dopamine receptors
  • PET
  • disease progression
  • UHDRS
  • FDG = [18F]2-fluoro-2-deoxyglucose
  • TFC = total functional capacity
  • UHDRS = Unified Huntington's Disease Rating Scale

Introduction

Huntington's disease is an autosomal dominant neurodegenerative disorder. The causative gene, a CAG repeat expansion in the IT15 gene on chromosome 4, was identified in 1993, enabling accurate predictive diagnostic testing for healthy, at-risk family members (Huntington's Disease Collaborative Research Group, 1993). The disease can present at any time from childhood to old age but has a peak incidence in mid-adult life. Clinically there is progressive cognitive dysfunction, psychiatric disturbance and movement disorder with chorea, dystonia, bradykinesia and rigidity. Death usually occurs 15–20 years after the onset of symptoms. In the Westphal variant, children and young adults present with rigidity and bradykinesia without ever demonstrating chorea (Marshall and Shoulson, 1997).

The pathological process is found to target striatal medium spiny neurons selectively (Graveland et al., 1985). These GABAergic striatopallidal projection neurons make up 90% of all striatal neurons and express dopamine D1 and D2 receptors. One of the striking histopathological features of Huntington's disease is that adjacent large aspiny cholinergic interneurons are preserved late into the disease (Ferrante et al., 1987).

Autoradiographic studies have shown that striatal dopamine receptor binding falls proportionally to the degree of cell loss in early Huntington's disease (Augood et al., 1997). Using PET and the selective dopamine D1 and D2 receptor antagonists [11C]SCH 23390 and [11C]raclopride, striatal dopamine receptor binding potentials can be measured in vivo. At tracer doses, binding potential is proportional to the ratio Bmax/KD, where Bmax is the total concentration of specific receptor binding sites and KD is the dissociation constant of the radioligand at equilibrium (Mintun et al., 1984). Previous PET studies using these ligands have demonstrated that all patients with Huntington's disease and a proportion of asymptomatic mutation carriers show significant and parallel reductions of striatal D1 and D2 receptor binding (Leenders et al., 1986; Hagglund et al., 1987; Brandt et al., 1990; Sedvall et al., 1994; Turjanski et al., 1995; Weeks et al., 1996; Ginovart et al., 1997). Patients with clinical disease show >40% loss of striatal dopamine binding (mean loss 60%), whereas asymptomatic mutation carriers show a loss of up to 50% of striatal dopamine binding (Turjanski et al., 1995; Antonini et al., 1996; Weeks et al., 1996). Reductions in dopamine D1 and D2 binding potentials are thought to represent loss of functioning receptors on intact but dysfunctioning striatal neurons as well as loss of receptors secondary to the eventual death of these striatal neurons.

Two previously reported studies have used PET to measure disease progression rates in asymptomatic Huntington's disease mutation carriers. Grafton and colleagues performed serial [18F]2-fluoro-2-deoxyglucose (FDG)-PET scans on eight asymptomatic mutation carriers and found a mean annual reduction in glucose metabolism of 3.1% per year for the caudate and 1.9% per year for the putamen (Grafton et al., 1992). FDG-PET measurements of striatal glucose metabolism reflect glial as well as neuronal activity and are influenced by cortical inputs and inflammatory processes. They are thus a less specific measure of intrinsic striatal neuronal function in Huntington's disease than PET measurements of dopamine receptor binding. Antonini and colleagues reported results of serial [11C]raclopride and FDG-PET findings for six asymptomatic mutation carriers and found a mean annual change of 6.3% in striatal D2 binding and of 2.3% in striatal glucose metabolism (Antonini et al., 1996); this suggests that FDG-PET may also be a less sensitive marker of disease progression than [11C]raclopride-PET. Hussey and colleagues have reported (in abstract) mean annual progression rates of 7.0 and 3.7% for clinically affected and asymptomatic Huntington's disease mutation carriers, respectively, using [11C]raclopride-PET (Hussey et al., 1998).

The Unified Huntington's Disease Rating Scale (UHDRS) was developed recently by the Huntington Study Group for repeated administration during clinical research studies. It gives a semi-objective score for motor performance, cognitive performance, behavioural abnormalities and functional capacity (Huntington Study Group, 1996). A longitudinal study in 78 patients showed significant decline in the motor score at 1 year and in total functional capacity at 2 years of follow-up (Siesling et al., 1998).

The aims of this study were (i) to compare the value of serial PET measures of striatal dopamine D1 and D2 receptor binding for the objective monitoring of disease progression in Huntington's disease mutation carriers (asymptomatic mutation carriers and clinically affected patients), and (ii) to determine whether levels of striatal D1 and D2 binding in Huntington's disease correlate with clinical severity of disease as assessed using the UHDRS.

Subjects and methods

Subjects

The study was partly prospective and partly retrospective in design. Twenty healthy subjects at risk for Huntington's disease who had attended neurogenetics clinics requesting predictive testing in 1994 were recruited into the study. All were symptom-free, with no signs of Huntington's disease on clinical examination (by R.A.W.). Fifteen of the 20 subjects had baseline [11C]SCH 23390- and [11C]raclopride-PET studies prior to their genetic tests and the remaining five were scanned subsequently. DNA analysis for the Huntington's disease mutation was performed in line with the recommendations of the UK Huntington's Disease Predictive Testing Consortium (Simpson and Harding, 1993; Davis et al., 1994). Of this cohort, nine were found to have the mutation, six were mutation-negative and five chose not to have the test and remained at risk (Weeks et al., 1996). Eight of the asymptomatic mutation carriers returned for repeat scanning, as did five mutation-negative and three at-risk subjects.

A second cohort of clinically affected patients had had [11C]SCH 23390- and [11C]raclopride-PET in 1992 as part of a separate study (Turjanski et al., 1995). Of the 10 patients studied, four agreed to return for repeat scans (two with akinetic rigid variant disease and two choreic). Of the others, two had died, two refused and two were lost to follow-up. One asymptomatic mutation carrier and two mutation-negative relatives who were scanned with this group agreed to have repeat scans.

Thus, in total we scanned a cohort of nine asymptomatic Huntington's disease mutation carriers, four clinically affected patients, three at-risk subjects and seven mutation-negative subjects on two occasions (Table 1). Ethical permission was restricted to repeating only one of the two PET scans in those subjects found to be mutation-negative.

View this table:
Table 1

Demographic summary of the four subject groups and of the groups combined

Status at baselineNumberSex (M : F)Age at follow-up in years (range)Scan interval in months (range)
HD = Huntington's disease.
Asymptomatic mutation carriers 9 4 : 539 (24–55)36 (28–67)
Clinically affected HD patients 4 1 : 347 (33–71)48 (37–57)
At-risk subjects 3 2 : 138 (29–45)36 (31–45)
Mutation-negative subjects 7 4 : 347 (37–56)42 (33–55)
Total2311 : 1243 (24–71)39 (28–67)

At the time of the repeat PET scans, 21 of the subjects had volumetric MRI and all subjects were formally assessed with the UHDRS (by T.C.A.). Clinical assessments were carried out blind to the genetic status of the subject and blind to PET scan findings. PET scans were analysed blind to genetic status and blind to whether scans were baseline or follow-up.

One of the affected subjects had been receiving tetrabenazine medication for 2 weeks at the time of his second raclopride scan and one of the asymptomatic mutation carriers was taking the serotonin reuptake inhibitor, fluoxetine, at the time of her repeat scan. These medications were withdrawn for the 12 h prior to PET. None of the other 21 subjects was receiving neuroleptics, serotonin reuptake inhibitors or dopamine-blocking drugs at the time of either study.

PET findings were not made available to the subjects. The study received ethical approval from the Ethics Committee of the Royal Postgraduate Medical School, Hammersmith Hospital. Permission to administer [11C]raclopride and [11C]SCH 23390 was obtained from the Administration of Radioactive Substances Advisory Committee of the UK (ARSAC). All subjects gave informed written consent in accordance with the declaration of Helsinki.

Scan procedure

T1-weighted volumetric MRI scans [repetition time (TR) 21 ms, echo time (TE) 6 ms] were acquired on a 1.0 Tesla Picker HPQ scanner. The [11C]raclopride-PET scans were all performed on a CTI 931/-08/12 scanner. Once reconstructed, the spatial resolution for 15 planes of image data was 7.0 mm axially and 8.5 × 8.5 mm transaxially (full width half maximum) (Spinks et al., 1988). The 16 subjects from the predictive genetic testing cohort also had [11C]SCH 23390 scans performed on this scanner. The [11C]SCH 23390 scans for clinically affected patients, one asymptomatic mutation carrier and two mutation-negative subjects were performed with a CTI/Siemens 953B scanner. The spatial resolution of this scanner for 31 planes of reconstructed image data in two-dimensional mode was also 8.5 × 8.5 mm transaxially but 3.5 mm axially (full width half maximum) (Spinks et al., 1992). All follow-up scans were performed with the same camera as the paired baseline scan.

Patients were aligned in the orbitomeatal plane. An inflatable head mould was used to restrict head movement. A 10 min transmission scan was obtained using a retractable external source of 68Ga/68Ge to correct for attenuation of gamma-radiation by the brain and skull. A background time frame of 30 s preceded the intravenous bolus injection of the ligand. The dynamic scan consisted of 22 serial time frames collected over 1 h (Lammertsma et al., 1996).

Data analysis

Image analysis was performed using Analyze software (version 7.5, BRU, Mayo Foundation, Rochester, Minn., USA) (Robb and Hanson, 1991) on Sun Sparc Ultra workstations. An integrated image (ADD) of the data from the last four time frames (20–60 min) was created for coregistration purposes. Each individual's MRI was coregistered separately to baseline and follow-up ADD images using coregistration software (MPR; Guy's Hospital, London, UK) (Studholme et al., 1997). The temporal order of the two PET scans and coregistered MRIs was unknown to the investigator analysing the images. The paired coregistered MRIs and corresponding ADD images were visualized and equivalent striatal and cerebellar planes were identified. Regions of interest were then traced around the ventral striatum, caudate and putamen on the MRI for all planes where these structures were clearly defined. Two circular regions of interest (32.8 mm in diameter) were placed on two or three equivalent planes of cerebellum. Regional time-activity curves were then obtained by applying regions of interest to the corresponding dynamic image series. Binding potentials for each region were calculated using a reference tissue model with a cerebellar input function (Lammertsma and Hume, 1996).

The total volume of each region of interest was recorded. As the same MRI, coregistered to PET, was used to define regions of interest for both baseline and repeat studies, the volumes of these regions of interest were unchanged. If there had been significant atrophy of striatal structures over the time course of the study, the true volumes of the caudate and putamen would be smaller at the time of the repeat PET scans. If the MRI estimate of the volume of a given structure, e.g. the caudate, is larger than the actual volume, the measured binding potential will be spuriously low because the PET signal from the white matter and CSF will be included in the region of interest (partial volume effect). However, as the MRI used for delineating the regions of interest was acquired at the time of the second scan, any partial volume effects should have been minimized.

In two subjects who did not have MRIs because of claustrophobia, regions of interest were traced directly on ADD images using equivalent planes of data. One of the nine asymptomatic mutation carriers did not have paired [11C]raclopride scans because of a failed transmission scan. One of the four clinically affected Huntington's disease patients did not feel well enough to come back for his repeat [11C]SCH 23390-PET study. This individual had the lowest striatal [11C]raclopride binding potential and was the most disabled. In two of the mutation-negative subjects the baseline PET had inadequately visualized the cerebellum and so could not be used.

Results

Rates of loss of striatal dopamine D1 and D2 binding potential

Figure 1 shows PET binding potential maps (Gunn et al., 1997) of [11C]SCH 23390 and [11C]raclopride uptake in two striatal planes for an individual asymptomatic Huntington's disease mutation carrier at baseline and follow-up 3 years later. Progressive loss of both striatal D1 and D2 binding is evident in this subject, especially in the more dorsal planes. Figure 2 shows the change in striatal D1 and D2 binding for each of the individual Huntington's disease mutation carriers over the time period of the study.

Fig. 1

Baseline and repeat [11C]SCH 23390 (D1) and [11C]raclopride (D2) PET binding potential maps from an asymptomatic mutation carrier. Scans were performed at an interval of 33 months. Loss of both D1 and D2 binding is demonstrated; this appears to be greatest in the more dorsal planes.

Fig. 2

Striatal [11C]SCH 23390 (D1) and [11C]raclopride (D2) binding potentials (BP) for each of the mutation carriers at baseline and repeat scans. Time interval between scans is shown in months on the x-axis. Asymptomatic mutation carriers are shown as square symbols and are numbered from 1 to 9 corresponding to subjects 1–9 in the text. Affected patients are shown as circles.

In the Huntington's disease asymptomatic mutation carrier group the mean annual change in striatal dopamine binding was –4.0% for D2 (n = 8) and –2.0% for D1 sites (n = 9). Striatal binding was calculated by averaging the binding potential from the caudate and putamen regions. In the clinically affected patient group the mean annual change in striatal binding was –3.0% for D2 (n = 4) and –5.0% for D1 binding (n = 3). The mean annual change for both D1 and D2 striatal binding in the at-risk group (n = 3) was less than –0.5% and in the mutation-negative group (D1, n = 2; D2, n = 3) it was +2.0%. Loss of striatal D1 and D2 receptor binding in the combined asymptomatic and clinically affected Huntington's disease mutation carrier group was significantly greater than that in the combined mutation-negative and at-risk groups (independent samples t test: D2, P = 0.001; D1, P < 0.05). (The mutation-negative and at-risk groups were combined in order to increase the number of degrees of freedom. The at-risk group may have included subjects who were mutation carriers; however, this would tend to decrease rather than increase the difference between the means of the mutation-negative/`at risk' and mutation carrier groups.) Loss of striatal D2 receptor binding in asymptomatic mutation carriers alone was also significantly greater than that in the combined mutation-negative and at-risk group (P < 0.01), but was not significantly greater for D1 binding (P = 0.09). However, there was one more mutation-negative subject in the D2 group and the variance in the at-risk group for D2 was less than for D1 binding. Table 2 shows the mean and the range for percentage annual change in D1 and D2 binding for each subject group.

View this table:
Table 2

Mean annual percentage changes in D1 and D2 binding potential

Mean (range) annual change in binding potential
StriatumCaudatePutamen
HD = Huntington's disease.
(A) D1 binding potential
Asymptomatic mutation carriers (n = 9)–2.1% (+1.0 to –8.1)–3.7% (+0.3 to –13.3)–1.2% (+1.8 to –2.8)
Active loss (total >5%) in asymptomatic mutation carriers (n = 4)–4.4% (–2.6 to –8.1)–6.7% (–1.7 to –13.4)–3.4% (+0.1 to –8.0)
Clinically affected HD patients (n = 3)–4.8% (–2.5 to –7.9)–6.2% (–1.4 to –13.1)–4.0% (–2.7 to –5.4)
At-risk subjects (n = 3)–0.5% (+1.6 to –3.2)–1.2% (+0.7 to –4.2)+0.1% (+2.3 to –2.4)
Mutation-negative subjects (n = 2)+2.0% (+3.2 to +0.9)+1.8% (+4.1 to –0.5)+2.6% (+3.2 to +2.0)
(B) D2 binding potential
Asymptomatic mutation carriers (n = 8)–4.1% (+1.0 to –8.3)–4.5% (–0.2 to –10.2)–3.8% (+1.4 to –7.7)
Active loss (total >5%) in asymptomatic mutation carriers (n = 5)–6.4% (–3.3 to –8.1)–7.1% (–3.5 to –10.2)–6.1% (–2.9 to –7.7)
Clinically affected HD patients (n = 4)–3.2% (–1.8 to –4.5)–6.7% (–0.3 to –9.4)–2.0% (–0.5 to –2.7)
At-risk subjects (n = 3)–0.1% (+2.2 to –0.4)–0.7% (+2.4 to –1.1)+0.2% (+4.6 to –3.3)
Mutation-negative subjects (n = 3)+1.9% (+3.3 to –0.1)+1.9% (+3.6 to –0.5)+1.5% (+3.9 to +0.5)

Previous PET studies have shown that striatal D2 receptor binding is lost at a rate of ~0.5% per year with normal ageing, although the rate of loss is far greater in younger subjects (i.e. subjects aged <30 years) (Antonini and Leenders, 1993), and that maximum scan–rescan variability is 10% with a mean variability of 4% (Volkow et al., 1993). The mutation-negative group in our study showed a total change in striatal dopamine binding ranging from –0.5 to +10.5% over the period of the study, with a mean annual change of +2.0% for both D1 and D2 sites. A total individual loss of binding in the mutation carrier group of >5% was taken to indicate active disease progression.

Five of the eight asymptomatic mutation carriers who had paired [11C]raclopride scans showed active loss of striatal D2 binding ranging from 9.0 to 19.5% over the 3–5 years of the study (mean annual loss 6.5%), while the remaining three showed <3.0% total change. Four of the nine asymptomatic mutation carriers showed loss of striatal D1 binding ranging from 8.5 to 22.5% (mean annual loss 4.5%) while the remaining five showed a total change in D1 binding of <3.0%. Two subjects with changes in binding potential of <10% are included in the actively progressing groups. One showed an 8.5% loss of D1 sites and a 19.5% loss of D2 and the other showed a 9.0% loss of D2 sites and a 12.0% loss of D1 sites. Thus, three asymptomatic mutation carriers showed active loss of both striatal D1 and D2 binding (subjects 3, 6 and 9 in Fig. 2), two showed active loss of D2 binding alone (subjects 2 and 5), while one showed active loss of D1 binding alone (subject 4). The remaining three showed <3% total change in striatal dopamine receptor binding over the study period (subjects 1, 7 and 8). All of the clinically affected patient group showed a loss of striatal D1 and D2 binding over 3–5 years that exceeded 7.5% (range 7.5–37.5%).

Caudate and putamen D1 and D2 binding were correlated at baseline and follow-up in the combined mutation carrier groups (Spearman rank correlation, r2 > 0.62, P < 0.01) but mean annual percentage changes in D1 and D2 binding in individual subjects were not correlated.

Loss of caudate dopamine binding was greater than that seen for the putamen in both the combined mutation carrier group and the asymptomatic group alone for D1 sites (Wilcoxon matched pairs test, P < 0.05) but did not reach significance for D2 sites for either group (P = 0.07; P = 0.21).

CAG repeat lengths were available for eight of the nine asymptomatic mutation carriers. The mean CAG repeat length was 43.8 (range 41–47). Despite this small range, CAG repeat length correlated with mean annual loss of striatal D2 binding potential (one-tailed Spearman rank correlation coefficient, r2 = 0.61, P < 0.05, n = 7). The patient group had a mean CAG repeat length of 47.5 (43–53). Across the combined mutation carrier group, CAG repeat length and mean annual change in striatal D1 and D2 binding potential were also correlated (P < 0.05; D1, r2 =0.34; D2, r2 = 0.47, n = 11) (Fig. 3). Mean annual loss of striatal D1 and D2 binding was multiplied by age of subject at the time of the baseline scan to estimate the total lifetime loss of binding, assuming a linear rate of loss. CAG repeat length correlated with the product of mean annual loss of D2 binding and age in asymptomatic mutation carriers (r2 = 0.50, P < 0.05, n = 7) and with the product of D1 binding and age (r2 = 0.39, P < 0.05, n = 11) but not with the product of D2 binding and age (r2 = 0.13) in the combined mutation carrier group.

Fig. 3

Linear correlations between CAG repeat number and percentage loss in striatal dopamine D2 binding potential (A) and D1 binding potential (B) for asymptomatic mutation carriers (squares) and clinically affected patients (circles).

Correlations with the UHDRS

At the time of follow-up PET, subjects were formally assessed using the UHDRS. This assessment was performed blind to genetic status and PET results. At the time of the second scan, five of the nine asymptomatic mutation carriers displayed some motor signs on examination, as scored on the UHDRS (subjects 4, 8, 3, 6 and 9 in order of increasing motor score). Motor signs included difficulty performing the Luria, mildly slowed or irregular alternating finger-taps and pronation/supination, jerky ocular pursuits and increased saccade latency, difficulty keeping the tongue fully protruded for 10 s, one or two deviations on tandem walking, increased tone with activation and increased restlessness or slight intermittent chorea.

Only one subject had developed clinically definite Huntington's disease on examination and considered herself to be symptomatic (subject 9), and one subject considered herself to have minor symptoms of disease but clinically was only `possibly affected' (subject 8). The other seven mutation carriers considered themselves to be asymptomatic, although two had developed reduced capacity for their normal work (subjects 6 and 7).

Table 3 shows the means and ranges for the motor and total functional capacity (TFC) scores of the UHDRS for the Huntington's disease patient and asymptomatic mutation carrier groups. None of the mutation-negative group had scores greater than 1 for motor function (possible scores 0–124) or <13 (possible scores 0–13) for TFC. One of the at-risk group had a motor score of 8.

View this table:
Table 3

UHDRS motor and total functional capacity scores at follow-up

Motor scoreTotal functional capacity
MeanRange (0–124)MeanRange (0–13)
HD = Huntington's disease.
Asymptomatic mutation carriers (n = 9) 5.3 0–14128–13
Clinically affected HD patients (n = 4)49.828–73 6.84–9

In the combined asymptomatic and clinically affected Huntington's disease mutation carrier group, UHDRS TFC and motor scores showed significant correlations with measures of both caudate and putamen dopamine D2 (Fig. 4A and B) and D1 binding at follow-up (Spearman rank correlation, P < 0.001). In the asymptomatic mutation carrier group alone, TFC and motor scores also showed significant correlations with measures of caudate and putamen D1 binding (P < 0.001) and TFC correlated with D2 binding (P < 0.05). The UHDRS was available only at the time of follow-up, so correlations between changes in clinical scores over time with this scale and loss of dopamine receptor binding potential could not be made.

Fig. 4

Putamen [11C]raclopride (D2) binding potentials (BP) at the follow-up scan for the combined Huntington's disease mutation carrier group plotted against UHDRS motor score (A) and total functional capacity (B). Mutation carriers who were asymptomatic with no signs of the disease at baseline are shown as squares and clinically affected patients are shown as circles. Asymptomatic mutation carriers are identified by subject number.

No significant correlations were found between age and TFC or motor score, or between age and striatal D1 and D2 binding in either the asymptomatic mutation carrier group or the combined mutation carrier groups.

Discussion

Current experimental therapeutic interventions with foetal striatal cell implants and neuroprotective agents aim to slow disease progression in Huntington's disease and restore basal ganglion function. In the future, with a better understanding of the pathogenesis of Huntington's disease, it may be possible to delay or prevent the onset of the neurodegenerative process. In order to ascertain the efficacy of a treatment, whether aimed at preclinical or symptomatic disease, it will be necessary to have an objective measure of disease progression and a better understanding of the time course of the disease.

This study has demonstrated that both [11C]SCH 23390- and [11C]raclopride-PET can be used to identify asymptomatic adult Huntington's disease mutation carriers in whom accelerated striatal degeneration is occurring, and can also be used to measure objectively the rate of dopamine binding loss, and thus of disease progression. Our asymptomatic Huntington's disease mutation carriers appeared to fall into two groups, a minority showing little change in striatal dopamine binding (total <3.0%) between baseline and repeat scans while the majority showed active loss ranging from –8.5 to –22.0%, with a mean annual change of –6.5% for D2 and –4.5% for D1 binding. Antonini and colleagues reported a 6.3% annual loss of striatal D2 binding in a cohort of asymptomatic Huntington's disease mutation carriers (Antonini et al., 1996). As five of their six subjects were actively progressing, we found a rate of loss of dopamine binding equivalent to that reported by these workers.

It is not known whether the disease is slowly progressive from early childhood with age of presentation determined by the rate of neuronal degeneration and the functional reserve of the striatum, or whether neurodegeneration starts later in life with progression to symptomatic disease over just a few years. Two of our asymptomatic mutation carriers with D1 and D2 binding within the normal range and a third with borderline uptake showed no significant progression over the 4 year time period of this PET study. Antonini and colleagues found that one of six asymptomatic mutation carriers showed no measurable change in PET measurements of raclopride binding or glucose metabolism (Antonini et al., 1996). These data suggest that loss of striatal function starts within the decade before symptom onset.

The length of the CAG repeat has been shown to correlate with age at presentation and the extent of neuropathology at death (Duyao et al., 1993; Furtado et al., 1996), but the influence of other genetic and environmental factors on age of onset and rate of disease progression is not known. It has been calculated that up to 70% of the variance in the age of presentation can be accounted for by the CAG repeat size (Rubinsztein et al., 1997). It would seem likely that the observed dose effect between CAG repeat length and age of onset relates to both the age at which neuronal degeneration starts and the rate of progression of the disease. In our study CAG repeat length correlated significantly with the mean annual rate of loss of dopamine D2 binding in the asymptomatic mutation carrier group and with the mean annual rate of loss of both D1 and D2 binding in the combined mutation carrier group. We also examined the relationship between CAG repeat length and the product of mean annual loss of binding and subject age. This analysis assumed a linear rate of loss over the lifetime of the subject. Significant correlations were again found for D2 in the asymptomatic mutation carrier group (although r2 was less than in the direct comparison) and for D1 in the combined mutation carrier group. However, no correlation was found for D2 in the combined group, suggesting that, for D2 at least, receptor loss does not follow a linear time course. Antonini and colleagues, in a cross-sectional study, found a correlation between CAG repeat length and the calculated percentage loss of striatal D2 binding below a normal mean divided by subject age in 10 asymptomatic mutation carriers and eight symptomatic patients (Antonini et al., 1996). These data suggest that CAG repeat length does influence the rate of disease progression. In the study by Antonini and colleagues, the slopes of the correlations for asymptomatic mutation carriers and patients were significantly different, suggesting that the neurodegenerative process is in fact non-linear.

The asymptomatic mutation carrier with the lowest striatal binding (D1 and D2 60% of normal mean) developed definite signs and symptoms of disease over a 3 year follow-up period, striatal D1 and D2 binding potentials falling to 0.8 (55%) and 1.1 (50%) respectively. In the three symptomatic patients with a clinical disease duration of <5 years, the linearly extrapolated striatal binding potential threshold for symptom onset was 0.7–0.8 for D1 and 0.9–1.0 for D2, i.e. a loss of 45–55% of striatal dopamine binding relative to the normal mean.

Our clinically affected Huntington's disease patients showed a mean annual change in striatal binding that was 3.2% for D2 and 5.0% for D1 sites. Hussey and colleagues, in an abstract, have reported a mean annual 7% loss of D2 binding for seven affected Huntington's disease patients (Hussey et al., 1998). The number in our cohort (four) is too small to draw any firm conclusions about the relative rates of progression in preclinical and clinical disease or about the differences between rates of D1 and D2 loss in akinetic rigid and choreic disease. It is possible that selection bias affected our data, as those patients with slower rates of disease progression may have selectively agreed to return for repeat scans. Two of our patients died in the interval between the baseline and follow-up PET; both of them had had symptoms for 1 year at the time of their first scans. This suggests both patients must have progressed rapidly. One other patient required admission to a nursing home during the study period and one refused to return for follow-up. Three of the four Huntington's disease patients who agreed to be scanned twice had had clinical disease for between 5 and 17 years at baseline PET, again suggesting that they may have had slower rates of disease progression than average.

Antonini and colleagues suggest that the rate of disease progression, as measured by [11C]raclopride-PET, slows later in the course of the disease (Antonini et al., 1998). A fall in disease progression rate with duration has also been reported in [18F]dopa-PET studies of Parkinson's disease (Morrish et al., 1998). A recent post-mortem autoradiographic study of Huntington's disease brains found that the density of D1 receptors per surviving striatal neuron was markedly reduced in Vonsattel grade 1 disease and then appeared to increase with increasing pathology (Augood et al., 1997). This finding could account for the apparent slowing of disease and was interpreted as reflecting ligand binding to a subset of striatal neurons not targeted by the disease—possibly cholinergic interneurons.

Models of basal ganglion connectivity have traditionally divided GABAergic striatal efferents into a direct pathway which projects to the internal segment of the globus pallidus and substantia nigra pars reticulata and an indirect pathway which projects to the external segment of the globus pallidus, the subthalamic nucleus and then to the globus pallidus (Albin et al., 1989). However, this is an oversimplified model; the external segment of the globus pallidus and substantia nigra pars reticulata are not simply relay stations in the indirect pathway but have widespread connections with other basal ganglion and subcortical nuclei (Chesselet and Delfs, 1996). In the striatum, dopamine D1 receptors are mostly expressed by the striatonigral and striato- pallidal substance P/dynorphin-containing neurons of the direct pathway, while D2 receptors are mostly expressed by the striatopallidal enkephalin-containing neurons of the indirect pathway (Ferre et al., 1997). Surmeier and colleagues, however, have challenged this view, showing that both D1 and D2 mRNA can be detected in individual striatal neurons (Surmeier et al., 1993). The Penney–Young model of Huntington's disease predicts that chorea results from early selective loss of the enkephalin (D2)-bearing neurons of the indirect pathway (Penney and Young, 1986; Albin et al., 1989). Post-mortem autoradiographic studies, however, have demonstrated parallel striatal D1 and D2 receptor loss in early disease, though with greater loss of D1 receptors from the striatum, globus pallidus and substantia nigra pars reticulata and of D2 receptors from the external segment of the globus pallidus (Richfield et al., 1991). In this study we found that binding potentials were striatal D1 and D2 tightly correlated in both baseline and repeat scans in our asymptomatic mutation carriers and in the combined asymptomatic/affected mutation carrier group.

Previous cross-sectional PET studies (Turjanski et al., 1995; Weeks et al., 1996) have also shown that there is a tight correlation between striatal D1 and D2 binding at all stages of the disease in vivo in both choreic and rigid phenotypes, arguing against the differential loss of one or other type of receptor. However, there was no correlation between annual percentage changes in striatal D1 and D2 binding in our cohort, so it is possible that, although still tightly correlated when compared across the group, subtler imbalances within an individual's D1/D2 direct/indirect neuronal function may contribute to the different symptoms of the disease.

Loss of caudate dopamine binding was greater than that seen for putamen binding for D1 but not for D2 sites. This finding probably reflects the fact that the pathological process preferentially affects the caudate relative to the putamen early in the disease. Partial volume effects, however, would also have a greater influence on the caudate than on the putamen. The distribution of D1 and D2 receptors within the caudate and putamen is controversial. Joyce and colleagues reported that D1 receptors were preferentially expressed by human striatal striosomes and D2 receptors by the matrix (Joyce et al., 1988). Rappaport and colleagues, however, reported a more complex distribution in the monkey striatum; D1 receptors were expressed preferentially in caudate striosomes but distributed homogeneously throughout the putamen while D2 receptors were expressed homogeneously in both caudate and putamen (Rappaport et al., 1993). The Rappaport model suggests that loss of D1 sites would occur from the caudate at a greater rate than from the putamen in asymptomatic mutation carriers, but that there would be a more even distribution of D2 receptor loss.

No striatal atrophy was evident in the asymptomatic Huntington's disease mutation carriers on MRI. However, formal partial volume correction of serial PET data was not possible in this study as subjects had volumetric MRIs only at follow-up. Previous studies using volumetric MRI have demonstrated mild striatal atrophy in some asymptomatic Huntington's disease mutation carriers (Aylward et al., 1994). However, Grafton and colleagues showed that the rate of change of caudate size on MRI did not correlate with changes in caudate metabolism in asymptomatic mutation carriers (Grafton et al., 1992). It is likely that partial volume effects do contribute to the measured loss of striatal dopamine binding in our cohort of asymptomatic and affected mutation carriers. However, as we are assessing the usefulness of PET as a technique for objectively measuring disease progression, the extent to which measured changes are due to volume effects as opposed to loss of function within remaining neurons is semantic.

Follow-up measures of caudate and putamen D1 and D2 binding were found to correlate significantly with UHDRS motor and total functional capacity scores both in the combined asymptomatic and clinically affected mutation carrier group and in the asymptomatic mutation carrier group alone. This suggests that the UHDRS and PET measures of dopamine receptor binding are both capable of detecting early disease in asymptomatic subjects. It does not necessarily follow, however, that abnormalities in the UHDRS arise as a direct consequence of striatal degeneration. Additionally, two of our asymptomatic Huntington's disease mutation carriers who showed accelerated loss of striatal dopamine binding had no soft signs, and so disease progression could not have been identified by the UHDRS in these subjects. Currently, it is unclear whether serial UHDRS scores provide as sensitive a measure as PET of disease progression in Huntington's disease asymptomatic and affected mutation carriers. PET, however, provides an independent biological measure of disease progression and is thus a valuable adjunct to clinical assessments.

Conclusion

In this study we have shown that [11C]raclopride and [11C]SCH 23390-PET can both be used to identify a subset of asymptomatic Huntington's disease mutation carriers who are actively progressing towards clinical disease, and to measure objectively the rate of loss of striatal function in preclinical and clinical disease. Due to the small number of symptomatic patients in this study we were unable to determine the relative sensitivity of these two measures for detecting disease progression in late-stage Huntington's disease, but the data suggest that in preclinical disease striatal D2 measures are more sensitive. It is intriguing that reductions in D1 and D2 binding in individual Huntington's disease mutation carriers did not correlate despite a close correlation between measures of absolute D1 and D2 binding at baseline and follow-up, suggesting that differential loss and dysfunction of the neurons that express these two receptor subtypes does occur and may underlie some of the different symptoms of Huntington's disease.

PET dopamine receptor binding studies have been used in vivo to demonstrate graft survival in a rat model of Huntington's disease (Torres et al., 1995). A large prospective [11C]raclopride-PET and clinical study is needed to determine the true variability in the rate of disease progression between individuals and within individuals over the time course of Huntington's disease. In the future, PET and the UHDRS will be useful for monitoring the effects of novel treatment interventions in Huntington's disease, such as foetal striatal cell implants and neuroprotective agents.

Acknowledgments

We wish to thank Drs M. Baraitser, E. Rosser and R. Barker for allowing us to study their patients, Ralph Myers and Adrian Lammertsma for help with data analysis, the chemistry section of the MRC Cyclotron Unit, Andrew Blythe and the PET radiographers, Leonard Schnorr for data reconstruction, Professor Graeme Bydder and the MR radiographers from the Robert Stein MRI unit, Dr Mary Davis and Mary Sweeney for DNA analysis, and all the participants in this research. T.C.A. was supported by grants from the MRC and the Huntington's Disease Association.

References

View Abstract