вход по аккаунту


Central levodopa metabolism in Parkinson's disease after administration of stable isotopeЧlabeled levodopa.

код для вставкиСкачать
Central Levodopa Metabolism in Parkinson’s
Disease after Administration of Stable
Isotope-Labeled Levodopa
Raymon Durso, MD,* James E. E,vans,i Ephraim Josephs,* George K. Szabo, BS,* Barbara A. Evans, BS,?
Joseph S. Handler, MD,* Dana Jennings, MD,* and Thomas R. Browne, MD*
We report the use of a new stable isotope-labeled form of levodopa (LD) to examine in vivo central LD metabolism in
Parkinson’s disease (PD). Eight patients representing a wide spectrum of disease severity were administered 50 mg of
carbidopa orally followed in 1 hour by an intravenous bolus of 150 mg of stable isotope-labeled LD (rimg1’,2’,3’,4’,5’,6’-13C6).
Serial blood samples were taken every 30 to 60 minutes and a lumbar puncture was performed 6
hours after the infusion. The average percentage of labeled homovanillic acid (HVA) in lumbar cerebrospinal fluid (CSF)
was 54% (SD, 9%; range, 34-67%). The mean CSF labeled HVA concentration was 34.7 ng/ml (SD, 20.2 ng/ml; range,
11.3-67.9 ng/ml). Area under the curve for labeled serum LD closely predicted CSF labeled HVA concentrations ( r =
0.747,p = 0.033).Labeled CSF HVA levels did not significantly correlate with either quality or duration of response to
the labeled LD dose. In a similar manner, labeled CSF HVA concentrations were not influenced by duration of disease
or previous daily LD dosage. These findings support the hypothesis that levodopa-induced benefit in PD is not severely
limited by a defect in central levodopa metabolism.
Durso R, Evans JE, Joseph E, Szabo GK, Evans BA, Handler JS, Jcnnings D, Browne TR.
Central levodopa metabolism in Parkinson’s disease after administration of stable
isotope-labeled levodopa. Ann Neurol 1937;42:300-304
The replacement of striatal dopamine (DA) by the oral
administration of the precursor levodopa (LD) represents the most common therapeutic approach in Parkinson’s disease (PD). However, the full extent to
which peripheral blood LD levels can affect central
DA production is essentially unknown. Within the
parkinsonian striatum, due to reductions in dopadecarboxylase (the enzyme converting LD to DA),
there may exist an inability to convert all peripherally
delivered LD to DA, resulting in a loss of drug response. Specifically, autopsy studies indicate that this
enzyme’s activity in putamen and caudate of affected
individuals is about 7% and 15?40,respectively, of controls [I]. Furthermore, animal PD models [2-41 demonstrate with increasingly greater nigral damage a tendency to accumulate unmetabolized LD in striatum
with resultant decreases in DA production. ‘To date,
the ability to directly examine in vivo central LD metabolism in patients with PD as a function of druginduced benefit has been limited in large measure by
an absence of a safe and adequate labeling technique.
W e have developed a labeling methodology for LD
by using stable isotopes. T o our knowledge, this is the
first reported use of such labeling techniques for human in vivo LD studies. Stable isotopes are naturally
occurring isotopes that usually differ from their parent
atom (the most abundant form of the element) by the
addition of one or more neutrons. These substances are
not radioactive. An example of one such stable isotope
is 13C (compared with the usually occurring I2C). The
margin of safety in applying
as a pharmacologic
label has been described as “virtually limitless” [ 5 ] . A
drug labeled with such isotopes can be quantitated
along with its metabolites by using gas chromatography-mass spectrometry (GC-MS). In our labeled LD
form, all carbons within the catechol ring were replaced with the stable isotope ’.3C,making the compound six atomic mass units greater than the unlabeled
analogue. Because conversion of LD to DA and hoinovanillic acid (HVA) does not involve alteration of
the ring, all relevant catabolic products of this labeled
LD form will similarly be 6 mass units greater than
From the *Department o f Neurology, Boston University School of
Medicine and Boston Veterans Administration Medical Center,
Boston; and tMass Specrrometry Laboratory, Eunice Kennedy
Shriver Ccntcr, Waitham, MA.
Address correspondence to Dr Durso, Department of Neurology,
Boston University School of Medicine/Boston VAMC, 150 South
Huiitington Avenue, Boston, MA 021 SO.
Received Nov 7, 1996, and in revised form J m 29, 1997. Accepted
for publication Mar 10, 1997.
0 1397 by the American Neurological Association
background unlabeled metabolite. W e are able to deliver a full pharmacologic dose of LD that is completely labeled because o u r form is not radioactive. All
quantitation of metabolites, therefore, represents actual
amounts produced from the administered dose a n d is
not subject to errors related to estimation based on tiny
radioactive tracer doses. Using labeled cerebrospinal
fluid (CSF) HVA as a marker for striatal LD a n d DA
metabolism, we now present data that suggests that
over a wide range of PD severity a n d a t serum LD
levels typical of clinical practice, central dopadecarboxylase activity is not saturated. In addition, w e
d o not detect a significant relationship between degree
of central LD metabolism to HVA a n d clinical d r u g
response. We believe these findings are most consistent
with the hypothesis that LD-induced benefit i n PD is
not severely limited by a defect in central LD metabolism.
Materials and Methods
A total of 8 PD patients participated in our infusion studies.
Subjects were all diagnosed with idiopathic PD by the principal author (R.D.) and at least one other neurologist. All
patients had experienced previous benefit from LD except for
1. The latter had been newly diagnosed and his first LD
exposure was the stable isotope-labeled LD given in our
study. The LD response characteristics for the remaining patients were (1) 4 patients, random “on-off‘ phenomena; (2)
2 patients, “wearing off”; and (3) 1 patient, stable response
(no fluctuations on 3 times per day dosing). The mean (standard deviation, range) age of our patient group was 60 years
(12 years, 44-74 years) and they averaged an 8-year (5 years,
0.5-15 years) duration of disease. Their mean Hoehn-Yahr
score was 3.5 (range, 2-5) in the “off” state. For patients
taking carbidopa (CD)/LD, their average daily dose before
study was 109 mg (44 mg, 37.5-175 mg)/543 mg (259 mg,
150-1,000 mg). Five patients were taking pergolide, 3 selegiline, and 1 bromocriptine before the stable isotope-labeled
LD infusion. All medications were discontinued on the day
of the infusion. The last CD/LD or DA agonist dose as part
of a daily medication regimen in all cases had been given at
least 12 hours before the study infusion.
After signing informed consent, patients were admitted to
the hospital on the afternoon before the day of study. Approval for the project had been given by the Research and
Development Committee and Human Studies Subcommittee at the Boston Veterans Administration Medical Center. A
Food and Drug Administration (FDA) investigational new
drug approval had been obtained for the stable isotopelabeled LD form by one of us (R.D.) before use. This compound was more than 99% isotopically pure at each labeled
site and chemical purity was 98.1%. This high degree of isotopic purity guaranteed that only +6 LD would be present
as a consequence of the labeled LD infusion [ie, there would
be no contamination by f O (unlabeled) LD unless residual
LD had been present from previous intake of medication].
The source of this labeled form was Cambridge Isotopes
Laboratory (Andover, MA). O n the day of investigation,
food and all medications were held from 12 midnight. At 7
subjects were connected to a cardiac monitor and rwo
heparin locks (one for blood drawing and one for the LD
infusion) were placed in arm veins (one lock in each arm).
All patients received 50 mg of C D orally at 8 A.M. At 9 A.M.
subjects were examined using a modified Unified Parkinson
Disease Rating Scale (UPDRS) and a blood sample was
drawn (heparin within the lock was first discarded). Immediately after the serum sample, a bolus infusion was started
of 150 mg (1 mg/ml) over 12 to 15 minutes of stable isotope-labeled LD (ring-l’,2’,3’,4’,5‘,6’-i’C6)
in 5% dextrose
injection USP solution (pH 4.5) using Harvard infusion
pumps (Harvard Apparatus Syringe Infusion Pump 22). Repeat blood drawings were done every 30 or 60 minutes for a
duration of 360 minutes. Samples were immediately spun
with resultant serum pipetted into polypropylene tubes and
then frozen (-70°C). The severity of patients’ parkinsonism
was measured every 30 minutes during the study. At 360
minutes a lumbar puncture was performed with CSF collected in sequential 5-ml aliquots (total, 25 ml).
Laboratory methodologies involved the use of both highperformance liquid chromatography (HPLC) and GC-MS.
HPLC was used for quantitation of total (labeled and unlabeled) HVA and LD concentrations. GC-MS was used to
determine ratio of labeled to unlabeled HVA. By knowing
the labeled/unlabeled ratio as well as total concentration, we
were able to calculate absolute amounts for labeled and unlabeled metabolite. Our HPLC assay was derived from a
previously published methodology [b] and allowed for the
simultaneous analysis of LD and HVA. Our GC-MS methodology involved a modification of the technique described
Statistical analysis for the
by De Jong and colleagues [7].
clinical pharmacologic data involved use of the Mann-Whitney U test and Pearson correlations.
O u r infusion produced blood levels of LD (Fig I) similar to those observed in PD patients taking an oral
251250 CD/LD dose [8]. Residual unlabeled LD was
Fig 1. Mean serum levels of levodopu (LO) (squares), labeled
homovunillic acid (HVA)(solid circles), and unlabeled HVA
(open circles) are shown for 8 patients ajier an intravenous
bolus infision of 150 mg of stable isotope-labeled LD. Error
bars represent standard deviation.
Durso et al: Levodopa Metabolism in PD
not present after the overnight fast (time 0). In addition, there was no increase in serum-unlabeled HVA
after the infusion, confirming the isotopic purity of our
labeled LD compound. All subjects received clinical
benefit after administering the labeled LD. Specifically,
the mean maximum improvement achieved (percentage
of change from baseline) was 65% (3O%, 13-100%).
The duration of response averaged 225 minutes (88
minutes, 120-360 minutes). These response parameters are very similar to those obtained after an oral 25/
250 CD/LD dose [9].
The average percentage of labeled HVA in lumbar
CSF (calculated as the simple mean of percentage of
labeled HVA in CSF tubes #1 and 85) 6 hours after
the bolus infusion was 54% (9%, 34-67%). When analyzing absolute levels of labeled HVA (Fig 2), a spinal
gradient was evident in all patients, as the mean concentration of labeled CSF HVA in tubes #5 was more
than double that in tubes #1 (49.2 compared with
20.2 ng/ml) ( p = 0.012, Mann-Whitney U test). The
average labeled CSF HVA concentration (calculated as
the simple mean of labeled HVA levels in tubes #I and
#5) correlated significantly with area under the curve
(AUC) for serum LD (see the Table). A significant correlation could not be found for average labeled CSF
HVA levels and either magnitude or duration of response to the labeled LD dose. In addition, labeled
CSF HVA concentrations were not influenced by duration of disease or previous daily L D dosage.
Tissue levels of DA are short-lived and present only in
trace amounts in human PD CSF [lo]. Because HVA
Fig 2. Absolute Levels of labeled homovanillic acid (HVA) in
lumbar cerebrospinal fluid (CSF) 6 hours after bolus infision
of 150 mg of labeled levodopa f i r 8 patients. Tube 1 represents the Jrst-to-fifih milliliters and tube 5, the 2Oth-to-25tb
milliliters o f CSFfiom tbe lumbar punctiue. Tbere is a signijcdnt difference between labeled HVA Levels in the two
tubes (p = 0.012, Maw-Whitney U test).
(49.2 ng/rnl)
Tube #1
Tube 85
302 Annals of Neurology
Vol 42
No 3
September 1997
represents the final stable metabolite of DA and closely
parallels levels of tissue DA after physiologic and pharmacologic manipulation [11-141, it is frequently used
as the preferred marker for DA. Such use of HVA in
CSF is based on the fact that brain extracellular fluid is
contiguous with CSF. It is known that stimulation of
the nigra (producing increases in striatal DA metabolism) results in the release of HVA into ventricular
CSF [15]. In human autopsy data, strong correlations
exist between ventricular CSF HVA and striatal tissue
HVA concentrations [16]. The most recent comprehensive review of HVA concludes that “CSF HVA appears to originate from a restricted region of brain,
mainly from the portion of the striatum adjacent to the
lateral ventricular walls” [ 171. Demonstration of a ventricular-lumbar CSF gradient for CSF HVA [18], as
well as observations that lumbar CSF HVA levels are
dramatically attenuated when ventricular-lumbar CSF
flow is disrupted [19], serve as evidence that lumbar
CSF HVA is almost exclusively derived from brain. In
our study, striatal DA production (as estimated by labeled CSF HVA) was directly proportional to levels of
LD attained in blood. This suggests that there is no
major impediment to central LD decarboxylation a t
LD blood levels typical of treated PD. O u r results indicate that higher systemic LD levels might have further increased central DA production. We note, however, it is uncertain that such an increase would have
resulted in greater benefit (ie, there was no correlation
between degree of central LD metabolism and clinical
response). The data in this study suggest that the “all
or none” response described by Nutt [20] (ie, systemic
LD doses beyond those required to produce an initial
response do not result in greater improvement) is unlikely due to a saturation of central dopa-decarboxylase
activity. Our results imply that factors other than central DA production (eg, storage capacity of nigrostriatal
axon terminals, DA receptor sensitivity) may play the
more important role in maximizing benefit derived
from LD. We also could not find a significant influence of duration of disease on central LD metabolism.
We believe these results do not conflict with L-[’*F]fluorodopa positron emission tomography (PET) studies that show diminished striatal activity with advancing
PD 121). These PET studies were interpreted by
their authors as primarily demonstrating a diminished
capacity to store DA in striatal terminals rather than a
limitation in the ability to synthesize fluoroDA from
an administered fluorodopa dose.
We were most surprised by the large percentage
(54%) of labeled HVA present in lumbar CSF after
only a single labeled LD dose. Previous studies in humans, with unlabeled and radioactive LD, had led us
to believe that less than 10% of the metabolite would
be tagged. Specifically, after a single unlabeled CD/LD
dose with serial sampling of ventricular CSF HVA, lit-
Tdble. Correlations Involving Labeled LD and HVA Concentrations
Variables X-Y
AUC serum labeled LD-average labeled CSF HVA
AUC serum labeled HVA-average labeled CSF HVA
AUC serum labeled LD-best improvement after infusion (% of change)
AUC serum labeled LD-duration of response to infusion
Average labeled CSF HVA-best improvement after infusion (% of change)
Average labeled CSF HVA-duration of response to infusion
Average labeled CSF HVA-age
Average labeled CSF HVA-duration of disease
Average labeled CSF HVA-previous daily levodopa dosage
LD = levodopa; HVA = homovanillic acid; AUC = area under the curve; CSF
tle or no change in CSF HVA concentrations can be
found [22-241. For radioactive LD, barely detectable
amounts of labeled HVA are reported in lumbar CSF 2
hours after an oral tracer dose [25].These previous reports, combined with our data, indicate that the CSF
HVA pool is far more dynamic than previously recognized. Total CSF HVA remains unchanged after an
unlabeled LD dose, because newly produced HVA is
probably replacing existing CSF HVA at a rapid rate.
HVA leaves the CSF pool via a CSF excretion mechanism for acidic dopaminergic metabolites [26]. We
believe the previously cited radioactive LD study was
unable to detect this large turnover probably due to a
combination of (1) the requirement that only tiny
tracer doses of labeled LD be used and (2) lumbar CSF
sampling being too soon after oral drug administration
(2 hours).
There is an abundance of literature that supports an
exclusive central origin for CSF HVA and the inability
of HVA to penetrate from blood into CSF [17]. Our
own data also support a central origin for CSF HVA.
The gradient for labeled CSF HVA in this study
strongly indicates that newly produced metabolite was
excreted into CSF at rostra1 levels rather than through
a diffuse blood-central nervous system-CSF transfer of
peripherally synthesized HVA. In addition, if the origin of labeled CSF HVA had been primarily from
transfer of labeled blood HVA into CSF rather than
from central metabolism, one might have expected a
significant correlation between levels of labeled HVA
in blood and those in CSF. This was not the case. W e
have previously reported such significant correlations
with serum and CSF LD [8], a substance that readily
crosses the blood-brain and presumably blood-CSF
barriers. W e also note that peak serum levels of labeled
HVA in our study were attained quickly (within 15
minutes) after the cessation of the LD infusion. Although there are no other in vivo human studies to
compare with, such rapid systemic conversion of LD to
HVA after parenteral LD administration has been observed in animals. Specifically, Rose and associates [27]
cerebrospinal fluid.
report that peak blood levels of HVA occur within 10
minutes after an intra-aortic dose of LD in rats. Even
after pretreatment with extremely large doses of CD
(25 mg/kg), a rapid rise in systemic HVA after intraaortic LD is present (ie, peak levels of HVA occur
within 30 minutes after the parenteral dose).
Supported in part by a VA Merit Review grant and a research grant
from the American Parkinson Disease Association.
1. Lloyd KG, Davidson L, Hornykiewicz 0. Disease: effect of
L-dopa therapy. J Pharmacol Exp Ther 1975;195:453-464
2. Spencer SE, Wooten FG. Altered pharmacokinetics of L-dopa
metabolism in rat striatum deprived of dopaminergic innervation. Neurology 1984;34:1105-1108
3. Goldberg A, Melamed E, Rosenthal J, et al. Synthesis of dopamine from exogenous levodopa in striatum: effect of graded destruction of dopaminergic nerve terminals by increasing doses
of MPTP. Neurology 1930;40(suppl 1):291 (Abstract)
4. Burns RS, Kopin I, Weiss V, et al. Decreased LAAD in striarum influences brain L-dopa metabolism in MPTP-treated rhesus monkey with severe parkinsonism. Neurology 1991;41
(suppl 1):378 (Abstract)
5. Speilmann H , Nau H. Embryotoxicity of stable isotopes and
use of stable isotopes in studies of teratogenetic mechanisms.
J Clin Pharmacol 1986;26:474-480
6. Szabo G, Davoudi HG, Durso R. High-performance liquid
chromatographic method for measuring homovanillic acid in
cerebrospinal fluid using electrochemical detection with internal
standardization. J Chromatogr 1988;430:112-I 17
7. De Jong APJM, Kok RM, Cramers CA, Wadman SK. Determination of acidic catecholamine metabolites in plasma and cerebrospinal fluid using gas chromatography-negative-ion mass
spectrometry. J Chromatogr 1986;382:19-30
8. Durso R, Szabo G, Davoudi H , Feldman RG. Magnitude of
response to levodopa in Parkinson disease as it relates to peripheral and central measurements of levodopa and associated
metabolites. Clin Neuropharmacol 1989;12:384-392
9. Durso R, Issac K, Perry L, et al. Age influences magnitude but
not duration of response to levodopa. J Neural Neurosurg Psychiatry 1993;56:65-68
10. LeWitt PA, Galloway MP, Matson W, et al. Markers of dopamine metabolism in Parkinson disease. Neurology 1932;42:
2 1 1 1-2 117
1 1 . Roth RH, Walters JR, Aghajanian GK. Effect of impulse flow
Durso et al: Levodopa Metabolism in PD
on the release and synthesis of dopamine in the rat striacum. In:
Usdin E, Snyder SH. eds. Frontiers in catecholamine research.
New York: Pergamon Press, 1973:567-574
12. Korf J, Grasdijk I,, Westerink BHC. Effects of electrical stimulnrion of die nigrostriatal parhway of the rar on dopamine metabolism. J Neurochem 1976;26:579-584
1.3. Extein I, Koth RH, Rowers MB. Accumulation of ’Hhomovunillic acid in rabbit hrain and cerebrospinal fluid following intravenous ‘H-~-dopa.Biol Psychiatry 1974;9:161-170
14. Bacopoiilos NG, Redmond DE, Baulu I, Roth RH. Chronic
haloperidol 01-fluphcnazine: effects on dopamine metabolism in
hrain, cerebrospinal fluid and plasma of Ce~copitl~ecus
(verver monkey). J I’harniacol Exp Ther 1380;212:1-5
15. I’ortig I)], Vogt M. Release to the cerebral ventricles of substances with possible transmitter function in the caudate nucleus. J Physiol (Lond) 1369;204:687-715
16. Wester P, Bergstrom U,Eriksson A, et al. Ventriculat cerehrospinal fluid monoamine cransinitter and metabolite concentrarions rctlect human brain neurochemistry in autopsy cases.
J Neurochem 1390;54:1148-1156
17. Amin k’,Davidson M, Davis KL. Homovanillic acid measurement in clinical research: a review of methodology. Schizophr
BLIH 1302;18:123-148
18. Moir ATB, Ashcroft GW, Crawford TBB, et al. Cerebral metabolites in cerebrospinal fluid as a biochemical approach to the
brairi. Brain 1970;93:357-368
19. Post RM, Goodwin FK, Gordon E, Watkin D M . Amine metabolites in human cerebrospinal fluid: effects of cord transection and spin,il fluid block. Science 1973;179:897-899
Annals of Neurology
Vol 42
No 3
September 1997
20. Nutt JG. On-off phenomenon: relation to levodopa pharmacokinetics and pharmacodynamics. Ann Neurol 1987;22:535-540
21. Leenders KL, Palmer AJ, Quinn N, et al. Brain dopamine metabolism in patienrs with Parkinson’s disease measured with
positron emission tomography. J Neurol Neurosurg Psychiatry
22. Olanow C W , Gauger LL, Cedarbaum JM. Temporal relationships between plasma and cerebrospinal fluid pharmacokinerics
of levodopa and clinical effect in Parkinson’s disease. Ann Neurol 1991;23:556-559
23. Moussa 2,Raftopoulos C, Przedborski S, Hildebrand J. Lack of
changes in ventricular cerebrospinal fluid concentrations of hoinovanillic acid following acute challenge with levodopa. Ann
Neurol 1992;31:113-114
24. Raftopoulos C, Dethy S, Laute A, et al. Slow increase of homovanillic acid in cerebrospinal fluid after levodopa administration. Mov Disord 1996;11:59-62
25. Morgan JP, Bianchine JR, Spiegel HE, et al. Metabolism of
levodopa in patients with Parkinson’s disease. Arch Neurol
1971;2S:39- 44
26. Ashcroft GW, Dow RC, Mori ATB. The active transport of
5-hydroxyindoleacetic acid and 3-methoxy-4-hydroxyphenylacetic acid from a recirculatory perfusion system of rhe cerebral
ventricles of the unanesthecized dog. J Physiol (Lond) 1768;
27. Rose S, Jenner P, Marsden CD. Peripheral pharmacokinetic
handling and metabolism of l-dopa in the rat: the effect of
route of administration and carhidopa pretreatment. J Pharm
Pharmacol 1991;43:325-330
Без категории
Размер файла
497 Кб
central, metabolico, isotopeчlabeled, administration, disease, parkinson, stable, levodopa
Пожаловаться на содержимое документа