12, 106-l 15 (1975)
of Nanogram Homovanillic
Central Nervous System Semiautomated Fluorometric BEN H. C. WESTERINK'
a Rapid Method
AND JAKOB KORF~
Laboratory for Pharmaceutical and Analytical Chemistry, Department of Clinical Chemistry,’ and Department of Biological Psychiatq, Groningen University,= The Netherlands Received
INTRODUCTION Since the development of fluorometric assay techniques for homovanillic acid (4 hydroxy-3-methoxyphenyl acetic acid, HVA, 1,2), a major dopamine (DA) metabolite, numerous reports have indicated the relevance of such measurements in nervous tissue. Moreover the levels of HVA in cerebrospinal fluid (CSF) and in the basal ganglia of the central nervous system (CNS) were found to be decreased in patients suffering from Parkinson’s disease as a result of degeneration of DA-containing neurons (3,4). Neuroleptics, including haloperidol, having marked effects on the turnover rate of DA in the CNS (5,6) also increase the levels of HVA in CNS of laboratory animals (7), as well as in CSF of patients during treatment with these drugs (8,9). As a consequence of such observations, one can conclude that HVA levels in CNS, including CSF, reflect at least in part, the rate of metabolism of DA in the CNS. Recent advanced techniques for the estimation of HVA in biological material are based on the use of gas-liquid chromatography (GLC) combined with electron-capture detection or with mass spectrometry (GLC-MS (10,11,12)). Although these techniques are very specific, at the present time most laboratories are not equipped with these instruments. GLC-MS requires experience so they are not used for routine measurements in most institutes easily. Current fluorometric analyses for HVA are rather specific, however relatively large amounts of CSF (2-10 ml) or brain tissue (pooled brain areas of several rats) being required for reliable measurements of HVA (1,13,14). The present method is based on a specific isolation procedure of HVA on small columns of Sephadex G 10, according to modifications of previously described methods (14,15), followed by an automated detection technique with a continuous flow system (Autoanalyzer?. This method enables us to determine HVA in samples reproducibly, con106 Copyright 0 1975 by Academic Press, Inc. All rights of reproduction in any form reserved.
taining no more than 2 ng (2 x lwg g). This is the first time to our knowledge that HVA levels in separate mesolimbic structures of the rat, including the nucleus accumbens and the olfactory tubercle, are reported. The reliabtility of the method is confirmed by treatment of rats with drugs, as the neuroleptic haloperidol and the monoamine oxidase inhibitor pargyline. MATERIALS Reagents All solutions were made with double glass distilled water. K,Fe(CN), reagent was prepared by dissolving 10 mg of K,Fe(CN), (Merck) in a mixture of 85 ml water and 15 ml concentrated ammonia (25% NH,, Merck). The solution was made up the day before use, unless otherwise indicated. Blank reagent: 1.5 ml concentrated ammonia (25% NHB) was diluted to 100 ml with water. Cysteine solution contained 40 mg cysteine (Merck) in 100 ml water and was used only on the same day. Phosphate bufler. Nine hundred milligrams of Na,HPO,. 2H,O (Merck) and 10 mg KH,PO, (Merck) were dissolved in 1 liter of water. Some batches of phosphate gave high blanks and were discarded. The pH of the phosphate buffer (0.5 mM) was 8.5. HVA-standards. Stock solutions of HVA (Fluka) contained 10 mg HVA/lOO ml, and were stored in vials of about 1 ml in the freezer (- 20°C). Standard solutions of HVA were made from one portion of the stock solution after appropriate dilution with the phosphate buffer. Usually standard curves were made with HVA concentration between 5-25 r&ml. KOH/formate solution was made by adding to 10 N KOH (Merck) concentrated formic acid (98% Merck) till the concentration of the latter was 5 N. Perchloric acid was purchased from Merck. Instrumentation We used the continuous flow analysis system (Autoanalyzef) of Technicon equipped with a fluorometer. The light source was a Mercury lamp and the following filters were used: primary filter Kodak Wratten 312 nm and a secondary filter Corning 373. The sampling to washing time ratio was determined with an electronic device, with which we could choose the various time intervals at steps of 10 sec. METHODS Isolation Procedures Cerebrospinal fluid. Sephadex G 10 columns (length 40 mm) were prepared in Pasteur pipettes (diameter 5 mm), with tips of about 125 mm. A piece of glass wool was placed at the bottom of the pipette and the
spontaneous flow through the column was prevented with [email protected]
). The pipette was filled with a suspension of Sephadex G 10, prepared in water one day before use. After the Sephadex was precipitated and the appropriate height was obtained, the [email protected]
was removed. In our hands this procedure produced columns of the Sephadex with optimal binding capacity for HVA and with sharp elution peaks. The column was washed with 3 ml 0.01 N ammonia and 3 ml 0.01 N formic acid. A CSFsample (0.5 ml), adjusted to a pH of 2.5 with a drop of formic acid (2.5%) was applied to the column. After the HVA was absorbed to the Sephadex, the column was washed with 1.5 ml 0.01 formic acid and 0.4 ml phosphate buffer. The HVA was eluted with 1.5 ml phosphate buffer and collected in an AutoanalyzeP cup. Regeneration of the column was performed with 0.01 N ammonia, 0.01 N formic acid and water. After this treatment the columns can be stored at 4°C several weeks before reuse. During the isolation and regeneration procedure columns are allowed to run dry. Tissue. Brain structures dissected according to Horn et al. (16, nucleus accumbens and olfactory tubercle) and Glowinski and Iversen (17) (corpus striatum) or according to a stereotaxic atlas (IS), were frozen on dry ice, weighted and stored at -8O”C, until analysis. Tissue (10-200 mg) was homogenized in 1 ml 0.4 M perchloric acid with a Potter-Elvehjem homogenizer. The excess of chlorate was precipitated by addition of two drops of KOH/formate solution. After centrifugation (20 min, 4OOOg, 4°C) the clear supernatant was decanted and stored at - 20°C until analysis (within 1 wk). Isolation of HVA in tissue extracts was performed on columns of Sephadex G 10 (5 x 70 mm) prepared as described in the previous section. After the tissue extract was passed through the column, 2.5 ml 0.01 N formic acid and 1.3 ml phosphate buffer were added consecutively. HVA was eluted with 1.5 ml phosphate buffer and collected in sample cups. Flow Diagram and Elution Projiles The flow diagram of the Autoanalyzer is shown in Fig. 1, and is based on the manual fluorometric procedure of Anden et al. (1). In this method a fluorophore of HVA is formed by oxidative dimerization by K,Fe(CN), under alkaline conditions. Sample blanks were obtained by using the residue of the samples; however, the K,Fe(CN), was replaced by the blank reagent. The Pasteur pipette, containing the Sephadex, was directly connected to the pump of the Autoanalyzer and the elution pattern was recorded at a flow rate of 0.32 ml/min. This flow rate is close to the flow rate of the columns used for CSF analysis. The correct proportion of reagents in the manifold was obtained by appropriate dilution of the eluate with
ASSAY OF HVA
Sampler sample:40 set wash : 30 set ml/min SMC
-cl fluorometer I exit.315nm recordel emis.>4lOnm
,. 0.16 : 010 _ 1.06
O,Ol%K3FpCN6 I 2NNH4OI.I 0,4%Cysteine
I SMC=single mixing coil DMC = double mixing coil
FIG. 1. Flow diagram of the automated determination (Technicon).
of HVA. Dl refers to standard fit-
additional tubing. Thus standard solutions or CSF were applied to the Sephadex columns; the volume of washing and elution solutions were as described for the CSF method. Drugs
Male rats of a Wistar derived strain (200-250 g, T.N.O., Zeist, The Netherlands) were injected, intraperitoneally, with haloperidol (0.5 mg/kg; Janssens Pharmaceutics) or pargyline. HCl (80 mg/kg, Abbott) 2 hr prior to decapitation. RESULTS AND
Adaptation of a manual assay to an automatic method, based on the continuous flow technique, usually results in an increased sensitivity, due in part to the following factors: smaller volumes can be used, both for the reaction and fluorometry, and the baseline, “reagent blank,” is very stable. Thus in the present assay we can reliably measure 10 times smaller amounts of HVA than with manual methods (1,2,13,14,15). A typical standard curve and the reproducibility of the present assay is shown in Fig. 3. With the time intervals for sampling (40 set) and washings (30 set) 98% of the maximal peak height, obtained after continuous sampling, was obtained. A problem of measuring HVA concentrations of
FIG. 2. Elution profiles of a Sephadex column after 0.5 ml CSF (50 ng HVA/ml) was applied. (The amount of phosphate buffer used was slightly more than used in the routine procedure.) Curve (a): in presence of K,Fe(CN&; (b): blank.
FIG. 3. Typical recordings of HVA standards. Numbers indicate concentration in rig/ml.
5 10 15 20 25 4. Standard curves obtained with reagent prepared the day before use (a) and with freshly prepared reagent (b). FIG.
2-10 rig/ml is the alinearity of the fluorometric method (15). However, acceptable standard curves were obtained, when the K,Fe(CN), reagent was made the day before use (Fig. 4). As indicated by Andin et al. (l), tissue blanks can be made by reversing the order of addition of the solutions of K,Fe(CN), and of cysteine, thus preventing the oxidation of HVA. With our method blanks were made by omitting the oxidant. Column
The maximal volume of 0.01 N formic acid necessary to remove interfering compounds with the fluorometric detection method depends on the length of the column. Exceeding a volume of 1.5 ml 0.01 N formic acid, used for CSF, gave lower recoveries of HVA and was less reproducible. The columns used for CSF were not applicable for tissue extracts, because the purification of HVA was insufficient. In contrast to the original procedure (14,1.5), we used a 1,000 times diluted phosphate buffer to elute HVA from the Sephadex. A typical elution profile of a CSF sample is shown in Fig. 2. Elution with a phosphate buffer is very specific (14); however, oxidation of HVA in such a buffer requires a laborious pH control (13). This step is not permissible for a routine assay, we preferring the oxidation procedure in strong alkaline (1). Under these conditions however phosphate buffer quenches fluorescence, if not diluted to 0.5 mM. Elution of HVA from Sephadex G 10 can also be performed with 0.01 N ammonia. However elution with ammonia gives an unacceptable high blank as Fig. 2 shows.
The recovery of HVA (25 ng) added to CSF or water was 92.6 +- 8.6% (mean & SD, number of experiments = 20) and 93.5 + 2.5% (n = lo), respectively. From 16 patients we compared CSF levels of HVA assayed according to the present procedure and our manual procedure (14). All patients received probenecid (14). Mean level of CSF-HVA obtained with the automated method was 106.0 r&ml compared to 94.3 rig/ml with the manual method. The 12% higher levels of CSF-HVA measured by the present method is due to the more reproducible isolation with better controlled recoveries of HVA on the small columns of Sephadex G 10. HVA in Tissue The distribution of HVA in some areas in the CNS is given in Table 1. As is to be expected from histochemical (19) and biochemical (4) studies, the highest levels of HVA were found in the corpus striatum. The method is sufficiently sensitive to measure HVA in the left and right striatal tissue, independently. Of the nucleus accumbens, the bilateral structures of one rat were combined, while olfactory tubercles of two rats were pooled for the assay. After haloperidol treatment the HVA levels in the various structures could be estimated in a single rat brain. The specificity of the assay is demonstrated after depletion of HVA by pargyline treatment. In this case HVA-levels found in all brain areas studied were virtually zero. A typical recording, including blanks, found after the treatment with pargyline and haloperidol is shown in Fig. 5. Even after haloperidol treatment we could not detect HVA in the hippocampus, hypothalamus, amygdala, cerebral cortex, cerebellum, and spinal cord, areas which have been shown to contain little or no DAcontaining nerve terminals. The found absence of HVA in the spinal cord supports the hypothesis that lumbar CSF of man contains only HVA derived from brain tissue (20). Fluorometric HVA assays have been criticized by Sjdquist and Anggard (lo), because of unspecificity. It is gratifying that the HVA CONTENT
TABLE 1 IN THREE DIFFERENT AREAS OF THE RAT BRAIN WITH HALOPERIDOL OR PARGYLINE
Corpus striatum Nucleus accumbens Olfactory tuber&
12 I? 8
41.6 + 6.9 20.1 t 3.4 18.3 f 3.5
Haloperidol mHVA/g 0.83 r 0.12 0.46 2 0.22 0.34 z 0.09
0.5 mgikg ip
4 6 5
47.8 t 8.3 21.0 2 4.3 9.1 t- 3.2
pgHVA& 3.30 -r 0.70 2.40 t 0.66 0.85 f 0.18
4 4 2
39.6 ? 4.5 15.8 2 1.60 23 and 24
up ~g!HVAip NV ND ND
Bilateral ulfdctory tuber&s of two rats were combined in control and pagyline treated animals. Content? of HVA in corpus striatum and weight< are given as the mean values of left and right structures Content in pgig wet weight i- SD. N = number of experiments: ND = not detectable; ip = intraperitoncally
ASSAY OF HVA
FIG. 5. HVA recordings (a) of different brain samples compared with blanks (b). Unilateral C. striatum tissue was applied. After haloperidol treatment HVA content in the C. striatum exceeded the recording capability.
levels in the corpus striatum obtained with GLC-MS (21) and the present assay technique are in good agreement (0.72 and 0.82 ,[email protected]
, respectively). Also the distribution of HVA throughout the CNS and its rapid disappearance after inhibition of monoamine oxidase demonstrates the specificity of the present fluorometric assay technique. The recovery of 25 ng HVA, added to 50 mg cerebellar tissue, which contains no HVA, and carried through the whole extraction and isolation procedure was 81.7 t 4.0% (mean 2 SD, 10 experiments). Interference of the fluorometric procedure by compounds related to HVA present in CSF or brain tissue seems unlikely as recently shown by Prasad and Fahn (22). At the time of completion of the development of the present procedure, a study appeared (23) on the use of an automated fluorometric detection of HVA after isolation on Sephadex G 10, this being based principally on our earlier described method (15). The method of Renaud et al. (23) is criticized on a number of points. The column procedure was inadequately described; Sephadex G 10 columns used were of too large a size and elution of HVA with diluted ammonia gives relatively high blanks (see Fig. 2) which do not permit determination of HVA below 10 rig/ml. This method (23) has not been applied to brain tissue and the specificity of the assay was not established.
The advantages of our present method are the specificity, sensitivity and reproducibility, enabling HVA to be measured in discrete areas of the rat brain without the need for pooling. Moreover the method is rapid: 80-120 samples in a working day can be handled. Replication of the method in other laboratories could be relatively simple. With our assay we can compare the effect of various neuroleptics and analgesics on different dopaminergic systems of the rat brain. Recently we applied the method for determining HVA output in a ventricular perfusate in rats, enabling HVA to be measured in 15 min fractions for several hours. SUMMARY A semiautomated fluorometric assay technique for homovanillic acid (HVA), a metabolite of dopamine, is described. The method is based on a rapid manually performed isolation of HVA on small columns of Sephadex G 10 and an automated fluorometric detection method (continuous flow). The method is highly sensitive: 2 ng HVA/ml can be reproducibly determined. The specificity of the method was confirmed by the effects of drugs on the HVA levels in the rat brain and by the distribution of HVA throughout the central nervous system: detectable levels of HVA were only found in the corpus striatum; nucleus accumbens and tuberculum olfactorius. The method is rapid (80-120 samples/day), reproducible (recoveries 80-95%) and can be applied to both brain tissue and cerebrospinal fluid. ACKNOWLEDGMENTS The authors thank Prof. Dr. D. A. Doornbos, Prof. Dr. J. S. Faber, Dr. A. Groen, Dr. D. F. Matheson, and Prof. Dr. H. M. van Praag for reading and criticizing the manuscript. The illustrations were drawn by Mr. W. Prummel.
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
Anden, N. E., Roos, B. E., and Werdenius, B.. Life Sci. 2, 448 (1963). Sharman, D. F., &it. J. Pharmacol. 20, 204 (1963). Olsson, R., and Roos, B. E., Nature (London) 219, 502 (1968). Hornykiewicz, O., Pharmacol. Rev. 18, 925 (1966). Nyback, H., and Sedvall, G., Europ. J. Pharmacol. 10, 193 (1970). Anden, N. E.. Butcher, S. G., Corrodi. H.. Fuxe, K., Ungerstedt, U.. Europ. J. Pharmacol. 11, 303 (1970). Anden, N. E., Roos, B. E., and Werdenius, B.. Life Sci. 3, 149 (1964). Persson, T., and Roos, B. E., Br. J. Psychiat. 115, 95 (1969). Fyro, B., Wode-Helgodt, B., Borg, S.. Sedvall, G., Acta Psychiat. Stand. Suppl. 243, 54 (1973). Sjoquist, B., and Anggard, E., Anal. Chem. 44, 2297 (1972). Bert&son, L., Life Sci. 13, 859 (1973). Fri. G., Wiesel, F., Sedvall, G.. Psychopharmacologia 35, 295 (1974). Gerbode, F. A., and Bowers, M. B., J. Neurochem. 15. 1053 (1968). Korf, J.. and van Praag, H. M., Brain Res. 35, 221 (1971). Korf, J.. Ottema, S., and Van der Veen. I. H., An&. Biochem. 40, 187 (1971).
16. Horn, A. S.. Cuello, A. C., and Miller, R. J., J. Neurochem. 22, 265 (1974). 17. Glowinsky, J., and Iversen, L. L., J. Neurochem. 13, 655 (1966). 18. K&rig, I. F. R., and Klippel, R. A., “A Stereotaxic Atlas of the Forebrain and Lower Parts of the Brainstem. The Rat Brain.” Williams and Wilkins Co.. Baltimore, 1963. 19. Ungerstedt. U., Acta Physiol. Stand. 82, Suppl. 367, 1 (1971). 20. Sourkes, ‘I. L.. J. Neurul Transmission 34, 153 (1973). 21. Wiesel, F.. Fri. C.. and Sedvall, G., Europ. J. Pharmacol. 23, 104 (1973). 22. Prasad, A. L. N., and Fahn, S., J. Neurochem. 21, 1551 (1973). 23. Renaud, B.. Quenin, P., and Quincy, C., C/in. Chim. Actn 52, 179 (1974).