- Email: [email protected]
Isocratic reverse-phase HPLC separation and RIA used in the analysis of neuropeptides in brain tissue
Neuropeprides (1991) 20,201-209 0 Longman Group UK Ltd 1991
Isocratic Reverse-Phase HPLC Separation and RIA Used in the Analysis of Neuropeptides in Brain Tissue M. L. DE CEBALLOS*t,
M. D. TAYLOR*
and P. JENNER*
*Parkinson’s Disease Society Experimental Research Laboratories, Pharmacology Group, Biomedical Sciences Division, King’s College London, Manresa Road, London SW3 6LX, UK; tDepartment of Neuropharmacology, Cajal Institute, C.S.I.C., Doctor Arce 37, Madrid 28002, Spain (Correspondence and reprint requests to PJ)
Abstract-A reverse-phase, high-performance liquid chromatographic (HPLC) method was employed to separate and characterise five neuropeptides from complex mixtures, with important advantages over methods employed earlier using combined HPLC-RIA studies. Peptides were separated using 0.5M pyridine - 0.5M formic acid buffer, pH 4, containing propan-l-01 14% (met-enkephalin, leu-enkephalin, neurotensin) or 20% (CCK-S-S, substance at a flow rate of l.Oml/min. Isocratic conditions, and volatile solvents, resulted in a highly reproducible method, producing samples in a form designed for subsequent RIA. The application and importance of the procedure is demonstrated by comparison of the measurements of apparent peptide levels in crude brain extracts with those of authentic peptides as determined after HPLC purification.
Introduction
Radioimmunoassay (RIA) measurements of the peptide content of brain extracts has several disadvantages. RIA has the appropriate sensitivity, but the lack of totally specific antisera raises the possibility of cross-reactivity with related peptides and precursors/metabolites. The low levels of some neuropeptides may require the use of high tissue concentrations for their quantitation raising
Date received 26 June 1991 Date accepted 12 August 1991
P)
the possibility of non-specific interference with the RIA procedure (1). so, a separation procedure prior to the quantification of individual neuropeptides is desirable. High performance liquid chromatography (HPLC) has been applied successfully to peptides separation. Some methods involve gradient conditions (l-6) and salt buffer systems (2,5,6, 7). The former has two disadvantages, the lack of reproducibility of chromatographic runs and slowness in the processing of the biological samples. Salt buffers can also produce interference in the subsequent RIA used for peptide quantitation.
201
202
NEUROPEPTIDES
Isocratic HPLC conditions using highly volatile solvents were employed in the present study to separate a number of neuropeptides: met- and leu-enkephalin (met- and Ieu-enk), neurotensin (NT), cholecystokinin (CCK-8-S) and substance P (SP), which were measured using specific RIA’s. These neuropeptides were selected because of their localisation in basal ganglia, and the changes in content observed in neurodegenerative disorders, such as Parkinson’s disease (see for example 8, 9). The reproducibility of the method was assessed using either unlabelled or tritiated peptides. The solvent system was selected for the high recovery of the compounds subjected to the procedure, and its compatibility with the RIA used for their measurement. Finally, when the method was applied to biological samples comparison was made between measurements from the crude extracts and HPLC purified samples from discrete areas of rat and human brain.
Methods High performance
liquid chromatography
The HPLC system employed consisted of two Perkin Elmer (PE) series 10 liquid chromatographs, commanded by a PE LC series 10 prodetector (PE LC 75 grammer, a UV spectrophotometric detector) and a LKB model 2212 Helirac programmable fraction collector. Chromatographic data were plotted and integrated by a PE 3600 Data Station. Reverse phase separations were carried out using a Spherisorb 5 ODS 2 column (5~., 25 x 0.4cm, Phase Separations), protected with a guard column Spherisorb 10 ODS 2 (lop., 5 X 0.4cm, Phase Separations). All the reagents used were of the highest purity. Buffers were filtered through Millipore filters (HVLP, 0.45 l.~)and degassed before use. In most experiments synthetic and brain tissue-derived peptides were separated using 0.5M pyridine0.5 M formic acid buffer, pH 4, containing propanl-01 14% (met-enk, leu-enk, NT) or 20% (CCK-S-S, substance P) at a flow rate of 1 ml/min. When synthetic peptides were monitored with the UV detector, pyridine was substituted for triethylamine (TEA) to circumvent the high UV absorbance of pyridine. When linear gradients
were used the actual conditions are expressed where appropriate. All HPLC procedures were performed at room temperature (25 f 3°C). Synthetic met-enkephalin, leu-enkephalin, neurotensin CCK-8-S, CCK-8-N& substance P and dynorphin 1-13 (Sigma) were dissolved in oxygenfree HPLC-grade water, and aliquots were stored at -20°C. Peptides were further diluted in HPLC buffer before injection onto the HPLC column. Tritiated peptides were similarly treated and stored at 4°C until used. Subsequently, amounts between 30-1OOfmol were subjected to HPLC separation. Tritiated peptides were from Amersham International, except 3H-neurotensin which was supplied by New England Nuclear. The eluent was monitored at 254-280nm (nmol sensitivity) or subjected to RIA, after freeze-drying the solvent (fmol sensitivity). In the case of radiolabelled peptides (fmol sensitivity), fractions were counted in a Beckman scintillation counter. To minimise possible contamination of subsequent runs, synthetic peptides were always chromatographed at the end of a series of experiments and the chromatographic system thoroughly washed by pumping 70% propan-l-01 through the chromatograph. Extraction of brain peptides
Male Wistar rats (200-25Og) were killed by cervical dislocation and decapitation, and the brain quickly dissected on a cold plate. Dissected brain areas were immediately frozen on dry ice. The samples were boiled for 15 min in 100 volumes of a mixture of 1 N acetic acid and 0.02 N HCI, containing 0.1% of 2-mercaptoethanol. Samples were homogenized with a Polytron (setting 5, for 5s), and centrifuged (12000 x g, 10min). Aliquots of the supernatant (equivalent 5 mg of wet weight tissue) were freeze dried and stored at -40°C until subjected to HPLC. Samples were reconstituted in HPLC buffer, microfuged for 3min and injected onto the column. Samples for CCK-8-S measurements were boiled in distilled water for 5min, homogenized and centrifuged as above. These extraction methods were also used for human brain tissue samples obtained at post-mortem from normal subjects. In preliminary experiments the possible formation of oxidised peptides, during homogeni-
HPLC-RADIOIMMUNOASSAY
203
(RIA) FOR NEUROPEPTIDES
sation of brain tissue, was assessed. In the presence of 2-mercaptoethanol no oxidised peptides were formed. There was only one peak, corresponding to met-enk, before or after sonication, as monitored by UV detection. Radioimmunoassays
Specific RIA’s were used for peptide quantitation. HPLC eluates containing individual peptides were reconstituted in the appropriate volume of buffer (50mM sodium phosphate buffer containing 0.2% gelatin and 10mM EDTA for the enkephalins, CCK-8-S and SP) pH 7.2. For NT RIA the assay buffer was 60mM sodium phosphate, pH 7.2, containing 0.3% bovine serum albumin and 10mM EDTA. An aliquot of samples or standards (100~1) was incubated with 100~1 of antisera (Amersham International) and ‘251-peptide (15 OOOcpm; Amersham International). The tubes were incubated at 4°C for 22-24h for met-enk, leu-enk and NT, and for up to 3 days CCK-8-S and SP to achieve appropriate sensitivity. Separation of bound and free peptide was achieved adding 250-500~1 of activated charcoal dextran in RIA buffer. After centrifugation (5000 x g, 10min) the supernatant was decanted, and counted in a LKB Minigamma counter for 1 min.
Table 1
Effects of different chromatographic conditions on the retention time of various neuropeptides Peptide
ox met-enk
met-enk leu-enk dyn 1-13 dyn l-8 ocrapeptide heptapeptide CCK-8-S CCK-8-NS ox SP SP neurotensin
A
6
c
D
5 7 15 12
5 7 14 9.6 14 16.3
5 7 13 8
9 14
ND ND ND
11 16 7 10 10
10.6 12.7 ND ND ND ND 19
ND
17
ND 6 6
ND
Chromatographic conditions: 0.5M TEAF buffer. pH 4, flow rate 1.Omi/min. A: + 12% propan-l-01; B: + 15% propan-l-01: C: +20% propan-l-01; D: + 15% acetonitrile. ND: not detected in runs up heptapeptide:
to 30min. octapeptide: met-enk-Arg-Gly-Leu: met-enk-Arg-Phe; ox: oxidised.
Freeze dried aliquots of tissue extracts were assayed with and without HPLC separation (the equivalent of 4mg of tissue being reconstituted in the same volume of RIA buffer in both cases). Cross reactivities of the antisera were as follows: the antisera did not display any cross-reactivity with dynorphin 1-13, met-enk-Arg-Phe, met-enkArg-Gly-Leu or CCK-8-S. Enkephalin antisera did not cross-react with antisera against NT, or vice versa (0.1%). Met-enk antisera cross-reacted to 6% with met-enk. The antiserum against SP had no detectable cross-reactivity with the following peptides: met-enk-Arg-Phe, met-enk-Arg-GlyLeu, eledoisin, physaelemin, neurokinin A, neurokinin B and only negligible cross-reactivity (0.02%) for kassinin. The RIA sensitivities (15% displacement of bound tracer) were 1.5fmol/tube for both enkephalins, 1fmoVtube for NT, 0.6fmoVtube for CCK-8-S and 1 fmol/tube for SP.
Results
The effects of varying the percent propan-l-01 on neuropeptide retention times are shown in Table 1. Several peptides were separated with 0.5M triethylamine-formate buffer (TEAF). pH 4, containing either 12% or 14% propan-l-01. Increasing concentrations of propan-l-01 reduced neuropeptide retention times. NT and dynorphin 1-13 were more sensitive to the amount of propanol contained in the buffer. Increased concentrations were necessary to elute CCK-8-S and SP (20% propan-1-ol), since both peptides were not detected in runs up to 30min using lower concentrations of propan-l-01. Interestingly, these chromatographic conditions also gave good resolution of CCK-8 sulphated and non-sulphated forms. When 15% acetonitrile was substituted for propan-l-01 in the buffer system, retention times of both enkephalins were greater, SP and oxidised SP were not separated, and CCK-8-S and NT were not eluted in runs of up to 30min. In Figure 1A and B a typical run using either isocratic conditions or a linear gradient with 0.5 M TEAF buffer containing propan-l-01 is shown. When a shallow gradient was used (12-18% propan-l-01) for 25min the neuropeptide retention
204
NEUROPEP’IIDES
Fig. 1 Resolution of a peptide mixture of met-enk (1.7nmol), leu-enk (2nmol) and NT (0.6nmol) using 0.5 M TEAF buffer, pH 4, containing propan-l-01, flow rate l.OmI/min. Peptides were monitored with a UV detector (280nm). A: isocratic conditions, 14% propan-l-01; B: linear gradient 12-18% propan-l-01 for 25min, gradient starting at injection time.
times were markedly increased, compared to isocratic conditions (14% propan-l-01). But if slightly (13-20% propan-l-01, concentrations higher 20min) were used, the good resolution in peptide separation was lost (Fig. 2A and B), since some peptides appeared to co-elute. Reproducibility of the chromatographic conditions were further explored using 3H-peptides. Identical retention times to those for unlabelled peptides were obtained (Fig. 3A, B and C). The recovery was greater than 90%
Extracts from biological samples were also subjected to HPLC separation. Fractions corresponding to retention times for 3H-peptides were collected and freeze dried. Serial dilutions of that material gave curves which were essentially parallel to those obtained with the synthetic peptide (data not shown except for leu-enk, Fig. 4). Results from RIA’s performed on each fraction in a complete HPLC run are displayed in Figures 5 and 6. Locations of immunoreactivity peaks for met- and leu-enk, and NT coincided with those of
HPLC-RADIOIMMUNOASSAY
205
(RIA) FOR NEUROPEFTIDES
Fig. 2 Resolution of a peptide mixture of met-enk (l.Snmol), dynorphin 1-13 (OSnmol), leu-enk (lSnmol), met-enk-Arg-GlyLeu (0, l.Onmol), met-enk-Arg-Phe (H, l.Onmol) and neurotensin (NT, O.Snmol). Chromatographic conditions: OSM TEAF buffer, pH 4, containing propan-l-01, flow rate l.Oml/min. Peptides were monitored with a UV detector (280nm). A: isocratic conditions, 14% propan-I-01; B: linear gradient 13-20% propan-t-01 for 20min, starting at injection time.
the tritiated peptide standards chromatographed prior to the experiment. The plots of leu-enk immunoreactivity also have major peaks at the same retention time as met-enk, which is in accordance with the 5% cross-reactivity of the leu-enk antiserum with met-enk. In Table 2 the results obtained when measuring neuropeptide content in the crude extracts of rat or human brain, and after HPLC separation are shown. There is a disparity between the immuno-
reactivity levels in the crude extracts when compared with those reflecting authentic peptides as determined by HPLC. Discussion
The results obtained in this work demonstrate the viability of the HPLC method developed to separate and characterize several neuropeptides in complex mixtures.
206
Fig. 3 HPLC separation of a mixture of 3H-peptides. Chromatographic conditions: 0.5M pyridine-formic acid buffer, pH 4.0, containing propan-l-01, flow rate l.Oml/min. A: isocratic conditions, 14% propan-l-01, met-enk, leu-enk and NT. B: isocratic conditions, 20% propan-I-ol, CCK-8-S. C: isocratic conditions, 20% propan-l-01, SP. DPM - disintegrations per min.
NEUROPEPTIDES
We have developed a method based on isocratic conditions, avoiding the use of gradients, which have poor reproducibility and are slow to perform. The HPLC buffer used in this study includes small amounts of a tertiary amine, either triethylamine or pyridine, which overcomes the problem of poor resolution and long runs to resolve several neuropeptides. In previous reports of HPLC separations (4, 6, 10) tertiary amines were used in linear gradient experiments but isocratic conditions were not employed. Another strategy to successfully resolve a complex mixture of peptides is the addition of different organic modifiers, acetonitrile has been commonly-used, again in gradients, but propan-l-01 has also been used by others (2,3, 7). In our hands propan-l-01 has several advantages. Peptides are more rapidly eluted compared with similar or greater proportions of acetonitrile. Usually the recovery is also greater. Propan-l-01 is volatile and can be removed by freeze drying before samples are subjected to the RIA. The reproducibility of the method has been proven by the independence of the elution pattern of each peptide from the quantity and purity of the sample loaded. The recovery of the compounds being separated is higher than 90% for all the peptides tested, at least in the fmol-pmol range. The HPLC profiles of biological samples, with peaks coinciding with those of synthetic peptides, strongly confirms the suitability of the extraction procedure, with no oxidation or cleavage of the The competition curves peptides occurring. demonstrate that the authentic peptides were measured and the antisera used in this study appeared to be relatively selective for each peptide. The importance of this combined HPLC-RIA method is revealed by the disparity between the measurements of apparent peptide levels in crude brain tissue extracts (from both rat striatum and human caudate nucleus) and those of authentic peptides as determined after HPLC purification. The fact that values were not obtained for NT in crude extracts at the desired tissue concentration equivalent is probably attributable to non-specific interference in the assay, as previously experienced for direct NT measurements in cortical extracts (1). A high proportion of the RIA measurement attributed to a specific peptide may
HPLC-RADIOIMMUNOASSAY
207
(RIA) FORNEUROPEPTIDES
Extract Dilution
--e-
l/16
118
114
112
111
I
I
I
I
I
ioo-
80 -
60 -
40 -
20 -
0 1
I
I
10
100 fmol Leu-enk
Fig. 4 Semilogarithmic plot of tracer displacement leu-enk from rat striatum extract.
Table 2
Comparison
of neuropeptide
content
f 1000
Y
curves for leu-enk antiserum using synthetic leu-enk or HPLC-purified
in crude extracts
and HPLC purified
samples from discrete areas of
rat and human brain
Tissue
Rat striatum Human caudate nucleus
Sample
Extract HPLC Extract HPLC
Substance P
178.8 115.4 52.3 34.5
Neuropeptide levels (pmollg) leu-enk met-enk CCK-8-S
Neurotensin
1331.7 830.6 1038.0 797.6
ND 3.2 ND 2.6
85.9 59.4 35.6 11.7
43.5 22.4 15.1 8.6
Individual, paired crude extracts versus HPLC results represent RIA measurements of apparent versus authentic neuropeptide level in the same brain tissue sample. ND: non-detected at a reconstitution concentration of 2.6mg tissue/lOOuI RIA buffer.
NEUROPEPTIDES
pared to direct RIA of the crude extracts. However, the possibility of quantitating several neuropeptides from the same sample is an advantage considering that the amount of tissue available is sometimes limited (for example, human brain material).
Fig. 5 Immunoreactive profile of rat striatal extracts after HPLC separation. Retention times of 3H-peptides are also indicated. A: met-enk; B: leu-enk; C: NT.
actually reflect cross-reactive material. Reliance on simple RIA peptide measurements may thus give rise to mis-leading data interpretation. With regard to the application of the method to the measurement of peptides in biological samples, a previous separation of the compounds to be measured prolongs the experiments, com-
Fig. 6 Immunoreactive profile of human caudate nucleus extracts after HPLC separation. Retention times of 3Hpeptides are also indicated. A: met-enk; B: leu-enk; C: NT.
HPLC-RADIOIMMUNOASSAY
209
(RIA) FOR NEUROPEPTIDES
Acknowledgements The authors are greatly indebted to Dr E. Paralta for his helpful suggestions. This study was supported by the Medical Research Council and the Parkinson’s Disease Society.
5.
6.
References Biggins, J. A., McDermott, J. R., Smith, A. I., Candy, J. M., Perry, R. H., Bateman, D. E. and Edwardson, J. W. (1984). Chromatographic characterization of neuropeptides in post-mortem human brain. Journal of Neurological Science 66: 47-58. Frey, P. (1983). Cholecystokinin octapeptide levels in rat brain are changed after subchronic neuroleptic treatment. European Journal of Pharmacology 95: 87-92. Morris. H. R., Etienne, A. T., Dell, A. and Alburquerque, R. (1980). A rapid and specific method for the high resolution, purification and characterization of neuropeptides. Journal of Neurochemistry 34: 574-582. Shults. C. W., Yajima, H., Gullner, H.-G., Chase, T. N. and O’Donohue. T. L. (1985). Demonstration and distribution of kassinin-like material (substance K) in the rat
7.
8
9
10
central nervous system. Journal of Neurochemistry 45: 552-558. Emson, P. C., Goedert, M., Horsfield, P., Rioux. F. and St. Pierre, S. (1982). The regional distribution and chromatographic characterisation of neurotensin-like immunoreactivity in the rat central nervous system. Journal of Neurochemistry 38: 992-999. Akil, H., Ueda, Y., Lin, H. L. and Watson, S. J. (1981). A sensitive coupled HPLURIA technique for separation of endorphins: multiple forms of 8-endorphin in rat pituitary intermediate vs. anterior lobe. Neuropeptides 1: 429-446. Dawson, R. Jr., Steves, J. P., Lorden, J. F. and Oparil. S. (1985). Reverse-phase separation and electrochemical detection of neuropeptides. Peptides 6: 1173-l 178. Studier. J., Javoy-Agid, F., Cesselin, F., Legrand, J. C. and Agid, Y. (1982). CCK-S-immunoreactivity distribution in human brain: selective decrease in the substantia nigra from parkinsonian patients. Brain Research 280: 379-382. Taquet, H., Javoy-Agid, F., Hamon, M., Legrand, J. C. and Cesselin, F. (1983). Parkinson’s disease affects differently Met- and Leu-enkephalin in the human brain. Brain Research 280: 379-382. Beinfield, M. C. (1981). An HPLC and RIA analysis of the cholecystokinin peptides in rat brain. Neuropeptides 1: 203-209.
Isocratic Reverse-Phase HPLC Separation and RIA Used in the Analysis of Neuropeptides in Brain Tissue M. L. DE CEBALLOS*t,
M. D. TAYLOR*
and P. JENNER*
*Parkinson’s Disease Society Experimental Research Laboratories, Pharmacology Group, Biomedical Sciences Division, King’s College London, Manresa Road, London SW3 6LX, UK; tDepartment of Neuropharmacology, Cajal Institute, C.S.I.C., Doctor Arce 37, Madrid 28002, Spain (Correspondence and reprint requests to PJ)
Abstract-A reverse-phase, high-performance liquid chromatographic (HPLC) method was employed to separate and characterise five neuropeptides from complex mixtures, with important advantages over methods employed earlier using combined HPLC-RIA studies. Peptides were separated using 0.5M pyridine - 0.5M formic acid buffer, pH 4, containing propan-l-01 14% (met-enkephalin, leu-enkephalin, neurotensin) or 20% (CCK-S-S, substance at a flow rate of l.Oml/min. Isocratic conditions, and volatile solvents, resulted in a highly reproducible method, producing samples in a form designed for subsequent RIA. The application and importance of the procedure is demonstrated by comparison of the measurements of apparent peptide levels in crude brain extracts with those of authentic peptides as determined after HPLC purification.
Introduction
Radioimmunoassay (RIA) measurements of the peptide content of brain extracts has several disadvantages. RIA has the appropriate sensitivity, but the lack of totally specific antisera raises the possibility of cross-reactivity with related peptides and precursors/metabolites. The low levels of some neuropeptides may require the use of high tissue concentrations for their quantitation raising
Date received 26 June 1991 Date accepted 12 August 1991
P)
the possibility of non-specific interference with the RIA procedure (1). so, a separation procedure prior to the quantification of individual neuropeptides is desirable. High performance liquid chromatography (HPLC) has been applied successfully to peptides separation. Some methods involve gradient conditions (l-6) and salt buffer systems (2,5,6, 7). The former has two disadvantages, the lack of reproducibility of chromatographic runs and slowness in the processing of the biological samples. Salt buffers can also produce interference in the subsequent RIA used for peptide quantitation.
201
202
NEUROPEPTIDES
Isocratic HPLC conditions using highly volatile solvents were employed in the present study to separate a number of neuropeptides: met- and leu-enkephalin (met- and Ieu-enk), neurotensin (NT), cholecystokinin (CCK-8-S) and substance P (SP), which were measured using specific RIA’s. These neuropeptides were selected because of their localisation in basal ganglia, and the changes in content observed in neurodegenerative disorders, such as Parkinson’s disease (see for example 8, 9). The reproducibility of the method was assessed using either unlabelled or tritiated peptides. The solvent system was selected for the high recovery of the compounds subjected to the procedure, and its compatibility with the RIA used for their measurement. Finally, when the method was applied to biological samples comparison was made between measurements from the crude extracts and HPLC purified samples from discrete areas of rat and human brain.
Methods High performance
liquid chromatography
The HPLC system employed consisted of two Perkin Elmer (PE) series 10 liquid chromatographs, commanded by a PE LC series 10 prodetector (PE LC 75 grammer, a UV spectrophotometric detector) and a LKB model 2212 Helirac programmable fraction collector. Chromatographic data were plotted and integrated by a PE 3600 Data Station. Reverse phase separations were carried out using a Spherisorb 5 ODS 2 column (5~., 25 x 0.4cm, Phase Separations), protected with a guard column Spherisorb 10 ODS 2 (lop., 5 X 0.4cm, Phase Separations). All the reagents used were of the highest purity. Buffers were filtered through Millipore filters (HVLP, 0.45 l.~)and degassed before use. In most experiments synthetic and brain tissue-derived peptides were separated using 0.5M pyridine0.5 M formic acid buffer, pH 4, containing propanl-01 14% (met-enk, leu-enk, NT) or 20% (CCK-S-S, substance P) at a flow rate of 1 ml/min. When synthetic peptides were monitored with the UV detector, pyridine was substituted for triethylamine (TEA) to circumvent the high UV absorbance of pyridine. When linear gradients
were used the actual conditions are expressed where appropriate. All HPLC procedures were performed at room temperature (25 f 3°C). Synthetic met-enkephalin, leu-enkephalin, neurotensin CCK-8-S, CCK-8-N& substance P and dynorphin 1-13 (Sigma) were dissolved in oxygenfree HPLC-grade water, and aliquots were stored at -20°C. Peptides were further diluted in HPLC buffer before injection onto the HPLC column. Tritiated peptides were similarly treated and stored at 4°C until used. Subsequently, amounts between 30-1OOfmol were subjected to HPLC separation. Tritiated peptides were from Amersham International, except 3H-neurotensin which was supplied by New England Nuclear. The eluent was monitored at 254-280nm (nmol sensitivity) or subjected to RIA, after freeze-drying the solvent (fmol sensitivity). In the case of radiolabelled peptides (fmol sensitivity), fractions were counted in a Beckman scintillation counter. To minimise possible contamination of subsequent runs, synthetic peptides were always chromatographed at the end of a series of experiments and the chromatographic system thoroughly washed by pumping 70% propan-l-01 through the chromatograph. Extraction of brain peptides
Male Wistar rats (200-25Og) were killed by cervical dislocation and decapitation, and the brain quickly dissected on a cold plate. Dissected brain areas were immediately frozen on dry ice. The samples were boiled for 15 min in 100 volumes of a mixture of 1 N acetic acid and 0.02 N HCI, containing 0.1% of 2-mercaptoethanol. Samples were homogenized with a Polytron (setting 5, for 5s), and centrifuged (12000 x g, 10min). Aliquots of the supernatant (equivalent 5 mg of wet weight tissue) were freeze dried and stored at -40°C until subjected to HPLC. Samples were reconstituted in HPLC buffer, microfuged for 3min and injected onto the column. Samples for CCK-8-S measurements were boiled in distilled water for 5min, homogenized and centrifuged as above. These extraction methods were also used for human brain tissue samples obtained at post-mortem from normal subjects. In preliminary experiments the possible formation of oxidised peptides, during homogeni-
HPLC-RADIOIMMUNOASSAY
203
(RIA) FOR NEUROPEPTIDES
sation of brain tissue, was assessed. In the presence of 2-mercaptoethanol no oxidised peptides were formed. There was only one peak, corresponding to met-enk, before or after sonication, as monitored by UV detection. Radioimmunoassays
Specific RIA’s were used for peptide quantitation. HPLC eluates containing individual peptides were reconstituted in the appropriate volume of buffer (50mM sodium phosphate buffer containing 0.2% gelatin and 10mM EDTA for the enkephalins, CCK-8-S and SP) pH 7.2. For NT RIA the assay buffer was 60mM sodium phosphate, pH 7.2, containing 0.3% bovine serum albumin and 10mM EDTA. An aliquot of samples or standards (100~1) was incubated with 100~1 of antisera (Amersham International) and ‘251-peptide (15 OOOcpm; Amersham International). The tubes were incubated at 4°C for 22-24h for met-enk, leu-enk and NT, and for up to 3 days CCK-8-S and SP to achieve appropriate sensitivity. Separation of bound and free peptide was achieved adding 250-500~1 of activated charcoal dextran in RIA buffer. After centrifugation (5000 x g, 10min) the supernatant was decanted, and counted in a LKB Minigamma counter for 1 min.
Table 1
Effects of different chromatographic conditions on the retention time of various neuropeptides Peptide
ox met-enk
met-enk leu-enk dyn 1-13 dyn l-8 ocrapeptide heptapeptide CCK-8-S CCK-8-NS ox SP SP neurotensin
A
6
c
D
5 7 15 12
5 7 14 9.6 14 16.3
5 7 13 8
9 14
ND ND ND
11 16 7 10 10
10.6 12.7 ND ND ND ND 19
ND
17
ND 6 6
ND
Chromatographic conditions: 0.5M TEAF buffer. pH 4, flow rate 1.Omi/min. A: + 12% propan-l-01; B: + 15% propan-l-01: C: +20% propan-l-01; D: + 15% acetonitrile. ND: not detected in runs up heptapeptide:
to 30min. octapeptide: met-enk-Arg-Gly-Leu: met-enk-Arg-Phe; ox: oxidised.
Freeze dried aliquots of tissue extracts were assayed with and without HPLC separation (the equivalent of 4mg of tissue being reconstituted in the same volume of RIA buffer in both cases). Cross reactivities of the antisera were as follows: the antisera did not display any cross-reactivity with dynorphin 1-13, met-enk-Arg-Phe, met-enkArg-Gly-Leu or CCK-8-S. Enkephalin antisera did not cross-react with antisera against NT, or vice versa (0.1%). Met-enk antisera cross-reacted to 6% with met-enk. The antiserum against SP had no detectable cross-reactivity with the following peptides: met-enk-Arg-Phe, met-enk-Arg-GlyLeu, eledoisin, physaelemin, neurokinin A, neurokinin B and only negligible cross-reactivity (0.02%) for kassinin. The RIA sensitivities (15% displacement of bound tracer) were 1.5fmol/tube for both enkephalins, 1fmoVtube for NT, 0.6fmoVtube for CCK-8-S and 1 fmol/tube for SP.
Results
The effects of varying the percent propan-l-01 on neuropeptide retention times are shown in Table 1. Several peptides were separated with 0.5M triethylamine-formate buffer (TEAF). pH 4, containing either 12% or 14% propan-l-01. Increasing concentrations of propan-l-01 reduced neuropeptide retention times. NT and dynorphin 1-13 were more sensitive to the amount of propanol contained in the buffer. Increased concentrations were necessary to elute CCK-8-S and SP (20% propan-1-ol), since both peptides were not detected in runs up to 30min using lower concentrations of propan-l-01. Interestingly, these chromatographic conditions also gave good resolution of CCK-8 sulphated and non-sulphated forms. When 15% acetonitrile was substituted for propan-l-01 in the buffer system, retention times of both enkephalins were greater, SP and oxidised SP were not separated, and CCK-8-S and NT were not eluted in runs of up to 30min. In Figure 1A and B a typical run using either isocratic conditions or a linear gradient with 0.5 M TEAF buffer containing propan-l-01 is shown. When a shallow gradient was used (12-18% propan-l-01) for 25min the neuropeptide retention
204
NEUROPEP’IIDES
Fig. 1 Resolution of a peptide mixture of met-enk (1.7nmol), leu-enk (2nmol) and NT (0.6nmol) using 0.5 M TEAF buffer, pH 4, containing propan-l-01, flow rate l.OmI/min. Peptides were monitored with a UV detector (280nm). A: isocratic conditions, 14% propan-l-01; B: linear gradient 12-18% propan-l-01 for 25min, gradient starting at injection time.
times were markedly increased, compared to isocratic conditions (14% propan-l-01). But if slightly (13-20% propan-l-01, concentrations higher 20min) were used, the good resolution in peptide separation was lost (Fig. 2A and B), since some peptides appeared to co-elute. Reproducibility of the chromatographic conditions were further explored using 3H-peptides. Identical retention times to those for unlabelled peptides were obtained (Fig. 3A, B and C). The recovery was greater than 90%
Extracts from biological samples were also subjected to HPLC separation. Fractions corresponding to retention times for 3H-peptides were collected and freeze dried. Serial dilutions of that material gave curves which were essentially parallel to those obtained with the synthetic peptide (data not shown except for leu-enk, Fig. 4). Results from RIA’s performed on each fraction in a complete HPLC run are displayed in Figures 5 and 6. Locations of immunoreactivity peaks for met- and leu-enk, and NT coincided with those of
HPLC-RADIOIMMUNOASSAY
205
(RIA) FOR NEUROPEFTIDES
Fig. 2 Resolution of a peptide mixture of met-enk (l.Snmol), dynorphin 1-13 (OSnmol), leu-enk (lSnmol), met-enk-Arg-GlyLeu (0, l.Onmol), met-enk-Arg-Phe (H, l.Onmol) and neurotensin (NT, O.Snmol). Chromatographic conditions: OSM TEAF buffer, pH 4, containing propan-l-01, flow rate l.Oml/min. Peptides were monitored with a UV detector (280nm). A: isocratic conditions, 14% propan-I-01; B: linear gradient 13-20% propan-t-01 for 20min, starting at injection time.
the tritiated peptide standards chromatographed prior to the experiment. The plots of leu-enk immunoreactivity also have major peaks at the same retention time as met-enk, which is in accordance with the 5% cross-reactivity of the leu-enk antiserum with met-enk. In Table 2 the results obtained when measuring neuropeptide content in the crude extracts of rat or human brain, and after HPLC separation are shown. There is a disparity between the immuno-
reactivity levels in the crude extracts when compared with those reflecting authentic peptides as determined by HPLC. Discussion
The results obtained in this work demonstrate the viability of the HPLC method developed to separate and characterize several neuropeptides in complex mixtures.
206
Fig. 3 HPLC separation of a mixture of 3H-peptides. Chromatographic conditions: 0.5M pyridine-formic acid buffer, pH 4.0, containing propan-l-01, flow rate l.Oml/min. A: isocratic conditions, 14% propan-l-01, met-enk, leu-enk and NT. B: isocratic conditions, 20% propan-I-ol, CCK-8-S. C: isocratic conditions, 20% propan-l-01, SP. DPM - disintegrations per min.
NEUROPEPTIDES
We have developed a method based on isocratic conditions, avoiding the use of gradients, which have poor reproducibility and are slow to perform. The HPLC buffer used in this study includes small amounts of a tertiary amine, either triethylamine or pyridine, which overcomes the problem of poor resolution and long runs to resolve several neuropeptides. In previous reports of HPLC separations (4, 6, 10) tertiary amines were used in linear gradient experiments but isocratic conditions were not employed. Another strategy to successfully resolve a complex mixture of peptides is the addition of different organic modifiers, acetonitrile has been commonly-used, again in gradients, but propan-l-01 has also been used by others (2,3, 7). In our hands propan-l-01 has several advantages. Peptides are more rapidly eluted compared with similar or greater proportions of acetonitrile. Usually the recovery is also greater. Propan-l-01 is volatile and can be removed by freeze drying before samples are subjected to the RIA. The reproducibility of the method has been proven by the independence of the elution pattern of each peptide from the quantity and purity of the sample loaded. The recovery of the compounds being separated is higher than 90% for all the peptides tested, at least in the fmol-pmol range. The HPLC profiles of biological samples, with peaks coinciding with those of synthetic peptides, strongly confirms the suitability of the extraction procedure, with no oxidation or cleavage of the The competition curves peptides occurring. demonstrate that the authentic peptides were measured and the antisera used in this study appeared to be relatively selective for each peptide. The importance of this combined HPLC-RIA method is revealed by the disparity between the measurements of apparent peptide levels in crude brain tissue extracts (from both rat striatum and human caudate nucleus) and those of authentic peptides as determined after HPLC purification. The fact that values were not obtained for NT in crude extracts at the desired tissue concentration equivalent is probably attributable to non-specific interference in the assay, as previously experienced for direct NT measurements in cortical extracts (1). A high proportion of the RIA measurement attributed to a specific peptide may
HPLC-RADIOIMMUNOASSAY
207
(RIA) FORNEUROPEPTIDES
Extract Dilution
--e-
l/16
118
114
112
111
I
I
I
I
I
ioo-
80 -
60 -
40 -
20 -
0 1
I
I
10
100 fmol Leu-enk
Fig. 4 Semilogarithmic plot of tracer displacement leu-enk from rat striatum extract.
Table 2
Comparison
of neuropeptide
content
f 1000
Y
curves for leu-enk antiserum using synthetic leu-enk or HPLC-purified
in crude extracts
and HPLC purified
samples from discrete areas of
rat and human brain
Tissue
Rat striatum Human caudate nucleus
Sample
Extract HPLC Extract HPLC
Substance P
178.8 115.4 52.3 34.5
Neuropeptide levels (pmollg) leu-enk met-enk CCK-8-S
Neurotensin
1331.7 830.6 1038.0 797.6
ND 3.2 ND 2.6
85.9 59.4 35.6 11.7
43.5 22.4 15.1 8.6
Individual, paired crude extracts versus HPLC results represent RIA measurements of apparent versus authentic neuropeptide level in the same brain tissue sample. ND: non-detected at a reconstitution concentration of 2.6mg tissue/lOOuI RIA buffer.
NEUROPEPTIDES
pared to direct RIA of the crude extracts. However, the possibility of quantitating several neuropeptides from the same sample is an advantage considering that the amount of tissue available is sometimes limited (for example, human brain material).
Fig. 5 Immunoreactive profile of rat striatal extracts after HPLC separation. Retention times of 3H-peptides are also indicated. A: met-enk; B: leu-enk; C: NT.
actually reflect cross-reactive material. Reliance on simple RIA peptide measurements may thus give rise to mis-leading data interpretation. With regard to the application of the method to the measurement of peptides in biological samples, a previous separation of the compounds to be measured prolongs the experiments, com-
Fig. 6 Immunoreactive profile of human caudate nucleus extracts after HPLC separation. Retention times of 3Hpeptides are also indicated. A: met-enk; B: leu-enk; C: NT.
HPLC-RADIOIMMUNOASSAY
209
(RIA) FOR NEUROPEPTIDES
Acknowledgements The authors are greatly indebted to Dr E. Paralta for his helpful suggestions. This study was supported by the Medical Research Council and the Parkinson’s Disease Society.
5.
6.
References Biggins, J. A., McDermott, J. R., Smith, A. I., Candy, J. M., Perry, R. H., Bateman, D. E. and Edwardson, J. W. (1984). Chromatographic characterization of neuropeptides in post-mortem human brain. Journal of Neurological Science 66: 47-58. Frey, P. (1983). Cholecystokinin octapeptide levels in rat brain are changed after subchronic neuroleptic treatment. European Journal of Pharmacology 95: 87-92. Morris. H. R., Etienne, A. T., Dell, A. and Alburquerque, R. (1980). A rapid and specific method for the high resolution, purification and characterization of neuropeptides. Journal of Neurochemistry 34: 574-582. Shults. C. W., Yajima, H., Gullner, H.-G., Chase, T. N. and O’Donohue. T. L. (1985). Demonstration and distribution of kassinin-like material (substance K) in the rat
7.
8
9
10
central nervous system. Journal of Neurochemistry 45: 552-558. Emson, P. C., Goedert, M., Horsfield, P., Rioux. F. and St. Pierre, S. (1982). The regional distribution and chromatographic characterisation of neurotensin-like immunoreactivity in the rat central nervous system. Journal of Neurochemistry 38: 992-999. Akil, H., Ueda, Y., Lin, H. L. and Watson, S. J. (1981). A sensitive coupled HPLURIA technique for separation of endorphins: multiple forms of 8-endorphin in rat pituitary intermediate vs. anterior lobe. Neuropeptides 1: 429-446. Dawson, R. Jr., Steves, J. P., Lorden, J. F. and Oparil. S. (1985). Reverse-phase separation and electrochemical detection of neuropeptides. Peptides 6: 1173-l 178. Studier. J., Javoy-Agid, F., Cesselin, F., Legrand, J. C. and Agid, Y. (1982). CCK-S-immunoreactivity distribution in human brain: selective decrease in the substantia nigra from parkinsonian patients. Brain Research 280: 379-382. Taquet, H., Javoy-Agid, F., Hamon, M., Legrand, J. C. and Cesselin, F. (1983). Parkinson’s disease affects differently Met- and Leu-enkephalin in the human brain. Brain Research 280: 379-382. Beinfield, M. C. (1981). An HPLC and RIA analysis of the cholecystokinin peptides in rat brain. Neuropeptides 1: 203-209.