Microdialysis of extracellular endogenous opioid peptides from rat brain in vivo

Neuroscience

Vol. 33, No. 3, pp. 349-337, 1989 Printed in Great Britain

0306-4322/89 $3.00 + 0.00 Pergamon Press plc 0 1989 IBRO

MICRODIALYSIS OF EXTRACELLULAR ENDOGENOUS OPIOID PEPTIDES FROM RAT BRAIN 1iV G’lV0 N. T. MAIDMENT,* D. R. BRUMBAUGH,V. D. RUDOLPH, E. ERDELYI and C. J. EVANS Pritzker Laboratory, Department of Psychiatry and Behavioral Sciences, Stanford University Medical Center, Stanford, CA 94305, U.S.A. Abstract-The combination of microdialysis and a highly sensitive radioimmunoassay was developed in order to monitor the in vivo extracellular levels of endogenous opioid peptides from discrete regions of the rat brain. The radioimmunoassay cross-reacts 100% with peptides with alpha N-acetyl Tyr.Gly.Gly.Phe-Met or -Leu at the N terminus and thus recognizes all known endogenous opioid peptide fragments following acetylation of the sample. The assay was conducted on solid phase with antibody bound via protein A to 96-well plates and provided a limit of detection of approximately 0.2 fmol. A variety of dialysis membranes were evaluated with respect to their eficiency in recovering opioid peptides in vitro. Custom-made probes (4 mm active length) manufactured from polyacrylonit~le membranes and commercially available polycarbonate membrane probes proved most suitable with relative recoveries for [Met]- and [Leulenkephalin in the range 610% at a flow rate of 2.7 pl/min. Probes implanted in the globus pallidus/ventral pallidum of halothane/N,O anaesthetized rats recovered approximately 1.5 fmol of immunoreactive opioid material per 30-min sample in the absence of peptidase inhibitors. The majority of this immunoreactivity co-eluted with [Met]- and [Leulenkephalin on reverse-phase high-performance liquid chromato~aphy. A Z-min pulse of 1OOmM K+-containing artificial cerebrospinal fluid in the perfusion medium during a 30-min sampling period increased the recovered immunoreactive material to 43.9 fmol + 12.4 S.E.M. A second stimulation 3 h later also resulted in elevated levels with an S2:Sl ratio of 0.64 + 0.03. The second stimulation was completely blocked by perfusion of a 1OmM EGTA-containing medium, basal release on average remaining unaffected. Repeated K+ stimulations (SOmM) at l-h intervals resulted in progressively diminishing elevations in opioid peptide release. Incorporation of veratridine (50 PM) in the perfusion medium for 2 min of a 30-min collection period increased the release to 16.8 fmol & 0.5 with an 52: Sl ratio of 0.62 & 0.17 after a 3-h interval. The second stim~ation was prevented by simultaneous perfusion of a 2 FM tetrodotoxin~ontaining medium, basal release remaining unaffected. The results demonstrate the potential of this technique for studying the regulation of opioid peptide release and metabolism from discrete regions of the.rat brain in vivo.

The majority of studies directed towards the analysis of endogenous opioid peptide release in the brain have been carried out in vitro using tissue slice preparations,1Z~*S~‘7~21~24~26~28~3’ Although this approach continues to provide important informati& on the release characteristics of these peptides, Ifurther insights into the regional regulation of their release and metabolism and interactions with other nenrotransmitter systems would be made’ poisible by the development of appropriate in z&o technology. The relatively few in oiuo studies carried out to date have employed push-pull cannulae.2*3,7 However, the advent of microdialysis has provided a preferred method for the in uiuo sampling of small molecular weight molecules from the brain- since it avoids exposing the tissue to the trauma of fast flowing fluids associated with push-pull perfusion.32 Although

*To whom correspondence should be addressed. ~~~rev~urj~~~: Arg, arginine; BSA, bovine serum albumen; CSF, cerebrospinal fluid; EGTA, ethyleneglycol tetraacetic acid; Gly, glycine; HPLC, ‘high-performance liquid chromatography; Leu, leucine; Met, methionine; Phe, phenylalanine; RIA, radioimmunoassay.

mi~rodialysis is now established as a method of choice for monitoring the release and metabolism of classical neurotransmitters,8~‘a~16~20~29~30~38 its application to the analysis of neuropeptide release has proven more difficult due, in part, to their low extracellular concentrations. An additional problem is the tendency for many peptides to adsorb nonspecifically to various polymers and hence recovering them across the dialysis membrane and through connecting tubing is a potential problem. Recent dialysis studies of neurokinin and substance P in the brain and spinal cord reported recovery of only sub-fmol amounts of immunoreactive material in 30:min samples even in the presence of a peptidase inhibitor.6*18.1q In the present study we examine the properties of several different polymeric membranes with regard to their ability to recover opioid peptides in vitro. The sensitivity issue is approached by developing a radioimmunoassay (RIA) that recognizes all the opioid active fragments of the three known endogenous opioid precursors namely, pro-opiomelanocortin, pro-enkephalin and pro-dynorphin with equal cross reactivity. Since both pro-enkephalin and 549

550

N. .F. MAWMENTC/ ttf

pro-dynorphin have multiple copies of the opioid core sequence Tyr-Gly-Gly-Phe-X the probability of detecting extracellular levels is maximized and interfacing this assay with reverse-phase highperformance liquid chromatography (HPLC) enables individual opioid peptides in the dialysates to be identified. We describe the successful application of this strategy to the monitoring of extracellular opioid peptides from the rat brain in vim Our initial experiments have focussed on the striatal-pallidal enkephalinergic pathway, the globus pal~idus/ventral pallidurn representing a region densely populated with enkephalinergic termina1s.9~27.33This pathway has previously been the focus of extensive biochemical and pharmacological studies’.“.37 and the major endogenous opioid peptides that have been identified in this area are derived from pro-enkephalin and include [Metlenkephalin, [Leu]enkephalin, [Metlenkephalin Arg.Gly.Leu, [Metlenkephalin Arg.Phe, Barn 18 and metorphamide. Opioid peptides derived from prodynorphin are present at approximately IO-fold lower concentrations. EXPERIMENTAL PROCEDURES Dialysis probe preparation

Dialysis probes were constructed to a concentric cannula design similar to that described previously” except that fused silica tubing (75 pm i.d., 150pm o.d., Polymicrotechnology) was used for the internal cannula. The outer cannula was manufactured from either 23-gauge or 24gauge thin-wall stainless steel tubing (Small Parts). The active part of the probes consisted of a 4-mm length of hollow tubular dialysis membrane sealed at the tip with epoxy resin. The following types of membrane were used: cellulose (Gambro), cellulose acetate (Cordis Dow), cuprophan (Enka Glanzstofl), polyacrylonitrile (Hospal) and polysulphone (Amicon). These probes were compared with the similarly constructed co~ercially avaifable polycarbonate variety (Carnegie Medecin~~ioanalyticai Systems). ArtificiaI cerebrospinal fluid (CSF), consisting of the following: glucose, 5mM; NaCl, l25mM; KCl, 2.5mM; NaHCO,, 27 mM; NaH,PG,, 0.5 mM; Na2HPG,, 1.2mM; CaCl, , 1.2 mM; MgCl,, 1 mM; ascorbic acid, 0.1 mM; bovine serum albumin (BSA). 0.025%. was bubbled with OJCO, (95%/5%) to pH 7.4 and filter’sterihzed. This was stored in sterile containers for a maximum of one week, aliquots being removed under a laminar flow hood before re-gassing to eliminate the risk of bacterial contamination of the stock solution. Such a precaution was considered necessary since bacteria are known to produce peptidases capable of degrading the already low levels of opioid peptides in the in uivo samples (for the same reason all perfusion lines were routinely flushed with 75% ethanol prior to each in vivo experiment). In the high potassium-containing artificial CSF NaCl concentrations were reduced accordingly. Similarly, CaCl, was removed and replaced with additional MgCl, together with EGTA (10mM) for experiments examining the calcium-dependency of the release. In vitro recovery experiments Probes, perfused continuously with artificial CSF, were immersed in 200~~1 or 5-ml sdlmions of opioid peptides dissolved in artificial CSF and in some cases ‘251-labeled

[Leu]enkephahn was also added. Reiative recoverres were therefore determined either by RIA or with a gamma counter. Factors examined included flow rate through the probe, concentration and type of opioid peptide in the bathing medium and time required to reach equilibrium. All experiments were conducted at room temperature. in vivo perfusion system Artificial CSF, contained in 5-ml gas-tight syringes (Hamilton), was delivered to the dialysis probes via one, or in some experiments two in series, HPLC injection valve(s) (Rheodyne) at a rate of 2.7 pl/min using a slow infusion pump (Harvard). One valve served as a simple liquid switch, allowing rapid changing of the perfusion medium for long periods using a second syringe. The second valve was equipped with a sample loop (5 or 75~1) to enable bolus infusion of drug-containing artificial CSF for approximately 2 or 30min, respectively. Both the connection between injection valve and dialysis probe inlet and the probe outlet employed narrow-bore tubing (120 Mm i.d., Carnegie Medecin/Bioanalyticai Systems) to minimize dead space. The time delay between drug injection and its appearance at the probe outlet was approximately 5 min. Surgical procedure

Male Sprague-Dawley rats (250-4OOg, Simonsen) were anesthetized with 223% halothane in O,/N,O (1:l) and placed in a stereotaxic frame (David Kopf). A hole was drilled in the skull and a continually perfused 4-mm dialysis probe (Carnegie M~~in~Bioanalytical Systems) slowly lowered into the brain over an approximate lo-min period. The coordinates used for the tip of the probe from bregma were: globus pallidus/ventral pallidum rostra1 .- I .Omm, lateral + 3.0 mm, ventral 9.0 mm.25 The halothane con~ntration was then reduced to approximately 1% for the remainder of the experiment. Sample collection commenced at 20-min, 30-min or lh intervals approximately 1 h following implantation. Samples were collected in Eppendorf tubes on ice and stored at -80°C prior to assay. Dialysis probes were removed at the end of the experiment and in some cases were used one or two more times following washing and storage in distilled water. Verl~cal~on of d~~ys~s probe placement

Rats were killed at the end of each experiment by anesthetic overdose, their brains removed and rapidly frozen (Accufreeze) and stored at - 80°C. Brain slices (30 pm) were cut in a freezing microtome, stained with Cresyl Violet and examined under a light microscope. Universal opioid plate radioimmunoassay

Antisera were raised to alpha N-acetyl B-endorphin in New Zealand white rabbits using established procedures.” Although the broad principle of the universal RIA for the analysis of rat brain extracts has been described previou~ly,~ a different antiserum and assay procedure were employed for the present study. The dilution of the antiserum selected for use in the plate RIA was approximately 1: 200,000. ruination

procedure

Alpha N-acetyl alpha endorphin iodinated via the iodogen method was used as a tracer in the assay. Borosilicate 12 x 75 mm glass tubes were coated with iodogen (Pierce) by adding 100 pl of a solution containing 1Opg iodogen dissolved in methylene chloride and evaporating the solvent under a stream of nitrogen. The coated tubes were stored at 4°C prior to the iodination. Carrier-free iodine (New England Nuclear), 100 FCi dissolved in 10 ~10.1 M NaOH, was added to an iodogen-precoated tube containing 1 gg alpha IV-acetyl alpha endorphin dissolved in 50 ~1 0.1 M KPO,, pH 7.0. After 10min the reaction mixture was diluted to 1 ml with 5% acetic acid then loaded directly onto

551

Microdialysis of opioid peptides in rat brain a 4.6 x 250 mm Ultrasphere ODS reverse-phase column pre-equilibrated in aqueous buffer consisting of 0.1 M KPG, adjusted to pH 2.7 with phosphoric acid. The peptides were eluted at a flow rate of 1 ml/mm with a two-step linear acetonitrile gradient comprising O-l 5% acetonitrile in 5 min followed by 15-50% acetonitrile in 60 min. Iodinated peptide eluting at approximately 31 min was used as a tracer in the assay. Acetylationprocedure Dialysis samples and standard con~ntrations of [Met]enkephalin were acetylated by addition of 5 ~1 acetic anhydridd and immediately vortexed. Standard [Met~nkepha~n solutions were made UD in artificial CSF used in the corresponding in viuo expehment previously stored at -80°C together with the samples. For the experiments involving veratridine a second standard curve was constructed using similarly stored veratridine-containing artificial CSF since

this solution occasionally caused some interference in the assay. Samples and standards were then dried in a “speed vat” and the residue re-dissolved in 100 ~1 6% ammonium hydroxide in order to hydrolyse remaining acetic anhydride and remove any phenolic acetyl esters caused by reaction of acetic anhydride with the side chain of tyrosine. The ammonia and ammonium acetate were then removed by ‘*speed vat” evaporation and the dried samples re-dissolved in 50 ~1 RIA buffer consisting of 0.15 M K,HPC&, 0.1% gelatin and 0.2% Tween 20 adjusted to pH 7.4. Assay procedure The solid phase of the assay was prepared by adding 0.1 pg Protein-A (Sigma, binding capacity 9-l 1 mg of human IgG per mg) dissolved in 100 ~1 0.1 M sodium bicarbonate pH 9 to Immulon 2 removable wells (VWR). The plates were generally stored in this state at 4°C for up to two weeks but could be used after incubation for 2 h at room temperature. The protein-A solution was dumped and the wells washed three times with RIA buffer. Immediately, 50 ~1 of the appropriate concentration of antiserum diluted in RIA buffer was added to each well, the dilution determined previously to provide approximately 20-40% binding. After 2 h incubation at room temperature, the antibody solution was dumped and the plates washed three times with RIA buffer, Samples and standards (SO~I), previously acetylated as described above, were added to the plates and allowed to preincubate with the antibody for 2 h at room temperature. After pre-incubation, 50~1 of RIA buffer containing approximately 5000 cpm ‘251-labeled trace was added to each well and the plates allowed to incubate overnight at 4°C. After washing three times with RIA buffer, the wells were separated and placed into 12 x 75 mm borosilicate tubes and counted for 4 min in a four-channel gamma counter. High -performance liquid chromatography ident&ation of recovered ~m~oreuctivity Samples of basal release of opioid peptide immunoreactivity were collected over a 10-h period. Each 1-h sample (160 ~1) was acidified by addition of 10~1 80% acetic acid. A proportion was removed from each tube for direct assay. The remainder was combined and subjected to HPLC analysis. Potassium (100 mM)-evoked release was collected for 1h and treated in the same way. A standard solution of Tyr.Gly.Gly.Phe, [Metlenkephalin, [Leujenkephalin, [MetJenkephalin Arg.Phe and [Metlenkephalin Arg.Gly.Leu was similarly injected onto the HPLC system. Reverse-phase chromatography was performed on 2 x 150mm columns, slurry-packed with 5-#m-wide-pore Hypersil CS (Shandon) using a Shandon column packer. Two identical columns with associated HPLC iniectors (Rheodvne) were olaced in parallel, thereby dividing the Row equally and ‘allowing semi-microbore chromato~aphy using the Altex 1lOA pumps with a flow rate through each column of 250 nl/min. This system had the advantage that standards could

be restricted to one injector and column to develop the separation gradient using U.V. detection. In this way contamination of the sample column could be avoided. Subsequently, as described above, low fmol quantities of standards were run on the sample column to confirm precise elution profiles. The buffer used consisted of 10mM KH,PO, adjusted to pH 2.8 with H,PG, and peptides were eluted with an acetonitrile gradient as illustrated in Fig. 2. Pyridine (27 ~1) was added to each l-min fraction, which was subsequently prepared for RIA as described above. Slat~sti~a~ analysis The analysis of the effect of EGTA on the K+-evoked release was carried out using a one tailed, unpaired t-test. The difference between the fmol amount of S2 and the corresponding baseline value was compared between the control and EGTA-treated groups. The same test was executed on S 1, The baseline value used for each animal was the mean of the three values immediately prior to the relevant stimulation. An identical procedure was employed to analyse the effect of tetrodotoxin on veratridinestimulated release. To assess the depletable nature of repeated K+ stimulations a one-tailed, paired r-test was used in which the difference between baseline and stimulation values was compared between stimulations. Baseline values were calculated as described above. RESULTS Plate radioimmunoassay reactivity

sensitivity

and

cross-

The mean and S.E.M. IC,, lcso and lcBo values for [Metlenkephalin carried through the acetylation procedure were 0.4 & 0.05, 1.5 + 0.2 and 9 &-2 fmol, respectively. A representative displacement curve demonstrating the low intra-assay variation is shown in Fig. 1. The cross-reactivity at the ICY and ICY was 100% for acetylated [~u~nkephalin, [Met]enkephalin Arg.Phe, [Metlenkephaiin Arg.GIy.Leu, j?-endorphin, dynorphin 1-8, alpha neo-endorphin, metorphamide and BAM 18 but only 20 and 7%, respectively, for acetylated Tyr.Gly.Gly.Phe. Less than 0.01% cross-reactivity was detected with acetylated Tyr.Gly.Gly and non-acetylated peptides with Tyr.Gly.Gly.Phe.Met or Leu at the N-terminus (data not shown). 1001

80%

60-

x p

40.

m

20 0 1

) .l

.

.

.._...

_ 1

met-enk

IYIYQ?L 10

100

(fmollwell)

Fig. 1. A typical displacement curve for [MetJenkephalin carried through the acetylation procedure demonstrating an tc,of approGmately 1.5 fmol. Standards were conducted in quadruplet and the points represent the mean and S.E.M. in each case. For several points the S.E.M. is too small to be clearlv* renresented. ‘

N. T. MAWMEW rf d

Table 1. Continuously perfused (2.7 pl:min) dialysis probes (4 mm active length) were immersed in 200 )x1of a 1nM [Leulenkephalin solution to which approx. I 0,000 cpm of high-performance liquid chromatography-puri~ed [iz51][Leu]enkephal~n were added Membrane Cellulose Cellulose acetate Cuprophan (6.5Frn wall) Cuprophan (19-pm wall) Polya~rylonitrile Polycarbonate

MW cut-off

O.D. (pm)

Recovery (%)

5000 70,000 12,000 12,000 40,000 20,000

250 230 215 320 300 500

3.2+0.1 2.6 * 0.2 5.6 + 0.6 2.6 * 0.2 8.0 2 0.7 6.4 & 0.8

After a 30-min equilibration period two successive 30-min samples were collected. These samples were counted in a gamma counter together with equal volumes of the external medium and the recovery expressed as a percentage. Each relative recovery value represents the mean and standard error of four probes. Recovery

in vitro

The results of the dialysis membrane comparison with regard to ‘Z51-labeled [Leulenkephalin relative recovery are shown in Table 1. The polysulphone membranes were found to be unsuitable for the probe design used in this study, ultrafiltration occurring at flow rates as low as 1 pl/min, and were therefore not examined further. The validity of the radiolabel method for estimating recovery was confirmed by simultaneous RIA of the collected samples (data not shown). The polyacrylonitrile and commercial polycarbonate probes were selected for further investigation on the basis of these data. As demonstrated in Table 2, the relative recovery values were independent of external concentration over a range estimated to occur in the in viva experiments described below (0.1-50 nM). Variations in flow rate produced the expected changes in relative recovery, absolute recovery per unit time remaining constant (Table 3). By transferring continuously perfused probes from artificial CSF into a solution of 1 nM [Leulenkephalin an equilibration point was shown to be reached by the second IO-min sample at the flow rate of 2.7 frl/min selected for. in vivo experiments. Similarly, repIa~ement of the probe in artificial CSF resulted in blank values in the second IO-min sample, demonstrating that no significant adsorption of the peptide to the probe was occurring (data not shown). Table 2. Continuously perfused’ (2.7 ~l~rni~) dialysis probes were immersed in increasing concentrations of [Leulenkephalin (5 ml) to which was added a constant amount of high-performance liquid chromatographypurified [‘251] [Leulenkephalin (approximately 200,000 cpm) [Leu}Enk (nM) 0.1 0.5 1.0 5.0 10.0 50.0

Recovery (%) Polyacrylonitrile Polycarbonate 9.3 9.1 1::; 8.9 10.3

8.2 10.2 10.5 7.1 8.2 7.2

Samples were collected. at 10&n intervals and relative recovery values determined as’ described in Table 1. Values represent a single probe in each case.

The comparison of [Metlenkephalin, [Leu]enkephalin, [Metlenkephalin Arg.Phe and [Metbnkephalin Arg.Gly.Leu recoveries across polycarbonate probes is shown in Table 4.

Examples of the opioid peptide immunoreactive profile of baseline and K+-evoked release are shown in Fig. 2. In both cases the majority of the immunoreactivity co-eluted with [Met]- and [Leulenkephalin with a small peak for the tetrapeptide Tyr.Gly.Gly.Phe. In some cases a smaller peak for either the heptapeptide [Metlenkephahn Arg.Phe or the octapeptide Met-enkephalin Arg.Gly.Leu was apparent but these were not consistent observations. In most cases the recovered immunoreactivity of these peaks accounted for approximately 60-70% of the amount injected as estimated from the fraction of the sample assayed directly. The ratio of [Metlenkephalin to [Leulenkephalin in basal and stimulated samples ranged from 1.3 : 1 to 3.0: 1, Table 3. A single polycarbonate

dialysis probe was immersed in a l-nM solution (5ml) of [Leulenkephalin together with 200,000 cpm of f’2fI][Leu~nkephaiin, perfused at various flow rates and 3O-min samples collected Flow rate (p lmin)

Relative recovery (X)

Absolute recovery (cpm/30 min)

0.5 I.2 2.4 5.3

29.2 13.5 6.9 3.6

198 204 207 213

Table 4. Polycarbonate membrane dialysis probes were immersed in I-nM solutions of opioid peptides and perfused at rate of 2.7 $/min Relative recovery (%} [MetJenkephalin {Leulenkephalin [Met]enkephalin.Arg,Phe [Met]enkephalin.Arg.Gly.Leu

9.5 + 2.0 7.6 & 2.0 8.4 * 1.8 3.7 + 1.0

Samples were collected at 30-min intervals together with aliquots of the bathing medium and relative recovery values were determined by RIA. Values represent the mean and S.E.M. of three probes.

5.53

Microdialysis of opioid peptides in rat brain

A

14

“at-EYK-Arg-Gig-l.aU 7 Tyr-Glq-Gly-Pbs

B

c

tides to 43.9 fmol + 12.4 S.E.M. A second stimulation 3 h later also resulted in elevated levels with an S2: Sl ratio of 0.64 _t 0.03 (Fig. 3A). In another series of experiments the second stimulation was completely prevented by simultaneous perfusion of a 10 mM EGTA-containing artificial CSF (Fig. 3B). This difference in the effectiveness of S2 compared to that in the control experiment was significant at the level P < 0.001, there being no signi~~ant difference in the effect of Sl between the two sets of experiments. On average, basal levels were unaffected by EGTA. It should be pointed out that there was a large variation in the absolute increase produced by K + as evidenced by the S.E.M. in Fig. 3B. Moreover, in approximately 30% of these experiments no initial or second K+-stimulated release was detected and these were not included in the data shown. However, in no instance was failure to elicit an initial stimulation followed by a subsequent stimulation or a successful initial stimulation accompanied by the absence of a second stimulation (except in the presence of EGTA). This failure rate was not explained by inapprop~ate placement of the dialysis probe as indicated by histological analysis and was not apparent in the veratridine experiments described below.

60

A fraction (min) Fig. 2. A. The profile of fmol quantities of opioid peptide standards injected onto reverse-phase HPLC as described in the text and eluted with the acetonitrile gradient shown. Each l-mm fraction was assayed by RIA as described in the text. Recovery of injected material was estimated as 80-100%c, B. An example of an elution profile of baseline release collected from the globus pallidus/ventral pallidum of a single rat over a 10-h period demonstrating clear peaks co-eluting with [Met]- and [Leulenkephalin together with smaller peaks coinciding with Tyr.Gly.Gly.Phe and [Metlenkephalin Arg.Gly.Leu. C. Elution profile of opioid peptide immunoreactivity recovered over a l-h period of continual stimulation with 100 mM K+-containing artificial CSF showing a similar profile to that under baseline con-

T

40

c ;

20

3 ‘T; t

15

z e

10

H

5

‘0

B T

5

6

7

20 15

SL z

10

%

Basal extracellular levels of opioid peptide represented, on average, approximately 1.5 fmol of immunoreactivity per 30-min sample (Figs 3 and 4) which, although appearing small in the figures, is at least five times higher than the limit of detection of the assay. These amounts were recovered in the absence of any peptidase inhibitors. Perfusion with 100 mM K+ for 2 min of a 30-min sample period increased the recovered opioid pep-

4

8 “,

of the evoked release

3

25

observed in the first lo-15 fractions but their precise elution times were not consistent.

Characterization

2

E =

ditions. Small peaks of immunoreactivity were occasionally

taking into account the slightly lower recovery of [Leulenkephalin across the dialysis membrane.

1

30

5 0

01234567

ttme thoure) Fig. 3. A. Time-course of experiments demonstrating the stability of basal extracellular opioid peptide levels in 30-min samples and the elevation of these levels evoked by 100 mM K+ included in the perfusion medium for 2 min of a collection period. Values represent the mean and S.E.M. of absolute amounts of total opioid peptide immunoreactivity recovered per sample (n = 4). B. A separate series of experiments demonstrating the effect of incorporation of EGTA in the perfusion medium of preventing the second stimulation by 100 mM K* (values represented as in A, n = 6). q Control; 1sI 100 mM K+ (2 mm); 8 fOmM EGTA; Q 10 mM EGTA, IOOmM K+ (2 mm).

554

N. T. MAIVMENT PI (I/.

A

20 2 $ t

15

.% 5

IO

2 z .o

5

z 0

012345

time

B

70 2 0

60

0

(hours)

Fig. 5. Time-course of repeated stimulations with 50mM K+ for 2 min of 20 min sampling periods demonstrating the progressively smaller increases in recovered opioid peptide above baseline during the course of the experiment. Values represent the mean and S.E.M. of each time point derived from six animals. 0 Control; 50 mM K+ (2 min).

01234567

time (hours) Fig. 4. A. Repeated stimulation of extracellular opioid peptide levels by incorporation of veratridine (50 PM) in the perfusion medium for 2 min of a 30-min sampling period. Values expressed as in Fig. 3, n = 4. B. Blockade of the second veratridine stimulation by continuous perfusion with tetrodotoxin (2 PM). No significant effect on basal levels was observed (n = 5). 0 Control; q 50 PM Verat (2 min); m 2 PM TTX; q 2 p M ITX, 50 PM Verat (2 min).

Incorporation of veratridine (50 p M) in the perfusion medium for 2 min of a 30-min collection increased the recovered peptides to period 16.8 fmol+ 0.5 with an S2:Sl ratio of 0.62 + 0.17 after a 3-h interval (Fig. 4A). In a separate series of experiments the second stimulation was prevented by simultaneous perfusion with a 2pM tetrodotoxincontaining artificial CSF (Fig. 4B). This difference in the effect of S2 between groups attained a level of significance of P -c 0.002. There was no significant difference in the effect of Sl between the two sets of experiments. Basal levels were unaffected by tetrodotoxin. Repeated stimulation with 50 mM K + for 2 min of a 20-min collection period at l-h intervals produced a trend of progressively smaller elevations in peptide levels (Fig. 5). Although a comparison of the amplitude of Sl with subsequent stimulations narrowly failed to demonstrate a statistically significant difference, the effect of S2 was significantly greater than that of both S3 and S4 (P < 0.05 in each case). Baseline levels were not affected by this treatment. DISCUSSION

The purpose of this study was to assess the feasibility of using microdialysis to monitor extracellular levels of endogenous opioid peptides from the rat brain in vivo. Although reports of successful in vivo measurement of [Met]- and [Leulenkephalin in the

cat’ and rat3.’ have appeared in the literature such studies have employed push-pull perfusion, Microdialysis can be considered a superior technique because brain tissue is not subjected directly to fast flowing fluid and therefore tissue disruption is reduced. However, the presence of the membrane has an associated disadvantage in that it provides a barrier to the recovery of the peptides. This exacerbates the problem of the already low concentration of these peptides in the extracellular environment. We therefore had to address the issues of both membrane permeability and assay sensitivity. Dialysis membrane evaluation

The results of our in vitro experiments suggest that the choice of membrane is important for maximixing peptide recovery. Clearly the issue is one of membrane material rather than simply molecular weight cut-off. Thus, while cellulose and cellulose acetate membranes with cut-offs ranging from SXlO70,000 have been used successfully in monoamine studies ‘3~20.32,35138 they exhibited poor recovery characteristic: for the enkephalins, presumably because they were adsorbing to the outside of the membrane. The wall thickness of the membrane may also be an issue as indicated by the different recovery values obtained with the two cuprophan membranes. Of the membranes tested, those made from polyacrylonitrile and the commercially available polycarbonate probes were the most successful. Although, largely for convenience, the commercial probes were used in these preliminary in vivo studies, the custom-made polyacrylonitrile probes have the advantage of being almost half the diameter of the commercial variety. We are therefore currently assessing the performance of these membranes in vivo. The flow rate of 2.7 Hl/min was selected for in vivo experiments largely because it produced manageable volumes in a 20or 30-min period for subsequent drying down and because future studies may involve taking aliquots for measurement of neurotransmitters by concentrationdependent assays where the low relative recoveries

Microdialysis of opioid peptides in rat brain associated with higher flow rates would be a disadvantage. Flow rates below 2.7 pl/min should, of course, prove equally effective but the small volumes associated with substantially lower flows will tend to result in greater handling errors unless the system is automated in some way. Radioimmunoassay

The plate RIA has proved to have significant advantages over other RIA procedures we have employed with respect to both sensitivity and convenience. The sub-fmol detection limit is an approximate four-fold improvement on that obtained using the second antibody separation procedure, even with equivalent assay volume, incubation time, number of counts added and antibody dilution.34 The increased sensitivity may be attributable to the quick separation which requires no incubation or centrifugation step and is simply achieved by dumping the contents of the well as with conventional enzyme-linked immunosorbant assay (ELISA) protocols. A second feature of the assay is the virtual absence of non-specific binding of the tracer (normally less than 0.5% of total counts added), the consequence being that a low number of specifically bound counts can be used to construct a reliable standard curve. The use of HPLC-purified tracer (presumably of the same specific activity as ‘25I itself) and delayed addition of the tracer were also important for maximizing the assay sensitivity, which is now close to the theoretical maximum for peptides labeled with a single lz51. Extracellular

opioid peptide proJik

The “universal” nature of the assay was considered advantageous for these initial studies in that it maximized the possibility of detecting basal levels of opioid peptide activity. Furthe~ore, when combined with HPLC, it offered the capability to observe the profile of opioid peptides present in the extracellular fluid. This revealed the expected major contribution from [Met]- and [Leulenkephalin. The ratio of [Met] : [Leulenkephalin was not consistent between animals, ranging from 1.5 : 1 to 3 : 1. The contribution from [Leulenkephalin in each case is therefore greater than would be predicted from the ratio of [Met] : [Leulenkephalin in the pro-enkephalin precursor and from whole-tissue measurements of peptide levels.‘,1’,37This discrepancy could not be explained by differentiai recoveries across the membranes or loss over the HPLC and consequently may represent differential degradation of [Met]- and [Leu]enkephalin. Alternatively, the releasable enkephalin pool may have a different composition to that stored in the tissue; either way this observation warrants further investigation. The possibility that some of the ~Leulenkephalin originates from dyno~phinergic neurons cannot be disregarded. Although small co-eluting with peaks [Metlenkephalin Arg.Phe and [Metlenkephalin Arg.Gly.Leu were occasionally observed, these were

555

not consistent findings. This is most readily explained by the absence of peptidase inhibitors in the perfusion medium since the heptapeptide and octapeptide have been reported to be more susceptible to degradation by such enzymes. 24 The tentative identification of a Tyr.GIy.Gly.Phe component to the immuno~active profile, even though the assay exhibits low crossreactivity with this peptide, would corroborate this conclusion. The other proposed major metabolite of the enkephalins-Tyr.Gly.Gly-could not be measured because of a total lack of cross reactivity in the assay. We are therefore developing a separate assay for this peptide which, when combined with the current assay, will permit a more thorough investigation of opioid peptide metabolism in vivo. Estimation

of basal extracellular

concentrations

It was considered important for future regulation studies to establish the possibility of measuring dialysate [Met]- and [Leulenkephalin levels in the absence of peptidase inhibitors since such agents, by artificially raising the level of the peptides in the synaptic cleft, may influence autoregulatory mechanisms under study. As far as we are aware this is the first report to demonstrate opioid peptide overflow in vivo in the absence of such drugs. The term “overflow” is perhaps more accurate than “release” since it highlights the fact that microdialysis is not sampling directly from the synapse but rather from a pool surrounding multiple synapses into which a proportion of released neurotransmitters and their metabolites are diffusing. It is, nevertheless, an interesting exercise to attempt to estimate the extracellular concentration of the opioid peptides measured under these conditions. The approximate concentration of the collected opioid peptides in the dialysates collected under basal conditions was 1.5 fmol/80 ~1. Ignoring small contributions from the tetra-, heptaand octapeptides and using an average relative recovery value of 8% for the two dominant peptides, produces a combined extracellular concentration for [Met]- and [Leulenkephalin of approximately 0.15 nM, which is approximately 10&300 times less than similarly computed values for dopamine in the caudate nucleus.3s However, it must be emphasized that such values are highly speculative since they rely on the essentially unfounded assumption that relative recovery estimations determined in vitro can be extrapolated to the in vivo environment. Indeed, recent work suggests that these values should be multiplied by a factor of approximately 12 to account for the tortuosity and volume fraction differences between brain tissue and solution.4 We are currently evaluating ways of making more accurate determinations of absolute extracellular concentrations with microdialysis. Basal and evoked overflow characteristics

The K+-evoked opioid peptide overflow exhibited several characteristics associated with synaptic

550

\

T. MAII)~~E~T (11t/i

release. Thus, repeated stimulation resulted in progressively smaller elevations above baseline showing that the releasable pool was depletable. It will be interesting to determine in future studies the timecourse of recovery from such depletion ~-the second 100 mM K + stimulation was consistently smaller than the first even with a 3-h interval. The reason for the failure of K + to elevate opioid peptide levels in a proportion of experiments is unclear but probably reflects a failure of extracellular K + concentrations to reach levels sufficient to produce widespread depolarization. It must be noted that the concentration attained around the probe will be significantly less than that perfused through the probe as a result of the less than 100% recovery of K’ across the membrane. Chelation of extracellular calcium with EGTA proved necessary to demonstrate the calciumdependent nature of the evoked release. Simply removing calcium from the perfusion medium failed to have any effect (data not shown). The most simple explanation for this observation is that omission of calcium from the perfusion medium is not sufficient to significantly reduce the concentration of this ion in the extracellular fluid surrounding the dialysis probe. However, such an omission has been reported to reduce the amount of dopamine recovered in dialysates of the caudate nucleus.‘J.3’ Depolarization of the neurons surrounding the probe with the voltage-dependent sodium channel activator, veratridine, also produced the predicted rise in extracellular opioid peptide levels which was reversed by the sodium channel blocker, tetrodotoxin. Of particular interest was the observation that neither EGTA or tetrodotoxin reduced the amount of opioid peptides recovered under basal conditions. The lack of calcium-dependency of basal release is also a common finding in in citro slice preparations peptide and other -neurotransfor both opioid

mittersLh.‘X.‘” and probably rcHects the prcscncc ok adequate intraneuronal stores for maintaining a low level of vesicular release. The insensitivity (11‘basal extracellular opioid peptide levels to tetrodotoxm has also been reported previously both in zGfro_“ and i/r rVc:o’ and suggests that a certain amount 01’“leakage” of opioid peptide from neurons (or from rxtraneuronal compartments) occurs independent of impulse flow. It is possible that such leakage only occurs as a result of tissue damage caused by the presence of the dialysis probe as has been suggested for dopamine when larger U-shaped probes are used.” However, this seems unlikely since one would expect to see such leakage regress over the period of an experiment whereas in the current studies basal levels remained reasonably constant throughout an 8-h recording period. Full resolution of this issue awaits similar analysis of basal opioid peptide overflow in the chronically implanted rat and such studies are currently in progress. CONCLUSION

We have demonstrated the feasibility of using microdialysis to continuously monitor basal and evoked extracellular levels of opioid peptides from a discrete region of the rat brain. We anticipate that this technique will prove valuable in evaluating the mechanisms regulating the release and metabolism of these neurotransmitters in Go. Acknowledgemenfs-The authors would like to thank Drs Jack Barchas, Kym Faull and William Dement for their continued support and Drs Joe Miller, Lee Phebus, Ivan Mefford, John Hsiao, Urban Ungerstedt and Agneta Eliasson for helpful discussions during the course of this work. This work was supported by a Lucille P. Markey Visiting Scholarship to NTM; NIDA No. DA-05010; NINCDS No. NS-23724; NSF No. BNS-8618972 and the Upjohn Company.

REFERENCES

H. and Walker J. M. (1984) A. Rev. Neurosci. 7, 1. Akil H., Watson S. J., Young E., Lewis M. E., Khachaturian 223-255. 2. Bayon A. and Anton B. (1986) Diurnal rhythm of the in oiuo release of enkephalin from the globus pallidus of the rat. Regul. Pept. 15, 63--70. R. R. and Bloom F. E. (1981) In viva release 3. Bayon A., Shoemaker W. J., Lugo L., Azad R., Ling N., Drucker-Colin of enkephalin from the globus pallidus. Neurosci. Lett. 24, E-70. J. Neurochem. 52, 1667-1679. 4. Benveniste H. (1989) Brain microdialysis. 5. Boarder M. R., Weber E., Evans C. J., Erdelyi E. and Barchas J. D. (1983) Measurement of total opioid peptides in rat brain and pituitary by radioimmunoassay directed at the a-N-acetyl derivative. J. Neurochem. 40, 1517-1522. 6. Brodin E., Linderoth B., Gazelius B. and Ungerstedt U. (1987) In viva release of substance P in cat dorsal horn studied with microdialysis. Neurosci. Let!. 76, 357-362. 7. Cesselin F., Soubrit P., Bourgoin S., Artaud F., Reisine T. D., Michelot R., Glowinski J. and Hamon M. (1981) In viva release of Met-enkephalin in the cat brain. Neuroscience 6, 301-313. of extracellular dopamine in Go. 8. Church W. H. and Justice J. B. Jr (1987) Rapid sampling and determination An&t. Chem. 59, 712-716. striopallidal pathway in rat brain. Nature 271, 9. Cue110 A. C. and Paxinos G. (1978) Evidence for a long Leu-enkephalin 1788180. release in freely moving 10. Damsma G., Westerink B., de Boar P., DeVries J. and Horn A. (1988) Basal acetylcholine rats detected by on-line trans-striatal dialysis: pharmacological aspects. Life Sci. 43, 1161-I 168. D. L. and Frederickson R. C. A. (1988) The opioid peptides. In The Opioid Receptors 11. Evans C. J., Hammond (ed. Pasternak G. W.), pp. 23-71. Humana Press, Clifton, NJ. G., Hughes J. and Kosterlitz H. W. (1978) In oifro release of Leu- and Met-enkephalin from the corpus 12. Henderson striatum. Nature 271, 677479.

Microdialysis of opioid peptides in rat brain

557

13. Hernandez L., Stanley B. G. and Hoebel B. G. (1986)A small, removable microdialysis probe. Life Sci. 39,262%2637. 14. Imperato A. and Di Chiara G. (1984) Trans-striatal dialysis coupled to reverse-phase high performance liquid chromatography with electrochemical detection: a new method for the study of the in uiuo release of endogenous dopamine and metabolites. J. Neurosci. 4, 966-984. 15. Iverson L. L., Iversen S. D., Bloom F. E., Vargo T. and Guillemin R. (1978) Release of enkephahn from rat globus pallidus in vitro. Nature 271, 679-681. 16. Kalen P., Strecker R., Rosengren E. and Bjorklund A. (1988) Endogenous release of neuronal serotonin and S-hydroxyindoleacetic acid in the caudate-putamen of the rat as revealed by intracerebral dialysis coupled to high performance liquid chromatography with fluorimetric detection. J. Neuroc~em. 51, 1422-1435. 17. Lindberg I. and Dahl J. L. (1981) Characterization of enkephahn release from rat striatum. J. Neurochem. 36,506-512. 18. Lindefors N., Brodin E. and Ungerstedt U. (1987) Micr~ialysis combined with a sensitive radioimmunoassay: a technique for studying in oivo release of neuro~ptides. J. Pha~uc. Meth. 17, 30.5-312. 19. Lindefors N., Yamamoto Y., Pantaleo T., Lagercrantz H., Brodin E. and Ungerstedt U. (1986) In vivo release of substance P in the nucleus tractus solitarii increases during hypoxia. Neurosci. Left. 69, 94-97. 20. Maidment N. T., Routledge C., Martin K. F., Braze11 M. P. and Marsden C. A. (1986) Identification of neurotransmitter autoreceptors using measurement of release in uivo. In Monitoring Neurofransmifter Release In Vioo (eds Joseph M. H., Fillenz M., Macdonald 1. A. and Marsden C. A.), pp. 73-93. Ellis Horwood, Chichester. 21. Nicolarakis K. E., Almeida 0. F. X. and Herz A. (1989) Multiple factors influencing the in vitro release of [Met’]-enkephalin from rat hypothalamic slices. J. Neurochem. 52, 428432. 22. Osborne H. and Herz A. (1980) K+-evoked release of Met-enkephalin from rat striatum in oitro: effect of putative neurotransmitters and morphine. Naunyn-Schmieakberg’s Arch. Pharmac. 310, 203-209. 23. Patey G., Cupo A., Giraud P., Chaminade M. and Rossier J. (1983) The heptapeptide Met-enkephalin-Arg, Phe is released from rat striatum in vitro by high potassium. Neurosci. Let?. 37, 267-271. 24. Patey G., Cupo A., Mazarguil H., Morgat J.-L. and Rossier J. (1985) Reiease of Proenkephalin-derived opioid peptides from rat striatum in v&o and their rapid degradation. Neuroscience 15, i035-1044. 25. Paxinos G. and Watson C. (1982) The Rat Brain in Stereotaxic Coordinates. Academic Press, New York. 26. Richter J. A., Wesche D. L. and Frederickson R. C. A. (1979) K-Stimulated release of leu- and met-enkephalin from rat striatal &es: lack of effect of morphine and naloxone. Eur. J. Pharmac. 56, 105-I 13. 27. Sar M., Stumpf W. E., Miller R. J., Chang K-J. and Cuatrecasas P. (1978) Immunohistoch~icai Iocalisation of enkephalin in rat brain and spinal cord. J. camp. Neural. 182, 17-38. 28. Sawynok J., Labella F. S. and Pinsky C. (1980) Effects or morphine and naloxone on the K+-stimulated release of methionine-enkephalin from slices of rat corpus striatum. Brain Res. 189,483-493. 29. Segovia J., Tossman U., Herrera-Marschitz M., Garcia-Munoz M. and Ungerstedt U. (1986) y-Aminobutyric acid release in the globus pallidus in viuo after a 6hydroxydopamine lesion in the substantia nigra of the rat. Neurosci. Lett. 70, 364-368. 30. Tossmann U., Eriksson S., Delin A., Hagenfeldt L., Law D. and Ungerstedt U. (1983) Brain amino acids measured by intracerebral dialysis in portacaval shunted rats. J. Neurochem. 41, 10461051. 31. Ueda H., Fukushima N., Kitao T., Ge M. and Takagi H. (1986) Low doses of naloxone produce analgesia in the mouse brain by blocking presynaptic autoinhibition of enkephalin release. Neurosci. Lett. 6!5, 247-252. 32. Ungerstedt U. (1984) Measurement of neurotransmitter release by intracranial dialysis. In &feasnrement ofNeuroiransmitfer Release In J&o (ed. Marsden C. A.), pp. 81-105. John Wiley, Chichester. 33. Wamsley J. K., Young W. S. and Kuhar M. J. (1980) Immunohist~hemjcal localisation of enkephalin in rat forebrain. Brain RES. 190, 153-174. 34. Weber E., Truscott R. J., Evans C., Sullivan S., Angwin P. and Barchas J. D. (1981) a-iV-acetyl /I-endorphins in the pituitary: immunohistochemical localisation using antibodies raised against dynorphin (I-13). J. Neurochem. 36, 1977--1985. 3% Westerink B. H. C. and De Vries J. B. (1988) Characterization of in vivo dopamine release as determined by brain microdialysis after acute and subchronic implantations: methodogical aspects. J. Neurochem. 51, 683-687.

36. Woodward J. J., Chandler L. J. and Leslie S. W. (1988) Calcium-dependent and independent release of endogenous dopamine from rat striatal synaptosomes. Brain Res. 473, 91-98. 37. Zamir N., Weber E., Palkovits M. and Brownstein M. (1984) Differential processing of prodynorphin and proenkephalin in specific regions of the rat brain. Proc. natn. Acad. Sci. U.S.A. 81, 6886-6889. 38. Zetterstrom T., Sharp T., Marsden C. A. and Ungerstedt U. (1983) In vivo measurement of dopamine and its metabolites by intracerebral dialysis: changes after d-amphetamine. 1. Neurochem. 41, 1769-1773. (Accepted 2 May 1989)