Neuroscience Vol. 40, No. 3, Pp. 657-671, 1991 Printed in Great Britain
0306-4522/91$3.00+ 0.00 Pergamon Press plc 0 1991IBRO
DISTRIBUTION OF DOPAMINERGIC RECEPTORS IN THE PRIMATE CEREBRAL CORTEX: QUANTITATIVE AUTORADIOGRAPHIC ANALYSIS USING [‘HIRACLOPRIDE, [3H]SPIPERONE AND [3H]SCH23390 M. S. Lmow,*
P. S.
D. W. GALLAGER and P. RAKIC
GOLDMAN-RAKIC,
Yale University, School of Medicine, Section of Neuroanatomy,
New Haven, CT 06510, U.S.A.
Abstract-A widespread distribution of dopamine D, receptors in the neocortex is well recognized. However, the presence of dopamine D, receptors in this structure has only recently been established [Martres er al. (1985) Eur. J. Pharmac. 118, 211-219; Lidow et al. (1989) Proc. mtn. Acad. Sci. U.S.A. fJ6, 6412-64161. In the present paper, a highly specific antagonist, [‘Hlraclopride, was used for autoradiographic determination of the distribution of D, receptors in 12 cytoarchitectonic areas of the frontal, parietal, and occipital lobes of the rhesus monkey. A low density of D,-specific [3H]raclopride binding (1.5-4.0 fmol/mg tissue) was detected in all layers of all cortical areas studied. Throughout the entire cortex, the highest density of binding was consistently found in layer V. This is a unique distribution not observed so far for any other neurotransmitter receptor subtype in monkey cerebral cortex, including D, receptor. In addition, a comparison was made of the distribution of [‘Hlraclopride and [3H]spiperone, which has been commonly used in previous attempts to label cortical D, receptors. We found marked differences in the distribution of these two radioligands. In the prefrontal cortex, the pattern of [3H]spiperone binding in the presence of ketanserin resembled the combined distribution of 5-HT,c serotoninergic and a,-adrenergic sites as well as D, receptors. Thus, [3H]raclopride provides a better estimation of the D, receptor distribution than does [3H]spiperone. The distribution of D,-specific binding of [3H]raclopride was also compared with the D,-specific binding of [3H]SCH23390 in the presence of mianserin to block labeling to 5-HT, and 5-HT,, sites. The density of D,-specific [)H]SCH23390 binding was lo-20 times higher than that of D,-specific [3H]raclopride binding throughout the cortex. The densities of both [‘Hlraclopride and [“H]SCH23390 binding sites display a rostral-caudal gradient with the highest concentrations in prefrontal and the lowest concentrations in the occipital cortex. However, the binding sites of these two ligands had different laminar distributions in all areas examined. In contrast to preferential [‘Hlraclopride binding in layer V, a bilaminar pattern of [3H]SCH23390 labeling was observed in most cytoarchitectonic areas, with the highest concentrations in supragranular layers I, II and IIIa and infragranular layers V and VI. Whereas [3H]raclopride binding was similar in all cytoarchitectonic areas, [‘H]SCH23390 exhibited some region-specific variations in the primary visual and motor cortex. The different regional and laminar distributions of D, and D, dopaminergic receptors indicates that they may subserve different aspects of dopamine function in the cerebral cortex.
An understanding of the distribution of dopamine receptors in the cerebral cortex depends on the availability of receptor subtype-specific ligands. However, most ligands have considerable problems of specificity. For example, [-‘H]SCH23390, an antagonist with high affinity for dopamine D, sites, has been used widely to evaluate the distribution of D, However, recently it has become receptors. 15-‘7~‘9-2’,49 clear that this ligand also labels 5-HTz and 5-HT,, serotonergic sites present in the cortex.6.46 Thus, the data from earlier studies employing [3H]SCH23390 *To whom correspondence
should be addressed. ( + )N,N-diethyl-N’-[(3a,4aa, lO~)-1,2,3,4,4a,10,10a-octahydro-6-hydroxy-l-propyl-3benzo-[Glquinolinyll-sulfamide; 5-HT, 5-hydroxytryptamine; SCH23390, (R)-( +)-8-chloro-2,3,4,5_tetrahydro3-methyl-5-phenyl-lH-3-benzazepin-7-01; SCH23982, (5R)-8-iodo-2,3,4,5-tetrahydro-3-methyl-5-phenyl-IH-3benzozepine 7,ol; SKF83566, ( + )7-bromo-8-hydroxy-3methyl- l-phenyl-2,3,4,5_tetrahydrolH-3-benzazepine hydrochloride; TH, tyrosine hydroxylase.
Abbreviations:
CV205-502,
should be re-evaluated. The detection of Dr receptors in the cerebral cortex has proven to be even more difficult. Indeed, the unreliability of cortical dopamine D, receptor labeling with available D,-sensitive radioligands has lead many investigators to doubt the presence of D, receptors in the cortex.‘~‘4*20~22~24~32 We recently addressed this problem and demonstrated that the D,-specific antagonist of the substituted benzamide group, [3H]raclopride, labels a population of cortical receptors with the pharmacological properties of D, sites.37 Furthermore, these D, sites were found to be widespread in the cerebral cortex of both rat and monkey.37 In the present paper, we report the results of an autoradiographic study of the distribution of D,specific [3H]raclopride binding in the monkey cerebral cortex and compare it to that of [3H]spiperone in the presence of ketanserin. The latter has been widely used in previous attempts to label cortical D, sites (for review see Ref. 47). We also compared the 657
658
M. S. LIDOW et al.
cortical localization o f D 2 sites with that o f D 1 receptors detected with [3H]SCH23390 in the presence o f mianserin, which prevents radioligand binding to 5-HT 2 and 5-HTlc serotonergic sites. In the course o f this analysis, the effect o f mianserin on the distribution [3H]SCH23390 binding was also examined. EXPERIMENTAL PROCEDURES
Tissue preparation The cerebral hemispheres of three adult rhesus monkeys (obtained from the Yale monkey colony) were prepared as described elsewhere. 38,48 Briefly, the animals were anaesthetized and perfused with ice-cold phosphate-buffered saline followed by 0.1% paraformaldehyde containing increasing concentrations of sucrose. The brains were rapidly removed, blocked and immersed in isopentane at - 4 0 ° C for 5 min before storing at -80°C. Brains were cut into 20-#mthick sections on a Bright Cryostat. Sections were mounted on acid cleaned, chrom~alum subbed slides and kept at - 8 0 ° C until assayed, usually within 2 weeks of the tissue being sectioned. Binding assay For labeling with [3H]raclopride, tissue sections were first preincubated for 20 min at room temperature in 50 mM Tris-HC1 buffer (pH 7.4) containing 150 mM NaC1. Sections were then incubated for 45 min at room temperature with 0.3 3.0 nM of radioligand in 50 mM Tris-HC1 buffer (pH 7.4) containing 150 mM NaC1 and 0.1% ascorbic acid. To determine non-specific binding, parallel sections were incubated in the presence of 1 #M (+)-butaclamol. At the end of the incubation, tissue sections were rinsed six times (60 s each) in ice-cold 50 mM Tris HC1 buffer (pH 7.4),
a
b
dipped in distilled water, dried and apposed to 3H-sensitive Ultrofilm for 8 months. Receptor binding with [3H]spiperone was conducted as described previously.38,48Tissue sections were incubated for 30 min at room temperature with 0.7q5.6 nM of radioligand in 170mM Tris-HC1 buffer (pH 7.7) containing 120mM NaCI, 5mM KC1, 2mM CaC12, 1 mM MgCI2, 0.1% ascorbic acid and 0.3 #M ketanserin. Non-specific binding was determined in the presence of 1 # M (+)-butaclamol or 1 0 # M (-)-sulpiride or 10#M dopamine. Sections were rinsed five times (60 min each) in ice-cold 170 mM Tris HC1 buffer (pH 7.7), dipped in distilled water and apposed against 3H-sensitive Ultrofilm for 2 months. Labeling of tissue sections with 1 10 nM [3H]SCH23390 was conducted for 90 min at room temperature in 50 mM Tris HC1 buffer (pH 7.4) containing 120 mM NaC1, 5 mM KC1, 2 mM CaC12, 1 mM MgC12 and 1 #M mianserin. To evaluate the effect of mianserin in the incubation buffer, some sections were incubated in mianserin-free media. Nonspecific binding was determined in the presence of 1 #M SKF83566. Sections were rinsed twice (10min each) in ice-cold 50 mM Tris-HC1 buffer (pH 7.4), dipped in distilled water and apposed to 3H-sensitive Ultrofilm for 3 months. After the films were developed, the tissue was stained with Cresyl Violet for identification of cortical layers. Tissue sections from each animal were assayed separately. For every animal, three consecutive sections were incubated with each concentration of radioactive ligand and the next two sections were used for evaluation of non-specific binding. Radioligands were obtained from New England Nuclear (Boston, MA). Other chemicals were purchased from Research Biochemicals (Natick, MA) and Sigma (St. Louis, MO).
Quantitative densitometry The autoradiographs were analysed with a computer imaging system which allows the overlay of the digitized
c
Fig. 1. Diagrams of the lateral (A) and medial (B) surfaces of the rhesus monkey cerebral hemisphere. Different patterns indicate the cortical areas where the distributions of muscarinic cholinergic receptors were examined. The cytoarchitectonic areas of prefrontal cortex (46, 9, 12 and 25) were identified according to Walker; ss the remaining areas (4, 3, 1, 2, 5, 7, 18 and 17) were identified according to Brodmann. s Line "a" shows the plane of sections presented in Fig. 2; line "b" shows the plane of sections presented in Fig. 4; and line "c" shows the plane of sections presented in Fig. 7.
Dopaminergic receptors in neocortex images of Cresyl Violet-stained sections and the corresponding autoradiograms on the computer screen in order to facilitate histological identification of specific layers on the autoradiographic images. Further, film images of sections with non-specific binding were substracted from those of adjacent sections with total binding, thus permitting the direct observation of images representing specific binding on
659
the screen. Our programs allow comparison of the optical densities of the film images of individual cortical layers with those of ‘H-standards (Amersham Corp., Arlington Heights, IL) that were apposed to the film along with the tissue sections and converts the optical densities of autoradiographs into concentrations of labeled compounds per tissue wet weight for each cortical layer. The optical
Fig. 2. Autoradiograms of the prefrontal cortex. (A) [‘H]SCH23390 labeling in the presence of mianserin (D, sites). (B) [3e]SCH23390 labeling in the absence ofmianserin (presumably Di , S-HT,, and 5-HT, sites). (C) I3HlRaclomide labeling (D, sites) and (D) I3Hlsuioerone binding in the nresence of ketanserin (presumably D2, a and 5-HT,, s&j. Roxes’mark the sites where quantitatiie measurement of radioligand distribution was made. (a) Area 46; (b) area 9; (c) area 12; (d) area 25. Arrow shows the slight increase in density of labeling in layer V corresponding with increase in [3H]raclopride binding in this layer. These and other photographs can be used only to compare the distributions of the ligands. They do not allow a comparison of the absolute densities of labeling between ligands.
660
M. S. Lmow e/ ul.
densities were between 0.08 and 0.80 (diffuse optical density) on all autoradiograms used in this study. In this range they are linearly related to tisue radioactivity on ‘H-sensitive Ultrofilm.3h We examined Walker’s areas 46, 9, 12 and 2555 in the prefrontal cortex, Brodmann’s area 48 in the precentral gyrus, Brodmann’s areas 1, 2, 3, 5 and 7 in the parietal cortex, and Brodmann’s areas 17 and 18 in the occipital cortex (Fig. 1). For the primary visual cortex (area 17) the cytoarchitectonic laminar divisions follow the criteria of Lund.4’ All other cortical areas were defined as described by WalkerSS and Brodmann.* Also, since the distribution of labeling in layer III was not homogeneous in all cortical areas, we found that an accurate description and quantification of the results required a division of layer III of most cytoarchitectonic areas into two strata: sublayer IIIaapproximately the upper one-third of the layer, and sublayer III&the lower two-thirds of the layer. Statisticalanalysis The analysis of saturation binding utilized the nonlinear curve-fitting computer programs KINETIC/EDBA/ LIGANDiLOWRY from Elsevier-BIOSOFT Co.. Cambridge, U.‘K. The analyses were based on concentrations of radioactive ligands specifically bound to tissue labeled with five different concentrations of ligands in incubating solutions. The number of data points employed for quantative analysis of saturation binding is a compromise between the desired accuracy and the availability of appropriate tissue. Five concentrations of free ligand in incubating solution is a minimal number which allows a relatively accurate estimation of B,,, and Kd for a one-site receptor model.9,39B,,,,, and Kdvalues obtained from different layers of each neocortical area were compared with the Gabriel modification of the GT2 method.s3 This method utilizes the data provided by KINETIC/EDBA/LIGAND/LOWRY and produces 95% comparison intervals for each Kd or B,,, value. K,, or Bmdxvalues with overlapping intervals are considered statistically identical; Kd or B,, with intervals which do not overlap are considered statistically different. Plots of these comparison intervals provide a comprehensive visual display of the degree of variation in Kd or B,, values between cortical layers and areas. We have previously demonstrated that sublayer IVb of area 17 and the deep strata of layer III, and layers V and VI of area 4 absorb significantly more ‘H-generated emissions than the rest of the cortex due to increased myelin content of these laminae.j6 The B,, values obtained for these layers and sublayers are underestimated and were therefore corrected.r6 RESULTS
[‘H]SCH23390
and [‘Hlraclopride
binding in the
cerebraI cortex D,-specific
labeling
of
[‘Hlraclopride
and
D,-
specific labeling of [3H]SCH23390 were observed in all layers of all cortical areas examined (Figs 2a,c, 4
and 7). We found no statistically significant changes in Kd values of either ligand between any cortical
layers or areas (Figs 3, 5, 6 and 8). The specific binding of [‘Hlraclopride constituted 4060% of the total binding and the specific binding of [3H]SCH23390, in the presence of mianserin, constituted 7&80% of the total binding. The density of [3H]SCH23390 binding was l&20 times higher than that of [‘Hlraclopride throughout the entire cortex (Figs 3, 5, 6 and 8). The prefrontal cortex had the highest overall density of both [‘Hlraclopride and [‘H]SCH23390 binding sites. In this area, the density of [‘Hlraclopride binding varied from 1.7 to 4 fmol/mg tissue (Fig. 3). In cytoarchitectonic areas 46, 9 and 12, the density of [‘H]SCH23390 binding varied from 18 to 40 fmol/mg. The binding of this ligand in area 25 was particularly high, reaching 50 fmol/mg tissue. In contrast, occipital cortex (areas 17 and 18) had the lowest density of binding for both ligands. [‘H]Raclopride binding varied from 0.6 to 2.7 fmol/mg and [3H]SCH23390 binding varied from 12 to 27 fmol/mg tissue (Fig. 8). Primary motor (area 4) and parietal cortex (areas 3, 1, 2, 5 and 7) had intermediate binding densities for both ligands (Figs 5 and 6). In these regions, [‘Hlraclopride binding was 1.4-9.5 fmol/mg tissue and [3H]SCH23390 binding was 18-38 fmol/mg tissue. The laminar distributions of [3H]SCH23390 and [‘Hlraclopride labeling were very different from each other in all cortical regions examined. The distribution of [‘Hlraclopride was the same in all cytoarchitectonic areas (Figs 2, 3, 4b, 5, 6, 7b and 8). In frontal, parietal and occipital cortex, this ligand consistently has the highest binding density in layer V. Other layers had approximately 4&50% lower density of binding sites. In contrast, [‘H]SCH23390 showed bilaminar patterns of labeling in a majority of cytoarchitectonic areas including all subdivisions of prefrontal and parietal cortex as well as prestriate region, area 18 (Figs 2a, 3, 4a, 5, 6, 7a and 8). The highest densities of labeling with this ligand were found in superficial layers I, II and IIIa and in deep layers V and VI. The middle strata of the cortex (sublayer IIIb and layer IV) were relatively poor in [‘H]SCH23390 binding sites. In parietal cortex, the density of [3H]SCH23390 binding sites in layers V and VI
Fig. 3. Histograms representing the distributions of specific [‘H]SCH23390 binding in the presence of mianserin (D, sites), specific [‘H]SCH23390 labeling in the absence of mianserin (presumably D, , 5-HT,, and 5-HT, sites), specific [3H]raclopride labeling (D2 sites) and specific [3H]spiperone binding in the presence of ketanserin (presumably D,, a and 5-HT,, sites) in area 46 of the prefrontal cortex. The data on the distribution of D, and D, receptors in areas 40, 9, 12 and 25 of monkey prefrontal cortex have already been published in Goldman-Rakic et a1.26Since the distributions of dopaminergic receptors were similar in all of these cortical areas, here we present only the data for area 46 as an example. Three histograms are presented for each Iigand. The middle histogram presents the distribution of B,, averaged for three animals + S.E.M. The Roman numerals indicate cortical layers. The histogram at the top shows B,, values with 95% comparison intervals generated by the GT2 test. The histogram below shows the distribution of Kd values with their 95% comparison intervals. If comparison intervals for any two Bm, or Kd values overlap, there are no statistically significant differences between them.
Area 46 [3 H]SCH23390 in the presence of
[3H]SCH23390
mianserin
in the absence of
(D1 receptors)
mianserin
40
~I ]~ ~ ~ ~
2O
2O
50"
5O
a~ 22' X
E
co
I
II
Ilia
IIIb
IV
V
Vl
'J
I
II
Ilia
IIIb
IV
V
VI
[3H]raclopride [3H]spiperone
(D2 receptors)
::3 o'J rJ)
8-
8"
O
6 ¸
X
E
co
4"
jlj
2'
i
i
II
Ilia
IIIb
IV
V
']
VI
,L
F i g . 3. 661
I
I
II
Ilia
IIIb
I
IV
V
I
VI
662
M. S. Lmow
L’/ u/
Fig. 4. Typical labeling patterns of [‘H]SCH23390 in the presence of mianserin (A) and [‘Hlraclopride (B) in area 4 of the frontal cortex and areas 3, I , 2, 5 and 7 of parietal cortex. Boxes mark the sites where quantitative measurement of radioligand distribution was made. (a) Area 4; (b) area 3; (c) areas 1 and 2; (d) area 5; (e) area 7. Due to differential absorption of 3H-generated emission, the patterns of optical densities of autoradiographic images of area 4 do not precisely reflect the distributions of radioligands; however, the distortions are insufficient to obscure the general pattern of ligand binding.36
was slightly (2&23%) lower than that in superficial cortical strata (Figs 4a, 5 and 6). In prefrontal and prestriate cortex, both supragranular and infragranular layers contained similar densities of binding sites (Figs 3 and 8). Exceptions to the general rule of bilaminar patterning of [‘H]SCH23390 labeling were found in the primary motor (area 4) and visual (area 17) cortices (Figs 4a, 6, 7a and 8). In the primary motor cortex, high binding density was observed in superficial layers I, II and IIIa where the binding was at least twice that in the deeper layers (Figs 4a and 5). In the visual cortex, on the other hand, [3H]SCH23390 labeling took on a trilaminar pattern with highest density in superficial layers I and II, middle sublayer IVa and deep layers V and VI (Figs 7a and 8). In other cortical layers of this area the density of labeling was 40-50% lower.
Comparison binding
qf
[’ Hlraclopride
and
[3Hlspiperone
The butyrophenone, [‘Hlspiperone, has been widely used to label cortical D, receptors (for review see Ref. 47). The labeling with this radioligand has generally been carried out in the presence of ketanserin to prevent binding to 5-HT, sites. We used three cold ligands as blanks to determine D2specific [‘Hlspiperone binding. Among them ( -)sulpiride and dopamine were unable to displace more than 5-15% of [‘Hlspiperone binding. In addition, the sections incubated with [3Hlspiperone in the presence of ketanserin and (-)-sulpiride or dopamine had labeling patterns similar to that obtained after incubation in (-)-sulpirideand dopamine-free media. Thus, we found it difficult to reliably determine the pattern of D,-specific binding
Fig. 5. Histograms representing the distribution of D,-specific binding of [‘H]SCH23390 and D,-specific binding of [3H]raclopride in areas 4, 3, 1 and 2. Conventions as in Fig. 3.
[3H]SCH23390
[3H]raclopride
(D1 receptors)
(D2 receptors)
] •. ~
t
0 4'
,60-
_¢ E
xl"
Ill 0
0306-4522/91$3.00+ 0.00 Pergamon Press plc 0 1991IBRO
DISTRIBUTION OF DOPAMINERGIC RECEPTORS IN THE PRIMATE CEREBRAL CORTEX: QUANTITATIVE AUTORADIOGRAPHIC ANALYSIS USING [‘HIRACLOPRIDE, [3H]SPIPERONE AND [3H]SCH23390 M. S. Lmow,*
P. S.
D. W. GALLAGER and P. RAKIC
GOLDMAN-RAKIC,
Yale University, School of Medicine, Section of Neuroanatomy,
New Haven, CT 06510, U.S.A.
Abstract-A widespread distribution of dopamine D, receptors in the neocortex is well recognized. However, the presence of dopamine D, receptors in this structure has only recently been established [Martres er al. (1985) Eur. J. Pharmac. 118, 211-219; Lidow et al. (1989) Proc. mtn. Acad. Sci. U.S.A. fJ6, 6412-64161. In the present paper, a highly specific antagonist, [‘Hlraclopride, was used for autoradiographic determination of the distribution of D, receptors in 12 cytoarchitectonic areas of the frontal, parietal, and occipital lobes of the rhesus monkey. A low density of D,-specific [3H]raclopride binding (1.5-4.0 fmol/mg tissue) was detected in all layers of all cortical areas studied. Throughout the entire cortex, the highest density of binding was consistently found in layer V. This is a unique distribution not observed so far for any other neurotransmitter receptor subtype in monkey cerebral cortex, including D, receptor. In addition, a comparison was made of the distribution of [‘Hlraclopride and [3H]spiperone, which has been commonly used in previous attempts to label cortical D, receptors. We found marked differences in the distribution of these two radioligands. In the prefrontal cortex, the pattern of [3H]spiperone binding in the presence of ketanserin resembled the combined distribution of 5-HT,c serotoninergic and a,-adrenergic sites as well as D, receptors. Thus, [3H]raclopride provides a better estimation of the D, receptor distribution than does [3H]spiperone. The distribution of D,-specific binding of [3H]raclopride was also compared with the D,-specific binding of [3H]SCH23390 in the presence of mianserin to block labeling to 5-HT, and 5-HT,, sites. The density of D,-specific [)H]SCH23390 binding was lo-20 times higher than that of D,-specific [3H]raclopride binding throughout the cortex. The densities of both [‘Hlraclopride and [“H]SCH23390 binding sites display a rostral-caudal gradient with the highest concentrations in prefrontal and the lowest concentrations in the occipital cortex. However, the binding sites of these two ligands had different laminar distributions in all areas examined. In contrast to preferential [‘Hlraclopride binding in layer V, a bilaminar pattern of [3H]SCH23390 labeling was observed in most cytoarchitectonic areas, with the highest concentrations in supragranular layers I, II and IIIa and infragranular layers V and VI. Whereas [3H]raclopride binding was similar in all cytoarchitectonic areas, [‘H]SCH23390 exhibited some region-specific variations in the primary visual and motor cortex. The different regional and laminar distributions of D, and D, dopaminergic receptors indicates that they may subserve different aspects of dopamine function in the cerebral cortex.
An understanding of the distribution of dopamine receptors in the cerebral cortex depends on the availability of receptor subtype-specific ligands. However, most ligands have considerable problems of specificity. For example, [-‘H]SCH23390, an antagonist with high affinity for dopamine D, sites, has been used widely to evaluate the distribution of D, However, recently it has become receptors. 15-‘7~‘9-2’,49 clear that this ligand also labels 5-HTz and 5-HT,, serotonergic sites present in the cortex.6.46 Thus, the data from earlier studies employing [3H]SCH23390 *To whom correspondence
should be addressed. ( + )N,N-diethyl-N’-[(3a,4aa, lO~)-1,2,3,4,4a,10,10a-octahydro-6-hydroxy-l-propyl-3benzo-[Glquinolinyll-sulfamide; 5-HT, 5-hydroxytryptamine; SCH23390, (R)-( +)-8-chloro-2,3,4,5_tetrahydro3-methyl-5-phenyl-lH-3-benzazepin-7-01; SCH23982, (5R)-8-iodo-2,3,4,5-tetrahydro-3-methyl-5-phenyl-IH-3benzozepine 7,ol; SKF83566, ( + )7-bromo-8-hydroxy-3methyl- l-phenyl-2,3,4,5_tetrahydrolH-3-benzazepine hydrochloride; TH, tyrosine hydroxylase.
Abbreviations:
CV205-502,
should be re-evaluated. The detection of Dr receptors in the cerebral cortex has proven to be even more difficult. Indeed, the unreliability of cortical dopamine D, receptor labeling with available D,-sensitive radioligands has lead many investigators to doubt the presence of D, receptors in the cortex.‘~‘4*20~22~24~32 We recently addressed this problem and demonstrated that the D,-specific antagonist of the substituted benzamide group, [3H]raclopride, labels a population of cortical receptors with the pharmacological properties of D, sites.37 Furthermore, these D, sites were found to be widespread in the cerebral cortex of both rat and monkey.37 In the present paper, we report the results of an autoradiographic study of the distribution of D,specific [3H]raclopride binding in the monkey cerebral cortex and compare it to that of [3H]spiperone in the presence of ketanserin. The latter has been widely used in previous attempts to label cortical D, sites (for review see Ref. 47). We also compared the 657
658
M. S. LIDOW et al.
cortical localization o f D 2 sites with that o f D 1 receptors detected with [3H]SCH23390 in the presence o f mianserin, which prevents radioligand binding to 5-HT 2 and 5-HTlc serotonergic sites. In the course o f this analysis, the effect o f mianserin on the distribution [3H]SCH23390 binding was also examined. EXPERIMENTAL PROCEDURES
Tissue preparation The cerebral hemispheres of three adult rhesus monkeys (obtained from the Yale monkey colony) were prepared as described elsewhere. 38,48 Briefly, the animals were anaesthetized and perfused with ice-cold phosphate-buffered saline followed by 0.1% paraformaldehyde containing increasing concentrations of sucrose. The brains were rapidly removed, blocked and immersed in isopentane at - 4 0 ° C for 5 min before storing at -80°C. Brains were cut into 20-#mthick sections on a Bright Cryostat. Sections were mounted on acid cleaned, chrom~alum subbed slides and kept at - 8 0 ° C until assayed, usually within 2 weeks of the tissue being sectioned. Binding assay For labeling with [3H]raclopride, tissue sections were first preincubated for 20 min at room temperature in 50 mM Tris-HC1 buffer (pH 7.4) containing 150 mM NaC1. Sections were then incubated for 45 min at room temperature with 0.3 3.0 nM of radioligand in 50 mM Tris-HC1 buffer (pH 7.4) containing 150 mM NaC1 and 0.1% ascorbic acid. To determine non-specific binding, parallel sections were incubated in the presence of 1 #M (+)-butaclamol. At the end of the incubation, tissue sections were rinsed six times (60 s each) in ice-cold 50 mM Tris HC1 buffer (pH 7.4),
a
b
dipped in distilled water, dried and apposed to 3H-sensitive Ultrofilm for 8 months. Receptor binding with [3H]spiperone was conducted as described previously.38,48Tissue sections were incubated for 30 min at room temperature with 0.7q5.6 nM of radioligand in 170mM Tris-HC1 buffer (pH 7.7) containing 120mM NaCI, 5mM KC1, 2mM CaC12, 1 mM MgCI2, 0.1% ascorbic acid and 0.3 #M ketanserin. Non-specific binding was determined in the presence of 1 # M (+)-butaclamol or 1 0 # M (-)-sulpiride or 10#M dopamine. Sections were rinsed five times (60 min each) in ice-cold 170 mM Tris HC1 buffer (pH 7.7), dipped in distilled water and apposed against 3H-sensitive Ultrofilm for 2 months. Labeling of tissue sections with 1 10 nM [3H]SCH23390 was conducted for 90 min at room temperature in 50 mM Tris HC1 buffer (pH 7.4) containing 120 mM NaC1, 5 mM KC1, 2 mM CaC12, 1 mM MgC12 and 1 #M mianserin. To evaluate the effect of mianserin in the incubation buffer, some sections were incubated in mianserin-free media. Nonspecific binding was determined in the presence of 1 #M SKF83566. Sections were rinsed twice (10min each) in ice-cold 50 mM Tris-HC1 buffer (pH 7.4), dipped in distilled water and apposed to 3H-sensitive Ultrofilm for 3 months. After the films were developed, the tissue was stained with Cresyl Violet for identification of cortical layers. Tissue sections from each animal were assayed separately. For every animal, three consecutive sections were incubated with each concentration of radioactive ligand and the next two sections were used for evaluation of non-specific binding. Radioligands were obtained from New England Nuclear (Boston, MA). Other chemicals were purchased from Research Biochemicals (Natick, MA) and Sigma (St. Louis, MO).
Quantitative densitometry The autoradiographs were analysed with a computer imaging system which allows the overlay of the digitized
c
Fig. 1. Diagrams of the lateral (A) and medial (B) surfaces of the rhesus monkey cerebral hemisphere. Different patterns indicate the cortical areas where the distributions of muscarinic cholinergic receptors were examined. The cytoarchitectonic areas of prefrontal cortex (46, 9, 12 and 25) were identified according to Walker; ss the remaining areas (4, 3, 1, 2, 5, 7, 18 and 17) were identified according to Brodmann. s Line "a" shows the plane of sections presented in Fig. 2; line "b" shows the plane of sections presented in Fig. 4; and line "c" shows the plane of sections presented in Fig. 7.
Dopaminergic receptors in neocortex images of Cresyl Violet-stained sections and the corresponding autoradiograms on the computer screen in order to facilitate histological identification of specific layers on the autoradiographic images. Further, film images of sections with non-specific binding were substracted from those of adjacent sections with total binding, thus permitting the direct observation of images representing specific binding on
659
the screen. Our programs allow comparison of the optical densities of the film images of individual cortical layers with those of ‘H-standards (Amersham Corp., Arlington Heights, IL) that were apposed to the film along with the tissue sections and converts the optical densities of autoradiographs into concentrations of labeled compounds per tissue wet weight for each cortical layer. The optical
Fig. 2. Autoradiograms of the prefrontal cortex. (A) [‘H]SCH23390 labeling in the presence of mianserin (D, sites). (B) [3e]SCH23390 labeling in the absence ofmianserin (presumably Di , S-HT,, and 5-HT, sites). (C) I3HlRaclomide labeling (D, sites) and (D) I3Hlsuioerone binding in the nresence of ketanserin (presumably D2, a and 5-HT,, s&j. Roxes’mark the sites where quantitatiie measurement of radioligand distribution was made. (a) Area 46; (b) area 9; (c) area 12; (d) area 25. Arrow shows the slight increase in density of labeling in layer V corresponding with increase in [3H]raclopride binding in this layer. These and other photographs can be used only to compare the distributions of the ligands. They do not allow a comparison of the absolute densities of labeling between ligands.
660
M. S. Lmow e/ ul.
densities were between 0.08 and 0.80 (diffuse optical density) on all autoradiograms used in this study. In this range they are linearly related to tisue radioactivity on ‘H-sensitive Ultrofilm.3h We examined Walker’s areas 46, 9, 12 and 2555 in the prefrontal cortex, Brodmann’s area 48 in the precentral gyrus, Brodmann’s areas 1, 2, 3, 5 and 7 in the parietal cortex, and Brodmann’s areas 17 and 18 in the occipital cortex (Fig. 1). For the primary visual cortex (area 17) the cytoarchitectonic laminar divisions follow the criteria of Lund.4’ All other cortical areas were defined as described by WalkerSS and Brodmann.* Also, since the distribution of labeling in layer III was not homogeneous in all cortical areas, we found that an accurate description and quantification of the results required a division of layer III of most cytoarchitectonic areas into two strata: sublayer IIIaapproximately the upper one-third of the layer, and sublayer III&the lower two-thirds of the layer. Statisticalanalysis The analysis of saturation binding utilized the nonlinear curve-fitting computer programs KINETIC/EDBA/ LIGANDiLOWRY from Elsevier-BIOSOFT Co.. Cambridge, U.‘K. The analyses were based on concentrations of radioactive ligands specifically bound to tissue labeled with five different concentrations of ligands in incubating solutions. The number of data points employed for quantative analysis of saturation binding is a compromise between the desired accuracy and the availability of appropriate tissue. Five concentrations of free ligand in incubating solution is a minimal number which allows a relatively accurate estimation of B,,, and Kd for a one-site receptor model.9,39B,,,,, and Kdvalues obtained from different layers of each neocortical area were compared with the Gabriel modification of the GT2 method.s3 This method utilizes the data provided by KINETIC/EDBA/LIGAND/LOWRY and produces 95% comparison intervals for each Kd or B,,, value. K,, or Bmdxvalues with overlapping intervals are considered statistically identical; Kd or B,, with intervals which do not overlap are considered statistically different. Plots of these comparison intervals provide a comprehensive visual display of the degree of variation in Kd or B,, values between cortical layers and areas. We have previously demonstrated that sublayer IVb of area 17 and the deep strata of layer III, and layers V and VI of area 4 absorb significantly more ‘H-generated emissions than the rest of the cortex due to increased myelin content of these laminae.j6 The B,, values obtained for these layers and sublayers are underestimated and were therefore corrected.r6 RESULTS
[‘H]SCH23390
and [‘Hlraclopride
binding in the
cerebraI cortex D,-specific
labeling
of
[‘Hlraclopride
and
D,-
specific labeling of [3H]SCH23390 were observed in all layers of all cortical areas examined (Figs 2a,c, 4
and 7). We found no statistically significant changes in Kd values of either ligand between any cortical
layers or areas (Figs 3, 5, 6 and 8). The specific binding of [‘Hlraclopride constituted 4060% of the total binding and the specific binding of [3H]SCH23390, in the presence of mianserin, constituted 7&80% of the total binding. The density of [3H]SCH23390 binding was l&20 times higher than that of [‘Hlraclopride throughout the entire cortex (Figs 3, 5, 6 and 8). The prefrontal cortex had the highest overall density of both [‘Hlraclopride and [‘H]SCH23390 binding sites. In this area, the density of [‘Hlraclopride binding varied from 1.7 to 4 fmol/mg tissue (Fig. 3). In cytoarchitectonic areas 46, 9 and 12, the density of [‘H]SCH23390 binding varied from 18 to 40 fmol/mg. The binding of this ligand in area 25 was particularly high, reaching 50 fmol/mg tissue. In contrast, occipital cortex (areas 17 and 18) had the lowest density of binding for both ligands. [‘H]Raclopride binding varied from 0.6 to 2.7 fmol/mg and [3H]SCH23390 binding varied from 12 to 27 fmol/mg tissue (Fig. 8). Primary motor (area 4) and parietal cortex (areas 3, 1, 2, 5 and 7) had intermediate binding densities for both ligands (Figs 5 and 6). In these regions, [‘Hlraclopride binding was 1.4-9.5 fmol/mg tissue and [3H]SCH23390 binding was 18-38 fmol/mg tissue. The laminar distributions of [3H]SCH23390 and [‘Hlraclopride labeling were very different from each other in all cortical regions examined. The distribution of [‘Hlraclopride was the same in all cytoarchitectonic areas (Figs 2, 3, 4b, 5, 6, 7b and 8). In frontal, parietal and occipital cortex, this ligand consistently has the highest binding density in layer V. Other layers had approximately 4&50% lower density of binding sites. In contrast, [‘H]SCH23390 showed bilaminar patterns of labeling in a majority of cytoarchitectonic areas including all subdivisions of prefrontal and parietal cortex as well as prestriate region, area 18 (Figs 2a, 3, 4a, 5, 6, 7a and 8). The highest densities of labeling with this ligand were found in superficial layers I, II and IIIa and in deep layers V and VI. The middle strata of the cortex (sublayer IIIb and layer IV) were relatively poor in [‘H]SCH23390 binding sites. In parietal cortex, the density of [3H]SCH23390 binding sites in layers V and VI
Fig. 3. Histograms representing the distributions of specific [‘H]SCH23390 binding in the presence of mianserin (D, sites), specific [‘H]SCH23390 labeling in the absence of mianserin (presumably D, , 5-HT,, and 5-HT, sites), specific [3H]raclopride labeling (D2 sites) and specific [3H]spiperone binding in the presence of ketanserin (presumably D,, a and 5-HT,, sites) in area 46 of the prefrontal cortex. The data on the distribution of D, and D, receptors in areas 40, 9, 12 and 25 of monkey prefrontal cortex have already been published in Goldman-Rakic et a1.26Since the distributions of dopaminergic receptors were similar in all of these cortical areas, here we present only the data for area 46 as an example. Three histograms are presented for each Iigand. The middle histogram presents the distribution of B,, averaged for three animals + S.E.M. The Roman numerals indicate cortical layers. The histogram at the top shows B,, values with 95% comparison intervals generated by the GT2 test. The histogram below shows the distribution of Kd values with their 95% comparison intervals. If comparison intervals for any two Bm, or Kd values overlap, there are no statistically significant differences between them.
Area 46 [3 H]SCH23390 in the presence of
[3H]SCH23390
mianserin
in the absence of
(D1 receptors)
mianserin
40
~I ]~ ~ ~ ~
2O
2O
50"
5O
a~ 22' X
E
co
I
II
Ilia
IIIb
IV
V
Vl
'J
I
II
Ilia
IIIb
IV
V
VI
[3H]raclopride [3H]spiperone
(D2 receptors)
::3 o'J rJ)
8-
8"
O
6 ¸
X
E
co
4"
jlj
2'
i
i
II
Ilia
IIIb
IV
V
']
VI
,L
F i g . 3. 661
I
I
II
Ilia
IIIb
I
IV
V
I
VI
662
M. S. Lmow
L’/ u/
Fig. 4. Typical labeling patterns of [‘H]SCH23390 in the presence of mianserin (A) and [‘Hlraclopride (B) in area 4 of the frontal cortex and areas 3, I , 2, 5 and 7 of parietal cortex. Boxes mark the sites where quantitative measurement of radioligand distribution was made. (a) Area 4; (b) area 3; (c) areas 1 and 2; (d) area 5; (e) area 7. Due to differential absorption of 3H-generated emission, the patterns of optical densities of autoradiographic images of area 4 do not precisely reflect the distributions of radioligands; however, the distortions are insufficient to obscure the general pattern of ligand binding.36
was slightly (2&23%) lower than that in superficial cortical strata (Figs 4a, 5 and 6). In prefrontal and prestriate cortex, both supragranular and infragranular layers contained similar densities of binding sites (Figs 3 and 8). Exceptions to the general rule of bilaminar patterning of [‘H]SCH23390 labeling were found in the primary motor (area 4) and visual (area 17) cortices (Figs 4a, 6, 7a and 8). In the primary motor cortex, high binding density was observed in superficial layers I, II and IIIa where the binding was at least twice that in the deeper layers (Figs 4a and 5). In the visual cortex, on the other hand, [3H]SCH23390 labeling took on a trilaminar pattern with highest density in superficial layers I and II, middle sublayer IVa and deep layers V and VI (Figs 7a and 8). In other cortical layers of this area the density of labeling was 40-50% lower.
Comparison binding
qf
[’ Hlraclopride
and
[3Hlspiperone
The butyrophenone, [‘Hlspiperone, has been widely used to label cortical D, receptors (for review see Ref. 47). The labeling with this radioligand has generally been carried out in the presence of ketanserin to prevent binding to 5-HT, sites. We used three cold ligands as blanks to determine D2specific [‘Hlspiperone binding. Among them ( -)sulpiride and dopamine were unable to displace more than 5-15% of [‘Hlspiperone binding. In addition, the sections incubated with [3Hlspiperone in the presence of ketanserin and (-)-sulpiride or dopamine had labeling patterns similar to that obtained after incubation in (-)-sulpirideand dopamine-free media. Thus, we found it difficult to reliably determine the pattern of D,-specific binding
Fig. 5. Histograms representing the distribution of D,-specific binding of [‘H]SCH23390 and D,-specific binding of [3H]raclopride in areas 4, 3, 1 and 2. Conventions as in Fig. 3.
[3H]SCH23390
[3H]raclopride
(D1 receptors)
(D2 receptors)
] •. ~
t
0 4'
,60-
_¢ E
xl"
Ill 0
