Laminar distributions of muscarinic acetylcholine, serotonin, GABA and opioid receptors in human posterior cingulate cortex

0306.4522:90 $3.00 + 0.00 Pergamon Press plc

.Yrurosciencr Vol. 36, No. I. pp. 165-174, 1990 Printed in Cireat Britain

('I1990

lBR0

LAMINAR DISTRIBUTIONS OF MUSCARINIC ACETYLCHOLINE, SEROTONIN, GABA AND OPIOID RECEPTORS IN HUMAN POSTERIOR CINGULATE CORTEX B. A. VOGT,* I/ M. D. PLAGER,? P. B. CRINO:!/

and

E. D. BIRD$

Departments

of *Anatomy, tPhysiology and $Behavioral Neuroscience, Boston University School of Medicine, 80 East Concord Street, Boston, MA 02118, U.S.A. $Brain Tissue Resource Center, McLean Hospital, 115Mill Street, Belmont, MA 02178. U.S.A. j/Veterans Administration Hospital, 200 Springs Road, Bedford, MA 01730, U.S.A.

Abstract-Experimental animal studies have demonstrated a number of receptor localizations on specific cortical afferents and neurons. The present study of human posterior cingulatc cortex evaluates the laminar distributions of particular receptors and their likely association with components of the neuropil. Coverslip autoradiographic and single grain counting techniques were used followed by heterogeneity analysis in which the layer of peak binding and an index of heterogeneity were determined for each ligand. The index was calculated by determining specific binding by layer as a percentage of binding in all layers. The differences from an absolutely homogeneous distribution, i.e. 11.1% for each of nine layers. were subtracted and the absolute laminar differences summed to form the index. High indices of over 15 reflected heterogeneous binding patterns in neocortex. The binding of ligands for muscarinic acetylcholine. serotonin, opioid, GABA and beta adrenoceptors was evaluated. Pirenzepine binding peaked in layer II of area 23a but was extremely homogeneous with an index of heterogeneity of 8.9. In contrast, oxotremorine-M binding had a peak in layer IlIc and an index of 16.4, while AF-DX I16 binding peaked in layer IIIa -b and had an index of 30.6. Of the ligands for serotonin uptake and receptor binding paroxetine binding was evenly distributed in layers I-111 and had a low index of heterogeneity of 9.8. Ketanserin binding was also homogeneous and, since it had an index of 8.9, this pattern was virtually the same as that for paroxetine. In contrast, serotonin and Il-hydroxy2-(di-~-propylamino)tetraiin binding peaked in layer II and had very high indices of 20.8 and 50.3. respectively, suggesting only a limited association with that of the paroxetine distribution. Finally. there were three layers which contained peaks in binding for ligands for opioid, GABA and beta adrenoceptors. Firstly, layer Ia had peak dynorphin-A binding, the latter of which had an index of 22.6. Secondly. Tyr-o-Ala-Gly-MePhe-Gly-ol and 2-D-penicillamine-5-D-penicillamine-enkephalin binding peaked in layer II and had indices of 8.6 and 17.4, respectively. Thirdly. muscimol and (-)-cyanopindolol binding peaked in layer IiIa-b and had indices of 29.6 and 11.1, respectively. When viewed in the context of experimental animal studies, it is likely that heterogeneities in oxotremorine-M and paroxetine binding are associated with the termination of the thalamic and raphe nuclei. respectively. While serotonin> receptors are co-distributed with serotonin uptake sites. serotonin,, receptors have a significant mismatch with these sites. Finally, postsynaptic receptors preferentially peak in different layers including kappa opioid receptors in layer la, M, acetylcholine. serotonin,,, mu and delta opioid receptors in layer II, GABA, and beta adrenoceptors in layer IIIa-b and serotonin, receptors in iayer 111~. Thus, ligand binding analyses provide new insights into the structure and connections of human cerebral cortex

There of

the

is an extensive primate

literature

brain,

detailing

particularly

connections those

in

the

is known, however, about connections brain because most tract tracing techniques are experimental in nature and cannot be conducted prior to death. In some instances neuronal degeneration following naturally occurring or monkey.

Little

in the human

Ahhreriation.s: AF-DX 116, 11[[2-[(diethyl-amino)methyl]I -piperidinyI]acetyl]-5, I 1-dihydro-6H-pyrido[2,3-6][ I ,4]benzodiazepine-6-ones DAGO, Tyr-D-Ala-Giy-MePh~Gly-ol; DPDPE, Z-u-penicillamine-5-o-penicillamineN-2-hydroxyethylpiperazineenkephalin; HEPES; N’-2-ethanesulfonic acid: S-HT. serotonin: S-OHDPAT. X-hydroxy-2-(di-n-propylamino)tetralin; OXOM~pirenzepine, [ 3H]oxotremorine-M coincubated with 50 nM unlabeled pirenzepine. 165

neurosurgicaily-placed lesions in human brain have been studied with cell degeneration,>“ axon demyelination ‘*z~J’ silver-stained axon degeneration’.“.” and imm;nohistochemical” techniques. Most of these approaches require the fortunate association of a restricted and well placed lesion with an approp~ate post-lesion survival time; requirements which greatly restrict their systematic application in studies of the human brain. The advent of neurotransmitter receptor subtype localization using cryomicrotome section autoradiography has provided the basis for systematic and high resolution studies of the structural organization of the human cerebral cortex. It is proposed here that this technique, in conjunction with heterogeneity analysis and comparison with experimental

localization findings. can be used to gain insights into specific connections of the human cerebral cortex. Heterogeneity analysis involves determining the layer(s) in which peak(s) in specific binding occur(s) and calculating an index of heterogeneity. This index is calculated by determining specific binding in each layer as a percentage of binding in all layers. The differences from an absolutely homogeneous distribution, i.e. 11.1% for each of nine layers in neocortex, are then subtracted from the former distribution. The absolute differences are summed to produce the index. Thus, the index of heterogeneity measures the extent to which one proportionate laminar distribution in receptor binding differs from another or from absolute homogeneity.” For example, ligands for muscarinic acetylcholine receptors have ditrerent laminar distributions in rat posterior cingulate cortex. Oxotremorine-M (0X0-M) binds preferentially to M2 acetylcholine receptors and this binding peaks in layers Ia and IV, while pirenzepine binds to M, receptors and is relativeiy holnogeneous throughout all layers. The index of heterogeneity can be calculated by subtracting the proportionate laminar distribution for pirenzepine from the same distribution for 0X0-M binding. The absolute values of these differences are summed to form the index. Higher indices reflect heterogeneous distributions of receptors, while low indices represent homogeneous distributions. Therefore. this procedure dissociates the binding in peak layers, i.e. heterogeneities. from that which is homogeneously distributed throughout all layers. There are two instances in which heterogeneities in receptor binding have been experimentally demonstrated to be associated with afferent connections and one in which they have been associated with neuronal dendrites. Firstly, ablations of rat anterior thalamic nuclei reduce binding of 0X0-M by 50% in layers Ia and IV of rat cingulate cortex. These are likely to be M, acetylcholine heteroreceptors.“’ Secondly, removal of raphe afferents to cingulate cortex in the rat with either ablation of the dorsal raphe nuclei or intraventricular 5,7-dihydroxytryptamine injections reduces peak binding of paroxetine to serotonin (5-HT) uptake sites by 70% in layer Ia.’ Thirdly, the peak in muscimol binding to GABA, receptors in layer Ia of rat cingulate cortex is reduced following neurotoxin-induced removal of cortical neurons, suggesting that these sites are on the apical dendritic tufts of pyramidal neurons.” The distribution of thalamic projections and dendrites of cortical neurons have been described in posterior cingulate cortex of non-human primates. Projections of the anterior thalamic nuclei terminate mainly in the granular layer of area 2936 and those of the medial pulvinar in layers 111~ and IV of area 23.4 The distribution of apical and basal dendrites of cortical pyramidal neurons as well as some of the features of non-pyramidal neurons have been describedJ5 and can be used in the context of

interpreting receptor binding patterns in thlh cortlcsl region. However, previous receptor binding studies of human brain have generally not provided ;L high degree of regional or laminar localization information so that it is difficult to draw conclusions about possible associations between particular components of the neuropil and specific receptors. For example, although autoradiographic studies have demonstrated the binding of paroxetine.” S-l-IT and 8-hydroxy-2-(di-~-propylamino)tetralill (&OHDPAT)‘7,‘8 and ketanserin’R,‘7 in human cerebral cortex, the laminar binding distributions of these ligdnds were not fully described in cingulate cortex and reports for other areas were only partial. The strategy for the present study was to evaluate laminar heterogeneities in receptor subtype binding as a probe for specific connections in human posterior cingulate cortex. Radiolabeled probes were used to analyse the laminar distributions of M, and M, acetylchoIine, 5-HT uptake and 5-HT,,, 5-HT,. GABA,. opioid and beta adrenoceptors. The distribution of thalamic and raphe afferents and intrinsic dendrites were then inferred from heterogeneities in the binding of ligands to these receptors. As probes for receptor subtypes are refined it may be expected that this type of analysis can be applied to a complete range of specific connections in the mammalian central nervous system including the human species.

Blocks of posterior cingulate cortex were obtained posf mortem from nine individuals without a clinical history of dementia and who died from non-neurological causes. In some cases blocks were taken from previously frozen, coronal, 1-2-cm-thick slabs and in other cases blocks were removed from brains in an ice slurry and then frozen to - 70°C. In all instances the brain was frozen to - 70°C and not rewarmed for dissection. The post mortem interval was 14 + 1.5 h (mean t: S.E.M.). There were seven male and two female cases whose age at death was 66 & 4.2 years and whose brains weighed 1336 + 52 g. Sections were cut on a cryomicrotome at a 16 hrn thickness for receptor binding or were cut at three times this thickness and counterstained with Thionin for cytoarchitectural analysis. The sections were mounted on chrome-alum subbed slides. Materials

Unlabeled pirenzepine was kindly provided by Boehringer Ingelheim, Ltd. Radiolabeled ligands were purchased from New England Nuclear and included the foIlowing: 13H]pirenzipine (specific activity 84 Ci/mM), [3HfoXO-M; (specific activity 85.1 CijmM), (3H]I 1[~2-(d~ethyl-amino)methyl]-lpi~ridinyl]a~tyl]-5,l l-dihydr~6H-py~do[2,3-6]~1,4]~nzodiazepine-&one ([)H]AF-DX I 16; specific activity 59.8 Ci/mM), 13HJparoxetine (specific activity 23.1 CijmM), [ ‘HIS-HT (specific activity 30 Ci/mM), [ 3H]8-OH*DPAT (specific activity 142.9 Ci/mM), [ ‘HI-ketanserin (specific activity 61 Ci/mM), [3H]muscimol (specific activity 23.2 Ci/mM), [3H]Tyr-o-Ala-Gly-MePhe-Gly-ol (DAGQ; specific activity 30.3 Ci/mM), [“H]2-D-penicillamine-!&Dpenicillamine-enkephalin (DPDPE; specific activity 43 Ci/mM), [3H]dynorphin-A (l-8) (specific activity

Receptor 27.6 CLmM) 2200 WmM).

and

[ “‘I]( - )-cyanopindolol

localization

(specific

in human

activity

During the course of these studies a total of nine cases were available for analysis. All nine cases were evaluated for pirenzepine. 0X0-M, 5-HT and muscimol binding. Six cases were used for &OH-DPAT, ketanserin. dynorphin-A and (-)-cyanopindolol binding. Four cases were employed for the analysis of AF-DX 116, DAGO, DPDPE and parcltoxine binding. The conditions for ligand binding experiments have been reported previously for both human and experimental animal brainsY,‘X.“.29.“.‘X.“Uand so will only be briefly stated here. All buffers were at pH 7.4. Pircpn:rpinr. Sections were incubated in I2 nM [ ‘Hjpircnzepine in Krebs-Henseleit buffer for 70 min at 25 C followed by two buffer washes at 4 C for 2 min each. Non-specific binding was assessed in a parallel series with I HIM atropine. O.wtremorinr-M. Sections were incubated in 0.1 or 5 nM [?H]OXO-M in 20mM HEPES Tris buffer with 10 mM Mg’ ’ and 50 nM pirenzepinc for 30 min at 25 C followed by two buffer washes at 4 ‘C for 2 min each and one 2-min water wash. Non-specific binding was determined with 1I’M atropine in a parallel series of sections. Purosetinc~. Sections were incubated in 100 pM ~~~]~ir~xetine in 50 mM Tris buffer with 120 mM NaCl and 5mM KC1 for 120min at 25 ‘C followed by two washes at 15 C for I h each and a I-min water wash at 4’C. Non-specific binding was assessed in a parallel series which inclu~~~d 100 ~5M Auoxetine. Erotonin. Sections were preincubated in 5OmM Tris buffer with 5.7 mM ascorbatc, IO,L~M pargyline and 4 mM CaCI, for 15 min at 25 C. They were then incubated in the Same buffer with 3 nM [‘HJS-HT at 25‘C for 40 min followed by three buffer washes at 4 C for 1min each. Non-specific binding was determined in a parallel series with I /lM 5-HT. K~~rcmsc~rin.Sections were preincubated in 50 mM Tris at 25 C for IOmin and then incubated in 2 nM [:H]ketanserin at 30 C for 30 min followed by three buffer washes at 4 C for I min each. 8-ff~~lrt)x~-2-(rli-n-~rc)~~luminn)terru~in. Sections were preiilcubated in I70 mM Tris with 4 mM CaCI, and 0.01% ascorbate at 25 C for 3Omin. The sections were then incubated in buffer with 2 nM I’HWOH-DPAT at 25 C for 60 min followed by two buf& ha&es at 4 C for 5 min each. Non-specific binding was evaluated in a parallel series of sections with 100 11M unlabeled 8-OH-DPAT. ~~z~~s~,~t?~~~~. Sections were preincubat~ in 50 mM Tris buffer at 20 C for 40min. They were then incubated in the same buffer with 2OnM [ ‘H]muscimol for 15 min at 20’C followed by one buffer wash at 4 C for 2 min and a Water wash at 4 C for 2min. 2-l~-Pf~ni~illumin~-S-u-~~mi~ilk~minr-clnke~halin. Sections were preincubated in 50mM Tris with 5mM MgCL, 2 my ml bovine serum albumin. ZO~g/ml bacitracin, IOOmM NaCl and 50pM GTP at 25 C for IS min. The sections were then washed twice in a similar buffer to remove the GTP at 25 C for S min each. Incubati(~n of [ ‘H]DPDPE was conducted in 50 mM Tris with 2 mg/ml bovine serum albumin at 25 C for 60 min followed by three butTer washes at 4 C for IO min each. Non-specific binding was assessed in a parallel series with 1JIM DPDPE. T,,r-u-Aku-Cam-M~Pllr-GI,-oi. Sections were incubated in 50 mM Tris with I nM [jH]DAGO at 25 C for 45 min followed by three buffer washes at 4 C for 1 min each. Non-specific binding was evaluated in a parallel series of sections with 1!tM levdllorphan which was kindly provided by Ho~tnanu-~ Roche. Inc. Dynorphm-A. Sections were incubated in 50 mM Tris with 5 nM [?H]dynorphin-A at 4 C for 90 min followed by four buffer washes at 4 C for 2 min each. Non-specific

cingulate

cortex

167

binding was evaluated in a parallel series with I JIM levaliorphan. (- )-C~anopindoW Sections were incubated in i 70 mM Tris buffer with 135 mM NaCl, I ,uM 5-HT and 65 pM [ “‘I]( -)-cyanopindoloi at 25 C for 120 min foilowed by three buffer washes at 4:C for 20min each and a 5-s water wash. Non-specific binding was determined in a parallel series with 30 it M isoproterenol. 11[[2-[(i3i~~f~~~-unlina)m~fh~i]l-pjperjdj~~l ]WPf~~l]-L,l I The dih.ydro- 6H -p.~rin’o[2,3 - 6][ I ,4]ben-_udiazupinP-hww

protocol for [jH]AF-DX I I6 binding included the following steps: pr~in~ub~tion in 50mM NajKPO, buffer (pH 7.4, 30 min, 25’ C); incubation in the same buffer with 5 nM [?H]AF-DX 1I6 (30 min, 25’ C); a wash in buffer (3 min. 4’ C); a wash in distilled water (1 min, 4’ C); rapid drying. Autoradiographs were prepared according to the method of Young and Kuhar.4’ Coverslips were acid cleaned and dipped in Kodak NTB-2 nuclear tract emulsion. The dried coverslips were attached to slides with cyanoacryiatc and exposed in the dark at -20 C for three weeks to eight months. All autoradiographs were developed in Kodak D-19 without hardener, fixed in Kodak Rapid Fixer and then counterstained with Thionin. Q~u~ti~~,at~on

of u~t~)rf~dj~~rup~l.~

The cytoarchitecture of posterior cingulate cortex in and human brains has been thoroughly monkey analvsed.0,2’,“4 For the present study the exact boundaries of areas in posterior cin&ate cortex”are those of Vogt’h and Vogt et ul.‘OThe mediclateral distribution of areas in the posterior cingulate cortex is shown in Fig. I. The five areas that were analysed in this study included areas 29 I, 29m, 30, 23a and 23b. Each cortical area and its layers and sublayers. i.e. layers la, Ic, If. II&b, 111~. IV. Va, Vb and VI were identified on bright-field illumination and then dark-field illumination was used so that a computerized image analysis system (Image Technology Model 1000, DonSanto Corp. were quantified per 2500pm’ of a cortical layer in three

i

Fig. I. Medial surface of the human brain. Sections were cut coronally through posterior cingulate cortex. The borders of each area are marked with arrows in the trdW#erSe section. Grain densities were quantified in each area at levels marked by the dotted lines. CC. corpus callosum: CS. cingulate sulcus; Sub, subiculum.

168

H A. Lo<;, <‘/ NI. Table

I. Indicts of heterogeneity

Area 291

t\rea

for pirenzepine

29m

Layer

%t

20%$

%

I4 .3“0

‘I/”

Ia

6.3 13.6

- 13.7 - 6.4

8.1 16.9

m-6.2 2.6

11.2 l2.9*

IC

II Illa-b IIIC IV Va Vb VI

(II -IV&

33.1**

13.1 (III)

(V)

28.8

8.4

18.3

- 1.7

Indices of heterogeneity

19.7**

3.4 0.5 -0.8 -4.9

0.1 I.8

%

Area

I I I%

Il.6 12.2*

3.5 2.5 0.2 -0.1 -2.1 -2.0 -3.3

23.8

43.3

Area 13~1

I I. I %

14.6** l3.6* 11.3 11.0 9.0 9.1 7.8

5.4

17.1* 14.8 13.5 9.4

binding

Area 30

0.5 I.1

13.2** 12.8* 11.5 I I.1 10.2 10.4 9.6

2. I 1.7 0.4 0 -0.9 -0.7 -1.5

15.6

I I 1Y,,

10.9 l2.4*

~0.2 I.?

13.5** l2.4* 12.2* 11.0 10.7 9.8 9.4

8.9

Table 2. Indices

Layer Ia Ic II IIIa-b IIIC IV Va Vb VI Indices of heterogeneity

Muscarinic acetylcholine receptor binding Pirenzepine binding. Specific pirenzepine binding in area 23a was at a peak level in layer II, high in adjacent layers Ic and IIIa-b and at moderate levels in other layers, as shown in Table 1. In the less cytoarchitecturally differentiated areas 29m and 291 peak pirenzepine binding was also associated with superficial layers that were adjacent to layer I, i.e. layers III and II-IV, respectively. Of the four protocols employed for evaluation of muscarinic acetylcholine receptor binding, pirenzepine binding was the most homogeneous in area 23a, i.e. had the smallest index of heterogeneity when compared with an absolutely homogeneous distribution of 11. I % binding in each layer, as shown in Table 2. Oxotremorine-M binding. The laminar distribution and index of heterogeneity for specific binding of 5 nM [ 3H]OXO-M co-incubated in 50 nM unlabeled

of heterogeneity

%t

Il.I%f

11.6 12.2* 13.2** l2.8* 11.5 II.1 10.2 10.4 9.6

0.5 1.1 2.1 1.7 0.4 0 -0.9 -0.7 - 1.5

8.8 II.2 11.6 12.3 13.9** 12.8* 13.1* 10.9 5.5

8.9

of layers

RESULTS

0X0-M, S/pirenzepine __-__ 11.1% %

Pirenzepine

2.4 1.3 I.1 -0.1 -0.4 - I.3 -~ 1.7 9.8

*Percentages within 10% of peak values. **Peak values were mostly associated with layer II; its border with layer I is emphasized with a dashed line. tpercentage of all binding in each layer. fpercentage from which binding in each layer was subtracted. This percentage varies depending on the number in an area. SLayem in parentheses are for those in areas which are poorly differentiated, i.e. areas 291 and 29m.

non-adjacent sections incubated with or without a blocker. The readings were visually corrected for miscounts due to overlapping grains and then non-specific binding subtracted from total binding to determine specific binding. The mean specific binding was calculated for each layer of each case. Specific binding for all layers in an area was then totalled and the proportion of specific binding in each layer calculated. These percentages were used to determine layers of peak binding by calculating the mean rt S.E.M. for all cases. Moderate levels of binding were defined as occurring in those layers which were within 10% of that in the peak layer. Indices of heterogeneity were determined as previously reported.38 This involved subtracting the proportionate distribution of paroxetine binding from similar distributions of 5-HT receptor ligands and then calculating the sum of the absolute differences to form an index of heterogeneity. Alternatively, the specific binding in each layer was reduced by 11.1% for those areas in which nine sublayers were sampled to form an index which deviated from absolute homogeneity, i.e. 100% divided by nine sample layers = 11.1%. In areas where laminar cytoarchitecture was not differentiated into the typical six-layer structure of neocortex, this measure was adjusted to account for the reduced number of sample layers.

?3b

u/o

-2.3 0.1 0.5 I.2 2.8 1.7 2.0 -0.2 -5.6 16.4

*Percentages within 10% of peak values. **Peak values. TPercentage of all binding in each layer. iDeviation of binding from 11.1%for each layer.

for area 23a

~

0X0-M, 0.1 /pirenzepine % 11.9’ 12.2* 11.9* 11.9* 13.0** 10.0 10.8 10.5 7.8

11.1% 0.8 I.1 0.8 0.8 1.9 - 1.1 -0.3 -0.6 -3.3 10.7

AF-DX % 8.5 11.6 12.7 18.0** 17.5* 8.5 9.5 6.4 7.4

116 11.1% -2.6 0.5 1.6 6.9 6.4 -2.6 - I.6 -4.7 -3.7 30.6

Receptor

localization

layers I-IV, which had highs of 12.9 k 1% (mean f S.E.M.) in layer Ia and 12.8 k 0.3% in layer 111~ and a low of 11.1 f 0.3% in layer IV. Layers V and VI had lower binding of between 8.9 + 0.7% and 9.5 + 0.3%. The index of heterogeneity for paroxetine binding was 9.8 as measured against an absolutely homogeneous distribution with 11.1% of specific binding in each layer. As binding of no other ligands for 5-HT receptors was this homogeneous and variations from it for other ligands are a measure of the extent to which a class of receptors may be innervated by serotoninergic afferents, the binding profiles of other ligands were measured against that for paroxetine. Serotonin binding. Specific 5-HT binding is shown for area 23a in Table 3. This binding was clearly heterogeneous with a peak in layer II of 16.7 k 0.4% and less binding in layer Ic of 15.1 & 0.4%. Although specific 5-HT binding was low in layers IIIc-VI, there was a secondary peak in layer VI of 9.3 k 0.4%. The index of heterogeneity for 5-HT binding calculated by subtracting the proportionate distribution of paroxetine from a similar distribution for 5-HT was 20.8, which was more than twice that for paroxetine binding alone. 8-Hydroxy-2-(di-n-propylamino)tetralin binding. The laminar distribution of specific 8-OH-DPAT binding was more heterogeneous than any other ligand analysed. As shown in Table 3. binding of 8-OH-DPAT reached a major peak in layer II of 23.3 k 1.9%, while in layers Ia and Ic were also high at 16.8 i 1.0% and 20.2 k 1.4%, respectively. While binding in layers IIIc-Vb was extremely low with a range of 4.1 k 0.9% to 6.6 k l.O%, there was a secondary peak in layer VI of 10.3 rfi I .3% which was present in all areas of posterior cingulate cortex. In area 29m, where layer II cannot be distinguished cytoarchitecturally, the primary peak was in layer III, while in area 291 this peak occurred in the undifferentiated layers II-IV. The index of heterogeneity for 8-OH-DPAT binding minus that for paroxetine in area 23a was extremely large at 50.3 due to the very

pirenzepine (0X0-M, S/pirenzepine) was very different from that for pirenzepine binding, as shown in Table 2. Peak 0X0-M. S/pirenzepine binding was in layer IIIc and binding values within 10% of that peak were in layers IV and Va. The index of heterogeneity of 16.4 was more than 50% higher than it was for pirenzepine binding. Tissue incubation in 0.1 nM [ ‘H]OXO-M coincubated with 50 nM unlabeled pirenzepine (0X0-M, O.l/pirenzepine) is a protocol which amplifies heterogeneities in the rat cortex and presumably has a higher specificity of binding for M, sites than the 0X0-M, S/pirenzepine protocol.‘8 Binding of 0X0-M, 0.l/pirenzepine peaked in layer 111~like 0X0-M, S/pirenzepine. However, binding of 0X0-M, 0. I/pirenzepine was very homogeneously distributed throughout layers Ia-IIIa-b and the index of 10.7 was very close to that for pirenzepine binding. It should be noted that in order to detect specific 0X0-M, O.l/pirenzepine binding the autoradiographs were exposed for very long periods of six to eight months and absolute grain densities were still low at about 25% those of the 0X0-M, S/pirenzepine series which were developed in about three months. Thus, 0X0-M, O.l/pirenzepine binding was both homogeneous and of low density. AF-DX 116 binding. Specific binding of AF-DX I I6 was the most heterogeneous of any ligand for muscarinic acetylcholine receptors, as shown in Table 2. Peak AF-DX 116 binding was in layer IIIaab, although binding in layer 111~was within 10% of that peak value. Binding was at a moderate level in layer II and quite low in layers IV-VI. The index of heterogeneity of 30.6 was almost 3.5 times that for pirenzepine. Serotonin

receptor binding

Parauetine binding. Heterogeneities in the distribution of 5-HT uptake sites is one means of estimating the laminar distribution of serotoninergic input to posterior cingulate cortex. Table 3 shows the distribution of paroxetine-labeled uptake sites in area 23a. There were only minor variations in binding among Table Paroxetine Layer la IC II

IIIammb IIIC IV Va Vb VI Indices of heterogeneity

%t 12.9** 11.5* 11.8* I1.6* 12.8** 11.1 9.9 9.4 9.2

3. Indices

of heterogeneity Serotonin

11.1% 1.8 0.4 0.7 0.5 1.6 0 - 1.2 - 1.7 ~ 1.9

% 13.8 15.1* 16.7** 12.7 9.0 8.5 7.9 7.5 9.3

9.8

*Percentages within 10% of peak values. **Peak values. tpercentage of all binding in each layer. IDeviation of binding from that of paroxetine

169

in human cingulate cortex

for area 23a R-OH-DPAT

Parox%$ 0.9 3.6 4.9 1.1 -3.7 -2.6 -2.0 - 1.9 0.1

% 16.8 20.2 23.3** 9.7 4.5 4.1 4.7 6.6 10.3

Parox%

%

Parox%

3.9 8.7 11.5 - 1.9 -8.2 -7.0 -5.2 -2.8 1.1

10.9 11.8* 12.6* 12.9* 13.1** 11.9* 10.8 8.8 7.4

-2.0 -0.3 0.8 1.3 0.4 0.8 0.9 -0.6 - 1.8

20.8

binding

Ketanserin

50.3

in each layer

8.9

I70

H. .A. V~,(;~r(‘I rri Table 4. Indices Area 291

for muscimol

Area 29m

binding

Area 30

Area 23a

Area ?3b

I 1 I 9”

20%

o/0

%

11.1%

%

I 1 1“j”

%I

3.s 10.0

- 16.5 -10.0

2.8 8.7

11.5 -5.6

9.6 II.6 17.1*

--I .5 0.5 6.0

I I.4 12.3 15.4*

0.3 1.2 4.3

10.6 Il.5 14.1*

m-o.5 0.4 3.0

44.0**

24.0

28.8**

14.5

17.7** 15.2 10.2 7.7 6.3 4.6

6.6 4.1 -0.9 -3.4 -4.8 -6.5

l6.5** 14.8 10.0 8.0 6.6 5.1

14.9** 14.1* I I.1 9.3 x.3 6.1

3.x 3.0 0 - 1.8 -2.8 -5.0

Layer

%

la IC II Illa-b (II-IV)? IIIC IV Va (V) Vb VI Indices of heteroeeneitv

of heterogeneity

29.6

(III)

22.6 17.2 12.3 7.5

9.6

12.9

14.3%

-7.1

8.3 2.9 -2.0 -6.8

67.2

51.6

5.4 3.7 -1.1 -3.1 -4.5 -6.0

34.3

29.6

*Percentages within 10% of peak values. **Peak values were associated with layer Illa-b or layers from which it is differentiated. tLavers in parentheses are for those in areas which are morlv differentiated. The dashed line emphasizes iayers C-IV, III and IIIa-b and the superficial layers. .

binding in layer II and very low binding in layers IIIc-Vb. Ketanserin binding. The laminar profile of specific ketanserin binding is presented in Table 3. Although a peak in binding occurred in layer 111~ at 13.1 + 0.7%, it was essentially part of a plateau in binding which occurred in layers 11-111~ at around 13%. Binding in layers I, IV and Va was close to 11%. The index of heterogeneity was only 8.9, suggesting a very close association of the distributions of paroxetine and ketanserin binding. In fact, the only major difference was in layer Ia where ketanserin binding was 10.9 f 0.9%, while that for paroxetine in the same layer was at 12.9 f 0.9%.

Opioid receptor binding. The laminar distribution of DAGO and DPDPE binding was essentially the same with the peak for each in layer II, as shown in Table 5. The index of heterogeneity, however, for DPDPE binding was more than twice that for DAGO binding. Specific dynorphin-A binding was unique when compared to all ligands analysed in this study. It had a single peak in layer Ia and moderate binding in layers Ic and II. The index of heterogeneity was 22.6 for dynorphin-A binding. The only other ligand with peak binding in layer Ia was paroxetine and, as already noted, paroxetine binding was also at peak levels in layer IIIc and had an index of only 9.8. Beta adrenoceptor binding. The distribution of specific ( - )-cyanopindolol binding is presented in

Peak specific binding of muscimol was associated with layer IIIa-b in all cingulate areas, as shown in Table 4. In areas 29 1 and 29m, where layer IIIa-b was not differentiated from other superficial cell layers, muscimol binding was highest in that layer which contained the undifferentiated layer IIIa-b, including layers II-IV and III, respectively. Progressively more differentiated areas had secondary peaks in adjacent layers including layer II in areas 30 and 23a and

Layer la IC II Illa-b IIIC IV Va Vb VI Indices of heterogeneity

xt l1.6* l2.7* 12.9** I I.5 II.2 10.9 IO.5 9.6 9.2

of heterogeneity DPDPE

DAGO 11.1% 0.5 I.6 1.8 0.4 0.1 -0.2 -0.6 - I.5 - 1.9

% 11.9 13.7* 14.0** 13.3* II.5 IO.1 8.6 9.0 8.2

8.6

*Percentages within 10% of peak values **Peak values. TPercentage of all binding in each layer.

between

Opioid and beta adrenoceptor binding

GABA, receptor binding

5. Indices

the border

layers II and IIIc in area 23b. The indices of heterogeneity progressively decreased from a high of 67.2 in area 291 to 20.3 in area 23b. This systematic decrease was due to an increase in the number of cytoarchitectural layers, broadening of the layer IIIa-b peak to adjacent layers II and 111~and an overall reduction of the proportion of layer I as a percentage of the full cortical thickness.

high

Table

20.3

for area 23a Dynorphin-A

1 I .I % 0.8 2.6 2.9 2.2 0.4 ~- I.0 -2.5 -2.1 -m2.9 17.4

% 14.8** 13.9; 14.3* 12.7 10.6 9.6 8.6 7.5 1.9

11.1% 3.7 2.8 3.2 I.6 -0.5 - I.5 -2.5 -3.6 -3.2 22.6

Cyanopindolol %

11.1%

10.6 I I.2 l2.8* 14.1** 12.0 10.6 10.0 9.9 9.0

-0.5 0.1 I.7 3.0 0.9 -0.5 -1.1 - I.2 -2.1 11.1

Receptor localization in human cinguhte cortex Table 5. It was similar to that for muscimol as both had peaks in layer IIIa-b and moderate binding in layer II of area 23a. The index of heterogeneity. however, was quite low for (-)-cyanopindolol binding at 11.1.

There are only a limited number of ways to analyse the structure and connections of the human brain and most of these approaches cannot be applied systematically because they depend on the rare occurrence of a relatively localized lesion and appropriate post-lesion survival times. Localization studies of neurotransmitter receptors have shown significant heterogeneities in neocortical receptor distributions and experimental studies have demonstrated associations between some receptors and particular connections and dendritic processes. It is proposed here that the laminar djstribut~ons of receptors provide a relatively direct and repeatable means of probing the cellular and connectional organization of the hutnan brain. As will be discussed shortly, OXOM~pirenzepine labeling of M2 a~tylcholine heteroreceptors may be a useful probe for thalamic afferents and paroxetine binding to 5-H-r uptake sites for raphe projections, while there are a number of probes for purely postsynaptic sites which are differentially localized in superficial cortical layers including pirenzepine labeling of M,, muscimol labeling of GABA,, ketanserin labeling of 5-HT: and DPDPE labeling of delta opioid receptors.

Before discussing specific localization issues, three potential drawbacks for using receptor autoradiography for studies of cortical architecture should be considered. Firstly, few receptors are exclusively pre- or postsynaptic in location. In order to circumvent this problem in the present study particular ligands were used which are relatively pure probes according to experimental studies including 0X0-M/ pirenzepine and paroxetine for presynaptic sites. A~iditi~nally, heterogeneity analysis often emphasizes pre- or postsyna~ti~ components of binding profiles because of the preferential laminar termination of an afferent which results in the enrichment of receptors in the same layer. Secondly. muscarinjc a~etylcholine receptors and 5-HT uptake sites have been identified on glial cells’4.“0and not just neuronal components of the neuropil. Removal of all serotoninergic afferents to posterior cingulate cortex leaves approximately 30% of paroxetine binding to uptake sites intact9 It is possible that as much as 30% of the binding to some neurotransmitter receptors is due to glial cells. Once again. however, since giial binding is homogeneously distributed thorughout cortex, it does not interfere with conclusions based on heterogeneity analysis. Thirdly, it is possible that quenching of the tritium signal alters the relative patterns of binding. It is

171

known that high concentrations of lipids can reduce the apparent binding of a tritiated jigand.‘,‘” There are different proportions of lipids in the cerebral cortex. Layer I, for example, has a particularly high density of myelinated axons and the constituent lipids might artificially reduce the tritium signal in this layer. One way of overcoming this problem is by calculating the index of heterogeneity on the basis of another tritiated ligand as was done with OXOM~pirenzepine binding when the pirenzepine pattern was subtracted from it. This procedure partially accounts for quenching because it is a part of the pirenzepine pattern of binding. Furthermore, a number of iigands had peak specific binding in layer I, including paroxetine and dynorphin-A, so that in spite of quenching a robust signal was preserved in this layer. Only area 29, where the stria of Lancisi is at its maximum,” is it possible that lipids in layer I massively reduced the tritium signal. It is also possible, however, that reduced layer I binding in area 29 also accurately reflects the distribution of postsynaptic receptors since there may be very few apical dendrites in this layer.‘$

In human cingufate cortex pirenzepine binding was highest in layer II and had a low index of heterogeneity as was true for the rat” and rabbit.” Neurotoxin lesions in the rat. which removed layers II&IV, greatly reduced pirenzepine binding not only in these layers but also in layer I, where most apical tuft dendrites of pyramidal neurons are located? In light of the strong similarity in pirenzepine binding across species and the latter observations in rat, it is likely that the bulk of piren~epine binding in human cortex is postsynapti~ in location and primarily associated with the dendrites of neurons in superficial layers. Indices of heterogeneity are always higher for 0X0-M binding than for pirenzepine and in rat at least half of this heterogeneity is due to receptors that are located on anterior thalamic afferents.‘” Peak binding of OXO-M, S/pirenzepine in human cingulate cortex occurs in layer IIIc with moderate amounts of binding abo in layer IV. These are the same layers in which the medial pulvinar nucleus terminates in monkey cingulate cortex.4 We conclude, therefore, that heterogeneities in 0X0-M. 5~pirenzepin~ binding serve as an excellent measure of the dist~but~on of thalamic axons in the human cerebral cortex. One possible confounding factor is that M2 sites have been iocalized to cholinergic terminals of basal forebrain neuronsZ4 Henke and LangIs have shown that in cases of Alzheimer’s disease the enzyme choline acetyltransferase is reduced in layers II and III of cingulate cortex and rodent studies suggest a broad laminar distribution of the projections of these neurons? Thus it is unlikely that the layer IIIc peak in OXU-M, Sjpirenzepine binding can be attributed to presynaptic autoreceptors on basal forebrain axon terminals.

172

H. A. VOW c/ ~11.

There is evidence for two pharmacologically unique subtypes of M, acetylcholine receptors. One site has a high affinity for the antagonist AF-DX 116 and constitutes 15% of all cortical muscarinic acetylcholine sites, while the other site has a low affinity for AF-DX 116 and constitutes 35% of cortical muscarinic sites in rat.‘“.41 Pharmacologically unique M1 sites may also occur in human cerebral cortex. Thus, AF-DX 116 binding peaks in layer IIIa-b, not layer 111~as is true for 0X0-M, S/pirenzepine binding, and AF-DX 116 has an index twice as high as that for the 0X0-M, S/pirenzepine index. The divergence in these two distributions suggests the presence of Mz subtypes. Additionally, the 5 nM and 0.1 nM OXOM/pirenzepine protocols produce two different laminar patterns in binding. 0X0-M, S/pirenzepine has a peak in layer IIIc, moderate levels in layers IV and Va and a relatively high index of heterogeneity at 16.4. As already noted, it is this distribution which most closely approximates the distribution of thalamic afferents. In contrast, specific 0X0-M, O.l/pirenzepine binding has a peak in layer 111~ but is extremely homogeneous with moderate binding in layers Ia-IIIa-b and an index of only 10.7. Neither this pattern nor that for AF-DX 116 approximates the termination zone of thalamic afferents. Serotonin receptors Paroxetine is a selective inhibitor of 5-HT uptake in vitro’4,25 and ablations of the raphe nuclei or serotoninergic terminals in rat reduce peak paroxetine binding in layer Ia by 70%.’ The remaining post-lesion component of paroxetine binding is likely to be associated with glial cells and is homogeneously distributed in all layers. Thus, we propose that heterogeneities in paroxetine binding delimit the distribution of raphe projections to cortex. In human cortex paroxetine binding occurs in a broad distribution in layers 1-111~ with small peaks in layers Ia and 111~. Heterogeneity analysis provides a basis for evaluating matches and apparent mismatches between raphe afferents and various postsynaptic receptors. Of the ligands analysed the ketanserin distribution most closely matched that for paroxetine. Ketanserin binding was homogeneous in layers I-III and had a small index of heterogeneity of 8.9. Since lesions which remove cortical serotoninergic axons do not alter ketanserin binding in rats,12.‘” it is likely that ketanserin binding is associated primarily with postsynaptic 5-HT, receptors. Therefore, there is a high correspondence in the laminar distributions of presynaptic 5-HT uptake sites and postsynaptic 5-HTz receptors. In contrast to the above noted correspondence, the binding of 5-HT is quite heterogeneous with an index of 20.8, which is 2.3 times greater than that for ketanserin. The heterogeneity of 5-HT binding is due to a major peak in layer II and a minor peak in layer VI. The mismatch between 5-HT uptake sites and

5-HT binding is due to the distribution of 5-HT,, receptors, since specific 8-OH-DPAT binding had an even greater peak in layer II and a very high index of 50.3, or 5.7 times that of the ketanserin distribution. These observations suggest that 5-HT,, receptors in layer II may receive a proportionately small serotoninergic innervation in contrast to 5-HT2 receptors. However, the high concentration of 5-HT,, receptors in layer II is likely to assure that 5-HT,,-mediated responses such as inhibition of adenylate cyclase activity” and neuronal hyperpolarization’ ‘” will predominate over 5-HT,-mediated responses such as phosphoinositide hydrolysis’ and neuronal depolarization.‘0 Intrinsic cortical organization Zilles43 proposes that analysis of the co-distribution of neurotransmitter receptors is one means of assessing the possible interactions among transmitter systems and measures co-distribution with Spearman rank correlation tests. According to this type of analysis in human visual cortex GABA, receptors co-distribute with M, acetylcholine and 5-HT, receptors. The present study of cingulate cortex shows that GABA, receptors are very heterogeneous in comparison with the other two classes and that each peaks in a different cortical layer: layer II for M, acetylcholine, layer IIIa-b for GABA, and layer 111~ for 5-HT,. These differences are likely to be due to the organizational variations between these two cortical areas and may suggest important functional differences as well. Another issue to be considered is whether or not co-distribution is a prerequisite for interactions among transmitter systems. It is possible that co-distribution is not necessary for interactions because they can occur on the dendrites of cortical neurons which span two or more layers. For example, a layer IIIa-b pyramidal neuron may have a high density of GABA, receptors on its soma and proximal dendrites in layer IIIa-b. Its apical dendrites in layer 11 may have many M, acetylcholine receptors in addition to some GABA, sites. Finally, its apical tuft dendrites in layer I could have many kappa opioid receptors. The instance in which co-distribution is required for transmitter interactions is when two receptors share the same protein for signal transduction such as GABA, and 5-HT,, receptors sharing the same guanine nucleotide binding protein.’ Classes of postsynaptic receptors are not strictly co-distributed in human posterior cingulate cortex suggesting variations in the functional architecture of different cortical layers. Thus, kappa opioid receptors are most dense in layer Ia, while M, acetylcholine, 5-HT,,, and mu and delta opioid receptors peak in layer II. Finally, GABA, and beta adrenoceptors both peak in layer IIIa-b. Acknowledgements-We thank Eugene L. Jensen and Ethan M. Herschman for their assistance in preparing and quantifying the autoradiographic material. This research was supported by NIH-NINDS grant no. NS18745.

Receptor

localization

in human

cingulate

cortex

173

REFERENCES

R. J., Bell R. D., Yu .I. and Renthal A. (1981) Quantitative measurement of local 1. Alexander G. M.. Schwartzman cerebral metabolic rate for glucose utilizing tritiated 2-deoxygiucose. Brain Res. 223, 59-67. 2. Andrade R., Maenka R. C. and Nicoll R. A. (1986) A G protein couples serotonin and GABA, receptors to the same channels in hippocampus. Science 243, 1261-1265. distinct actions of serotonin on single pyramidal neurons of 3. Andrade R. and Nicoll R. A. (1987) Pharmacologically the rat hippocampus recorded in dro. J. Phtisiol. 394, 99-124. 4. Baleydier C. and Mauguiere F. (1985) Anatomical evidence for medial pulvinar connections with the posterior cingulate cortex, the retrospieniai area, and the posterior pa~hippocampal gyrus in monkeys. J. camp. Neuraf. 232, 219-228. of the central pain pathway in man: the conscious appreciation of pain. Bruin 80, 5. Bowsher D. (1957) Termination 6066622. 6. Braak H. (1979) Pigment architecture of the human teiencephalic cortex. IV. Regio retrospleniaiis. Cc/l 7%. Rex. 204, 43 l-440. 7. Conn P. J. and Sanders-Bush E. (1986) Regulation of serotonin-stimulated phosphoinositide hydrolysis: relation to the serotonin S-HT, binding site. f. Neurosci. 6, 366993675. 8. Corms R., Soriano E., Pazos A., Probst A. and Palacios J. M. (1988) Autoradiography of antidepressant binding sites in the human brain: localization using ‘H-imipramine and 3H-paroxetine. Neuroscience 27, 473496. 9. Crino P. B., Vogt B. A., Volicer L. and Wiley R. G. (1990) Cellular localization of serotonin IA, IB and uptake sites in cingulate cortex of the rat. J. Pharmac. esp. Ther. 252, 651-656. 10. Davies M. F., Deisz R. A., Prince D. A. and Peroutka S. J. (1987) Two distinct effects of Shydroxytryptamine on single cortical neurons. Brain &ss. 423, 347-352. Il. De Vivo M. and Maayani S. (1986) Characterization of the 5-hydroxytryptamine,. receptor-mediated inhibition of forskoiin-stimulated adcnylate cyciase activity in guinea pig and rat hippocampal memb~nes. J. Pharmuc. exp. T&r. 238, 248-253. 12. Fischette C. T.. Neck B. and Renner K. (1987) Effects of 5,7-dihydroxytryptamine on serotonin, and serotonin, receptors throughout the rat central nervous system using quantitative autoradiography. Brain Rex. 421, 263-279. 13. Giraldo E., Hammer R. and Ladinsky H. (1987) Distribution of muscarinic receptor subtypes in rat brain as determined in binding studies with AF-DX I16 and pirenzepine. L(fe Sci. 40, 8333840. 14. Habert E., Graham D., Tahraoui L., Claustre Y. and Langer S. (1985) Characterization of H-paroxetine binding to rat cortical membranes. Eur. J. Pharmac. 118, 107-I 14. IS. Henke H. and Lang W. (1983) Choljnergic enzymes in neocortex. hippo~mpus and basal forebrain of non-neurological and senile dementia of Alzheimer-type patients. Brain Res. 267, 281.-291. 16. H&Ii E. and H&h L. (1988) Autoradiographic localization of binding sites for muscarinic and nicotinic agonists and antagonists on cultured astrocytes. Expi Brain RPS. 71, 450-454. 1-i. Hoyer D.. Pazos A., Probst A. and Palacios J. M. (1986) Serotonin receptors in the human brain. I. Characterization and autoradiographic localization of 5-HT,, recognition sites. Apparent absence of SHT,, recognition sites. Brain Res. 376, 85-96. 18. Hoyer D., Pazos A.. Probst A. and Palacios J. M. (1986) Serotonin receptors in human brain. Il. Characterization and autoradiographic localization of S-HT,, and 5-HT, recognition sites. Bruin Res. 376, 97.-107. 19. Hyman B. T., Kromer L. J. and Van Hoesen G. W. (1988) A direct demonstration of the perforant pathway terminal zone in Alzheimer’s disease using the monoclonal antibody ALZ-50. Brain Res. 450, 392--397. 20 Kimelberg II. K. and Katz D. M. (1985) High affinity uptake of serotonin into immunocytochen?ically identilied astrocytes. Science 228, 889-890. ?I Kuypers H. G. J. M. (1958) Corticobular connexions to the pons and lower brain-stem in man. Brain 81, 364.-388. 22 Locke S., Kruper D. C. and Yakovlev P. I. (1964) Limbic nuclei of thalamus and connections of limbic cortex. Arch.7 Akurol. Ii, 57 I-582. 33 Locke S. and Yakovlev P. 1. (1965) Transcallosal connections of the cingulum of man. Arrhs h&d. 13, 471 486. 24 Mash D. C., Flynn D. D. and Potter L. 7. (1985) Loss of MZ muscarinic receptors in the cerebral cortex in Alzheimer’s disease and experimental cholinergic denervation. Science 228, I 115--lI 17. 25. Mellerup E. T. and Plenge P. (1986) High affinity binding of ‘H-paroxetine and ‘H-imipramine to rat neuronal membranes. P.~~*chopharmacoio~ia89, 436-439. 26. Mesulam M.-M. (1979) Tracing neural connections of human brain with selective silver impregnation. Art&r Ncurol. 36.814-818. 27. Pazos A.. Probst A. and Palacios J. M. (1987) Serotonin receptors in the human brain-IV. Autoradiographic mapping of serotonin-2 receptors. Neuroscience 21, 123-139. 28. Pazos A., Probst A. and Palacios J. M. (1987) Serotonin receptors in the human brain-III. Auto~diog~phic mapping of serotonin-1 receptors. Neuroscience 21, 97-122. 29. Plager M. D. and Vogt B. A. (1988) Mu and delta opioid receptor binding peaks and kappa homogeneity in the molecular layers of rat hippocampai formation. Bruin Res. 460, 150-154. 30. Quik M. and Azmitia E. (1983) Selective destruction of the serotonergic fibers of the fornix-hmbria and cingulum bundle increases 5-HT, but not S-NT2 receptors in rat midbrain. Eur. J. Pharmac. 90, 377-384. 31. Rundles R. W. and Papez J. W. (1937) Connections between the striatum and substantia nigra in a human brain. ,ilr&s Nwd. P.yvchial. 38, 550-563. 32. Saper C. (1984) Organization of cerebral cortical afferent systems in the rat. II. Magnocellular basal nucleus, J. conrp, Neurof. 222, 343 352. 33. Slesinger P. A., Lowenstein P. R., Singer H. S., Walker L. C., Casanova M. F., Price D. L. and Coyle J. T. (1988) Development of B, and & adrenergic receptors in baboon brain: an autoradiographic study using “‘l-iodocyanopindolol. J. conrp. ~~u~[)i. 273, 31S329. 34. Torch W. C., Hirano A. and Solomon transneuronal degeneration in the limbic system: . . __S. (1977) _~ Anterograde /. cllnico-anatonllc correlatron. Neurofog_r 27, I 157-I 163. 35. Vogt B. A. (1976) Retrosplenial cortex in the rhesus monkey: a cytoarchitectonic and Golgi study. J. camp. Neural. 169, 63-98.

174

H. ,A, vo(;r

C’l


36. Vogt B. A. (1985) Cingulate cortex. In (‘ereh~crl C’ortt2.vteds Peters A. and Jones E. G.), Vol. 4. pp. 89 li’) Plenum Press, New York. 37. Vogt B. A. (19X8) Localization of cortical muscarinic receptor subtypes. Tier& pharmuc. &i. SuppI. III, J9 5.3. 38. Voet B. A. and Burns D. L. (1988) Experimental localization of muscarinic receptor subtypes . to cingulntc cortical affe&rts and neurons. J. Nkosci. ‘8, 643 652. 39. Vogt B, A. and Hedberg T. G. (1988) Autoradiographic localization of muscimol and baclofen binding in rat cmgulate cortex. E.$ Bruin Rex 71, 208-214. 40. Vogt B, A., Pandya D. N. and Rosene D. L. ( 1987) Cingulate cortex of rhesus monkey. I. Cytoarchitecture and thalamic afferents. J, camp. )Verrroi. 262, 256 270. 41. Wang J. X., Roeske W. R., Gulva K.. Wang W. and Yamamura H. I. (1987) ‘H-AF-DX 116labeis subsets of muscarinic cholinergic receptors in rat b&in and heart. Lisp Sci. 41, 1751.-1760. 42. Young W. S. III and Kuhar M. J. (1979) A new method for receptor autoradiography ‘H-opioid receptor labeling in mounted tissue sections. Bruin Res. 179, 255-270. 43. Zilles K. (1990) Codistribution of receptors in the human cerebral cortex. In Receptors in the Human Nervous Sq’stem (eds Mendelshn F. A. 0. and Paxinos G.). Academic Press, New York. 44. Zilles K., Armstrong E., Schaug G. and Schleicher A. (1986) Quantitative cytoarchitectonics of the posterior cingulate cortex in primates. J. camp. Neurol. 253, 514-524. 45. Zilles K.. Schleicher A., Rath M. and Bauer A. (1988) Quantitative receptor autoradiograph~ in the human brain. Methodical aspects. ~i.~r~~~~?e?~j.~rr~, 90, 129 137. (Accepted

17 Jutwar,)’ 1990)