Meuroseienee Vol. 18, No. 2, pp. 329-346, 1986 Printed in Great Britain

03~522186 %3.00 + 0.00 Peqamon Joumais Ltd

0 1% IBRO

DISTRIBUTION OF SOMATOSTATIN RECEPTORS IN THE HUMAN BRAIN: AN AUTO~DIO~~PHIC

STUDY

J. C. REUBI,* R. CORTI~, R. MAURER, A. PRomt and J. M. PALACIUS

Preclinical Research, Sandoz Ltd., CH 4002 Etaset, S~~rl~~ tDepartment of Pathology, Division of Ne~o~~olo~, University of Rasel, CH 4002 Basel, Switzerland

Ahatraet-High affinity somatostatin receptors have been measured in postmortem brains from 18 neurologically asymptomatic patients (mean age: 67 years) using the stable somatostatin analog 12rI-204-090, D~he-~s-‘~yr-D’ltpLy8-‘lbr-~~s-~(ol), as radioligand. In homogenates from human frontal cortex, high affinity (& = 0.52 nM; B,, = 557 fmol/mg protein) receptors with pharmacological specificity for somatostatin, [~T~~matos~tin and somatostatin-28 were found. The CNS distribution of these receptors was studied by auto~~o~phy. ~matostatin receptors were dist~but~ in varying densities throughout the whole brain. High concentrations are found in all cortical layers, the deeper layers (V-VI) being usually more dense than the superticial layers (I-III). The limbic system is heavily labeled, in particular hippocampus (CAl, dentate gyms), most of the nuclei of the amygdala, and the habenula. Also parts of the basal ganglia am very rich in somatostatin receptors: the nucleus caudatus as well as the nucleus accumbens are very dense, whereas the globus pallidus is virtually unlabeled. Interestingly, significant amounts of somatostatin receptors are found in the human cerebellum, which is devoid of endogenous ~matos~tin. Other discrete areas of the CNS are enriched with somatostatin receptors: locus coeruleus, tuberal nuclei of the h~~~arnus, claustum, tuberculum olfactorium as well as spinal trigeminal nucleus and substantia gelatinosa of the spinal cord. The substantia innominata is poor in somatostatin receptors. In general there is a good correlation in the distribution of somatostatin receptors in the human and rat brain and there is a reasonable correlation with endogenous somatostatin levels in human brain tissue, particularly in the larger structures.

The very high density and the specific localization of somatostatin receptors in strategic key points in the CNS such as cortex, basal ganglia, limbic system and substantia gelatinosa suggests an important role of ~~tos~tin in cognitive, sensory and extrapy~~dal motor functions. The ~~ifi~~ of somato- statin receptors in the human cerebellum remains to be elucidated.

The tetradecapeptide somatostatin, first isolated by Brazeau et at.4 on the basis of its growth hormone release in~bi~ng properties, has been shown to have not only multiple peripheral effects, i.e. in pituitary, pancreas or gut, but also extensive effects within the CNS. Strong evidence suggests that in mammals it acts as a neurotransmitter or neuromodulator.‘“36 In the human brain, the presence of somatostatin has been determined in postmortem material using either sensitive radioimmunoassays or immunohistochem- istry. The peptide appears to be widely distributed within the human CNS, being localized, in addition to the hypothalamus, mainly in cortical and limbic structures. The basal ganglia and specific regions such as the locus coeruleus, inferior colhcuIus, peri- aqueductal grey and substantia gelatinosa of the spinal cord are also rich in somatostatin immu- noreactivity.2*5*9*‘4.2” A limited number of immu- nohistochemical studies have shown the localization of somatostatin within neuronal elements in these regions. 10,25,#,su,55

Increasing evidence suggests that various CNS

*Address for correspondence: Sandoz Research Unit, Wan- der Research Institute, P.O. Box 2747, CH 3001 Bern, Switzerland.

diseases are associated with disturbances of CNS somatos~tin levels. Indeed, in ~stmortem brains or in cerebrospinal fluid% from patients with CNS dis- orders as varied as Alzheimer’s disease, Huntington’s chorea, multiple sclerosis, schizophrenia, depression or parkinsonian dementia, abnormal levels of somatostatin have been measured.1*3~8*‘z~13*z’~3’~47~49~s’ Moreover, CNS-active drugs such as carbamazcpine can alter somatostatin levels in the ~rebrospinal fluid.35 Finally, a close relationship has been found between somatostatin neurons and pathological pro- cesses such as neuritic plaques and neurofibrillary tangles seen in Alzheimer’s disease.29*4s To better understand the pathophysiolo~cal mechanisms in such diseases, a better knowledge of the human CNS somatostatin system is necessary.

In animals, somatostatin has been shown to act through high affinity and saturable receptors specific for somatostatin located in various regions of the CNS.n*43*s2 These studies were performed using various kinds of somatostatin analogs as iodinated radioligands, such as lTyr’]-somatostatin, fryr* ‘I- somatostatin, (I.&,D-TrpZ2,Tyrzs]somatostatin-28 or the stable somatostatin analogs CGP-23-996 or 204-090. 7~1’*37~3**43*s2 Recent autoradiographical studies have permitted a detailed visualization of these recep-

329

330 J. C. hJB1 et al.

tors in the rat brain.25,27.39,53 High densities have been found in particular in all cortical areas, the hippocampus, the amygdala, the septum, the locus coeruleus, the inferior colliculus and the substantia

gelatinosa of the spinal cord. In the human, no detailed information concerning

specific CNS somatostatin receptors has been re- ported so far, although human pituitary somatostatin receptors are described? We report here the detailed

mapping of somatostatin receptors in postmortem human brain using a membrane binding assay and quantitative autoradiography techniques at the light microscopic level using ‘2sI-204-090 as radioligand. This ligand was chosen due to its particularly high resolution for the measurement of somatostatin receptors.38*39

EXPERIMENTAL PROCEDURES

Tissue preparation

Human brains from a total of 18 patients (8 males and IO females, mean age 67 years) were obtained at autopsy after an average postmortem delay of 6 h. None of the cases studied showed clinical or neuropathological evidence of a brain disease. Individual characteristics of these subjects, including sex, age, interval between death and brain freez- ing, premortem condition and cause of death are listed in Table 1. After removing the brain, several 4-6mm thick tissue blocks from one cerebraI and one cerebeIIar hemi- sphere, as well as tissue blocks from the brainstem and the upper cervical cord (Cl-C2 segments), were stored at -80°C to be later cut in a cryostate-microtome or proceased to membrane homogenates. The remainder of the CNS was immersed in buIfered formaldehyde (4%) and further processed for routine neuropathological examination.

Radiolabeling o/ the ligand

The [Ty*SMS 201-995 analog u)4090 [H-DPhe- Cys- I yr-D’t rp-~ys-~~~tu-Cjs-Thr(ol)] has been iodinated by the chloramine T method and purified using high pressure

liquid chromatography according to Reubi.‘* The high pressure liquid chromatography purification of the radio- ligand allows the assumption that its specific activity is near the theoretically calculated one. The iodinated compound was routinely characterized in in oitro homogenate binding assays from rat cortex before being used in auto- radiography.

Posrmortem studies in rut brain

To test whether somatostatin receptors survive post- mortem modifications of brain chemistry, a preliminary study was performed with rat brains left unfrozen in situ for various periods of time after decapitation. Autoradiography of somatostatin receptors in rat brain tissue was performed as described previou~ly.~~~~~

Binding studies on human brain homogenates

Frozen blocks from frontal cortex of a limited number of control patients listed in Table 1 were taken for membrane homogenate preparation as described previously for rat brain.“,V Binding studies using ‘2’1-204-090 as radioligand were performed as described by Reubi.% In a limited number of experiments, the radioligand ‘zJI-(Tyr”]- somatostatin was used for comparis0n.r’~”

Autoradiography

Frozen tissue blocks were brought to -20°C and moun- ted onto microtome chucks. Ten pm thick sections were cut using a cryostat (Leitz 1720, from L&z, W&Jar, F.R.G.) and mounted onto gelatin coated ghus slides. Somatostatin receptors were labeled as previously deacrIbed.z’39 Tiiue sections were preincubated in Tris-HCl buffer (5OmM pH 7.4) containing CaCI, (2 mM) and KC1 (5 mM) for 10min at room temperature and then washed twice for 2min in the same buffer without addition of salts. Incu- bation was carried out routinely at ambient tempaature for 2 h in Tris-HCI buffer (0.17 M, pH 7.4) containing I% bovine serum albumin, bacitracin (4O&nJ) and MgClr (5 mM) to inhibit endogenous proteases in the presence of the iodinated ligand (0.12 x IO”dpm/ml, approx. 60 PM). Excess ligand was removed during two consecutive Smin washes in cold incubation buffer containing 0.25% bovine serum albumin and a final dip in cold doubIedistiIIed water to remove excess salts before drying the sections under a stream of air.

Table 1. Sources of brain tissue

Case Se.x Age Postmortem delay Immediate cause of death Premortem condition

A B C D

E F

G H I J K L

M

N 0

9

F 49 6h30min Pulmonary embolism F 80 3h Cardiac failure M 51 5h3Omin Cardiac failure F 65 6h Cardiac failure

F 80 5h Cardiac failure F 84 8h30min Cardiac failure

M 53 F 71 M 79 F 48

12h 2h 6h45min 5h 6h3Omin 5h30min

Cardiac failure Bronchopneumonia Cardiac failure Acute pyelonephritis Cardiac failure Bleeding from dissecting aneurysm of the aorta Bronchopneumonia

M 70 M 83

F 60 6h

M 54 6h30min M 92 13 h ISmin

F 61 6h M 58 4h

Chfdiae failure Cardiac failure

Chronic renal failure Ischemic heart disease Ischemic cardiopathy Chronic bronchitis, Pulmonary emphysema. Cor puhnonale Carcinoma of the stomach Pyelonephritis, Arterial hypertension Coronary artery disease Rhabdomyosarcoma of the uterus Carcinoma of the gaIlbIadder Carcinoma of the breast Coronary artery disease Arterial hypertension

Non-Hodgkin’s lymphoma (extra cranial) Carcinoma of the lung Arterial hypertension, Senile cardiac amyloidosis Carcinoma of the lung Carcinoma of the urinary bladder

F 73 3h3Omin BroncboImeumonia Chronic respiratory infection

Human brain somatostatin receptors 331

400-

e Cortex

8,00- ‘5 300 N_$z$;

h

i

Thalamus loo- * -, ~ ,

0 , I I I 0 8 16 24

Postmortem time ( h )

CAI

Fig. 1. Postmortem stability of SRIF receptors in 3 different regions of the rat brain. Brains from rats left at 4°C for 0, 4, 8, 16 and 24 h after decapitation were processed for autoradiography using L251-204-090 as radioligand. Receptor densities were measured with densitometry and are

expressed as dpm per mg protein.

Sections were apposed to ‘H-sensitive Ultrotilme and exoosed for 1 week in X-rav cassettes.% Non-specific bind- ---r---- ~~

ing was determined by admixing unlabeled 2&&090 at a concentration of 1 PM. Quantification of autoradiograms was performed using a computer-assisted image processing system, as previously described.6 Lacking tissue standards for iodinated compounds, we used ‘H-standards prepared with pig brain grey matter to calibrate the densitometer.” Thus, the values do not represent exact receptor densities. Brain areas and nuclei were identified using atlases of the human brain.32*33

Materials Nau51 was obtained from New England Nuclear

(Dreieich, F.R.G.), bovine serum albumin (Research grade) from Serva and ‘H-sensitive Ultrolilms from LKB (Sweden). 204-090 was synthesized by W. Bauer in our chemical research laboratories.

RESULTS

Postmortem studies in rat brain

Somatostatin binding sites in rat brain are not altered under postmortem conditions (Fig. 1). If brains were left in situ for 24 h after decapitation while stored at 4”C, no changes in distribution or density of somatostatin binding sites were seen.

Binding of 1251-204-090 to human brain membranes

‘251-204-090 has saturable binding to a human cortical membrane preparation taken from a frontal cortex sample of case R (Fig. 2). Scatchard analysis reveals a high affinity component with a K,, of 0.525 nM and a B_ of 557 fmol/mg protein. The time course of association, shown in Fig. 3 for the same membrane preparation, reveals that equilibrium is reached at approximately 45-60min incubation time. Figure 4 shows the pharmacological specificity of 1251-204-090 binding sites for somatostatin. The biologically active compounds somatostatin-28, [DTrp’]-somatostatin, somatostatin-14, and 204-090 have all a high aIIinity for ‘251-204-090 binding sites, [DTrpr]-somatostatin and somatostatin-28 being the most potent. The biologically inactive compounds, however, such as [Des-Trp*]-somatostatin or somatostatin-28( 1 - 12), or the unrelated peptide Luteinizing-Hormone-Releasing-Hormone (LHRH) have no affinity for the binding sites.

Autoradiography

Somatostatin receptors were heterogenously dis- tributed throughout the human brain. High densities of receptors were found in some grey matter struc-

0.15

0.10

0.5 1.0 1.5 0.01 0.03 0.05 ‘0.07

‘25I-2O4-O9o (nM) 0 (nM)

Kd : 0.525 nM

E mol : 557 fmoles/mg protein

Fig. 2. Binding of ‘251-204-O!90 to human cortical membranes. (a) Saturation curve using a human cortex membrane prepartion (case R) incubated for 45 min with increasing concentration of ‘251-204-090. Specific binding is total binding minus binding persisting in presence of 108 nM 204-090 (non-specific). Points are average of triplicate tubes. (b) Scatchard plot from the data shown in (a). yd =0.525 nM;

B_ = 557 fmol/mg protein.

332

30 60 90

Time Imin)

Fig. 3. Time curve of apa!& binding of ‘251-#)4-090 to human cortical membranes. 70,OOOcgm were iacuktod with cortical membrane preparation (approx. 12 pg proti)

for various periods of time.

tures. while in white matter tracts no specific binding was detected.

In several brain regions the densities of somato- statin receptors were determined semiquantitatively using ‘H-standards, and are summarized in Table 2.

The non-specific binding, as determined by the addition of 1 PM of the somatostatin analog 204490, was homogenous in all regions studied, ranging from 20,000 to 40,000 dpm/mg protein, with the exception of the cell bodies of the locus coeruleus and the substantia nigra, where it was markedly higher. The

J. C. REULU et al.

01 I IO nM(pepti&e)

Fig. 4. Competitive displacement of ‘%204-O!% by various SRIF analogs. Human cortical mc&mztm (e to 12 fig protein) wewinahted for 45 lain *&#W+n rah@nd (‘%-). Total binding ir itI%. Spa&h biading (total bhding minus binding pariitine in preaacc of 1OOnM 2o4-090) is 76% of total. Poiits are average of

O-O, 2O4-090; O-0, =F; A-A, x-x, SRIF-2& @-J# SW-D (l-12);

A-A,, IDaa-Trpc)sRIF; t-a.’ L#&WWi?ormow- RekWng-Hormone (CHRH).

inferior olivary nucleus also presented somewhat elevated kvels of non-specific binding.

One of the highest densities of somatostatin recep- tors measured in the human brain corrqxm&d to the pyramidal cell layer of tbe CA1 hippocan+1 subfield. The dentate gyrus presented high de&ties as well, mostly restricted to the molecular layer. On the other hand, only low levels of bisuling occmred in the CA3 SUW (Fig. 5).

In the amygdala the distribution of SRIF receptors was heterogenous (Fig. 6). High densities of binding

Table 2. Somatoatatin receptor densities in the human brain

Mean density f SEM Brain area N (dpm x 10-y/mg protein)

Basal ganglia and basal forebrain Nucleus caudatus I 187.3 f 14.1 Putamen 8 126.7 + 15.9 Globus pallidus 6 27.6 f 10.0 Nucleus accumbens 2 182.8, 261.9* Substantia innominata 4 33.7 k 14.6 Septum 2 9.9, 59.6’ Claustrum I 147.9 f 24.2

Thalamus Anterior nucleus 3 9.2 + 4.2 Medial nucleus 5 14.3 f 3.8 Nucleus paramediaxms rotundocelklaris 3 77.6 f 11.9 L.atefal liuclear group 3 9.5 f 3.8 Pulvinar nuclei 2 18.9, 23.5* Habenula 2 89.7, 209.6, Nucleus subthalamicus 2 1.1, 7.1*

Hrpotlraaamus Lateralarea 6 26.8k 11.3 Iwerior nucleus 7 62.9 f 18.5 Tuheral nuclei 3 161.2& 11.4 Mammillary bodies 3 6.0& 1.9

Human brain somatostatin receptors 333

Table 2. (cont.)

Brain area Mean density f SEM

N (dpm x IO-“/mg protein)

Amygdala Lateral nucleus Basal nucleus Transitory zone Granular nucleus Accessory basal nucleus Cortical nucleus Accessory cortical nucleus

Hippocampus Dentate gyrus CA3, pyramidal cell layer CA 1, pyramidal cell layer

Cerebral cortex Frontal layers I-III

layers V-VI Temporal layers I-III

layers V-VI Precentral layers I-III

layer V layer VI

Postcentral layers I-III layers V-VI

occipital (A17) layers I-III layers V-VI

Cingular layers I-III layers V-VI

Entorhinal layers I-III layers V-VI

Cerebellum Molecular layer Granular cell layer Dentate nucleus

Midbrain Superior colliculus Inferior colliculus Central grey Substantia nigra Substantia nigra, cell bodies Rapht centralis

Pons Locus coeruleus Motor nucleus of the trigeminal nerve Nucleus reticularis pontis oralis Nucleus reticularis tegmenti pontis Griseum pontis

Medulla oblongata Inferior olive

Main nucleus Medial accessory nucleus Dorsal accessory nucleus

Formatio reticularis lateralis Solitary nucleus Hypoglossal nucleus Nucleus medullae oblongatae centralis Spinal nucleus of the trigeminal nerve

Substantia gelatinosa Nucleus proprius

Nucleus cuneatus medialis

Cervical cord Substantia gelatinosa Motor cells

91.4k6.5 135.4 f 8.9 73.7k21.1

121.1 f 11.4 83.1 f 16.9 44.1, 88.3*

173.6 f 14.1

5 184.8 f 13.3 5 80.5* 11.7 5 234.3 f 21.0

: 8 8 3 3 3 2 2 4 4 5 5 6 6

105.8 f 19.4 179.1 * 15.0 103.9 f 15.8 166.9 f 26.2 112.6 f 16.0 93.8 f 19.7

166.1 f 17.5 30, 110’

152, 155’ 33.2 f 15.9

122.2 + 23.8 108.5 + 8.9 201.6 f 6.6 113.2 + 5.5 244.3 f 5.5

2 2 2

3 2 5 4 3 4

3 2 4 2 3

3 3 4 3 3 4 4 5

5 5 5

4 4

69.1, 70.4* 28.4, 37.5*

6.3, 16.0*

10.0 + 3.6 8.2, 8.5*

17.7 k9.1 11.1 + 1.8 54.1 f 2.3 45.8 f 14.1

76.1 + 10.4 9.9, 18.1*

12.3 f 4.7 4.5, 5.5;

10.3 + 3.9

26.4 + 7.1 25.2 * 11.7 23.3 f 6.2 46.9 f 21.1 46.9k21.1 37.5 f 12.4 21.6f3.1 14.2 rt 1.7

60.0 f 10.9 18.5 + 4.9

1.3 f 0.9

106.5 f 13.3 12.7 f 3.5

*Individual values are given.

334 J. C. &UBl et al.

sites were present in the accessory cortical nucleus. The granular, basal, accessory basal and lateral nuclei contained intermediate densities. Other regions in- cluding cortical nucleus and transitory zone were associated with low levels of binding.

Most thalamic nuclei (Fig. 7) including anterior, dorsomedial, ventral, lateral, pulvinar and intra- laminar formations as well as the reticular nucleus and medial and lateral geniculate bodies presented low concentrations of somatostatin receptors. Significantly higher levels of binding were seen in the nucleus paramedianus rotundocehularis of the para- ventricular thalamic formation (Fig. 7), although these were only intermediate to low densities, when compared to the hippocampal CA1 subfield. In the epithalamus, high to very high numbers of somato- statin receptors were present in the medial habenular nucleus (Fig. 8), in contrast to the low densities observed in the lateral habenular nucleus.

In the hypothalamus high densities of binding sites were observed over the tuberal nuclei (Fig. 7), whereas, in other areas such as the anterior, lateral, ventromedial, periventricular and infundibular nuclei as well as the mamillary bodies very low densities were measured. Somewhat higher levels of binding were observed in the dorsomedial and posterior nuclei (Fig. 7).

Intermediate to low densities of silver grains were associated with the preoptic area, both lateral and medial nuclei. More laterally, very low concen- trations of binding sites were observed along the nuclei of the diagonal band of Broca and over the substantia innominata. The septum presented very low densities as well.

The olfactory tubercle, claustum, caudate, puta- men (Fig. 9) and nucleus accumbens had high to very high densities of somatostatin receptors, while only

very low densities were associated with the medial and lateral globus pallidus.

The cerebral cortex was rather enriched in SRIF receptors (Fig. 10). In all cortices analyzed, SRIF receptors were concentrated in the deep layers (lami- nae V and VI), while outer layers (laminae l--III) contained lower densities of receptors. Outer layers of the primary visual cortex (A17) (Fig. IOA) presented very low concentrations of SRIF receptors, values being among the lowest observed in the neocortex (Table 2). In the remainder of the cortex, including the parastriate area (A18) (Fig. lOA), intermediate concentrations of SRIF receptors were seen over these layers. In the inner layers the levels of binding ranged from intermediate (occipital cortex) to high [frontal (Fig. IOB), temporal, pre-, and postcentral cortices (Fig. lOC)] or very high (cingulate gyrus and entorhinal area; Fig. 5).

The cerebellum presented intermediate densities of receptors, the highest level of binding being observed over the molecular layer of the cerebellar cortex (Fig. 11).

In general the brainstem was associated with very low concentrations of SRIF receptors. Slightly higher densities were found in the raphe centralis, the for- matio reticularis lateralis, the solitary nucleus and the locus coeruleus. The latter structure contained the highest density of SRIF receptors detected in the brainstem, although it also presented unusually high levels of non-specific binding. The substantia nigra was poor in SRIF receptors; however, somewhat higher levels of specific binding were present over the cell bodies. Over these cell bodies, as well as over the cell bodies of the locus coeruleus, the non-specific binding was higher than in all other brain areas analyzed.

At the level of the pyramidal decussation,

A ABa ACo

Z CG De

E E Fx G Gr HbL HbM

Abbreviations used in figures

anterior nucleus of the thalamus Hil hilus accessory basal nucleus of the amygdala La lateral nuckus of the amygdala accessory cortical nucleus of the amygdala LH lateral area of the hypothalamus basal nucleus of the amygdala M medial nucleus of the tbalamus caudate nucleus MB mamillary bodies central grey MO molecular layer of the cerebellum dentate nucleus of the cerebellum Pm nucleus paramedianus rotundocelhdaris of the thal- dentate gyms amus dorsomedial nucleus of the hypothalamus PO posterior nucleus of the hypothalamus entorhinal cortex Pro nucleus proprius of spinal trigeminal nucleus fomix Put putamen granular cell layer of the cerebeIlum granular nucleus of the amygdala

T fg.id& layer of the hippocampal CA1

lateral habenular nucleus !Xie substantia &atinosa of the spinal trigeminal nucleus medial habenular nucleus Tub tuberal nuclra of the hypothalamus

Figs 5-12. Bright field photomicrographs of autoradiograms o issues labekd with ‘“I-204-OW. Dark regions represent high densities of ceptors.

e ~ofmyhigh&nsitiesofSRIF 0 *layeroftbs~tegyrW(WJ*

of the dentate gyms (Hil), CA3 subfkld, ofb@?!EEl~iK?IE&Z!~tZ and s&kulum (S) and pman receptors over two small regions (arrows) of ~&al scarring with total loss of nerve cells (old vascular

lesions). Bar = 3 mm.

Fig. 5.

335

336

Fig. 7. Autoradiogram of a coronal section through the diencephalon. In the thalamus low densities of SRIF receptors are present in the anterior (A) and ventral anterior (VA) nuclei; the nucleus paramedianus rotundocellularis (Pm) contains somewhat higher levels of binding. In the hypothalamus SRIF receptors are heterogenousry distributed, densities being very low in the mamillary bodies (MB) and the lateral nucleus (LH), somewhat higher in the posterior (PO) and dorsomedial (DM) nuclei and markedly high

in the tuberal nuclei (Tub). Bar = 5 mm.

331

-. ..-_.- -

Fig. 8. Photomicrograph of au autoradiogram of a section through the e ‘thakmw and t&urn. Note the ptesence of very hi& levels of birkding in the me&al h&em&w nwkus 8 MU@ in conMast t4ie lateral habenula (HbL) which contains low levels of binding. Other structures such as central grey (CG), &turn

and medial thalamic nucleus (M) contain low densities of labeling. Bar = 5 mm.

338

Fig. 9. Photomicrograph from an autoradiogram image of the striatum. Very high levels of binding are observed in the caudate nucleus (Cd) and putamen (Put). Bar = 5 mm.

339

Fig. IO.

340

Fig. 11. Autoradiogram of the cerebellum showing the presence of SRIF receptors in the molecular layers (MO) of the cortex. Lower densities of receptors are observed along the granular layer (G) and over the

dentate nucleus (&I. Bar = 3 mm.

Fig 10. Autoradiographic images of three cortical areas of the human brain. (A) Visual cortex; the arrow indicates the transition from striate (A-17) to parastriate (A-18) areas; (B) frontal cortex (area 10 of Brodmana); (C) pre- aed postcentral cortkes (middle thifd). Note the more dense labeiing over inner

layers (VI) in contrast to more superficiaf layers, in all cortim shown. Bar = 5 mm.

341

Fig. 12. Distribution of SRIF receptors in the medulla oblongata at the level of the pyramidal decussation. Almost no specific binding is present in any structures, with the exception of the substantia gelatinosa

of the trigeminal nucleus (SGe). Bar = 3 mm.

342

Human brain somatostatin receptors 343

significant although low densities of SRIF receptors characterized the substantia gelatinosa of the spinal trigeminal nucleus (Fig. 12).

In the spinal cord the substantia gelatinosa in particular was enriched in SRIF receptors, account- ing for intermediate densities at the first cervical segment. At the same level, very low densities were observed over the motor cells of the ventral horn.

One brain area presenting pathological features has been included in the study. The comu Ammonis of the hippocampus of patient 0 presented two small vascular lesions, which appeared as areas of glial scarring, and were totally devoid of neurons (Fig. 5). A complete loss of SRIF receptors was observed over these lesions.

DISCUSSION

This is the first detailed report describing the presence of SRIF binding sites in the human brain. It shows first in membrane homogenates from human frontal cortex that the binding sites are saturable and of high affinity. Furthermore, they are specific for SRIF since their pharmacological characteristics re- semble those of SRIF receptors measured in freshly prepared membranes from rat brain.3**43 Their phar- macology also correlates with that of SRIF receptors from pituitary’l or from pancreatic /I-cells” and with the corresponding bioassay profiles.‘* This strongly suggests the presence of specific SRIF receptors in the human brain homogenates.

The present autoradiographic results represent the first mapping of the distribution of such high affinity SRIF receptors, labeled with ‘251-204-O!90, in the brain of individuals without evidence of neurological or psychiatric disorders. The study is validated by our investigation of the postmortem fate of rat brain SRIF receptors, which did not show significant alter- ations even after 24 h postmortem storage of the brain at 4°C. In general, a very high concentration of SRIF receptors was found in several regions of the human brain, in particular in all cortical regions, hippocampus, basal ganglia, amygdala, habenula, tuberculum olfactorium, locus coeruleus and sub- stantia gelatinosa of the spinal cord. The highest concentrations were found in the CA1 region of the hippocampus, the deep layers of the cingulate and entorhinal cortices as well as the caudate and accumbens nuclei.

Correlation with SRIF receptor distribution in rat brain

The overall distribution of SRIF receptors ob- served in this study is in agreement with previous results obtained in the CNS of experimental animals by microdissection techniques38~43~52 and by light microscopic autoradiographic techniques.2527*39~53 Like in these studies, we observed that the distri- bution of SRIF receptors was highly regionally localized. Furthermore, a rostrocaudal decreasing

gradient of receptor concentration was found in the whole brain. In particular, a good correlation be- tween man and rat can be seen in the hippocampus, amygdala, habenula, thalamus, hypothalamus, tuber- culum olfactorium, locus coeruleus or substantia gelatinosa of the spinal cord. In the basal ganglia, the SRIF receptor density is clearly higher in human than in the rat, despite similar incubation conditions. Good correlation is found in all cortical areas except for the distribution of SRIF receptors in the superficial cortical layers (I-III), where SRIF recep- tors are virtually lacking in the rat. In contrast to the absence of SRIF receptors in the rat cerebellum and nucleus accumbens, in the human brain, interestingly, these regions, especially the nucleus accumbens, are enriched with SRIF receptors. On the other hand, the human brain has fewer SRIF receptors in the central grey and the inferior colliculus than the rat. Inter- estingly, in the human septum, the concentration of SRIF receptors is extremely low, which is in sharp contrast with the high density found in the rat.39

Correlation with SRIF distribution in human brain

There is an incomplete correlation between the endogenous levels of SRIF in the human brain and the SRIF receptor distribution. Since immunohisto- chemical studies on the localization of SRIF- containing neurons are limited to selected regions like the neocortex or hypothalamus’0*50 a comparison has to be based on reports dealing with microdissection of brain and SRIF measurement with radio- immunoassay.5~9*‘4 Since with this method the SRIF neuronal cell bodies and nerve terminals cannot be considered separately, a correlation study between endogenous levels of SRIF and SRIF receptors is only of limited value. Nevertheless, in the basal ganglia, a very good correlation is found, with high amounts of SRIF and SRIF receptors in the caudate and putamen but low levels in the globus pallidus. The very high number of SRIF receptors in the nucleus accumbens correlate well with immunohisto- chemically stained SRIF fiber network in the rat.19 Good correlation exists also for the amygdala, hippo- campus, locus coeruleus or substantia gelatinosa of the spinal cord, where both high SRIF levels and high SRIF receptor density are found in the human. In the thalamus, substantia nigra, superior colliculus and inferior olive both low SRIF levels and low receptor concentration are detected. In the cortex, where a high density of SRIF-positive neuronal cell bodies and fibers are visualixed”sM through all layers, a high density of SRIF receptors is also found. Although SRIF receptors are found in all cortical layers, there is a clear difference between the superficial layers (I-III) and the deeper layers (V-VI), which are much more massively labeled except in the pre- and post- central region. Furthermore, an often abrupt change of the specific laminar pattern of the SRIF receptor distribution within a single cortical region can frequently be found. This feature, not observed in the

344 J. C. REUSI et al.

rat cortex, is unexplained at the moment, but could reflect the laminar variation of SRIF immu- nohistochemically visualized in human cortex.30 Such an irregular layering has also been observed for opiate receptors in the cortex of monkeys and shown to differ in rat, monkey and man5’ In addition, the concentration of SRIF receptors in the cortical deep layers (V and VI) resembles that of k-opiate recep- tors.” We have, however, excluded previously3’ that the SRIF receptors labeled with ‘2sI-204-090 have any affinity for opiates. Therefore, cortical layers V and VI are rich in both opiates and SRIF receptors, suggesting that the major role of these two neu- rotransmitters in these layers could be the same, namely a major influence on cortical neurons projecting to the thalamus. Pyramidal cells in layer VI also project to the brain stem whose reticular activating system may play a role in the level of alertness influenced by SRIF.16

The substantia innominata, which is the source of cholinergic fibers to the neocortex and is known to degenerate in Alzheimer’s disease,& does not have significant amounts of SRIF receptors.

In several regions there is no correlation between endogenous SRIF and SRIF receptor density. The most striking difference is seen in the hypothalamus, where only small amounts of SRIF receptors are measured, except in the nucleus tuberalis. This low level of SRIF receptors is likely due to the fact that most endogenous SRIF in the hypothalamus acts on extrahypothalamic receptors in the pituitary. It should also be mentioned that there are low receptor levels in the nucleus subthalamicus, colliculus inferior and central grey, all structures with relatively high SRIF levels. On the other hand, the finding of somatostatin receptors in the cerebellum is surprising since this structure was found to be devoid of SRIF.5~‘4 Apart from the cerebellum, this report reveals a few discrete regions with high levels of SRIF receptors, such as the claustrum, nucleus para- medianus rotundocellularis of the thalamus and tu- beral nuclei of the hypothalamus, which had not been associated with SRIF systems before. It should be mentioned that in general a discrepancy between endogenous SRIF levels and SRIF receptors is less pronounced in man than in the rat.)’ A discrepancy between the distribution of the neurotransmitter and the localization of its receptors has been found for many substances,22 especially in the case of pep- tides.26,48 Several explanations have been proposed to justify this fact. 22 It is possible that several members of a peptide family bind to the receptor site, without being reactive to a particular antisera, as has been proposed for the case of substance P.26 The same is probably true for SRIF, since it has been shown that various molecular forms of the peptide are present within neuronal elements. 28 The existence of several subtypes of SRIF receptors must also be kept in mind.27.37.38.54 a minor proportion of them might not be labeled under the present conditions.27,38,39

Cellular localization

An important question is that related to the cellular localization of these binding sites. Although their heterogenous distribution suggest a neuronal local- ization, no data have been reported so far regarding this point. By means of lesion experiments in the rat we have recently shown that SRIF receptors are mainly located neuronally in the hippocampus (in preparation). A similar neuronal localization of SRIF receptors in the human is suggested by our present observation, in that a discrete vascular lesion in the hippocampus of patient 0, leading to a neuronal death, induced a complete loss of SRIF receptors. These results strongly suggest that at least in some brain regions like the hippocampus, a large part of the SRIF receptors are located on neurons.

Functional implicutions

An attempt to correlate the presently reported distribution of SRIF receptors with functional char- acteristics of this peptide are extremely problematic. In the human, no information is available on the function of SRIF in the CNS. In the rat, inter- pretation of the available studies has to take into account the necessity of intracerebroventricular injec- tion of SRIF and the use of pharmacological doses to compensate for rapid degradation. Nevertheless, the wide limbic localization of SRIF receptors coutd well explain the general limbic activation observed in rats after intracerebroventricular SRIF appIication” whereas some of the striatal SRIF receptors could be mediating the intracerebroventricular effect of SRIF seen on motor coordination” It is of interest to note that from the many sensory systems such as olfac- tory, visual and auditory, as well as the somatic and visceral afferent systems shown to be rich in SRIF receptors in the rat,39 it is mainly the spinal trigeminal nucleus and the substantia gelatinosa of the spinal cord in the human which are particularly rich in SRIF receptors. In these regions a remarkably precise distribution of SRIF receptors is seen. It is not yet clear whether these receptors are postsynaptic to primary afferent” or to descending SRIF neurons.2’ In both cases it can be speculated that SRIF receptors play some role in modulating sensory inputs. Also the locus coeruleus is a region with a particularly high density of SRIF receptors. They could explain the effect of intracerebroventricularly injected SRIF in increasing arousal and decreasing sleep.16

Of more interest for the understanding of the functional relevance of SRIF and its receptors in the human brain will be the demonstration of alterations of the SRIF system in speciiic CNS disease. It is highly interesting that in Alzheimer’s disease, a con- dition related to massive cognition and memory deficits, there is a massive diminution of SRIF in the cortex and hippocampus. “*” In Huntington’s Chorea. a disease associated with motor coordination diaabil- ity, there is an increase in striatal SRIF.‘.‘.” Other

Human brain somatostatin receptors 345

diseases such as multiple sclerosis during relapse, depression, dementia associated to Parkinson, or even schizophrenia, seem to be accompanied by deficits of the brain SRIF system either measured in the CSF or in postmortem brain samples.‘2~‘3~20~31~49~5’ It is unknown, at the moment, whether the SRIF receptors are also affected in these diseases.

CONCLUSION

The present report, which describes an extremely dense network of SRIF receptors in several areas of the human brain, in particular in such regions known to be pathologically altered in certain neuro-

psychiatric diseases, is a first step towards in- vestigations of the status of SRIF receptors in those diseases. Our preliminary result from cases of Alzheimer’s disease seems to indicate that indeed in such patients some of the pathologically affected CNS regions such as hippocampus or neocortex show a reduced amount of SRIF receptors (in prepara- tion). However, there is by no means a complete loss of receptors, leaving the possibility of a potential substitution therapy open.

Acknowledgements-We would like to thank Dr J. Rivier, San Diego, for the gift of m-Trps]-SRIF. We acknowledge Mr B. Thomi, MS R. Mousson and D. Groeppelin for technical assistance and Mr R. W. Foote for reading the manuscript.

REFERENCES

1. Aronin N., Copper P. E., Lorenz L. J., Bird E. D., Sagar S. M., Leeman S. E. and Martin J. B. (1983) Somatostatin is increased in the basal ganglia in Huntington disease. Ann. Neural. 13, 519-526.

2. Aubert M. L., Grumbach M. M. and Kaplan S. L. (1977) The ontogenesis of human fetal hormones. IV. Somatostatin, luteinixing hormone releasing factor and thyrotropin releasing factor in hypothalamus and cerebral cortex of human fetuses 10-22 weeks of age. J. clin. Endocr. Metab. 44, 1130-1141.

3. Beal M. F., Bird E. D., Langlais P. J. and Martin J. B. (1984) Somatostatin is increased in the nucleus accumbens in Huntington’s disease. Neurology, N.Y. 34, 663-666.

4. Braaeau P., Vale W., Burgus R., Ling N., Butcher M., Rivier J. and Guillemin R. (1973) Hypothalamic polypeptide that inhibits the secretion of immunoreactive pituitary growth hormone. Science, Wash. 179, 77-79.

5. Copper P. E., Femstrom M. H., Rorstad 0. P., Leeman S. E. and Martin J. B. (1981) The regional distribution of somatostatin, substance P and neurotensin in human brain. Brain Res. 218, 219-232.

6. Cot&s R., Probst A. and Palacios J. M. (1984) Quantitative light microscopic autoradiographic localization of cholinergic muscarinic receptors in the human brain: brainstem. Neuroscience 12, 1003-1026.

7. Czemik A. J. and Petrack B. (1983) Somatostatin receptor binding in rat cerebral cortex. J. biof. Chem. 258,5225-5539. 8. Davies P., Katzman R. and Terry R. D. (1980) Reduced somatostatin-like immunoreactivity in cerebral cortex from

cases of Alxheimer disease and Alxheimer senile dementia. Nature 288, 279-280. 9. Davies P., Josef J. and Thompson A. (1981) Anterior to posterior variations in the concentration of somatostatin-like

immunoreactivitv in human basal aanalia. Brain Res. Bull. 7. 365-368. 10.

11.

12.

13.

14.

15.

16.

17.

:::

20.

21.

22. 23.

24.

25.

26.

Desy L. and Pelietier G. (1977) Immu~ohistochemical localization of somatostatin in the human hypothalamus. Cell Tiss. Res. 184, 491-497. Epelbaum J., Arancibia L. T., Kordon C. and Enjalbert A. (1982) Characterization, regional distribution, and subcellular distribution of [12JI]Tyr’]-somatostatin-binding sites in rat brain. J. Neurochem. 38, 1515-1523. Epelbaum J., Ruberg M., Moyse E., Javoy-Agid F., Dubois B. and Agid Y. (1983) Somatostatin and dementia in Parkinson’s disease. Brain Res. 278, 376-379. Ferrier I. N., Roberts G. W., Crow T. J., Johnstone E. C., Owens D. G., Lee Y. C., G’Shaughnessy D., Adrian T. E., Polak J. M. and Bloom S. R. (1983) Reduced cholecystokinin-like and somatostatin-like immunoreactivity in limbic lobe is associated with negative symptoms in schizophrenia. Life Sci. 33, 475482. Geola F. L., Yamada T., Warwick R. J., Tourtelotte W. W. and Her&man J. M. (1981) Regional distribution of somatostatin-like immunoreactivity in the human brain. Brain Res. 229, 3H2. Goodman R. R. and Snyder S. H. (1982) K opiate receptors localized by autoradiography to deep layers of cerebral cortex: relation to sedative effects. Proc. natn Acad. Sci. U.S.A. 79, 5703-5707. Havlicek V., Rexek M. and Friesen H. (1976) Somatostatin and thyrotropin releasing hormone: central effect on sleep and motor system. Pharmac. Biochem. Behav. 4, 455-459. Hendry S. H. C., Jones E. G. and Emson P. C. (1984) Morphology, distribution and synaptic relations of somatostatin and neuropeptide Y-immunoreactive neurons in rat and monkey neocortex. J. Neurosci. 4, 2497-2517. Iversen L. L. (1983) Nonopioid neuropeptides in mammalian CNS. A. Rev. Pharmac. Toxicol. 23, l-27. Johansson O., Hiikfelt T. and Elde R. P. (1984) Immunohistochemical distribution of somatostatin-like immuno- reactivity in the central nervous system of the adult rat. Neuroscience 13, 265-339. Kohler J., Schroter E. and Cramer H. (1982) Somatostatin-like immunoreactivity in the cerebrospinal fluid of neurological patients. Arch Psychiat. Neru Krankh. 231, 503-508. Krisch B. (1981) Somatostatin-immunoreactivity fiber projections into the brain stem and the spinal cord of the rat. Cell Tbs. Res. 217, 531-552. Kuhar M. J. (1985) The mismatch problem in receptor mapping studies. TINS 8, 19&191. Kuhar M. J. and Unnerstall J. R. (1985) Quantitative receptor mapping by autoradiography: some current technical problems. TINS 8, 49-53. Langevin H. and Emson P. C. (1982) Distribution of substance P, somatostatin and neurotensin in the human hypothalamus. Brain Res. 246, 65-69. Leroux P. and Pelletier G. (1984) Radioautographic localization of somatostatin-14 and somatostatin-28 binding sites in the rat brain. Peptides 5, 503-506. Mantyh P. W., Hunt S. P. and Maggio J. E. (1984) Substance P receptors: localization by light microscopic autoradiography in rat brain using rH]SP as the radioligand. Brain Res. 307, 147-165.

346 J. C. REUSI er al.

27. Maurer R. and Rcubi 3. C. (1985) Brain somatostatin receptor subpopulation visualized by autoradiography. Erarn Res. 333, 178-181.

28. Morrison J. H., Benoit R., Magistretti P. J. and Bloom F. E. (1983) lmmunohistcchcmical distribution of pro-somatostatin-related peptides in cerebral cortex. Bruin Res. 262, 344-351.

29. Morrison J. H., Rogers J., Schcrr S., Benoit R. ancl Bloom F. E. (1985) Somatostatin immunorcactivity in neuritic plaques of Alzheimcr’s patients. Nature 314, 90-92.

30. Morrison J. H., Foote S. L., Bloom F. E. and Rogers J. (1985) Organization of monoamine and peptide systems in primate brain and implications for Alzhcimcr’s ‘disease. Neurosci. Lett., Suppl., S I >S 14.

31. Nemeroff C. B., Youngblood W. W., Manberg P. J., Prange A. J. Jr and Kizer J. S. (1983) Regional brain concentrations of ncuropeptidcs in Huntington’s chorea and schizophrenia. Science 221, 972--975.

32. Nieuwenhuis R.. Voogd J. and Van Hiizcn C. (1978) The Human Central Nemous System. A Swopsis and Ailac. Springer Verlag, Berlin.

33. Olszewski J. and Baxter D. (1982) Cytoarchitecture of rhe Human Bra&stem, 2nd edn. S. Karger, Basal. 34. Pate1 Y.. Rao K. and Rcichlin S. (1977) Somatostatin in human cerebrosoinal fluid. New Enal. J. Med. 2%, 529-533. 35. Post R. h., Uhdc T. W., Rubinow-D. i., Ballenger J. C. and Gold P. W. (1983) Biochemical &cts of carbamazcpinc:

relationship to its mechanism of action in affective illness. Prog. Neuropsychopharmac. Biol. Psychiat. 7, 263-271. 36. Reichlin S. (1983) Somatostatin. In Brain Peptides (eds Krieger D. T., Brownstein M. J. and Martin J. B.), pp. 71 l-752.

John Wiley, New York. 37. Rcubi J. C. (1984) Evidence of two somatostatin-14 receptors types in rat brain cortex. Neurosci. Left. 49, 259-263. 38. Reubi J. C. (1985) Novel selective radioligand for one population of rat cortex somatostatin receptors. Li/e Sci. 36,

1829-1836. 39. Rcubi J. C. and Maurcr R. (1985) Autoradiographic mapping of somatostatin receptors in the rat central nervous

system and pituitary. Neuroscience 15, 1183-l 193. 40. Reubi J. C. and Landolt A. (1984) High density of somatostatin receptors in pituitaries of acromegalic patients. J. c/in.

Endocr. Metab. 59. 1148-1151. Reubi J. C., Per& M., Rivier J. and Vale W. (1982) High affinity binding sites for somatostatin to rat pituitary. Biochem. Biophys. Res. Commun. 105, 1538-1545.

41.

42.

43.

44.

45.

46. 47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

Reubi J. C., Rivier J., Perrin M., Brow-n M. and Vale W. (1982) Sm high allinity binding sites for somatostatin-28 on pancreatic B-cells: differences with brain somatostatin nceptors. Et&c&o&~ 110, IW-1051. Reubi J. C., Perrin M. H., Rivier J. E. and Vale W. (1981) High affinity binding sites of a somatostatin-28 analog in rat brain. Life Sci. 28, 2191-2198. Roberts G. W.. Alien Y.. Crow T. J. and Polak J. M. (1983) Immunocvtochemical localization of neurowptides in the fornix of rat, monke; and man. Brain Res. 263, 15i-ISi. -

_ _

Roberts G. W., Crow T. J. and Polak J. M. (1985) Location of neuronal tangles in somatostatin neurones in Alzheimer’s disease. Nature 314, 92-94. Rossor M. N. (1982) Neurotransmitters and CNS disease. The Luncet, November 27. Rossor M. N., Emson P. C., Motmtjoy C. Q., Roth M. and Iversen L. L. (1980) Reduced amounts of immunoreactivc somatostatin in the temporal cortex in sen4e dementia of Aizh&mer type. Neurosci. Lets. 20, 373-377. Rothman R. B.. Herkenham M.. Pert C. B.. Liana T. and Cascieri M. A. (1984) Visualization of rat brain receptors for the neuropebtide substance i. Brain Rei. 3@9;47-54. Rubinow D. R., Gold P. W., Post R. M. and Ballenger J. C. (1983) CSF somatostatin in affective illness. Archs gen. Psychiat. 40, 409-412. Sorensen K. V. (1982) Somatostatin: localization and distribution in the cortex and the subcortical white matter of human brain. Neuroscience 7, 1227-1232. Sorensen K. V., Christensen S. E., Dupont E., Hansen A. P., Pedcrsen E. and Orskov H. (1980) Low somatostatin content in cercbrospinal fluid in multiple sclerosis. Acta Neural. 61, 186191. Srikant C. B. and Pate1 Y. C. (1981) Somatostatin receptors: identification and characterization in rat brain membranes. Proc. natn. Acad. Sci. U.S.A. 70, 3930-3934. Tran V. T., Uhl G. R., Perry D. C., Manning D. C., Vale W. P., Perrin M. H., Rivicr J. E., Martin J. B. and Snyder S. H. (1984) Autoradiographic local&&ion of somatostatin receptors in rat brain. Eur. J. Phurmac. 101, u)7---309. Tran V. T.. Beal M. F. and Martin J. B. (1985) Two types of somatostatin receptors differentiated by cyclic somatostatin analogs. Science 228, 492-495. Vincent S. R.. Johansson 0.. Hokfelt Meverson B., Sachs C., Elde R. P.. Terenius L. and Kimmel J. (1982) Neuropeptide coexistence in human co&l n&ones. Pkre Z%, 65-67. Young W. S. and Kuhar M. J. (1981) Ncurotensin receptor localization by light microscopic autoradiography in rat brain. Brain Res. 206, 273-285. Wise S. P. and Herkenham M. (1982) Opiate receptor distribution in the cerebral cortex of the rhesus monkey. Science 218. 387-389.

(Accepted 9 December 1985)