Immunolocalization of tripeptidyl peptidase II, a cholecystokinin-inactivating enzyme, in rat brain

Pergamon

PII:

Neuroscience Vol. 88, No. 4, pp. 1225–1240, 1998 Copyright  1998 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/99 $19.00+0.00 S0306-4522(98)00257-7

IMMUNOLOCALIZATION OF TRIPEPTIDYL PEPTIDASE II, A CHOLECYSTOKININ-INACTIVATING ENZYME, IN RAT BRAIN P. FACCHINETTI,* C. ROSE,† PH. ROSTAING,‡ A. TRILLER‡ and J.-C. SCHWARTZ†§ *Laboratoire de Physiologie, Faculte´ de Pharmacie, 4 Avenue de l’Observatoire, 75006 Paris, France †Unite´ de Neurobiologie et Pharmacologie (U.109) de l’INSERM, Centre Paul Broca, 2ter rue d’Ale´sia, 75014 Paris, France ‡Laboratoire de Biologie Cellulaire de la Synapse, INSERM CJF 9410, Ecole Normale Supe´rieure, 46 rue d’Ulm, 75005 Paris, France Abstract––Tripeptidyl peptidase II (EC 3.4.14.10) is a serine peptidase apparently involved in the inactivation of cholecystokinin octapeptide [Rose C. et al. (1996) Nature 380, 403–409]. We have compared its distribution with that of cholecystokinin in rat brain, using a polyclonal antibody raised against a highly purified preparation for immunohistochemistry at the photon and electron microscope levels. Tripeptidyl peptidase II-like immunoreactivity was mostly detected in neurons, and also in ependymal cells and choroid plexuses, localizations consistent with a possible participation of the peptidase in the inactivation of cholecystokinin circulating in the cerebrospinal fluid. Immunoreactivity was mostly detected in cell bodies, large processes and, to a lesser extent, axons of various neuronal populations. Their localization, relative to that of cholecystokinin terminals, appears to define three distinct situations. The first corresponds to neurons with high immunoreactivity in areas containing cholecystokinin terminals, as in the cerebral cortex or hippocampal formation, where pyramidal cell bodies and processes surrounded by cholecystokinin axons were immunoreactive. A similar situation was encountered in many other areas, namely along the pathways through which cholecystokinin controls satiety, i.e. in sensory vagal neurons, the nucleus tractus solitarius and hypothalamic nuclei. The second situation corresponds to cholecystokinin neuronal populations containing tripeptidyl peptidase II-like immunoreactivity, as in neurons of the supraoptic or paraventricular nuclei, axons in the median eminence or nigral neurons. In both situations, localization of tripeptidyl peptidase II-like immunoreactivity is consistent with a role in cholecystokinin inactivation. The third situation corresponds to areas with mismatches, such as the cerebellum, a region devoid of cholecystokinin, but in which Purkinje cells displayed high tripeptidyl peptidase II-like immunoreactivity, possibly related to a role in the inactivation of neuropeptides other than cholecystokinin. Also, some areas with cholecystokinin terminals, e.g., the molecular layer of the cerebral cortex, were devoid of tripeptidyl peptidase II-like immunoreactivity, suggesting that processes other than cleavage by tripeptidyl peptidase II may be involved in cholecystokinin inactivation. Tripeptidyl peptidase II-like immunoreactivity was also detected at the ultrastructural level in the cerebral cortex and hypothalamus using either immunoperoxidase or silver-enhanced immunogold detection. It was mainly associated with the cytoplasm of neuronal somata and dendrites, often in the vicinity of reticulum cisternae, Golgi apparatus or vesicles, and with the inner side of the dendritic plasma membrane. Hence, whereas a fraction of tripeptidyl peptidase II-like immunoreactivity localization at the cellular level is consistent with its alleged function in cholecystokinin octapeptide inactivation, its association with the outside plasma membrane of neurons remains to be confirmed.  1998 IBRO. Published by Elsevier Science Ltd. Key words: EC 3.4.14.10, cholecystokinin octapeptide, electron microscope immunolocalization, light microscope immunolocalization, satiety, serine proteases.

In contrast with the mechanisms responsible for turning off the signals generated by amino acids and monoamines, of which the inactivating enzymes or transporter systems are well identified, those responsible for turning off neuropeptide signals are, in §To whom correspondence should be addressed. Abbreviations: BSA, bovine serum albumin; CCK, cholecystokinin; CCK-8, cholecystokinin octapeptide; LI, -like immunoreactivity; PBS, phosphate-buffered saline; TBS, Tris-buffered saline; TPPII, tripeptidyl peptidase II.

most cases, ill defined. One possibility is that the generally long-lived peptidergic signals are turned off by diffusion and/or hydrolysis of the neurotransmitter by a series of non-specific peptidases.28 In some cases, however, well-defined peptidases were found to play a critical role in the inactivation of the neuropeptide due to a combination of adequate localization and catalytic properties. The first and still best example is that of enkephalins,47 whose physiological inactivation is attributable to the action of enkephalinase (neprilysin, EC 3.4.24.11)23 and

1225

1226

P. Facchinetti et al.

aminopeptidase M (EC 3.4.11.2),12 two membranebound metallopeptidases which have been immunolocalized in some detail in the rat brain.24,25,35,36,51 Until recently, little was known about the corresponding inactivation mechanisms of cholecystokinin octapeptide (CCK-8), one of the most abundant neuropeptides in the brain.20,37,38,55 Several studies had shown that an ill-characterized acidic or neutral aminopeptidase,7,8 neprilysin,9,26,30,58 a thiolendopeptidase27 or a metalloendopeptidase52 may participate in the hydrolysis of exogenous CCK-8 by cerebral membranes, with cleavages occurring at various peptide bonds within the molecule. However, since peptidases generally display rather broad substrate specificity and numerous peptidases are present in cerebral membranes, metabolic pathways of exogenous neuropeptides may not reliably reflect those responsible for endogenous neuropeptide inactivation. We have therefore studied the metabolic fate of endogenous CCK-8 released by depolarization of brain slices, a model which largely retains the integrity of the native tissue.13,22,32 This model allows the released peptide, at near-physiological concentrations, to come into contact initially with physiologically relevant peptidases, presumably located close to the nerve endings from which CCK-8 originates. The use of this model ultimately led to the identification of a membrane-bound form of tripeptidyl peptidase II (TPPII, EC 3.4.14.10) as the responsible enzyme.3,41–43 TPPII, a serine peptidase whose function was previously unknown,53 first releases the N-terminal tripeptide of CCK-8 and the resulting pentapeptide CCK-5 is further cleaved into inactive fragments. Purified TPPII displays a rather high degree of specificity toward the sulfated octapeptide and CCK-5, but is still able to cleave a few other neuropeptides, although at a significantly lower rate. The role of TPPII as a major CCK-inactivating enzyme was finally established using butabindide, a rationally designed potent and selective TPPII inhibitor, which was shown to protect almost completely endogenous CCK-8 from hydrolysis in the brain slice model which, in its absence, was inactivated by about 85%. In addition, in rodents, butabindide elicited satiety, a typical CCK-like effect, prevented by administration of a CCKA receptor antagonist, thereby demonstrating, in a physiological model, the implication of TPPII in endogenous CCK-8 inactivation in vivo. Preliminary localization studies of TPPII by in situ hybridization of its mRNA or by immunohistochemistry using polyclonal antibodies raised against the purified enzyme showed its presence in cerebral neurons and sensory neurons of the nodose ganglion of the vagus nerve which were not inconsistent with its alleged functions.43 Here, using the same antibodies, we have undertaken to establish in a more detailed manner the distribution of cells displaying TPPII-like immunoreactivity (LI) in rat brain, particularly in relation-

ship with that of CCK-LI, as investigated previously in numerous studies, e.g., those reviewed by Vanderhaeghen.54 In addition, we have studied TPPII-LI at the ultrastructural level to further assess the association with neuronal membranes of a fraction of this predominantly cytosolic enzyme, as suggested by previous biochemical and ultrastructural studies.43 EXPERIMENTAL PROCEDURES

Antibodies The TPPII antibody was raised starting from a highly purified rat liver enzyme preparation.43 Two Albino New Zealand rabbits (Charles River) were injected subcutaneously into multiple sites on the back with 0.1 mg of TPPII, emulsified with Freund’s complete adjuvant for the first injection and Freund’s incomplete adjuvant for subsequent injections. The first three injections were given at weekly intervals and booster injections every month, followed by bleeding one week later. The immunoglobulin G fraction was purified from the serum using the method of Danielsen et al.5 and stored at 4C in phosphate-buffered saline (PBS; 0.1 M phosphate buffer, pH 7.4, 0.9% NaCl) containing 0.02% sodium azide. For non-specific staining controls, the antibody (0.3 µg protein/ml) was preabsorbed with purified TPPII (1.8 µg protein/ml). The specificity of the TPPII antibody was assessed using western blot analysis. The polyclonal CCK-8 antibody (RAS 7181N) was purchased from Peninsula Laboratories (Belmont, CA, U.S.A.). Tissue processing Male Wistar rats (150–200 g; Iffa Credo, St Germain l’Arbresle, France) were anesthetized with pentobarbital (50 mg/kg, i.p.) and perfused through the ascending aorta with 10 ml of saline (0.9% NaCl), followed either by an ice-cooled 4% paraformaldehyde solution for light microscopy processing, or by 4% paraformaldehyde/0.1% glutaraldehyde solution for electron microscopy processing in 0.1 M phosphate buffer (pH 7.4; 1 ml/g body weight). The brain was dissected out and postfixed in the same fixative solution for 2 h at 4C and rinsed in PBS. Serial coronal sections (50–60 µm thickness) were made with a Leica VT1000 Vibratome and collected at 0.3-mm intervals, either for indirect immunoperoxidase histochemistry or for Nissl staining, the latter performed to facilitate the identification of structures. Immunohistochemistry at the light microscope level Free-floating sections were washed and permeabilized with 0.05% Triton X-100 in PBS and preincubated for 30 min at room temperature in PBS containing 2% bovine serum albumin (BSA) and 1% normal goat serum. Sections were then incubated overnight at 4C with the primary TPPII antibody (1:5000) or CCK-8 antibody (1:1000) dissolved in PBS containing 2% BSA. After three washes in PBS (10 min each), the sections were incubated in 0.3% hydrogen peroxide for 30 min and, following several rinses in PBS, incubated at room temperature for 2 h with a biotinylated goat anti-rabbit antibody (Biosys; 1:200) and, for 45 min, with an avidin–biotin– peroxidase complex (Elite, Biosys). These latter two steps were followed by three rinses in PBS. Finally, the peroxidase activity was detected by incubating the sections in 50 mM Tris buffer (pH 7.6) containing 0.05% 3,3 -diaminobenzidine tetrahydrochloride (Sigma) and 0.01% hydrogen peroxide. The reaction was monitored under a light microscope and stopped by transfer into fresh Tris buffer.

Cholecystokinin-inactivating peptidase The specificity of TPPII-LI staining was demonstrated in two ways: firstly, by replacing the primary antibody with a preimmune serum and, secondly, by carrying out primary incubations in the presence of a saturating concentration of antigen, i.e. 1.8 µg/ml of the purified TPPII enzyme preparation. After immunohistochemical staining, sections were mounted on to glass slides coated with 0.1% gelatin and prepared for light microscopic examination, being dehydrated, defatted in xylene, coverslipped with acrytol and viewed using a Zeiss (axiophot) microscope. Immunohistochemistry at the electron microscope level Free-floating sections were washed in 0.1 M Tris-buffered saline (TBS; pH 7.6), incubated for 30 min in TBS containing 1% BSA and 5% normal goat serum, then incubated overnight at 4C with the primary TPPII antibody (1:5000) in 1% BSA/5% normal goat serum in TBS. After three washes in TBS, some sections were incubated with a goat anti-rabbit immunoglobulin G horseradish peroxidase antibody (Amersham), and peroxidase activity was detected as described in the preceding section. Other sections were incubated successively in 0.3% gelatin and 1% BSA for 15 min, then in (auroprobe one, Amersham) goat antirabbit immunoglobulin G bound to 1-nm colloidal gold diluted 1:200 in 0.3% gelatin and 1% BSA for 4 h, in 2% glutaraldehyde for 20 min and intenSEM (Amersham) for silver intensification for 20 min. All incubations and washes were carried out in 0.1 M phosphate buffer (pH 7.6), except for rinses in 0.2 M citrate buffer performed before and after silver intensification. Sections were then mounted on slides for light microscopic examination. For electron microscopy, labeled sections were postfixed for 15 min in 2% osmium tetroxide in 0.1 M phosphate buffer, dehydrated and flat-embedded in Epon 812. Four embedded sections showing optimal immunolabeling were selected under the light microscope and glued on the top of prepolymerized Epon blocks for ultrathin sectioning. In each section, the ventromedial hypothalamic nucleus and the cerebral cortex were trimmed on one side and serial thin sections were cut from the surface of the block using a Reichert ultramicrotome. Sections were collected on grids, counterstained with uranyl acetate and lead citrate, and examined with a Philips electron microscope. To evaluate the possible binding of the colloidal gold anti-immunoglobulin G, some sections were processed either by omitting the primary antibody or by using the preimmune antibody at the same dilution. RESULTS

The specificity of the antibody used in most experiments, a polyclonal antibody raised against a highly purified TPPII preparation, has been assessed previously using western blot analysis.43 In addition, when the primary TPPII antibody was either omitted or applied in the presence of an excess of purified TPPII, all immunohistochemical signals were suppressed (not shown). Immunolocalization at the photon microscope level Observation of serial sections of rat brain at the light microscope level showed TPPII-LI largely but heterogeneously distributed in many areas (Figs 1– 4). In most cases, the immunoreactivity was detected in the cytoplasm of neuronal cell bodies, axons and dendrites, but some glial elements, such as

1227

ependymal cells and epithelial cells of the choroid plexuses, were also immunoreactive. The relative intensity of TPPII-LI in major immunoreactive areas was independently rated by two persons and is reported in Table 1. For purposes of comparison, this table and Figs 1–3 also show corresponding distributions of CCK-LI, derived mostly from the literature54 or, in some cases, from personal observations (e.g., Figs 1G, 3F), and of CCK receptor (type A or, generally, B) binding from the literature.29 Olfactory regions. In these areas, TPPII-LI of intermediate abundance was found. In the main olfactory bulb (not shown), TPPII-LI neuronal cell bodies and processes (presumably dendrites) were found in all layers except the glomerular and internal plexiform layers. The highest density of immunoreactivity was observed in the mitral cell layer, at the level of triangularly shaped (presumably mitral) cells. Less numerous cells were found in the external plexiform layer (intensely labeled tufted cells) and the internal granular layer, where small immunoreactive soma were scattered. CCK-LI and mRNAs seem to be absent from mitral cells and mainly present in tufted cells.45,54,57 In the anterior olfactory nucleus, TPPII-LIpositive neuronal soma and processes were most numerous in pars dorsal and ventral. In the piriform cortex, a strong TPPII-LI was observed, mainly in pyramidal neurons of layer III (Fig. 1F). In the olfactory tubercle, which displays a high density of CCK-LI-containing axons, mainly in the medial part,16 and of CCKB receptors,29 a high density of TPPII-LI was detected, namely at the level of the pyramidal cell layer (Fig. 1). In the tenia tecta, numerous cells expressed TPPII-LI that were surrounded by scattered CCK-LI-containing neurons (Fig. 1G, H). With the sensitive technique of in situ hybridization, many tenia tecta neurons were found to express CCK receptor mRNA.19 Cerebral cortex. Here, TPPII-LI-containing neurons were detected in all neocortical areas, their abundance being highest in frontal and occipital areas and lowest in cingulate and entorhinal areas (partially shown in Fig. 1A). A laminar distribution of TPPII-LI was observed in all neocortical areas, as shown in Fig. 3D for the parietal or cingulate cortex. Whereas no immunoreactivity was detected in the molecular layer and positive cells were scarce in layer IV, numerous positive pyramidal cell bodies and their thick emerging dendrites were present in layers III and V (Fig. 3E). In layer II, only numerous thick immunoreactive processes were found, which are likely to correspond to long dendrites of the pyramidal cells emanating from underlying layers. In layer VI, some cells of irregular shape (pyramidal or ovoid) and size were immunoreactive. In comparison, CCK-LI is not found in pyramidal cells, but in varicose axons which

1228

P. Facchinetti et al.

Table 1. Distribution of tripeptidyl peptidase II in areas of rat brain and comparison with cholecystokinin-like immunoreactivity and CCKB receptor binding Brain structures

TPPII-LI

CCK-LI

CCK receptor

2–4+ 2–4+ 4+ 4+ 4+

1–2+ 1+ 4+ 3+ 2+

2–4+ 1+ 4+ 2+ 1+

Cerebral cortex (frontal and occipital) Layer I Layer II Layer III Layer IV Layer V Layer VI

0 4+ 3+ 1+ 4+ 2+

3+ 4+ 4+ 2+ 2+ 2+

3+ 3+ 4+ 3+ 3+ 3+

Cerebral cortex (parietal and temporal) Layer I Layer II Layer III Layer IV Layer V Layer VI

0 2+ 2+ 1+ 2+ 2+

3+ 4+ 4+ 2+ 2+ 2+

2+ 2+ 2+ 3+ 2+ 3+

2+

2+

1–3+

0 0 2+ 1+

0–1+ 1+ 3+ 3+

1+ 1+ 1+ 1+

0 0 3+

1+ 0 3+

1+ 1+ 3+

2+

4+

1–3+

Septum

3–4+

3–4+

1+

Basal ganglia Nucleus accumbens Caudate–putamen Globus pallidus Islands of Calleja

2+ 1–2+ 3–4+ 3+

4+ 2+ 1+ 4+

4+ 2–3+ 0–1+ 2+

Thalamus Anterior, median, lateral and intralaminar nuclei Midline nuclei, epithalamus, subthalamus and metathalamus

2–3+ 3–4+

0 3+

0 1+

Hypothalamus Anterior, lateral nuclei Arcuate nucleus Median eminence, ventromedial and dorsomedial nuclei Mammillary nuclei Supra- and premammillary nuclei and posterior hypothalamus

3–4+ 4+ 4+ 4+ 4+

2–4+ 1+ 4+ 2+ 3–4+

2+ 1+ 2–3+ 0 2+

Mesencephalon Colliculi Substantia nigra Ventral tegmental area Interpeduncular nuclei

1–3+ 4+ 3+ 3+

3+ 3+ 3+ 3+

1+ 1+ 0 1+

Metencephalon Tegmental nuclei Superior olivary complex and pontine reticular nuclei Precerebellar nuclei

3–4+ 3–4+ 3+

2+ 2+ 3+

1+ 0–1+ 1+

4+ 4+ 4+

2+ 3+ 3+

1+ 2–3+ 0

Olfactory regions Olfactory bulb Anterior olfactory nuclei Piriform cortex Olfactory tubercle Tenia tecta

Hippocampal formation Subicular complex Ammon’s horn Lacunosum molecular layer Oriens Pyramidal cell layer Radiatum layer Dentate gyrus Granular layer Molecular layer Polymorph layer Amygdala

Medulla Parabrachial nucleus Nucleus tractus solitarius Gracile nucleus

Cholecystokinin-inactivating peptidase

1229

Table 1. Continued. Brain structures

TPPII-LI

CCK-LI

CCK receptor

Medulla Cuneate nucleus Area postrema Cochlear and vestibular nuclei Raphe nuclei Medullary reticular nuclei Cranial nerve nuclei (10, 12, 5)

4+ 4+ 4+ 2+ 4+ 4+

1+ 4+ 1+ 2+ 4+ 3+

1+ 4+ 1+ 1+ 0 1+

Cerebellum Purkinje cells Granule cells

4+ 4+

0 0

0 0

The intensity of TPPII-LI was rated by observation of serial frontal sections similar to those shown in Figs 1–4, whereas ratings of CCK-LI and CCKA or CCKB binding intensities were derived from Refs 11, 29 and 54.

appear to surround the latter (Fig. 3F). Pyramidal cells, however, were reported to express CCK mRNA.2,15,45,57 CCK-LI cell bodies are known to be present in layers II, III, V and VI, and are local circuit neurons, predominantly medium-sized bipolars and multipolars with processes coursing perpendicular to the pial surface.11,34,48,54 In the cingulate cortex, high TPPII-LI was observed in layer III, which expresses high CCKB receptor binding.29 Hippocampal complex. In this region, TPPII-LI was present in all fields, although in a heterogeneous manner. In the subiculum, TPPII-immunoreactive neurons were present in moderate abundance in all layers (not shown) where CCK-LI has been detected.54 In the dentate gyrus, the molecular layer and granule cells were not labeled. Numerous interneurons of the hilar region which innervate granule cells were highly immunoreactive, a pattern which resembles that of CCK-LI and CCKB receptor binding.16,29,54 In Ammon’s horn, some pyramidal neurons of the CA1, CA2 or CA3 fields and their

dendrites were highly labeled, whereas multipolar neurons in the stratum radiatum were occasionally labeled and no immunoreactivity was detected in the oriens layer (Fig. 3A, C). In comparison, CCK-LI was detected in interneurons of strata oriens and radiatum (Fig. 3B), and CCK mRNA in most pyramidal cells.2,45,57 Amygdaloid complex. Here, a moderate and rather homogeneous level of TPPII-LI was found in neurons from a number of nuclei, such as the medial, cortical, lateral, basomedial or basolateral nuclei, and the nucleus of the lateral olfactory tract, as well as in the amygdalohippocampal area. In most cases, moderate immunoreactivity was found in the soma and processes of small-sized multipolar neurons. CCK-LI (and CCK mRNA) is known to be prevalent in intra-amygdaloid projecting neurons of the medial, central, basolateral and cortical nuclei, but the amygdala also receives projections from neurons in the substantia nigra, ventral tegmental area, raphe nuclei and rostral linear nucleus.40,45,49

Abbreviations used in the figures 10 12 Acb sh AI AP Arc CA1–CA3 Ce Cg Co CPu Cu d DG Fr Gr Gra Hb ICj IO LD LR Me

dorsal motor nucleus vagus hypoglossal nucleus accumbens nucleus, shell agranular insular cortex area postrema arcuate hypothalamic nucleus fields CA1–CA3 of Ammon’s horn central amygdaloid nucleus cingulate cortex cortical amygdaloid nucleus caudate–putamen cuneate nucleus dendrites dentate gyrus frontal cortex gracile nucleus granular layer habenular nucleus islands of Calleja inferior olive nuclei laterodorsal thalamic nucleus lateral reticular nuclei medial amygdaloid nucleus

ME Mol Or Par Pir Pu PV Py Rad Re Rh S SO Sol Sp5 TT TU VL VM VMH VP VPL VPM

median eminence molecular layer stratum oriens parietal cortex piriform cortex Purkinje cells paraventricular thalamic nucleus pyramidal cell layer stratum radiatum reuniens thalamic nucleus rhomboid thalamic nucleus neuronal somata supraoptic nucleus nucleus of the solitary tract spinal trigeminal nuclei tenia tecta olfactory tubercle ventrolateral thalamic nucleus ventromedial thalamic nucleus ventromedial hypothalamic nucleus ventral pallidum ventral posterolateral thalamic nucleus ventral posteromedial thalamic nucleus

Fig. 1. Localization of TPPII-LI on a frontal section of rat brain prepared at an interaural anterior level (IA 10.7 mm; A) and comparison with CCK-LI (B) and CCKB receptor binding (C). Enlargements of A are shown at the levels of the islands of Calleja (D), the ventromedial region of the nucleus accumbens shell (E), piriform cortex (F) and the tenia tecta (G). In H, CCK-LI cells in the region of the tenia tecta are shown. The schematic drawing of CCK-LI (B) is based on data from Vanderhaeghen54 and Fallon and Seroogy11 expressed in increasing fiber densities (1+ to 4+), whereas nerve cells are represented by dots. The schematic drawing of CCKB receptor binding is based on data from Moran and McHugh,29 variations in density being graded from 1+ to 4+.

Cholecystokinin-inactivating peptidase

1231

Fig. 2. Localization of TPPII-LI immunoreactivity on a frontal section of rat brain (IA 6.2 mm; A) and comparison with CCK-LI (B) and CCKB receptor binding (C), represented as in Fig. 1. Enlargements of A are shown at the levels of the ventromedial hypothalamus (D), medial amygdaloid nuclei (E), ventral posteromedial thalamic nucleus (F) and supraoptic nucleus (G).

Septal region. In this region, strongly immunoreactive neurons, medium-sized with a triangular shape and several fine processes, were found in the medial and lateral parts, as well as in the bed nucleus of the stria terminalis. In all of these areas, a high density of CCK-LI is present in cell bodies and axons (some of which correspond to projections of dopaminergic neurons).

Basal ganglia. In this area (Fig. 1), TPPII-LI was distributed in a highly heterogeneous fashion. The highest level was found in the globus pallidus, particularly its ventral part. In the latter, TPPII-LI was prominent in numerous clusters of large multipolar neurons with thick processes, whereas in the remaining parts of the globus pallidus, it was in homogeneously distributed neurons of medium size

1232

P. Facchinetti et al.

Cholecystokinin-inactivating peptidase

and with fine processes. In the islands of Calleja, TPPII-LI of intermediate intensity was detected in the perikaryon and thick processes of hilar cells, whereas granule cells were never labeled. In the nucleus accumbens, TPPII-LI was found in a few small neurons with unlabeled processes in the core and in more numerous, scattered, medium-sized (20 µm) neurons in the shell. The nucleus accumbens contains a dense network of CCK-LI fibers, corresponding to dopaminergic projections of ventral tegmental neurons.16–18 In the caudate–putamen, low TPPII-LI was present. In this structure, CCK-LI displays a heterogeneous pattern representing fibers of different types and origin, including dopaminergic axons.16 Thalamus. Here (Fig. 2), the TPPII-LI distribution displayed a rostrocaudal gradient, with the anterior, median and lateral nuclear groups moderately labeled, and the intralaminar and midline nuclear groups more heavily labeled, e.g., the epithalamic, subthalamic and metathalamic nuclei. Intense immunoreactivity was found in the periventricular and dorsolateral geniculate nuclei, which also contain a high density of CCK-LI neurons.11,54 All TPPII-LIstained neurons displayed a similar morphological profile, consisting of large triangular cells with thick processes that were more concentrated at the lateral edges of the thalamus. Hypothalamus. In this brain region (Fig. 2), a very high level of TPPII-LI was observed in many nuclei, extending from the rostral to the caudal part of the structure. Immunoreactivity was present in a large variety of neuronal types. For instance, magnocellular neurons of the supraoptic (Fig. 2G) and paraventricular nuclei, known to contain CCK-LI and CCK mRNA, displayed high TPPII-LI, and this was also the case for magnocellular neurons of the tuberomammillary nucleus, which appear to be CCK immunoreactive.21 High TPPII-LI was also detected in medium-sized neurons of the dorsomedial and ventromedial nuclei, in which few CCK-positive somas and numerous CCK-positive fibers are present. In the arcuate nucleus, in which fibers containing CCK-LI seem scarce,21 TPPII-LI was detected in the soma of small and round-shaped neurons. In contrast, TPPII-LI was present in the median eminence, where some processes were labeled, particularly in the internal layer. CCK-LI is known to be present in nerve fibers of both the internal and external layers of the median eminence, and may correspond to

1233

projections of magnocellular and parvocellular neurons from the paraventricular and periventricular nuclei. Mesencephalon. In this area (not shown), a high level of TPPII-LI was found in the inferior colliculus, where the positive cell population in the central nucleus consisted mainly of small- to medium-sized ovoid or round neurons (10–15 µm), intermingled with larger and even more intensely labeled neurons (20 µm). In the external cortex and dorsal cortex of the inferior colliculus, large immunoreactive cells were predominant. TPPII-LI was less intense in the superior colliculus, being localized in only a few neurons in the deep part. In the substantia nigra, pars compacta and pars lateralis, virtually all cell bodies (15–20 µm, fusiform or triangular shape) and their processes were labeled. Numerous neurons were also labeled in the ventral tegmental area, whereas only a few cells, scattered within a dense network of positive processes, displayed CCK-LI in the substantia nigra pars reticulata.16 Metencephalon. Here, a high level of TPPII-LI was detected in various tegmental nuclei, e.g., the locus coeruleus and tegmental, mesencephalic trigeminal or reticulotegmental nuclei. Whereas most immunoreactivity in the locus coeruleus and mesencephalic trigeminal nucleus was found in cell bodies, it was mostly found in processes within the tegmental nucleus. In the pontine reticular nucleus, high TPPII-LI was detected in large pyramidal (30–40 µm) and small cell bodies (10 µm), as well as in numerous processes. In the superior olivary complex and pontine reticular nucleus, numerous ovoid cells and processes were labeled. In the inferior olivary complex, where CCK-LI fibers have been reported,11,54 TPPII-LI was high. In the dorsal raphe nucleus, medium-sized (15–20 µm) round or ovoid neurons and processes were moderately labeled. The same nucleus also contains CCK-LI cell bodies.11,54 In the parabrachial nucleus, TPPII-LI was high, mainly in ovoid, medium-sized (12–18 µm) neurons. Medulla oblongata. In the medulla oblongata (Fig. 4), very high TPPII-LI was detected, particularly in relay nuclei and other areas, e.g., the gracile and cuneate nuclei, trigeminal nucleus and medullary reticular nucleus, where CCK-LI cell bodies and/or networks were described. In addition, in the area postrema, where CCK-LI fibers are abundant, numerous small cells (<10 µm) were immunostained

Fig. 3. Comparison of TPPII-LI and CCK-LI in the hippocampus (A–C) and cerebral cortex (D–F). (A) TPPII-LI pyramidal cell layer in the CA1 field of the hippocampus. (B) CCK-LI interneurons in stratum oriens and stratum radiatum of the CA1 field. (C) Enlargement of A, showing TPPII-LI within the perikaryon and dendrites of a pyramidal neuron. (D) TPPII-LI cell bodies and fibers in various layers of the parietal cortex. (E) Enlargement of D, showing TPPII-LI in a pyramidal cell body in layer III. (F) CCK-LI nerve fibers forming a delicate network around a pyramidal cell (arrow) in layer III of the parietal cortex.

1234

P. Facchinetti et al.

Fig. 4. Localization of TPPII-LI on a frontal section of rat brain (A; IA
4.7 mm) and comparison with CCK-LI (B) and CCKB receptor binding (C), represented as in Fig. 1. Enlargements of A show TPPII-LI in structures surrounding the fourth ventricle, e.g., the area postrema, nucleus of the solitary tract, cuneate nucleus, several cranial nerve nuclei (D), the nucleus of the solitary tract (E) and the cerebellum (F), where granule cells and Purkinje cells are highly immunoreactive (G).

Cholecystokinin-inactivating peptidase

in a way suggesting that the entire neuronal population contained TPPII-LI. In the nucleus of the solitary tract, which contains a dense network of CCK-immunoreactive fibers, numerous small ovoid cells were stained. Several cranial nerve nuclei contained TPPII-LI, e.g., the hypoglossal nucleus, where very large (40–60 µm), most often triangular cells, associated with a dense network of thick processes, were labeled. Cerebellum. In the cerebellum (Fig. 4), a high level of TPPII-LI was found in all Purkinje cell somata, along with dendritic arborization. In the granule cell layer, granule cells themselves were strongly immunoreactive, but not Golgi cells, and the molecular layer was devoid of immunoreactivity. CCK-LI and mRNA are absent from the cerebellum. Spinal cord. In the cervical spinal cord (not shown), high TPPII-LI, confined to the gray matter, was found. It was predominant in the ventral horn, where large pyramidal or fusiform-shaped neurons (25–30 µm) and their thick processes were labeled, whereas only scarce CCK-immunoreactive cell bodies were reported. In the dorsal horn, where CCK-LI is concentrated in nerve fibers of layers I and II of Rexed, lesser TPPII-LI was detected in smaller (12–18 µm) ovoid cells. Around the central canal, a dense network of TPPII-positive processes was present. Immunolocalization at the ultrastructural level The ultrastructural features of TPPII-LI were analysed in the cerebral cortex and hypothalamus. Using immunoperoxidase, we found that it was restricted to neuronal somata and dendrites (Fig. 5A). Some very thin dendrites were also labeled, suggesting that the TPPII-LI extended quite far in dendrites. Axons were always devoid of labeling. Other cell types such as astrocytes, oligodendrocytes or endothelial cells were not labeled. The electron-dense immunoenzymatic product was distributed unevenly in the cytoplasm. It tended to aggregate along cisternae of the reticulum, mitochondria and at the cytoplasmic side of the plasma membrane. As reported previously,43 the enzymatic reaction product was also detected on the postsynaptic side of synaptic complexes. However, this pattern of labeling could result from a secondary translocation and non-specific adsorption of the oxidized diaminobenzidine. In order to define the relationship of TPPII with the plasma membrane precisely, we have also used silver-enhanced immunogold. With this method, we found that, in the soma (data not shown), gold particles were within the cytosol, often found in the vicinity of reticulum cisternae, Golgi apparatus or vesicles, or mitochondria, but not in close association with the membranes of these organelles. Within dendrites, the silver-enhanced gold particles were either

1235

within the cytosol (Fig. 5B) or associated with the inner side of the dendritic plasma membrane. In this case, they could be detected in front of glial processes (Fig. 5C, D). DISCUSSION

The major aim of the present immunohistochemical study was to assess whether TPPII localization in the rat brain was consistent with its alleged role in the inactivation of endogenous CCK-8. For this purpose, we have undertaken to determine whether the distribution of cells in which TPPII-LI was observed was parallel to that of CCK-immunoreactive terminals at both the regional and cellular levels. In addition, we have addressed the question of subcellular distribution of TPPII by immunolocalization studies at the electron microscope level. Our results reveal the extensive distribution of cells containing TPPII in the brain, whose immunoreactivity was detected mainly in the cytoplasm of neuronal cell bodies and their attached processes. Nevertheless, some non-neuronal cells were also immunoreactive in the brain and, to a larger extent, in peripheral tissues. Indeed, strong immunoreactivity, consistent with the detection of TPPII mRNA hybridization signals and TPPII catalytic activity, was detected in tissues such as the liver, kidney, intestine, thymus or testis,43 which, with the exception of the intestine and testis,33 do not contain any detectable CCK-LI or mRNA. In the brain, ependymal cells lining all the ventricles as well as epithelial cells in the choroid plexuses displayed TPPII-LI. Such a localization could be consistent with a role of TPPII in the inactivation of CCK-8 circulating in the cerebrospinal fluid.6,39 Neurons displaying TPPII-LI were extremely pleiomorphic in size (from small cells in various areas to magnocellular neurons mainly found in the hypothalamus) and shape (round, pyramidal, etc.). When these neurons are considered in connection with their alleged function, i.e. their vicinity with CCK-LI terminals, in order to assess whether their topography may allow them to receive or respond to endogenous CCK-8, three main situations are encountered: (i) in many cases, the cellular distribution of TPPII-LI and CCK-LI is ‘‘complementary’’, i.e. fully consistent with this hypothesis; (ii) in several neuronal populations in which we detected the peptidase, the presence of the neuropeptide was also reported; (iii) in some areas, our data do not support this hypothesis, since one marker occurs in the absence of the other. These three situations will be illustrated by a few typical examples taken from those described more extensively in the Results section and summarized in Table 1. The first situation was encountered in a number of areas where CCK-LI and TPPII-LI occurred at high levels (in some cases together with high CCK

1236

P. Facchinetti et al.

Fig. 5. Ultrastructural detection of TPPII-LI in the cerebral cortex and hypothalamus. (A) Immunoperoxidase detection showing the presence of electron-dense reaction product in neuronal somata and dendrites (arrows). Inset: higher magnification showing the association of electron-dense reaction product with the reticulum (crossed arrows), the outer membrane of mitochondria (small arrows) and the inner side of the plasma membrane (arrowhead). (B–D) Silver-enhanced immunogold detection (arrowheads) showing the presence of TPPII-LI in dendrites (B) and its association with non-synaptic dendritic membrane (C, D). Scale bars=5 µm (A), 1 µm (B), 0.5 µm (C, D), 1 µm (inset).

receptor binding) and with strikingly similar distributions at the regional level, as was the case in the piriform or cingulate cortex, olfactory tubercles, median eminence, areas surrounding the fourth ventricle, etc. At the cellular level as well, many examples of complementary distributions can be underlined. In the main olfactory bulb, TPPII-LI is found in mitral cells and CCK-LI in tufted cells, the apical dendrites of which, like those of mitral cells, establish synapses with dendrites of granule cells, an organization consistent with a possible role of the peptidase in CCK-8 inactivation therein. In the cerebral cortex, in which TPPII-LI is distributed in a laminar fashion, pyramidal cells of layers III and V and thick emerging dendrites were positive, whereas a large number of local circuit

neurons, with processes coursing perpendicularly to these dendrites, was CCK positive. A parallel situation seems to occur in the hippocampus, where TPPII-LI is detected in pyramidal cells, whereas CCK neurons are interneurons modulating the activity of these cells.29 In areas of the septal region and basal ganglia, e.g., in the bed nucleus of the stria terminalis or nucleus accumbens shell, TPPII-LI was present in medium-sized neurons, a category of cells innervated by dopaminergic neurons mainly emanating from the ventral tegmental area and well known to co-express CCK-LI.16,18 Electrophysiological, neurochemical and behavioral studies suggest that dopamine and CCK-8 interact at these levels (reviewed in Ref. 4). The localizations of TPPII-LI which are, perhaps, the most supportive for a role of the peptidase in

Cholecystokinin-inactivating peptidase

CCK inactivation are those observed all along the pathways through which the endogenous or exogenous neuropeptide is thought to control food intake by promoting satiety. In agreement, TPPII-LI is present in sensory vagal neurons, as shown at the level of the nodose ganglion,43 in the nucleus of the solitary tract, where numerous cells were labeled, and in the paraventricular and ventromedial hypothalamic nuclei. At all these levels where CCKA receptors were demonstrated by binding and in situ hybridization studies,19,29 local administration of CCK induces a pro-satieting effect mediated by CCKA receptors, whereas selective CCKA receptor antagonists have the opposite effects (reviewed in Ref. 50). In addition, inhibition of TPPII induces satiation in rodents selectively via a CCKA receptor-mediated effect. Furthermore, the neuropeptide CCK-8 is an excellent TPPII substrate, whereas the hormone CCK-33 is not.43 Taken together, these various observations support the view that CCK-8 released from enteric neurons during digestion affects vagal afferents projecting to the nucleus tractus solitarius; this information is further processed by CCK neurons projecting to the hypothalamus, and CCKergic transmission along this chain is mediated by CCKA receptors and controlled by TPPII. The second situation corresponds to the peptidase apparently being present in CCK neurons themselves, as shown for magnocellular and parvocellular neurons of the hypothalamic supraoptic and/or paraventricular nucleus. Co-existence of CCK and oxytocin in magnocellular neurons, and the presence of CCK-immunoreactive terminals in the median eminence and posterior pituitary were demonstrated.54,56 We show here that TPPII-positive axons are also present in the internal layer of the median eminence, and in situ hybridization detected TPPII mRNA in the anterior lobe of the pituitary.43 It therefore seems that TPPII may participate in the inactivation of CCK, affecting the liberation of releasing factors in the median eminence or reaching the pituitary via either the portal or the neuronal system. In comparison, it is well established that acetylcholinesterase is expressed not only by cholinoceptive, but also by cholinergic neurons, and that neprilysin, an enkephalin-degrading peptidase, is detected in enkephalinergic neurons.35 Other classes of neuron in which CCK-LI and TPPII-LI may co-exist include pyramidal neurons in the cerebral cortex or neurons in the ventral tegmental area and pars compacta or lateralis of the substantia nigra.15,57 In contrast, the sensory cranial nerve nuclei ganglia, which are TPPII positive and contain CCK-LI, were more recently found not to contain authentic CCK-8, but rather a cross-reacting substance, presumably calcitonin gene-related peptide.16 The third situation corresponds to a number of areas displaying either CCK-LI or TPPII-LI in the

1237

absence of the other ‘‘partner’’. As an example, in the molecular layers of the cerebral cortex or CA1 hippocampal region, in which CCK-LI terminals54 and CCKA and CCKB receptors are detected,19,29 TPPII-LI was not detectable. TPPII-LI also seems absent from the granule cells of the dentate gyrus, which are depolarized by CCK.10 In such areas, processes other than cleavage by TPPII may be involved in CCK inactivation, e.g., diffusion or hydrolysis by other peptidases which remain to be identified. The reverse situation is well illustrated in the case of the cerebellum, where TPPII immunoreactivity as well as TPPII catalytic activity (Rose C., unpublished observation) is not accompanied by any detectable CCK-LI or CCK receptor binding. The function of cerebellar TPPII, if any, may be related to the hydrolysis of a peptide distinct from CCK. Indeed, as a rule, peptidases are less specific than other neurotransmitter-inactivating devices, i.e. transporters or monoamine-metabolizing enzymes. Purified TPPII cleaves only CCK-8 with a high specificity constant (the ratio of Kcat/Km), but a few other neuropeptides, which are still cleaved at a reasonable rate (e.g., neurokinin, somatostatin, vasopressin, dynorphin A), may represent physiological substrates in some neuronal systems. The various criteria for a peptidase to be considered as a neurotransmitter-inactivating enzyme comprise the association with the external side of the plasma membrane.47 That TPPII is partly an ectopeptidase was suggested by its subcellular distribution in the brain and studies of CCK hydrolysis by extensively washed brain slices.3,41–43 Nevertheless, it was of interest to confirm this localization by studies at the ultrastructural level. TPPII-LI was found to be partly associated with the somatodendritic plasma membrane. However, the localization of the epitope on the cytosolic side of the plasma membrane hardly seems compatible with its postulated inactivating activity of presynaptically released CCK.43 One possibility is that this protein would have been washed out from the external side of the membrane during the fixation process. This is a possibility, since TPPII does not contain any hydrophobic stretch compatible with a transmembrane domain and appears to be bound to the plasma membrane through a glycosyl phosphatidyl anchor.43 However, the question remains of how TPPII could be translocated from the cytosolic to the external side of the plasma membrane. TPPII does not contain a signal peptide. Therefore, it is not likely that this molecule follows the classical secretion pathways, passing sequentially through the rough endoplasmic reticulum, Golgi apparatus, trans-Golgi network and vesicles. In any case, the silver-enhanced gold particles were never detected within a membrane-bound structure. Another possibility would be that the latter would be fragile and destroyed during the fixative procedure. Such is the case of caveolae, which have been proposed to be involved in signal transduction;

1238

P. Facchinetti et al.

furthermore, caveolae-like microdomains are found in neurons (reviewed in Ref. 31). Indeed, such structures are enriched in glycosyl phosphatidyl-anchored proteins.14 Alternatively, some proteins, such as basic fibroblast growth factor1 or interleukin-1,44 lack signal peptides, but are secreted. Whatever the mechanisms, the presence of TPPII close to the inner side of the plasma membrane would allow a rapid externalization of the molecule.

CONCLUSION

The cellular localization of TPPII in many (but not all) brain areas seems consistent with its hypothesized role, whereas the present ultrastructural studies could not confirm its ectopeptidase localization. In any case, the function of a major fraction of the enzyme associated with the cytoplasm of neurons remains to be clarified.

REFERENCES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Abraham J. A., Mergia A., Whang J. L., Tumolo A., Friedman J., Hjerrild K. A., Gospodarowicz D. and Fiddes J. C. (1986) Nucleotide sequence of a bovine clone encoding the angiogenic protein, basic fibroblast growth factor. Science 233, 545–548. Burgunder J. M. and Young W. S. (1990) Cortical neurons expressing the cholecystokinin gene in the rat: distribution in the adult brain, ontogeny and some of their projections. J. comp. Neurol. 300, 26–46. Camus A., Rose C. and Schwartz J.-C. (1989) Role of a serine endopeptidase in the hydrolysis of exogenous cholecystokinin by brain slices. Neuroscience 29, 595–602. Crawley J. N. and Corwin R. (1994) Biological actions of cholecystokinin. Peptides 15, 731–755. Danielsen E. M., Sjo¨stro¨m M., Noren O. and Dabelsteen E. (1977) Immunoelectrophoretic studies on pig intestinal brush border proteins. Biochem. biophys. Acta 494, 332–342. Della Fera M. A., Baile C. A. and Beinfeld M. C. (1982) Cerebral ventricular transport and uptake: importance for CCK-mediated satiety. Peptides 3, 963–968. Deschodt-Lanckman M., Bui N. D., Koulischer D., Paroutoud P. and Strosberg A. D. (1983) Cholecystokinin octaand tetrapeptide degradation by synaptic membranes. II. Solubilization and separation of membrane-bound CCK-8 cleaving enzyme. Peptides 4, 71–78. Deschodt-Lanckman M., Goe Diem Bui N., Noyer M. and Christophe J. (1981) Degradation of cholecystokinin-like peptides by a crude rat brain synaptosomal fraction: a study by high pressure liquid chromatography. Regul. Pept. 2, 15–30. Deschodt-Lanckman M., Koulischer D., Przedborski S. and Lauwereys M. (1984) Cholecystokinin octa- and tetrapeptide degradation by synaptic membranes. III. Inactivation of CCK-8 by a phosphoramidon-sensitive endopeptidase. Peptides 5, 649–651. Dodd J. and Kelly J. S. (1981) The actions of cholecystokinin and related peptides on pyramidal neurons of the mammalian hippocampus. Brain Res. 205, 337–350. Fallon J. H. and Seroogy K. B. (1985) The distribution and some connections of cholecystokinin neurons in the rat brain. Ann. N. Y. Acad. Sci. 448, 121–132. Giros B., Gros C., Solhonne B. and Schwartz J.-C. (1985) Characterization of aminopeptidases responsible for inactivating endogenous (Met5)enkephalin in brain slices using peptidase inhibitors and anti-aminopeptidase M antibodies. Molec. Pharmac. 29, 281–287. Giros B., Llorens-Cortes C., Gros C. and Schwartz J.-C. (1986) The endogenous tripeptide Tyr-Gly-Gly as a possible metabolite of opioid peptides in rat brain: identification, regional distribution, effects of lesions and formation in depolarized slices. Peptides 7, 669–677. Gorodinski A. and Harris D. A. (1995) Glycolipid-anchored proteins in neuroblastoma cells form detergent-resistant complexes devoided of caveolin. J. Cell Biol. 129, 619–627. Ho¨kfelt T., Cortes R., Schalling M., Ceccatelli S., Pelto-Huikko M., Persson H. and Villar M. J. (1991) Distribution patterns of CCK and CCK mRNA in some neuronal and non-neuronal tissues. Neuropeptides 19, 31–43. Ho¨kfelt T., Herrera-Marchitz M. and Seroogy K. (1988) Immunohistochemical studies on cholecystokinin (CCK)immunoreactive neurons in the rat using sequence specific antisera and with special reference to the caudate nucleus and primary sensory neurons. J. chem. Neuroanat. 1, 11–52. Ho¨kfelt T., Rehfeld J. F., Skirboll L., Ivemark B., Goldstein M. and Markey K. (1980) Evidence for coexistence of dopamine and CCK in mesolimbic neurons. Nature 285, 476–477. Ho¨kfelt T., Skirboll L., Rehfeld J. F., Goldstein M., Markey K. and Dann O. (1980) A subpopulation of mesencephalic dopamine neurons projecting to limbic areas contains a cholecystokinin-like peptide: evidence from immunohistochemistry combined with retrograde tracing. Neuroscience 5, 2093–2124. Honda T., Wada E., Battey J. F. and Wank S. A. (1993) Differential gene expression of CCKA and CCKB receptors in the rat brain. Molec. cell. Neurosci. 4, 143–154. Hughes J., Dockray G. J. and Woodruff G. (eds) (1989) The Neuropeptide Cholecystokinin (CCK), pp. 1–244. Ellis Horwood, Chichester. Lantos T. A., Gorcs T. J. and Palkovits M. (1995) Immunohistochemical mapping of neuropeptides in the premammillary region of the hypothalamus in rats. Brain Res. Rev. 20, 209–249. Llorens-Cortes C., Gros C. and Schwartz J.-C. (1986) Steady state level and turnover rate of the tripeptide Tyr-Gly-Gly as indexes of striatal enkephalin release in vivo and their reduction during pentobarbital anesthesia. Proc. natn. Acad. Sci. U.S.A. 83, 6226–6230. Malfroy B., Swertz J. P., Guyon A., Roques B. P. and Schwart J.-C. (1978) High affinity enkephalin-degrading peptidase in brain is increased after morphine. Nature 276, 523–526. Marcel D., Pollard H., Verroust P., Schwartz J.-C. and Beaudet A. (1990) Electron microscopic localization of immunoreactive enkephalinase (EC 3.4.24.11) in the neostriatum of the rat. J. Neurosci. 10, 2804–2817.

Cholecystokinin-inactivating peptidase 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

50. 51. 52. 53. 54. 55. 56. 57.

1239

Matsas R., Kenny A. J. and Turner A. J. (1986) An immunohistochemical study of endopeptidase 24.11 (‘‘enkephalinase’’) in the pig nervous systems. Neuroscience 18, 991–1012. Matsas R., Turner A. J. and Kenny A. J. (1984) Endopeptidase 24.11 and aminopeptidase activity in brain synaptic membranes are jointly responsible for the hydrolysis of cholecystokinin octapeptide (CCK-8). Fedn. Eur. biochem. Socs Lett. 175, 124–128. McDermott J. R., Dodd P. R., Edwardson J. A., Hardy J. A. and Smith A. J. (1983) Pathway of inactivation of cholecystokinin octapeptide (CCK-8) by synaptosomal fractions. Neurochem. Int. 5, 641–647. McKelvy J. F. and Blumberg S. (1986) Inactivation and metabolism of neuropeptides. A. Rev. Neurosci. 9, 415–434. Moran T. H. and McHugh P. R. (1990) Cholecystokinin receptors. In Handbook of Chemical Neuroanatomy: Neuropeptides in the CNS (eds Bjo¨rklund A., Ho¨kfelt T. and Kuhar M. J.), Vol. 9, Part 2, pp. 455–476. Elsevier Science, Amsterdam. Najdovski T., Collette N. and Deschodt-Lanckman M. (1985) Hydrolysis of the C-terminal octapeptide of cholecystokinin by rat kidney membranes: characterization of the cleavage by solubilized endopeptidase 24.11. Life Sci. 37, 827–834. Parton R. G. (1996) Caveolae and caveolins. Curr. Opin. Cell Biol. 8, 542–548. Patey G., De la Baume S., Schwartz J.-C., Gros C., Roques B. P., Fournie´-Zaluski M. C. and Soroca-Lucas E. (1981) Selective protection of methionine enkephalin released from brain slices by enkephalinase inhibition. Science 212, 1153–1155. Persson H., Rehfeld J. F., Ericsson A., Schalling M., Pelto-Huikko M. and Ho¨kfelt T. (1989) Transient expression of the cholecystokinin gene in male germ cells and accumulation of the peptide in the acrosomal granule: possible role of cholecystokinin in fertilisation. Proc. natn. Acad. Sci. U.S.A. 86, 6166–6170. Peters A., Miller M. and Kimerer L. M. (1983) Cholecystokinin-like immunoreactive neurons in rat cerebral cortex. Neuroscience 3, 431–448. Pollard H., Bouthenet M. L., Moreau J., Souil E., Verroust P., Ronco P. and Schwartz J.-C. (1989) Detailed immunoautoradiographic mapping of enkephalinase (EC 3.4.24.11) in rat central nervous system: comparison with enkephalins and substance P. Neuroscience 30, 339–376. Pollard H., Llorens-Cortes C., Couraud J. Y., Ronco P., Verroust P. and Schwartz J.-C. (1987) Enkephalinase (EC 3.4.24.11) is highly localized to a striatonigral pathway in rat brain. Neurosci. Lett. 77, 267–271. Rehfeld J. F. (1978) Immunochemical studies on cholecystokinin. II. Distribution and molecular heterogeneity in the central nervous system and small intestine of man and dog. J. biol. Chem. 253, 4022–4030. Rehfeld J. F. (1985) Neuronal cholecystokinin: one or multiple transmitters? J. Neurochem. 44, 1–10. Rehfeld J. F. and Kruse-Larsen C. (1978) Gastrin and cholecystokinin in human cerebrospinal fluid. Immunochemical determination of concentrations and molecular heterogeneity. Brain Res. 155, 19–26. Roberts G. W., Woodhams P. L., Polak J. M. and Crow T. J. (1982) Distribution of neuropeptides in the limbic system of the rat: the amygdaloid complex. Neuroscience 7, 99–132. Rose C., Camus A. and Schwartz J.-C. (1988) A serine peptidase responsible for the inactivation of endogenous cholecystokinin in brain. Proc. natn. Acad. Sci. U.S.A. 85, 8326–8330. Rose C., Camus A. and Schwartz J.-C. (1989) Protection by serine peptidase inhibitors of endogenous cholecystokinin released from brain slices. Neuroscience 29, 583–594. Rose C., Vargas F., Facchinetti P., Bourgeat P., Bambal R. B., Bishop P. B., Chan S. M. T., Moore A. N. J., Ganellin C. R. and Schwartz J.-C. (1996) Characterization and inhibition of a cholecystokinin-inactivating serine peptidase. Nature 380, 403–409. Rubartelli A., Cozzolino F., Talio M. and Sitia R. (1990) A novel secretory pathway for interleukin-1 beta, a protein lacking a signal sequence. Eur. molec. Biol. Org. J. 9, 1503–1510. Schiffmann S. N. and Vanderhaeghen J. J. (1991) Distribution of cells containing mRNA encoding cholecystokinin in the rat central nervous system. J. comp. Neurol. 304, 219–233. Schwartz J.-C. (1983) Metabolism of enkephalins and the inactivating neuropeptidase concept. Trends Neurosci. 6, 45–48. Schwartz J.-C., Malfroy B. and De la Baume S. (1981) Biological inactivation of enkephalins and the role of enkephalin-dipeptidyl-carboxypeptidase (enkephalinase) as neuropeptidase. Life Sci. 29, 1715–1740. Seroogy K., Fallon J. H., Loughlin S. E. and Leslie F. M. (1985) Few cortical cholecystokinin immunoreactive neurons have long projections. Expl Brain Res. 59, 533–542. Seroogy K., Schalling M., Brene S., Dagerling A., Chai S. Y., Ho¨kfelt T., Persson H., Brownstein M., Huan R., Dixon J., Filler D., Schlessinger D. and Goldstein M. (1989) Cholecystokinin and tyrosine hydroxylase messenger RNAs in neurons of rat mesencephalon: peptide/monoamine coexistence studies using in situ hybridization combined with immunocytochemistry. Expl Brain Res. 74, 149–162. Smith G. P. and Gibbs J. (1994) Satiating effect of cholecystokinin. Ann. N. Y. Acad. Sci. 713, 236–241. Solhonne B., Gros C., Pollard H. and Schwartz J.-C. (1987) Major localization of aminopeptidase M in rat brain microvessels. Neuroscience 22, 225–232. Steardo L., Knight M., Tamminga C. A., Barone P., Kask A. M. and Chase T. N. (1985) CCK-26-33 degrading activity in brain and nonneuronal tissue: a metalloendopeptidase. J. Neurochem. 45, 784–790. Tomkinson B. and Jonsson A. K. (1991) Characterization of cDNA for human tripeptidyl peptidase II: the N-terminal part of the enzyme is similar to subtilisin. Biochemistry 30, 168–174. Vanderhaeghen J. J. (1985) Neuronal cholecystokinin. In Handbook of Chemical Neuroanatomy: GABA and Neuropeptides in the CNS (eds Bjo¨rklund A. and Ho¨kfelt T.), Vol. 4, Part 1, pp. 406–432. Elsevier Science, Amsterdam. Vanderhaeghen J. J. and Crawley J. N. (1984) Neuronal cholecystokinin. Ann. N. Y. Acad. Sci. 448, 1–697. Vanderhaeghen J. J., Lotstra F., Vandesande F. and Dierickx K. (1981) Coexistence of cholecystokinin and oxytocin-neurophysin in some magnocellular-hypothalamo-hypophyseal neurons. Cell Tiss. Res. 221, 227–231. Vanderhaeghen J. J. and Schiffmann S. N. (1992) Distribution of brain neuronal CCK. In Multiple Cholecystokinin Receptors in the CNS (eds Dowrish C. T., Cooper S. J., Iversen S. D. and Iversen L. L.), pp. 38–56. Oxford University Press, New York.

1240 58.

P. Facchinetti et al. Zuzel K. A., Rose C. and Schwartz J.-C. (1985) Assessment of the role of ‘‘enkephalinase’’ in cholecystokinin inactivation. Neuroscience 15, 149–158. (Accepted 28 April 1998)