240 Brain Research, 604 (1993) 240-250 © 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.00

BRES 18552

Comparative autoradiographic distribution of central to (benzodiazepine) modulatory site subtypes with high,

intermediate and low affinity for zolpidem and alpidem

J. Benav ides a, B. P e n y a, D. R u a n o b j . Vi tor ica b and B. Sca t ton a

a Department of Biology, Synthelabo Recherche - - LERS, Bagneux (France) and b Department of Biochemistry, School of Pharmacy, Sevilla (Spain)

(Accepted 15 September 1992)

Key words: Benzodiazepine (to) modulatory site subtype; Autoradiographic distribution; Adult; Newborn; Rat; Central nervous system

Pharmacological characterization of [3H]benzodiazepine binding to membrane preparations of adult rat hippocampus and neonatal rat brain have demonstrated, in addition to the to I and w 2 populations of central to benzodiazepine binding sites associated with GABA A receptors, the existence of binding sites with p.M affinity for the imidazopyridines zolpidem and alpidem. In the present study we have investigated their comparative autoradiographic distribution using [3H]flumazenil as a ligand. In the neonatal rat CNS, the imidazopyridine derivatives zolpidem and alpidem were found to discriminate two [3H]flumazenil binding site populations with an ICso value ratio of more than 200-fold. In the different regions investigated (spinal cord, striatum, neocortex and inferior colliculus) the low affinity component had IC5o values of 20-40 p.M (zolpidem) and 5-15 /~M (alpidem) and accounted for ca. 50% of the total binding site population. In the adult rat, these imidazopyridine derivatives displayed a greater displacing potency in the cerebellum (ICs0 = 6 and 36 nM, respectively) than in the hippocampus (IC50 = 37 and 403 nM, respectively). In the cerebellum, [3H]flumazenil binding was fully displaced by 1 /zM of either compound and Hill coefficients of displacement curves were close to unity. In the hippocampus, 25% of [3H]flumazenil binding were resistant to 3 tzM zolpidem or 1 /~M alpidem, but were displaced by 100 p.M of either compound. CL 218,872 also displayed a greater displacing potency in the cerebellum (IC50 = 83 nM) than in the hippocampus (ICs0 = 711 nM) but [3H]flumazenil binding in the hippocampus was fully displaced by 10/~M of this compound. In adult rat hippocampus, zolpidem and alpidem were found to discriminate between three central to site subtypes which display high (ICs0 = 31 and 6.1 nM, respectively; tol sites), intermediate (ICs0 = 480 and 122 nM, respectively; to2t) and low (ICs0 = 20100 and 2300 nM, respectively; W2L) affinity for these imidazopyridine derivatives. In contrast, CL 218,872 discriminated between to1 and w 2 sites but not between two ~o 2 receptor subpopulations, to1 sites were mainly localized in layer IV of the sensorimotor cortex, cerebellum, substantia nigra, olfactory bulb and inferior colliculus. ~o2t sites were present in the cortical mantle (with higher levels in the cingulate and olfactory than in the sensorimotor cortex) and in subcortical (hippocampus, hypothalamus and nucleus accumbens) limbic structures. In the hippocampus, hypothalamus, spinal cord and nucleus accumbens, to2L sites accounted for more than 25% of the specific [3H]flumazenil binding; the density of these sites was minor in the cortex and in most pyramidal and extrapyramidal system structures. In conclusion, the present results indicate that in adult and neonatal rat CN8 sections zolpidem and alpidem can discriminate between 3 populations of [3H]flumazenil binding sites that may be akin to to binding sites associated with GABA A receptors containing al-subunits ( tol) , a 2- or a3-subunits (to2i) and as-subunits ((O2L) , respectively.

I N T R O D U C T I O N

The neuropharmacological effects of the benzodi- azepines are mediated by their interaction with central to (benzodiazepine) modulatory sites associated with the GABA A receptor complex. Biochemical and phar- macological studies have clearly established the hetero- geneity of these s i tes 2°'23'29. Furthermore, recent mo- lecular cloning data have provided a molecular basis for this diversity by demonstrating the hetero-

oligomeric nature of functional GABA a receptors in transfection experiments using GABA A subunit combi- nants constituted of a, /3, 7, and/or 6 subunit v a r i a n t s 6,24,26.

Displacement studies of [3H]flumazenil binding to transfected cells have shown that the imidazopyridine derivatives zolpidem z7 and alpidem (D. Pritchett, per- sonal communication) possess high affinity for the a 1 subunit-containing GABA A receptors, an intermediate affinity for the a 2- or a3-containing variants and very

Correspondence: J. Benavides, Department of Biology, Synthelabo Recherche (LERS), 31 Avenue Paul Vaillant-Couturier, 92220 Bagneux, France. Fax: (33) (1) 45-36-20-00.

241

low (IzM) affinity for the as-containing variants when these subunits are expressed together with /33 and Y2 subunits. In these transfected receptors, the triazolopy- ridazine CL 218,872 possesses a greater affinity for the a 1- or as-subunit containing variants than for GABA A receptors containing ot 2- or Ota-subunits 27. None of these compounds has affinity for the a6-subunit con- taining c o n s t r u c t s 17. *

These transfection studies suggest that to modula- tory sites possess a greater heterogeneity than that previously recognized from displacement studies of [3H]benzodiazepine binding to native to sites in rodent and primate membranes and sections. Indeed, in these latter studies, zolpidem, alpidem 1,2,a334, quazepam, oxoquazepam 3'7, CL 218,87211 and some /3-carboline derivatives 4'5'1° were found to discriminate only two to modulatory (tox and to2) site subtypes. It should be mentioned, however, that recent studies of [3H]fluni- trazepam binding to rat neonatal cortical or adult hippocampal membranes using a wide range of zolpi- dem concentrations 28, or to rat spinal cord membranes using alpidem as displacing agent t9, have demonstrated the presence of a /zM affinity to site for these imida- zopyridines, thereby suggesting further to subytpe het- erogeneity. However it is not know which is the subunit composition of native GABA A receptor subtypes and which could be the functional relevance (if any) of each of these subunit combinations.

The main objective of the present work was thus, to determine the regional distribution of multiple native to modulatory sites with differing affinities for zolpi- dem and alpidem in the adult and neonatal rat brain. Although the autoradiographic distribution in the rat CNS of high (to1) and low (to E) affinity binding sites for zolpidem has been previously characterized 22, the re- cent demonstration of the heterogeneity of the low affinity (to2) binding site in rat brain membranes 28 rendered necessary a reevaluation of the anatomical distribution of these three binding site populations. To this end, we have studied by quantitative autoradiogra- phy the displacement of [3H]flumazenil binding by selective and non-selective to site subtype ligands (zolpidem, alpidem, CL 218,872, flumazenil) in the

neonatal (4-day-old) rat CNS as well as in the adult rat cerebellum and hippocampus. In the latter region as well as in the neonatal rat brain, membrane binding experiments have demonstrated the existence of three to site populations and in situ hybridization studies have revealed the presence of high levels of mRNA encoding the as-subunit 1s'31, in addition to a 1- and ae-subunit mRNAs. Finally, by using zolpidem concen- trations which displaced either the high and intermedi- ate affinity components or the high affinity component exclusively, we have established the autoradiographic distribution of these populations of [3H]flumazenil binding sites in the neonatal and adult rat brain and spinal cord.

MATERIALS AND METHODS

Adult (male, 150 g body weight) or 4-day-old Sprague-Dawley rats were obtained from Charles River, France. Animals were sacri- ficed by decapitation, the brain and spinal cord were removed rapidly, frozen in isopentane chilled at - 4 8 ° C and mounted onto a microtome chuck. Fifteen-/xm-thick sagittal or coronal sections were then prepared at - 2 0 ° C using a Leitz 1720 microtome-cryostat and were thaw-mounted onto chromalum-gelatin subbed microscope glass slides. Tissue sections were incubated for 60 min at 4°C in 4 ml of a Tris-HCl solution (50 mM, pH 7.4) in the presence of 1 nM of [3H]flumazenil (spec. act. 92 Ci/mmol) . Non specific binding was defined by incubating adjacent tissue sections with 1 /xM of fluni- trazepam and represented less than 10% of total binding. The incubation was terminated by rinsing sections twice for 5 min in the cold incubation buffer. Sections were then dipped briefly in cold distilled water and dried rapidly under a stream of cold air. Autora- diographs were prepared from sections which together with radioac- tivity standards (Amersham, Buckingamshire, UK) were apposed to a radiographic film ([3H]Ultrofilm) for 2 weeks 2. Local binding densi- ties were quantified using a variable frame computerized densitome- ter (Biocom, Les Ulis, France).

For competition experiments, agents to be tested were added to the incubation medium from a 1000-fold concentrated ethanolic solution. Controls were incubated in the presence of an equivalent concentration of ethanol (0.1%).

CL 218,872, zolpidem, alpidem and flumazenil were synthetised in the Chemistry Department of Synth61abo Recherche (Bagneux, France). [3H]Flumazenil was purchased from NEN (Bad Homburg, FRG).

Competitive inhibition curves were analysed using a RS1 scien- tific calculation package installed in a VAX 5000 computer.

IC50 values were computed by fitting experimental data to the following equations:

y = 1 0 0 X l C s 0 / ( x + IC50); (one site)

y = a × IC 50u / ( x + IC 50H ) -}- (100 -- a ) X IC501 / ( x + IC501 );

* Nomenclature note. Benzodiazepine (to) modulatory sites coupled to GABA A receptors have been classified into to 1 (BZ1, type I) and toe (BZ2, type II) sites according to their respective high and low affinity for CL 218,872, several fl-carboline derivatives, zolpidem, alpidem, quazepam and oxoquazepam 12,23. However, a larger heterogeneity of to sites has been recently revealed by displacement studies in transfected GABA A receptors. To take into account this diversity of to modulatory sites the IUPHAR Committee on Receptor nomenclature has proposed to classify to sites according to the a-subunit to which they are associated, i.e. tox (ax), tOE (0/2), 0)3 (Ot3) and so on. Moreover, GABA A receptor will be named according to the pharmacology of their to sites as GABAAo (no to site), GABAAI (tol site), GABAA2 (to2 site) and so on. Because the exact subunit composition of each to modulatory site subtype is still unknown and since more than one a-subunit variant may exist in a GABAA-to receptor complex 16, we have termed those [3H]flumazenil binding sites that possess high, intermediate and low affinity for zolpidem and alpidem as to1, to2i and toEL sites, respectively. The as-subunit corresponds to the ot 4 sequence of MacLennan et a l) 8.

242

(two sites) or

y = a x IC5o H / ( x + IC50H) + b × ICso I / ( x + ICsoI)

+ (100 - a - b) × ICs0 L / ( x + ICsoL); (three sites)

where y is the percent binding at concentration x; a and b are the percentages of each component and ICs0H, ICs01 and ICs0 L the ICs0 value for each component (high, intermediate and low affinity).

The adequacy of the fit to one, two or three sites was evaluated by using the F-ratio test 21.

RESULTS

Neonatal rat CNS

Competitive inhibition curves of [3H]flumazenil binding by cold flumazenil in sections from several neonatal CNS regions were monophasic, with Hill slopes close to unity (Fig. 1 and Table I). A similar affinity of flumazenil was observed in the four struc- tures examined (cerebral cortex, striatum, inferior col- liculus and spinal cord). Flumazenil displayed a slightly lower displacing potency in the neonatal than in the adult rat brain (compare Table I and Table II).

Hill coefficients of [3H]flumazenil displacement curves by CL 218,872 were smaller than unity in all the neonatal structures examined but no regional differ- ence in displacing potency was seen (Table I and Fig. 1). Fitting of displacement curves indicated the pres- ence of two components with a 20-fold difference in affinity and each amounting to approximately half of the total specific [3H]flumazenil binding (Table I).

Zolpidem and alpidem displacing curves were shal- lower than those of CL 218,872 and Hill coefficients were always less than 0.5 (Fig. 1 and Table I). These curves fitted well to a two site model with a more than 200-fold difference in displacing potency between the two components. The respective percentages of high- and low-affinity sites were similar to those observed for CL 218,872. The autoradiographic distribution of these

two sites is illustrated in Fig. 2.

Adult rat CNS

Monophasic inhibition curves of [3H]flumazenil binding by cold fulmazenil were also observed in the

t9 Z t-4

Z I-d ¢fl

rD b4

H ¢0

60

{3 Z H

z

m

¢D b4 hq H ¢D

o~O

A w

(9 8 z

H

Z H

m

t) H 50

H (3

In

4e

o o , i

[ FLUMAZENIL ], (nM).

10o,

t9 Z

Z H 0~ u H 50

I-'4

¢D ja~ ¢/)

80

# 1' ,'o Io'oo ' ioooo

[ ZOLPIDEM ], (nM) .

,// ,~ i lOO lOOO ioooo

[ CL 218,872 ] , (nM).

\

' / / i' 1'0 1o'00 ' IOOOO

[ ALPIDEM ], (nM) .

O SPINAL CORD

• STRIATUM

........ A CORTEX

-- -- • INF.COLLICULUS

Fig. 1. Displacement of [3H]flumazenil binding to CNS sections from neonatal rats by flumazenil, CL 218,872, zolpidem and alpidem. Sections were incubated in the presence of 1 nM [3H]flumazenil and increasing concentrations of displacing agents. Results are the mean of triplicate

determinations. Displacement curves were fitted to a one-site (flumazenil) or two-site (other compounds) displacement equation.

adult rat cerebellum and hippocampus and the displac- ing potency of this compound was similar in both structures (Table II and Fig. 3). In contrast, in the adult rat hippocampus, the displacement curve of [3H]flumazenil binding by CL 218,872 gave a Hill value significantly smaller than unity (Table II) and could be resolved into two components.

Displacement curves of [3H]flumazenil binding by zolpidem and alpidem in the adult rat hippocampus (CA fields) were very shallow (Hill slope = 0.5) and showed a net inflexion at concentrations ranging from 1 to 10 /xM (Fig. 3). Fitting of these curves to a biphasic or a triphasic displacement model was signifi- cantly better than fitting to a single component (Table II). The high-affinity component had an ICs0 about 100 times lower than the low affinity component which displays /xM affinity for zolpidem and alpidem. These sites with /xM affinity for the imidazopyridine deriva- tives represented ~ 25% of the total binding. Similar results were obtained when displacement curves in several other structures (ventromedial hypothalamus, dentate gyrus, nucleus accumbens and olfactory nu- cleus (results not shown)). In contrast, displacement curves of [3H]flumazenil binding to the adult rat cere- bellum by CL 218,872, zolpidem and alpidem were monophasic with Hill slopes close to unity. (Fig. 3 and Table II).

243

The comparative anatomical distribution of [3H]- flumazenil binding sites with high intermediate and low affinity for zolpidem was determined by autoradiogra- phy in sections incubated in the presence of 0, 100 and 3000 nM of zolpidem (Fig. 4 and Table III). A calcula- tion based on the IC50 values of each component indicated that 3000 nM zolpidem displaced more than 85% of the intermediate affinity component but less than 10% of the low-affinity component. A zolpidem concentration of 100 nM was subsequently selected in order to discriminate the high and intermediate com- ponents but, due to the smaller difference in affinity between these sites (16-fold), some cross-contamina- tion is to be expected.

As previously reported 22, [3H]flumazenil binding sites displaced by 100 nM zolpidem (o9 t sites) ac- counted for more than 50% of specific [3H]flumazenil binding in most of the CNS structures (Table III and Fig. 4). The highest levels of these high-affinity zolpi- dem binding sites were found in the olfactory bulb, followed by the inferior colliculus, ventral pallidum, molecular layer of the cerebellum, substantia nigra and the neocortex where densities in layer IV were higher than in the other layers. Intermediate densities were found in the hippocampus, amygdala, superior collicu- lus, nucleus accumbens, olfactory tubercule, zona in- certa and central grey, and low levels were quantified

TABLE I

Parameters of the displacement of [~H]flumazenil binding to brain and spinal cord sections from 4-day-old rats by flumazenil, CL 218,872, zolpidem and alpidem

Displacement parameters were calculated by iterative fitting of data from Fig. 1. P: statistical significance of a better fit to two sites as compared to a single site.

Displacing agent One site P (2S/1S) Two sites

IC5o (nM) nH IC5o (nM); % IC5o (nM); %

Flumazenil Spinal cord 2.9_+ 0.8 0.93_+ 0.06 > 0.50 - - Striatum 4.0_+ 0.8 0.78_+0.04 > 0.50 - - Cortex 3.7_+ 1.2 0.86_+ 0.10 > 0.50 - - Inferior colliculus 3 .8+ 0.4 0.91 _+0.12 > 0.50 - -

CL 218,872 Spinal cord 754 _+135 0.64_+0.05 0.04 178+55; (57%) 5300_+ 710; (43%) Striatum 797 _+146 0.61_+0.03 0.03 150_+16; (51%) 5300_+ 750; (49%) Cortex 864 _+ 122 0.62_+ 0.04 0.04 135 _+ 55; (47%) 4315 _+ 1400; (53%) Inferior colliculus 496 + 87 0.70-+ 0.0 0.06 187-+ 33; (68%) 4185 + 1410; (32%)

Zolpidem Spinal cord 2 350 + 342 0.43 -+ 0.05 0.02 205 -+ 21; (54%) 41500-+ 6 750; (46%) Striatum 12 943 -+ 620 0.46 _+ 0.09 0.04 629 _+ 29; (42%) 44 300 _+ 14 700; (58%) Cortex 4 640 _+ 525 0.46 -+ 0.04 0.04 234 _+ 65; (44%) 35 400 _+ 10 400; (56%) Inferior colliculus 459 _+ 68 0.42_+ 0.01 0.01 89 _+ 14; (64%) 23 300_+ 7 400; (36%)

Alpidem Spinal cord 182 _+ 61 0.38_+0.03 0.03 25-+ 8; (62%) 9900+ 2500; (38%) Striatum 967 + 234 0.45 _+ 0.04 0.01 77_+ 17; (52%) 13 400_+ 2400; (48%) Cortex 545 _+ 49 0.44_+0.04 0.01 57_+ 12; (57%) 12000_+ 3300; (43%) Inferior colliculus 75 _+ 6 0.46_+0.05 0.02 29_+ 4; (61%) 5900_+ 1600; (39%)

244

=

~9

d

~a

E 8

©

8

o

o

~q

8

~o

+1

+1

oo

oo +1

+1

+1

c5 ii

v

+1

+1

+1

S ~ +r

+1

~5

II

V

+1

v

+[

A A A II A II A II

+1 +1 +1 +1 +1 +1 +1 +1

+1 +1 +1 +I +1 +1

~o~ ~

+1 +1

in the nucleus accumbens, striatum, hypothalamus and spinal cord.

[3H]Flumazenil binding sites with intermediate affinity for zolpidem (resistant to 100 nM but displaced by 3 /xM; 0)21 ) represented a lower portion of specific [3H]flumazenil binding (45-20%) than high-affinity sites (Table III and Fig. 4). The highest concentrations of these sites were found in the olfactory bulb followed by the dentate gyrus, the hippocampal CA fields, neo- cortex (with higher levels in the cingulate and olfactory than in the sensorimotor cortex), superior and inferior colliculi and thalamus. Structures endowed with inter- mediate densities of these sites included the amygdala, nucleus accumbens, ventral pallidum and hypothala- mus. In most of the other structures, low concentra- tions of [3H]flumazenil binding sites with intermediate affinity for zolpidem were found.

[3H]Flumazenil binding sites with /zM affinity for zolpidem (resistant to 3 /xM of zolpidem; w2L) repre- sented a minor component of [3H]flumazenil binding in most of the structures examined (Table III and Fig. 4). Structures in which (.oZL sites represented more than 20% of total [3H]flumazenil binding were the ventromedial hypothalamus, hippocampus, striatum,

3H-FLUMAZENIL BINDING SITES

3H-FLUMAZENIL BINDING SITES WITH HIGH AFFINITY FOR ZOLPIDEM

3H-FLUMAZENIL BINDING SITES WITH LOW AFFINITY FOR ZOLPIDEM

Fig. 2. Comparative autoradiographic distribution of [3H]flumazenil binding sites with high and low affinity for zolpidem in the neonatal rat brain. Virtual images were generated by computerized subtrac- tion of autoradiograms obtained from sections incubated in the presence of 1 nM [3H]flumazenil with or without 1000 nM zolpidem. [3H]Flumazenil binding sites with high and low affinity for zolpidem are defined as those displaced by or resistant to 1000 nM zolpidem.

245

nucleus accumbens and olfactory nucleus. In absolute terms, the highest binding densities (25-50 fmo l /mg tissue) of (.ozL sites were present in the olfactory bulb, hippocampus, dentate gyrus and ventromedial hypotha- lamus, whereas intermediate densities (15-25 fmol /mg of tissue) were found in the cingulate cortex, nucleus accumbens and thalamus. Most of the other structures examined displayed binding densities ranging between 5 and 15 fmol /mg of tissue but in the cerebellum, substantia nigra and inferior colliculus binding densi- ties were less than 5 fmol /mg of tissue.

DISCUSSION

The present results demonstrate the existence of [3H]flumazenil binding sites with txM affinity for zolpi- dem and alpidem in sections from neonatal and adult rat CNS. In the neonatal rat brain, these sites repre- sent around a half of the total [3H]flumazenil binding site population, while in the adult rat brain, they ac- count for a minor component ( -- 20%) of these binding sites.

In the neonatal rat, more than 50% of specific [3H]flumazenil binding was displaced by CL 218,872 with an ICs0 similar to that of the high-affinity compo- nent in the adult hippocampus and to that in the adult cerebellum. In contrast, zolpidem and alpidem exhib- ited a 6-8 fold greater IC50 for these high-affinity sites in the developing rat CNS than in the adult rat brain. A possible explanation for this discrepancy may lie in the very small difference (4-fold) in affinity of CL 218,872 for off and as-subunit-containing transfected GABA A receptors as compared to the imidazopy- ridines investigated (more than 40 fold) 27. Thus, in the developing rat CNS, the high-affinity component for CL 218,872 may be mainly accounted for by receptors containing ozs-subunits which cannot be differentiated by this compound from those containing al-SUbunits, whereas the high-affinity component for imidazopy- ridines may be associated to GABA A receptors con- taining the al-, oz 2- or a3-subunits. It is also expected that binding sites with low affinity for CL 218,872 would be associated to GABA A receptors containing a 2 or a3-subunits whereas those with low affinity for zolpidem and alpidem are likely associated to GABA A receptors containing as-subunits. Thus, although dis- placement curves of [3H]flumazenil binding by zolpi- dem, alpidem and CL 218,872 in neonatal brain struc- tures fit to a 2 site model, it is likely that each of these sites is heterogenous.

As previously demonstrated in hippocampal mem- branes 28, the analysis of the displacement curve of [3H]flumazenil binding to rat hippocampal sections by

246

zolpidem or alpidem demonstrate the existence of binding sites with txM affinity for these compounds, whilst in the CL 218,872 displacement curve this very low-affinity component is absent. The high-affinity component of CL 218,872 accounts for a fraction of [3H]flumazenil binding site population close to that of the high-affinity component of zolpidem and alpidem, while the low-affinity component of CL 218,872 has a binding capacity close to that of the summation of the intermediate and low-affinity components of these imi- dazopyridines.

The comparison of the relative affinities of zolpidem and alpidem for these three components with those for transfected GABA A receptors (Table IV) suggests that high, intermediate and low-affinity sites correspond to

GABA A receptors containing al-, a 2- (or a3-) and as-subunits, respectively. This potential association is also supported by the close correspondence observed

between the anatomical CNS distribution of these binding site populations (as assessed by autoradiogra- phy) and of mRNAs encoding different GABA A re- ceptor o~ subunits (as determined by in situ hybridiza- tion~5,18,3°'31). Thus, [3H]flumazenil binding sites with

high affinity for zolpidem (o) 1 sites) and m R N A encod- ing for the oq-subunit have a widespread distribution in the rat CNS with the highest levels found in the olfactory bulb, olfactory and cerebral cortex, ventral pallidum, substantia nigra, inferior colliculum and cerebellum. Likewise, low levels of both ozFsubunit m R N A and wl sites are found in the nucleus accum- bens, caudate-putamen, thalamus, hypothalamus and spinal cord. The highest levels of the m R N A encoding

for the az-subunit were quantified in the hippocampus and olfactory bulb, two regions also endowed with high levels of [3H]flumazenil binding sites with intermediate

affinity for zolpidem (w21 sites).

z H

z H izq

L) H

H 0

(,9

o A [] w

~ 8 1oo. ...... 8 [] clHZt~

H

o H 50"

H . O

O0

' / / o'o~ o., ,

FLUMAZENIL ], (riM) .

0 "1:I

0 '13 b " "0

1ooo

,n

o'-Z---- o I ioooo

[ CL 218.872 ], (nM) .

i()o- © z H z H cq

u H 50" H u p-I

O9 ~o

• B ' (_9

a Z H

H

CI • 1~ H [a4 H

o

0 . U]

• ' . . . . .

// , | i iolo ioi00 ioooo 0 io

[ ZOLPIDEM ], (nM).

- - O CEREBELLUM

........ [] HIPPOCAMPUS

0 0 ""ar:'l .,..

El "11

. / / : .'o ' iooo

" 1 3 . .11 . . . .

o o

I ioooo

[ ALPIDEM ], (nM).

Fig. 3. Displacement of [3H]flumazenil binding to adult rat brain sections by flumazenil, CL 218,872, zolpidem and alpidem. Sections were incubated in the presence of 1 nM [3H]flumazenil and increasing concentrations of displacing agents. Results are the mean of triplicate determinations. Displacement curves in the cerebellum were fitted to one site. Displacement curves in the hippocampus were fitted to one site

(flumazenil), two-site (CL 218,872) or three-site (zolpidem and alpidem) displacement equations.

O Z

Z

Z

¢

¢/) H / I--- o~

¢",1 :

¢.0 ILl I--- ( o

i!iiiiiii~iiii~i~i ~i ~

o9 ILl I---

( o

,,q:

"i~iijiiiiiiii!iiiiiii~iiiiiii~iiii~ijiiii~iiii!iiiii~i!~iiii~ 17 • ~ ~ ii~i~ii~ ii ¸~ ¸ i~

iii~ ~ ~ ~ ~!~i!i I , i ~ / ~ ~ilil ~

2 4 7

Zo ~ ~:~ ~..~ q

~= ~ ' ~

~ m

"~N~ 0 ~ ,....~ ~

a N . , = ~

.~ ~ ~

• ~ ~ ~-~

~ ~ ~

~ - ~ ~ , ~

m ~

N

248

TABLE Ill

Quantitative autoradiographic determination of the anatomical distribution of [3H]flumazenil binding sites with high, intermediate and low affinity for zolpidem in the adult rat CNS

High-affinity sites were defined as the fraction of specific [3H]flumazenil binding displaced by 100 nM zolpidem. Intermediate affinity sites were defined as the fraction of specific [3H]flumazenil binding resistant to 100 nM zolpidem but displaced by 3/xM zolpidem. Low-affinity sites were defined as the fraction of [3H]flumazenil binding resistant to 3/zM zolpidem. Values are the mean of the results obtained from 3 rats.

Structure Specific 3 H flumazenil fmol / mg tissue binding (% specific binding) ( fmol / mg tissue) High Intermedia te Low

affinity affinity affinity

Cortex Frontal 98 41 (42%) 42 (43%) 15 (15%) Cingulate 163 104 (64%) 42 (26%) 17 (11%) Frontoparietal layer I - I l l 149 96 (65%) 42 (28%) 11 (7%) Frontoparietal layer IV 204 139 (68%) 51 (25%) 14 (6%) Frontoparietal layer V-VI 135 87 (65%) 34 (24%) 14 (11%) Posterior cingulate 118 81 (69%) 28 (24%) 9 (8%)

Olfactory tubercle 122 79 (65%) 29 (24%) 14 (11%) Olfactory bulb 418 264 (63%) 117 (28%) 37 (9%) Olfactory nucleus 55 21 (37%) 22 (41%) 12 (23%) Amygdala 110 61 (56%) 36 (32%) 13 (12%) Nucleus accumbens 64 21 (33%) 26 (41%) 17 (27%) Striatum 42 17 (41%) 13 (32%) 12 (27%) Ventral pallidum 173 129 (75%) 32 (19%) 12 (6%) Septum 45 23 (51%) 12 (27%) 10 (22%) Globus pallidus 81 59 (73%) 14 (17%) 8 (11%) Thalamus 139 79 (57%) 42 (30%) 18 (13%) Hippocampal formation

Hippocampus proper 140 51 (37%) 45 (31%) 44 (32%) Dentate gyrus 183 87 (48%) 60 (33%) 36 (19%)

Hypothalamus Ventromedial 90 28 (32%) 34 (36%) 28 (32%) Dorsomedial 80 41 (52%) 24 (31%) 15 (17%) Lateral 65 41 (64%) 16 (23%) 8 (13%)

Zona incerta 77 53 (69%) 17 (23%) 7 (8%) Colliculi

Superior (superficial layer) 156 79 (51%) 49 (31%) 28 (18%) Inferior 180 129 (71%) 48 (27%) 3 (2%)

Central grey 101 63 (63%) 31 (31%) 7 (7%) Substantia nigra 123 94 (76%) 25 (21%) 4 (3%) Cerebellum

Molecular layer 115 88 (77%) 27 (23%) 0 (0%) Granular layer 50 34 (69%) 16 (30%) 0 (0%)

Spinal cord 40 14 (36%) 16 (38%) 10 (26%)

T h e s t r u c t u r e s w i t h t h e h i g h e s t l eve l s o f w2L s i tes

( [ 3 H ] f l u m a z e n i l b i n d i n g s i t es w i t h / xM a f f in i ty fo r

z o l p i d e m ) a r e t h e h i p p o c a m p u s , d e n t a t e gy rus a n d

o l f a c t o r y b u l b . T h e s e s t r u c t u r e s a r e a l so c h a r a c t e r i z e d

b y t h e g r e a t e s t h y b r i d i z a t i o n s i g n a l fo r t h e a s - s u b -

u n i t 18,3a. M o r e o v e r , r e g i o n s w i t h i n t e r m e d i a t e l eve l s o f

TABLE IV

Comparative affinity of zolpidem and CL 218,872 for native and transfected central o~ site subtypes

Tissue Drug Displacement of [3H]flumazenil binding; IC5o (nM)

High affinity Intermediate affinity Low affinity

Adult rat CNS sections

Neonatal rat CNS sections

Transfected GABA A receptors a

zolpidem CL 218,872 zolpidem CL 218,872 zolpidem CL218,872

36 280 83 1314 - 234

135 4300 al 83 T2 ~2 83 ~2 a3 83 T2

19 450 398 120 1786 1495

20100

35400

a5 83 ~2 > 15000

49O

a Data from Pritchett and Seeburg 25.

this mRNA (olfactory nucleus, pyriform cortex and amygdala) display appreciable amounts of [3H]fluma- zenil binding resistant to 3 /xM zolpidem. Another analogy between the distribution of w2L sites and of the mRNA encoding for the 0/5 subunit is the higher levels of these markers in deep than in superficial cortical layers. The fact that in the neonatal rats CNS, okZL sites account for a much greater percentage of specific [3H]flumazenil binding than in the adult ani- mal adds further support to the association of these sites with GABA A receptors containing 0/5-subunits since much greater amounts of mRNA encoding this subunit are present in the neonatal than in the adult brain 18.

Notwithstanding this global correlation some mis- matches are observed, however, between the distribu- tion of the different oJ binding sites and of the mRNA 15'1s'3°'31 encoding for the GABA A receptor to

which they are potentially associated. For example, much higher levels of the 0/l-subunit mRNA were detected in the cerebellum than in the cerebral cortex or in the olfactory bulb whereas the opposite was true for the distribution of o91 sites. Moreover, in several regions there were low, though detectable, levels of [3H]flumazenil binding resistant to 3 /xM zolpidem but no detectable hybridization signal for mRNA encoding the a 5 subunit. This is the case for the ventromedial hypothalamus which displays a 0)2L binding site density greater than that found in the cerebral cortex but no hybridization signal for the 0/5-subunit mRNA TM. The reasons for these discrepancies are as yet unknown but they suggest that [3H]flumazenil binding sites with high, intermediate and low affinity for zolpidem may also be associated with 0/-subunits other than those predicted on the basis of their affinity for transfected GABA A receptors. Alternatively, the failure to detect mRNA encoding the 0/5 subunit in several regions (spinal cord, striatum, nucleus accumbens) in which low but signifi- cant amount of (.O2L binding sites were quantified, may be due to the limited sensitivity of the hybridization technique. The regional mismatches observed may also result from the fact that mRNAs are exclusively lo- cated in cell bodies whereas ~0 sites are also present in the dendritic fields or on presynaptic terminals far remote from the soma. This may account for the obser- vation that in the hippocampus, the mRNA encoding for the 0/5 subunit is restricted to the pyramidal cell layer whereas WzL binding sites are also seen in the dendritic fields of these neurons.

The differential autoradiographic distribution of high, intermediate and low-affinity binding sites for zolpidem provides additional support for the existence of at least three native o) receptor populations. That

249

these binding site populations are not interconvertible conformations is indicated by the poor correlation that exists between the regional distribution of these sites (low vs. intermediate; r = 0.27; P = 0.60; low versus high; r = 0.21; P = 0.40; high vs. intermediate r = 0.32; P = 0.25).

The distinct anatomical distribution of these sites may provide some indications about their possible functional role. We have previously hypothetized 8 that the preferential distribution of w 2 sites in the hip- pocampus may explain the lack of impairment by zolpi- dem of the f9 rhythm and memory 9. Moreover, the low myorelaxant potential of zolpidem 25 may relate to the fact that in the spinal cord, o~ sites display low affinity for this imidazopyridine. That an important proportion of the o~ 2 sites in the hippocampus and spinal cord are akin to the W2L subtype suggests that the activation of these sites by benzodiazepine derivatives may play a role in the memory disrupting and myorelaxant effects of these molecules.

A substantial amount of in vivo [3H]flumazenil bind- ing to the rodent hippocampus and spinal cord is not displaced by zolpidem even at doses that give rise to low/xM concentrations of this compound in the brain 2. This may explain the lack of myorelaxant and memory disrupting effects of zolpidem and supports the sugges- tion that W2L sites are implicated in these unwanted effects of benzodiazepine derivatives.

High levels of ~02L are also found in the hypothala- mus, particularly in its ventromedial part, and it is tempting to suggest that the activation of these sites by non selective benzodiazepine derivatives could affect the hypothalamic control of food intake and the re- lease of hormones by the pituitary gland. The low affinity of zolpidem for these sites may explain why this compound, in contrast to benzodiazepine derivatives, does not increase food intake 32.

In conclusion, the present data demonstrate that the heterogeneity of central oJ sites observed in transfected GABA A receptors and in rat brain membranes can also be revealed by quantitative autoradiography in rat CNS tissue sections. The three w-binding site subtypes revealed in displacement studies of [3H]flumazenil binding by zolpidem and alpidem have a differential autoradiographic distribution which is consistent with their potential association with GABA A receptors con- taining the al-SUbunit (w 1 sites), the 0/2 or 0/3 subunits (Wzi sites) and 0/5 subunits (WzL sites).

Acknowledgments. D. Ruano was supported by a fellowship from the 'Junta de Andalucia'. We thank Dr. D. Graham for critical reading of this manuscript. The secretarial work of Mrs S. Leflond is also acknowledged.

250

R E F E R E N C E S

1 Arbilla, S., Allen, J., Wick, A. and Langer, S.Z., High affinity [3H] zolpidem binding in the rat brain: an imidazopyridine with agonist properties at central benzodiazepine receptors, Eur. J. Pharmacol., 130 (1986) 257-263.

2 Benavides, J., Peny, B., Dubois, A., Perrault, G., Morel, E., Zivkovic, B. and Scatton, B., In vivo interaction of zolpidem with central benzodiazepine (BZD) binding sites (as labeled by [3H] Ro 15-1788) in the mouse brain. Preferential affinity of zolpidem for the w I (BZD 1) subtype, J. Pharmacol. Exp. Ther., 245 (1988) 1033-1041.

3 Billard, W., Crosby, G., Iorio, L., Chipkin, R. and Barnett, A. Selective affinity of the benzodiazepines quazepam and 2- oxoquazepam for BZ~ binding site and demonstration [3H]2- oxoquazepam as BZ I selective radioligand, Life Sci., 42 (1988) 179-187.

4 Braestrup, C., Schmiechen, R., Neef, G., Nielsen, M. and Pe- tersen, E.N., Interaction of convulsive ligands with benzodi- azepine receptors, Science, 216 (1982) 1241-1243.

5 Braestrup, C. and Nielsen, M., [3H Propyl /3-carboline-3- carboxylate as a selective radioligand for the BZD~ benzodi- azepine receptor subclass, J. Neurochem., 37 (1981) 333-341.

6 Bureau, M. and Olsen R.W., Multiple distinct subunits of gamma aminobutyric acid A receptor protein show different ligand-bind- ing affinities, Mol. Pharmacol., 37 (1990) 497-502.

7 Corda, M.G., Giorgi, O., Longoni, B., Ongoni, E., Pesce, G., Cruciani, R., and Biggio, G., Functional coupling of GABA A receptors and benzodiazepine recognition site subtypes in the spinal cord of the rat, Eur. J. PharmacoL, 169 (1989) 205-213.

8 Dennis, T., Dubois, A., Benavides, J. and B. Scatton., Distribu- tion of central oJ~ (benzodiazepine a) and °)2 (benzodiazepine~) receptor subtypes in the monkey and human brain. An autoradio- graphic study with [3H] flunitrazepam and the w 1 selective ligand [3H] Zolpidem, J. Pharmacol. Exp. Ther., 247 (1988) 309-322.

9 Depoortere, H., Zivkovic, B., Lloyd, K.G., Sanger, D.J., Perrault, G., Langer, S.Z. and Bartholini, G. Zolpidem, a novel nonbenzo- diazepine hypnotic. I. Neuropharmacological and behavioral ef- fects, J. Pharmacol. Exp. Ther., 237 (1986) 649-658.

10 Jensen, L.H., Pertersen, E.N., Braestrup, C., Honore, T., Kehr, W., Stephens, D.N., Schneider, H., Seidelmann and Schmiechen, R., Evaluation of the /3 carboline ZK 93426 as a benzodiazepine receptor antagonist, Psychopharmacology, B (1984) 249-256.

11 Klepner, C.A., Lippa, A.S. Benson, D.I., Sano, M.C., and Beer, B., Resolution of two biochemically and pharmacologically dis- tinct benzodiazepine receptors, Pharmacol. Biochem. BehaL'., 11 (1979) 457-462.

12 Langer, S.Z. and Arbilla, S. Imidazopyridines as a tool for the characterization of benzodiazepine receptors: a proposal for a pharmacological classification as omega receptors, Pharmacol. Biochem. Behau., 29 (1988) 763-767.

13 Langer, S.Z., Arbilla, S., Benavides, J. and Scatton, B. Subtypes of ~o-receptors involved in the modulation of GABA A receptor function by the imidazopyridines zolpidem and alpidem, Fidia Research Foundation Symposium Series, Vol. 7, in press.

14 Langer, S., Arbilla, S., Tan, S., Lloyd, K.G., George, P., Allen, J. and Wick, A.E. Selectivity for omega-receptor subtypes as a strategy for the development of anxiolytic drugs, Pharmacopsychi- atry, 23 (1990) 102-107.

15 Laurie, D.J., Seeburg, P.H. and Wisden, W. The distribution of 13 GABA A receptor subunits mRNA in the rat brain. II. Olfac- tory bulb and cerebellum, J. Neurosci., 12 (1992) 1063-1076.

16 LiJddens, H., Killisch, I. and Seeburg, P.H., More than one alpha variant may exist in a GABA A/benzodiazepine receptor com- plex, J. Receptor Res., 11 (1991) 535-551.

17 Liiddens, H., Pritchett, D.B., K6hler, M., Killisch, I., Keinfinen, K., Monyer, H., Sprengel, R., and Seeburg, P.H., Cerebellar GABA A receptor selective for a behavioural alcohol antagonist, Nature, 346 (1990) 648-651.

18 MacLennan, A.J., Brecha, N., Khrestchatisky, M., Sternini, C., Tillakaratne, N.J.K., Chiang, M.Y., Anderson, K., Lai, M. and Tobin, J., Independent cellular and ontogenetic expression of mRNAs encoding three o~ polypeptides of the rat GABA A recep- tor, Neuroscience, 43 (1991) 369-380.

19 Macguire, P.A., Davies, F., Villar H.O., Loew, G.H., Evidence for more than two central benzodiazepine receptors in the spinal cord, Eur.J. Pharmacol., 214 (1992) 85-88.

20 Massoti, M., Schlichting, J.L., Antonacci, M.D., Giusti, P., Memo, M., Costa, E. and Guidotti, A., y-aminobutyric acid A receptor heterogeneity in rat central nervous system: studies with clon- azepam and other benzodiazepine ligands, J. Pharmacol. Exp. Ther., 256 (1991) 1154-1160.

21 Munson, P.J. and Rodbard, D., LIGAND: a versatile computer- ized approach for characterization of ligand-binding systems, Anal. Biochem., 107 (1980) 220-239.

22 Niddam, R., Dubois, A., Scatton, B., Arbilla, S. and Langer, S.Z., Autoradiographic localization of 3H-zolpidem binding sites in the rat central nervous system. Comparison with the distribution of 3H-flunitrazepam binding sites, J. Neurochem., 49 (1987) 890-899.

23 Olsen, R.W., McCabe, R.T. and Wamsley, J.K. GABA A receptor subtypes: autoradiographic comparison of GABA, benzodi- azepine, and convulsant binding sites in the rat central nervous system, J. Chem. Neuroanat., 3 (1990) 59-76.

24 Olsen, R.W. and Tobin, A.J., Molecular biology of GABA A receptors, FASEB J., 4 (1990) 1469-1480.

25 Perrault, G., Morel, E., Sanger, D.J. and Zivkovic, B., Differ- ences in pharmacological profiles of a new generation of benzodi- azepine and non-benzodiazepine hypnotics, Eur. J. Pharmacol., 187 (1990) 487-494.

26 Pritchett, D.B., Luddens, H. and Seeburg P.H., Type I and type lI GABAa-benzodiazepine receptors produced in transfected cells, Science, 245 (1989) 1389-1392.

27 Pritchett, D.B. and Seeburg, P.H., y-Aminobutyric acid A recep- tor a5-subunit creates novel type II benzodiazepine receptor pharmacology, J. Neurochem., 54 (1990) 1802-1804.

28 Ruano, D., Vizuete, M., Cano, J., Machado, D.A. and Vitorica, J., Heterogeneity in the allosteric interaction between the GABA A receptor and the benzodiazepine receptor in the rat nervous system, J. Neurochem., 58 (1992) 485-493.

29 Sieghart, W., Multiplicity of GABAA-benzodiazepine receptors, TIPS, 10 (1989) 407-411.

30 Wisden, W., Gundlach, A.L., Barnard, E.A., Seeburg, P.H. and Hunt, S.P., Distribution of GABA A receptor subunit mRNAs in rat lumbar spinal cord, Mol. Brain Res., 10 (1991) 179-183.

31 Wisden, W., Laurie, D.J., Monger, H. and Seerburg, P.H., The distribution of 13 GABA a receptor subunit mRNA in the rat brain. I. Telencephalon, diencephalon, mesencephalon, J. Neu- rosci., 12' (1992) 1040-1062.

32 Yerbury, R.E. and Cooper, S.J., Novel benzodiazepine receptor ligands: palatable food intake following zolpidem, CGS 17867A, or Ro 23-0364, in the rat, Pharmacol. Biochem. Behav., 33 (1989) 303-307.