Autoradiographic localization of adenosine uptake sites in rat brain using [3H]nitrobenzylthioinosine

The adenosine uptake site has been localized in rat brain by an in vitro light microscopic autoradiographic method, using [3H]nitrobenzylthioinosine ( [3H]NBI) as the probe. The binding characteristics of [3H]NBI on slide-mounted sections are comparable to those seen in studies performed on brain homogenates. A very high density of uptake sites occurs in the nucleus tractus solitarius, in the superficial layer of the superior colliculus, in several thalamic nuclei, and also in geniculate body nuclei. A high density of sites are also observed in the nucleus accumbens, the caudate putamen, the dorsal tegmentum area, the substantia nigra, and the central gray. The localization of the adenosine uptake site in brain may provide information on the functional activity of the site and suggests the involvement of the adenosine system in the central regulation of cardiovascular function.

There is a growing body of evidence to suggest that adenosine acts as a neuromodulator in the CNS (for review, see Daly, 1983). In common with various classical neurotransmitters, adenosine is stored within nerve terminals and is released in a calcium-dependent manner from these structures when depolarized. The released pool of adenosine acts at specific receptors, causing a change in intracellular cyclic AMP (Van Calker et al., 1979) and/or calcium levels (Ribeiro, 1979). Two distinct adenosine receptors coupled to the adenylate cyclase have recently been identified (Van Calker et al., 1979;Cooper et al., 1980). They are the A-l receptor, involved in the adenosinemediated inhibition of the cyclase activity, and the A-2 type, mediating the stimulation of adenylate cyclase. More recently, adenosine receptors have been identified by binding studies (Bruns et al., 1980;Pate1 et al., 1982). A specific adenosine receptor involved in the modulation of calcium fluxes has not yet been identified. different from the adenosine A-l and A-2 receptors (Marangos et al., 1982;Pate1 et al., 1982).
In the present report, using [3H]NBI, we have labeled adenosine uptake sites in rat brain slices and have visualized them by means of autoradiography.
In this manner we were able to demonstrate the precise localization of [3H]NBI-binding sites in the rat brain. A regional distribution comparison of [3H] NBI-binding sites and those of adenosine A-l receptor is also presented.
The autoradiographic procedure involves incubating slide-mounted tissue sections with 13H1NBI (16 Ci/mol: Moravek Biochemicals) to label the binding sites' anb appoiing t&&-sensitive film (LKB Uitrofilm). The details of this procedure have been reported previously (Herkenham and Pert, 1982). Male Sprague-Dawley rats (250 to 300 em) were decapitated, and the brains were removed quickly and frozen Like other classical neurotransmitters, a mechanism is present for the removal of adenosine from the synaptic cleft. This is achieved by reuptake involving an energy-dependent transport system (Shimizu et al., 1972;Kuroda and MacIlwain, 1974;Bender, 1980;Barberis et al., 1981). Deamination of adenosine by adenosine deaminase may also play a role in the inactivation of adenosine (Pull and MacIlwain, 1974;Arch and Newsholme, 1978). Previous studies have demonstrated that the adenosine uptake site can be labeled using [3H]nitrobenzylthioinosine ([3H]NBI), a potent inhibitor of adenosine uptake (Picard and Patterson, 1972;Jarvis and Young, 1980 and then placed in apposition to LKB Ultrofilm in a light-proof cassette (Wolf Picker). After 9 weeks of exposure, the films were developed with Kodak D-19 (4 min at 22"C), then rinsed in running water for 30 min, cleaned with deionized water, and dried at room temperature. The autoradiograms were screened and optical density measurements were performed using a computer image-processing system previously described (Goochee et al., 1980). The identification of the brain microregions was made with the aid of adjacent stained sections, using as reference the atlases of Paxinos and Watson (1982) and Koenig and Klippel (1963). For further comparison, one set of rat brain sections was incubated with ["Hlcyclohexyladenosine, using conditions previously described (Lewis et al., 1981).

Results
Biochemical properties of r3H]NBI binding to brain slices. ["H]NBI binding to rat brain sections is saturable (Fig. 1).

BOO 900
Nonspecific binding represents 30% of the total binding for ["HINBI concentrations below 1 nM. Scatchard analysis of the binding shows a dissociation constant (&) of 0.39 nM and a maximal number of sites (B,,,) of approximately 150 fmol/mg of protein (Fig. 1). These values are comparable to those observed in rat brain homogenates (Marangos et al., 1982). To ensure that the binding in brain sections involved the same site that had been characterized previously in brain membranes, we evaluated the displacement of [3H]NBI by NBI, NG-cyclohexyladenosine (CHA), a specific ligand of the adenosine receptor, and dipyridamole.
The If& values (data not shown) obtained were 2 nM for NBI, 6 PM for dipyridamole, and 15 PM for CHA, all of which are comparable to those observed with membrane preparations (Marangos et al., 1982). The binding to brain sections reached an equilibrium TABLE I Regional distribution of r3H]NBZ binding in rat brain Vol. 5, No. 2, Feb. 1985 Mean optical densities were measured by outlining each anatomical structure with the image processor; stained sections were used comparatively for precise localization. The optical density values are averages obtained for three to five measurements of the same structure on contiguous sections of each animal's brain. The value of the SEM was 50.01. The mean optical density of the film background was .08. For hippocampus and dentate gyrus the optical density reported is the mean value of all layers. after 20 min, and more than 70% of the initial specific binding remained after a 20-min wash in 500 ml of 50 mM Tris buffer (data not shown). Regional distribution of [3H]NBI-binding sites. The optical densities of the various brain areas measured are summarized in Table I. In the forebrain several structures show high NBI binding density, among which is the pyriform cortex. All other cortical areas show a very low density of binding sites with a more elevated level in layer IV, with the exception of cingulate cortex which shows a moderate to low density of sites (Fig. 2, B and D). Caudate putamen and nucleus accumbens show a homogeneous high density. In the septum only the dorsal part of the lateral septum displays a high density of binding sites (Fig. 3~). In the amygdala, moderate optical densities are observed in medialis and cortical nuclei. The hippocampus shows a very low density of sites (Fig. 20). In the midbrain, high densities are observed in the superficial layer of the superior colliculus, the central gray matter in the substantia nigra, the dorsal tegmentum area, where both nuclei parabrachialis superior and inferior and locus ceruleus show higher density than does the nucleus tegmentum dorsalis of Gudden (Figs. 3 and  4). In the hindbrain the localization of NBI-binding sites is very heterogenous. The nucleus tractus solitarius shows the highest density of sites in the whole brain. A very high density of sites is observed in the commissural part of that nucleus, the rostra1 part shows lower density. (Fig. 5, a and b). The substantia gelatinosa of the nucleus of the spinal tract of the trigemini nerve also shows a high density of NBI sites (Fig. 5b). Moderate levels of binding are observed in the superior olivary complex (Fig. 5b). The cerebellum has a low homogeneous level of sites (Fig. 2F)

Discussion
In this study we show that the binding of [3H]NBI to brain tissue sections has characteristics of the adenosine uptake site in brain homogenates (Marangos et al., 1982). The binding is saturable with kinetic constants compatible to those observed in homogenate preparations and shows comparable pharmacological properties. The autoradiographic study shows a heterogeneous distribution of adenosine uptake sites throughout the brain that is basically consistent with the regional distribution observed in brain homogenate studies (Marangos et al., 1982). In order to evaluate the meaning of this distribution, we have compared the localization of thz adenosine uptake site with the localization of the adenosine. A-l receptor (Table II). For this comparison we have used previous reports describing the anatomical distribution of adenosine receptor (Lewis et al., 1981;Goodman and Snyder, 1982), and we have performed an autoradiographic study on a set of sections adjacent to those used for [3H]NBI binding. The results of this comparison are summarized in Table I and illustrated in Figure 2. Some structures that show a very high density of both sites are the caudate putamen, nucleus accumbens, dorsal septum, pyriform cortex, substantia nigra, superficial layer of the superior colliculus, and substantia gelatinosa of the nucleus of the spinal tract of the trigeminal nerve. In the thalamus both the adenosine receptor and the adenosine uptake site are also present at high densities, but some differences exist in their relative distribution among the different thalamic nuclei. Higher adenosine receptor densities are observed in the anterior nuclei, gelatinosa, and medial nuclei. Higher levels of adenosine uptake sites are seen in paraventricular nucleus and the rhomboid and reuniens nuclei (see Fig. 2). In contrast, some structures which possess a very high density of adenosine receptors show low amounts of aden- osine uptake sites. In the cerebellum and the hippocampus, where adenosine receptors are present at high densities on specific layers, the uptake sites show homogeneous low levels without differences between the different layers (see Fig. 2).
The same disparity is observed in the amygdala complex, where only the central nucleus possesses adenosine receptors, but the adenosine uptake site is present in high amounts in both the cortical and medial nuclei. The relative density of both the receptor and uptake site could represent one index of functional activity of the adenosine system, high activity being associated with high number of adenosine receptor and uptake sites. In light of this, an active adenosinergic system might be expected in caudate putamen, nucleus accumbens, pyriform cortex, septum, substantia nigra, superior colliculus (superficial layer), and substantia gelatinosa of the trigeminal nerve. In contrast, cerebellum and hippocampus, although rich in the adenosine receptor, might display a lower level of adenosinergic function. Further physiological studies are of course required to substantiate these speculations. A third group of structures containing a high density of the adenosine uptake site with very low levels of adenosine receptor also exists. In this group are found the nucleus of the tractus  solitarius, the central gray nuclei, the hypothalamus (in its It is also possible that the [3H]NBI-binding sites seen in dorsomedial part), and the dorsal tegmentum. The uptake site some brain areas function to take up other nucleosides, since in these areas could be related to the presence of an adenosine receptor not susceptible to [3H]CHA binding. Since [3H]CHA this has been shown to be possible (Hammond and Clanachan, probably binds selectively to A-l adenosine receptors, it is 1983). However, it has been shown that the adenosine uptake possible that these areas contain A-2 receptors. Previously, we site does favor adenosine (Bender, 1980). have shown (Marangos et al., 1983) that the adenosine antag-The presence of a very high density of adenosine uptake sites onist [3H]diethylphenylxanthine has binding properties differ-in the nucleus of the tractus solitarius could be related to some ent from [3H]CHA has, with a higher number of sites (possibly central effect of adenosine. This nucleus, particularly its comreflecting its ability to interact with both A-l and A-2 recep-missural part, represents the main central control structure of tors). It is, therefore, possible that [3H]CHA is not labeling all blood pressure. The closely related nucleus parasolitarius is adenosine receptors in brain.
also involved in the control of respiration (Palkovits and Za-   et al., 1982), coupled with the high density of the uptake site in the nucleus tractus solitarius, suggests that the hypotensive effects of adenosine are in part centrally mediated. This is also supported by the observation that phenylsulfotheophylline, an adenosine antagonist that does not cross the blood-brain barrier, does not fully reverse the hypotensive effect of adenosine agonists (unpublished observations). It has been suggested recently that benzodiazepines (BZs) may act at the adenosine uptake site (Phillis et al., 1981;Hammond and Clanachan, 1982). The distributions of the BZ receptor (Young and Kuhar, 1980) and the adenosine uptake site show some similarities.
Both sites are present in important amounts in certain structures such as the pyriform cortex, superior colliculus, central gray, substantia nigra, and the dorsal tegmentum, substantia gelatinosa of the trigeminal nerve. However, there are many discrepancies between the localization of BZ receptors and adenosine uptake sites. For instance, the cerebellum and the hippocampus show high densities of BZ receptors and very low density of adenosine uptake sites. The high densities of BZ receptor observed in frontal cortex are also not matched by adenosine uptake sites. In the septum, BZ receptors are concentrated in the nucleus medialis, whereas the uptake site is present only in the dorsal part of the septum.
The highest density of adenosine uptake site is in the pars commissuralis of the nucleus tractus solitarius. BZ receptors are only present in very low density in that area.
The adenosine uptake site is, therefore, highly localized in rodent brain with a regional distribution that is different from that of the adenosine receptor labeled by [3H]CHA and the BZ receptor. It appears, therefore, that certain brain nuclei have a greater density of adenosine uptake sites, and it is possible that these areas may represent brain regions which contain adenosinergic neurons.