Introduction

Top Abstract Introduction Methods Results Discussion References

In keeping with the innervation by 5-HT of particular anxiety-related regions (e.g., hippocampus, amygdala) and its closed links with GABAergic systems, it has been initially suggested that central 5-HT plays a key role in the etiology of anxiety (Iversen, 1984). In confirmation, early experiments, mainly conducted by means of nerve lesions or 5-HT synthesis inhibition, have suggested an anxiogenic influence of 5-HT activity (Chopin and Briley, 1987; Handley, 1995). More recently, the recognition of different 5-HT receptor subtypes, and in turn, the progressive availability of more or less selective agonists/antagonists for each of these 5-HT receptor subtypes, has allowed analyses of their role in anxiety (Griebel, 1995; Handley, 1995). Actually, pharmacological analyses using ethological (unconditioned), but not conflict (conditioned), procedures have provided results consistent with the hypothesis that 5-HT activity is associated with anxiogenesis (Griebel, 1995; Handley, 1995).

Among the 5-HT receptor subtypes analyzed so far, the so-called 5-HT_1A receptor has received much attention (De Vry, 1995) due to (1) the early recognition of selective agonists for this receptor, including 8-OH-DPAT and the azapirones ipsapirone and buspirone, (2) the identification of 5-HT_1A receptors in critical regions such as the hippocampus but also in the midbrain raphe nuclei, where they act as autoreceptors that control negatively 5-HT nerve firing and synthesis/release of 5-HT and (3) the observation that azapirones were effective both in conditioned and unconditioned animal models of anxiety and in human studies. There is substantial evidence for the key role of 5-HT_1A autoreceptors in mediating the anxiolytic effects of azapirones and 8-OH-DPAT; on the other hand, the extent to which postsynaptic (e.g., hippocampal) 5-HT_1A receptors participate in the anxiolytic effects of 5-HT_1A receptor agonists remains a matter of controversy (De Vry, 1995).

As illustrated above, pharmacological tools help in identifying the serotonergic mechanisms involved in anxiety processes. Thus, besides 5-HT_1A receptors, the so-called 5-HT_2A, 5-HT_2B/2C and 5-HT₃ receptors have also been claimed to play a role in anxiogenesis because the blockade of one or another of these receptor subtypes often, but not always, either reduces the anxiety-related behaviors in numerous animal models or diminishes human anxiety (Griebel, 1995; Handley, 1995). Except for human studies in which the therapeutic properties of different serotonergic agonists/antagonists may truly be assessed, it is, however, noteworthy that a compound is often primarily defined as anxiogenic/anxiolytic on the basis of its effects in animal models of anxiety. Although this parallel may prove to be correct (however, see Treit, 1985), the need for additional models allowing the recognition of the mechanisms, including the serotonergic ones, involved in anxiety is obvious both to allow further validation of the hypotheses tested and to raise new hypotheses that may then prove to be relevant to human disorders.

Results of recent human studies have shown that personality is affected by both environmental factors and the genetic status of the individual (Bouchard, 1994), with the respective impacts of each of these factors being characteristic of a given individual. Actually, this statement holds true for rodents, as illustrated by the genetic analysis of mouse behaviors in an unconditioned model of anxiety, such as the elevated plus-maze (Flint et al., 1995).

In keeping with the latter observation, we recently conducted a thorough investigation of anxiety-related behaviors in six inbred rat strains on placement in novel environments, including the elevated plus-maze, the light/dark box and the social interaction test (i.e., so-called unconditioned animal models of anxiety) (Ramos et al., 1997). As far as the elevated plus-maze is concerned, our analysis allowed the recognition of a highly discriminant pair of strains, the LEW strain and the SHR strain, characterized by high and low anxiety scores, respectively (Ramos et al., 1997). In keeping with this observation, we conducted the present series of experiments to analyze whether this genetic difference between SHR and LEW rats was associated with strain-related differences in central serotonergic systems. To this end, male and female SHR and LEW rats were compared in vitro for the activity of the rate-limiting enzyme in 5-HT biosynthesis (i.e., tryptophan hydroxylase) (Boadle-Biber, 1982; Hamon et al., 1981) and ex vivo (in males only) for 5-HT synthesis. In addition, both [³H]8-OH-DPAT binding kinetics at hippocampal and midbrain 5-HT_1A receptors (Gozlan et al., 1983) and [³H]ketanserin binding kinetics at cortical and striatal 5-HT_2A receptors (Leysen et al., 1982) were determined. Because 5-HT reuptake sites play a key role in the regulation of synaptic levels of 5-HT (Blier and De Montigny, 1994) and SSRIs (e.g., citalopram) promote anxiety on acute administration but anxiolysis on repeated treatment (Cadogan et al., 1992; Handley and McBlane, 1993; Nutt, 1995), [³H]citalopram binding at 5-HT reuptake sites (D'Amato et al., 1987) was studied in the midbrain (i.e., on cell bodies/dendrites) and in the hippocampus (i.e., at nerve terminals).

Actually, because chronic administration of SSRIs reduces different 5-HT_2B/2C receptor-mediated events, such as hypolocomotion (Kennett et al., 1994; Maj and Moryl, 1992), a change thought to contribute to their anxiolytic effects, the hypolocomotor effect of acute 5-HT_2B/2C receptor stimulation (Kennett and Curzon, 1988; Lucki et al., 1989) was also assessed in males. Last, male SHR and LEW rats were compared for their 5-HT_1A autoreceptor functions (as assessed on the basis of 8-OH-DPAT-induced inhibition of tryptophan hydroxylase activity) (Hamon et al., 1984; Hjorth and Magnusson, 1988) but also for their postsynaptic 5-HT_1A receptor and 5-HT_2A receptor functions, as assessed on the basis of 8-OH-DPAT-elicited forepaw treading and flat body posture (Tricklebank et al., 1984) and the number of head shakes promoted by the stimulation of 5-HT_2A receptors (Schreiber et al., 1995), respectively.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals. Male and female SHR and LEW rats (IFFA CREDO, Les Oncins, France; 5 weeks old on arrival) were housed four to a cage (one strain/one sex per cage) under a constant temperature (22 ± 1°C) and a 12-hr/12-hr light/dark cycle (lights on, 7:00 a.m.). Food and water were available ad libitum. All rats were tested at 3 to 4 weeks after their arrival. Except for the rats that underwent an initial elevated plus-maze test, all rats were used only once. Males and females were used for the biochemical studies (radioligand binding, in vitro tryptophan hydroxylase), whereas only males were used for the determination of 5-HT and 5-HIAA levels and for the functional studies (ex vivo 5-HT synthesis, behaviors). In these functional studies, body weights of male SHR ranged from 230 to 270 g, whereas LEW rats weighed 250 to 300 g.

Elevated plus-maze tests. One week before their final tests, some of the rats were randomly removed from their home cages and tested for 5 min in an elevated plus-maze to ensure the persistence of strain-related differences in anxiety levels.

Tissue preparations. Rats were killed, and their frontal cortex, striatum, hippocampus and midbrain were rapidly dissected out on an ice-cold plate. All the structures were immediately plunged in dry ice and stored thereafter at 80°C until analyses. Except for the midbrains, which were individually assayed, two structures were pooled (same strain, same sex) per assay. Membrane (radioligand binding assays) and cytosol (in vitro tryptophan hydroxylase assays) preparations were obtained as previously described (Kulikov et al., 1995). Briefly, the different structures were homogenized in ice-cold Tris-acetate buffer, pH 7.6, containing 2 mM dithiotreitol and centrifuged (12,000 × g for 20 min at 4°C). The resulting supernatants were immediately stored at 80°C for subsequent tryptophan hydroxylase activity. Cortical, striatal and part of midbrain and hippocampal pellets to be used for 5-HT_1A and 5-HT_2A receptor binding assays were suspended in 40 vol of cold Tris·HCl, pH 7.7, whereas part of midbrain and hippocampal pellets to be used for 5-HT transporter binding assays were suspended in 40 vol of cold Tris·HCl, pH 7.4. The samples were then homogenized and centrifuged (15,000 × g for 10 min at 4°C). The resulting pellets were resuspended in 40 vol of their respective buffers, homogenized and incubated at 35°C for 15 min. Thereafter, samples were centrifuged (15,000 × g for 10 min at 4°C), and the resulting pellets were stored at 80°C until radioligand binding analyses. All protein concentrations were estimated using bovine serum albumin as standard (Bradford, 1976).

In vitro tryptophan hydroxylase activity. Tryptophan hydroxylase activity was performed as previously described (Kulikov et al., 1995). The initial supernatants (see above) were mixed (1.5:2.5 v/v) with 50 mM Tris-acetate buffer, pH 7.6, containing 1 mM dithiotreitol, 50 units of catalase, 1 mM NSD 1015, 0.025 to 0.8 mM L-tryptophan, and 0.3 mM 6-methyl-5,6,7,8-tetrahydropterin (Sigma-Coger, Paris, France). After 15 min of incubation at 35°C, the reaction was stopped by the addition of one-fourth volume of trichloracetic acid (50%). The samples were centrifuged at 10,000 × g for 20 min; the resulting supernatants were diluted in 1% trichloracetic acid; and the 6-methyl-5,6,7,8-tetrahydropterin-protected solutions were kept at 20°C until analysis (2-3 days later) of 5-HTP (the reaction product of tryptophan hydroxylase).

Radioligand binding analyses. All radioligand binding analyses were performed as previously described (Kulikov et al., 1995), except that 10 µM 5-HT (and pargyline in the buffer) were used (instead of 10 µM bufotenin) for the estimation of [³H]8-OH-DPAT nonspecific binding. For [³H]8-OH-DPAT and [³H]ketanserin binding analyses, the pellets (see above) were suspended in 40 volumes of a 50 mM Tris·HCl buffer, pH 7.7, either containing ([³H]8-OH-DPAT binding) or not ([³H]ketanserin binding) containing 5 mM CaCl₂ and 0.1% ascorbic acid. The suspension was transferred to glass tubes (total volume, 0.5 ml), and the reaction was carried out for 15 min at 35°C in the presence of six concentrations (0.25-4 nM) of either [³H]8-OH-DPAT (154 Ci/mmol) or [³H]ketanserin (85 Ci/mmol). Nonspecific binding was carried out in the presence of 10 µM 5-HT (Sigma-Coger) and 10 µM methysergide (Sandoz, Paris, France), respectively. For [³H]citalopram binding to 5-HT transporters, the pellets were suspended in 40 volumes of a 50 mM Tris·HCl buffer, pH 7.4, containing 120 mM NaCl and 5 mM KCl. The suspension was transferred to glass tubes (total volume, 0.25 ml), and the reaction was carried out for 60 min at 25°C in the presence of six concentrations (0.5-16 nM) of [³H]citalopram (81 Ci/mmol) with and without 1 µM paroxetine (SmithKline & Beecham, Harlow, England). Tritiated ligands were all purchased from DuPont-NEN (Les Ulis, France). Reactions were stopped by the addition of 4 ml of the respective buffer, followed by a rapid filtration through Whatman GF/B glass-fiber filters. The filters were washed twice with 4 ml of the buffer, and radioactivity was measured by liquid scintillation. All samples were assayed in duplicate, and the data were analyzed by means of Scatchard plots.

5-HT_1A autoreceptor-mediated inhibition of 5-HTP synthesis. Male SHR and LEW rats were weighed and administered subcutaneous 1 ml/kg injections with 0.9% NaCl or 8-OH-DPAT (125 or 250 µg/kg; RBI-BioBlock, Illkirch, France) and returned to their home cages. Thirty minutes later, all the rats were injected with NSD 1015 (100 mg/kg i.p.; 1 ml/kg) and returned to their home cages for another 30-min period (Hjorth and Magnusson, 1988). Thereafter, rats were killed, and midbrains and hippocampi were collected and stored at 80°C until analysis. Samples were then sonicated in 0.4 N perchloric acid (containing mercaptoethanol 1:50 v/v) and centrifuged (15,000 × g for 10 min). Each supernatant was then analyzed for its 5-HTP contents through the use of high performance liquid chromatography coupled to an electrochemical detection, as reported above.

Midbrain and hippocampus 5-HT and 5-HIAA levels. Male SHR and LEW rats were removed from their home cages and killed, and their midbrains and hippocampi were rapidly dissected and stored at 80°C until analysis. Samples were prepared and analyzed as reported above, except that supernatant 5-HT and 5-HIAA levels were determined instead of 5-HTP levels.

Analyses of 5-HT-related behaviors. Male SHR and LEW rats were transferred from their home cages to individual cages made of transparent Perspex (24 × 24 × 30 cm) for subsequent analyses of 5-HT_1A, 5-HT_2A and 5-HT_2B/2C receptor-related behaviors, as previously described (Zamfir et al., 1992). In a first series of experiments, 3 SHR and 3 LEW rats were assigned codes to allow blind experiments and injected subcutaneously with the 5-HT_1A receptor agonist 8-OH-DPAT (0.5 or 1 mg/kg) at 10 to 15 min after their placement. Observation sessions of 30 sec began 5 min later and were repeated every 3 min over a 15-min period. Forepaw treading and flat body posture were scored using a four-point ranked intensity scale (0 indicates absent and 3 indicates intense), and the scores were summed over the five observation periods. This protocol was then repeated twice with naive rats to achieve groups of 9 animals per strain per dose of 8-OH-DPAT. In a second series of experiments, 1 SHR and 1 LEW rat were injected intraperitoneally (10-15 min after their placement) with DOI (0.5-1 mg/kg; RBI-BioBlock, Illkirch, France) or quipazine (5 mg/kg; RBI-BioBlock). Thereafter, the number of head shakes was counted for four successive 5-min periods. This series of experiments included 18 subseries to achieve groups of 6 rats per strain per treatment.

Statistical analysis. All data are expressed as mean ± S.E.M. Radioligand binding and tryptophan hydroxylase data were analyzed by means of linear regressions using the least-squares method and compared by two-way analyses of variance followed, if significant, by Tukey's multiple comparison tests. Except for the head shakes and mCPP-induced hypolocomotion, which were respectively compared by means of two- and three-way analyses of variance (followed by Tukey's tests), all behavioral data were analyzed by Kruskal-Wallis tests, followed, if significant, by Mann-Whitney U tests.

	Results

Top Abstract Introduction Methods Results Discussion References

Elevated plus-maze behaviors of SHR and LEW rats. In a first series of experiments involving SHR and LEW rats of both sexes (to be used thereafter for biochemical analyses), it was found that the four rat groups differed in their percentage of number of open arm entries (P < .0001) and the percentage of time spent therein (P < .0001) but not in the number of closed arm entries (fig. 1). Furthermore, total (open plus closed) arm entries did not differ significantly between groups (10.5 ± 1.1 and 13.1 ± 1.5 in male and female SHR, respectively; 9.6 ± 1.4 and 12.1 ± 1.4 in male and female LEW rats, respectively). Post hoc tests revealed that the first two variables were higher in SHR than in LEW rats (fig. 1). Although females entered more frequently (and for a longer duration) in the open arms than did their male counterparts, this difference proved to be significant in SHR only (fig. 1).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Behaviors of male and female SHR and LEW rats placed for 5 min in an elevated plus-maze. Values are mean ± S.E.M. of 8 rats. ***P < .001 for the difference between SHR and LEW rats within each sex group. +P < .05 for the difference between males and females within each strain.

In vitro tryptophan hydroxylase activity in SHR and LEW rats. Table 1 depicts the respective strain and sex influences on tryptophan hydroxylase activity in serotonergic cell bodies (midbrain) and serotonergic nerve terminals. Strain effects on V_max could be observed in midbrain (P = .0144), hippocampus (P = .0115), frontal cortex (P = .0143) and striatum (P = .0038), with SHR displaying a lower tryptophan hydroxylase activity than LEW rats. These changes were associated with strain-related differences in tryptophan hydroxylase K_M values in midbrain (P = .0418) and in striatum (P = .0056), with SHR displaying lower K_M values than LEW rats (table 1). In midbrain and hippocampus, neither the sex factor nor the strain × sex interaction had a significant impact on tryptophan hydroxylase V_max and K_M values; on the other hand, in frontal cortex, sex had an influence on V_max (P = .0287) and K_M (P = .0019), with females displaying higher enzyme activity and affinity (for its substrate) than males (table 1). Last, in striatum, the V_max (P = .0463) and K_M (P = .0053) values were affected in opposite ways by the sex factor, with males displaying higher activity and affinity than their female counterparts (table 1).

Midbrain and hippocampal [³H]8-OH-DPAT binding in SHR and LEW rats. In midbrain, in which serotonergic cell bodies are located, the binding of [³H]8-OH-DPAT to 5-HT_1A autoreceptors was affected by neither the rat strain nor the sex (fig. 2). This was also true for the respective K_D values, which varied from 1.40 ± 0.22 to 1.57 ± 0.26 nM (data not shown). On the other hand, in hippocampus, [³H]8-OH-DPAT binding at 5-HT_1A postsynaptic receptors was influenced by the sex (P = .0175), but not the strain, of the animals; therefore, females displayed higher binding than males (fig. 2). Last, K_D values (which varied between 1.80 ± 0.23 and 2.01 ± 0.34 nM) proved to be resistant to both strain and sex factors.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Midbrain and hippocampal Bmax values of [3H]8-OH-DPAT (specific) binding at 5-HT1A receptors in male and female SHR and LEW rats. Values are the mean ± S.E.M. of 6 or 7 independent experiments. +P < .05 for the difference between males and females within each strain.

[³H]Ketanserin and [³H]citalopram binding in SHR and LEW rats. Although trends toward decreases could be observed in LEW rats compared with SHR and in females compared with males, neither the strain nor the sex significantly affected the B_max values of [³H]ketanserin binding in frontal cortex and striatum (table 2). Regarding [³H]citalopram binding at 5-HT transporters, there was a strain effect (P = .0183) in midbrain, with SHR displaying slightly lower B_max values than LEW rats (table 2). Besides, neither midbrain K_D values nor B_max and K_D values of [³H]citalopram binding in the hippocampus proved to be sensitive to the strain or sex factors (table 2).

Midbrain and hippocampus 5-HT and 5-HIAA levels in SHR and LEW rats. In midbrain, neither 5-HT levels (3.68 ± 0.15 and 4.19 ± 0.28 nmol/g in 4 SHR and 4 LEW rats) nor 5-HIAA levels (3.07 ± 0.32 and 3.09 ± 0.21 nmol/g in 4 SHR and 4 LEW rats) differed between strains. Confirmingly, in hippocampus, there was no strain-related difference regarding 5-HT (2.02 ± 0.13 and 2.02 ± 0.2 nmol/g in 4 SHR and 4 LEW rats) and 5-HIAA (1.91 ± 0.22 and 1.89 ± 0.12 nmol/g in 4 SHR and 4 LEW rats).

Midbrain 5-HT_1A (auto)receptor sensitivity in SHR and LEW rats. To investigate whether SHR and LEW rats differed in the function of their 5-HT_1A autoreceptors (independently from changes in the number and affinity of these receptors), we analyzed the inhibitory influence of 5-HT_1A autoreceptor stimulation on 5-HT synthesis (as assessed by 5-HTP accumulation) in serotonergic cell bodies and nerve terminals. As shown in figure 3, 8-OH-DPAT dose-dependently (P < .0001) decreased NSD 1015-elicited 5-HTP accumulation in midbrain and hippocampus; on the other hand, neither baseline 5-HTP accumulation nor the inhibitory effect of 8-OH-DPAT on 5-HTP accumulation proved to be different between strains.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Midbrain and hippocampal 5-HTP accumulation in male SHR and LEW rats pretreated either with saline or 8-OH-DPAT. Values are mean ± S.E.M. of 6 or 7 rats. ++P < .01 for the difference between 8-OH-DPAT and saline within each strain.

5-HT_1A, 5-HT_2A and 5-HT_2B/2C receptor-mediated behaviors in SHR and LEW rats. After a significant Mann-Whitney analysis (P < .0001), it was observed that the intensity of 8-OH-DPAT-elicited forepaw treading was (1) dose- dependent (P = .0012 and P = .0315 for the difference between the two doses of 8-OH-DPAT in SHR and LEW rats, respectively) and (2) strain-dependent (SHR > LEW); this reached significance for the highest dose of 8-OH-DPAT used (fig. 4). With regard to 8-OH-DPAT-elicited flat body posture, groups were found to differ (P < .0001), but such a difference was accounted for by strain- (LEW > SHR), but not dose-related, changes (fig. 4).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Forepaw treading and flat body posture responses to 8-OH-DPAT administration in male SHR and LEW rats. Values are mean ± S.E.M. of 9 rats. *P < .05 and **P < .01 for the difference between SHR and LEW rats. For clarity, statistical differences between 0.5 and 1 mg/kg doses of 8-OH-DPAT are not indicated.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Head shake responses to DOI or quipazine administration in male SHR and LEW rats. Values are mean ± S.E.M. of 6 rats. *P < .05 and **P < .01 for the differences between SHR and LEW rats.

	Discussion

Top Abstract Introduction Methods Results Discussion References

The observation that SHR and LEW rats, whether males or females, behaved differently in the elevated plus-maze is in agreement with our previous studies (Ramos et al., 1997). Thus, on the basis of the respective scores regarding the percent number of visits to and the percentage of time spent in open arms (these two ratios providing indices of anxiety as assessed by the respective effects of anxiogenic and anxiolytic drugs: Pellow et al., 1985), we proposed that LEW rats are more anxious than SHR. Indeed, this difference cannot be accounted for by strain-related differences in early experience because recent experiments using SHR and LEW rats bred in our laboratory confirmed that SHR are less anxious than LEW rats¹ (in preparation).

On the other hand, the lack of a genetic difference in the number of closed arm entries (i.e., an index of activity) (File, 1991; Ramos et al., 1997) confirms that the two strains do not differ in locomotor activity (Ramos et al., 1997). This was confirmed in the mCPP-related experiments (see below), in which it was found that strains did not differ regarding locomotion during the first 5 min of testing. In keeping with these results, it is noteworthy that the aforementioned genetic difference in the elevated plus-maze extends to other anxiety indices (i.e., the number of transitions in the light/dark box and open-field inner locomotion) (Ramos et al., 1997). However, in another anxiety model (i.e., the social interaction test) (File, 1980), SHR and LEW rats do not differ, thereby confirming that different anxiety tests may capture different anxiety dimensions (File, 1991; Ramos et al., 1997). It is worthy of mention that among other behavioral differences, LEW rats may spend more time immobile than SHR on a single exposure to a forced swimming test (Lahmane and Armario, 1996).

Because SHR and LEW rats differ in numerous physiological aspects, such as hypertension in SHR and susceptibility to inflammatory agents in LEW rats, links between these physiological differences and the respective strain-related differences in anxiety behaviors could be proposed. Actually, we recently observed in 200 F2 rats bred from initial SHR × LEW and LEW × SHR crosses that systolic blood pressure levels did not correlate with any of the variables measured in the elevated plus-maze (Ramos et al.¹). Whether there is a link between the immune status of the rats and their anxiety scores is, however, still unknown.

As stated in the introduction, the study of rat strains differing in their respective anxiety-related behaviors is a highly relevant approach to validate previous pharmacological hypotheses or identify new mechanisms. In keeping with the former, we began to investigate whether SHR and LEW rats differ in several components of their central serotonergic systems (i.e., systems that play a key role in the etiology of anxiety) (Griebel, 1995; Handley, 1995).

Studies involving lesions of central serotonergic neurons or 5-HT synthesis inhibition have reported positive effects of these procedures on anxiety levels (at least as assessed through unconditioned models), that is, anxiolysis (Griebel, 1995; Handley, 1995). In keeping with this overview and the reports that repeated stressors, but not all, may increase 5-HT synthesis through tryptophan hydroxylase hyperactivity (Chaouloff, 1993), we first investigated whether the two strains differed in their 5-HT synthesis rates. To this end, we first determined the in vitro activity of the rate-limiting enzyme in 5-HT biosynthesis (i.e., tryptophan hydroxylase) in serotonergic cell bodies (midbrain) and nerve terminals (hippocampus, frontal cortex, striatum). In most regions analyzed, LEW rats displayed increased in vitro tryptophan hydroxylase activity, as assessed on the basis of V_max values, thus suggesting that 5-HT synthesis has a greater capacity to be increased in LEW rats compared with SHR. In addition, within each strain, there were regional differences in tryptophan hydroxylase characteristics (V_max and/or K_M), thus reflecting probable differences in gene expression and/or the ratio of the active to inactive (phosphorylated to nonphosphorylated) forms of the enzyme. Besides, tryptophan hydroxylase activity, as assessed in vitro, showed discrete (albeit nonuniform) increases in females compared with males.

Although this sex-related difference was not investigated ex vivo, our results are in line with past findings showing that 5-HT synthesis/metabolism, as assessed ex vivo, is higher in females than in males (Carlsson and Carlsson, 1988). Taken together, our in vitro analyses of tryptophan hydroxylase activity could suggest strain-related differences. However, it is noteworthy that in most cases, strain-related differences in tryptophan hydroxylase activity were associated with concomitant decreases in the affinity of the enzyme toward its substrate. Actually, in the ex vivo experiment, it was found that regardless of the region analyzed, neither 5-HTP accumulation nor 5-HT and 5-HIAA levels differed between male SHR and LEW control rats. Confirmingly, in a recent ex vivo study aimed at analyzing 5-HT_1A autoreceptor-mediated inhibition of 5-HT synthesis in fed and fasted SHR and LEW rats, NSD 1015-elicited 5-HTP accumulation in saline-injected fed and fasted rats did not differ between strains (Chaouloff et al., 1997). It is relevant to add that in the latter study, midbrain levels of tryptophan (the precursor of 5-HTP) were significantly increased by 13% in fed SHR compared with fed LEW rats (Chaouloff et al., 1997).

These results thus indicate that in vivo, 5-HT synthesis/metabolism is not different between strains, although the respective importance of tryptophan levels on the one hand and tryptophan hydroxylase activity on the other hand may differ between strains (see above). The lack of strain-related differences on ex vivo 5-HT synthesis, as revealed by NSD 1015-induced 5-HTP accumulation, further suggests that basal release of 5-HT, including that in critical regions such as the hippocampus, may not differ between SHR and LEW rats. Besides this suggestion, one important question that remains unanswered concerns putative strain-related changes in the amount of 5-HT released during acute stimuli, such as during exposure to the elevated plus-maze (File et al., 1993).

As indicated above, there is experimental and clinical evidence for a crucial role of 5-HT_1A presynaptic and, possibly, postsynaptic receptors in the control of anxiety processes (De Vry, 1995). Thus, local administration of 5-HT_1A receptor agonists in the raphe nuclei (in which 5-HT_1A autoreceptors are located) has been repeatedly found to elicit anxiolysis (De Vry, 1995). This behavioral response is thought to be due to the inhibitory effects of 5-HT_1A autoreceptor stimulation on (1) serotonergic nerve firing (Sprouse and Aghajanian, 1987) and, in turn, (2) 5-HT (synthesis and) release at nerve terminals (e.g., cortex, hippocampus) (Hutson et al., 1989). Herein, neither radioligand binding at 5-HT_1A receptors in midbrain (which contains the different raphe nuclei) nor the ability of these receptors to exert their inhibitory influence on tryptophan hydroxylase activity (Hamon et al., 1984; Hjorth and Magnusson, 1988), as demonstrated by 8-OH-DPAT-induced inhibition of 5-HTP accumulation, differed between male SHR and LEW rats. One could argue that 8-OH-DPAT may have had differential strain-related effects on tryptophan levels and tryptophan hydroxylase activity (see above), thus impeding any conclusion. Actually, this is unlikely because 8-OH-DPAT administration does not affect brain tryptophan levels in Wistar rats (Chaouloff et al., 1992) or in SHR and LEW rats (Chaouloff et al., 1997). Taken together, our data thus suggest that the above-mentioned strain-related behavioral difference in the elevated plus-maze is independent of changes at the 5-HT_1A autoreceptor level, although one cannot dismiss that the present study did not address the possibility of strain-related differences in the number and/or the morphology of midbrain raphe nuclei.

The role of hippocampal (postsynaptic) 5-HT_1A receptors in anxiety processes has been the subject of intense investigation, but a clearcut appreciation of this role is still lacking (De Vry, 1995). In keeping with the observation that azapirones, which are anxiolytic, are partial agonists for the 5-HT_1A receptor, it has been initially proposed that the anxiolytic effect of 5-HT_1A receptor agonists involved stimulation of 5-HT_1A autoreceptors and concomitant blockade of postsynaptic 5-HT_1A receptors (Traber and Glaser, 1987). Confirmingly, acute intrahippocampal injection of 5-HT_1A receptor agonists has been shown to elicit anxiety; however, such a procedure may also promote anxiolysis or be without significant effects, depending on the anxiety model, the doses used, and possible drug diffusion to other key regions (De Vry, 1995). On repeated administration of 5-HT_1A receptor agonists endowed with clinical albeit delayed anxiolytic profiles, some studies, but not all, reported down-regulation of hippocampal 5-HT_1A receptors and/or desensitization of their functional responses, thus giving support to the hypothesis that stimulation of hippocampal 5-HT_1A receptors may have anxiogenic consequences (De Vry, 1995). Herein, the SHR and LEW strains did not differ regarding [³H]8-OH-DPAT binding at hippocampal 5-HT_1A receptors, whereas at the opposite sex had an influence, with females displaying higher B_max values than males.

This last observation is in line with past autoradiographic studies in Sprague-Dawley rats showing higher retention of [³H]8-OH-DPAT in the hippocampal CA1 formation of females compared with males (Mendelson and McEwen, 1991). Interestingly, male LEW rats tended to display a higher B_max value than SHR, a trend that could suggest some discrete differences between strains that could have consequences on anxiety levels. Actually, the observation that female SHR and LEW rats displayed equal B_max values but strikingly different levels of anxiety strongly suggests that the strain-related difference in anxiety levels is independent of the number of [³H]8-OH-DPAT binding sites in hippocampus. However, this does not exclude a possible strain-related difference in the function of 5-HT_1A receptors (e.g., through a quantitatively different coupling between the binding site and its second messengerinhibition of adenylate cyclase activity). To examine this crucial hypothesis, work aimed at determining the effects of 8-OH-DPAT on forskolin-stimulated adenylate cyclase activity in the hippocampi of SHR and LEW rats are in progress in our laboratory.

Among other tools to assess the functional status of 5-HT_1A postsynaptic receptors, 8-OH-DPAT-elicited forepaw treading and flat body posture have been frequently used. Repeated administration of 8-OH-DPAT has been shown to decrease the intensity of these behavioral responses to an acute 8-OH-DPAT challenge (De Vry, 1995), thereby indicating that this paradigm may provide indices on 5-HT_1A receptor desensitisation. Furthermore, because some chronic stressors, but also corticosterone administration, decrease forepaw treading and/or flat body posture intensities after the acute administration of 5-HT_1A receptor agonists (Chaouloff, 1993), it has been proposed that these behavioral responses allow the recognition of glucocorticoid- and/or stress-related effects on postsynaptic 5-HT_1A receptors. Indeed, it is of first importance to add that forepaw treading and flat body posture changes, either on repeated stimulation of 5-HT_1A receptors (De Vry, 1995) or after nerve lesions (Wieland et al., 1990), are not necessarily associated with parallel changes in radioligand binding at 5-HT_1A receptors in the central nervous system. Unfortunately, SHR displayed increased forepaw treading but decreased flat body posture compared with LEW rats, rendering difficult any interpretation regarding strain differences in postsynaptic 5-HT_1A receptors. Besides the observation that the intensity of forepaw treading increased in a dose-dependent manner yet flat body posture was already maximal with the 0.5 mg/kg dose of 8-OH-DPAT, in confirmation of previous data (Tricklebank et al., 1984; Zamfir et al., 1992), our finding regarding the heterogeneous strain-related changes in two putative indices of 5-HT_1A receptor function is noteworthy. Interestingly, we recently observed a similar pattern in a comparison of Fischer 344 and LEW rats (i.e., strains that differ in hippocampus but not midbrain 5-HT_1A receptor binding) (Fischer 344 > LEW: Burnet et al., 1992; Chaouloff et al., 1995). Thus, 8-OH-DPAT-induced flat body posture and forepaw treading were increased and decreased in LEW rats, respectively, compared with Fischer 344 rats (data not shown).

The entire set of data could then suggest that SHR and LEW rats differ (1) in the respective function of the 5-HT_1A receptors that mediate flat body posture or forepaw treading (and are thought to be in the brainstem and/or the spinal cord: Jacobs and Klemfuss, 1975; Wieland et al., 1990), (2) in nonserotonergic mediators of 5-HT_1A receptor-mediated forepaw treading and flat body posture (e.g., catecholaminergic: Tricklebank et al., 1984) and/or (3) in the availability of 8-OH-DPAT, when injected systemically, at its respective targets. If one or all of these hypotheses are true, the real impact on strain-related differences in anxiety levels remains to be determined.

As reviewed previously, acute or repeated administration of 5-HT_2A receptor antagonists often, but not always, leads to anxiolysis, including during the elevated plus-maze test (Griebel, 1995; Handley, 1995). This result is in keeping with the observation that down-regulation of 5-HT_2A receptors (e.g., in the frontal cortex) and/or functional desensitization of these receptors may occur on treatment with serotonergic compounds (e.g., SSRIs, 5-HT_2A/2C receptor antagonists) endowed with anxiolytic effects in the elevated plus-maze (Benjamin et al., 1992; Cadogan et al., 1992; Johnson, 1991; however, see Hrdina and Vu, 1993). In the present study, [³H]ketanserin binding in the frontal cortex and the striatum did not differ significantly between strains; conversely, DOI- and quipazine-elicited head shaking (i.e., a 5-HT_2A receptor-mediated behavior) (Schreiber et al., 1995) was increased in SHR compared with in LEW rats. Furthermore, DOI elicited a dose-dependent effect in SHR but not LEW rats. Because the strain-related effect of DOI was shared by another 5-HT_2A receptor agonist (i.e., quipazine), although these are different chemical entities, the higher head shake response to DOI in SHR compared with LEW rats cannot be accounted for by strain-related differences in drug pharmacokinetics.

Taken together, the results could thus suggest a genetic difference in the effector coupling to the binding site. As mentioned, second-messenger studies (i.e., phosphatidyl inositol in the case of 5-HT_2A receptors), which we will begin, would provide an examination of this hypothesis. However, one must keep in mind that (1) 5-HT_2A receptor-mediated head shakes are regulated by numerous transmitter systems (e.g., the noradrenergic system) (Handley and Singh, 1986). Therefore, strain-related differences in these systems but not in 5-HT_2A receptor-effector coupling could well account for our behavioral observation (2) because [³H]ketanserin binds to both G protein-coupled and -uncoupled forms of the 5-HT_2A receptor (i.e., high- and low-affinity states of the receptor), whereas agonists such as DOI bind only to the G protein-coupled form of the receptor (Teiteler et al., 1990), it may be that SHR and LEW rats differ only in the high-affinity state of the protein. Therefore, if SHR display a higher proportion of G protein-coupled 5-HT_2A receptors than do LEW rats, this difference would fit with the aforementioned difference in head shake responses to DOI.

Recent studies in animals and humans have reported anxiolytic effects of 5-HT reuptake blockers on repeated administration (Cadogan et al., 1992; Handley and McBlane, 1993; Nutt, 1995). This effect may appear paradoxical in view of the (1) stimulatory effects of these repeated treatments on 5-HT transmission (Blier and De Montigny, 1994) and (2) possible anxiogenic consequences of increased 5-HT activity, at least in unconditioned models (Griebel, 1995; Handley, 1995). Indeed, the anxiolytic effect of SSRIs has been tentatively assigned to indirect (i.e., 5-HT-mediated) 5-HT_2B/2C receptor down-regulation, as demonstrated by decreases in various 5-HT_2B/2C receptor-mediated behaviors, including mCPP-induced hypolocomotion (Kennett et al., 1994; Maj and Moryl, 1992). Interestingly, chronic treatment with SSRIs may be associated with decreases in the genomic expression of and/or radioligand binding at 5-HT reuptake systems (Lesch et al., 1993; Pineyro et al., 1994; but see Hrdina and Vu, 1993).

In keeping with the aforementioned results, we thus investigated (1) [³H]citalopram binding in serotonergic cell bodies and hippocampal nerve terminals and (2) mCPP-induced hypolocomotion. We found that [³H]citalopram binding, which was higher in midbrain than in the hippocampus, thus confirming previous results (D'Amato et al., 1987), did not significantly differ between strains. In keeping with the lack of genetic difference regarding 5-HT synthesis, our results could suggest that in hippocampus, extracellular levels of 5-HT, which depend on release and reuptake of the amine, are identical in SHR and LEW rats. Indeed, this suggestion could extend to serotonergic cell bodies because 5-HT_1A autoreceptor sensitivity did not differ between strains (see above). As far as putative genetic differences in 5-HT_2B/2C receptor-mediated hypolocomotion are concerned, the observation that both 15-min horizontal and vertical activities were diminished in saline-treated LEW rats compared with saline-treated SHR could impede any conclusion. However, the lack of a significant strain × drug interaction when data were expressed as percentage of change from control values allows us to suggest that 5-HT_2B/2C receptors, at least those mediating the hypolocomotor effects of mCPP (Kennett and Curzon, 1988), do not differ between SHR and LEW rats. This statement is reinforced by the observation that a 0.5 mg/kg dose of MK-212 (Lucki et al., 1989), another 5-HT_2B/2C receptor agonist (Kennett, 1993), decreased locomotion to similar extents in SHR and LEW rats (22% and 26%, respectively; data not shown). Whether this lack of a genetic difference extends to other 5-HT_2B/2C receptor-mediated functions, including anxiety (Kennett, 1993), is a question of prime importance. Although this possibility merits consideration (although it is difficult to check with agonists such as mCPP due to the low base-line levels displayed by LEW rats in the plus-maze), it is worthy of mention that 5-HT_2B/2C receptor agonists/antagonists are effective in the social interaction test (Kennett, 1993) (i.e., a paradigm in which SHR and LEW rats did not differ).

In conclusion, in the present study, we addressed the possibility that central serotonergic systems play a role in the differences displayed by SHR and LEW rats in the elevated plus-maze. Actually, neither 5-HT synthesis nor radioligand binding at several key receptors differed between strains. Furthermore, our study indicates that 5-HT_1A autoreceptors are not involved in the behavioral difference between SHR and LEW rats. On the other hand, this study shows that future experiments aimed at analyzing 5-HT_1A and 5-HT_2A receptor-effector coupling could prove fruitful; however, whether these genetic differences in receptor-effector coupling, if any, underlie the respective anxiety scores in SHR and LEW rats will require thorough investigation.