Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The use of selective 5-hydroxytryptamine (serotonin) (5-HT) ) reuptake inhibitors (SSRIs), such as fluoxetine, in the treatment of neuropsychiatric disorders has become increasingly widespread (Goldberg, 1995; Fuller, 1996). Inhibition of 5-HT reuptake increases 5-HT levels in the synaptic cleft, thereby prolonging the activation of postsynaptic 5-HT receptors. Therapeutic effects, however, are not attained in patients until 2 to 3 weeks after the beginning of treatment (Artigas et al., 1996; Quitkin et al., 1996). Adaptive changes in 5-HT receptors may underlie the therapeutic effectiveness of SSRIs (Le Poul et al., 1995; Blier and de Montigny, 1996; Li et al., 1996; Li et al., 1997a, 1997b).

Repeated exposure to SSRIs induces desensitization of somatodendritic 5-HT_1A autoreceptors in the raphe region within 3 days (Le Poul et al., 1995). Because these autoreceptors serve as feedback inhibitors of serotonergic cell firing, their desensitization leads to increased serotonergic neurotransmission. In vivo microdialysis studies indicate that chronic treatment with fluoxetine increases the levels of 5-HT in the forebrain (Rutter et al., 1994). Higher levels of 5-HT in the synaptic cleft lead to a subsequent desensitization of the postsynaptic 5-HT_1A receptors in forebrain regions like the hypothalamus (Lesch et al., 1991; Li et al., 1993b, 1994, 1996; Lerer et al., 1997; Li et al., 1997b). Hence, neuroadaptive changes produced by repeated exposure to SSRIs include 5-HT_1A receptors.

We have consistently shown in rats that daily injections of 10 mg/kg/day fluoxetine or paroxetine for 14 to 21 days completely inhibit the oxytocin, ACTH, and corticosterone responses to the 5-HT_1A agonists 8-OH-DPAT or ipsapirone (Li et al., 1993b, 1994, 1996, 1997b). This fluoxetine dose is about 10 times higher than the clinically relevant doses. For example, a reduced ACTH response to ipsapirone is observed in normal humans and in patients with obsessive compulsive disorder taking 20 to 60 mg/day fluoxetine (approximately equivalent to 0.3-1 mg/kg/day) for 4 weeks to 3 months (Lesch et al., 1991; Lerer et al., 1997). The seemingly high dose of fluoxetine (10 mg/kg/day) in rats is the minimum required to produce a complete inhibition of 5-HT uptake (Fuller et al., 1978); it is unknown what dose is required in humans. If the degree of desensitization of the postsynaptic 5-HT_1A receptors is dependent on the degree of blockade of 5-HT transporters, then it should depend on the dose of SSRI administered. The present study investigated chronic exposure to lower doses of fluoxetine to determine if the degree of desensitization of hypothalamic 5-HT_1A receptors is dose-dependent.

We have previously shown that exposure to fluoxetine does not affect the density or the affinity of hypothalamic 5-HT_1A receptors, suggesting that the mechanism of desensitization involves alterations in signal transduction (Li et al., 1993b). 5-HT_1A receptors are coupled to the G_i family of G proteins, which include G_i1, G_i2, G_i3, G_o, and G_z proteins (Raymond et al., 1993; Barr et al., 1997). Hypothalamic G_z proteins mediate the ACTH and oxytocin responses to activation of 5-HT_1A receptors (Serres et al., 1998; Van de Kar et al., 1998). Therefore, to understand the mechanism of fluoxetine-induced desensitization of 5-HT_1A receptor systems, we investigated changes in hypothalamic G_z protein levels as a likely component of the 5-HT_1A signaling system to be altered by fluoxetine. Our previous time course study indicates that daily injections of fluoxetine produce a gradual reduction in hypothalamic G_i1 and G_i3 protein levels that matches the time course of desensitization of hypothalamic 5-HT_1A receptors (as measured from the reduction in the neuroendocrine responses to 8-OH-DPAT) (Li et al., 1996). Levels of G_z protein in the hypothalamus have not been examined. Therefore, we compared the dose-dependent changes in G_i proteins with the G_z protein levels in the hypothalamus and with the neuroendocrine responses to 8-OH-DPAT. To begin to determine the mechanisms underlying the reduction in the levels of membrane-associated G_z and G_i1 proteins, we investigated whether there was a redistribution of membrane-bound G proteins to the cytosol.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals

Male Sprague-Dawley rats (225-250 g) were purchased from Harlan Laboratories (Indianapolis, IN). Animals were housed 2 per cage in a room controlled for temperature, humidity, and lighting (lights on 0700-1900). Food and water were available ad libitum at all times. All procedures were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals as approved by the Loyola University Institutional Animal Care and Use Committee.

Drugs

Fluoxetine HCl was donated by Eli Lilly and Co. (Indianapolis, IN). A fresh fluoxetine solution for the highest dose in each experiment was made daily by dissolving fluoxetine in the vehicle (0.9% saline). Serial dilutions of the fluoxetine solution were then prepared to correspond to the lower doses in each experiment, allowing all animals to receive injections (i.p.) in a volume of 2 ml/kg. 8-OH-DPAT was purchased from Research Biochemical Inc. (Natick, MA); it was dissolved in saline and injected (s.c.) in a volume of 1 ml/kg.

Animal Drug Treatments

In experiment 1, rats were injected (i.p.) once daily for 14 days with either 0 (saline), 0.3, 1, 3, or 10 mg/kg fluoxetine. In experiment 2, rats were injected (i.p.) once daily for 14 days with either 0 (saline), 5, 7.5, or 10 mg/kg fluoxetine. Doses of fluoxetine were chosen in experiment 1 to cover two logarithmic scales. In experiment 2, the choice of doses was based on the results obtained in experiment 1. All other procedures were similar in experiments 1 and 2. Body weight was recorded throughout the injection period. Rats were then challenged 18 h after the last fluoxetine injection with either saline or 8-OH-DPAT (50 µg/kg, s.c.) and decapitated 15 min later. 8-OH-DPAT is a 5-HT_1A agonist with a high affinity for 5-HT_1A receptors and at least a 10-fold lower affinity for other 5-HT receptor subtypes (Boess and Martin, 1994). By activating hypothalamic 5-HT_1A receptors, 8-OH-DPAT stimulates the secretion of oxytocin, ACTH, and corticosterone (Bagdy, 1996). Maximal stimulation of the secretion of oxytocin and ACTH is reached by 8-OH-DPAT doses of 200 to 500 µg/kg (s.c.) (Li et al., 1993b). A submaximal dose of 8-OH-DPAT was used in the present study (50 µg/kg, s.c.) to prevent activation of other 5-HT receptor subtypes. This dose of 8-OH-DPAT was previously shown to be the most effective dose in demonstrating desensitization of 5-HT_1A receptors (Li et al., 1996). Trunk blood was collected in centrifuge tubes containing 0.5 ml of a 0.3 M ethylenediamine tetraacetic acid (EDTA; pH 7.4) solution. The plasma was separated and stored at -70°C until measurement of hormone levels. The hypothalamus was removed from the brain immediately after decapitation and frozen in liquid nitrogen, then stored at -70°C for the measurement of G protein levels.

Radioimmunoassay of Hormones

Plasma oxytocin, ACTH, and corticosterone concentrations were determined in all animals from both experiments by radioimmunoassays as previously described in detail (Li et al., 1993b, 1997b).

Immunoblots of G Proteins

Cytosolic and Membrane-Bound Protein Fractionation. G protein levels were measured in hypothalamic tissue from animals in experiment 2 that received a saline challenge. All procedures were conducted at 4°C unless otherwise indicated. Briefly, the tissues were homogenized in 0.5 ml of a 50 mM Tris buffer, pH 7.4, containing 150 mM NaCl, 10% sucrose, and 0.5 mM phenylmethanesulfonyl fluoride (PMSF). After centrifugation at 20,000 g for 60 min, the supernatant was collected and stored for determination of cytosolic G protein levels. The pellets (containing the membrane-bound proteins) were solubilized in a 20 mM Tris buffer, pH 8 (containing 1 mM EDTA, 100 mM NaCl, 1% sodium cholate, and 1 mM dithiothreitol) in a ratio of 3 µl buffer/mg tissue. The resuspended homogenates were incubated and shaken for 1 h, followed by centrifugation at 100,000 g for 60 min. The supernatant (containing the membrane-bound G proteins) was collected and stored for the determination of G protein levels. Protein concentrations were measured using bovine serum albumin as a standard. The total-protein concentrations were between 3.6 to 5.4 µg/µl.

Quantification of G proteins. The cytosolic and membrane-bound solubilized proteins were resolved by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis, using 0.75 mm thick Tris-Glycine denaturing reducing gels, containing 0.1% SDS, 12.5% acrylamide/bisacrylamide (30:0.2), 4.6 M urea and 375 mM Tris, pH 8.7 (Mullaney and Miligan, 1990). Three samples from each treatment group and three randomly selected controls were loaded on each gel. Each sample was measured on three independent gels. The proteins were then electrophoretically transferred for 2 hours to nitrocellulose membranes. After drying, the membranes were incubated in a blocking solution containing 5% nonfat dry milk, 0.05% NP40, 50 mM Tris, and 150 mM NaCl, pH 7.4 for 1 h, washed and incubated overnight at 4°C with polyclonal antisera for G_i1/2 (AS/7, Du-Pont NEN, Boston, MA, 1:2,500 dilution), G_i3 (Anti-G_i3, Upstate Biotechnology Inc., Lake Placid, NY, 1:2,000 dilution), G_o (GC/2, Du-pont NEN, 1:10,000 dilution), or G_z (I-20, Santa Cruz Biotechnology Inc., Santa Cruz, CA, 1:6,000 dilution) at 4°C overnight. The membranes were then incubated with a secondary antibody (goat anti-rabbit serum, Cappell, Organon Teknika Corp, Durham, NC, 1:5,000 dilution), followed by an incubation with rabbit peroxidase-antiperoxidase (Cappell, Organon Teknika, 1:10,000). The membranes were incubated with the ECL chemiluminescence substrate solution (Amersham, Arlington Heights, IL) and then exposed to Kodak X-ray film.

Data Analysis. Films were analyzed densitometrically using NIH Image (v. 1.57) for Macintosh computers. The gray-scale density readings were calibrated using a transmission step wedge standard. The integrated optical density (IOD) of each band was calculated as the sum of optical densities of all the pixels within the area of the band outlined. An area adjacent to the G protein bands was used to calculate the background optical density of the film. The IOD for the film background was subtracted from the IOD for each band. The resulting IOD for each G protein band was then divided by the amount of protein loaded on the corresponding lane. The IOD/µg protein values obtained from treated rats were divided by the mean IOD/µg protein values obtained from control rats in each gel to determine the relative amounts of the G proteins. The data for each rat were the means obtained from the three gels.

High-Pressure Liquid Chromatography (HPLC) of Fluoxetine and Norfluoxetine

Plasma Preparation. Fluoxetine and norfluoxetine levels were measured in plasma from treated animals that received a saline challenge before sacrifice. Extraction of fluoxetine and norfluoxetine from plasma was carried out at room temperature. In a 2-ml polypropylene microfuge tube, 5 µl of a 100 µM fluvoxamine solution in 0.01 M H₃PO₄ (internal standard) was added to 500 µl of thawed plasma; 500 µl of 1 M NaOH was added and tubes were mixed; 600 µl of chloroform was then added and tubes were shaken vigorously for 3 min. Tubes were then centrifuged at 9,000g for 1 min, and the aqueous phase was aspirated and discarded. Compounds were back-extracted from the chloroform phase by adding 400 µl of 0.1 M H₃PO₄, shaking for 3 min and centrifuging at 9,000 g for 1 min; 375 µl of the aqueous phase was collected, placed in new tubes, and evaporated at 43°C in a SpeedVac (Savant Instruments; Holbrook, NY). The dried residue was reconstituted in 100 µl of 0.005 M H₃PO₄ and centrifuged at 9,000 g for 1 min. The clear extract was placed in autosampler vials in a refrigerated autosampler (SIL-10A; Shimadzu Scientific Instrument; Wood Dale, IL) for injection of 50 µl into the HPLC. Preliminary studies using spiked control plasma demonstrated that the recovery of fluvoxamine, fluoxetine, and norfluoxetine were similar and averaged 66 ± 2%.

HPLC Conditions. Fluoxetine and norfluoxetine were separated by reversed-phase HPLC. The mobile phase consisted of 0.02 M NaH₂PO₄, 0.02 M sodium citrate, 200 mM EDTA, 35% acetonitrile, 4.4 mM tetraethylammonium (hydrogen sulfate salt), and was brought to an apparent pH of 4.5 using H₃PO₄. The stationary phase consisted of a 7.5 cm × 4.6 mm (i.d.) Ultrasphere ODS 3 mm reversed-phase column (Beckman Instruments, CA). Flow rate was adjusted to 0.9 ml/min. Under these conditions, the retention times of fluvoxamine, norfluoxetine, and fluoxetine were 3.0, 3.8, and 4.5 min, respectively. Compounds were detected by ultraviolet absorbance at 230 nm using a UV detector (SPD-10A; Shimadzu Scientific Instruments). Quantitation was achieved by comparing peak heights of samples with peak heights of standards. The limit of sensitivity was 75 nM for fluoxetine and 25 nM for norfluoxetine.

Statistical Analyses

Hormone data were analyzed by two-way analysis of variance (ANOVA). G protein levels were analyzed by one-way ANOVA. Group means were compared by Newman-Keuls' multiple range test. Linear correlation analysis also was carried out for within-group relationships between plasma hormone and hypothalamic G protein levels in untreated rats (Steel and Torrie, 1960). GB-STAT software (Dynamic Microsystems, Inc., Silver Spring, MD) was used for all statistical analyses.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Fluoxetine and Norfluoxetine in Plasma

Fluoxetine and norfluoxetine concentrations were determined in the plasma of rats 18 h after the last injection. The data are presented in Table 1. Consistent with the short half-life (7.7 h) of fluoxetine in rats (Caccia et al., 1990), no detectable levels of fluoxetine were observed in plasma even after treatment with the highest fluoxetine dose (10 mg/kg/day). The levels of norfluoxetine in plasma were elevated in a dose-dependent manner (Table 1). The minimum dose of fluoxetine that resulted in detectable levels of norfluoxetine in plasma was 5 mg/kg/day.

Body Weight Changes

Daily injections of 10 mg/kg/day fluoxetine significantly inhibited weight gain during the 14 days of injections (Fig. 1). Doses of fluoxetine between 0.3 and 3 mg/kg/day had no effect on weight gain (Fig. 1A), whereas fluoxetine doses of 5, 7.5, and 10 mg/kg/day inhibited weight gain in a dose-dependent manner (Fig. 1B).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Dose-response effect of fluoxetine on body weight gain. A, experiment 1 using fluoxetine doses of 0.3, 1, 3, and 10 mg/kg/day; B, experiment 2 using fluoxetine doses of 5, 7.5 and 10 mg/kg/day. The data represent the mean ± S.E.M. of 20 rats per group. A, the two-way ANOVA indicated the following: main effect of days F(8,693) = 523.76, p < 0.01; main effect of fluoxetine dose F(4,693) = 64.83, p < 0.01; days × fluoxetine dose F(32,693) = 3.79, p < .01. B, the two-way ANOVA indicated the following: main effect of days F(7,608) = 142.7, p < .01; main effect of fluoxetine dose F(3,608) = 77.27, p < .01; days × fluoxetine dose F(21, 608) = 2.80, p < 0.01. The post hoc Newman-Keuls' tests indicated the following: +significant difference compared with the saline group, p < 0.05; *significant difference compared with the saline group, p < 0.01.

Hormone Responses to 8-OH-DPAT

8-OH-DPAT significantly increased plasma oxytocin (Fig. 2), ACTH (Fig. 3), and corticosterone (Fig. 4) levels. Consistent with our previous results, chronic injections of 10 mg/kg/day fluoxetine completely blocked plasma oxytocin (Fig. 2), ACTH (Fig. 3), and corticosterone (Fig. 4) responses to 8-OH-DPAT without altering basal hormone levels. Doses of fluoxetine between 0.3 and 3 mg/kg/day did not alter the effect of 8-OH-DPAT on plasma hormones (Figs. 2A, 3A, and 4A).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Dose-response effect of fluoxetine on the oxytocin response to 8-OH-DPAT (50 µg/kg s.c.). A, experiment 1 using fluoxetine doses of 0.3, 1, 3, and 10 mg/kg/day; B, experiment 2 with fluoxetine doses of 5, 7.5, and 10 mg/kg/day. The data represent the mean ± S.E.M. of 8 to 12 rats per group. A, the two-way ANOVA indicated the following: main effect of fluoxetine dose F(4,60) = 4.75, p < 0.01; main effect of 8-OH-DPAT F(1,60) = 48.60, p < 0.01; fluoxetine dose × 8-OH-DPAT, F(4,60) = 4.86, p < 0.01. B, the two-way ANOVA indicated the following: main effect of fluoxetine dose F(3,66) = 16.36, p < 0.01; main effect of 8-OH-DPAT F(1,66) = 86.30, p < 0.01; fluoxetine dose × 8-OH-DPAT, F(3,66) = 17.44, p < 0.01. The posthoc Newman-Keuls' tests indicated the following: *significant difference from the rats challenged with saline (0 dose of 8-OH-DPAT), p < 0.01; #significant difference from the vehicle rats (0 dose of fluoxetine) challenged with 8-OH-DPAT, p < 0.01.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Dose-response effect of fluoxetine on the ACTH response to 8-OH-DPAT (50 µg/kg s.c.). A, experiment 1 using fluoxetine doses of 0.3, 1, 3, and 10 mg/kg/day; B, experiment 2 with fluoxetine doses of 5, 7.5, and 10 mg/kg/day. The data represent the mean ± S.E.M. of 8 to 12 rats per group. A, the two-way ANOVA indicated the following: main effect of fluoxetine F(4,70) = 5.86, p < 0.01; main effect of 8-OH-DPAT F(1,70) = 89.83, p < 0.01; fluoxetine × 8-OH-DPAT, F(4,70) = 5.93, p < 0.01. B, the two-way ANOVA indicated the following: main effect of fluoxetine F(3,68) = 13.64, p < 0.01; main effect of 8-OH-DPAT F(1,68) = 75.90, p < 0.01; fluoxetine X 8-OH-DPAT, F(3,68) = 12.15, p < 0.01. The post hoc Newman-Keuls' tests indicated the following: *significant difference from the rats challenged with saline (0 dose of 8-OH-DPAT), p < 0.01; #significant difference from the vehicle rats (0 dose of fluoxetine) challenged with 8-OH-DPAT, p < 0.01.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Dose-response effect of fluoxetine on the corticosterone response to 8-OH-DPAT (50 µg/kg s.c.). A, experiment 1 using fluoxetine doses of 0.3, 1, 3, and 10 mg/kg/day; B, experiment 2 with fluoxetine doses of 5, 7.5, and 10 mg/kg/day. The data represent the mean ± S.E.M. of 8 to 12 rats per group. A, the two-way ANOVA indicated the following: main effect of fluoxetine F(4,67) = 3.22, p < 0.05; main effect of 8-OH-DPAT F(1,67) = 51.60, p < 0.01; fluoxetine X 8-OH-DPAT, F(4,67) = 3.23, p < 0.05. B, the two-way ANOVA indicated the following: main effect of fluoxetine F(3,66) = 10.23, p < 0.01; main effect of 8-OH-DPAT F(1,66) = 72.86, p < 0.01; fluoxetine × 8-OH-DPAT, F(3,66) = 7.28, p < 0.01. The posthoc Newman-Keuls' tests indicated the following: * significant difference from the rats challenged with saline (0 dose of 8-OH-DPAT), p < 0.01; # significant difference from the vehicle rats (0 dose of fluoxetine) challenged with 8-OH-DPAT, p < 0.01.

However, fluoxetine doses between 5 and 10 mg/kg/day produced a dose-dependent decrease in the effect of 8-OH-DPAT on plasma oxytocin (Fig. 2B), ACTH (Fig. 3B), and corticosterone (Fig. 4B), without altering basal hormone levels. The minimum effective dose of fluoxetine that reduced the neuroendocrine responses to 8-OH-DPAT was 5 mg/kg/day. Exposure to 7.5 mg/kg/day fluoxetine produced a greater, although incomplete, reduction in the oxytocin, ACTH, and corticosterone responses to 8-OH-DPAT. In both experiments, a complete blockade of the oxytocin, ACTH, and corticosterone responses to 8-OH-DPAT was attained after exposure to 10 mg/kg/day fluoxetine (see Table 2 for reduction percentages).

G Proteins in Hypothalamus

Examples of immunoblots of membrane-bound and cytosolic G_i1, G_i2, and G_z proteins in the hypothalamus are presented in Fig. 5 (A-D). The specificity of the antibodies for G_i1, G_i2, G_i3, and G_o proteins has been verified in previous studies (Li et al., 1996). The specificity of the G_z antiserum was confirmed by showing that preabsorption of the antibody with a G_z blocking peptide prevented the antiserum from binding to G_z protein on immunoblots (Fig. 5E).

View larger version (40K): [in this window] [in a new window]   Fig. 5.   A-D, examples of immunoblots of membrane-bound (left) and cytosolic (right) G proteins in the hypothalamus of rats treated with increasing doses (0, 5, 7.5, and 10 mg/kg/day, i.p.) of fluoxetine for 14 days. A, membrane-bound Gi1 and Gi2 proteins; B, cytosolic Gi1 and Gi2 proteins; C, membrane-bound Gz proteins; D, cytosolic Gz proteins; E, a preabsorption control, indicating that a Gz blocking protein prevents the antibody from binding to the Gz protein on the blot.

Membrane Proteins. Hypothalamic levels of G_z protein (Fig. 6A) were significantly reduced after exposure to the 7.5 and the 10 mg/kg/day doses of fluoxetine. Consistent with previously reported results (Li et al., 1996), repeated daily injections of 10 mg/kg/day fluoxetine produced a significant decrease in G_i1 protein levels (Fig. 6B). G_i3 protein levels were not significantly reduced after chronic exposure to 10 mg/kg/day fluoxetine. However, exposure to 5 and 7.5 mg/kg/day fluoxetine significantly increased the levels of G_i3 proteins in the hypothalamus (Fig. 6C). The levels of G_i2 protein (Fig. 6D) and G_o protein (Fig. 6E) were not altered by chronic exposure to any fluoxetine doses.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   Dose-response effect of fluoxetine on membrane-bound (left) and cytosolic (right) G proteins. From top to bottom, the G proteins are Gz, Gi1, Gi3, Gi2, and Go. The data represent the mean ± S.E.M. of 5 to 6 rats per group. For membrane G proteins, the one-way ANOVAs indicated the following: Gz, F(3,19) = 4.13, p < 0.05; Gi1, F(3,20) = 5.40, p < 0.01; Gi3, F(3,32) = 4.464, p < 0.01; Gi2, F(3,21) = 2.21, p > 0.10; Go, F(3,23) = 0.40, p > 0.10. For cytosolic G proteins, the ANOVAs indicated the following: Gz, F(3,19) = 0.86, p > 0.10; Gi1, F(3,17) = 2.36, p > 0.10; Gi3, F(3,18) = 0.05, p > 0.10; Gi2, F(3,20) = 3.06, p > 0.05; Go, F(3,18) = 0.63, p > 0.10. The post hoc Newman-Keuls' tests indicated the following: *significant difference from the vehicle rats, p < 0.05; **significant difference from the vehicle rats, p < 0.01.

Cytosolic Proteins. Levels of G_z, G_i1, G_i3, G_i2, and G_o proteins in the cytosol were unaltered by exposure to fluoxetine (Fig. 6 F-J).

Comparison of Hormone Responses to Hypothalamic G Proteins. Table 2 presents the percent reduction as compared to controls for the hormone responses to 8-OH-DPAT and hypothalamic G protein levels for the 5, 7.5, and 10 mg/kg/day doses of fluoxetine. Overall, reductions in the magnitudes of hormone responses to 8-OH-DPAT followed a similar pattern for all three hormones (i.e., partial and significant reduction after 5 mg/kg/day fluoxetine, further reduction after 7.5 mg/kg/day, complete reduction after 10 mg/kg/day). In contrast, the effect of fluoxetine on G protein subunits were variable. Reductions in G_z protein levels were the most similar to hormone responses, while reductions in G_i1 proteins only occurred with the highest fluoxetine dose.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The most important and novel findings in these studies are: 1) fluoxetine dose-dependently reduces the neuroendocrine response to a 5-HT_1A agonist, suggesting a dose-dependent desensitization of hypothalamic postsynaptic 5-HT_1A receptors; 2) fluoxetine is more effective in reducing the hypothalamic levels of G_z proteins, which couple 5-HT_1A receptors to the secretion of ACTH and oxytocin, than it is in reducing the levels of other G_i and G_o proteins; 3) fluoxetine does not induce a translocation of G proteins from the membranes into the cytosol; and 4) to our knowledge, this is the first in vivo study showing changes in G_z protein levels in the brain.

This study provides a systematic examination of fluoxetine's dose-response effect on postsynaptic 5-HT_1A receptor systems in the hypothalamus of rats. Although many studies using fluoxetine in rats have used doses of 10 to 30 mg/kg/day (Nestler et al., 1990; Trouvin et al., 1993; Gardier et al., 1994), others have used lower doses of 2.5 to 5 mg/kg/day (Lesch and Manji, 1992; Nibuya et al., 1996). Because of this large variability in daily fluoxetine doses, there could be disparities in the effects of fluoxetine on serotonergic function. Our studies provide information on the minimum dose that will produce adaptive changes in postsynaptic 5-HT_1A receptors in rats.

Consistent with previously reported results in rats (Li et al., 1993b, 1996), the present study demonstrates that repeated daily injections of a 10 mg/kg/day dose of fluoxetine produces a complete inhibition of the neuroendocrine responses to a submaximal dose of 8-OH-DPAT, suggesting a desensitization of hypothalamic 5-HT_1A receptors. Furthermore, the current findings indicate a dose-dependent effect of fluoxetine on the hormone responses to the 5-HT_1A agonist. The lowest dose of fluoxetine necessary to produce desensitization of hypothalamic postsynaptic 5-HT_1A receptors after 14 days of exposure is 5 mg/kg. This dose in rats was reported to produce a 77% inhibition of 5-HT uptake in vivo, whereas a fluoxetine dose of 3 mg/kg only produced a 27% inhibition of 5-HT uptake (Fuller et al., 1978). These observations suggest that a high percentage of 5-HT uptake sites need to be blocked in order to produce desensitization of postsynaptic 5-HT_1A receptors. Furthermore, as inhibition of 5-HT uptake sites becomes more complete, the degree of 5-HT_1A receptor desensitization increases.

As mentioned in the introduction, the doses necessary to achieve desensitization of 5-HT_1A receptors in rats are about 5- to 10-fold higher than the doses necessary in humans. The reason for this difference in dose effectiveness is not clear. It is possible that the difference in doses is caused by differences in pharmacokinetics of fluoxetine between rats and humans. In rats, the half-life after a 10 mg/kg dose of fluoxetine is 7.7 h for the parent drug and 15.8 h for its pharmacologically active metabolite, norfluoxetine (Caccia et al., 1990). In contrast, the half-life of fluoxetine in humans is 4 to 6 days, and of norfluoxetine is 4 to 16 days (DeVane, 1994). Because humans take fluoxetine (20-60 mg) daily, the accumulation of fluoxetine and norfluoxetine in their brain could be as high or even higher than it would be with higher doses in rats. Examination of plasma levels of fluoxetine 18 h after the last fluoxetine injection revealed no detectable levels of fluoxetine. This observation is consistent with data reported by Caccia et al. (1990). Fluoxetine treatment resulted in a dose-dependent increase in the levels of norfluoxetine in plasma. The levels of norfluoxetine in plasma correspond to the changes in body weight and the desensitization of hypothalamic 5-HT_1A receptor signaling. Thus, the data suggest that fluoxetine doses below 5 mg/kg/day do not produce a sufficient level of fluoxetine and norfluoxetine that would produce a desensitization of hypothalamic 5-HT_1A receptors.

This study focused on signal transduction mechanisms as possible targets of SSRI action because repeated injections of fluoxetine and other SSRIs do not alter the density or the affinity of hypothalamic postsynaptic 5-HT_1A receptors (Hensler et al., 1991; Li et al., 1993b). A novel finding in the present study is the fluoxetine-induced reduction in the levels of membrane-associated G_z protein at both the 7.5 and the 10 mg/kg/day doses. G_z proteins are pertussis toxin insensitive members of the G_i protein family that have a high affinity for 5-HT_1A receptors in vitro (Barr et al., 1997). Our studies using pertussis toxin and G_z antisense oligodeoxynucleotides indicate that hypothalamic G_z proteins mediate 5-HT_1A receptor-stimulated ACTH and oxytocin secretion (Serres et al., 1998; Van de Kar et al., 1998). Therefore, G_z proteins might play a role in fluoxetine-induced desensitization of hypothalamic postsynaptic 5-HT_1A receptors.

Consistent with our previous findings (Li et al., 1996), membrane-associated G_i1 protein levels in the hypothalamus are also reduced after exposure to the 10 mg/kg/day dose of fluoxetine. This is a pertussis toxin-sensitive G protein that probably couples 5-HT_1A receptors in the raphe and hippocampus to hyperpolarization of neurons and possibly to other physiological responses (Clarke et al., 1987; Blier et al., 1993; Penington et al., 1993). Hence, the fluoxetine-induced reduction in G_i1 proteins in the hypothalamus might be relevant to other physiological phenomena linked to 5-HT_1A receptor desensitization. In contrast with our previous observation (Li et al., 1996), fluoxetine did not produce a significant reduction in G_i3 proteins. The inconsistency in effects of fluoxetine on the levels of G_i3 proteins suggest that G_i3 proteins are not significant to the desensitization of 5-HT_1A receptor systems in the hypothalamus. In the present study, levels of G_i2 and G_o proteins in the hypothalamus also were not altered after exposure to any dose of fluoxetine, further confirming the lack of involvement of G_i2 or G_o proteins in SSRI-induced desensitization of postsynaptic 5-HT_1A receptors in the hypothalamus (Li et al., 1996, 1997b).

Several mechanisms could underlie the fluoxetine-induced reduction in the levels of G proteins in the hypothalamus. These mechanisms could include reduced synthesis or increased degradation of G proteins. An additional mechanism responsible for the apparent reduction in hypothalamic levels of G_z and G_i1 proteins could be their redistribution from the membrane to the cytosol as has been shown to occur for G_z proteins in vitro (Hallak et al., 1994). However, cytosolic levels of the G proteins, including G_i1 and G_z proteins, were not altered after repeated exposure to any dose of fluoxetine, indicating that fluoxetine-induced reductions in membrane-bound proteins were not a result of redistribution to the cytosol.

Table 2 indicates that no relationship exists between the decrease in G protein levels and reduced hormone responses to 8-OH-DPAT; the best correspondence occurs with the levels of G_z proteins. One possible interpretation of these results is that, in addition to reductions in levels of G proteins, other mechanisms are also involved in the desensitization of hypothalamic 5-HT_1A receptors. Such a mechanism could include post-translational modifications of G_z proteins that would reduce the coupling efficiency of 5-HT_1A receptors to their second messengers. These possibilities will be investigated in future studies.

Table 3 shows that only G_z protein levels were significantly correlated with basal plasma levels of ACTH and oxytocin in untreated rats. We do not believe, however, that G_z proteins are the sole regulators of basal levels of ACTH and oxytocin because fluoxetine treatment reduced the levels of G_z proteins without changing the basal ACTH and oxytocin levels. Furthermore, the resting plasmas levels of hormones normally are maintained by multiple neurotransmitter and feedback mechanisms. Therefore, one would not expect a single factor to correlate with basal levels of hormones.

An alternative explanation for the incongruity between the degree of reduction in G_z proteins and hormone responses to 8-OH-DPAT is that the methodology used for quantifying G protein levels is not sensitive enough to detect small changes between groups. The immunoblot technique currently used by most researchers can detect gross changes but may not be sensitive enough to detect smaller changes. Because the maximum reduction in G protein levels after the highest dose of 10 mg/kg fluoxetine is only about a 35% decline, methodological issues would impede detecting small dose-dependent changes if there were indeed some. Therefore, the possibility that changes in G protein levels may be involved with functional desensitization of 5-HT_1A receptors cannot be excluded.

The importance of the results on change in body weight should be emphasized in light of the fact that most other classes of antidepressant drugs increase body weight, whereas SSRIs reduce body weight, or at least do not alter body weight substantially (Connolly et al., 1995; Greeno and Wing, 1996). The current results also support findings in rodents that weight gain is inhibited after chronic exposure to fluoxetine (Dryden et al., 1996; Halford and Blundell, 1996) and further indicate that this inhibition is dose-dependent. It could be argued that the inhibition in weight gain may contribute to the decline in neuroendocrine responses to the 5-HT_1A agonist 8-OH-DPAT. However, chronic exposure to fluoxetine produces a heightened hormone responses to the 5-HT_2C agonist MK-212 and the 5-HT_2A/2C agonist (±)-1-(2,5-Dimethoxy-4-lodophenyl)-2-aminopropane HCl (DOI) (Li et al., 1993a). Hence, chronic exposure to fluoxetine does not produces a generalized reduction in neuroendocrine systems.

In conclusion, the results of the present study indicate that fluoxetine-induced desensitization of hypothalamic postsynaptic 5-HT_1A receptors is dose-dependent, suggesting that the degree of 5-HT uptake blockade determines the degree of desensitization of postsynaptic 5-HT_1A receptors. Furthermore, changes in hypothalamic levels of G_z proteins (that couple 5-HT_1A receptors to the secretion of ACTH and oxytocin) may play a role in this desensitization. Altered synthesis or degradation of G_z proteins likely underlies the reduced levels of membrane associated G_z protein because G_z proteins were not redistributed to the cytosol.