OUP user menu

Brain-derived neurotrophic factor-deficient mice exhibit a hippocampal hyperserotonergic phenotype

Bruno P. Guiard, Denis J. P. David, Thierry Deltheil, Franck Chenu, Erwan Le Maître, Thibault Renoir, Isabelle Leroux-Nicollet, Pierre Sokoloff, Laurence Lanfumey, Michel Hamon, Anne M. Andrews, René Hen, Alain M. Gardier
DOI: http://dx.doi.org/10.1017/S1461145707007857 79-92 First published online: 1 February 2008


Growing evidence supports the involvement of brain-derived neurotrophic factor (BDNF) in mood disorders and the mechanism of action of antidepressant drugs. However, the relationship between BDNF and serotonergic signalling is poorly understood. Heterozygous mutants BDNF +/− mice were utilized to investigate the influence of BDNF on the serotonin (5-HT) system and the activity of the serotonin transporter (SERT) in the hippocampus. The zero net flux method of quantitative microdialysis revealed that BDNF +/− heterozygous mice have increased basal extracellular 5-HT levels in the hippocampus and decreased 5-HT reuptake capacity. In keeping with these results, the selective serotonin reuptake inhibitor paroxetine failed to increase hippocampal extracellular 5-HT levels in BDNF +/− mice while it produced robust effects in wild-type littermates. Using in-vitro autoradiography and synaptosome techniques, we investigated the causes of attenuated 5-HT reuptake in BDNF +/− mice. A significant decrease in [3H]citalopram-binding-site density in the CA3 subregion of the ventral hippocampus and a significant reduction in [3H]5-HT uptake in hippocampal synaptosomes, revealed mainly a decrease in SERT function. However, 5-HT1A autoreceptors were not desensitized in BDNF +/− mice. These results provide evidence that constitutive reductions in BDNF modulate SERT function reuptake in the hippocampus.

Key words
  • Brain-derived neurotrophic factor
  • dorsal raphe nucleus
  • hippocampus
  • serotonin transporter
  • 5-HT1A receptors


Growing evidence suggests that brain-derived neurotrophic factor (BDNF) is involved in both the pathophysiology of stress-related disorders such as depression and anxiety and the mechanism of action of serotonin selective reuptake-inhibiting antidepressants (SSRIs) (Duman and Monteggia, 2006). Accordingly, various clinical studies report reduced BDNF in the brains of unmedicated depressed patients (Dwivedi et al., 2006). Moreover, in animal studies, BDNF mRNA levels are increased in the hippocampus after chronic antidepressant treatment (Nibuya et al., 1995). These and many other studies suggest that serotonergic neurotransmission regulates BDNF expression (for review see Duman, 1998). Furthermore, reciprocal interactions between BDNF and serotonin (5-HT) in the central nervous system have been proposed (Mattson et al., 2004), however, the molecular mechanisms by which BDNF regulates 5-HT neurotransmission are also not well understood. BDNF has been shown to stimulate regenerative sprouting of serotonergic neurons after neurotoxic injury (Mamounas et al., 2000), and to promote serotonergic neurotransmission (Mamounas et al., 2000; Siuciak et al., 1996). In mice with partial constitutive reductions in BDNF expression (BDNF +/− mice), alterations in behaviours regulated by 5-HT have been observed including increased aggressiveness, hyperphagia and weight gain (Lyons et al., 1999). BDNF +/− mice also exhibit changes in the expression of several 5-HT post-synaptic receptor subtypes in various forebrain areas and accelerated age-related loss of serotonergic innervation to the hippocampus (Luellen et al., 2006; Lyons et al., 1999). However, the neurochemical bases underlying such alterations remain unclear.

In the present study, we used in-vivo microdialysis in young adult BDNF +/− mice (Guillin et al., 2001; Korte et al., 1995) to examine the role of reductions in endogenous BDNF in the regulation of 5-HT neurotransmission in the hippocampus. A previous study showed no changes in extracellular 5-HT ([5-HT]ext) levels in BDNF +/− mice in the striatum or frontal cortex (Szapacs et al., 2004). However, much of the evidence suggesting a co-regulatory relationship between 5-HT and BDNF points to the hippocampus as an important brain region for modulating these effects. Because [5-HT]ext levels reflect a balance between 5-HT release and reuptake, we determined the density and functional status of the serotonin transporter (SERT) using [3H]citalopram autoradiography and [3H]5-HT uptake from hippocampal synaptosomes in BDNF +/+ vs. BDNF +/− mice. Moreover, since somatodendritic 5-HT1A autoreceptors located within the dorsal raphe nucleus (DRN) are thought to play a key role in the therapeutic onset of efficacy of SSRIs (Artigas et al., 1996); we examined 5-HT1A autoreceptor function in BDNF +/− mice. Overall, the results demonstrate that constitutive reductions in BDNF selectively alter serotonergic neurotransmission in the hippocampus via a mechanism involving reductions in SERT expression and function.

Materials and Methods


Male wild-type BDNF +/+ and heterozygous mutant BDNF +/− mice (age 3–4 months, 25–30 g body weight) were bred on a mixed S129/Sv×C57BL/6 genetic background (Korte et al., 1995) and raised at the animal facility of the university of Paris XI (Chatenay-Malabry, France). Heterozygous adult mice with one functional BDNF allele (BDNF +/−) exhibited reduced BDNF protein levels in the hippocampus (data not shown). All animals were genotyped by polymerase chain reaction as described previously (Guillin et al., 2001). Animals were housed in groups of six mice per cage under standard conditions (12:12 h light–dark cycle, 22±1°C ambient temperature, 60% relative humidity, food and water ad libitum) and handled according to the guidelines of the French Agriculture Ministry (permission no. 92–196 to A.M.G.).


8-Hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT) hydrobromide was obtained from Sigma (L'Isle d'Abeau, France). Paroxetine hydrochloride was a gift from GlaxoSmithKline (Harlow, UK).

In-vivo microdialysis

Concentric dialysis probes (0.30 mm outer diameter) were constructed of cuprophane and set up as described previously (Guiard et al., 2004; Malagie et al., 2001). Probes were implanted into either the ventral hippocampus, the frontal cortex (active length of 1.6 mm) or the dorsal raphe nucleus (DRN) (active length 1.0 mm) of anaesthetized mice [chloral hydrate, 400 mg/kg, intraperitoneally (i.p.)] according to the mouse brain atlas of Paxinos and Franklin (2001). The coordinates from Bregma (mm) were: A=−3.4, L=3.4, V=4.0 for the ventral hippocampus; A=+1.6, L=1.3, V=1.6 for the frontal cortex and, A=−4.5, L=0, V=3.5 for the DRN. The next day, after recovery from surgery, probes were continuously perfused with artificial cerebrospinal fluid (aCSF) in awake animals at a flow rate of 1.5 µl/min in the ventral hippocampus and frontal cortex and at a flow rate of 0.5 µl/min in the DRN. Dialysate samples were collected every 15 min in both serotonergic terminal regions and every 30 min at the levels of cell bodies in the DRN for the measurement of 5-HT using high-performance liquid chromatography coupled to an amperometric detector (1049A, Hewlett-Packard, Les Ulis, France) The limit of sensitivity for 5-HT was ∼0.5 fmol/sample (signal-to-noise ratio 2).

Zero net flux microdialysis method

In most cases, four fractions were collected to determine basal hippocampal 5-HT levels before local perfusion of increasing concentrations of 5-HT (0, 5, 10 and 20 nm). The dialysate 5-HT concentrations (Cout) obtained during perfusion of the various concentrations of 5-HT (Cin) were used to construct a linear regression curve for each animal. The net change in 5-HT (CinCout) was plotted on the y-axis against Cin on the x-axis. [5-HT]ext levels and the extraction fraction of the probe (Ed) were determined as described by Parsons et al. (1991). The 5-HT concentration in the extracellular space ([5-HT]ext) is estimated from the concentration at which CinCout=0 and corresponds to a point at which there is no net diffusion of 5-HT across the dialysis membrane. The extraction fraction (Ed) is the slope of the linear regression curve and has been shown to provide an estimate of changes in transporter-mediated 5-HT uptake (Gardier et al., 2003; Parsons et al., 1991).

Conventional microdialysis

In separate animals, after 1 h of stabilization necessary to reach uniform 5-HT concentrations in the dialysate, four samples were collected to measure basal 5-HT values into either the ventral hippocampus, frontal cortex or DRN. Paroxetine was then injected intraperitoneally (4 and/or 8 mg/kg) at t=0 and subsequent dialysate fractions were collected. At the end of the experiments, placement of microdialysis probes was verified histologically.

[3H]Citalopram autoradiography

The protocol was adapted from d'Amato et al. (1987). After sacrifice, the brains of the mice were removed, frozen in isopentane at −30°C and stored at −80°C until sectioning. Coronal 20-µm tissue sections were cut on a cryostat at −22°C (Leica, Bensheim, Germany) throughout the brain from the anterior cortex to the brainstem and thaw-mounted onto chrome-alum 5% gelatin/0.5% coated glass slides, dehydrated at room temperature for 30 min and stored at −80°C. Brain sections were brought to room temperature and preincubated for 15 min at 37°C in Tris buffer (pH 7.4). To assess total binding, incubation was performed at room temperature for 1 h in Coplin jars containing 30 ml for 10 slides of the same buffer with 1 nm [3H]citalopram (PerkinElmer NEN, Courtaboeuf, France; sp. act. 84.2 Ci/mmol) in the presence of the protease inhibitor bacitracin at 0.5 mg/ml. Non-specific binding was evaluated in adjacent sections incubated under the same conditions in the presence of 10 µm fluoxetine (Lilly, Indianapolis, IN, USA). The slices were then dipped into buffer, rinsed twice in the same buffer at 4°C during 10 min, and dipped into distilled water. Slices were dried under a cold air stream for 30 min and allowed to dry and dehydrate overnight. The sections were apposed to Biomax MR films (Kodak, Courbevoie, France) for 5 wk, together with [3H] microscales (Amersham Biosciences, GE Healthcare Europe, Saclay, Orsay, France). SERT binding in the ventral hippocampus, striatum, frontal cortex and DRN of BDNF +/+ mice and BDNF +/− mice were quantified on a computerized image analysis system (Tribvn, Chatillon, France). The [3H]citalopram-binding-site densities are expressed as fmol/mg of tissue equivalent.

[3H]5-HT uptake in synaptosomes

Synaptosomes were prepared according to the method described by Whittaker with minor modifications (Whittaker, 1971). The hippocampus of mice were dissected at 4°C and homogenized in 1 ml ice-cold 0.32 m sucrose. Homogenates were diluted (1:2) in sucrose buffer and centrifuged at 1000 g for 10 min. The resulting pellets were resuspended in 1 ml of uptake buffer. Incubations in triplicate were started by adding 150 µl of synaptosome suspensions to uptake buffer [50 µl: 10 mm Hepes, 120 mm NaCl, 5.4 mm KCl, 1.2 mm CaCl2, 1.2 mm MgSO4, 1 mg/ml ascorbic acid, 5 mm d-glucose (pH 7.4)] containing 20 nm [3H]5-HT (116 Ci/mmol, Amersham) and increasing concentrations of unlabelled 5-HT (0, 5×10−8, 1×10−7 and 2.5×10−7 nm). Non-specific [3H]5-HT uptake was determined in the presence of 2 µm of the SSRI, citalopram (Lundbeck A/S, Copenhagen, Denmark). After 10 min incubation at 37°C, uptake was stopped by the addition of 3 ml ice-cold uptake buffer and immediate filtration through Whatman GF/B filters (presoaked in 0.05% polyethyleneimine). [3H]5-HT taken up was quantified by scintillation counting. Data are expressed as fmol of 5-HT taken up per mg of protein/min. The levels of protein in each sample were obtained from the optical density (OD) vs. concentration curve of bovine serum albumin diluted in Coomassie Blue.

8-OH-DPAT-induced hypothermia

The procedures used for these studies are based on those described by Bill et al. (1991). Body temperature was measured in gently restrained mice with a thermistor probe (M-99 electro-therm, Phymep, Paris, France) inserted 2 cm into the rectum. Basal body temperatures were measured every 15 min, starting half an hour before either vehicle (NaCl) or 8-OH-DPAT injection [0.5 and 1 mg/kg, subcutaneously (s.c.)]. The hypothermic response to 8-OH-DPAT was measured over a period of 2 h. The hypothermic response to 8-OH-DPAT was measured as the maximal decrease in body temperature recorded in the latter period.


Immediately after removal from the skull, brains were immersed in an ice-cold buffer continuously bubbled with carbogen (95% O2/5% CO2) to maintain the pH at 7.3. A block of tissue containing the DRN was cut into sections (400 µm thick) in the aCSF using a vibratome. Brainstem slices were immediately immersed in oxygenated aCSF at room temperature (22°C). A single slice was then placed on a nylon mesh, completely submerged in the recording chamber and continuously superfused with oxygenated aCSF (34°C) at a constant flow rate of 2–3 ml/min.

Extracellular recordings of the firing of DRN serotonergic neurons were made using glass microelectrodes filled with 2 m NaCl (12–15 MΩ). Cells were identified as 5-HT neurons according to previously described criteria (Lanfumey et al., 1999). Firing was evoked in the otherwise silent neurons by adding the α1 adrenoceptor agonist phenylephrine (3 µm) into the superfusing aCSF.

Baseline activity was recorded for at least 10 min prior to the perfusion of drugs. The effects of the 5-HT1A receptor agonist ipsapirone were evaluated by comparing the mean discharge frequency during the 2 min prior to its application with that recorded at the peak action of the drug, i.e. 2–3 min after its removal from the perfusing aCSF. Data are expressed as percentages of the baseline firing rates. Nonlinear regression was carried out using Prism 4.0 software (GraphPad Software Inc., San Diego, CA, USA) for the calculation of EC50 values of the drugs.

Statistical analysis

All data are reported as mean±s.e.m. Microdialysis data for basal dialysate levels corrected or not for recovery are reported in nm or fmol per fraction, respectively. Following linear regression of the data for each animal in the zero net flux experiments, unpaired two-tailed Student's t tests were used to assess the effects of genotype on [5-HT]ext levels in the hippocampus and Ed. For conventional microdialysis experiments, statistical analyses were performed on the area under the curve (AUC) values for the amount of 5-HT outflow collected during the 0–120 min period (in the ventral hippocampus and frontal cortex) or during the 0–180 min period (in the DRN) after the administration of vehicle (i.p.) or paroxetine. To compare different AUC values in each group of mice, a one-way ANOVA with treatment factor (NaCl, paroxetine 4 or 8 mg/kg) followed by Fisher's protected least significance difference (PLSD) post-hoc test was conducted. In addition, 5-HT levels in the ventral hippocampus, frontal cortex and DRN across all animals involved in conventional microdialysis studies have been compared by using Student's t test for differences between genotypes. For the 5-HT uptake experiments, results were analysed by nonlinear regression and the uptake capacity (Vmax) and Km of [3H]5-HT were calculated. Two-way ANOVA was then performed with 5-HT concentration as a within-subject variable and genotype as a between-subject variable. For the autoradiography study, the OD of selected brain region was measured and converted into fmol/mg tissue using the standard curve. Non-specific binding was substracted from total binding to evaluate specific binding in each brain region of each animal. Measurements were made on three sections from each brain region, and the values were averaged for each animal. The values for each region for each animal were then analysed by Student's t test for differences between genotypes. Maximal decreases in body temperature were analysed by a two-way ANOVA, with drug treatment and genotype as the independent variables, followed by Fisher's PLSD post-hoc test. Finally, extracellular recordings were analysed by two-way ANOVA followed by Bonferroni's post-hoc t test, and Student's t test to compare the experimental groups with their controls. For all data, significance level was set at p⩽0.05. All analyses were conducted using Statview 5.0 (JMP Software, Cary, NC, USA).


[5-HT]ext levels are increased in the ventral hippocampus in BDNF +/− mice

The method of zero net flux was used to evaluate basal [5-HT]ext levels in the ventral hippocampus of BDNF +/− and BDNF +/+ mice. Here, [5-HT]ext levels corrected for in-vivo recovery were significantly higher in BDNF +/− mice (1.56±0.4 nm, n=10) compared to wild-type mice (0.57±0.2 nm, n=12, p=0.012; Figure 1a, b). Thus, constitutive deletion of a single copy of the BDNF gene is associated with an increase in 5-HT levels in the ventral hippocampus and this effect may reflect either increase in hippocampal 5-HT release and/or decrease in 5-HT uptake.

Figure. 1

Zero net flux analysis of 5-HT levels in the ventral hippocampus of BDNF +/+ vs. BDNF +/− mice. The plots in (a) show the mean±s.e.m. gain or loss of 5-HT (CinCout) as a function of Cin (0, 5, 10, 20 nm of 5-HT) and the average linear regression of the data in BDNF +/− mice (–––) and BDNF +/+ mice (- - -). The Cin at which CinCout=0 equals the extracellular 5-HT levels ([5-HT]ext), and the slope of linear regression corresponds to the extracellular fraction of the probe (Ed). Statistically significant differences were observed between the two genotypes studied regarding (b) mean±s.e.m. of [5-HT]ext; and (c) mean±s.e.m. of the slope Ed. * p<0.05 compared to wild-type controls (n=10–12 mice per group).

Previous studies have shown that manipulations that decrease neurotransmitter uptake also decrease the recovery of neurotransmitter from the tissue as reflected in the extraction fraction, Ed (Parsons et al., 1991). Here, Ed was significantly lower in BDNF +/− mice (0.32±0.05, n=10) compared to wild-type mice (0.49±0.04, n=15, p=0.02; Figure 1a, c).

Conventional microdialysis data confirmed that constitutive decreases in BDNF expression produce an elevation of basal dialysate 5-HT concentrations (Table 1). In addition, by using the same method we observed that the 5-HT metabolite 5-hydroxyindoleacetic (5-HIAA) was significantly reduced in the ventral hippocampus in BDNF +/− mice (0.97±0.2 pmol/sample) compared to BDNF +/+ mice (1.8±0.2 pmol/sample, p=0.005). Together, these data suggest that increases in [5-HT]ext levels in BDNF +/− mice result from a decrease in 5-HT uptake at serotonergic nerve terminals located in the ventral hippocampus.

View this table:
Table 1

Paroxetine-induced increases in [5-HT]ext levels are blunted in the hippocampus in BDNF +/− mice

One-way ANOVA on AUC values for 5-HT outflow in the ventral hippocampus revealed a statistically significant effect of treatment factor in BDNF +/+ mice (F2,18=29.1, p<0.001), but not in BDNF +/− mice (F2,25=1.3, p=0.2). In the ventral hippocampus of BDNF +/+ mice, 4 and 8 mg/kg paroxetine increased [5-HT]ext levels compared to the corresponding group of wild-type mice treated with vehicle (p<0.001 for both doses) (Figure 2c).

Figure 2

Dose–response effects of paroxetine on 5-HT outflow in the ventral hippocampus of BDNF +/+ vs. BDNF +/− mice. Data are mean ± s.e.m. of extracellular 5-HT ([5-HT]ext) levels expressed as fmol/sample in (a) BDNF +/+ and (b) BDNF +/− mice following treatment with vehicle (≤), 4 mg/kg paroxetine (Prx) (Embedded Image) or 8 mg/kg paroxetine (▴). (c) Area under the curve values (AUC, mean±s.e.m.). □, Vehicle; Embedded Image, 4 mg/kg paroxetine; ■, 8 mg/kg paroxetine. *** p<0.001 relative to the corresponding vehicle-treated group (n=7–8 mice per group).

Interestingly, the statistical analysis also revealed a significant effect of treatment factor in the frontal cortex of BDNF +/+ and BDNF +/− mice (F1,13=4.1, p<0.05; F1,12=4.9, p<0.05, respectively) as well as in the DRN of both strains (F1,10=9.7, p<0.01; F1,8=8.7, p<0.01, respectively) (Figure 3c, f).

Figure 3

Effects of paroxetine on 5-HT outflow in the frontal cortex and dorsal raphe nucleus (DRN) of BDNF +/+ vs. BDNF +/− mice. Data are means ± s.e.m. of extracellular 5-HT levels expressed as fmol/sample in the frontal cortex (a, b) and the DRN (d, e) of BDNF +/+ and BDNF +/− mice following treatment with either the vehicle (⋄) or 8 mg/kg paroxetine (▴). (c, f) Area under the curve values (AUC, mean±s.e.m.). Embedded Image, 8 mg/kg paroxetine; □, vehicle. * p<0.05 and ** p<0.01 relative to the corresponding vehicle-treated group (n=5–6 mice per group).

[3H]Citalopram-binding sites are reduced in the hippocampus of BDNF +/− mice

To further explore the underlying causes of the decreases in extraction fraction in hippocampal 5-HT reported above for BDNF +/− mice, we performed autoradiography to determine SERT binding across a number of different brain regions. Examination of [3H]citalopram-binding-site densities revealed that the ventral hippocampus displayed a significant reduction in the number of [3H]citalopram-binding sites in BDNF +/− mutants compared to BDNF +/+ mice (Figure 4). In particular, a significant decrease in [3H]citalopram-binding sites was measured in the CA3 (p<0.01, Table 2), but not in the DG and CA1 (p>0.05, Table 2) subregions of the hippocampus in BDNF +/− mutants compared to BDNF +/+ mice. Moreover, no differences in the density of labelling were noted in the frontal cortex (p=0.71), striatum (p=0.24) and DRN (p=0.22) with respect to genotype (Table 2). These data show that constitutive BDNF down-regulation is associated with a decrease in SERT levels specifically in the ventral hippocampus.

Figure 4

Autoradiographic labelling of the serotonin transporter by [3H]citalopram in the ventral hippocampus. Representative coronal sections showing total hippocampal binding sites from wild-type BDNF +/+ mice and BDNF +/− mutant mice.

View this table:
Table 2

[3H]5-HT uptake in hippocampal synaptosomes from BDNF +/− mice is reduced

[3H]5-HT uptake by synaptosomes prepared from the hippocampus was decreased in BDNF +/− mice compared to BDNF +/+ mice (Figure 5). Constitutive reductions in BDNF affected Vmax (528±32 vs. 942±59 pmol/mg protein per min in BDNF +/− mice and BDNF +/+ mice, respectively, p=0.007), but was without significant effect on Km values for [3H]5-HT uptake (35±3 vs. 59±11 nm in BDNF +/− mice and BDNF +/+ mice, respectively; p>0.05).

Figure 5

[3H]5-HT uptake in hippocampal synaptosomes from BDNF +/+ (□) vs. BDNF +/− (▪) mice. Eadie–Hofstee analysis of the uptake of [3H]5-HT into hippocampal synaptosomes of BDNF +/+ mice and BDNF +/− mice. Non-specific [3H]5-HT uptake was determined in the presence of citalopram (n=5 mice per group).

Functional status of 5-HT1A autoreceptors is unaltered in BDNF +/− mice

We assessed presynaptic 5-HT1A receptor status by examining 8-OH-DPAT-induced changes in body temperature. Two-way ANOVA (genotype x treatment) of maximal decreases in body temperature revealed a statistically significant main effect of treatment (F2,33=12.89, p<0.001) but not genotype (F1,33=0.35, p=0.55). 8-OH-DPAT dose-dependently reduced the maximal decreases in body temperature in both genotypes compared with the corresponding group of mice treated with vehicle (−2.5±0.6°C, p>0.05 and −2±0.3°C, p>0.05 for BDNF +/+ and BDNF +/− mice, respectively after 0.5 mg/kg 8-OH-DPAT) and (−3.8±0.3°C, p<0.001 and −3.4±0.3°C, p<0.01 for BDNF +/+ and BDNF +/− mice, respectively for 1 mg/kg 8-OH-DPAT) (Figure 6a).

Figure 6

Functional status of 5-HT1A autoreceptors in BDNF +/− mice compared to wild-type littermates. (a) Body temperatures following either vehicle or 8-OH-DPAT injection (0.5 and 1 mg/kg, s.c.). The hypothermic response was measured as the maximum fall in body temperature following 8-OH-DPAT administration. (b) Integrated firing rate histograms (spikes/10 s) show the effects of increasing concentrations of the 5-HT1A receptor agonist ipsapirone on the electrical activity of a DRN 5-HT neuron brainstem slice from BDNF +/− and BDNF +/− mice. (c) Ipsapirone-induced inhibition is expressed as a percentage of the baseline firing rate. Each point is the mean±s.e.m. of data obtained from 8 to 10 individual cells. The vertical lines projecting onto the x-axis illustrate the EC50 values of ipsapirone in BDNF +/+ and BDNF +/− mice and in their littermates. No statistical differences were observed between genotypes (n=5 animals per group). *** p<0.001 and ** p<0.01 (n=10–12 mice per group); n.s., not statistically different.

Electrophysiological recording of the firing rates of dorsal raphe serotonergic neurons in brainstem slices indicated that the 5-HT1A receptor agonist ipsapirone (1–1000 nm) caused a significant concentration-dependent inhibition of firing in both BDNF +/+ and BDNF +/− mice (Figure 6b, c). Indeed, ipsapirone induced a significant decrease in both genotypes (F1,55=63.2, p<0.0001) by using a two-way ANOVA followed by a Bonferroni's t test. In addition, the genotype difference did not influence the response to ipsapirone (F1,55=3.6, p>0.05), i.e. a similar decrease in the firing rate occurred in both genotypes. The EC50 of ipsapirone to inhibit DRN firing was 53±10 nm in BDNF +/+ mice (n=9) and 73±10 nm in BDNF +/− mice (n=6) and these values were not significantly different (p>0.05).


The present study assessed whether constitutive deletion of one copy of the BDNF gene during development in male mice affects hippocampal 5-HT neurotransmission in adulthood. Our in-vivo and in-vitro approaches demonstrate that 3- to 4-month-old male BDNF +/− mice display a significant and selective elevation in basal hippocampal [5-HT]ext levels that appears to occur in response to reduced 5-HT reuptake. A reduction in SERT function and/or density in the adult hippocampus may account for alterations in 5-HT clearance from hippocampal nerve terminals leading to a blunted neurochemical response to the acute administration of paroxetine. These changes are not associated with a functional desensitization of 5-HT1A autoreceptors in the DRN since the capacity of 5-HT1A receptor agonists to decrease body temperature or DRN 5-HT neuronal activity was unchanged in BDNF +/− mice compared to wild-type littermates.

A prior study using zero net flux microdialysis demonstrated that changes in [5-HT]ext do not occur in the frontal cortex or striatum of adult BDNF +/− mice (Szapacs et al., 2004). Here, we used zero net flux microdialysis to investigate alterations in [5-HT]ext levels in the ventral hippocampus. This brain region is of particular interest because high levels of BDNF protein are produced in the hippocampus (Kolbeck et al., 1999). Furthermore, 5-HT reuptake-inhibiting antidepressants, as well as antidepressant drugs of other classes are thought to increase transcription of BDNF in the hippocampus as part of their therapeutic mechanisms (Duman and Monteggia, 2006; Nibuya et al., 1995). We found here that basal dialysate levels of 5-HT, which are not corrected for in-vivo recovery, as well as [5-HT]ext corrected for in-vivo recovery in the zero net flux method of quantitative microdialysis are both significantly increased in the ventral hippocampus of BDNF +/− mice. The in-vivo extraction fraction (Ed), which is a measure of in-vivo recovery and is sensitive to changes in transporter-mediated uptake was decreased in BDNF +/− mice suggesting that increases in [5-HT]ext in the ventral hippocampus may be a result of decreased 5-HT reuptake. Zero net flux microdialysis has already proven to be a useful technique for comparing the neurotransmitter reuptake capacity of presynaptic monoaminergic nerve terminals in wild-type vs. mutant mice for dopamine and 5-HT in 5-HT1B receptor knockout mice (Gardier et al., 2003; Shippenberg et al., 2000). 5-HT clearance in vivo is largely mediated by the SERT and faster clearance rates of 5-HT are related to greater densities of functional SERT (Montañez et al., 2002). Consequently, the present data suggest that the elevated hippocampal basal 5-HT levels occurring in response to constitutive partial deletion of the BDNF gene in the male mouse is related to a down-regulation in plasma membrane localization of the SERT and/or reduced its 5-HT reuptake capacity.

We also observed that dialysate 5-HIAA levels resulting from intracellular 5-HT enzymatic degradation by a type A monoamine oxidase (MAO) are decreased in the ventral hippocampus of BDNF +/− mice. Reductions in 5-HIAA may also reflect a reduced 5-HT reuptake capacity, the extracellular capacity of which was decreased in the striatum of mice with constitutive loss of SERT expression (Mathews et al. 2004). However, no changes in total 5-HIAA levels were reported previously in the hippocampus of BDNF +/− mice prior to 18 months of age (Chourbaji et al., 2004; Lyons et al., 1999; Ren-Patterson et al., 2005). The observation that 5-HIAA levels were also unaltered in the hippocampus of SERT +/− mice (Kim et al., 2005) suggests that brain tissue measurements of 5-HIAA levels (mainly intracellular content) do not necessarily predict changes in extracellular concentrations of the main metabolite of 5-HT. Adaptive changes in the activities of the key enzymes of 5-HT metabolism such as the tryptophan hydroxylase (TH) and/or the MAO could also occur in constitutive BDNF +/− mutant mice and explain the lack of decrease in tissue hippocampal 5-HIAA levels. To our knowledge, such studies have not yet been conducted in either constitutive or inducible knockout mice for the BDNF gene.

The hyperserotonergic phenotype reported herein should lead to changes in spontaneous behavioural responses. These animals develop enhanced inter-male aggressiveness and hyperphagia accompanied by significant weight gain in early adulthood; these behavioural abnormalities are known to correlate with 5-HT dysfunction (Lyons et al., 1999). However, constitutive BDNF +/− mice displayed normal levels of depression-like behaviour in the forced swimming test (FST) (Saarelainen et al., 2003; MacQueen et al., 2001). Since the above-mentioned behavioural studies included females, but did not directly compare the two sexes, it can be proposed that the apparent discrepancies between neurochemical and behavioural data are gender related. Nevertheless, in a recent report it was shown that male conditional BDNF KO −/− mice also display normal depression-related behaviours (Monteggia et al., 2007). The period, level and site of BDNF mutation (constitutive vs. conditional knockout mice) may be important factors influencing the neurochemical and antidepressant-like responses of these mutants to SSRIs and should hold our attention for future investigations.

To investigate the alteration of the SERT further, we conducted in-vivo microdialysis experiments to determine whether the neurochemical effects of a single systemic dose of the SSRI paroxetine are influenced by decreased BDNF expression. The effects of paroxetine to increase hippocampal [5-HT]ext levels by blocking SERT in BDNF +/− mice were blunted and these data are in agreement with decreased SERT function in these mice. Our results are consistent with recent in-vivo neurochemical data showing that male BDNF +/− mice at age 5 months are less sensitive to the effect of the SSRI fluvoxamine to block 5-HT uptake and to prolong the time-course of 5-HT clearance (Daws et al., 2007). This apparent decreased sensitivity of BDNF +/− mice to the SSRI was also reported in behavioural paradigms. Indeed, constitutive BDNF+/− mice are resistant to the acute response to imipramine (Saarelainen et al., 2003) in the FST. The brain region involved in serotonergic antidepressant drug efficacy has not yet been resolved. In inducible BDNF KO lines of mice, in which the BDNF gene was deleted in the hippocampus and cortex, an attenuated response to desipramine was also observed (Monteggia et al., 2004, 2007) confirming that normal forebrain BDNF levels are essential for antidepressant efficacy in mice (D'Sa and Duman, 2002).

To test further our hypothesis that 5-HT reuptake is reduced in BDNF +/− mice and to determine whether this alteration is associated with decreased activity and/or expression of SERT in the mouse hippocampus, we conducted two types of in-vitro experiments. In hippocampal synaptosome, the intensity of [3H]5-HT uptake (i.e. Vmax) was significantly decreased in BDNF +/− mice compared to their wild-type littermates. These results parallel recent in-vivo electrochemical data reporting a marked attenuation of exogenous 5-HT clearance rate in the hippocampus of 5-month-old BDNF +/− mice (Daws et al., 2007). Because Vmax value is a reliable index of the number of SERT molecules present in the synaptosomal preparation, the possibility that BDNF may control SERT expression has been raised. Autoradiography of [3H]citalopram binding indicates a reduced SERT density in the hippocampus of adult male BDNF +/− mice. Regional differences in SERT density between both strains of mice were restricted to the CA3 subregion of the ventral hippocampus. These results stand in contrast with the fact that [3H]cyano-imipramine binding to the SERT in CA3 of the hippocampus does not differ between BDNF +/− and BDNF +/+ mice (Daws et al., 2007). However, the binding properties of [3H]cyano-imipramine and [3H]citalopram differ (Leake et al., 1991). [3H]Cyano-imipramine does not bind directly to the 5-HT uptake site but binds with high affinity to a distinct but closely associated region. Moreover, [3H]cyano-imipramine is known as a lipophilic agent labelling not only surface membrane SERT but also transporters translocated to the intracellular side of the membrane. The possibility that such an effect occurred in the present study with [3H]citalopram cannot be ruled out and would explain, at least in part, why we did not detect binding differences in other subregions of the ventral hippocampus such as CA1 and dentate gyrus. Indeed, it is somewhat unexpected that a decrease in SERT density in the CA3 subregion accounts for a reduced in-vivo and in-vitro 5-HT reuptake throughout the whole hippocampus. Thus, it appears more reasonable to assume that an attenuation of SERT function rather than expression, has participated in the reduction of 5-HT transport to increase local 5-HT neurotransmission. Interestingly, [3H]citalopram binding was unaltered in the frontal cortex, striatum and DRN of BDNF +/− mice compared with wild-type controls. This is consistent with our neurochemical data showing that the basal [5-HT]ext levels and paroxetine-induced increase in dialysate 5-HT levels in both the frontal cortex and DRN were not affected by the partial loss of BDNF. Nevertheless, the underlying reason for this regional specificity of action of BDNF could involve attenuated sprouting and growth of 5-HT nerve terminals in the hippocampus, but not the serotonergic cell bodies and dendrites located in the DRN. In line with this hypothesis, it has been reported that infusion of BDNF directly into the cortex increases 5-HT axon density in the rat brain (Mamounas et al., 1995), as well as the extension and ramification of serotonergic neurites (Rumajogee et al., 2002). However, decreased hippocampal serotonergic innervation was only reported to occur in 12-month-old BDNF +/− mice (Lyons et al., 1999).

The present data support the conclusion that serotonergic neurotransmission of adult BDNF +/− mice is selectively increased in the ventral hippocampus by a mechanism involving decreased 5-HT reuptake. This phenomenon is in agreement with the finding that post-synaptic 5-HT1A receptor-mediated [35S]GTP-γ-S binding is attenuated in the hippocampus of BDNF +/− mice (Hensler et al., 2003). Moreover, a body of literature indicates that conditions that decrease hippocampal BDNF expression, such as chronic social stress (Aleisa et al., 2006) or chronic restraint stress (Murakami et al., 2005; Smith et al., 1995), also produce desensitization or down-regulation of hippocampal post-synaptic 5-HT1A receptors. Thus, early neurochemical and behavioural changes in BDNF +/− mice may result from the effects of prolonged increases in 5-HT concentrations in extracellular space, specifically in the hippocampus. This is somewhat different from what occurs in SERT-deficient mice (KO SERT −/−) in which [5-HT]ext levels increased in many brain regions (Mathews, et al. 2004).

In the DRN, somatodendritic 5-HT1A receptors are involved in serotonergic modulation of their own neuronal activity via a negative feedback autoregulatory mechanism (Guilloux et al., 2006). Since increases in [5-HT]ext levels detected in SERT −/− mice have consistently been associated with presynaptic 5-HT1A receptor desensitization (Fabre et al., 2000; Gobbi et al., 2001; Mathews et al., 2004), we also tested an alternate hypothesis that 5-HT1A autoreceptors exert reduced inhibitory control of 5-HT release, thus leading to a hyperserotonergic tone in the hippocampus of BDNF +/− mice. Here, two different techniques have been used to detect 5-HT1A autoreceptor desensitization – determination of changes in body temperature (Gardier et al., 2001; Trillat et al., 1998) and recordings of DRN 5-HT neuronal activity (Froger et al., 2001) – in response to a 5-HT1A receptor agonist. Evidence that body temperature is mediated by the presynaptic population of this receptor in mice was repeatedly demonstrated (Bill et al., 1991; Gross et al., 2002). Thus, in the present study, we report that the 5-HT1A receptor agonist 8-OH-DPAT produced a dose-dependent hypothermic response in BDNF +/+ control mice, while this effect was not altered in BDNF +/− mice. Since 8-OH-DPAT-induced hypothermia in rodents could be mediated by both 5-HT1A and 5-HT7 receptors (Hedlund et al., 2004), we also assessed the functional status of 5-HT1A autoreceptors by using an electrophysiological approach. Data from these experiments did not reveal alterations in 5-HT1A autoreceptor sensitivity since the capacity of ipsapirone to inhibit DRN 5-HT neuronal activity was not different between BDNF +/− and BDNF +/+ mice. The lack of desensitization of presynaptic 5-HT1A autoreceptors is also supported by the findings that BDNF +/− mice do not have alterations in stimulated 5-HT neurotransmission in the DRN. Indeed, it is well known that the functional desensitization of DRN 5-HT1A receptors in mice can potentiate SSRI-induced changes in dialysate 5-HT at the somotodendritic level (Bortolozzi et al., 2004; Guilloux et al., 2006). These results further specify the hyperserotonergic effects of constitutive reductions in BDNF expression to the hippocampus and are in agreement with those of Hensler et al. (2003) that show no decreases in 5-HT1A autoreceptor-stimulated [35S]GTP-γ-S binding in the DRN and no changes in the number of 5-HT1A receptors in BDNF +/− mice compared to wild-type controls.

The present study demonstrates that a 50% reduction in BDNF levels is sufficient to cause measurable increases in basal [5-HT]ext levels specifically in the ventral hippocampus and that these changes are related to a reduced 5-HT reuptake occurring in the same brain region. Human gene variants occurring at the SERT gene loci were implicated in susceptibility to mood disorders (Caspi et al., 2003) and altered responses to serotonergic antidepressant drugs (Serretti et al., 2005). A BDNF gene single nucleotide polymorphism was also recently described (Egan et al., 2003) that reduces BDNF secretion and hippocampal activity. In addition, a knock-in mouse was generated (BDNFMet/Met) that reproduces some of the phenotypic hallmarks present in humans expressing the equivalent polymorphism including an increase in anxiety and a decreased response to the SSRI fluoxetine (Chen et al., 2006). In conclusion, the results of the present study clearly indicate a relationship between constitutive alterations in BDNF levels and the expression and function of the SERT that are unique to the adult hippocampus. Further studies will be aimed at investigating the molecular mechanisms by which these two neural systems interact to produce these effects.


During the performance of this work, B. P. Guiard was the recipient of a fellowship from La Fondation pour la Recherche Médicale (FRM). We are grateful to GlaxoSmithKline for the generous gifts of paroxetine.

Statement of Interest



View Abstract