Long-term fluoxetine treatment induces input-specific LTP and LTD impairment and structural plasticity in the CA1 hippocampal subfield

Abstract

Antidepressant drugs are usually administered for several weeks for the treatment of major depressive disorder. However, they are also prescribed in several additional psychiatric conditions as well as during long-term maintenance treatments. Antidepressants induce adaptive changes in several forebrain structures which include modifications at glutamatergic synapses. We recently found that repetitive administration of the selective serotonin reuptake inhibitor (SSRI) fluoxetine to naïve adult male rats induced an increase of mature, mushroom-type dendritic spines in several forebrain regions. This was associated with an increase of GluA2-containing α-amino-3-hydroxy-5-methylisoxazole-4-propionate receptors (AMPA-Rs) in telencephalic postsynaptic densities. To unravel the functional significance of such a synaptic re-arrangement, we focused on glutamate neurotransmission in the hippocampus. We evaluated the effect of four weeks of 0.7 mg/kg fluoxetine on long-term potentiation (LTP) and long-term depression (LTD) in the CA1 hippocampal subfield. Recordings in hippocampal slices revealed profound deficits in LTP and LTD at Schaffer collateral-CA1 synapses associated to increased spine density and enhanced presence of mushroom-type spines, as revealed by Golgi staining. However, the same treatment had neither an effect on spine morphology, nor on LTP and LTD at perforant path-CA1 synapses. Cobalt staining and immunohistochemical experiments revealed decreased AMPA-R Ca(2+) permeability in the stratum radiatum (s.r.) together with increased GluA2-containing Ca(2+) impermeable AMPA-Rs. Therefore, 4 weeks of fluoxetine treatment promoted structural and functional adaptations in CA1 neurons in a pathway-specific manner that were selectively associated with impairment of activity-dependent plasticity at Schaffer collateral-CA1 synapses.

Keywords: LTD; LTP; antidepressants; dendritic spines; glutamate receptors.

Figure 1

Long-term fluoxetine treatment induces morphological changes in proximal dendritic spines of CA1 pyramidal neurons. (A) Left: A representative image of a Golgi stained neuron from a fluoxetine-treated rat. CA1 stratum radiatum (s.r.) subfield was chosen for analysis. Scale bar: 50 μm. Right: representative dendritic segments of saline and fluoxetine-treated rats under larger amplification. Scale bar: 10 μm. (B) The selected dendritic segments in (A) (right panels) are shown with identified spine types: filopodia/thin (T), stubby (S), and mushroom/branched (M). The asterisks indicate two superimposed M-type spines that were resolved by observation at different focal planes. Bar graphs show spine density and the abundance of spines in each of the three shape categories. Results are presented as mean ± SEM and were determined from 11 cells per condition, obtained from eight saline (total spines, 874)- and nine fluoxetine (total spines, 957)-treated rats. Data were statistically evaluated with the Mann–Whitney U-test, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 2

Long-term fluoxetine treatment does not affect distal dendritic spine morphology of CA1 pyramidal neurons. (A) Left: Golgi stained CA1 pyramidal neuron indicating the stratum lacunosum moleculare (s.l.m.) where the morphological analysis was performed. ^#At this level can be observed granule neuron dendrites coming from dentate gyrus. Scale bar: 100 μm. Right: representative dendritic segments of saline and fluoxetine-treated rats under larger amplification. Scale bar: 10 μm. (B) The number of spines and the morphology of spines were analyzed and plotted as spine density and % of spine type, respectively. Filopodia/thin (T), stubby (S), and mushroom/branched (M) shape categories are shown. Results are presented as mean ± SEM and were determined from 10 to 11 cells per condition, obtained from five saline (total spines, 979)- and six fluoxetine (total spines, 942)-treated rats. Data were not statistically significant after a Mann–Whitney U-test analysis.

Figure 3

Four weeks of fluoxetine administration affects basal synaptic transmission and synaptic plasticity at CA3-CA1 synapses. Experiments were performed in slices from saline (sal), single (flx 24 h), and repeated (flx 4 weeks) fluoxetine treatment groups. (A) The diagram shows the stimulating electrode inserted into Schaffer collaterals (SC) and the recording electrode placed in the CA1 s.r. (B) Input/output curves. The fEPSP slopes vs. increasing presynaptic fiber volley (FV) amplitudes were plotted. Data were fit to a linear regression and compared using covariate analysis. The slope obtained in repetitive fluoxetine-treated rats (0.48 ± 0.05, r² = 0.99) was significantly different (p < 0.01) compared to saline-treated rats (0.23 ± 0.04, r² = 0.99) or to rats treated with fluoxetine only once (0.23 ± 0.05, r² = 0.97). Saline, n = 6 (17 slices); fluoxetine 24 h, n = 6 (14 slices); fluoxetine 4 weeks, n = 8 (20 slices). (C) Representative traces of extracellular field recordings are shown before and after LTP (upper panels) or LTD (lower panels) induction. (D) Long-term potentiation (LTP) induced by theta burst stimulation (TBS) and (E) long-term depression (LTD) induced by low frequency stimulation (LFS) were blocked in hippocampal slices of rats treated for 4 weeks with fluoxetine. Bar graphs represent the mean % change in fEPSP slopes compared to baseline, from 45 to 60 min after LTP (D inset) or LTD induction (E inset). LTP: sal, n = 4 (14 slices); flx 4 weeks, n = 6 (14 slices); flx 24 h, n = 6 (12 slices). LTD: sal, n = 4 (12 slices); flx 4 weeks, n = 5 (15 slices); flx 24 h, n = 4 (9 slices). Data from D and E were statistically evaluated by one-way ANOVA, followed by a Bonferroni post hoc test. ^**p < 0.01, ^***p < 0.001.

Figure 4

Four weeks of fluoxetine administration does not affect synaptic plasticity at perforant pathway-CA1 synapses. Experiments were performed in slices from saline- (sal) and fluoxetine- (flx 4 weeks) treated rats. (A) The diagram shows the stimulating electrode inserted into the temporoammonic pathway (TA) and the recording electrode in the CA1 s.l.m. EC, entorhinal cortex. The dashed line indicates the place of the physical de-afferentation of Schaffer collateral inputs to CA1. (B) The fEPSP slopes vs. increasing presynaptic fiber volley (FV) amplitudes were plotted to analyze the input/output curves. No significant difference between saline- and repetitive fluoxetine-treated rats was found after data were fit to a linear regression and compared using covariate analysis. Saline, n = 5 (12 slices); fluoxetine 4 weeks, n = 8 (19 slices). (C,D) Sample traces of extracellular field recordings are shown before and after LTP (E) or LTD (F) induction. (E) Long-term potentiation (LTP) induced by theta burst stimulation (TBS) were unaffected after perforant pathway stimulation. No significant difference was found when the mean fEPSP integrated between 45 and 60 min after LTP induction was compared (p = 0.143; unpaired t-test). LTP: sal, n = 3 (8 slices); flx 4 weeks, n = 3 (6 slices). (F) Similar LTD was also induced by low frequency stimulation (LFS) of the perforant path in slices from saline- and fluoxetine-treated animals (p = 0.519; unpaired t-test).

Figure 5

GluA1 and GluA2 were detected in the CA1 region by immunofluorescence and immunohistochemical staining. (A) Fluorescence images of GluA1 and GluA2 immunostainings reveal a change in the distribution of subunits between the cell soma (s.p.) and dendritic (s.r.) compartments. (B) Relative staining intensity (s.r./s.p.) of immunohistochemical GluA1 and GluA2 stainings revealed higher GluA2 intensity in the s.r. of fluoxetine-treated rats (^*p < 0.05, n = 5−7, Student t-test). s.r., stratum radiatum; s.p., stratum pyramidale. Scale bars: 50 μm.

Figure 6

Long-term fluoxetine administration decreased Ca²⁺-permeable AMPA-Rs in the CA1 stratum radiatum (s.r.). Hippocampal slices from saline- (sal) and fluoxetine- (flx) treated rats were stimulated with 100 μM quisqualic acid (QA) to detect Co²⁺ uptake. The AMPA-R blocker CNQX (20 μM) was used as a negative control (QA+CNQX). (A) Hippocampal sections are shown in each experimental condition. Bar: 500 μm. (B) Increased magnification was used to quantify relative staining in the two CA1 subfields: stratum pyramidale (s.p.) and s.r. Bar: 50 μm. Bar graphs: the mean optical density was plotted. Data were statistically evaluated with the Mann–Whitney U-test, ^*p < 0.05, ^***p < 0.001. (n = 9 group).