Chronic Fluoxetine Treatment Induces Maturation-Compatible Changes in the Dendritic Arbor and in Synaptic Responses in the Auditory Cortex

Abstract

Fluoxetine is a selective serotonin reuptake inhibitor (SSRI) used to treat mood and anxiety disorders. Chronic treatment with this antidepressant drug is thought to favor functional recovery by promoting structural and molecular changes in several forebrain areas. At the synaptic level, chronic fluoxetine induces an increased size and density of dendritic spines and an increased ratio of GluN2A over GluN2B N-methyl-D-aspartate (NMDA) receptor subunits. The "maturation"-promoting molecular changes observed after chronic fluoxetine should also induce structural remodeling of the neuronal dendritic arbor and changes in the synaptic responses. We treated adult rats with fluoxetine (0.7 mg/kg i.p. for 28 days) and performed a morphometric analysis using Golgi stain in limbic and nonlimbic cortical areas. Then, we focused especially on the auditory cortex, where we evaluated the dendritic morphology of pyramidal neurons using a 3-dimensional reconstruction of neurons expressing mRFP after in utero electroporation. With both methodologies, a shortening and decreased complexity of the dendritic arbors was observed, which is compatible with an increased GluN2A over GluN2B ratio. Recordings of extracellular excitatory postsynaptic potentials in the auditory cortex revealed an increased synaptic response after fluoxetine and were consistent with an enrichment of GluN2A-containing NMDA receptors. Our results confirm that fluoxetine favors maturation and refinement of extensive cortical networks, including the auditory cortex. The fluoxetine-induced receptor switch may decrease GluN2B-dependent toxicity and thus could be applied in the future to treat neurodegenerative brain disorders characterized by glutamate toxicity and/or by an aberrant network connectivity.

Keywords: NMDA receptors; antidepressant; auditory cortex; dendritic architecture; neuronal segmentation.

Figure 1

Chronic fluoxetine treatment (28 days) decreased the dendritic length of pyramidal neurons in limbic cortices. Golgi staining was performed to analyze the prelimbic (PrL) and retrosplenial granular b (RSGb) cortices. Neurons were drawn with camera lucida and digitally scanned to visualize the morphology. (A) and (D) The analyzed areas are schematically shown (Paxinos and Watson, 1998). (B) and (E) Representative drawing of pyramidal neurons obtained from vehicle treated (control) or fluoxetine-treated rats. In the PrL cortex, layer II–III and V neurons were analyzed, while in the RSGb cortex, the analysis was restricted to layer V neurons. In each case, the neuron at the left is from the control (vehicle treated), while the neuron at the right is from the fluoxetine-treated animal. Scale bar: 20 µm. (C) and (F) Quantifications of the total, apical, and basal dendritic lengths of neurons in the PrL cortex (layers II/III in the left panel and layer V in the right panel) and RSGb cortices. Values represent the mean ± SEM. For each condition, we analyzed at least 10 neurons obtained from four animals per condition for the PrL cortex and three animals per condition for the RSGb cortex. *p ≤ 0.05, Mann–Whitney U-test.

Figure 2

Chronic fluoxetine treatment (28 days) decreased the dendritic length of pyramidal neurons in the motor and auditory cortices (layer II–III). Golgi-stained neurons of the II–III layer of the secondary motor cortex (M2) and the primary auditory cortex (AUD1) were drawn with camera lucida and digitally scanned. (A) and (D) Scheme of the analyzed areas (Paxinos and Watson, 1998). (B) and (E) Representative drawings of neurons obtained from vehicle or fluoxetine-treated animals. In each case, the neuron at the left is from the control (vehicle-treated), while the neuron at the right is from the fluoxetine-treated animal. Scale bar: 20 µm. (C) and (F) Quantifications of the total, apical, and basal dendritic lengths. For each condition, we analyzed at least 10 neurons obtained from three animals per group (in the case of M2) and from three (control) and four (fluoxetine-treated) animals in the case of AUD1. Values represent mean ± SEM. *p ≤ 0.05; **p < 0.01 ***p < 0.001 Mann–Whitney U-test.

Figure 3

Repetitive flx treatment decreases the dendritic length of pyramidal neurons in the primary auditory cortex of rats. (A) Schematic representation of in utero electroporation to target gene delivery into neurons in the auditory cortex. The inset shows mRFP electroporated pyramidal neurons selected for analysis in layer V of the primary auditory cortex in coronal slices of 150 µm. (B) The 3-dimensional reconstruction of confocal images is shown for isolated neurons obtained from vehicle and fluoxetine-treated animals. Upper panel: raw image; middle panel: skeletonized form; and lower panel: the reduced skeletonized form. (C) Quantifications (mean ± SEM) of the total, apical, and basal dendrites, measured in the skeletonized form of neurons after 3-dimensional reconstruction (left grouped bars) or measured after Golgi staining of the same area (right grouped bars). (D) Quantification of the number of bifurcations according to increasing orders of processes. Value represent mean ± SEM. For 3-dimensional reconstruction, eight control neurons obtained from three animals and seven neurons obtained from three animals were analyzed. *p ≤ 0.05; **p < 0.01 Mann–Whitney U-test.

Figure 4

Chronic fluoxetine treatment modifies the postsynaptic response in the auditory cortex of rats. Different parameters of the fEPSPs were measured after stimulation in layer I and recording in layer II–III: (A) Representative recordings in slices from vehicle or fluoxetine-treated animals (left) followed by the NMDA-mediated response (in the presence of CNQX to block AMPA receptors) (middle) and of the GluN2A-mediated response (in the presence of CNQX + ifenprodil to block AMPA and GluN2B-containing receptors) (right). (B) Peak fEPSPs, (C) rise time (or the time at which the 67% of the peak potential is achieved), (D) total postsynaptic potential (area under the curve), and (E) percentage of blockade of the total fEPSP by CNQX (NMDA component) or by CNQX + ifenprodil (GluN2A component). Slices were perfused with Mg-free aCSF (control recording) or with aCSF in the presence of CNQX and then in the presence of CNQX + ifenprodil. For each condition, at least 32 slices were recorded, obtained from 16 animals (n = 8 control and n = 8 fluoxetine). Figures show the mean ± SEM. One way-ANOVA followed by post hoc Bonferroni test was used. *p ≤ 0.05; **p < 0.01 relative to the control within each drug condition (i.e., within each pair of red vs. white bar). For indicated comparisons among groups, unpaired t-tests were used ^# p ≤ 0.05; ^## p < 0.01, ^### p < 0.001 relative to the first bar (aCSF-control without drug addition).