5-HTR_2A and 5-HTR_3A but not 5-HTR_1A antagonism impairs the cross-modal reactivation of deprived visual cortex in adulthood

Animals

In total 54 C57Bl/6 J mice (Janvier Elevage, Le Genest-St-Isle, France) of either sex (32 male/22 female) were used in this study. All mice were housed under standard laboratory conditions with constant room temperature and humidity, an 10/14-h dark/light cycle with food and water available ad libitum. All experiments have been approved by the Ethical Research Committee of KU Leuven and were in strict accordance with the European Communities Council Directive of 22 September 2010 (2010/63/EU) and with the Belgian legislation (KB of 29 May 2013). Every effort was made to minimize animal suffering and to reduce the number of animals. Figure 1 illustrates the experimental manipulations and the number of mice used per condition (Fig. 1b, d). The different phases of cortical plasticity under study have been determined previously based on the impact of either visual stimulation via the spared eye, or somatosensory deprivation/stimulation based on whisker clipping/natural whisker use during the exploration of new toys in complete darkness, on neuronal activity in the visual cortex of adult ME mice [15]. Specifically, 1 week post-ME (1wME) mice are in an ongoing unimodal open-eye potentiation phase. Mice with a 3 week post-ME recovery period (3wME) are at the end of the open-eye potentiation phase, which restores normal visually driven activity levels in an extended binocular zone (Bz). 5 weeks post-ME (5wME) mice are in an ongoing cross-modal phase whereas mice with a 7 week post-ME recovery period (7wME) have undergone maximal cross-modal visual cortex reactivation in which normal activity levels are restored in the monocular zone of the visual cortex, especially medial to the Bz (Mmz), only now relying on whisker inputs. Cortical regions of interest therefore are the visual cortex, Bz and Mmz, and the primary somatosensory barrel field (S1BF).

Monocular enucleation paradigm and tissue preparation

The removal of the right eye, or monocular enucleation (ME), was performed as described previously [14]. Briefly, adult (P120) mice were anaesthetized by intraperitoneal injection of a mixture of ketamine hydrochloride (75 mg/kg, Dechra Veterinary Products, Eurovet) and medetomidine hydrochloride (1 mg/kg Orion Corporation, Janssen Animal Health). Eye ointment (Tobrex, Alcon) was administered to the left eye to prevent dehydration. The right eye was carefully removed and the orbit was filled with hemostatic cotton wool (Qualiphar, Bornem, Belgium) in case of bleeding. Analgesics were injected subcutaneously (Metacam, 2 mg/kg, 0.1 mL) and atipamezol hydrochloride was administered to reverse anaesthesia (1 mg/kg i.p., Orion Corporation, Elanco Animal Health). Following ME, the mice were housed in their home cages under standard laboratory conditions for a 1 to 7 week recovery period. Control mice are age-matched (AMC) to the 7 week ME mice, the time point of maximal recovery of neuronal activity [56]. At the end of the ME period, the mice were sacrificed by cervical dislocation. For HPLC-based determination of the total 5-HT concentration in visual and somatosensory tissue samples, the brains were rapidly extracted and immediately frozen in 2-methylbutane (Merck, Overijse, Belgium) at a temperature of − 40 °C. All brains were stored at − 80 °C until sectioning. A tissue blotting device (Model PA 002 Mouse Brain Blocker, 1 mm, David KOPF instruments, California) was used to prepare approximately 1 mm-thick coronal slices to subsequently isolate the whole visual cortex and the primary somatosensory cortex (S1BF) with sterile scalpels. Separate analysis of medial monocular and binocular cortex was technically impossible with this method. For Western blotting experiments, 100 μm-thick coronal cryosections were collected on baked glass slides and stored at − 80 °C. Specifically for radioactive in situ hybridization (ISH) experiments, the mice were placed overnight in their home cages in a dark room to reduce zif268-expression to basal levels. The following day, the mice were placed in a high-lit environment to upregulate sensory driven zif268-mRNA expression. After 45 min, at peak zif268-mRNA expression levels, the mice were sacrificed by cervical dislocation. The brains were rapidly removed, immersed in 2-methylbutane at a temperature of − 40 °C (Merck, Overijse, Belgium) and stored at − 80 °C until sectioning. Serial sections with a thickness of 25 μm were prepared on a cryostat (HM 500 OM, Microm, Thermo Scientific, Walldorf, Germany), mounted on 0.1% poly-L-Lysine-coated (Sigma-Aldrich) slides, and stored at − 20 °C until further processing.

Quantification of the total serotonin content in mouse brain homogenates

The total serotonin (5-HT) content in the visual and barrel cortex (HPLC total n = 15, AMC: n = 9; 7wME: n = 6) was measured based on previously reported methods [57–59]. In summary, after weighing cortical tissue, 190 μL of an antioxidant solution (0.1 M acetic acid, 3.3 mM L-cysteine, 0.27 mM Na₂EDTA and 0.0125 mM ascorbic acid) and 10 μL of an internal standard solution (3,4-dihydroxybenzylamine solution 1 μg/mL in antioxidant) were added to the tissue. After homogenization, the samples were centrifuged (20 min, 9500 g, 4 °C). The supernatant was diluted 5-fold in 0.5 M acetic acid and 20 μL was injected automatically on a reversed phase liquid chromatography system (autosampler ASI-100 and HPLC pump P680 A HPG/2, Dionex, Amsterdam, The Netherlands) with electrochemical detection (potential = + 700 mV) (Amperometric Detector LC-4C, BAS, Indiana, USA). With this liquid chromatography method, we are able to measure within one run the monoamines noradrenaline, dopamine and 5-HT, some of their major metabolites (such as 3,4-dihydroxyphenylacetic acid; 5-hydroxy-indoleacetic acid; homovanillic acid) as well as the internal standard 3,4-dihydroxybenzylamine (Fig. 2). The separation between the different compounds was achieved using a narrowbore C18 column (Alltech®, Alltima™, 5 μm, 150 × 2.1 mm, Grace, Deerfield, IL, USA). The mobile phase buffer contained 0.1 M sodium acetate, 20 mM citric acid, 1 mM sodium octane sulfonic acid, 1 mM dibutylamine and 0.1 mM Na₂EDTA adjusted to pH 3.7 (mobile phase composition: 97 buffer / 3 methanol (v/v)). For this study, we specifically quantified the 5-HT content of the different samples. Tissue concentration was expressed as ng 5-HT /g wet tissue (ng/g).

Western analysis

Western blotting (WB) was used to investigate the changes in relative expression of pre- and postsynaptic proteins involved in serotonergic neurotransmission. Time-course samples were prepared to separately examine the protein expression in the Mmz, Bz and S1BF over a 7-week period. The experimental conditions included: 1, 3, 5, 7 weeks post-ME. For each of the experimental conditions, as well as for the control mice age-matched to the 7wME mice (AMC), at least 4 mice were included (Figs. 3 and 4, WB total n = 24, AMC: n = 4; 1wME: n = 5; 3wME: n = 5; 5wME: n = 5; 7wME: n = 5). All ME samples were analyzed relative to the AMC samples (absolute OD-values measured in the Mmz, Bz or S1BF of the AMC as mean ± SEM for vMAT2 (Mmz: 2.824 ± 0.483; Bz: 0.691 ± 0.044; S1BF: 0.953 ± 0.130), 5-HTR_1A (Mmz: 0.855 ± 0.126; Bz: 0.667 ± 0.076; S1BF: 0.700 ± 0.073) and 5-HTR_3A (Mmz: 0.955 ± 0.087; Bz: 0.854 ± 0.138; S1BF: 0.764 ± 0.273)). We chose to analyze vMAT2 protein expression levels since they provide information on the presynaptic loading of vesicles with 5-HT and because vMAT2 expression levels positively correlate with the total 5-HT concentration present in a given brain region of interest [60–62].

Protein extraction from tissue slices

Based on the mouse brain atlas of Paxinos and Franklin (2013) [63], the primary somatosensory barrel field (S1BF: 0.5-(− 2); 3–4.5; 1, relative to Bregma in A-P; M-L; D), the medial monocular zone (Mmz, comprising monocular V1 and V2M: − 2.7-(− 4.7); 1–2.5; 1, relative to Bregma and the binocular zone (Bz, comprising binocular V2L and V1: − 2.7-(− 4.7); 2.5–4; 1, relative to Bregma) were microscopically excised and collected separately from 100 μm-thick coronal cryosections (Fig. 1c). Tissue was collected in a mix of 4 μL of complete protease inhibitor cocktail (Roche Diagnostics, GmbH) and 100 μL ice-cold lysis buffer (2% SDS, 65 mM Tris-HCl in MQ) optimized for the enrichment of membrane (−associated) proteins [64, 65]. Proteins were extracted from the tissue by mechanical homogenization using drill-driven, sterile disposable pestles (Argos Technologies), sonication (5 × 10 s), incubation at 70 °C (5 min) and centrifugation (15 min, 13000 rpm, 4 °C). The supernatant was collected and the total protein concentration was determined using the Qubit fluorometer (Invitrogen). Samples were stored at − 80 °C.

Immunoblotting

To obtain the optimal primary antibody working concentration, a protein dilution series ranging from 5 to 30 μg was analyzed. A concentration within the linear range of the detection system that resulted in a good signal to noise ratio was chosen for monocular, binocular and somatosensory samples separately. For vMAT2, 5-HTR_1A and 5-HTR_3A analysis, this resulted in 15 μg for samples of all three regions. 5-HTR_2A was excluded from the analyses due to the lack of a specific antibody. Reference sample (pool) consisting of a mixture of equal amounts of each prepared tissue sample was run for monocular, binocular and somatosensory cortex, with the same optimal amount of protein on each gel to gauge blot-to-blot variability. After the addition of 5 μl reducing agent (10×, Invitrogen, Paisley, United Kingdom) and 2 μl LDS sample buffer (4×, Invitrogen), the samples were denatured (10 min, 70 °C). The protein samples were separated on 4–12% Bis-Tris Midi-gels in the XCell4 SureLock Midi-Cell (Invitrogen). The Spectra™ Multicolor Broad range protein ladder (ThermoScientific) was used as molecular weight standard. Subsequently, the samples were transferred to a polyvinylidene fluoride (PVDF) membrane. After 1–2 h incubation in a 5% ECL blocking agent (GE Healthcare, Buckinghamshire, UK) in Tris-saline (0.01 M Tris, 0.9% NaCl, 0.1% TX-100, pH 7.6), the membrane was incubated overnight with a primary antibody rabbit anti-vMAT2 (1:1000, R&D Systems, Novus Biologicals), with rabbit anti-5-HTR_1A (1:200, Alomone Labs), with rabbit anti-5-HTR_3A (1:200, Alomone Labs). The next day, the blots were successively washed in Tris-saline (4 × 5 min), 30 min incubated with HRP-conjugated secondary antibody (goat anti-rabbit IgG, 1:50.000, Dako, Glostrup, Denmark), rinsed in Tris-saline (5 × 5 min) and Tris-stock (1 × 5 min) (0.05 M Tris, pH 7.6). The immunoreactive bands were visualized using a chemiluminescent reaction (1 × 5 min, Super Signal West Dura, ThermoScientific, Pierce) combined with the BIO-RAD ChemiDoc™ MP Imaging System. In order to correct for intra- and inter-gel variability and to normalize the concentration of the specific detected proteins to the total amount of protein present, we performed a total protein stain (TPS) with Swift Membrane Stain (G-Biosciences) according to manufacturer’s instructions. Immediately after the staining, blots were scanned with the BIO-RAD ChemiDoc™ MP Imaging System.

Semi-quantitative Western analysis

The immunostained protein bands were semi-quantitatively evaluated by densitometry (Image Lab™ software) separately for monocular, binocular and somatosensory cortex samples (Figs. 3 and 4). First, to account for intra-gel and inter-gel variability including loading differences or incomplete transfer onto the membrane, a TPS was employed rather than the use of a single reference protein [53, 66, 67]. For each protein of interest, the optical density value per mouse was normalized to its corresponding TPS. Also, to compare samples between different gels, normalized data were expressed relative to the region-specific reference sample (pool).

Pharmacology

ME mice (pharmacology total n = 15, saline: n = 6; WAY-100635: n = 3; ketanserin: n = 3; ondansetron: n = 3) received daily injections (i.p.), at 2 pm exactly, of a specific 5-HT receptor antagonist at a dose based on prior literature, and specifically during the 3-week period of cross-modal plasticity from week 5 to 7 post-ME (Fig. 1d). Systemic drug delivery was chosen because a long-term treatment with a locally implanted slow release system (e.g. Alzet minipump) would lead to cortical tissue damage and scar formation, potentially influencing the local neuromodulator levels [68]. Drugs used were the 5-HTR_1A antagonist WAY-100635 maleate (Abcam Ab120550, 1 mg/kg, 0.2 mL i.p.) [37, 36, 69, 70], the 5-HTR_2A antagonist ketanserin tartrate (R&D Systems, Tocris Bioscience, 5 mg/kg, 0.2 mL i.p.) [71–74] and the 5-HTR_3A antagonist ondansetron hydrochloride (R&D Systems, Tocris Bioscience, 5 mg/kg, 0.2 mL i.p.) [75]. Control animals were housed under the same standard conditions and received saline injections (0.9%, i.p.) following the same injection regimen. All animals were sacrificed by cervical dislocation. Brains were rapidly removed and stored as described above.

In situ hybridization for zif268-mRNA

High-throughput radioactive ISH experiments were performed on series of coronal brain sections between Bregma levels − 1.5 mm and − 5 mm and changes in the mRNA expression level of the immediate early gene (IEG) zif268, a proven excellent activity reporter gene in the mammalian brain, were quantified (mouse: [15, 16, 51, 53, 56, 76–78], cat: [79–82]. As such, the spatial extent and the exact anatomical location of experience-induced, predominantly excitatory [6, 83–87], neuronal activity changes were analyzed and compared throughout all cortical layers of the visual and somatosensory neocortex. This high-throughput approach allows the molecular visualization of cortical reactivation patterns in response to ME. ISH for zif268-mRNA was performed with a mouse-specific synthetic oligonucleotide probe (Eurogentec, Seraing, Belgium) with sequence 5′-ccgttgctcagcagcatcatctcctccagtttggggtagttgtcc-3′. As described previously [51, 53, 77, 88], each probe was 3′-end labeled with [³³P] dATP using terminal deoxynucleotidyl transferase (Invitrogen, Carlsbad, CA). Unincorporated nucleotides were separated from the labeled probe by means of miniQuick Spin Oligo Columns (Roche Diagnostics, Vilvoorde, Belgium). The cryostat sections were fixed, dehydrated and delipidated. The radioactively labeled probes were added to a hybridization cocktail (50% (v/v) formamide, 4× standard SSC buffer, 1× Denhardt’s solution, 10% (w/v) dextran sulfate, 100 μg/ml herring sperm DNA, 250 μg/ml tRNA, 60 mM dithiothreitol, 1% (w/v) N-lauroyl sarcosine, and 20 mM NaHPO₄, pH 7.4) and applied to the cryostat sections (10⁶ cpm/section). After an overnight incubation at 37 °C in a humid chamber, the sections were rinsed in 1× standard SSC buffer at 42 °C, dehydrated, air-dried and exposed to an autoradiographic film (Biomax MR; Kodak, Rochester, NY). After 7 days, the films were developed in EMS replacement for Kodak developer D-19 (Electron Microscopy Sciences, Hatfield) and fixed in Rapid fixer (Ilford Hypam; Kodak). Autoradiographic images of adjacent sections per examined cortical area per mouse were scanned at 1200 dpi (CanoScan LIDE 600F; Canon, Tokyo, Japan), and all images were similarly adjusted for brightness and contrast in Adobe Photoshop Elements 2018 (version 16.0, × 64, Adobe Systems Incorporated).

Histology and localization of visual and somatosensory areal boundaries with Nissl patterns

Upon ISH, the cryostat sections were Nissl-counterstained (1% cresyl violet; Fluka, Sigma-Aldrich) according to standard procedures to visualize cortical boundaries between different visual and somatosensory areas and to aid the interpretation of the zif268-activity patterns (Fig. 1c). Images of the stained coronal sections were obtained at 5× (NA: 0.16) with a light microscope (Zeiss Axio Imager Z1) equipped with an AxioCam MRm camera (1388 × 1040 pixels) using the software program Zen (Zen Pro 2012, Carl Zeiss, Benelux). Comparisons were made with the stereotaxic mouse brain atlas [63] to delineate visual and somatosensory cortical borders as described previously [53, 56]. In all figures illustrating visual or somatosensory cortex, large arrowheads indicate the total extent of the cortex, whereas small arrowheads indicate the interareal borders. In the visual cortex, five subregions can be distinguished from lateral to medial (Figs. 1c, 5): the lateral extrastriate cortex (V2L), which is subdivided into a monocular (V2Lm, segments 1–4) and binocular region (V2Lb, segments 4–8), the primary visual cortex (V1) which is subdivided further into a binocular (V1b, segments 8–15) and monocular region (V1 m, segments 15–21), and the medial extrastriate cortex (V2M, segments 21–24) [51, 53, 56, 89]. For the zif268 analysis, we focused specifically on the Bz (V2Lb-V1b) and the Mmz (V1 m-V2M) as these regions undergo open-eye potentiation and cross-modal plasticity respectively. In the somatosensory cortex, we distinguished the primary somatosensory barrel field (S1BF) from the more lateral secondary somatosensory cortex (S2) and the more medial primary somatosensory cortex (S1), as the primary receiver of whisker inputs (Fig. 6) [63].

Quantitative analysis of ISH results

To quantify the optical density (OD; mean gray value per pixel) of the ISH autoradiograms, a custom-made Matlab (Matlab R2017a; Mathworks) script was used as described previously [51, 53]. We analyzed at least three mice per condition. Per mouse, three ISH sections with an inter-distance of 100 μm were investigated (for visual cortex: − 3.5-(− 3.7); 1–4; 1, relative to Bregma; for somatosensory cortex: − 1.7-(− 1.9); 2.5–4; 1, relative to Bregma). The region of interest in the left hemisphere was demarcated by determining the top edge of the cortex, the boundary between the supra-and granular layers (II-III and IV) and the infragranular layers (V and VI), and the border between the infragranular layers and the corpus callosum. The region of interest was then divided equally into 24 segments from lateral to medial to create two lattices of 24 quadrangles, corresponding to the upper (II-IV) and lower (V-VI) layers. To compensate for possible variation in brain size and morphology, the lattices were translated on each autoradiogram over the cortical curvature, fixing the border of a specific segment to an areal border (border segment 20/21 is the area border V1 m/V2M). For each segment created this way, the relative OD was calculated as the mean gray value of all pixels contained within a particular quadrangle and was normalized to the mean gray value of a square measured in the corpus callosum (a defined region with no zif268-mRNA expression above background) in order to compare autoradiograms across experiments. Relative neuronal activity was expressed in percentages based on the following formula: 1 – (cortical zif268/background) × 100. Results are presented as profiles of neuronal activity per brain area and are illustrated in subregion-specific bar graphs. Pseudo-color maps were generated through a second custom-made Matlab script (Matlab R2017a; Mathworks, Natick, MA) and represent a false color coding of the gray values ranging from black (0) to white (225); high gray values are represented in white/yellow on the false color scale bar (0–50), medium high gray values are represented in red (50–100) or blue (100–150) and low gray values are represented in green/black (150–255).

Statistics

All HPLC results, Western blotting data and relative OD-values in ISH-sections were presented as mean ± SEM. Normal distribution and in parallel, equal variance between groups was tested. A non-parametric test (Mann-Whitney) was applied for pairwise comparison. A Kruskal-Wallis test followed by a Dunn’s multiple comparisons post-hoc test was used to determine the recovery time-dependent modulations of protein expression levels upon ME. For all tests, a probability level (α level was set to 0.05) of < 0.05 was accepted as statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001). Statistical analyses were performed using GraphPad Prism 5.01 (GraphPad Software, Inc).

Reduced visual cortex 5-HT concentration in response to monocular enucleation

To investigate if 5-HT is involved in adult ME-induced cross-modal plasticity and to better understand its role therein, we first examined whether the total 5-HT concentration was affected in the two cortices that functionally adapt upon ME. The total 5-HT concentration was analyzed with High-Performance Liquid Chromatography (HPLC) on whole tissue homogenates of the visual cortex and S1BF 7 weeks after performing ME in adult mice (7wME) (Fig. 2a, b). At this 7wME endpoint, when the cross-modal recruitment of the initially deprived visual cortex is completed [15], we observed a decreased 5-HT concentration in the visual cortex (mean ± SEM, AMC: 82.54 ± 4.77; 7wME: 65.50 ± 4.77, Mann-Whitney test, p = 0.018, Fig. 2c). In S1BF, the cortical area in which compensatory plasticity is triggered by vision loss, no significant difference was reached (AMC: 113.40 ± 8.06; 7wME: 86.41 ± 10.49; Mann-Whitney test, p = 0.114, Fig. 2d). These results suggest brain region-specific modulations of total 5-HT content in response to visual deprivation.

ME induces pre- and postsynaptic changes in proteins related to 5-HT signaling

Parallel to the ME-induced decrease in total 5-HT concentration in the visual cortex, the impact of partial vision loss might manifest at the level of 5-HT loading into readily releasable vesicles or at the level of 5-HT receptors. To investigate possible brain region-specific expression changes in these pre- and postsynaptic proteins involved in 5-HT signaling, we performed Western blotting experiments on time-course samples from the visual Mmz and Bz separately and from S1BF. Post-ME recovery periods of respectively 1 week (1wME: ongoing open-eye potentiation phase), 3 weeks (3wME: end of the open-eye potentiation phase) and 5 weeks (5wME: ongoing cross-modal phase) were chosen as intermediate time points towards the time point of maximal visual cortex reactivation (7wME) for the evaluation of possible post-ME recovery time-dependent modulations in protein expression in addition to endpoint evaluation (Fig. 1) [15].

We chose to analyze vMAT2 protein expression levels since they provide information on the presynaptic loading of vesicles with 5-HT and because vMAT2 expression levels positively correlate with the total 5-HT concentration present in a given brain region of interest [60–62]. vMAT2 expression levels were significantly decreased in the Mmz in 1wME, 5wME and 7wME mice (represented as * within the bar, bar graphs in Figs. 3 and 4) compared to AMC (represented as a dashed line in Figs. 3 and 4, mean ± SEM, 1wME: 0.34 ± 0.08, p = 0.0159; 3wME: 0.80 ± 0.56, p = 0.4127; 5wME: 0.45 ± 0.02, p = 0.0159; 7wME: 0.53 ± 0.28, p = 0.0317, Mann-Whitney test). No statistically significant modulations occurred as a function of post-ME recovery time (Kruskal-Wallis test: p = 0.0159, with no statistical difference in the Dunn’s post-hoc test for pairwise comparison) (Fig. 3a). Western analysis of the vMAT2 expression in the Bz indicated increased vMAT2 expression in 1wME and 3wME mice compared to AMC mice (mean ± SEM, 1wME: 1.87 ± 0.19, p = 0.0317; 3wME: 1.45 ± 0.06, p = 0.0159; 5wME: 1.07 ± 0.07, p = 0.4127; 7wME: 1.01 ± 0.04, p = 0.7302, Mann-Whitney test). As for the Mmz, also for the Bz no significant recovery time-dependent changes were observed (Kruskal-Wallis test: p = 0.0095, with no statistical differences in the Dunn’s post-hoc test for pairwise comparison). In this case, however, a trend of decreasing vMAT2 expression was observed over the 1w to 7w post-ME time course (Fig. 3b). In S1BF, vMAT2 expression was significantly increased in 1wME mice compared to AMC (mean ± SEM, 1wME: 1.58 ± 0.11, p = 0.0317; 3wME: 1.50 ± 0.10, p = 0.0635; 5wME: 0.77 ± 0.09, p = 0.1905; 7wME: 1.29 ± 0.11, p = 0.4127, Mann-Whitney test, Fig. 3c). Furthermore, in 1wME and 3wME mice, the vMAT2 expression levels were higher compared to 5wME mice (Kruskal-Wallis test: p = 0.0081). These results point towards a return to baseline vMAT2 expression levels specifically at the time when the cross-modal whisker take-over of the deprived visual cortex starts to occur [15].

To define the impact of ME on the serotonergic modulation of cortical inhibition through specific 5-HT receptors, we performed Western blotting experiments for 5-HTR_1A and 5-HTR_3A. Besides their role in different types of visual cortex plasticity, reasons supporting these receptor choices, are the fact that the 5-HTR_1A receptor is expressed both on pre-and postsynaptic neurons, respectively acting as an autoreceptor and a Gi-coupled GPCR mediating inhibitory neurotransmission upon ligand binding [90], while 5-HTR_3A is a Na^+/K^+/Ca²⁺ permeable ion channel exclusively expressed on the 5-HTR_3A expressing inhibitory interneurons [91]. We did not observe any ME-induced or recovery time-dependent changes in the expression of 5-HTR_1A or 5-HTR_3A in the Mmz (mean ± SEM, 5-HTR_1A, 1wME: 1.48 ± 0.21, p = 0.2857; 3wME: 0.93 ± 0.15, p = 0.7302; 5wME: 1.00 ± 0.07, p = 0.7302; 7wME: 0.83 ± 0.06, p = 0.7302, 5-HTR_3A, 1wME: 1.29 ± 0.17, p = 0.2857; 3wME: 1.03 ± 0.13, p = 0.9048; 5wME: 0.98 ± 0.08, p = 1.000; 7wME: 0.78 ± 0.06, p = 0.1905, Mann-Whitney test; Kruskal-Wallis test, 5-HTR_1A: p = 0.2493; 5-HTR_3A: p = 0.1395, Fig. 4a, b). In the Bz on the other hand, 1wME and 3wME mice showed increased 5-HTR_1A expression levels (mean ± SEM, 5-HTR_1A, 1wME: 1.69 ± 0.12, p = 0.0317; 3wME: 1.44 ± 0.06, p = 0.0317; 5wME: 1.12 ± 0.06, p = 0.2857; 7wME: 0.99 ± 0.02, p = 0.9048, Mann-Whitney test, Fig. 4c) and a gradual decrease in 5-HTR_1A expression between 1wME and 7wME mice (Kruskal-Wallis test: p = 0.0039). In addition, 5-HTR_3A expression levels were significantly increased in the Bz of the 3wME mice compared to the AMCs (mean ± SEM, 5-HTR_3A, 1wME: 1.59 ± 0.17, p = 0.1111; 3wME: 2.10 ± 0.13, p = 0.0159; 5wME: 1.45 ± 0.13, p = 0.1111; 7wME: 0.87 ± 0.08, p = 0.5556, Mann-Whitney test, Fig. 4d) and decreased over time to reach the AMC level at 7 weeks post-ME (Kruskal-Wallis test: p = 0.0038). In S1BF, 5-HTR_1A expression was increased in 1wME and 3wME mice, gradually decreased to reach the AMC level at 5 weeks post-ME (Kruskal-Wallis test: p = 0.0039), and was again increased at 7 weeks post-ME (mean ± SEM, 5-HTR_1A, 1wME: 1.84 ± 0.09, p = 0.0159; 3wME: 1.56 ± 0.08, p = 0.0159; 5wME: 1.17 ± 0.07, p = 0.4127; 7wME: 1.60 ± 0.12, p = 0.0317, Mann-Whitney test, Fig. 4e). We did not observe significant modulations of 5-HTR_3A protein expression in S1BF (mean ± SEM, 5-HTR_3A, 1wME: 1.88 ± 0.19, p = 0.1905; 3wME: 2.13 ± 0.34, p = 0.1111; 5wME: 1.28 ± 0.13, p = 0.5556; 7wME: 2.26 ± 0.43, p = 0.1111, Mann-Whitney test; Kruskal-Wallis test: p = 0.1325, Fig. 4f). Taken together, these observations indicate brain region-specific alterations in vMAT2-mediated presynaptic monoamine vesicle loading, as well as post-ME recovery time-dependent modulations in 5-HTR_1A and 5-HTR_3A expression levels specifically in the Bz and in S1BF but not in the Mmz.

5-HTR_2A and 5-HTR_3A but not 5-HTR_1A antagonists suppress ME-induced cross-modal plasticity

We hypothesized that, in case 5-HT receptor function is involved in the whisker-mediated reactivation of the deprived visual cortex, the ME-induced cross-modal recruitment could be suppressed by systemic administration of the 5-HTR_1A antagonist WAY-100635 maleate, the 5-HTR_2A antagonist ketanserin tartrate, or the 5-HTR_3A antagonist ondansetron hydrochloride during the last 3 weeks of recovery of the Mmz in ME mice (Fig. 5a-e). As the 5-HT receptors influence the excitability of brain circuits by modulating excitatory and inhibitory neurotransmission and thus by mediating the cortical E/I balance, we expected to observe effects of these drugs on cortical activity, and ultimately on cortical plasticity [36, 54, 55]. Because it was previously established that the most pronounced impact of whisker inputs on neuronal activity involves the infragranular layers of the visual cortex, as based on the expression of neuronal activity reporter gene zif268, these layers were interrogated separately from the supra- and granular layers.

Suppression of the 5-HTR_1A receptor function by long-term administration of WAY-100635 from week 5 to 7 post-ME did not alter the neuronal activity levels reached throughout the Mmz of 7wME mice, compared to saline-injected age-matched control ME mice (Mmz, supra- and granular layers: p = 0.9048, infragranular layers: p = 0.9048, Bz, supra- and granular layers: p = 0.0476, infragranular layers: p = 0.2619, Mann-Whitney test, Fig. 5f). However, it did decrease the neuronal activity in the supra- and granular layers of the Bz (Mean ± SEM, supra- and granular layers, saline, Bz: 71.58 ± 1.26; Mmz: 60.14 ± 1.48, WAY-100635, Bz: 65.73 ± 1.58; Mmz: 57.95 ± 3.77, infragranular layers, saline, Bz: 71.31 ± 1.57; Mmz: 56.97 ± 1.84, WAY-100635, Bz: 67.72 ± 3.35; Mmz: 55.96 ± 2.12, Fig. 5f), indicative of Bz-specific 5-HTR_1A expression changes during the preceding open-eye potentiation phase.

Blocking 5-HTR_2A receptor function by ketanserin administration from week 5 to 7 post-ME significantly decreased the neuronal activity across all layers of the Mmz (Mmz, supra- and granular layers: p = 0.0476, infragranular layers: p = 0.0238; Bz, supra- and granular layers: p = 0.0952, infragranular layers: p = 0.2619, Mann-Whitney test, Fig. 5g) but did not induce any change in neuronal activity in the Bz (Mean ± SEM, supra- and granular layers, saline, Bz: 71.58 ± 1.26; Mmz: 60.14 ± 1.48, ketanserin, Bz: 66.15 ± 2.83, Mmz: 51.18 ± 3.25, infragranular layers, saline, Bz: 71.31 ± 1.57; Mmz: 56.97 ± 1.84, ketanserin, Bz: 66.18 ± 3.05; Mmz: 48.85 ± 1.46, Fig. 5g). Suppression of the 5-HTR_3A receptor function by administration of ondansetron only reduced the neuronal activity specifically in the supra- and granular layers of the Mmz (Mmz, supra- and granular layers: p = 0.0476, infragranular: p = 0.5476; Bz, supra- and granular layers: p = 0.7143, infragranular layers: p = 0.5476, Mann-Whitney test; Mean ± SEM, supra- and granular layers, saline, Bz: 71.58 ± 1.26; Mmz: 60.14 ± 1.48, ondansetron, Bz: 71.63 ± 1.40, Mmz: 51.95 ± 2.04, infragranular layers, saline, Bz: 71.31 ± 1.57; Mmz: 56.97 ± 1.84, ondansetron, Bz: 72.64 ± 0.94; Mmz: 54.70 ± 0.40, Fig. 5h). Taken together, these results point out that modulation of the 5-HTR_2A and 5-HTR_3A receptor function, influences the late cortical response to ME, suggesting that these receptors play an important role in the cross-modal reactivation phase of the visual cortex following ME.

Previous work has shown that ME induces increased neuronal activity in the spared sensory brain areas adjacent to the deprived visual cortex. For somatosensory cortex, the zif268-mRNA expression levels were found to be significantly higher in adult 7wME mice compared to AMC mice [16, 53]. In order to assess the impact of pharmacological modulation of serotonergic neurotransmission on S1BF neuronal activity, next to mapping the extent of visual cortex reactivation, we compared S1BF activity levels between saline-injected 7wME mice and 7wME mice that were treated with the specific 5-HTR_1A, 5-HTR_2A or 5-HTR_3A antagonists.

Analysis of the zif268-mRNA expression levels (− 1.7-(− 1.9); 2.5–4; 1, relative to Bregma) (Fig. 6a-e) indicated that the neuronal activity levels in S1BF of WAY-100635 treated 7wME mice remained unaltered across all cortical layers compared to the saline-injected 7wME mice (S1BF, supra- and granular layers: p = 0.9048, infragranular layers: p = 0.7143, Mann-Whitney test, Fig. 6f). The mice thus displayed the expected ME-induced increase in S1BF activity, despite a suppressed 5-HTR_1A receptor function (Mean ± SEM, supra- and granular layers, saline: 64.94 ± 0.49; WAY-100635: 64.90 ± 1.76, infragranular layers, saline: 65.17 ± 0.66; WAY-100635: 66.19 ± 1.67, Fig. 6f). Treatment of 7wME mice with ketanserin (Mean ± SEM, supra- and granular layers, saline: 64.94 ± 0.49; ketanserin: 69.77 ± 1.13, infragranular layers, saline: 65.17 ± 0.66; ketanserin: 68.68 ± 1.63, Fig. 6g) and similarly, with ondansetron (Mean ± SEM, supra- and granular layers, saline: 64.94 ± 0.49; ondansetron: 69.35 ± 1.11, infragranular layers, saline: 65.17 ± 0.66; ondansetron: 71.78 ± 2.38, Fig. 6h), also did not prevent the normal ME-induced increase in S1BF activity. On the contrary, it even induced higher neuronal activity across all cortical layers of S1BF, as the zif268-mRNA expression levels were significantly higher in ketanserin and ondansetron treated 7wME mice compared to the saline-injected control group (ketanserin, supra- and granular layers: p = 0.0238, infragranular layers: p = 0.0476; ondansetron, supra- and granular layers: p = 0.0238, infragranular layers: p = 0.0476, Mann-Whitney test). Thus, S1BF activation seems preserved in the drug-treated mice, and even intensified as visualized with 5-HTR_2A antagonist ketanserin and 5-HT_3A antagonist ondansetron.

Late-onset ME modulates 5-HT and vMAT2 levels in adult visual cortex

Our HPLC data imply a lowered whole tissue 5-HT concentration specifically in the visual cortex of long-term ME mice. Since 5-HT release is triggered by neuronal activity in the sensory cortex, the observed decrease in 5-HT concentration in the visual cortex of 7wME mice is possibly due to the ME-induced loss of visual input in the Mmz [102]. A similar decrease in total 5-HT levels was also observed in the deprived visual cortex of adult cats in response to retinal lesions, an animal model for age-related macular degeneration [59]. For more in depth interpretation of brain region-specific alterations in 5-HT signaling, we still depend on new methodological developments that would allow reliable longitudinal analysis of the true local 5-HT release from only 1 mm thick cortical tissue, such as the Bz or the Mmz in our mouse model. For now, predictions based on longitudinal vMAT2 expression changes allow to formulate some possibilities. Tong et al., (2011) indeed already showed that vMAT2 levels are positively correlated with the tissue concentrations of total monoamines (5-HT, dopamine and noradrenaline), as measured in human tissue by means of HPLC analysis [62]. By deduction, the significantly decreased vMAT2 protein expression already shortly after the introduction of ME, may indicate decreased 5-HT concentrations in the Mmz immediately upon the loss of all its visual input. Opposite to the Mmz, the Bz displayed only a temporary increase in vMAT2 expression from week 1 to 3 weeks post-ME, at a time when its remaining open-eye inputs become potentiated. Spared cortical territory seems to share such an early 5-HT mediated response, in relation to compensating for the sensory deficit. Indeed, a similar pattern of early increased vMAT2 expression in S1BF, could go hand in hand with a fast, initial increase in 5-HT concentration and might indicate a compensatory, cortical plasticity mechanism similar to the one described by Jitsuki et al., (2011) after only 2 days of visual deprivation in S1BF of young rats [32]. Once a new functional cortical balance is established for the spared sense, it may be able to subsequently recruit nearby sensory deprived cortical territory. In our model of acquired blindness, somatosensation indeed seems to be capable of recruiting the Mmz once fine-tuned whisker-information processing has been established in S1BF.

5-HTR_1A mediates unimodal open-eye potentiation in the Bz of the visual cortex

Cortical plasticity depends on both excitation and inhibition levels, which is determined by the distribution of excitatory and inhibitory receptors. An established E/I balance defines the stability and efficacy of neuronal circuits and is therefore required for proper processing of sensory information [36, 43, 103, 104]. A study, investigating the serotonergic regulation of the two predominant types of inhibition, respectively the short intermittent bursts of phasic inhibition and the constant long-lasting tonic inhibition, aside from its effect on the pyramidal neurons in the rat visual cortex, showed that 5-HT suppressed tonic inhibition through the 5-HTR_1A receptor and Protein Kinase A signaling while phasic inhibition was enhanced through 5-HTR_2A receptor function and CaM Kinase II [105, 106]. Clearly, alterations in 5-HT receptor expression and function can exert a major impact on the excitability of brain circuits involved in brain plasticity, as this would influence the inhibitory neurotransmission and the regulation of the cortical E/I balance [36, 54, 55].

Our observations indicate ME-induced increases in 5-HTR_1A receptor expression in the spared cortices, the Bz and S1BF, confirming a regionally restricted and modulatory role in maintaining cortical excitability immediately upon insult. In the Bz, the increase in 5-HTR_1A expression occurred at the start of the open-eye potentiation phase, pointing towards an early involvement of 5-HTR_1A-mediated tonic inhibition, which implies the consistent activation of extra- and perisynaptic GABA_AR_δ1 receptors on pyramidal and inhibitory neurons by ambient GABA, to regulate the overall cortical excitability during this thalamocortical plasticity process [107]. Of note, such enhanced tonic inhibition also triggers ocular dominance plasticity in the mouse visual cortex during the critical period early in life [108].

Many studies previously reported on 5-HT-mediated induction of unimodal brain plasticity in adulthood. For example, adult rats were again susceptible for monocular deprivation-induced ocular dominance plasticity after cortical infusion of 5-HT directly into the visual cortex [37]. Also in rodents, treatment with the SSRI fluoxetine, was found to induce both hippocampal synaptic plasticity and unimodal ocular dominance plasticity in the visual cortex [26, 37, 109, 110]. In line with this, in the human visual system, long-term systemic treatment with the SSRI sertraline was shown to induce synaptic plasticity by increasing the amplitude of stimulus-induced visually evoked potentials [111]. Fluoxetine even emerged as a possible treatment for persistent amblyopia, or lazy eye, in adult humans [112, 113]. Maya-Vetencourt et al., (2011) could previously also show that the ocular dominance shift, normally occurring in the Bz shortly after monocular deprivation, can be prevented by blocking this receptor with WAY-100635 [37]. Furthermore, local application of this same selective 5-HTR_1A antagonist in V1 by means of reverse-phase microdialysis was found to facilitate long-term potentiation (LTP) after theta-burst stimulation (TBS) in adult rats, while LTP was inhibited in juvenile rats, confirming an age-dependent role for 5-HTR_1A in gating V1 plasticity [27].

We were not surprised by the observation that 5-HTR_1A expression levels remained close to the AMC level in the Mmz, as our lab previously described that well-defined levels of GABA_AR_α1-mediated phasic inhibition, but not GABA_AR_δ1-mediated tonic inhibition, are crucial for the establishment of the cross-modal reactivation of the Mmz in adult ME mice [53]. Our interpretation, that 5-HTR_1A is of central importance to ME-induced open-eye potentiation in the Bz but not to the late cross-modal plasticity phase in the Mmz, is further supported by the observed downregulation of 5-HTR_1A receptors in the Bz from the moment that open-eye potentiation is completed, and by the absence of suppressive drug-effects in the Mmz itself during the cross-modal phase. Indeed, the normal zif268-mRNA expression levels observed across the upper and lower cortical layers of the Mmz argue against a substantial contribution of this receptor subtype to the cross-modal take-over of the deprived cortical territory.

Antagonism of 5-HTR_2A and 5-HTR_3A to spatiotemporally control cross-modal brain plasticity

In this study, next to WAY-100635, we also chronically injected ketanserin or ondansetron to respectively block 5-HTR_2A and 5-HTR_3A receptor function during the entire time window of cross-modal plasticity, and following a normal period of open-eye potentiation. Although applying a local slow drug-release approach might have allowed a better discrimination of true cortical events from for example subcortical contributions to the observed plasticity phenomena, we still chose the i.p. injection approach because this systemic approach holds a greater prediction value with regards to human clinical trials than an invasive cortical infusion system. Also, long-term cannula implantation may have led to cortical tissue damage and glial scar formation. Such side-effects would have influenced the local neuromodulator levels [63] hampering the read-out as well as the interpretation of the results.

By blocking the 5-HTR_2A receptor with ketanserin, we successfully suppressed adult cross-modal visual cortex reactivation across all layers of the Mmz. This observation accords with literature describing a role for 5-HTR_2A in compensatory cross-modal plasticity [32] and with our previous observations, that post-ME cross-modal reactivation of the Mmz only occurs under well-defined levels of GABA_AR_α1-mediated phasic inhibition [53], which is considered to be strictly regulated by upstream 5-HTR_2A function [105].

5-HTR_3A interneurons constitute a heterogenous population. One type of interneurons dependent on 5-HT signaling are the Vasoactive Intestinal Peptide (VIP)-expressing 5-HTR_3A interneurons. VIP-mediated disinhibition is a well-described mechanism in which VIP interneurons inhibit somatostatin and parvalbumin expressing interneurons and may underlie ME-induced cortical plasticity [44, 46–48, 114, 115]. VIP interneurons exert an important function in controlling sensory processing and are target cells of long-range projections that integrate sensory information originating from different brain regions [116, 117]. This interneuron type thus constitutes an ideal candidate for 5-HT to shape cross-modal brain circuits and plasticity patterns via the 5-HTR_3A receptor. In a recent study [4], convergence of sensory and neuromodulatory information onto 5-HTR_3A cells was shown to vastly contribute to the shaping of critical period plasticity in the primary auditory cortex of the mouse. This research group discovered that a topographic map is formed already early in life, regulated by the non-VIP expressing 5-HTR_3A cell population in cortical layer I. As predicted based on the 5-HTR_3A cortical distribution pattern, which is predominantly in L1 of the cortex [4], our results indeed indicate that 5-HTR_3A receptor antagonism with ondansetron during the cross-modal reactivation phase, specifically suppressed the neuronal activity in the upper layers of the Mmz. It will be interesting to further investigate how and to what extent both the VIP and non-VIP 5-HTR_3A cells can spatiotemporally regulate adult cross-modal plasticity as this may enable steering brain plasticity towards the desired outcome as a future therapy, being complete functional recovery of the lost primary sense. Ex-vivo and in vivo electrophysiology and fast-scan voltammetry, should allow dissecting the local molecular cascade and cellular circuit involved in the different cortical plasticity phenomena.

A possible strategy to improve the success rate of bionic implants

The success of bionic implants in restoring sensory function depends on our understanding of how the brain responds to sensory loss. Often, by the time the neuro-electric device is implanted, the brain has already compensated for the loss of sensory input through mechanisms of cross-modal plasticity, thereby possibly impairing the cortical translation of inputs, transmitted by the device, into meaningful sensory information [118]. During the last decades, the attempt to fully restore primary sensory function received much attention, giving rise to very promising neurobionics such as the cochlear implants to restore auditory function [119–121], visual prostheses [122–126] and brain-computer interfaces [127–129]. In addition, patients suffering from progressive vision loss, as caused by the neurodegenerative diseases Retinitis pigmentosa or age-related macular degeneration, can nowadays benefit from the recently developed, fully organic, subretinal prosthesis [130], which tackles some of the common mechanical, manufacturing and technical difficulties generally occurring in epiretinal [131], subretinal [132] and suprachoroidal [133] prosthetics.

Given the fact that 5-HT modulates sensory information processing and integration [134] and based on our findings that chronic, systemic administration of the 5-HTR_2A antagonist ketanserin and 5-HTR_3A antagonist ondansetron can suppress the cross-modal reorganization after partial vision loss in a cortical brain region- and layer-specific manner, we predict that a pharmacological monoamine-oriented strategy in combination with a bionic implant may significantly increase their success rate, as this approach would better allow to spatiotemporally control different forms of cortical plasticity that precede and co-occur with its functional integration.