Altered serotonin innervation in the rat epileptic brain

Accepted Manuscript Title: Altered serotonin innervation in the rat epileptic brain Authors: Gisela H. Maia, Joana I. Soares, Sérgio G. Almeida, Juliana M. Leite, Helena X. Baptista, Alisa N. Lukoyanova, Cátia S. Brazete, Nikolai V. Lukoyanov PII: DOI: Reference: S0361-9230(18)30804-9 https://doi.org/10.1016/j.brainresbull.2019.07.009 BRB 9726 To appear in: Brain Research Bulletin Received date: Revised date: Accepted date: 16 October 2018 28 June 2019 8 July 2019 Please cite this article as: { https://doi.org/ This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Altered serotonin innervation in the rat epileptic brain Gisela H. Maiaa,b,c,d,e,*, Joana I. Soaresa,b,c,*, Sérgio G. Almeidae, Juliana M. Leited,e, Helena X. Baptistad,e, Alisa N. Lukoyanovad,e, Cátia S. PT Brazetec, Nikolai V. Lukoyanova,b,c,d,e a RI Instituto de Investigação e Inovação em Saúde, Universidade do Porto, b SC Portugal Neuronal Networks Group, Instituto de Biologia Molecular e Celular da N U Universidade do Porto, Porto, Portugal c A Departamento de Biomedicina, Faculdade de Medicina da Universidade do d M Porto, Porto, Portugal Medibrain, Vila do Conde, Portugal ED e PT Brain Research Institute, Porto, Portugal * CC E These authors equally contributed to this work Highlights A  Kainate-induced epilepsy is associated with neuroplastic changes in the raphe nuclei  Serotonin fibers in epileptic rats show increased density of larger-sized varicosities  Chronic epilepsy can trigger strong structural reorganization in ascending serotonergic pathways 2 Abstract Studies in animal models of epilepsy revealed compromised serotonin (5HT) transmission between the raphe nuclei and the brain limbic system. The goal of the present study was to evaluate the effects of epilepsy on the structural PT integrity of the dorsal (DR) and median (MnR) raphe nuclei and on the RI morphology of serotonergic fiber terminals in the dentate gyrus (DG), infralimbic cortex (IL) and medial septum (MS). The study was performed in SC adult Wistar rats using the kainate (9.5 mg/kg) status epilepticus (SE) model. U Four months post-SE, the brainstem sections of the animals were N immunostained for 5-HT, whereas the forebrain sections were immunostained A for serotonin transporter (SERT). Stereological analysis revealed that epileptic M rats, as compared with controls, had approximately 30% less 5-HT-stained cells ED in the interfascicular part of the DR, but twice as many 5-HT-stained cells in the MnR. Another finding was the reorganization of the 5-HT fiber network in all PT target areas analyzed, as indicated by the rightward shift of the density-size CC E distribution histograms of SERT-stained fiber varicosities. Nonlinear regression analysis of these histograms revealed that SERT-stained varicosities were A represented by two subpopulations characterized by distinct cross-sectional areas. The areal density of the small-sized varicosities was decreased in the DG (hilus and molecular layer), IL cortex (layers II/III) and MS, while that of the larger-sized varicosities was increased. The present results support the 3 hypothesis that chronic epilepsy can trigger profound structural reorganization of the ascending serotonergic pathways in the rat brain. Keywords: 5-Hydroxytryptamine; Serotonin transporter protein; Temporal lobe RI PT epilepsy; Kainic acid; Stereology; Fiber varicosities Introduction SC Parahippocampal and hippocampal neuronal circuits are highly U hyperexcitable, which is thought to be a risk factor for the formation of N epileptogenic foci in this brain region (Vismer et al., 2015). Indeed, temporal A lobe epilepsy (TLE), characterized by pathophysiological alterations in the M hippocampal formation (HF), is the most common subtype of focal epilepsy ED (Engel, 1989). TLE can be associated with progressive neuron loss with gliogenesis (hippocampal sclerosis) (Cavanagh and Meyer, 1956; Margerison PT and Corsellis, 1966) and a number of comorbid neuropsychiatric disorders, such CC E as cognitive deficits (Alessio et al., 2004), anxiety, major depression and other mood disorders (Piazzini et al., 2001; Tellez‐Zenteno et al., 2007). Furthermore, A seizures originated in the medial temporal lobe region are often resistant to existing antiepileptic drugs (Hauser, 2001). However, the cellular mechanisms underlying hippocampal epileptogenesis and TLE development are not yet clearly defined. It has been found, in both humans and animal models, that the 4 loss of injury-vulnerable mossy cells located in the hilus of the dentate gyrus (DG) is related to aberrant sprouting of the granule cell mossy fibers into the DG molecular layer (Nissinen et al., 2000; Sutula et al., 1988; Sutula et al., 1989). This recurrent excitatory drive onto granule cells is one of the putative PT mechanisms by which injury-induced neuron loss might trigger hippocampal RI epileptogenesis (Buckmaster and Dudek, 1997; Buckmaster et al., 2002; Lynch and Sutula, 2000). However, recent studies showed that blockade of mossy fiber SC sprouting and synaptogenesis within the DG molecular layer does not prevent U epileptogenesis (Heng et al., 2013; Yamawaki et al., 2015). Although these data N do not rule out the epileptogenic role for mossy fiber sprouting (Zeng et al., A 2009), they suggest that other cellular mechanisms might also be involved. M The forebrain serotonin (5-HT) transmission system consists of several ED fiber tracts ascending from the rostral group of the brainstem raphe nuclei (Jacobs and Azmitia, 1992; Michelsen et al., 2008) and their target cells, both PT neurons and astrocytes (Azmitia et al., 1996), which express various types of CC E 5-HT receptors (Barnes and Sharp, 1999). Serotonin is a potent regulator of neuronal excitability (Celada et al., 2013; Ropert and Guy, 1991; Varga et al., A 2009) and has long been implicated in controlling seizure activity (Bagdy et al., 2007; Browning et al., 1978; Racine and Coscina, 1979). Moreover, it has been reported that the binding potency of 5-HT receptors is reduced in TLE patients (Hasler et al., 2007; Theodore et al., 2012) and a limited number of clinical studies revealed beneficial effects of selective serotonin reuptake inhibitors 5 (SSRI) on seizure frequency (Favale et al., 1995; Favale et al., 2003; Specchio et al., 2004). However, the contribution of this neurotransmitter to epilepsy and epilepsy-related comorbid disorders remains largely underexplored. Evidence from animal models indicates that serotoninergic transmission PT between the raphe nuclei and the HF is compromised in chronic epilepsy RI (Mazarati et al., 2008) and that treating epileptic rats with SSRIs is capable of reducing brain excitability (Mazarati et al., 2008) and seizure frequency SC (Hernandez et al., 2002; Vermoesen et al., 2012). In addition, the basal levels of U 5-HT and its turnover were found to be lower in hippocampal tissue of epileptic N rats as compared to control rats (Mazarati et al., 2008). Similarly, amygdala A kindling produced a reduction of 5-HT content in the rat ipsilateral M hippocampus (Lewis et al., 1987). Consistent with this, we have shown (Maia et ED al., 2016) that experimental epilepsy in rats is associated with loss of a subset of 5-HT-IR neurons in the interfascicular part of the dorsal raphe nucleus (DRI) PT known to project to various limbic brain regions, likely including the HF (Hale CC E and Lowry, 2011). However, several studies failed to detect changes in hippocampal 5-HT levels in animal models of chronic epilepsy (Cavalheiro et A al., 1994; Szyndler et al., 2005). Moreover, studies using the pentylenetetrazole kindling in rats (Szyndler et al., 2010) and amygdala kindling in cats (Shouse et al., 2001) showed that increased seizure activity correlates rather with increased levels of hippocampal serotonin. From a clinical perspective, it has been shown that 5-HT levels are increased in actively spiking versus non-spiking specimens 6 of temporal cortex obtained from TLE patients (Louw et al., 1989). Recently, hippocampal 5-HT levels in TLE patients were shown to negatively correlate with a history of generalized tonic-clonic convulsions, but not with other epilepsy-related variables, including seizure frequency and the presence of PT neuropsychiatric disorders (da Fonseca et al., 2015). RI The majority of the forebrain serotonergic afferents come from two brainstem nuclei, the dorsal raphe (DR) and the median raphe (MnR) (Jacobs SC and Azmitia, 1992). The terminal fields of these afferents in the forebrain are U both complementary and partly overlapping (Michelsen et al., 2008; Vertes and N Linley, 2008). Therefore, it is possible that the decreased serotonergic activity A in one of these nuclei can be compensated by a respective increase in another M nucleus, which might be a confounding factor when establishing causal ED relationships between epileptic seizures and central 5-HT levels (see above). However, this possibility has not been tested so far. In this study, using the PT kainate-induced post-status epilepticus (post-SE) model of TLE in rats, we CC E compared the effects of chronic epilepsy on the total number of 5-HT cells in the DRI, a subdivision of DR in which we previously observed epilepsy-related A loss of 5-HT staining, and in the MnR. We chose this model, because it reproduces most of the neuropathological features of human TLE (Buckmaster and Dudek, 1997; Maia et al., 2014, 2016). Prior studies have shown that serotoninergic fiber terminals may show different morphologies depending on the cytoarchitectonic organization of their target area and on their origin 7 (Kosofsky and Molliver, 1987; Mamounas et al., 1991). Therefore, we attempted to distinguish between the effects of epilepsy on serotonergic fibers with different morphologic characteristics by estimating the densities of fiber varicosities ranked according to their cross-sectional areas. For this analysis, the PT following structures of interest were chosen: the DG, as a possible site for RI epileptogenic processes, the infralimbic cortex (IL), implicated in the neuropathology of depression (Drevets et al., 2008), and the medial septum N U (Dannenberg et al., 2015; Freund and Gulyás, 1997). SC (MS), which has a particular role in the regulation of hippocampal excitability M A Material and Methods ED Ethical statement The handling and care of the animals were conducted according to the PT “Principles of laboratory animal care" (NIH publication No. 86-23, revised CC E 1985) and Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes. A The experimental protocol has been approved by the Ethics Committee of the Faculty of Medicine of the University of Porto and the General Veterinary Direction for the FCT application grant PTDC/SAU-NSC/115506/2009. All efforts were made to minimize the number of animals used and their suffering. 8 Animals and treatments Male Wistar rats, maintained individually under standard laboratory conditions were used in this study. The kainic acid-induced At 10 weeks of age, they were randomly divided into two groups: KA group (n=14) and control PT group (n=8). In the first group, the rats were injected with 9.5 mg/kg of KA RI (i.p., Sigma) to induce convulsive SE, which was defined as the appearance of behavioral symptoms corresponding to stage 3, 4 or 5 seizures on the Racine SC scale (Racine, 1972), i.e. bilateral forelimb clonus, rearing, and rearing with U falling. In 12 rats, SE developed within the first 2h following the treatment and N lasted for 3-6 hours. Two rats, which showed only a few wet-dog-shake seizures A (without SE), were excluded from the study. The animals were periodically M injected with saline (s.c.) during the first 48 hours of the recovery period. The ED rats that refused to eat or drink were hand-fed using a plastic syringe. Rats in the PT control group were injected with saline alone. CC E Behavioral monitoring Following the treatments, all animals were given a 4-month recovery A period. In the last week of each month, the behavior of rats in their home cages was videorecorded during the light phase of the 12-h light/dark cycle (between 08:00h and 20:00h) over a period of 5 consecutive days. The recording was performed using a digital video camera Sony DCR-SR58E (Sony Corporation, Japan), which was positioned above the cages. The video files were then 9 analyzed in a fast-motion mode by a person blind to treatment groups using the VLC media player (VideoLAN, France). In addition, the rats were daily, except weekends, observed for spontaneous behavioral seizures during 2-h intervals PT between 09:00h and 11:00h. RI Tissue preparation At 6 months of age, all rats were deeply anesthetized with pentobarbital SC (90 mg/kg) and perfused transcardially with 150 ml of 0.1 M phosphate buffer U (pH 7.4) followed by 250 ml of a fixative solution containing 4% N paraformaldehyde in phosphate buffer. The perfusion procedures were A performed between 15.00h and 17.00h. Only animals that did not show M behavioral seizures starting from 9.00h and until the time of perfusion were ED selected. Rat that had at least one seizure during the time interval from 9.00h to 15.00h were submitted to perfusion on next day. The brains were removed from PT the skulls, immersed for 2 h in the fixative, and infiltrated during 36 h in 10% CC E sucrose solution at 4°C. From each brain, two tissue blocks were dissected: one containing the forebrain and another containing the midbrain-hindbrain region. A From the forebrain blocks, the occipital poles and the anterior parts of the frontal poles were removed. The blocks were mounted on a vibratome and sectioned in the coronal plane at 40 μm. When cutting the forebrain blocks, every third section was collected in phosphate-buffered saline (PBS) to form two independent sets of adjacent sections to be used, respectively, for 10 immunostaining for serotonin transporter protein (SERT) and for Nissl staining. The tissue blocks containing the midbrain-hindbrain region were mounted on a vibratome and sectioned in the coronal plane at 40 μm. In this case, every second section was collected using a systematic random sampling procedure. PT All sections were stored until use at –20°C in cryoprotectant (30% sucrose, 30% RI ethylene glycol, 0.25 mM polyvinylpyrrolidone in PBS). All sections obtained from the midbrain-hindbrain region were SC immunostained for 5-HT as previously described (Maia et al., 2016). Briefly, U sections were washed twice in PBS, treated with H2O2 and incubated during 72 N h at 4°C with the polyclonal rabbit antibody against 5-HT (Newmarket A Scientific, United Kingdom; 1:15000 dilution in PBS). Thereafter, the sections M were washed twice and incubated with biotinylated anti-rabbit antibody (Vector ED Laboratories, Burlingame, CA, USA; 1:400 dilution in PBS). Sections were then treated with avidin-biotin peroxidase complex (Vectastain Elite ABC kit, PT Vector Laboratories; 1:800 dilution in PBS). In the two last steps, the incubation CC E was carried out for at least 1h at room temperature. Following treatment with the peroxidase complex, sections were incubated in 0.05% diaminobenzidine A solution to which H2O2 (0.01%) was added. Triton X-100 (0.5%) was added to PBS that was used in all immunoreactions and washes. Following termination of the staining procedures, sections were mounted on gelatin-coated slides, air-dried, dehydrated in a series of ethanol solutions (50%, 70%, 90% and 100%) and coverslipped using Histomount (National Diagnostics, Atlanta, GA, 11 USA). One set of sections collected from the forebrain regions was processed for immunohistochemistry for SERT. The sections were pretreated with H2O2 to inactivate endogenous peroxidase and incubated during 72h at 4°C with the PT polyclonal rabbit antibody against SERT (Newmarket Scientific, United RI Kingdom; 1:12500 dilution in PBS). Thereafter, the sections were washed twice and incubated with biotinylated anti-rabbit antibody (Vector Laboratories, SC Burlingame, CA, USA; 1:400 dilution in PBS). Sections were then treated with U avidin-biotin peroxidase complex (1:800 dilution in PBS). Following treatment N with the peroxidase complex, sections were incubated in 0.05% A diaminobenzidine (Sigma) solution to which H2O2 was added. Sections were M rinsed with PBS for at least 15 min between each step. To increase the tissue ED penetration, Triton X-100 (0.5%) was added to PBS that was used in all immunoreactions and washes. Specificity of the immune reactions was PT controlled by omitting the incubation step with primary antisera. All CC E immunochemical reactions and washings described above were carried out simultaneously in four 12-well tissue culture plates, 6 sections in each well, to A assure that staining of the sections from all animals analyzed was performed under identical conditions. Following termination of the staining procedures, sections were mounted on gelatin-coated slides and air-dried. They were then dehydrated and coverslipped. The remaining sections were stained with a Giemsa, dehydrated and coverslipped. 12 Morphological analysis The brainstem sections stained for 5-HT were visualized using an Olympus BX-53 microscope equipped with a color digital camera and a PT computer-controlled motorized stage system (MBF Bioscience, Williston, RI USA). The boundaries of the dorsal raphe nucleus and of its subdivisions, particularly DRI, and of the median raphe nucleus were consistently defined in SC the 5-HT-immunostained brainstem sections using previously described U cytoarchitectonic criteria (Hale and Lowry, 2011; Jacobs and Azmitia, 1992; N Maia et al., 2016). The stereotaxic coordinates of the DR subdivisions with A reference to a standard rat brain atlas of Paxinos and Watson (Paxinos and M Watson, 1998) have been previously determined (Abrams et al., 2004; Lowry et ED al., 2008). In this study, we estimated the rostrocaudal boundaries of the two regions of interest at approximately the following coordinates (relative to PT bregma): DRI, from -7.64 to -8.54; MnR, from -7.20 to -8.70. The total numbers CC E of neurons in DRI and MnR were estimated using the optical fractionator probe (West et al., 1991) of the Stereo Investigator software (MBF Bioscience). The A regions of interest were delineated with a 40× objective lens at all levels along the rostrocaudal axis of the brainstem. Cell counting was carried out with a 100× oil-immersion lens. Beginning at a random starting position within the region of interest, counting frames were systematically sampled using a raster pattern procedure. Tissue thickness was estimated at each counting frame and 13 guard zones of 2 μm were implemented. The nucleus of the neurons was used as the counting unit. The coefficients of error in the individual estimates, ranging between 0.05 and 0.09, were calculated according to Gundersen et al. (Gundersen et al., 1999). PT The brain sections immunostained for SERT and Nissl-stained sections RI were visualized using a Zeiss Axioskop microscope equipped with a Leica EC3 color digital camera. The boundaries of the IL, MS and HF were delineated with SC a 40× objective lens on the basis of the rat brain atlas of Paxinos and Watson U (Paxinos and Watson, 1998) and using the adjacent Nissl-stained sections. Four N consecutive SERT-stained sections containing the IL region were sampled for A each brain. From each section, 16 photomicrographs of cortical layers II and III M were taken with a 100× objective lens. A counting frame of 1274 µm2 was ED applied for each image and the number and the cross-sectional area of all varicosities that fell within this frame were measured using Fiji PT image-processing software (http://rsb.info.nih.gov/ij/). Six consecutive CC E SERT-stained sections containing the MS region were sampled for each brain. From each section, 2 photomicrographs from the central part of the MS, close to A the midline, were taken with a 100× objective lens. Two non-overlapping counting frames of 1274 µm2 were applied for each image and the number and the cross-sectional area of all varicosities that fell within each frame were measured. Six consecutive SERT-stained sections containing the middle portion of the dorsal HF were sampled for each brain. In each section, the following 14 photomicrographs were taken with a 100× objective lens: central hilus – 6 images; inner (IML), intermediate (middle, MML) and outer (OML) molecular layers (adjacent to the suprapyramidal blade of the dentate granule cell layer) – 4 images per layer. In this case, three non-overlapping counting frames of 1274 PT µm2 were applied for each photomicrograph and the number and the RI cross-sectional area of all varicosities that fell within each frame were measured. Data obtained for each region of interest were normalized to an SC arbitrary chosen area of 30000 µm2 and averaged across all sections per animal. U Only fiber varicosities with the cross-sectional area ranging from 0.04 µm2 to N 1.44 µm2 were including in the analysis. The varicosities were grouped in 35 A bins of 0.04 µm2 each according to their cross-sectional area and their density Statistical analysis ED M was plotted as a distribution histogram. PT The numbers of 5-HT-immunoreactive cells in the DRI and MnR nuclei CC E were analyzed using MANOVA and, when appropriate, the Tukey HSD test was applied. The frequency distribution of varicosities as a function of their A cross-sectional area was analyzed using repeated-measures ANOVA followed by the Newman–Keuls post hoc test. The nonlinear regression analysis of the histograms for Gaussian and bi-Gaussian distribution were performed using the least-squares fitting method. Differences were considered as significant at the p<0.05 level. 15 Results Monitoring during the recovery period PT Despite all the efforts made to reduce the mortality rate in KA-treated rats, RI from the 10 rats that had experienced convulsive SE 2 rats have died within 48 h following the treatment. Repeated generalized seizures of stage 3 to 5 on SC the Racine scale were observed in 2 rats starting from the end of the first month U of recovery and in 3 other rats starting from the third month of observation (in a N total of 5 rats). One of the KA-treated rats showed only stage 1 to 2 seizures on A the Racine scale, that is unilateral forelimb clonus and repeated stereotyped M orofacial movements, frequently coincident with periods of sudden immobility. ED No signs of behavioral seizures (stages 1-5 on the Racine scale) were detected in the remaining 2 KA-treated rats, which were therefore excluded from the PT study. Thus, the final group size for KA rats was n=6. No behavioral CC E abnormalities were observed in the control group (n=8). A Total neuron numbers in the DRI and MnR nuclei As illustrated by representative photomicrographs shown in Fig. 1A-D, the DRI region of epileptic rats had fewer 5-HT-stained neurons compared with control rats. Visual examination of these sections additionally revealed that the ventrally located MnR nucleus of post-SE rats displayed, in contrast, more 16 5-HT-immunoreactive cells than those of control rats (Fig. 1E-H). The stereological estimates of the total number of neurons in the DRI and MnR nuclei are shown graphically in Fig. 2. MANOVA of these data yielded a significant main effect of treatment (Rao R2,9=12.11, p<0.005). Furthermore, PT post-hoc tests for multiple comparisons revealed significant effects of treatment RI on the total number of 5-HT-positive neurons in both nuclei (DRI, p<0.005; MnR, p<0.001). Our estimates showed that epileptic rats, as compared to SC controls, had approximately 30% less 5-HT cells in the DRI, but twice as many N U 5-HT cells in the MnR. A Areal density of SERT-immunoreactive fiber varicosities M The rat dentate gyrus possesses a dense network of SERT-immunoreactive ED fibers, which bear numerous varicosities of different shapes and sizes (Fig. 3A). However, the distribution patterns of the fibers and varicosities appeared to PT differ considerably between control and post-SE rats (Fig. 3A,D). In particular, CC E the representative photomicrographs taken from the molecular layer of a control rat (Fig. 3B) display less large varicosities than those of an epileptic rat (Fig. A 3E). This difference was particularly evident in the images taken from the dentate hilus of the control rat (Fig. 3C) and of the epileptic rat (Fig. 3F), which showed loss of small-sized varicosities in epilepsy and a parallel increase in the density of large varicosities. The distribution histograms of the SERT-stained DG varicosities as a 17 function of their cross-sectional area are shown in Fig. 4 (IML: A,B,C; MML: D,E,F; OML: G,H,I and DG hilus: J,K,L). Repeated-measures ANOVA revealed that there were significant effects of size of varicosities on their density in all DG subdivisions (IML, F34,340=102.37, p<0.00001; MML, F34,340=87.46, PT p<0.00001; OML, F34,340=80.31, p<0.00001; hilus, F34,340=75.40, p<0.00001). RI The main effect of treatment on the density of SERT varicosities was significant only in the DG hilus (F1,10=75.40, p<0.005). However, the interaction between SC the two effects was significant both in the dentate hilus and in the molecular U layer (hilus, F34,340=9.18, p<0.00001; IML, F34,340=3.54, p<0.0001; MML, N F34,340=1.65, p<0.05; OML, F34,340=1.67, p<0.05). Post-com comparisons A applied on these interactions showed that there was a significant rightward shift M in the density-size distribution of SERT-immunoreactive varicosities in all DG ED regions. Both visual examination and nonlinear regression analysis of these histograms revealed that the density-size distributions of SERT varicosities PT could not be described by a single Gaussian equation, but, instead, they could be CC E well fitted with a two-component Gaussian function (r2 range 0.93-0.99). This analysis additionally demonstrated that the entire population of the SERT A varicosities in the DG is actually composed of two independent subpopulations with the mean cross-sectional areas of 0.37-0.39 µm2 and 0.55-0.64 µm2, respectively (Fig. 4, panels B,E,H,K). Furthermore, the results of this analysis also confirmed that the development of epilepsy was associated with a loss of the small-sized varicosities and generation of new larger-sized varicosities (Fig. 18 4, panels C,F,I,L). Fig. 5 and Fig. 6A,B,C summarize quite similar results for the IL cortex. The representative photomicrograph taken from IL layers II/III of an epileptic rat shows the presence of relatively large SERT-stained varicosities, while PT practically no varicosities of this size can be seen in the respective image taken RI from a control rat (Fig. 5C,D). However, the density of the small-sized varicosities does not differ appreciably between the control rat and the epileptic SC rat. The distribution histograms of the SERT-stained varicosities as a function U of their cross-sectional area are shown in Fig. 6A. Repeated-measures ANOVA N revealed that, in the IL cortex, there were a significant main effect of treatment A on the density of SERT varicosities (F1,10=5.26, p<0.05), a significant effects of M size of varicosities on their density (F34,340=59.61, p<0.00001) and a significant ED interaction between the two effects (F34,340=2.38, p<0.00001). Post-hoc comparisons applied on these interactions showed that there was a significant PT rightward shift in the density-size distribution of SERT-immunoreactive CC E varicosities (Fig. 6A). Similarly to what was found in the DG, the density-size distributions of SERT varicosities could be best fitted with a bi-Gaussian A equation (r2 range 0.98-0.99). In addition, the results of the regression analysis suggested that, in this region, the density of the small-sized varicosities, rather than of larger varicosities, was elevated post-SE (Fig. 6B,C). However, this impression is partly misleading, because the two populations of varicosities detected in the IL cortex differ substantially between the control rats and the 19 epileptic rats. More specifically, the mean cross-sectional areas of the varicosities found in post-SE rats were greater than in control rats (0.36 and 0.74 µm2 versus 0.29 and 0.55 µm2, respectively), which is consistent with the overall rightward shift in the distribution histogram of SERT varicosities (Fig. PT 6A). RI Fig. 7 shows representative images of the SERT-stained sections cut through the MS region of a control rat and of an epileptic rat. As can be seen in SC the inset shown in Fig. 7D, the epileptic state was associated with an increase in U the density of larger-sized varicosities when compared to controls (Fig. 7C). N The distribution histograms of SERT varicosities in the MS of control and A epileptic rats are shown in Fig. 6D. Repeated-measures ANOVA applied on M these data revealed that there were a significant main effect of treatment on the ED density of SERT varicosities (F1,10=9.73, p<0.01), a significant effects of size of varicosities on their density (F34,340=32.72, p<0.00001) and a significant PT interaction between the two effects (F34,340=1.60, p<0.05). Epileptic rats, when CC E compared to controls, had significantly more varicosities with the cross-sectional area of 0.56-1.12 µm2, suggesting a significant rightward shift in A the density-size distribution of SERT-positive varicosities in the MS. Nonlinear regression analysis of these histograms revealed the existence of two independent subpopulations of SERT varicosities with the mean cross-sectional areas of 0.37 µm2 and 0.73-0.74 µm2, respectively (r2=0.94-0.97; Figs. 6E,F). The effect of the epileptic state on the SERT varicosities in the MS was similar 20 to that described for the DG, that is, it produced loss of the small-sized varicosities and greatly increased the number of large varicosities (Figs. 6E,F). PT Discussion Two key observations emerge from this experiment. First, the total number RI of the serotonin-IR neurons in the rostrally projecting raphe nuclei does not SC appear to decline in this model of epilepsy, despite their significant loss in DRI U region, because in another raphe subdivision, the MnR nucleus, the number of N serotonin-stained cells was increased. The second finding was the A reorganization of the 5-HT network in all target areas analyzed, that is, the DG, M IL cortex and MS, as indicated by the rightward shift of the density-size ED distribution histograms of SERT-stained fiber varicosities. In other words, this finding suggest that chronic epilepsy was associated with a loss of serotonergic PT fibers with small-sized varicosities, which was commensurate with increased CC E density of larger-sized varicosities. We previously reported that development of epilepsy in the KA-induced A SE model may lead to loss of 5-HT staining in the DRI nucleus (Maia et al., 2016), which is known to project to various limbic regions implicated in epileptogenesis and neuropathology of cognitive and affective disorders (Azmitia et al., 1996; Drevets et al., 2008). Mazarati et al. (Mazarati et al., 2008) have shown that serotonergic neurotransmission between the raphe nuclei 21 and the hippocampus is compromised in the Li-pilocarpine model of epilepsy. However, a number of studies failed to detect epilepsy-related depletion of hippocampal serotonin in both animal models (Cavalheiro et al., 1994; Shouse et al., 2001; Szyndler et al., 2005; Szyndler et al., 2010) and human epilepsy PT (Louw et al., 1989). This discrepancy prompted us to re-examine the effects of RI chronic epilepsy on serotonergic neurons of the rostral raphe nuclei, this time including the MnR nucleus. Although we expected to detect some SC treatment-related morphological alterations in the MnR, given its connectivity U with the hippocampal formation and other midline regions of the forebrain N (Vertes and Linley, 2008), the results were somewhat surprising as they A revealed a two-fold increase in the total number of 5-HT-stained cells in this M nucleus. In fact, after summing the cell numbers counted in both nuclei, the ED resulting values did not differ between the two groups, suggesting that adding 5-HT-rich cells to the MnR nucleus fully compensated for their loss in DRI, at PT least in numerical terms. CC E To test potential links between the observed phenomena in the raphe nuclei and changes in serotonergic transmission in their target forebrain areas, we A performed a morphometric analysis of fiber varicosities immunoreactive to the serotonin transporter (SERT) in the DG, IL cortex and MS. In our pilot immunocytochemical experiments, we found that 5-HT is well retained in cell bodies, but less well in fibers and fiber varicosities, probably, due to a rapid removal during the perfusion and/or histological procedures. Therefore, to 22 ensure uniform staining of serotonergic fibers across subjects, we opted to use an anti-SERT antibody. The results of the morphometric analysis, that is, the areal density of the SERT-stained varicosities and their cross-sectional area, were plotted as the distribution histograms. It was found that, in either target PT area, the SERT-stained varicosities are represented by two independent RI subpopulations characterized by distinct cross-sectional areas. Interestingly, the effects of epilepsy on these subpopulations were opposite: the small-sized SC varicosities were generally reduced in number in epileptic rats, whereas the U density of the larger-sized varicosities tended to increase. Although it is difficult N based only on these data to explain the epilepsy-related shift toward larger-sized A varicosities, several possibilities can be considered. Firstly, there is a large body M of evidence showing that morphological features of 5-HT terminals vary ED substantially depending on their target field, e.g., region of the brain (for review, see Descarries et al., 2010). In addition, the ratio between the 5-HT PT terminals that establish synaptic contacts versus those that act via the “volume CC E transmission” appears also to depend on the local environment in the target domain (Descarries et al., 2010; Gaspar and Lillesaar, 2012). Of note, it has A been shown that 5-HT terminals forming junctional complexes are in average larger than non-junctional terminals (Oleskevich et al., 1991). Thus, it is possible, that diffuse KA-induced neurodegeneration changed neuronal and glial microenvironment in such a way that serotonergic volume transmission became deficient in several brain regions. Secondly, because it has been 23 previously shown that a subpopulation of the MnR neurons possesses axons bearing distinctly large spherical varicosities, 2-5 μm in diameter (Kosofsky and Molliver, 1987; Mamounas et al., 1991), it is possible that increased number of larger-sized varicosities in epileptic rats is due to the fiber sprouting specifically PT in this neuronal population. However, this explanation is not fully plausible, RI given that our staining method did not reveal distinctive large SERT-IR varicosities in either brain region analyzed. Thirdly, the existence of varicosities SC with different sizes may be related to the fact that many serotonergic neurons U located in the ventral DR, DRI and MnR coexpress vesicular glutamate N transporter 3 (VGLUT3) (Hioki et al., 2010; Jackson et al., 2009), i.e. possess A fiber terminals with both 5-HT- and glutamate-containing vesicles, whereas M fiber terminals of other cells bear only 5-HT vesicles (Wang et al., 2019). Both ED types of fibers project to various forebrain regions, including the hippocampus and medial septum (Jackson et al., 2009), but the 5-HT-plus-Glu terminals PT appear to make synaptic junctions more frequently than the 5-HT-only terminals CC E (Wang et al., 2019). Therefore, it is possible that some unidentified epilepsyrelated factors can downregulate 5-HT-only neurons and upregulate those which A coexpress 5-HT and glutamate. Finally, the relative prevalence of larger varicosities in the rat epileptic brain can be related to some general neuroinflammatory or degenerative processes in the nervous tissue. For example, SERT-IR fibers bearing enlarged, twisted or swollen varicosities have been described in the brain of patients with some neurodegenerative diseases, 24 such as Parkinson’s disease, frontal lobe dementia and Lewy-body dementia (Azmitia and Nixon, 2008). However, robust sprouting of serotonergic fibers possessing normal morphology has been found in the pallidum of Parkinsonian monkeys (Gagnon et al., 2018). Whether similar changes in 5-HT fibers occur PT in the brain of epileptic patients remains to be demonstrated. RI The anatomical plasticity of the 5-HT network in the adult brain has been first discovered after partial removal of serotonergic afferents from the SC cingulum bundle (Azmitia et al., 1978; Zhou and Azmitia, 1986). It has been U further shown that nonspecific excitotoxic lesions of either the striatum, N hippocampus (Zhou et al., 1995), or basal forebrain (Harkany et al., 2000) A induce a rapid sprouting of serotonergic fibers into the lesioned zone. M Subsequent studies revealed that at least three trophic factors present in ED damaged nervous tissue, S-100 (Zhou et al., 1995), BDNF and NT-3 (Grider et al., 2005), are capable of triggering the outgrowth of serotonergic axons. PT Consistent with this, we observed that epilepsy was associated with a noticeable CC E increase in the density of SERT varicosities in the dentate hilus, an area known for its heightened vulnerability to seizure-induced damage. However, we also A observed the redistribution of SERT-stained varicosities in two other brain regions, the IL cortex and MS, which undergo less dramatic anatomical changes in epilepsy. Thus, the increased density of the larger-sized varicosities in the IL and MS may be partly due to collateral sprouting of serotonergic axons whose primary target was the hippocampal region. This hypothesis is supported by the 25 neuroanatomical evidence showing that a part of the MnR neurons project to both the hippocampus and the MS (Acsady et al., 1996). In fact, collateral sprouting of 5-HT axons to the hippocampus and MS has been previously demonstrated in the lesions studies (Ueda et al., 1991; Zhou and Azmitia, 1986). PT A more widespread reorganization of the 5-HT network was found following β– RI amyloid-induced lesions of the basal forebrain cholinergic neurons (Harkany et al., 2000). Furthermore, it has been hypothesized that the lesion-induced SC redistribution of the forebrain 5-HT afferents can contribute, at least partly, to U respective behavioral abnormalities, including increased anxiety and changes in N spontaneous behavior (Harkany et al., 2000; Harkany et al., 2001). It will be A interesting to see in future studies whether this explanation is applicable to the M behavioral changes observed in the KA model of epilepsy, namely, cognitive ED deficits, altered anxiety levels (Maia et al., 2014), and depression-like symptoms (Maia et al., 2016). PT The increased number of 5-HT-immunoreactive cells in the MnR nucleus CC E of epileptic rats is likely to reflect another facet of functional neuroplasticity of central serotonergic systems. It has been previously shown in rats that A tryptophan hydroxylase (TPH), the rate-limiting enzyme in 5-HT synthesis, is downregulated by adrenalectomy and restored by subsequent treatment with dexamethasone (Azmitia et al., 1993). A number of recent studies have shown that the expression of mRNA for two known TPH isoforms, TPH-1 and TPH-2, is strongly affected by the adrenocortical hormones (Clark et al., 2005) and by 26 the exposure to chronic stress (Abumaria et al., 2007). In addition, there is evidence that glucocorticoids are capable of modulating the functional state of the presynaptic 5-HT1A autoreceptor located on the membrane of serotonergic neurons (Man et al., 2002; Pineda et al., 2011). Dysregulation of the PT hypothalamo–pituitary–adrenal axis, characterized by elevated circulating levels RI of glucocorticoids, is well documented in experimental models of epilepsy (Mazarati et al., 2009; Pineda et al., 2010) and in patients with TLE (Kanner, SC 2012). Thus, it appears plausible that chronically elevated levels of the U adrenocortical hormones, by acting on intracellular targets, can modify the N neurochemical phenotype of some of the raphe neurons. A There are several caveats that need to be kept in mind when interpreting the M present findings. One of them is related to the fact that both the DR and MnR ED nuclei contain neurons of distinct chemotypes, i.e., serotonergic, GABAergic, glutamatergic and dopaminergic (Hale and Lowry, 2011; Lowry et al., 2008). PT Complicating the picture, many serotonergic neurons, as already mentioned CC E above, use glutamate as a co-neurotransmitter (Jackson et al., 2009; Wang et al., 2019). However, in this study we estimated only the number of 5-HT-IR cells, A which leaves some uncertainty regarding the fate of those neurons that were undetected in DRI of epileptic rats. Indeed, this result can indicate either inhibition of 5-HT synthesis in DRI neurons, e.g., due to overactivation of 5HT1A autoreceptors, or their actual death. To clarify this issue, it would be important to examine whether induction of epilepsy in this model produces 27 changes in the total DRI cell number, as wells as in numbers of cells of other chemical phenotypes. Similarly, the increased number of 5-HT-IR cells in the MnR can be due to increased 5-HT synthesis in a subset of dual-transmitter neurons, that is, neurons, which co-express SERT and VGLUT3. Indeed, PT growing body of evidence indicates that the dual-transmitter phenotype is RI plastic and can be modulated under both normal and pathological conditions (Prado et al., 2013; Vaaga et al., 2014). Should this be the case, we can expect a SC decrease in the number of VGLUT3-IR cells in the MnR of epileptic rats (and U respective increase in the DRI), suggesting even more profound neurochemical N reorganization in the DR-MnR region. A Another limitation of this study is that it does not shed light on the M molecular mechanisms underlying plasticity of 5-HT neurons and their ED forebrain projections. One possibility, as discussed above, might be related to the ability of the adrenocortical hormones, whose levels are chronically PT increased in epilepsy, to modulate the functional state of the 5-HT1A CC E autoreceptors (Clark et al., 2005; Man et al., 2002; Pineda et al., 2011). Thus, suppression of the enhanced activity of the hypothalamo–pituitary–adrenal axis A might have a normalizing effect on serotonergic nuclei, possibly leading to amelioration of seizures and depressive symptoms. However, there are other possibilities. For example, there is evidence that the brainstem noradrenergic and cholinergic cell groups are involved in epilepsy (Englot et al., 2017; Le Saux et al., 2002; Soares et al., 2018) and that they are closely interconnected 28 with both DRI and MnR (Lechin et al., 2006; Vertes and Linley, 2008). Therefore, it is likely that pharmacological manipulations employing adrenergic and cholinergic agents are capable of indirectly stabilizing the 5-HT synthesis and release in DR/MnR neurons. Thus, further studies are warranted to unravel PT the role of the hypothalamo–pituitary–adrenal axis and of the adrenergic and RI cholinergic neurotransmissions in modulating the functional state and neurochemical plasticity of serotonergic neurons. SC In summary, the present study was designed to test the hypothesis that U chronic epilepsy can trigger structural reorganization of the ascending N serotonergic pathways in the rat brain. We found that epileptic rats had fewer A 5-HT-IR cells in the DRI region, whereas their number in the MnR nucleus was M considerably increased. We additionally observed that the density of the ED SERT-stained small-sized fiber varicosities was decreased in the DG, IL cortex and MS, while that of the larger-sized varicosities was increased. The global PT reorganization of the 5-HT network can be neuroprotective in nature, but it may CC E also be a contributing factor in epileptogenesis as well as in the development of A comorbid neuropsychiatric disorders. Author Contributions: General idea and study design: G.M., J.S. and N.L. Data acquisition and analysis: G.M., S.A., J.L., C.B. and J.S. Writing of the manuscript: G.M., J.S. and N.L. Obtaining funding. G.M and N.L. 29 Conflicts of Interest Authors declare no conflicts of interest related to this study. PT Acknowledgments This work was supported by FEDER founds through the Programa RI Operacional Factores de Competitividade – COMPETE and National Funds SC through FCT – Fundação para a Ciência e a Tecnologia within the scope of the U Project PTDC/SAU-SNC/115506/2009-FCOMP 01-0124-FEDER-015919. This N work was also supported by FCT grant (SFRH/BD/87886/2012) to Joana A CC E PT ED M A Soares. 30 References Abrams, J.K., et al., 2004. Anatomic and functional topography of the dorsal raphe nucleus. Ann N Y Acad Sci. 1018, 46-57. Abumaria, N., et al., 2007. Effect of chronic citalopram on serotonin-related and PT stress-regulated genes in the dorsal raphe nucleus of the rat. Eur RI Neuropsychopharmacol. 17, 417-429. Acsady, L., et al., 1996. Topographic distribution of dorsal and median raphe SC neurons with hippocampal, septal and dual projection J Acta Biologica U Hungarica. 47, 9-19. N Alessio, A., et al., 2004. Differences in memory performance and other clinical A characteristics in patients with mesial temporal lobe epilepsy with and M without hippocampal atrophy. Epilepsy Behav. 5, 22-27 ED Azmitia, E.C., Buchan, A.M., Williams, J.H., 1978. Structural and functional restoration by collateral sprouting of hippocampal 5-HT axons. Nature. PT 274, 374. CC E Azmitia, E.C., Liao, B., Chen, Y., 1993. Increase of tryptophan hydroxylase enzyme protein by dexamethasone in adrenalectomized rat midbrain. J A Neurosci. 13, 5041-5055. Azmitia, E.C., et al., 1996. Cellular localization of the 5-HT1A receptor in primate brain neurons and glial cells. Neuropsychopharmacology. 14, 35 Azmitia, E.C., Nixon, R., 2008. Dystrophic serotonergic axons in neurodegenerative diseases. Brain Res. 1217, 185-94. 31 Bagdy, G., et al., 2007. Serotonin and epilepsy. J Neurochem. 100, 857-873. Barnes, N.M., Sharp, T., 1999. A review of central 5-HT receptors and their function. Neuropharmacology. 38, 1083-1152. Browning, R.A., Hoffmann, W.E., Simonton, R.L., 1978. Changes in seizure PT susceptibility after intracerebral treatment with 5, 7‐dihydroxytryptamine: RI role of serotonergic neurons. Ann N Y Acad Sci. 305, 437-456. SC Buckmaster, P.S., Dudek, F.E., 1997. Neuron loss, granule cell axon U reorganization, and functional changes in the dentate gyrus of epileptic N kainate‐treated rats. J Comp Neurol. 385, 385-404 A Buckmaster, P.S., Zhang, G.F., Yamawaki, R., 2002. Axon sprouting in a model M of temporal lobe epilepsy creates a predominantly excitatory feedback ED circuit. J Neurosci. 22, 6650-6658. Cavalheiro, E., et al., 1994. Spontaneous recurrent seizures in rats: amino acid PT and monoamine determination in the hippocampus. Epilepsia. 35, 1-11. CC E Cavanagh, J.B., Meyer, A., 1956. Aetiological aspects of Ammon's horn sclerosis associated with temporal lobe epilepsy. British Medical Journal. A 2, 1403. Celada, P., Puig, M., Artigas, F., 2013. Serotonin modulation of cortical neurons and networks. Front Integr Neurosci. 7, 25. Clark, J.A., et al., 2005. Differential hormonal regulation of tryptophan hydroxylase-2 mRNA in the murine dorsal raphe nucleus. Biol Psych. 57, 32 943-946. da Fonseca, N.C., et al., 2015. Hippocampal serotonin depletion is related to the presence of generalized tonic–clonic seizures, but not to psychiatric disorders in patients with temporal lobe epilepsy. Epilepsy Res. 111, 18- PT 25. RI Dannenberg, H., et al., 2015. Synergy of direct and indirect cholinergic septohippocampal pathways coordinates firing in hippocampal networks. J SC Neurosci. 35, 8394-8410. U Descarries, L., Riad, M., Parent, M., 2010. Ultrastructure of the serotonin N innervation in the mammalian central nervous system. In: Handbook of A behavioral neuroscience. Vol. 21. Elsevier, pp. 65-101. M Drevets, W.C., Price, J.L., Furey, M.L., 2008. Brain structural and functional ED abnormalities in mood disorders: implications for neurocircuitry models of depression. Brain Struct Funct. 213, 93-118. PT Engel, J.J., 1989. Seizures and epilepsy, Vol., Philadelphia, FA Davis CC E Company. Englot, D.J., et al., 2017. Functional connectivity disturbances of the ascending A reticular activating system in temporal lobe epilepsy. J Neurol Neurosurg Psychiatry. 88, 925-932. Favale, E., et al., 1995. Anticonvulsant effect of fluoxetine in humans. Neurology. 45, 1926-1927. Favale, E., et al., 2003. The anticonvulsant effect of citalopram as an indirect 33 evidence of serotonergic impairment in human epileptogenesis. Seizure. 12, 316-318. Freund, T.F., Gulyás, A.I., 1997. Inhibitory control of GABAergic interneurons in the hippocampus. Can J Physiol Pharmacol. 75, 479-487. PT Gagnon, D., et al., 2018. Evidence for Sprouting of Dopamine and Serotonin RI Axons in the Pallidum of Parkinsonian Monkeys. Front Neuroanat. 12, 38. SC Gaspar, P., Lillesaar, C., 2012. Probing the diversity of serotonin neurons. U Philos Trans R Soc Lond B Biol Sci. 367, 2382-94. N Grider, M., et al., 2005. In situ expression of brain‐ derived neurotrophic factor A or neurotrophin‐ 3 promotes sprouting of cortical serotonergic axons M following a neurotoxic lesion. J Neurosci Res. 82, 404-412. ED Gundersen, H.J.G., et al., 1999. The efficiency of systematic sampling in stereology—reconsidered. J Microsc 193, 199-211. PT Hale, M.W., Lowry, C.A., 2011. Functional topography of midbrain and CC E pontine serotonergic systems: implications for synaptic regulation of serotonergic circuits. Psychopharmacology (Berl). 213, 243-264. A Harkany, T., et al., 2000. Increased amyloid precursor protein expression and serotonergic sprouting following excitotoxic lesion of the rat magnocellular nucleus basalis: neuroprotection by Ca2+ antagonist nimodipine. Neuroscience. 101, 101-114. Harkany, T., et al., 2001. β-Amyloid (1-42)-induced cholinergic lesions in rat 34 nucleus basalis bidirectionally modulate serotonergic innervation of the basal forebrain and cerebral cortex. Neurobiol Dis. 8, 667-678. Hasler, G., et al., 2007. 5-HT1A receptor binding in temporal lobe epilepsy patients with and without major depression. Biol Psych. 62, 1258-1264. PT Hauser, W.A., 2001. Epidemiology of intractable epilepsy. In: Epilepsy surgery. RI Vol., C.Y. Lurdes HO, editors. Lippincott Williams and Wilkins, SC Philadelphia, pp. 55.61. Heng, K., Haney, M.M., Buckmaster, P.S., 2013. High‐dose rapamycin blocks A epilepsy. Epilepsia. 54, 1535-1541 N U mossy fiber sprouting but not seizures in a mouse model of temporal lobe M Hernandez, E.J., Williams, P.A., Dudek, F.E., 2002. Effects of fluoxetine and ED TFMPP on spontaneous seizures in rats with pilocarpine‐induced epilepsy. Epilepsia. 43, 1337-1345. PT Hioki, H., et al., 2010. Vesicular glutamate transporter 3-expressing CC E nonserotonergic projection neurons constitute a subregion in the rat midbrain raphe nuclei. J Comp Neurol. 518, 668-86. A Jackson, J., Bland, B.H., Antle, M.C., 2009. Nonserotonergic projection neurons in the midbrain raphe nuclei contain the vesicular glutamate transporter VGLUT3. Synapse. 63, 31-41. Jacobs, B.L., Azmitia, E.C., 1992. Structure and function of the brain serotonin system. Physiol Rev. 72, 165-229 35 Kanner, A.M., 2012. Can neurobiological pathogenic mechanisms of depression facilitate the development of seizure disorders? Lancet Neurol. 11, 10931102. Kosofsky, B.E., Molliver, M.E., 1987. The serotoninergic innervation of PT cerebral cortex: different classes of axon terminals arise from dorsal and RI median raphe nuclei. Synapse. 1, 153-168. Le Saux, F., Besson, M.-J., Maurin, Y., 2002. Abnormal postnatal ontogeny of SC the locus coeruleus in the epileptic mutant mouse quaking. Brain Res Dev U Brain Res. 136, 197-205. N Lechin, F., van der Dijs, B., Hernández-Adrián, G., 2006. Dorsal raphe vs. A median raphe serotonergic antagonism. Anatomical, physiological, M behavioral, neuroendocrinological, neuropharmacological and clinical ED evidences: relevance for neuropharmacological therapy. Prog Neuropsychopharmacol Biol Psychiatry. 30, 565-585. PT Lewis, J., Westerberg, V., Corcoran, M.E., 1987. Monoaminergic correlates of CC E kindling. Brain Res. 403, 205-212. Louw, D., et al., 1989. A study of monoamine metabolism in human epilepsy. A Can J Neurol Sci. 16, 394-397. Lowry, C.A., et al., 2008. Topographic organization and chemoarchitecture of the dorsal raphe nucleus and the median raphe nucleus. In: Serotonin and sleep: molecular, functional and clinical aspects. Springer, pp. 25-67. Lynch, M., Sutula, T., 2000. Recurrent excitatory connectivity in the dentate 36 gyrus of kindled and kainic acid–treated rats. J Neurophysiol. 83, 693704 Maia, G.H., et al., 2014. Loss of hippocampal neurons after kainate treatment correlates with behavioral deficits. PLoS One. 9, e84722. PT Maia, G.H., et al., 2016. Altered taste preference and loss of limbic-projecting RI serotonergic neurons in the dorsal raphe nucleus of chronically epileptic rats. Behav Brain Res. 297, 28-36. SC Mamounas, L.A., et al., 1991. Dual serotoninergic projections to forebrain in N U the rat: Morphologically distinct 5‐HT axon terminals exhibit differential A vulnerability to neurotoxic amphetamine derivatives. J Comp Neurol. M 314, 558-586. ED Man, M.-S., Young, A.H., McAllister-Williams, R.H., 2002. Corticosterone modulation of somatodendritic 5-HT1A receptor function in mice. J PT Psychopharmacol. 16, 245-252. CC E Margerison, J.H., Corsellis, J.A.N., 1966. Epilepsy and the temporal lobes. Brain. 89, 499-530 A Mazarati, A., et al., 2008. Depression after status epilepticus: behavioural and biochemical deficits and effects of fluoxetine. Brain. 131, 2071-2083. Mazarati, A.M., et al., 2009. Elevated plasma corticosterone level and depressive behavior in experimental temporal lobe epilepsy. Neurobiol Dis. 34, 457-461. 37 Michelsen, K.A., Prickaerts, J., Steinbusch, H.W.M., 2008. The dorsal raphe nucleus and serotonin: implications for neuroplasticity linked to major depression and Alzheimer's disease. Prog Brain Res. 172, 233-264 Nissinen, J., et al., 2000. A new model of chronic temporal lobe epilepsy PT induced by electrical stimulation of the amygdala in rat. Epilepsy Res. 38, RI 177-205 Oleskevich, S., et al., 1991. Ultrastructural features of the serotonin innervation SC in adult rat hippocampus: an immunocytochemical description in single U and serial thin sections. Neuroscience. 42, 777-91. A Academic Press, San Diego. N Paxinos, G., Watson, C., 1998. The rat brain in stereotaxic coordinates, Vol., M Piazzini, A., et al., 2001. Depression and anxiety in patients with epilepsy. ED Epilepsy Behav. 2, 481-489 Pineda, E., et al., 2010. Comorbidity between epilepsy and depression: PT experimental evidence for the involvement of serotonergic, CC E glucocorticoid, and neuroinflammatory mechanisms. Epilepsia 51, 110114. A Pineda, E.A., et al., 2011. Plasticity of presynaptic and postsynaptic serotonin 1A receptors in an animal model of epilepsy-associated depression. Neuropsychopharmacology. 36, 1305. Prado, V.F., et al., 2013. Regulation of cholinergic activity by the vesicular acetylcholine transporter. Biochem J. 450, 265-274. 38 Racine, R., 1972. Modification of seizure activity by electrical stimulation: II. Motor seizure. Electroencephalogr Clin Neurophysiol. 32, 281-294. Racine, R., Coscina, D.V., 1979. Effects of midbrain raphe lesions or systemic p-chlorophenylalanine on the development of kindled seizures in rats. PT Brain Res Bull. 4, 1-7. RI Ropert, N., Guy, N., 1991. Serotonin facilitates GABAergic transmission in the CA1 region of rat hippocampus in vitro. J Physiol. 441, 121-136. SC Shouse, M.N., et al., 2001. Long-lasting effects of feline amygdala kindling on U monoamines, seizures and sleep. Brain Res. 892, 147-165. N Soares, J.I., et al., 2018. The pedunculopontine and laterodorsal tegmental A nuclei in the kainate model of epilepsy. Neuroscience Letters. 672, 90-95. M Specchio, L.M., et al., 2004. Citalopram as treatment of depression in patients ED with epilepsy. Clin Neuropharmacol. 27, 133-136. Sutula, T., et al., 1988. Synaptic reorganization in the hippocampus induced by PT abnormal functional activity. Science. 239, 1147-1150 CC E Sutula, T., et al., 1989. Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Ann Neurol. 26, 321-330 A Szyndler, J., et al., 2005. Behavioral, biochemical and histological studies in a model of pilocarpine-induced spontaneous recurrent seizures. Pharmacol Biochem Behav. 81, 15-23. Szyndler, J., et al., 2010. Time course of changes in the concentrations of monoamines in the brain structures of pentylenetetrazole-kindled rats. J 39 Neural Transm. 117, 707-718. Tellez‐Zenteno, J.F., et al., 2007. Psychiatric comorbidity in epilepsy: a population‐based analysis. Epilepsia. 48, 2336-2344 PT Theodore, W.H., et al., 2012. Serotonin 1A receptors, depression, and memory RI in temporal lobe epilepsy. Epilepsia. 53, 129-133. Ueda, S., Sano, Y., Kawata, M., 1991. Collateral sprouting of serotonergic U hippocampal lesions. Brain Res. 556, 329-332. SC fibers in the cingulate cortex and the septum following cortical- N Vaaga, C.E., Borisovska, M., Westbrook, G.L., 2014. Dual-transmitter neurons: M Neurobiol. 29, 25-32. A functional implications of co-release and co-transmission. Curr Opin ED Varga, V., et al., 2009. Fast synaptic subcortical control of hippocampal circuits. Science. 326, 449-453. PT Vermoesen, K., et al., 2012. The antidepressants citalopram and reboxetine CC E reduce seizure frequency in rats with chronic epilepsy. Epilepsia. 53, 870878. A Vertes, R.P., Linley, S.B., 2008. Efferent and afferent connections of the dorsal and median raphe nuclei in the rat. In: Serotonin and sleep: molecular, functional and clinical aspects. Springer, pp. 69-102. Vismer, M.S., et al., 2015. The piriform, perirhinal, and entorhinal cortex in seizure generation. Front Neural Circuits. 9, 27 40 Wang, H.L., et al., 2019. Dorsal Raphe Dual Serotonin-Glutamate Neurons Drive Reward by Establishing Excitatory Synapses on VTA Mesoaccumbens Dopamine Neurons. Cell Rep. 26, 1128-1142 e7. West, M., Slomianka, L., Gundersen, H.J.G., 1991. Unbiased stereological PT estimation of the total number of neurons in the subdivisions of the rat RI hippocampus using the optical fractionator. Anat Rec. 231, 482-497. Yamawaki, R., Thind, K., Buckmaster, P.S., 2015. Blockade of excitatory SC synaptogenesis with proximal dendrites of dentate granule cells following U rapamycin treatment in a mouse model of temporal lobe epilepsy. J Comp N Neurol. 523, 281-297 A Zeng, L.-H., Rensing, N.R., Wong, M., 2009. The mammalian target of M rapamycin signaling pathway mediates epileptogenesis in a model of ED temporal lobe epilepsy. J Neurosci. 29, 6964-6972 Zhou, F.C., Azmitia, E.C., 1986. Induced homotypic sprouting of serotonergic PT fibers in hippocampus. II. An immunocytochemistry study. Brain Res. CC E 373, 337-348. Zhou, F.C., Azmitia, E.C., Bledsoe, S., 1995. Rapid serotonergic fiber sprouting A in response to ibotenic acid lesion in the striatum and hippocampus. Brain Res Dev Brain Res. 84, 89-98. 41 Figure Legends Figure 1. Representative photomicrographs of coronal sections from the midbrain of a control (A, C, E, G) and an epileptic (B, D, F, H) rats, showing main subdivisions of the dorsal raphe nucleus (A, B) and the median raphe PT nucleus (E, F). The sections were immunostained using a primary antibody for RI 5-HT. High-power photomicrographs of the interfascicular region of the control rat (inset in A) and of the epileptic rat (inset in B) are shown in (C) and (D), SC respectively. High-power images taken from the central part of the median U raphe nucleus of the control rat (inset in E) and of the epileptic rat (inset in F) N are shown in (G) and (H). Note that the density of the 5-HT-immunostained A cells appears to be lower in the DRI of the epileptic rat when compared with the M control rat. The opposite can be observed in the median raphe, where the ED density of stained cells is markedly increased in the epileptic rat (H versus G). Abbreviations: Aq, cerebral aqueduct; DRD, dorsal division of DR; DRV, PT ventral division of DR; DRVL, ventrolateral division of DR; DRI, CC E interfascicular division of DR; mlf, medial longitudinal fasciculus; MnR, median raphe nucleus. Scale bar = 200 µm (A, B, E, F) and 25 µm (C, D, G, A H). A ED PT CC E PT RI SC U N A M 42 43 Figure 2. Total number of the DRI and MnR cells immunoreactive to 5-HT in control and post-SE (kainate-treated) rats. Chronic epilepsy was associated with loss of 5-HT-stained neurons in the DRI (approximately 30%, p<0.005 vs. control). However, in the MnR nucleus epileptic rats had more than twice as PT many 5-HT-stained cells compared with controls (p<0.001) Columns represent CC E PT ED M A N U SC RI means and vertical bars represent one SD. A Figure 3. Representative photomicrographs of coronal sections from the dentate gyrus (DG) of a control (A, B, C) and an epileptic (D, E, F) rats. The sections were immunostained using a primary antibody for serotonin transporter (SERT). The boxed insets shown in (A) and (B) indicate regions of the dentate molecular 44 layer (ML, including the inner, intermediate and outer sublayers) and hilus, from which the respective high-power images were taken (B, E – dentate molecular layer and C, F – hilar region). Note that the section derived from the epileptic rat show a higher density of larger-sized SERT-stained fiber PT varicosities (indicated by rows) compared with the control. Other abbreviations: RI GL, granular layer; CA3, pyramidal subfield of the hippocampus proper. Scale A CC E PT ED M A N U SC bar = 200 µm (A, D) and 25 µm (B, C, E, F). Figure 4. Graphic representation of the density of fiber varicosities 45 immunostained for serotonin transporter (SERT) as a function of their crosssectional area in the inner molecular layer (IML - A, B, C), middle molecular layer (MML - D, E, F), outer molecular layer (OML - G, H, I), and dentate gyrus hilus (DG - J, K, L). Statistical analysis of these data revealed a PT significant rightward shift in the density-size distribution of SERT varicosities RI in epileptic rats versus control rats (panels A, D, G, J; *p<0.05). Regression analysis of the data revealed that the density-size distribution of SERT SC varicosities could be best fitted with a bi-Gaussian function both in the control U group (green line in B, E, H, K, also shown in A, D, G, J) and in the post-SE N group (orange line in C, F, I, L, also shown in A, D, G, J), suggesting the A existence of two distinct size populations of SERT varicosities. The relative M contribution of the smaller-sized varicosities (blue line) and the larger-sized ED varicosities (red line) to the overall population of SERT-positive fiber varicosities are separately shown for the control group (B, E, H, K) and the PT epileptic group (C, F, I, L). The coefficients of correlations (r2) are shown in A CC E respective panels. A ED PT CC E PT RI SC U N A M 46 47 Figure 5. The laminar organization of the infralimbic (IL) prefrontal cortex (Nissl stain) and respective laminar distribution of SERT-immunoreactive fibers are shown in two representative coronal brain sections obtained from a control rat (A and B, respectively). The arrows indicate the superior and inferior PT borders of the IL area. Layers I to VI of the IL cortex are shown in (B). High- RI power photomicrographs taken from IL layers II/III of SERT-stained sections obtained from a control rat (C) and an epileptic rat (D). The arrows in (D) SC indicate larger-sized SERT-stained fiber varicosities, which are not seen in the A CC E PT ED M A N U control rat (C). Scale bar = 200 µm (A, B) and = 25 µm (C, D). A ED PT CC E PT RI SC U N A M 48 49 Figure 6. Graphic representation of the density of fiber varicosities immunostained for serotonin transporter (SERT) as a function of their crosssectional area in the infralimbic cortex (IL – A, B, C) and medial septum (MS – D, E, F). Statistical analysis of these data revealed a significant rightward shift PT in the density-size distribution of SERT varicosities in epileptic rats versus RI control rats (A, D; *p<0.05). Regression analysis showed that the density-size distribution of SERT varicosities could be best fitted with a bi-Gaussian SC function both in the control group (green line in B, E, also shown in A, D) and U in the post-SE group (orange line in C, F, also shown in A, D), suggesting the N existence of two distinct size populations of SERT varicosities. The relative A contribution of the smaller-sized varicosities (blue line) and the larger-sized M varicosities (red line) to the overall population of SERT-positive fiber ED varicosities are separately shown for the control group (B, E) and the epileptic A CC E panels. PT group (C, F). The coefficients of correlations (r2) are shown in respective ED M A N U SC RI PT 50 PT Figure 7. Representative photomicrographs of coronal sections from the medial septum of a control (A, C) and an epileptic (B, D) rats. The sections were CC E immunostained using a primary antibody for serotonin transporter (SERT). The boxed insets shown in (A) and (B) indicate regions of the medial septum, from A which the high-power images were taken (C, D). Note that the section derived from the epileptic rat show a higher density of larger-sized SERT-stained fibers and fiber varicosities (D) compared with the control (C). Scale bar = 200 µm (A, B) and 25 µm (C, D). A ED PT CC E PT RI SC U N A M 51