Raphe serotonin neurons are not homogenous: Electrophysiological, morphological and neurochemical evidence

Abstract

The median (MR) and dorsal raphe (DR) nuclei contain the majority of the 5-hydroxytryptamine (5-HT, serotonin) neurons that project to limbic forebrain regions, are important in regulating homeostatic functions and are implicated in the etiology and treatment of mood disorders and schizophrenia. The primary synaptic inputs within and to the raphe are glutamatergic and GABAergic. The DR is divided into three subfields, i.e., ventromedial (vmDR), lateral wings (lwDR) and dorsomedial (dmDR). Our previous work shows that cell characteristics of 5-HT neurons and the magnitude of the 5-HT_1A and 5-HT_1B receptor-mediated responses in the vmDR and MR are not the same. We extend these observations to examine the electrophysiological properties across all four raphe subfields in both 5-HT and non-5-HT neurons. The neurochemical topography of glutamatergic and GABAergic cell bodies and nerve terminals were identified using immunohistochemistry and the morphology of the 5-HT neurons was measured. Although 5-HT neurons possessed similar physiological properties, important differences existed between subfields. Non-5-HT neurons were indistinguishable from 5-HT neurons. GABA neurons were distributed throughout the raphe, usually in areas devoid of 5-HT neurons. Although GABAergic synaptic innervation was dense throughout the raphe (immunohistochemical analysis of the GABA transporters GAT1 and GAT3), their distributions differed. Glutamate neurons, as defined by vGlut3 anti-bodies, were intermixed and co-localized with 5-HT neurons within all raphe subfields. Finally, the dendritic arbor of the 5-HT neurons was distinct between subfields. Previous studies regard 5-HT neurons as a homogenous population. Our data support a model of the raphe as an area composed of functionally distinct subpopulations of 5-HT and non-5-HT neurons, in part delineated by subfield. Understanding the interaction of the cell properties of the neurons in concert with their morphology, local distribution of GABA and glutamate neurons and their synaptic input, reveals a more complicated and heterogeneous raphe. These results provide an important foundation for understanding how specific subfields modulate behavior and for defining which aspects of the circuitry are altered during the etiology of psychological disorders.

This article is part of a Special Issue entitled ‘Serotonin’.

Introduction

The median raphe (MR) and dorsal raphe (DR) nuclei contain the 5-hydroxytryptamine (5-HT, serotonin) cell bodies that provide the majority of 5-HT innervation of the forebrain. These 5-HT neurons have been implicated in mediating numerous homeostatic functions, i.e., stress responses, sleep–wake cycles, arousal, pain, learning and memory, and temperature regulation (Abrams et al., 2004, Buhot, 1997, Herman and Cullinan, 1997, Lopez et al., 1999, Lowry, 2002, Meneses, 1998, Wang and Nakai, 1994). Additionally, MR and DR are also implicated in the etiology and treatment of pathophysiological processes, in particular mood disorders and psychoses (Kroeze and Roth, 1998, Meltzer, 1999, Mourilhe and Stokes, 1998). Understanding how specific raphe circuits and neuron populations control particular behaviors requires a mechanistic description at the cellular and circuit level. Many investigators have proposed that the different subfields of DR and MR have differential roles in terms of mediating stress, anxiety and depression (Adell et al., 1997, Andrade and Graeff, 2001, Andrade et al., 1999, Andrews et al., 1997, Andrews et al., 1994, File and Gonzalez, 1996, Gonzalez and File, 1997, Gonzalez et al., 1998, Graeff et al., 1996, Lowry, 2002). Raphe circuits involve direct modulation of the HPA axis as well as indirectly mediated influences orchestrated by raphe projections to other limbic structures (e.g., the hippocampus, amygdala, and medial prefrontal cortex) and brainstem areas that regulate the autonomic nervous system.

MR and DR project to many of the same forebrain regions but also have distinct projections; additionally DR subfields project to different brain regions (Azmitia and Segal, 1978, Datiche et al., 1995, Imai et al., 1986a, Imai et al., 1986b, Johnson et al., 2008, Lowry, 2002, Vertes, 1991, Vertes et al., 1999, Vertes and Martin, 1988). For example, the rostral DR projects to the caudate–putamen and substantia nigra, the middle DR to the amygdala, whereas the caudal DR projects to the lateral and medial septum, ventral hippocampus, bed nucleus of the stria terminalis, locus coeruleus and hypothalamus. The MR projects to limbic regions such as the habenula, medial and lateral septum, medial prefrontal cortex, and dorsal and ventral hippocampus. Subfield projections (i.e, dmDR, vmDR and lwDR) of the middle and caudal DR project to the amygdala, medial prefrontal cortex and parts of the autonomic nervous system, all regions that are involved in uncontrollable stress and anxiety (Abrams et al., 2005, Abrams et al., 2004, Hale and Lowry, 2011, Johnson et al., 2008, Petrov et al., 1992, Sawchenko et al., 1983).

Non-5-HT-containing neurons are present and occur in equal or greater number to 5-HT neurons in DR and MR (Descarries et al., 1982, Kiss et al., 2002, Kohler and Steinbusch, 1982, Li et al., 2001, Van Bockstaele et al., 1993). For any given neurotransmitter, however, the number of neurons is lower, i.e., one third to one tenth less, than the number of 5-HT neurons. Most of non-5-HT neurons are differentially distributed within DR, e.g., GABA in lwDR and CRF cell bodies in dmDR (Allers and Sharp, 2003, Commons et al., 2003, Day et al., 2004). The evidence is less clear for non-5-HT neurons in MR. Many report that non-5-HT neurotransmitters are co-localized with 5-HT (Amilhon et al., 2010, Fremeau et al., 2002, Fu et al., 2010, Ma and Bleasdale, 2002). Thus both 5-HT and the non-5-HT neurotransmitters may be co-released within the raphe as well as in projection areas (Cardin et al., 2010, Varga et al., 2009).

Previous studies characterizing the active and passive membrane characteristics of 5-HT neurons in DR have been tentative since most did not use neurochemical identification. Of the three studies primarily cited as a basis for identifying putative 5-HT neurons (Aghajanian and Lakoski, 1984, Aghajanian and Vandermaelen, 1982, Vandermaelen and Aghajanian, 1983) only one used neurochemical identification. The hallmark characteristics include a firing rate of 1–5 Hz, an action potential with a long duration and large afterhyperpolarization (AHP) and a hyperpolarizing response to 5-HT_1A receptor activation. Additionally, several laboratories have proposed different subtypes of 5-HT neurons (Gartside et al., 2000, Hajos and Sharp, 1996, Kocsis et al., 2006). We have recently used whole-cell recording techniques in concert with immunohistochemistry to identify the cellular characteristics of both 5-HT and non-5-HT neurons in rat vmDR and MR (Beck et al., 2004). Findings in these two regions have emphasized the necessity to examine these characteristics in all raphe subfields. Results from these studies indicate that differences between 5-HT and non-5-HT neurons are not great enough to identify the neurons electrophysiologically; immunohistochemical identification is required since the electrophysiological properties overlap.

The present report completes our investigation of 5-HT and non-5-HT cell characteristics within the rat raphe subfields. Additionally we report the regional mapping of GABA and glutamate cell bodies and synaptic boutons within the raphe as well as differences in dendritic arborization of 5-HT cells across subfields. Understanding the unique characteristics of 5-HT and non-5-HT neurons within the different raphe subfields in concert with their anatomy and topography leads to a greater understanding of the mechanisms which govern raphe signaling, through which raphe neurons regulate homeostatic processes and that may be altered in pathological states.