CHAPTER 2

Neurotransmitters and Receptors in Psychiatric Disorders

Carolyn M. Drazinic, M.D., Ph.D.

Steven T. Szabo, M.D., Ph.D.

Todd D. Gould, M.D.

Husseini K. Manji, M.D., F.R.C.P.C.

This chapter serves as a primer on the recent advances in our understanding of neural function both in health and in disease. It is beyond the scope of this chapter to cover these important areas in extensive detail, and readers are referred to outstanding textbooks that are entirely devoted to the topic (Kandel 2013; Nestler et al. 2015; Squire 2013). Here, we focus on the principles of neurotransmission that are critical for an understanding of the biological bases of major psychiatric disorders, as well as the mechanisms by which effective treatments may exert their beneficial effects. In particular, our goal is to lay the groundwork for the subsequent chapters in this volume that focus on individual disorders and their treatments.

Although this chapter is intended to provide a general overview on neurotransmitter function, whenever possible, we emphasize the neuropsychiatric relevance of specific observations. We outline principles that are of fundamental importance to the study and practice of psychopharmacology. The figure legends contain additional details for the interested reader.

What Are Neurotransmitters?

Several criteria have been established for a neurotransmitter, including that 1) it is synthesized and released from neurons; 2) it is released from nerve terminals in a chemically or pharmacologically identifiable form; 3) it interacts with postsynaptic receptors and brings about the same effects as are seen with stimulation of the presynaptic neuron; 4) its interaction with the postsynaptic receptor displays a specific pharmacology; and 5) its actions are terminated by active processes (Kandel 2013; Nestler et al. 2015; Squire 2013). However, our growing appreciation of the complexity of the central nervous system (CNS) and of the existence of numerous molecules that exert neuromodulatory and neurohormonal effects has blurred the classical definition of neurotransmitters somewhat, and even well-known neurotransmitters do not meet all these criteria under certain situations (Cooper et al. 2003).

Most neuroactive compounds are small polar molecules that are synthesized in the CNS via local machinery or are able to permeate the blood–brain barrier. To date, more than 50 endogenous substances have been found to be present in the brain that appear to be capable of functioning as neurotransmitters. There are many plausible explanations for why the brain would need so many transmitters and receptor subtypes to transmit messages. Perhaps the simplest explanation is that the sheer complexity of the CNS results in many afferent nerve terminals impinging onto a single neuron. This requires a neuron to be able to distinguish the multiple information-conveying inputs. Although this can be accomplished partially by spatial segregation, it is accomplished in large part by chemical coding of the inputs—that is, different chemicals convey different information. Moreover, as we discuss in detail later, the evolution of multiple receptors for a single neurotransmitter means that the same chemical can convey different messages depending on the receptor subtypes it acts on. Additionally, the firing pattern of neurons is a means of conveying information; thus, the firing activities of neurons in the brain differ widely, and a single neuron firing at different frequencies can even release different neuroactive compounds depending on the firing rate (e.g., the release of peptides often occurs at higher firing rates than that which is required to release monoamines). These multiple mechanisms to enhance the diversity of responses—chemical coding, spatial coding, frequency coding—are undoubtedly critical in endowing the CNS with its complex repertoire of physiological and behavioral responses (Kandel 2013; Nestler et al. 2015). Finally, the existence of multiple neuroactive compounds also provides built-in safeguards to ensure that vital brain circuits are able to partially compensate for loss of function of particular neurotransmitters.

Receptors

An essential property of any living cell is its ability to recognize and respond to external stimuli. Cell surface receptors have two major functions: recognition of specific molecules (neurotransmitters, hormones, growth factors, and even sensory signals) and activation of “effectors.” Binding of the appropriate agonist (i.e., neurotransmitter or hormone) externally to the receptor alters the conformation (shape) of the protein. Cell surface receptors use a variety of membrane-transducing mechanisms to transform an agonist’s message into cellular responses. In neuronal systems, the most typical responses ultimately (in some cases rapidly, in others more slowly) involve changes in transmembrane voltage and hence neuronal changes in excitability. Collectively, the processes are referred to as transmembrane signaling or signal transduction mechanisms. This process is not restricted to neurons. For instance, astrocytes were once thought to be uninvolved in neurotransmission, but they have since been shown to possess volume-regulated Cl^– anion channels, which work together with gap junction/hemichannels to permit efflux of amino acids such as taurine, glutamate, and aspartate in response to swelling due to brain injury (Mulligan and MacVicar 2006; Ye et al. 2009).

Interestingly, although increasing numbers of potential neuroactive compounds and receptors continue to be identified, it has become clear that translation of the extracellular signals (into a form that can be interpreted by the complex intracellular enzymatic machinery) is achieved through a relatively small number of cellular mechanisms. Generally speaking, these transmembrane signaling systems, and the receptors that use them, can be divided into four major groups (Figure 2–1):

1. Those that are relatively self-contained in structure and whose message takes the form of transmembrane ion fluxes (ionotropic receptors)
2. Those that are multicomponent in nature and generate intracellular second messengers (G protein–coupled receptors)
3. Those that contain intrinsic enzymatic activity (receptor tyrosine kinases and phosphatases)
4. Those that are cytoplasmic and translocate to the nucleus to directly regulate transcription (gene expression) after they are activated by lipophilic molecules (often hormones) that enter the cell (nuclear receptors)

FIGURE 2–1. Major receptor subtypes in the central nervous system.

See Plate 4 to view this figure in color.

This figure depicts the four major classes of receptors in the CNS. (A) Ionotropic receptors. These receptors comprise multiple protein subunits that are combined in such a way as to create a central membrane pore through this complex, allowing the flow of ions. This type of receptor has a very rapid response time (milliseconds). The consequences of receptor stimulation (i.e., excitatory or inhibitory) depend on the types of ions that the receptor specifically allows to enter the cell. Thus, for example, Na⁺ entry through the NMDA (N-methyl-D-aspartate) receptor depolarizes the neuron and brings about an excitatory response, whereas Cl^– efflux through the γ-aminobutyric acid type A (GABA_A) receptor hyperpolarizes the neuron and brings about an inhibitory response. Illustrated here is the NMDA receptor regulating a channel permeable to Ca²⁺, Na⁺, and K⁺ ions. The NMDA receptors also have binding sites for glycine, Zn²⁺, phencyclidine, MK801/ketamine, and Mg²⁺; these molecules are able to regulate the function of this receptor. (B) G protein–coupled receptors (GPCRs). The majority of neurotransmitters, hormones, and even sensory signals mediate their effects via seven transmembrane domain–spanning receptors that are G protein–coupled. The amino terminus of the G protein is on the outside of the cell and plays an important role in the recognition of specific ligands; the third intracellular loop and carboxy terminus of the receptor play an important role in coupling to G proteins and are sites of regulation of receptor function (e.g., by phosphorylation). All G proteins are heterotrimers (consisting of α, β, and γ subunits). The G proteins are attached to the membrane by isoprenoid moieties (fatty acid) via their γ subunits. Compared with the ionotropic receptors, GPCRs mediate a slower response (on the order of seconds). Detailed depiction of the activation of G protein–coupled receptors is given in Figure 2–2. Here we depict a receptor coupled to the G protein G_s (the s stands for stimulatory to the enzyme adenylyl cyclase [AC]). Activation of such a receptor produces activation of AC and increases in cyclic adenosine monophosphate (cAMP) levels. G protein–coupled pathways exhibit major amplification properties, and, for example, in model systems researchers have demonstrated a 10,000-fold amplification of the original signal. The effects of cAMP are mediated largely by activation of protein kinase A (PKA). One major downstream target of PKA is CREB (cAMP response element–binding protein), which may be important to the mechanism of action of antidepressants. (C) Receptor tyrosine kinases. These receptors are activated by neurotrophic factors and are able to bring about acute changes in synaptic function, as well as long-term effects on neuronal growth and survival. These receptors contain intrinsic tyrosine kinase activity. Binding of the ligand triggers receptor dimerization and transphosphorylation of tyrosine residues in its cytoplasmic domain, which then recruits cytoplasmic signaling and scaffolding proteins. The recruitment of effector molecules generally occurs via interaction of proteins with modular binding domains SH2 and SH3 (named after homology to the src oncogenes–src homology domains); SH2 domains are a stretch of about 100 amino acids that allow high-affinity interactions with certain phosphotyrosine motifs. The ability of multiple effectors to interact with phosphotyrosines is undoubtedly one of the keys to the pleiotropic effects that neurotrophins can exert. Shown here is a tyrosine kinase receptor type B (TrkB), which upon activation produces effects on the Raf, MEK (mitogen-activated protein kinase/ERK), extracellular response kinase (ERK), and ribosomal S6 kinase (RSK) signaling pathway. Some major downstream effects of RSK are CREB and stimulation of factors that bind to the AP-1 site (c-Fos and c-Jun). (D) Nuclear receptors. These receptors are transcription factors that regulate the expression of target genes in response to steroid hormones and other ligands. Many hormones (including glucocorticoids, gonadal steroids, and thyroid hormones) are able to rapidly penetrate into the lipid bilayer membrane, because of their lipophilic composition, and thereby directly interact with these cytoplasmic receptors inside the cell. Upon activation by a hormone, the nuclear receptor–ligand complex translocates to the nucleus, where it binds to specific DNA sequences, referred to as hormone responsive elements (HREs), and regulates gene transcription. Nuclear receptors often interact with a variety of coregulators that promote transcriptional activation when recruited (coactivators) and those that attenuate promoter activity (corepressors). However, nongenomic effects of neuroactive steroids have also been documented, with the majority of evidence suggesting modulation of ionotropic receptors. This figure illustrates both the genomic and the nongenomic effects. ATF1=activation transcription factor 1; BDNF=brain-derived neurotrophic factor; CaMKII=Ca²⁺/calmodulin–dependent protein kinase II; CREM=cyclic adenosine 5′-monophosphate response element modulator; D₁=dopamine₁ receptor; D₅=dopamine₅ receptor; ER=estrogen receptor; GR=glucocorticoid receptor; GRK=G protein–coupled receptor kinase; P=phosphorylation; PR=progesterone receptor.

A more extensive and continuously updated synopsis of these and many other receptors and ligands can be found at the International Union of Basic and Clinical Pharmacology/British Pharmacological Society Guide to Pharmacology Web site (www.guidetopharmacology.org) and associated publications (Alexander et al. 2013a, 2013b, 2013c, 2013d, 2013e, 2013f). We review the four major groups in the following subsections.

Ionotropic Receptors

The first class of receptors contains in their molecular complex an intrinsic ion channel. Receptors of this class include those for several amino acids, including glutamate (e.g., the NMDA [N-methyl-D-aspartate] receptor), GABA (γ-aminobutyric acid via the GABA_A receptor), and the nicotinic acetylcholine (nACh) receptor and the serotonin type 3 (5-HT₃) receptor. Ion channels are integral membrane proteins that are directly responsible for the electrical activity of the nervous system by virtue of their regulation of the movement of ions across membranes. Receptors containing intrinsic ion channels have been called ionotropic and are generally composed of four or five subunits that open transiently when a neurotransmitter binds, allowing ions to flow into (e.g., Na⁺, Ca²⁺, Cl^–) or out of (e.g., K⁺) the neuron, thereby generating synaptic potential (see Figure 2–1).

Often, the ionotropic receptors are composed of different combinations of various subunits, thereby providing the system with considerable flexibility. For example, there is extensive research into the potential development of an anxiolytic that is devoid of sedative effects by targeting GABA_A receptor subunits present in selected brain regions (Salerno et al. 2012; Taliani et al. 2009). In general, neurotransmission that is mediated by ionotropic receptors is very fast, with ion channels opening and closing within milliseconds, and these receptors regulate much of the tonic excitatory (e.g., glutamate-mediated) and inhibitory (e.g., GABA-mediated) activity in the CNS; as we discuss in the “Neurotransmitter and Neuropeptide Systems” section later in this chapter, many of the classic neurotransmitters (e.g., monoamines) exert their effects on a slower time scale and are therefore often considered to be modulatory in their effects.

G Protein–Coupled Receptors

Most receptors in the CNS do not have intrinsic ionic conductance channels within their structure but instead regulate cellular activity by the generation of various “second messengers.” Receptors of this class generally do not interact directly with the various second-messenger-generating enzymes but transmit information to the appropriate “effector” by the activation of interposed coupling proteins. These are the G protein–coupled receptor families, and they provide a slower onset, but longer duration of signaling, compared with ionotropic receptors (Squire 2013). The G protein–coupled receptors (GPCRs, which constitute more than 80% of all known receptors in the body and number about 800 in humans) all span the plasma membrane seven times and have three intracellular loops and three extracellular loops (see Figure 2–1) (Alexander et al. 2013b). G proteins are so named because of their ability to bind the guanine nucleotides guanosine triphosphate and guanosine diphosphate. Receptors coupled to G proteins include those for dopamine, serotonin, acetylcholine, various peptides, and even sensory signals such as light and odorants.

GPCRs have increasingly become the focus of extensive research in psychiatry (Catapano and Manji 2007). The amino terminus is on the outside of the cell and plays a critical role in recognition of the ligand, which can be a small molecule, peptide, or large protein; the carboxy terminus and third intracellular loop are inside the cell and regulate coupling to different G proteins, “cross-talk” between receptors, and desensitization (see Figure 2–1) (Alexander et al. 2013b). Although the bimodal model of ligands switching the GPCR “on” or “off” is appealing, an individual GPCR can actually assume many different conformations, which influences the nature of the ligand–receptor interaction and the predominant complex signal generated in a particular cell type; this concept is called ligand-induced selective signaling (Millar and Newton 2010). Differential oligomerization, differential phosphorylation, signaling through molecules other than G proteins, and second-messenger independent signaling together add even more complexity for future GPCR research (Kandel 2013; Millar and Newton 2010).

In conclusion, many classes and subtypes of G proteins exist, playing key roles in amplifying and integrating signals.

Autoreceptors and Heteroreceptors

Autoreceptors are receptors located on neurons that produce the endogenous ligand for that particular receptor (e.g., a serotonergic receptor on a serotonergic neuron). By contrast, heteroreceptors are receptor subtypes that are present on neurons that do not contain an endogenous ligand for that particular receptor subtype (e.g., a serotonergic receptor located on a dopaminergic neuron).

Two major classes of autoreceptors play very important roles in fine-tuning neuronal activity. Somatodendritic autoreceptors are present on cell bodies and dendrites and exert critical roles in regulating the firing rate of neurons. In general, activation of somatodendritic autoreceptors (e.g., α₂-adrenergic receptors for noradrenergic neurons, 5-HT_1A receptors for serotonergic neurons, or dopamine type 2 [D₂] receptors for dopaminergic neurons) inhibits the firing rate of the neurons by opening K⁺ channels and by reducing cyclic adenosine monophosphate (cAMP) levels, both of which may be important in psychiatric disease. For instance, TREK-1 is a background K⁺ channel regulator protein important in serotonin transmission and potentially in moodlike behavior regulation in mice (Heurteaux et al. 2006). This exemplifies how fundamental mechanisms of neuronal transmission such as K⁺ channels, which regulate membrane potentials, may relate to global alterations in brain functioning relevant to psychiatry.

The second major class of autoreceptors, nerve terminal autoreceptors, play an important role in regulating the amount of neurotransmitter released per nerve impulse, generally by closing nerve terminal Ca²⁺ channels. Both of these types of autoreceptors are typically members of the GPCR family. Neurotransmitter release is known to be triggered by influx and alterations of intracellular calcium, with functioning of three types of SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein [SNAP] receptor) proteins mediating a critical role. The distinct kinetics of neurotransmitter release modulators, such as botulinum and tetanus neurotoxins, induce prominent alterations in synaptobrevin and syntaxin, leading to calcium-independent mechanisms of neurotransmitter regulation (Sakaba et al. 2005). Most synapses are dependent on influx of Ca²⁺ through voltage-gated calcium channels for presynaptic neurotransmitter release; however, in the retina, this influx of calcium occurs via glutamatergic α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors (Chávez et al. 2006). Beyond the receptor level, presynaptic SAD (synapses of amphids defective, originally identified as a regulator of neuronal polarity in the Caenorhabditis elegans worm model), an intracellular serine threonine kinase, is associated with the active zone cytomatrix that regulates neurotransmitter release, and SAD kinases in presynaptic neurons also control the maturation and arborization of synapses in the central and peripheral nervous systems (Inoue et al. 2006; Lilley et al. 2014). These data further exemplify the dynamic nature of basic processes involved in neurotransmitter regulation that may possibly aid in advancing treatment of psychopathology.

GPC Receptor Regulation and Trafficking

The mechanism by which GPCRs translate extracellular signals into cellular changes was once envisioned as a simple linear model. It is now known, however, that the activity of GPCRs is subject to at least three additional principal modes of regulation: desensitization, downregulation, and trafficking (Carman and Benovic 1998) (Figure 2–2). Desensitization, the process by which cells rapidly adapt to stimulation by agonists, is generally believed to occur by two major mechanisms: homologous and heterologous.

FIGURE 2–2. G protein–coupled receptors and G protein activation.

See Plate 5 to view this figure in color.

All G proteins are heterotrimers consisting of α, β, and γ subunits. The receptor shuttles between a low-affinity form that is not coupled to a G protein and a high-affinity form that is coupled to a G protein. (A) At rest, G proteins are largely in their inactive state, namely, as αβγ heterotrimers, which have GDP (guanosine diphosphate) bound to the α subunit. (B) When a receptor is activated by a neurotransmitter, it undergoes a conformational (shape) change, forming a transient state referred to as a high-affinity ternary complex, comprising the agonist, the receptor in a high-affinity state, and the G protein. A consequence of the receptor interaction with the G protein is that the GDP comes off the G protein α subunit, leaving a very transient empty guanine nucleotide binding domain. (C) Guanine nucleotides (generally GTP) quickly bind to this nucleotide binding domain; thus, one of the major consequences of active receptor–G protein interaction is to facilitate guanine nucleotide exchange—this is basically the “on switch” for the G protein cycle. (D) A family of GTPase-activating proteins for G protein–coupled receptors has been identified, and they are called regulators of G protein signaling (RGS) proteins. Since activating GTPase activity facilitates the “turn off” reaction, these RGS proteins are involved in dampening the signal. Binding of GTP to the α subunit of G proteins results in subunit dissociation, whereby the α-GTP dissociates from the βγ subunits. Although not covalently bound, the β and γ subunits remain tightly associated and generally function as dimers. The α-GTP and βγ subunits are now able to activate multiple diverse effectors, thereby propagating the signal. While they are in their active states, the G protein subunits can activate multiple effector molecules in a “hit and run” manner; this results in major signal amplification (i.e., one active G protein subunit can activate multiple effector molecules). The activated G protein subunits also dissociate from the receptor, converting the receptor to a low-affinity conformation and facilitating the dissociation of the agonist from the receptor. The agonist can now activate another receptor, and this also results in signal amplification. Together, these processes have been estimated to produce a 10,000-fold amplification of the signal in certain models. (E) Interestingly, the α subunit has intrinsic GTPase activity, which cleaves the third phosphate group from GTP (G-P-P-P) to GDP (G-P-P). Since α-GDP is an inactive state, the GTPase activity serves as a built-in timing mechanism, and this is the “turn off” reaction. (F) The reassociation of α-GDP with βγ is thermodynamically favored, and the reformation of the inactive heterotrimer (αβγ) completes the G protein cycle.

Homologous desensitization is receptor specific; that is, only the receptor actively being stimulated becomes desensitized. This form of desensitization occurs via a family of kinases known as G protein–coupled kinases. When a receptor activates a G protein and causes dissociation of the α subunit from the βγ subunits, the βγ subunits are able to provide an “anchoring surface” for the G protein–coupled kinases to allow them to come into the proximity of the activated receptor and phosphorylate it. This phosphorylation then recruits another family of proteins known as arrestins, which physically interfere with the coupling of the phosphorylated receptor and the G protein, thereby dampening the signal. This form of desensitization is very rapid and usually transient (i.e., the receptors get dephosphorylated and return to the baseline state). However, if the stimulation of the receptor is excessive and prolonged, it leads to an internalization of the receptor, and often its degradation, a process referred to as downregulation.

Heterologous desensitization is not receptor specific and is mediated by second-messenger kinases such as protein kinase A (PKA) and protein kinase C (PKC). Thus, when a receptor activates PKA, the activated PKA is capable of phosphorylating (and thereby desensitizing) not only that particular receptor but also other receptors that are present in proximity and have the correct phosphorylation motif, thereby producing heterologous desensitization.

After prolonged or repeated activation of receptors by agonist ligands, the process of receptor downregulation is observed. Downregulation is associated with a reduced number of receptors detected in cells or tissues, thereby leading to attenuation of cellular responses (Carman and Benovic 1998). The process of GPCR sequestration is mediated by a well-characterized endocytic pathway involving the concentration of receptors in clathrin-coated pits and subsequent internalization and recycling back to the plasma membrane (Tsao and von Zastrow 2000). Endocytosis can thus clearly serve as a primary mechanism to attenuate signaling by rapidly and reversibly removing receptors from the cell surface. However, endocytosis and receptor trafficking also mediate GPCR signaling by way of certain effector pathways, most notably mitogen-activated protein (MAP) kinase cascades. Evidence also indicates that endocytosis of GPCRs may be required for certain signal transduction pathways leading to the nucleus (Tsao and von Zastrow 2000). These diverse functions of GPCR endocytosis and trafficking lead to unexpected insights into the biochemical and functional properties of endocytic vesicles. Indeed, there is considerable excitement about our growing understanding of the diverse molecular mechanisms for signaling specificity and receptor trafficking and the possibility that this knowledge could lead to new selective therapeutics.

Receptor Tyrosine Kinases

The receptor tyrosine kinases, as their name implies, contain intrinsic tyrosine kinase activity and are generally used by growth factors, such as neurotrophic factors, and cytokines. Binding of an agonist initiates receptor dimerization and transphosphorylation of tyrosine residues in its cytoplasmic domain (Patapoutian and Reichardt 2001) (see Figure 2–1). The phosphotyrosine residues of the receptor function as binding sites for recruiting specific cytoplasmic signaling and scaffolding proteins. The ability of multiple effectors to interact with phosphotyrosines is undoubtedly one of the keys to the pleiotropic effects that neurotrophins can exert. These pleiotropic and yet distinct effects of growth factors are mediated by varying degrees of activation of three major signaling pathways: the MAP kinase pathway, the phosphoinositide-3 (PI₃) kinase pathway, and the phospholipase C (PLC)–γ1 pathway. These pathways are part of complex secondary messaging systems of the cell, which are beyond the scope of this chapter.

Nuclear Receptors

Nuclear receptors are transcription factors that regulate the expression of target genes in response to steroid hormones and other ligands. Many hormones (including glucocorticoids, gonadal steroids, and thyroid hormones) are able to rapidly penetrate into the lipid bilayer membrane because of their lipophilic composition, and thereby directly interact with these cytoplasmic receptors inside the cell (see Figure 2–1). On activation by a hormone, the nuclear receptor–ligand complex translocates to the nucleus, where it binds to specific DNA sequences referred to as hormone-responsive elements, and subsequently regulates gene transcription (Mangelsdorf et al. 1995; Truss and Beato 1993). Nuclear receptors often interact with a variety of coregulators that promote transcriptional activation when recruited (coactivators) and those that attenuate promoter activity (corepressors). Numerous nuclear receptors have been identified, as reviewed elsewhere (Alexander et al. 2013e).

With this broad overview of neurotransmitters and receptor subtypes, we now turn to a discussion of selected individual neurotransmitters and their receptors.

Neurotransmitter and Neuropeptide Systems

Serotonergic System

Largely on the basis of the observation that most effective antidepressants and antipsychotics target these systems, the monoaminergic systems (e.g., serotonin, norepinephrine, dopamine) have been extensively studied. Serotonin was given that name because of its activity as an endogenous vasoconstrictor in blood serum (Rapport et al. 1948). It was later acknowledged as being the same molecule (secretine) that is found in the intestinal mucosa and that is “secreted” by chromaffin cells (Brodie 1900). Following these findings, serotonin later was characterized as a neurotransmitter in the CNS (Bogdanski et al. 1956).

Serotonin-producing cell bodies in the brain are localized to the central gray, in the surrounding reticular formation, and in cell clusters located in the center, and thus the name raphe (from Latin, meaning “midline”) was adopted (Figure 2–3A). The dorsal raphe, the largest brain stem serotonin nucleus, contains approximately 50% of the total serotonin neurons in the mammalian CNS; in contrast, the medial raphe comprises 5% (Descarries et al. 1982; Wiklund and Björklund 1980). Serotonergic neurons project widely throughout the CNS rather than to discrete anatomical locations (as the dopaminergic neurons appear to do; see Figure 2–4A later in this chapter), leading to the suggestion that serotonin exerts a major modulatory role throughout the CNS (Reader 1980). Interestingly, evidence suggests that infralimbic and prelimbic regions of the ventromedial prefrontal cortex (vmPFC) in rats are responsible for detecting whether a stressor is under the organism’s control. When a stressor is controllable, stress-induced activation of the dorsal raphe nucleus is inhibited by the vmPFC, and the behavioral sequelae of the uncontrollable stress response are blocked (Amat et al. 2005). The ability to regulate serotonin neuron activity and function has been a major ongoing focus of psychiatric disorder research and treatments. Lysergic acid diethylamide, a hallucinogen, has a chemical structure similar to that of serotonin, and monoamine oxidase (MAO) inhibitors (MAOIs), which are classic antidepressants, increase the levels of monoamines such as serotonin in the synapse (Squire 2013).

FIGURE 2–3. The serotonergic system.

See Plate 6 to view this figure in color.

This figure depicts the location of the major serotonin (5-HT)–producing cells (raphe nuclei) innervating brain structures (A), and various cellular regulatory processes involved in serotonergic neurotransmission (B). 5-HT neurons project widely throughout the CNS and innervate virtually every part of the neuroaxis. L-Tryptophan, an amino acid actively transported into presynaptic 5-HT-containing terminals, is the precursor for 5-HT. It is converted to 5-hydroxytryptophan (5-HTP) by the rate-limiting enzyme tryptophan hydroxylase (TrpH). This enzyme is effectively inhibited by the drug p-chlorophenylalanine (PCPA). Aromatic amino acid decarboxylase (AADC) converts 5-HTP to 5-HT. Once released from the presynaptic terminal, 5-HT can interact with a variety (15 different types) of presynaptic and postsynaptic receptors. Presynaptic regulation of 5-HT neuron firing activity and release occurs through somatodendritic 5-HT_1A (not shown) and 5-HT_1B,1D autoreceptors, respectively, located on nerve terminals. Sumatriptan is a 5-HT_1B,1D receptor agonist. (The antimigraine effects of this agent are likely mediated by local activation of this receptor subtype on blood vessels, which results in their constriction.) Buspirone is a partial 5-HT_1A receptor agonist that activates both pre- and postsynaptic receptors. Cisapride is a preferential 5-HT₄ receptor agonist that is used to treat irritable bowel syndrome as well as nausea associated with antidepressants. The binding of 5-HT to G protein receptors (G_o, G_i, etc.) that are coupled to adenylyl cyclase (AC) and phospholipase C–β (PLC-β) will result in the production of a cascade of second-messenger and cellular effects. Lysergic acid diethylamide (LSD) likely interacts with numerous 5-HT receptors to mediate its effects. Pharmacologically this ligand is often used as a 5-HT₂ receptor agonist in receptor-binding experiments. Ondansetron is a 5-HT₃ receptor antagonist that is marketed as an antiemetic agent for chemotherapy patients but is also given to counteract side effects produced by antidepressants in some patients. 5-HT has its action terminated in the synapse by rapidly being taken back into the presynaptic neuron through 5-HT transporters (5-HTTs). Once inside the neuron, it can either be repackaged into vesicles for reuse or undergo enzymatic catabolism. The selective 5-HT reuptake inhibitors (SSRIs) and older-generation tricyclic antidepressants (TCAs) are able to interfere/block the reuptake of 5-HT. 5-HT is then metabolized to 5-hydroxyindoleacetic acid (5-HIAA) by monoamine oxidase (MAO), located on the outer membrane of mitochondria or sequestered and stored in secretory vesicles by vesicular monoamine transporters (VMATs). Reserpine causes a depletion of 5-HT in vesicles by interfering with uptake and storage mechanisms (depressive-like symptoms have been reported with this agent). Tranylcypromine is an MAO inhibitor (MAOI) and an effective antidepressant. Fenfluramine (an anorectic agent) and 3,4-methylenedioxymethamphetamine (MDMA; “Ecstasy”) are able to facilitate 5-HT release by altering 5-HTT function. cAMP=cyclic adenosine monophosphate; DAG=diacylglycerol; IP₃=inositol-1,4,5-triphosphate.

Source. Adapted from Cooper JR, Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology, 8th Edition. New York, Oxford University Press, 2001. Copyright 1970, 1974, 1978, 1982, 1986, 1991, 1996, 2001 by Oxford University Press, Inc. Used by permission of Oxford University Press, Inc. Modified from Nestler et al. 2001.

FIGURE 2–4. The dopaminergic system.

See Plate 7 to view this figure in color.

This figure depicts the dopaminergic projections throughout the brain (A) and various regulatory processes involved in dopaminergic neurotransmission (B). The amino acid L-tyrosine is actively transported into presynaptic dopamine (DA) nerve terminals, where it is ultimately converted into DA. The rate-limiting step is conversion of L-tyrosine to L-dihydroxyphenylalanine (L-dopa) by the enzyme tyrosine hydroxylase (TH). α-Methyl-p-tyrosine (AMPT) is a competitive inhibitor of tyrosine hydroxylase and has been used to assess the impact of reduced catecholaminergic function in clinical studies. The production of DA requires that L-aromatic amino acid decarboxylase (AADC) act on L-dopa. Thus, the administration of L-dopa to patients with Parkinson’s disease bypasses the rate-limiting step and is able to produce DA quite readily. DA has its action terminated in the synapse by rapidly being taken back into the presynaptic neuron through DA transporters (DATs). DA is then metabolized to dihydroxyphenylalanine (DOPAC) by intraneuronal monoamine oxidase (MAO; preferentially by the MAO-B subtype) located on the outer membrane of mitochondria, or is sequestered and stored in secretory vesicles by vesicular monoamine transporters (VMATs). Reserpine causes a depletion of DA in vesicles by interfering and irreversibly damaging uptake and storage mechanisms. γ-Hydroxybutyrate inhibits the release of DA by blocking impulse propagation in DA neurons. Pargyline inhibits MAO and may have efficacy in treating parkinsonian symptoms by augmenting DA levels through inhibition of DA catabolism. Other clinically used inhibitors of MAO are nonselective and thus likely elevate the levels of DA, norepinephrine, and serotonin. Once released from the presynaptic terminal (because of an action potential and calcium influx), DA can interact with five different G protein–coupled receptors (D₁–D₅), which belong to either the D₁ or D₂ receptor family. Presynaptic regulation of DA neuron firing activity and release occurs through somatodendritic (not shown) and nerve terminal D₂ autoreceptors, respectively. Pramipexole is a D₂/D₃ receptor agonist and has been documented to have efficacy as an augmentation strategy in cases of treatment-resistant depression and in the management of Parkinson’s disease. The binding of DA to G protein receptors (G_o, G_i, etc.) positively or negatively coupled to adenylyl cyclase (AC) results in the activation or inhibition of this enzyme, respectively, and the production of a cascade of second-messenger and cellular effects (see diagram). Apomorphine is a D₁/D₂ receptor agonist that has been used clinically to aid in the treatment of Parkinson’s disease. (SKF-82958 is a pharmacologically selective D₁ receptor agonist.) SCH-23390 is a D₁/D₅ receptor antagonist. There are likely physiological differences between D₁ and D₅ receptors, but the current unavailability of selective pharmacological agents has precluded an adequate differentiation thus far. Haloperidol is a D₂ receptor antagonist, and clozapine is a nonspecific D₂/D₄ receptor antagonist (both are effective antipsychotic agents). Once inside the neuron, DA can either be repackaged into vesicles for reuse or undergo enzymatic catabolism. Nomifensine is able to interfere/block the reuptake of DA. The antidepressant bupropion has affinity for the dopaminergic system, but it is not known whether this agent mediates its effects through DA or possibly by augmenting other monoamines. DA can be degraded to homovanillic acid (HVA) through the sequential action of catechol-O-methyltransferase (COMT) and MAO. Tropolone is an inhibitor of COMT. Evidence suggests that the COMT gene may be linked to schizophrenia (Akil et al. 2003). cAMP=cyclic adenosine monophosphate; DOPA=dihydroxyphenylalanine; MT=methoxytyramine;

Source. Adapted from Cooper JR, Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology, 8th Edition. New York, Oxford University Press, 2001. Copyright 1970, 1974, 1978, 1982, 1986, 1991, 1996, 2001 by Oxford University Press, Inc. Used by permission of Oxford University Press, Inc.

The precursor for serotonin synthesis is L-tryptophan, an amino acid that comes primarily from the diet and crosses the blood–brain barrier through a carrier for large neutral amino acids. Tryptophan hydroxylase is the rate-limiting enzyme in serotonin biosynthesis (Figure 2–3B), and polymorphisms in this enzyme have been extensively investigated in psychiatric disorders, with equivocal results. A more fruitful research strategy in humans has been tryptophan depletion via dietary restriction to study the role of serotonin in the pathophysiology and treatment of psychiatric disorders (Bell et al. 2001; van Donkelaar et al. 2011). These studies have indicated that acute tryptophan depletion (ATD) produces a rapid depressive relapse in patients taking selective serotonin reuptake inhibitors (SSRIs) but not in those taking norepinephrine reuptake inhibitors; the data suggesting induction of depressive symptoms in remitted patients or individuals with family histories of mood disorders are more equivocal (Bell et al. 2001). Findings similar to those from ATD depression studies also have been identified in numerous studies of panic disorder, whereby healthy subjects with no history of panic disorder, or patients with panic disorder in remission, were not very sensitive to ATD, but patients with unremitted panic disorder were extremely sensitive to carbon dioxide challenges (Sobczak and Schruers 2014). Some of the inconsistencies of the ATD findings in the literature may have been due to the various formulations of the tryptophan depletion amino acid cocktails used, but more surprising are suggestions that the classic ATD cocktails may not have been depleting extracellular synaptic serotonin levels after all but may have been working through other molecular mechanisms as described elsewhere (Sobczak and Schruers 2014; van Donkelaar et al. 2011). Nevertheless, numerous studies in the literature suggest that, in general, low levels of serotonin in the plasma and/or cerebrospinal fluid (CSF) correlate with depression, impulsivity, aggression, alcohol dependence, suicide attempts, completed suicides, and violent suicides (Bortolato et al. 2013; Moberg et al. 2011; Oo et al. 2016; Purselle and Nemeroff 2003; Ruljancic et al. 2011, 2013; Spreux-Varoquaux et al. 2001). Serotonin dysfunction–related endophenotypes that could predict patients at risk for suicide are being explored, including fenfluramine challenges leading to poor prolactin release, increased time spent in rapid eye movement (REM) sleep stages, increased loudness-dependent auditory evoked potentials, and increased frontal cortex P300 (Lee and Kim 2011). Furthermore, postmortem studies in depressed bipolar patients who completed suicide found lower serotonin and norepinephrine activity (Wiste et al. 2008).

More sophisticated molecular approaches targeting individual serotonin receptors and transporters are described in the following subsections.

Serotonin Transporters

As is the case for many classic neurotransmitters, termination of the effects of serotonin in the synaptic cleft is brought about in large part by an active reuptake process mediated by the serotonin transporter (5-HTT). Serotonin is taken up into the presynaptic terminals, where it is metabolized by the enzyme MAO or sequestered into secretory vesicles by the vesicle monoamine transporter (see Figure 2–3B). This presumably underlies the mechanism by which MAOIs initiate their therapeutic effects; that is, the blockade of monoamine breakdown results in an increase in the available pool for release when an action potential invades the nerve terminal. It is now well established that many tricyclic antidepressants and SSRIs exert their initial primary pharmacological effects by binding to the 5-HTT and blocking serotonin reuptake, thereby increasing the intrasynaptic levels of serotonin, which initiates a cascade of downstream effects (see Figure 2–3B for details). It has been hypothesized that the first step in serotonin transport involves the binding of serotonin to the 5-HTT and then a cotransport with Na⁺, and the second step involves the translocation of K⁺ across the membrane to the outside of the cell. SSRIs bind to the same site on the transporter as serotonin itself. Elegant biochemical and mutagenesis experiments have elucidated a leucine transporter from bacterial species, providing information that helped unravel the mechanism by which mammalian transporters couple ions and substrates to mediate neurotransmitter clearance. The crystal structure for sodium- and chloride-dependent neurotransmitter transporters (including transporters for serotonin [SERT], dopamine [DAT], norepinephrine [NET], glycine [GlyT1b], and GABA [GAT1]) with the L-leucine binding sites for Na⁺ ions also has been elucidated (Henry et al. 2006; Yamashita et al. 2005).

In the brain, 5-HTTs have been radiolabeled with [³H]-imipramine (Hrdina et al. 1985; Langer et al. 1980) and with SSRIs such as [³H]cyanoimipramine (Wolf and Bobik 1988), [³H]paroxetine (Habert et al. 1985), and [³H]citalopram (D’Amato et al. 1987). The regional distribution of 5-HTT corresponds to discrete regions of rat brain known to contain cell bodies of serotonin neurons and synaptic axon terminals, most notably the cerebral cortex, neostriatum, thalamus, and limbic areas (Cooper et al. 2003; Hrdina et al. 1985; Madden 2002). The specific cellular localization of 5-HTT in the CNS also has been accomplished by using site-specific antibodies (Lawrence et al. 1995a). Immunohistochemical studies that used antibodies against the serotonin carrier have reported both neuronal and glial staining in areas of the rat brain containing serotonin somata and terminals (i.e., dorsal raphe and hippocampus) (Lawrence et al. 1995b). Experimental alterations of 5-HTT in young mice for a brief period during early development indicate abnormal emotional behavior in the same mice later in life, similar to the phenotype in mice in which 5-HTT is deficient throughout life (Ansorge et al. 2004). This suggests the necessity of serotonin early in emotional development and provides a possible mechanism by which genetic changes in the 5-HTT system may lead to susceptibility to develop psychiatric diseases such as depression (Caspi et al. 2003). Furthermore, serotonin uptake ability has been documented in primary astrocyte cultures (Kimelberg and Katz 1985) and has been postulated to account for considerable serotonin uptake in the frontal cortex and periventricular region (Ravid et al. 1992).

Because 5-HTT is transcribed from a single copy gene, abnormalities in platelet 5-HTT are thought to reflect CNS abnormalities (Owens and Nemeroff 1998). Several studies on platelet 5-HTT density have been undertaken using [³H]imipramine binding or [³H]paroxetine binding in mood disorders. Although the results of these studies are not entirely consistent, in total the results suggest that the receptor density value for platelet serotonin density is significantly lower in depressed subjects compared with healthy control subjects (Owens and Nemeroff 1998). The distribution of SERT (5-HTT) in the postmortem human brain was found to be highest in the thalamus, amygdala, putamen, globus pallidus externa, lateral geniculate body, hippocampus, and caudate, with the lowest amounts in the cerebral cortex and minimal levels in the cerebellum and white matter (Kish et al. 2005).

Numerous studies have examined suicide risk related to individual genetic polymorphisms in SERT. For example, the serotonin-transporter-linked polymorphic region (5-HTTLPR) short S allele, which decreases presynaptic 5-HTT expression and thereby decreases serotonin reuptake and has an interaction with childhood abuse to increase lifetime depression risk, has nevertheless yielded contradictory results related to suicide risk, likely because of other confounding polymorphisms (Purselle and Nemeroff 2003; Ressler et al. 2010). In patients with major depressive disorder (MDD) who attempted suicide, positron emission tomographic (PET) scanning with the [¹¹C]N,N-dimethyl-2-(2-amino-4-cyanophenylthio) benzylamine ([¹¹C]DASB) ligand indicated that SERT (5-HTT) levels were lower in the midbrain regions (Miller et al. 2013). In bipolar I patients with depression who completed suicide, a nearly 50% decrease in serotonin and norepinephrine activity was observed related to the locus coeruleus in postmortem brain studies, as compared with unipolar depressed patients who completed suicide and matched control subjects (Wiste et al. 2008). Three different antidepressants, sertraline (SSRI), citalopram (SSRI), and venlafaxine (serotonin–norepinephrine reuptake inhibitor), at clinical dosages in living patients were shown with PET ligand [¹¹C]DASB to have 85% occupancy of SERT, effectively blocking SERT-mediated reuptake of free serotonin in the synapse and increasing serotonin synaptic levels (Voineskos et al. 2007). In summary, multiple mechanisms can lead to insufficient levels of serotonin in neuronal synapses, which contributes significantly to the risk for depression and for suicide.

Serotonin Receptors

In 1957, the existence of two separate serotonin receptors was first proposed, primarily because of the opposing phenomenon this neurotransmitter produces, in reference to cholinergic mediation of smooth muscle contraction (Gaddum and Picarelli 1957). Through the use of more precise molecular cloning and pharmacological and biochemical studies, seven distinct serotonin receptor families have been identified (5-HT_1–7), many of which contain several subtypes. With the exception of the 5-HT₃ receptor, which is an excitatory ionotropic receptor, all the other serotonin receptors are GPCRs. The 5-HT_1A,B,D,E,F receptors are negatively coupled to adenylyl cyclase; the 5-HT_2A,B,C subtypes are positively coupled to PLC; and the 5-HT₄, 5-HT₅, 5-HT₆, and 5-HT₇ subtypes are positively coupled to adenylyl cyclase (see Figure 2–3B) (Humphrey et al. 1993; Nestler et al. 2015). When all types and subtypes are counted, 13 serotonin receptors are identified in humans (Nestler et al. 2015).

5-HT₁ receptors. 5-HT_1A receptors are found in particularly high density in several limbic structures, including the hippocampus, septum, amygdala, and entorhinal cortex, as well as on serotonergic neuron cell bodies, where they serve as autoreceptors regulating serotonin neuronal firing rates (Blier et al. 1998; Cooper et al. 2003; Pazos et al. 1985). The highest density of labeling is found in the dorsal raphe, with lower densities observed in the remaining raphe nuclei (Pazos et al. 1985). The density and messenger RNA (mRNA) expression of 5-HT_1A receptors appear insensitive to reductions in serotonin transmission associated with lesioning the raphe or administering the serotonin-depleting agent p-chlorophenylalanine. Similarly, elevation of serotonin transmission, resulting from chronic administration of an SSRI or MAOI, does not consistently alter 5-HT_1A receptor density or mRNA in the cortex, hippocampus, amygdala, or hypothalamus. In contrast to the insensitivity to serotonin, 5-HT_1A receptor density is downregulated by adrenal steroids. Postsynaptic 5-HT_1A receptor gene expression is under tonic inhibition by adrenal steroids in the hippocampus and some other regions. Thus, in rodents, hippocampal 5-HT_1A receptor mRNA expression is increased by adrenalectomy and decreased by corticosterone administration or chronic stress. The stress-induced downregulation of 5-HT_1A receptor expression is prevented by adrenalectomy. Mineralocorticoid receptor stimulation has the most potent effect on downregulating 5-HT_1A receptors, although glucocorticoid receptor stimulation also contributes to this effect.

In addition to being expressed on neurons, postsynaptic 5-HT_1A receptors are also abundantly expressed by astrocytes and some other glia (Whitaker-Azmitia et al. 1990) (see Figure 2–7 later in this chapter). Stimulation of astrocyte-based 5-HT_1A sites causes astrocytes to acquire a more mature morphology and to release the trophic factor S-100β, which promotes growth and arborization of serotonergic axons. Administration of 5-HT_1A receptor antagonists, antibodies to S-100β, or agents that deplete serotonin produces similar losses of dendrites, spines, and/or synapses in adult and developing animals—effects that are blocked by administration of 5-HT_1A receptor agonists or SSRIs. These observations have led to the hypothesis that a reduction of 5-HT_1A receptor function may play an important role in mood disorders that are known to be associated with glial reductions (Manji et al. 2001). The use of conditional knockouts of the 5-HT_1A receptor, in which gene expression is altered only in particular anatomical regions and/or during particular times, has illustrated the caution necessary in attributing complex behaviors to simple “too much” or “too little” neurotransmitter or receptor hypotheses. One report used a knockout/rescue approach with regional and temporal specificity to show that the anxiety-related effect of the 5-HT_1A receptor knockout was actually developmental. Specifically, expression limited to the hippocampus and cortex during early postnatal development was sufficient to counteract the anxious phenotype of the mutant, even though the receptor was still absent in adulthood (Gross et al. 2002). As is discussed in the chapters on antidepressants (see Chapters 8–23), there is interest in the observation that antidepressants enhance hippocampal neurogenesis (Duman 2002; Malberg et al. 2000). It is noteworthy that data suggest that 5-HT_1A receptor activation is required for SSRI-induced hippocampal neurogenesis in mice (Jacobs et al. 2000). Altering serotonin levels with the SSRI fluoxetine does not affect division of stem cells in the dentate gyrus, but rather increases symmetric divisions of an early progenitor cell class that exists after stem-cell division (Encinas et al. 2006).

5-HT_1A receptors are now known to use a variety of signaling mechanisms to bring about their effects in distinct brain areas. Thus, somatodendritic 5-HT_1A receptors appear to inhibit the firing of serotonergic neurons by opening a K⁺ channel through a pertussis toxin–sensitive G protein (likely G_o, discussed later in the section on G proteins) (Andrade et al. 1986), as well as by reducing cAMP levels. Postsynaptic 5-HT_1A receptors appear to exert many of their effects by inhibiting adenylyl cyclase via G_i (De Vivo and Maayani 1990) but also have been found to potentiate the activity of certain adenylyl cyclases (Bourne and Nicoll 1993) and to stimulate inositol-1,4,5-triphosphate (IP₃) production and activate PKC (Liu and Albert 1991). Structurally, the 5-HT_1A receptors are more related to D₂ receptors than to other serotonin receptors (Squire 2013). Functional polymorphisms in the promotor of 5-HT_1A receptors influencing 5-HT_1A receptor expression did not lead to changes in binding of 5-HT_1A receptor antagonists in healthy subjects based on two PET studies, suggesting that there are compensatory mechanisms in the brains of healthy living human subjects (Bortolato et al. 2013; David et al. 2005; Lothe et al. 2010). PET studies also reported that 5-HT_1A receptors decrease with advancing age in men, which may potentially explain the increased risk of suicide in this population (Moses-Kolko et al. 2011). A functional promoter polymorphism, rs6295, for which the G allele causes decreased transcription of 5-HT_1A, may have a role in increasing risk for depression and suicide, because it is overrepresented in patients with depression, although the mechanism is not clear, because postmortem brain samples from depressed individuals who committed suicide showed more equal expression of the G allele with the C allele (Donaldson et al. 2016; Savitz et al. 2009). Buspirone is a 5-HT_1A receptor agonist indicated for treating generalized anxiety disorder (Nestler et al. 2015).

5-HT_1D receptors are virtually absent in rodents but have been detected in guinea pigs and humans (Bruinvels et al. 1993). On the basis of an approximately 74% sequence homology, it has been proposed that 5-HT_1B receptors are the rodent homologue of 5-HT_1D receptors (Saxena et al. 1998). Furthermore, the distribution of the 5-HT_1D receptors in guinea pigs and humans is approximately equivalent to that of the 5-HT_1B receptors in rats (Bruinvels et al. 1993). Both 5-HT_1B and 5-HT_1D receptors have been proposed to represent the major nerve terminal autoreceptors regulating the amount of serotonin released per nerve impulse (Piñeyro and Blier 1999) (see Figure 2–3B). Like 5-HT_1A receptors, 5-HT_1B and 5-HT_1D receptors inhibit cAMP formation and stimulate IP₃ production and activate PKC (Schoeffter and Bobirnac 1995). This appears to be the case for many receptors coupled to G_i and G_o (Table 2–1). The α subunits of the G protein (α_i and α_o) inhibit adenylyl cyclase and regulate ion channels, respectively, whereas the βγ subunits activate PLC isozymes to stimulate IP₃ production and activate PKC. Examples of 5-HT_1B/1D receptor agonists include the triptan family of medications used to treat migraines, which are incidentally contraindicated in patients who have comorbid coronary artery disease because of the presence of 5-HT_1B receptors causing constriction of coronary arteries (Lambert 2005; Nestler et al. 2015). 5-HT_1B receptor agonists have shown promise for decreasing reactive aggression and alcohol intake in animal models, but human polymorphism studies searching for a relationship between HTR1B polymorphisms and aggression or suicide risk have yielded inconsistent results (Bortolato et al. 2013).

TABLE 2–1. Key features of G protein subunits

G protein class	Members	Effector(s)/functions	Examples of receptors
αi	Gαi1–3, Gαo	AC (+)	α2, D2, A1, μ, M2, 5-HT1A
		Ligand-type Ca2+ channels (+)	Olfactory signals
	Gαz, Gαt1–2	K+ channels (+)
		Ca2+ channels (−)a	GABAB
		cGMP	Retinal rods, cones (rhodopsins)
		Phosphodiesterase (+) (Gαt1–2)
αq	Gαq, Gα11, Gα14, Gα15, Gα16	PLC-β (+)	TxA2, 5-HT2C, M1, M3, M5, α1
α12	Gα12, Gα13	RGS domain–containing rho exchange factors	TxA2, thrombin
βb	β (×5)	AC type I (−); AC types II, IV (potentiation)
		PLC (+)
		Receptor kinases (+)
		Inactivates αs
γ	γ (×12)	β γ subunits required for interaction of α subunit with receptor
Note. AC=adenylyl cyclase; A1, A2=adenosine receptor subtypes; β1, α1, α2=adrenergic receptor subtypes; cGMP=cyclic guanosine monophosphate; D1, D2=dopamine receptor subtypes; Gαo=olfactory, but also found in limbic areas; Gαt=transducin; GABAB=γ-amino-butyric acid receptor subtype; 5-HT1A, 5-HT2C=serotonin receptor subtypes; M1, M2, M3, M5=muscarinic receptor subtypes; μ=opioid μ receptor; PLC=phospholipase C; RGS=regulators of G protein signaling; TxA2=thromboxane A2 receptor. aAlthough regulation of Na+/H+ exchange and Ca2+ channels by Gα1–2 and Gα1–3 has been demonstrated in artificial systems in vitro, these findings await definitive confirmation. bEffectors are regulated by β γ subunits as a dimer.

The 5-HT_1C receptors have structural and transductional similarities to the 5-HT₂ receptor class (Hoyer et al. 1986).

5-HT₂ receptors. The three subtypes of 5-HT₂ receptors are 5-HT_2A, 5-HT_2B, and 5-HT_2C. The highest level of 5-HT_2A binding sites and mRNA for these receptors exists in the cortex, and these receptors have been implicated in the psychotomimetic effects of agents such as lysergic acid diethylamide, as reviewed by Aghajanian and Marek (1999). In addition, lesioning serotonin neurons with 5,7-DHT does not reduce the 5-HT₂ receptor density reported in brain regions (Hoyer et al. 1986), indicating that these receptors are primarily (if not exclusively) postsynaptic. Autoradiography performed with the potent and selective radioligand [³H]MDL 100,907 has localized 5-HT_2A receptors to many similar brain regions in the rat and primate brain (López-Giménez et al. 1997). Hallucinogenesis is likely a result of cortical 5-HT_2A receptor activation, because experiments in mice expressing 5-HT_2A receptors in the cortex only incur receptor signaling and behavior changes to hallucinogenic drugs similar to that of genetically unaltered mice (González-Maeso et al. 2007). Competition studies with other radioligands (Westphal and Sanders-Bush 1994) and their mRNA distribution indicate that 5-HT_2C receptors are considerably widespread throughout the CNS, with the highest density in the choroid plexus (Hoffman and Mezey 1989). 5-HT_2B receptors are detected sparingly in the brain and are more prominently located in the fundus, gut, kidney, lungs, and heart (Hoyer et al. 1986).

Evidence from animal experiments in which cortical 5-HT_2A receptors are disrupted indicates a specific role of these receptors in modulation of conflict anxiety without affecting fear conditioning and depression-like behaviors (Weisstaub et al. 2006). Furthermore, chronic administration of many antidepressants downregulates 5-HT₂ receptors, suggesting that this effect may be important for their efficacy (Scott and Crews 1986). However, chronic electroconvulsive shock appears to upregulate 5-HT₂ expression, precluding a simple mechanism for antidepressant efficacy. The obesity seen in 5-HT_2C knockout animals suggests that in addition to histamine receptor blockade, 5-HT_2C receptor blockade may play a role in the weight gain observed with certain psychotropic agents; this is an area of considerable research. In keeping, evidence suggests that the weight gain “orexigenic” properties of atypical antipsychotics are likely due to potent activation of hypothalamic AMP-kinase through histamine type 1 (H₁) receptors (Kim et al. 2007). Aggressive and impulsive behaviors are also linked to the serotonin system, in which a neuropeptide Y (NPY) and serotonin interaction from studies in NPY knockout mice detected circuits responsible for aggressive behavior linked to feeding (Emeson and Morabito 2005). The regulation of 5-HT₂ receptors is intriguing, as not only is it important in psychiatric disorders and therapeutic benefit, but also both agonists and antagonists appear to cause an internalization of the receptor. Moreover, data suggest that mRNA editing may play an important role in regulating the levels and activity of this receptor subtype (Niswender et al. 1998). All of the 5-HT₂ receptor subtypes are linked to the phosphoinositide signaling system, and their activation produces IP₃ and diacylglycerol (DAG), via PLC activation (Conn and Sanders-Bush 1987) (see Figure 2–3B).

A pharmacogenetic study searched for genetic predictors of treatment outcome in 1,953 patients with MDD who received the antidepressant citalopram in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study and were prospectively assessed (McMahon et al. 2006). In a split-sample design, a selection of 68 candidate genes was genotyped with 768 single-nucleotide-polymorphism markers chosen to detect common genetic variation. A significant and reproducible association was found between treatment outcome and a marker in HTR2A (P=1×10^–6 to 3.7×10^–5 in the total sample). The “A” allele (associated with better outcome) was six times more frequent in white than in black participants, for whom treatment was also less effective in this sample (McMahon et al. 2006). The A allele thus may contribute to racial differences in outcomes of antidepressant treatment. Taken together with prior neurobiological findings, these genetic data make a compelling case for a key role of HTR2A in the mechanism of antidepressant action.

A leading hypothesis for the mechanism of action of atypical antipsychotic agents suggests that the ratio of D₂-to-5-HT₂ blockade confers “atypicality” properties of many currently available antipsychotic agents (Meltzer 2002). Several antidepressants (e.g., mianserin, mirtazapine) and atypical antipsychotics (e.g., clozapine, risperidone, olanzapine) are antagonists of 5-HT_2A receptors, raising the possibility that blockade of 5-HT₂ receptors may play an important role in the therapeutic efficacy of these agents (Nestler et al. 2015). Studies comparing the various compounds and the state of 5-HT_2A/2C receptors based on psychiatric phenotype would be enhanced by the development of a highly specific PET ligand, [¹⁸F]FECIMBI-36, which was confirmed in postmortem human brain tissue sections to bind at high levels to the prefrontal cortex (PFC), temporal cortex, and hippocampus for 5-HT_2A receptors and to the choroid plexus for 5-HT_2C receptors (Prabhakaran et al. 2015).

5-HT_2C receptors are expressed on pro-opiomelanocortin (POMC) neurons of the arcuate nucleus of the hypothalamus (Nestler et al. 2015). The selective 5-HT_2C receptor agonist lorcaserin is used for the treatment of obesity (Nestler et al. 2015). Many second-generation antipsychotic drugs cause weight gain through 5-HT_2C receptor antagonism (Nestler et al. 2015).

5-HT_3–7 receptors. Unlike the other serotonin receptors, 5-HT₃ receptors are ligand-gated ion channels capable of mediating fast synaptic responses, although the opening of the channel is relatively slower compared with other ligand-gated ion channels (see Figure 2–3B) (Squire 2013; Yun and Rhim 2011). The cis–trans isomerization and molecular rearrangement at proline 8 is the structural mechanism that opens the 5-HT₃ receptor protein pore (Lummis et al. 2005). 5-HT₃ receptors are present in multiple brain areas, including the hippocampus, dorsal motor nucleus of the solitary tract, and area postrema (Laporte et al. 1992). The 5-HT₃ receptor, located mostly in the peripheral nervous system, is effectively modulated by a variety of compounds, such as alcohols and anesthetics, and antagonists of this receptor are used as effective antiemetics in patients who are undergoing chemotherapy (e.g., ondansetron) (Squire 2013).

5-HT₄, 5-HT₆, and 5-HT₇ are GPCRs that are preferentially coupled to G_s and activate adenylyl cyclases (see Figure 2–3B). 5-HT₄ receptors are able to modulate the release of monoamines and GABA in the brain, appear to improve memory and learning in mice, and mediate pituitary prolactin release in the presence of estrogen (Darcet et al. 2016; Papageorgiou and Denef 2007). 5-HT₄ receptor subtypes include 5-HT_4A, 5-HT_4B, 5-HT_4C, 5-HT_4D, 5-HT_4E, 5-HT_4F, 5-HT_4G, 5-HT_4H, and 5-HT_4HB (Yun and Rhim 2011). 5-HT₅ receptors are located in the hypothalamus, hippocampus, corpus callosum, fibra, cerebral ventricles, and glia (Hoyer et al. 2002). The 5-HT_5A receptor is negatively coupled to adenylyl cyclase, whereas the 5-HT_5B receptor is not functional because of stop codons (Grailhe et al. 2001). 5-HT₆ receptors are located in the striatum, amygdala, nucleus accumbens, hippocampus, cortex, and olfactory tubercle (Hoyer et al. 2002). Of interest, many antipsychotic agents and antidepressants are high-affinity antagonists of 5-HT₆ and 5-HT₇ receptors, and 5-HT₆ receptor antagonism is being specifically investigated for its ability to improve cognition in patients with Alzheimer’s disease (Parker et al. 2015; Roth et al. 1994; Yun and Rhim 2011). 5-HT₇ receptors, made up of subtypes 5-HT_7A, 5-HT_7B, 5-HT_7C, and 5-HT_7D, have been localized to the cerebral cortex, medial thalamic nuclei, substantia nigra, central gray, and dorsal raphe nucleus (Hoyer et al. 2002; Yun and Rhim 2011). Chronic treatment with antidepressants downregulates 5-HT₇ receptors, whereas acute stress has been reported to increase 5-HT₇ expression (Sleight et al. 1995; Yau et al. 2001). Lurasidone and vortioxetine, agents used to treat schizophrenia and depression, respectively, are also nonselective 5-HT₇ receptor antagonists that work through increased AMPA-mediated neurotransmission (Andreetta et al. 2016; Fountoulakis et al. 2015; Sanchez et al. 2015).

In the future, the subtle differences between the different serotonin receptors and their subtypes will be elucidated as specific PET ligands, genotype–phenotype correlations, and receptor-selective medications are developed (Kumar and Mann 2014).

Dopaminergic System

Dopamine was originally thought to be simply a precursor of norepinephrine and epinephrine synthesis, but the demonstration that its distribution in the brain was quite distinct from that of norepinephrine led to extensive research establishing its role as a critical, unique neurotransmitter. Dopamine synthesis requires transport of the amino acid L-tyrosine across the blood–brain barrier and into the cell. Once tyrosine enters the neuron, the rate-limiting step for dopamine synthesis is conversion of L-tyrosine to L-dihydroxyphenylalanine (L-dopa) by the enzyme tyrosine hydroxylase; L-dopa is readily converted to dopamine and, hence, is used as a precursor strategy to correct a dopamine deficiency in the treatment of Parkinson’s disease (Figure 2–4B). The activity of tyrosine hydroxylase can be regulated by many factors, including the activity of catecholamine neurons; furthermore, catecholamines function as end-product inhibitors of tyrosine hydroxylase by competing with a tetrahydrobiopterin cofactor (Cooper et al. 2003).

In contrast to the widespread serotonin and norepinephrine projections, dopamine neurons form more discrete circuits, with the nigrostriatal, mesolimbic, tuberoinfundibular, and tuberohypophysial pathways constituting the major CNS dopaminergic circuits (Figure 2–4A). The nigrostriatal circuit is composed of dopamine neurons from the mesencephalic reticular formation (region A8) and the pars compacta region of the substantia nigra (region A9) of the mesencephalon. These neurons give rise to axons that travel via the medial forebrain bundle to innervate the caudate nucleus and putamen (Andén et al. 1964; Ungerstedt 1971). The dopamine neurons that make up the nigrostriatal circuit have been assumed to be critical for maintaining normal motor control, because destruction of these neurons is associated with Parkinson’s disease; however, it is now clear that these projections subserve a variety of additional functions. For instance, evidence from human brain imaging studies indicates that drugs that modulate striatal dopamine receptor activation correlate with the subject’s ability to choose gradations of rewarding actions during instrumental learning tasks. This further implies that the dopamine reward pathway in the brain is likely convergent on many discrete brain circuits and neurotransmitter alterations and shows that striatal activity can also account for how the human brain proceeds toward making future decisions based on reward prediction (Pessiglione et al. 2006).

The mesolimbic dopamine circuit consists of dopamine neurons located in the midbrain just medial to the A9 cells in an area termed the ventral tegmental area (VTA) (Cooper et al. 2003; Nestler et al. 2015; Squire 2013). This circuit shares some similarities to the nigrostriatal circuit in that it is a parallel circuit consisting of axons that make up the medial forebrain bundle. However, these axons ascend through the lateral hypothalamus and project to the nucleus accumbens; olfactory tubercle; bed nucleus of the stria terminalis; lateral septum; and frontal, cingulate, and entorhinal regions of the cerebral cortex (Cooper et al. 2003). This circuit innervates many limbic structures known to play critical roles in motivational, motor, and reward pathways and has therefore been implicated in a variety of clinical conditions, including psychosis and drug abuse (Cooper et al. 2003). Data also suggest a potential role of dopamine—and, in particular, mesolimbic pathways—in the pathophysiology of bipolar mania, as well as bipolar and unipolar depression (Beaulieu et al. 2004; Dunlop and Nemeroff 2007; Goodwin et al. 2007; Roybal et al. 2007). It is perhaps surprising that the role of the dopaminergic system in the pathophysiology of mood disorders has not received greater study, because it represents a prime candidate on several theoretical grounds. The motoric changes in bipolar disorder are perhaps the most defining characteristics of the illness, ranging from near catatonic immobility to the hyperactivity of manic states. Similarly, loss of motivation is one of the central features of depression, whereas anhedonia and “hyperhedonic states” are among the most defining characteristics of bipolar depression and mania, respectively. In this context, it is noteworthy that the midbrain dopaminergic system is known to play a critical role in regulating not only motoric activity but also motivational and reward circuits. It is clear that motivation and motor function are closely linked and that motivational variables can influence motor output both qualitatively and quantitatively. Furthermore, considerable evidence indicates that the mesolimbic dopaminergic pathway plays a crucial role in the selection and orchestration of goal-directed behaviors, particularly those elicited by incentive stimuli (Goodwin et al. 2007).

The firing pattern of mesolimbic dopamine neurons appears to be an important regulatory mechanism; thus, in rats, electrical or glutamatergic stimulation of medial PFC elicits a burst firing pattern of dopaminergic cells in the VTA and increases dopamine release in the nucleus accumbens (Murase et al. 1993; Taber and Fibiger 1993). The burst firing of dopamine cell activity elicits more terminal dopamine release per action potential than the nonbursting, pacemaker firing pattern (Roth et al. 1987). The phasic, burst firing of dopamine neurons and accompanying rise in dopamine release normally occur in response to primary rewards (until they become fully predicted) and reward-predicting stimuli. Such a role also has been postulated to provide a neural mechanism by which PFC dysfunction could alter hedonic perceptions and motivated behavior in mood disorders (Drevets et al. 2002). Studies indicate that the amygdala has importance in the learning of new cocaine drug-seeking responses and its habit-forming properties (Lee et al. 2005). The supraphysiological levels of dopamine induced by cocaine and other drugs of abuse lead to powerful reinforcement of drug-seeking behavior, by co-opting the dopamine reward circuit of the brain, as reviewed elsewhere (Volkow and Morales 2015).

Dopamine Transporters

As with serotonin, the dopamine signal in the synaptic cleft is terminated primarily by reuptake into the presynaptic terminal. The DAT comprises 12 transmembrane domains and is located somatodendritically as well as on dopamine nerve terminals (see Figure 2–4B). Like other monoamine transporters, the DAT functions as a Na⁺/K⁺ pump to clear dopamine from the synaptic cleft on its release. However, data suggest that many drugs of abuse are capable of altering the function of these transporters. Thus, the amphetamines are thought to mediate their effects, in part, by reversing the direction of the transporter so that it releases dopamine. Cocaine is capable of blocking the reuptake of DAT, leading to an increase in dopamine in the synaptic cleft. Of interest, altered neuronal long-term potentiation in the VTA in response to chronic cocaine exposure has been linked to drug-associated memory and likely contributes to the powerful addictive potential of this drug of abuse (Liu et al. 2005). Dopamine in the medial frontal cortex is taken up predominantly by the norepinephrine transporter, which goes against the dogma of transporters being able to selectively take up only their respective neurotransmitter. Furthermore, this provides a mechanism by which norepinephrine reuptake–inhibiting antidepressants also may increase synaptic levels of dopamine in the frontal cortex, an effect that may be therapeutically very important. Interestingly, a meta-analysis of single photon emission computed tomography (SPECT) scans examining the DAT gene SLC6A3 variable-number tandem repeat (VNTR) polymorphism did not find significant changes in the levels of the dopamine transporters in the brains of patients with schizophrenia, attention-deficit/hyperactivity disorder (ADHD), and even Parkinson’s disease (Costa et al. 2011). However, another DAT polymorphism of note, a VNTR in the nontranslated 3′ end of exon 15, causes a 25% decreased density of DAT in humans with a VNTR of 9 repeats instead of the more common 10 repeats (Lacerda-Pinheiro et al. 2014). SPECT scanning of the brain with the presynaptic DAT radioligand is now being used to distinguish Parkinson’s disease syndromes from other causes of parkinsonism such as antipsychotic-induced parkinsonism, the latter of which does not show a deficit in DAT binding in the caudate and putamen (Tatsch and Poepperl 2013).

In patients with MDD and bipolar disorder, SPECT studies showed that DAT availability is higher—and, by inference, synaptic dopamine is lower—in patients with depression (Camardese et al. 2014a, 2014b). PET studies identified decreased binding potential of the DAT PET ligand to the dopamine transporter in the striatum in MDD patients and more specifically in the caudate in bipolar patients (Anand et al. 2011; Meyer et al. 2001; Savitz and Drevets 2013). Postmortem studies in patients with MDD showed lower DAT levels in the amygdala and higher D₂ and D₃ receptor levels (no change in D₁ receptors), consistent with similar observations in rats with dopamine depletion (Klimek et al. 2002).

To treat refractory depression, the new triple transporter reuptake inhibitor compound BMS-820836, reported by PET studies to specifically inhibit monoamine transporters DAT, SERT, and NET, was developed for the goal of increasing synaptic levels of the respective monoamines dopamine, serotonin, and norepinephrine and thereby decreasing depressive symptoms (Risinger et al. 2014). Moderate to severe MDD or atypical MDD (with reversed neurovegetative symptoms) was found by PET studies to be associated with higher levels of MAO-A, an enzyme that catabolizes and thus decreases synaptic dopamine, serotonin, and norepinephrine, presumably increasing depressive symptoms (Chiuccariello et al. 2014). Classic MAOIs increase all three monoamines in the synapse by blocking the catabolizing enzyme MAO (both A and B enzyme subtypes).

Bupropion is widely believed to increase synaptic dopamine by blocking the DAT-mediated reuptake, but a PET study showed that it has at most only 22% occupancy of the DAT (Meyer et al. 2002). Given the observation of a decreased binding potential of 15% of DAT in patients with MDD, it remains unknown whether bupropion occupies the same binding site on DAT as the PET ligand, whether 22% occupancy of DAT is sufficient for bupropion to work (as compared with 80% binding to SERT for SSRIs), and whether bupropion works by another uncharacterized mechanism to achieve its antidepressant properties (Meyer et al. 2001, 2002).

Dopaminergic Receptors

The existence of two subtypes of dopamine receptors, D₁ and D₂, was initially established with classic pharmacological techniques in the 1970s (Stoof and Kebabian 1984). Subsequent molecular biological studies have shown that the D₁ family contains both the D₁ and the D₅ receptors, whereas the D₂ family contains the D₂, D₃, and D₄ receptors (Cooper et al. 2003). D₁ receptor family members were originally defined solely on the ability to stimulate adenylyl cyclase, whereas the D₂ family inhibited the enzyme. Interestingly, dopamine receptors complexed with subunits from other subclasses of dopamine receptors within a receptor family are able to form distinct hetero-oligomeric receptors also termed kissing cousin receptors. Notably, hetero-oligomeric D₁–D₂ receptor complexes in the brain require binding to active sites of both receptor subtypes to induce activation of the hetero-oligomeric receptor complex. These receptors have been shown to use traditional D₁ receptor intracellular signaling components of G_q/11 and Ca²⁺/calmodulin–dependent protein kinase II (CaMKII) second-messenger activation as seen in the nucleus accumbens (Rashid et al. 2007). This opens up possibilities for the brain to use different receptor subunit proportions to further fine-tune brain neurophysiology. Similar to the DAT exon 15 VNTR, the D₂ dopamine receptor gene (DRD2) has a TaqIA polymorphism (allele A1+) in the 3′ nontranslated region that reduces D₂ receptor density (Lacerda-Pinheiro et al. 2014).

D₁ and D₅ receptors. The D₁ and D₅ receptors stimulate adenylyl cyclase activity via the activation of G_s or G_olf (a G protein originally thought to be present exclusively in olfactory tissue but now known to be abundantly present in limbic areas) (see Figure 2–4B). Other second-messenger pathways also have been reported to be activated by D₁ receptors, an effect that may play a role in the reported D₁–D₂ cross-talk (Clark and White 1987). The frontal cortex contains almost exclusively D₁ receptors (Clark and White 1987), suggesting that this receptor may play an important role in higher cognitive function and perhaps in the actions of medications such as methylphenidate. A D₁/D₅ receptor agonist, dihydrexidine, has been developed and tested for its ability to improve cognition and working memory in patients with schizophrenia but was prematurely terminated because of hypotension and potentially increased seizure risk (Arnsten et al. 2017). Although the D₅ receptor is a neuron-specific receptor located primarily in limbic areas of the brain, no compounds have been developed so far that pharmacologically distinguish it from the homologous D₁ receptor (Arnsten et al. 2017). Drugs such as cocaine and methylphenidate cause peak dopamine activation of D₁ receptors on striatal GABAergic medium spiny neurons, which plays a key role in the direct striatal pathway mediating expectations of reward (Volkow and Morales 2015).

D₂ receptors. Four types of D₂ receptors have been identified: two subtypes of D₂ receptors (the short and long splice variants, D_2S and D_2L, respectively) and D₃ and D₄ receptors. Although a seemingly identical pharmacological profile for these receptors exists, there are undoubtedly physiological differences between the two subtypes. D₂ receptors mediate their cellular effects via the G_i/G_o proteins and thereby several effectors (see Figure 2–4B). In addition to the well-characterized inhibition of adenylyl cyclase, D₂ receptors in different brain areas also regulate PLC, bring changes in K⁺ and Ca²⁺ currents, and possibly regulate phospholipase A₂. D₂ receptors are located on cell bodies and nerve terminals of dopamine neurons and function as autoreceptors. Thus, activation of somatodendritic D₂ receptors reduces dopamine neuron firing activity, likely via the opening of K⁺ channels, whereas activation of nerve-terminal D₂ autoreceptors reduces the amount of dopamine released per nerve impulse, in large part by closing voltage-gated Ca²⁺ channels. As discussed extensively elsewhere in this volume (see Chapter 24, “Classic Antipsychotic Medications,” by Nasrallah and Tandon), D₂ receptors have long been implicated in the pathophysiology and treatment of schizophrenia. In one study, transgenic mice overexpressing D₂ receptors in the striatum had many phenotypic hallmarks of schizophrenia (Kellendonk et al. 2006). However, the potential role of D₂ receptors in contributing to depression is a comparatively more recent recognition, confirmed by the observation that D₂ receptor binding is increased by PET ligand [¹¹C]raclopride in the putamen of patients with MDD and psychomotor retardation, reflecting decreased dopamine levels in the synapse (Meyer et al. 2006).

The typical antipsychotics are generally defined by their high-affinity antagonism or blockade of the D₂ receptor, which substantially prevents synaptic dopamine from activating the postsynaptic receptor but also has a significant risk for extrapyramidal side effects (EPS) such as dystonia and akathisia in the short term and parkinsonism and tardive dyskinesia in the longer term (see Figure 2–4B). Medications with high anticholinergic properties, such as diphenhydramine and benztropine, are frequently used to avoid development of EPS, but tardive dyskinesia can develop nevertheless after many years of antipsychotic exposure. Cocaine and methylphenidate stimulation of D₂ receptors in the indirect, inhibitory striatal pathway, which mediates punishment, leads to sustained motivation to procure these drugs in the future (Volkow and Morales 2015).

D₃ receptors. D₃ receptors possess a different anatomical distribution than D₂ receptors and, because of their preferential limbic expression, have been postulated to represent an important target for antipsychotic drugs. Numerous studies have investigated the position association of a polymorphism in the coding sequence of the D₃ receptor with schizophrenia, with equivocal results. It has been suggested that brain-derived neurotrophic factor (BDNF) may regulate behavioral sensitization via its effects on D₃ receptor expression (Guillin et al. 2001). Cariprazine, an antipsychotic with possible efficacy in treating acute mania, has antagonist/partial agonist binding to D₂ and D₃ receptors, with a much higher affinity for D₃ receptors compared with other antipsychotics, but it still has significant D₂-related EPS (Altinbaş et al. 2013). Buspirone, a 5-HT_1a receptor partial agonist, also has D₃/D₄ receptor antagonist activity, so it is being considered for treating methamphetamine and cocaine dependence, given the observation of increased brain D₃ receptors by PET studies in addicted individuals and the attenuation of self-administration in animals given D₃ receptor antagonists (Paterson et al. 2014).

D₄ receptors. The D₄ receptor has received much interest in psychopharmacological research because of the fact that clozapine has a high affinity for this receptor. More selective D₄ antagonists are being explored as adjunctive agents in the treatment of schizophrenia. Furthermore, considerable attention has focused on the possibility that genetic D₄ variants may be associated with thrill-seeking behavior (Zuckerman 1985), ADHD (Roman et al. 2001), and responsiveness to clozapine (Van Tol et al. 1992). A potent, selective antagonist of the D₄ receptor called ML398 has been developed that is still being investigated (Berry et al. 2014).

Noradrenergic System

Named sympathine because it was initially encountered as being released by sympathetic nerve terminals, the molecule was later given the name norepinephrine after meeting the criteria for a neurotransmitter in the CNS (Cooper et al. 2003). Norepinephrine is produced from the amino acid precursor L-tyrosine found in neurons in the brain, chromaffin cells, sympathetic nerves, and ganglia. The enzyme dopamine β-hydroxylase converts dopamine to norepinephrine, and as is the case for dopamine synthesis, tyrosine hydroxylase is the rate-limiting enzyme for norepinephrine synthesis (Figure 2–5B). The dietary depletion of tyrosine and α-methyl-p-tyrosine (a tyrosine hydroxylase inhibitor) has played an important part in efforts aimed at delineating the role of catecholamines in the pathophysiology and treatment of mood and anxiety disorders (Coupland et al. 2001; McCann et al. 1995).

FIGURE 2–5. The noradrenergic system.

See Plate 8 to view this figure in color.

This figure depicts the noradrenergic projections throughout the brain (A) and the various regulatory processes involved in norepinephrine (NE) neurotransmission (B). NE neurons innervate nearly all parts of the neuroaxis, with neurons in the locus coeruleus being responsible for most of the NE in the brain (90% of NE in the forebrain and 70% of total NE in the brain). The amino acid L-tyrosine is actively transported into presynaptic NE nerve terminals, where it is ultimately converted into NE. The rate-limiting step is conversion of L-tyrosine to L-dihydroxyphenylalanine (L-dopa) by the enzyme tyrosine hydroxylase (TH). α-Methyl-p-tyrosine (AMPT) is a competitive inhibitor of tyrosine hydroxylase and has been used to assess the impact of reduced catecholaminergic function in clinical studies. Aromatic amino acid decarboxylase (AADC) converts L-dopa to dopamine (DA). L-dopa then becomes decarboxylated by decarboxylase to form DA. DA is then taken up from the cytoplasm into vesicles, by vesicular monoamine transporters (VMATs), and hydroxylated by dopamine β-hydroxylase (DBH) in the presence of O₂ and ascorbate to form NE. Normetanephrine (NM), which is formed by the action of catechol-O-methyltransferase (COMT) on NE, can be further metabolized by monoamine oxidase (MAO) and aldehyde reductase to 3-methoxy-4-hydroxyphenylglycol (MHPG). Reserpine causes a depletion of NE in vesicles by interfering with uptake and storage mechanisms (depressive-like symptoms have been reported with this hypertension). Once released from the presynaptic terminal, NE can interact with a variety of presynaptic and postsynaptic receptors. Presynaptic regulation of NE neuron firing activity and release occurs through somatodendritic (not shown) and nerve-terminal α₂ adrenoreceptors, respectively. Yohimbine potentiates NE neuronal firing and NE release by blocking these α₂ adrenoreceptors, thereby disinhibiting these neurons from a negative feedback influence. Conversely, clonidine attenuates NE neuron firing and release by activating these receptors. Idazoxan is a relatively selective α₂ adrenoreceptor antagonist primarily used for pharmacological purposes. The binding of NE to G protein receptors (G_o, G_i, etc.) that are coupled to adenylyl cyclase (AC) and phospholipase C–β (PLC-β) produces a cascade of second-messenger and cellular effects (see diagram and later sections of the text). NE has its action terminated in the synapse by rapidly being taken back into the presynaptic neuron via NE transporters (NETs). Once inside the neuron, it can either be repackaged into vesicles for reuse or undergo enzymatic degradation. The selective NE reuptake inhibitor and antidepressant reboxetine and older-generation tricyclic antidepressant desipramine are able to interfere/block the reuptake of NE. On the other hand, amphetamine is able to facilitate NE release by altering NET function. Green spheres represent DA neurotransmitters; blue spheres represent NE neurotransmitters. cAMP=cyclic adenosine monophosphate; DAG=diacylglycerol; DOPA=dihydroxyphenylalanine; IP₃=inositol-1,4,5-triphosphate; NA=nucleus accumbens.

Source. Adapted from Cooper JR, Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology, 8th Edition. New York, Oxford University Press, 2001. Copyright 1970, 1974, 1978, 1982, 1986, 1991, 1996, 2001 by Oxford University Press, Inc. Used by permission of Oxford University Press, Inc. Modified from Nestler et al. 2001.

The mammalian CNS has seven norepinephrine cell groups, designated A1 through A7. In the brain stem, these are the lateral tegmental neurons (A5 and A7) and the locus coeruleus neurons (A6) (Dahlström 1971) (see Figure 2–5B). In general, the projections from A5 and A7 are more restricted to brain stem areas and do not interact with those of A6. The term locus coeruleus (LC) was derived from the Greek because of its saddle shape and its “blueish color” (caeruleum). The LC is the most widely projecting CNS nucleus known (Foote et al. 1983), responsible for approximately 90% of the norepinephrine innervation of the forebrain and 70% of the total norepinephrine in the brain (Figure 2–5A). Indeed, the LC norepinephrine neurons, although small in number, constitute a diffuse system of projections to widespread brain areas via highly branched axons. The extensive efferent innervation suggests that the LC plays a modulatory and integrative role rather than a role in specific sensory or motor processing (Foote et al. 1983).

Several physiological roles have been ascribed to the LC, notably in the control of vigilance and the initiation of adaptive behavioral responses (Foote et al. 1983). Considerable data support the hypothesis that norepinephrine neurons in the LC constitute a CNS response or defense system, because the neurons are activated by “challenges” in both the behavioral/environmental and the physiological domains (Jacobs et al. 1991). Thus, various sensory stimuli are capable of increasing LC activity, but noxious or stressful stimuli are particularly potent in this regard. Moreover, considerable evidence also supports a role for LC norepinephrine neurons in the learning of aversively motivated tasks and in the conditioned response to stressful stimuli (Rasmussen and Jacobs 1986; Rasmussen et al. 1986), with obvious implications for a variety of psychiatric conditions (Gould et al. 2002; Szabo and Blier 2001). Indeed, tonic activation of the LC occurs preferentially in the response to stressful stimuli, in contrast to stimuli limited to simply evoking activation or arousal (Rasmussen and Jacobs 1986; Rasmussen et al. 1986). Preclinical studies have found increased sensitivity of the LC norepinephrine system in females compared with males, which may help explain the increased incidence of posttraumatic stress disorder (PTSD) and depression in human females (Bangasser et al. 2016). Of note, the LC also releases neuropeptides such as galanin, NPY, cocaine- and amphetamine-related transcript, and BDNF; galanin specifically co-localizes with norepinephrine in LC neurons as they project to the PFC, hippocampus, and VTA and mediate stress responses and addictive behaviors (Weinshenker and Holmes 2016). Increased norepinephrine and activation of the LC may be involved in both fear and anger responses, and excessive anger may be a “vent” for fear of the unexpected, as suggested in one review (Gu et al. 2016).

Norepinephrine Transporter

NET, the first of the monoamine transporters to be cloned in humans, transports norepinephrine from the synaptic cleft back into the neuron (Pacholczyk et al. 1991). Like other monoamine transporters, the NET comprises 12 putative transmembrane domains, and autoradiography with various norepinephrine reuptake inhibitors has been used to determine the brain distribution of NET. A high level of NET is found in the LC, with moderate to high levels found in the dentate gyrus, raphe nuclei, and hippocampus (Tejani-Butt and Ordway 1992; Tejani-Butt et al. 1990). This pattern of expression is consistent with the norepinephrine innervation to these structures. The NET is expressed mainly on norepinephrine terminals, as shown by a drastic reduction in labeling following norepinephrine destruction with the neurotoxin 6-hydroxydopamine or N-(2-chloroethyl)-N-ethyl-2-bromobenzylamine (DSP-4) (Tejani-Butt and Ordway 1992; Tejani-Butt et al. 1990).

The NET is dependent on extracellular Na⁺ to mediate norepinephrine reuptake and the effectiveness of norepinephrine reuptake inhibitors in inhibiting norepinephrine reuptake (Brüss et al. 1997, 1999; Harder and Bönisch 1985). The uptake of norepinephrine is chloride dependent, meaning that the electrogenic process of norepinephrine transport is Na⁺ and Cl^– driven (Harder and Bönisch 1985). In addition to the electrogenic process, the NET has properties of a channel-like pore, in that it can transport norepinephrine showing an infinite stoichiometry that can be blocked by cocaine and desipramine (Galli et al. 1995, 1996). Several studies suggest that the NET can be regulated by diverse stimuli, neuronal activity, and peptide hormones, as well as protein kinases. Indeed, studies have shown that all monoaminergic transporters (5-HTT, DAT, and NET) are rapidly regulated by direct or receptor-mediated activation of cellular kinases, particularly PKC (Bauman et al. 2000). PKC activation results in an activity-dependent transporter phosphorylation and sequestration. Protein phosphatase–1/2A (PP-1/PP-2A) inhibitors, such as okadaic acid and calyculin A, also promote monoaminergic transporter phosphorylation and functional downregulation (Bauman et al. 2000). These phenomena that occur beyond the receptor level may well be important in the long-term actions of psychotropic drugs known to regulate protein kinases (Chen et al. 1999; Manji and Lenox 1999). Given that norepinephrine neurons co-localize and release orexins, it is of interest that these neuropeptides have been implicated in sleep disorders and hypoglycemia through the glucose-sensing tandem membrane receptor K⁺ channels (K_2P type) affecting coordinated arousal (Scott et al. 2006). It is also interesting that enkephalin, an endogenous opioid peptide, co-localizes with NET on the axon terminals of the basolateral amygdala, as indicated by transmission electron microscopy and immunohistochemistry (Zhang and McDonald 2016). The latest PET radiotracer developed for human brain NET showed high uptake in the LC as well as the raphe nucleus, red nucleus, and thalamus, which could aid in monitoring disease progression of various conditions with NET dysregulation, including Parkinson’s disease, Alzheimer’s disease, epilepsy, ADHD, depression, and anxiety (Adhikarla et al. 2016). Atomoxetine, a selective NET inhibitor and the first nonstimulant medication approved for the treatment of ADHD in humans, was found to occupy both the NET and the SERT at clinically relevant doses when tested in rhesus monkeys (Ding et al. 2014).

Adrenergic Receptors

The α and β catecholamine receptors were first discovered more than 50 years ago (Ahlquist 1948) and were later subdivided further into α₁, α₂, β₁, β₂, and β₃ adrenoreceptors—all of which are GPCRs—on the basis of molecular cloning and pharmacological and biochemical studies (see Figure 2–5B). The crystal structures of these receptors were subsequently solved, along with those of other GPCRs, as reviewed elsewhere (Millar and Newton 2010).

α Receptors. The three subtypes of α₁ receptors are denoted 1A, 1B, and 1D; they are all positively coupled to PLC and possibly phospholipase A₂ (see Figure 2–5B). The α₂ family comprises the 2A/D, 2B, and 2C subtypes, which couple negatively to adenylyl cyclase and regulate K⁺ and Ca²⁺ channels (see Figure 2–5B). The 2A, 2B, and 2C adrenoceptors correspond to the human genes ADRA2A, ADRA2B, and ADRA2C, respectively. The bovine, guinea pig, rat, and mouse α_2D adrenoceptor is thought to be a species homologue or variant of the human α_2A adrenoceptor (Bylund et al. 1994) and is often referred to as α_2A/D. The α₂ receptors represent autoreceptors for norepinephrine neurons, and blockade of these autoreceptors results in increased norepinephrine release—a biochemical effect that has been postulated to play a role in the mechanisms of action of selected antidepressants (e.g., mianserin, mirtazapine) and antipsychotics (e.g., clozapine). In the LC, α₂-adrenergic receptors converge onto similar K⁺ channels as μ opioid receptors, and this convergence has been postulated to represent a mechanism for the efficacy of clonidine (an α₂ agonist) in attenuating some of the physical symptoms of opioid withdrawal. In addition, clonidine and another α₂ receptor agonist, guanfacine, have both been approved as nonstimulant medications to treat ADHD (Chan et al. 2016). The α₂ receptor antagonist yohimbine, which robustly increases norepinephrine neuron firing and norepinephrine release, has been used as a provocative challenge in clinical studies of anxiety disorders and as an antidepressant-potentiating agent. Yohimbine also has been used to explore the role of α₂ adrenoceptors in modulating different types of pain in healthy humans given high-frequency electrical stimulation (Vo and Drummond 2016).

β Receptors. The β receptor family comprises β₁, β₂, and β₃ adrenoceptors, which are all positively coupled to adenylyl cyclase (Bylund et al. 1994) (see Figure 2–5B). As is discussed in greater detail in the chapters on antidepressants (see Chapters 8–23), most effective antidepressants produce a downregulation/desensitization of β₁ receptors in rat forebrain, leading to the suggestion that these effects may play a role in their therapeutic efficacy. Interestingly, β receptors also have been shown to play a role in regulating emotional memories, leading to the proposal that β antagonists may have utility in the treatment of PTSD (Cahill et al. 1994; Przybyslawski et al. 1999). Propranolol, a combined β₁–β₂ receptor antagonist, has been used to try to prevent consolidation of traumatic memories in people who had just experienced trauma, such as rape victims or soldiers, but a meta-analysis suggested that it was not effective (Argolo et al. 2015). More recently, propranolol is being proposed to mitigate the development of intensive care unit–induced PTSD, which commonly stems from the trauma of hospitalization in an intensive care unit (Gardner and Griffiths 2014). β₃ receptors are not believed to be present in the CNS but are abundantly expressed on brown adipose tissue (BAT), where they exert lipolytic and thermogenic effects. Not surprisingly, active researchers are attempting to develop selective β₃ agonists for the treatment of obesity, and they are now called Class I BAT activators (Mukherjee et al. 2016).

Cholinergic System

Acetylcholine (ACh) is the only major low-molecular-weight neurotransmitter substance that is not derived from an amino acid (Kandel 2013). ACh is synthesized from acetyl coenzyme A and choline in nerve terminals via the enzyme choline acetyltransferase (ChAT). Choline is transported into the brain by uptake from the bloodstream and enters the neuron via both high-affinity and low-affinity transport processes (Cooper et al. 2003). In addition to the “standard” ChAT pathway, ACh can be synthesized by several possible mechanisms; the precise roles of these additional pathways and their physiological relevance in the CNS remain to be fully elucidated (Cooper et al. 2003). The highest activity of ChAT is observed in the interpeduncular nucleus, caudate nucleus, corneal epithelium, retina, and central spinal roots. In contrast to the other transmitters discussed thus far (which are most dependent on reuptake mechanisms), ACh has its signal terminated primarily by the enzyme ACh esterase, which degrades ACh (Figure 2–6B). Not surprisingly, therapeutic strategies to increase synaptic ACh levels (e.g., for the treatment of Alzheimer’s disease) have focused on inhibiting the activity of cholinesterases (Nestler et al. 2015).

FIGURE 2–6. The cholinergic system.

See Plate 9 to view this figure in color.

This figure depicts the cholinergic pathways in the brain (A) and various regulatory processes involved in cholinergic neurotransmission (B). Choline crosses the blood–brain barrier to enter the brain and is actively transported into cholinergic presynaptic terminals by an active uptake mechanism (requiring adenosine triphosphate [ATP]). This neurotransmitter is produced by a single enzymatic reaction in which acetyl coenzyme A (AcCoA) donates its acetyl group to choline by means of the enzyme choline acetyltransferase (ChAT). AcCoA is primarily synthesized in the mitochondria of neurons. Upon its formation, acetylcholine (ACh) is sequestered into secretory vesicles by vesicle ACh transporters (VATs), where it is stored. Vesamicol effectively blocks the transport of ACh into vesicles. An agent such as β-bungarotoxin or AF6₄A is capable of increasing synaptic concentration of ACh by acting as a releaser or a noncompetitive reuptake inhibitor, respectively. In turn, agents such as botulinum toxin are able to attenuate ACh release from nerve terminals. Once released from the presynaptic terminals, ACh can interact with a variety of presynaptic and postsynaptic receptors. In contrast to many other monoaminergic neurotransmitters, the ACh signal is terminated primarily by degradation by the enzyme acetylcholinesterase (AChE) rather than by reuptake. Interestingly, AChE is present on both presynaptic and postsynaptic membranes and can be inhibited by physostigmine (reversible) and soman (irreversible). Currently, AChE inhibitors such as donepezil and galantamine are the only classes of agents that are FDA approved for the treatment of Alzheimer’s disease. ACh receptors are of two types: muscarinic (G protein–coupled) and nicotinic (ionotropic). Presynaptic regulation of ACh neuron firing activity and release occurs through somatodendritic (not shown) and nerve terminal M2 autoreceptors, respectively. The binding of ACh to G protein–coupled muscarinic receptors that are negatively coupled to adenylyl cyclase (AC) or coupled to phosphoinositol hydrolysis produces a cascade of second-messenger and cellular effects (see diagram). ACh also activates ionotropic nicotinic acetylcholine (nACh) receptors. ACh has its action terminated in the synapse through rapid degradation by AChE, which liberates free choline to be taken back into the presynaptic neuron through choline transporters (CTs). Once inside the neuron, it can be reused for the synthesis of ACh, can be repackaged into vesicles for reuse, or undergoes enzymatic degradation. There are some relatively new agents that selectively antagonize the muscarinic receptors, such as CI-1017 for M₁, methoctramine for M₂, 4-DAMP for M₃, PD-102807 for M₄, and scopolamine (hardly a new agent) for M₅ (although it also has affinity for the M₃ receptor). Nicotine receptors (or nACh receptors) are activated by nicotine and the specific alpha(4)beta(2*) agonist metanicotine. Mecamylamine is an ACh receptor antagonist. cAMP=cyclic adenosine monophosphate; DAG=diacylglycerol; IP₃=inositol-1,4,5-triphosphate; PLC=phospholipase C.

Source. Adapted from Cooper JR, Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology, 8th Edition. New York, Oxford University Press, 2001. Copyright 1970, 1974, 1978, 1982, 1986, 1991, 1996, 2001 by Oxford University Press, Inc. Used by permission of Oxford University Press, Inc. Modified from Nestler et al. 2001.

In brief, cholinergic neurons can act as local circuit neurons (interneurons) and are found in the caudate, putamen, nucleus accumbens, olfactory tubercle, and islands of Calleja complex (Cooper et al. 2003). They also function as projection neurons that connect different brain regions; one fairly well-characterized pathway runs from the septum to the hippocampus (Figure 2–6A). The cholinergic complex is composed of eight cholinergic nuclei from which cholinergic neurons originate, designated Ch1 to Ch8. The basal forebrain cholinergic nuclei include the medial septal nucleus (Ch1), the vertical nucleus of the diagonal band (Ch2), the horizontal limb of the diagonal band (Ch3), and the nucleus basalis of Meynert (Ch4); brain stem cholinergic nuclei include the pedunculopontine nucleus (Ch5), the laterodorsal tegmental nucleus (Ch6), the medial habenula (Ch7), and the parabigeminal nucleus (Ch8) (Nestler et al. 2015). These nuclei project cholinergic neurons to the entire nonstriatal telencephalon, pontomesencephalotegmental cholinergic complex, thalamus, and other diencephalic loci (see Figure 2–6A). Descending cholinergic projections from these nuclei also innervate pontine and medullary reticular formations, deep cerebellar and vestibular nuclei, and cranial nerve nuclei (Cooper et al. 2003; Nestler et al. 2015). Basal forebrain cholinergic nuclei significantly degenerate in Alzheimer’s disease, which can be mitigated to some extent with cholinergic agents that increase synaptic ACh, such as donepezil; conversely, anticholinergic agents such as benztropine and trihexyphenidyl, through brain stem cholinergic interneurons, modulate striatal dopaminergic neurons and partially counteract motor symptoms of Parkinson’s disease (Nestler et al. 2015).

Cholinergic Receptors

The two major distinct classes of cholinergic receptors are the muscarinic and nicotinic receptors. Five muscarinic receptors (M₁ through M₅) have been cloned (Kandel 2013). These receptors are G protein–coupled and act either by regulating ion channels (in particular, K⁺ or Ca²⁺) or through being linked to second-messenger systems. Generally speaking, M₁, M₃, and M₅ are coupled to phosphoinositol hydrolysis, whereas M₂ and M₄ are coupled to inhibition of adenylyl cyclase and regulation of K⁺ and Ca²⁺ channels (Cooper et al. 2003) (see Figure 2–6B). M₁, M₃, and M₄ receptors are located in the cerebral cortex and hippocampus, with the M₁ and M₄ receptors concentrated in the striatal motor and reward circuits; the M₅ receptors are expressed widely at low levels; and the M₂ presynaptic autoreceptors (which decrease ACh release into the synapse) are located in the basal forebrain (Cannon et al. 2011; Nestler et al. 2015). Many older medications such as chlorpromazine (antipsychotic), benztropine (used for EPS), tricyclic antidepressants, and antihistamines also are antagonists of muscarinic receptors, and they cause common side effects such as dry mouth and constipation. One older muscarinic antagonist, scopolamine, commonly known to help with seasickness, more recently was found in humans to have rapid antidepressant properties, and in rats, a single dose reversed chronic unpredictable stress anhedonia, potentially through a burst increase in glutamate, mTORC1 (mechanistic target of rapamycin complex 1), and synaptic spines in the medial PFC (Drevets et al. 2013; Navarria et al. 2015). Furthermore, one study showed that a TT homozygous polymorphism rs324650 in the fifth intron of the presynaptic M₂ autoreceptor gene (CHRM2) resulted in more severe illness, including increased risk for suicide, that was potentially related to the observed reduced total volume of distribution (V_T) of M₂ receptors in the cingulate cortex of depressed bipolar patients, as measured by the high-affinity PET ligand M₂ receptor agonist [¹⁸F]FP-TZTP (Cannon et al. 2011). A conflicting postmortem observation detected no difference between bipolar patients and control subjects with the M₂/M₄ receptor antagonist [³H]AFDX (Smith and Jakobsen 2009). This negative result with [³H]AFDX, as compared with the positive result with [¹⁸F]FP-TZTP, was potentially explained by a difference in the binding properties of the two radioligands (Cannon et al. 2011).

By contrast, the nicotinic receptors are ionotropic receptors, and at least 12 different functional receptors (based on different subunit composition) have been identified (Nestler et al. 2015). Biochemical and biophysical data indicate that the nicotinic receptors in the muscle are formed from five protein subunits around a central pore, with the stoichiometry of α₂βγδ (Kandel 2013). The binding of ACh molecules on the α subunit is necessary for channel activation. By contrast, neuronal nicotinic receptors contain only two types of subunits (α and β), with the α occurring in at least eight different forms and the β in three (Cooper et al. 2003; Nestler et al. 2015). The most common nicotinic receptor subtypes are α4β2-nACh and α7-nACh (Horti et al. 2013). Nicotinic receptors are the targets of considerable cross-talk, as a variety of kinases (including PKA, PKC, and tyrosine kinases) are able to regulate the sensitivity of this receptor. Several regulatory mechanisms exist. For example, the mammalian prototoxin lynx1 acts as an allosteric modulator of the nicotinic receptor (Miwa et al. 2006). Curare (a poisonous full nicotinic receptor antagonist used in poison-tipped arrows and so forth) and succinylcholine (a weak partial nicotinic receptor agonist and routine surgical muscle relaxant) are two examples of compounds affecting nicotinic receptors (Nestler et al. 2015).

From a clinical standpoint, a long-standing observation is that patients with schizophrenia have a rate of nicotine use disorder that is substantially higher than that in the general population, in both the prevalence (~90%) and the quantity of cigarettes smoked (chain smoking). This observation has led to numerous studies of medication targeting nicotinic receptors to try to alleviate symptoms of psychosis, including the use of nicotine replacement, but the results have been disappointing in terms of treating psychosis even though the nicotine replacement maintains cognitive functioning (AhnAllen et al. 2015). One report (Freedman et al. 1997) determined that in a cohort of patients with schizophrenia, abnormal P50 auditory evoked potentials were linked to a susceptibility locus for this disease on chromosome 15. Notably, this is where a nicotinic receptor subunit is found, providing indirect genetic and phenotypic support for the long-standing contention that the high rates of cigarette smoking in patients with schizophrenia may represent some attempt by patients to self-medicate for their underlying nicotinic receptor defect. Varenicline, a partial agonist at the α4β2-nACh receptor subtype developed for smoking cessation, helped patients with schizophrenia quit smoking but did not help them with their psychotic symptoms or cognitive deficits (Smith et al. 2016).

Although the PET radioactive ligand [¹¹C]nicotine was one of the earliest human radioligands developed, it had problematic qualities as a PET ligand because of poor specificity and fast metabolism; the latest generation nicotinic ligand [¹⁸F]AZAN has some of the highest specific activity observed for the α4β2-nACh receptor, and it was shown to be only partially blocked in the human brain by nicotine gum or secondhand smoke but completely blocked by the specific agonist varenicline (Horti et al. 2013). Despite the well-publicized long-term negative health consequences of smoking, tobacco products that contain nicotine remain one of the most widely used addictive legal substances in the world.

Glutamatergic System

Glutamate and aspartate are the two major excitatory amino acids in the CNS and are present in high concentrations (Nestler et al. 2015; Squire 2013). As the principal mediators of excitatory synaptic transmission in the mammalian brain, they participate in wide-ranging aspects of both normal and abnormal CNS function. Physiologically, glutamate appears to play a prominent role in synaptic plasticity, learning, and memory. However, glutamate can also be a neuronal excitotoxin under a variety of experimental conditions, triggering either rapid or delayed neuronal death. Unlike the monoamines, which require transport of amino acids through the blood–brain barrier, glutamate and aspartate cannot adequately penetrate into the brain from the periphery and are produced locally by specialized brain machinery. The metabolic and synthetic enzymes responsible for the formation of these nonessential amino acids are located in glial cells and neurons (Squire 2013).

The major metabolic pathway in the production of glutamate is derived from glucose and the transamination of α-ketoglutarate; however, a small proportion of glutamate is formed directly from glutamine. The latter is actually synthesized in glia, via an active process (requiring adenosine triphosphate [ATP]), and is then transported to neurons, where glutaminase is able to convert this precursor to glutamate (Figure 2–7). Following release, the concentration of glutamate in the extracellular space is highly regulated and controlled, primarily by a Na⁺-dependent reuptake mechanism involving several transporter proteins.

FIGURE 2–7. The glutamatergic system.

See Plate 10 to view this figure in color.

This figure depicts the various regulatory processes involved in glutamatergic neurotransmission. The biosynthetic pathway for glutamate involves synthesis from glucose and the transamination of α-ketoglutarate; however, a small proportion of glutamate is formed more directly from glutamine by glutamine synthetase. The latter is actually synthesized in glia and, via an active process (requiring adenosine triphosphate [ATP]), is transported to neurons, where in the mitochondria glutaminase is able to convert this precursor to glutamate. Furthermore, in astrocytes glutamine can undergo oxidation to yield α-ketoglutarate, which can also be transported to neurons and participate in glutamate synthesis. Glutamate is either metabolized or sequestered and stored in secretory vesicles by vesicle glutamate transporters (VGluTs). Glutamate can then be released by a calcium-dependent excitotoxic process. Once released from the presynaptic terminal, glutamate is able to bind to numerous excitatory amino acid (EAA) receptors, including both ionotropic (e.g., NMDA [N-methyl-D-aspartate]) and metabotropic (mGlu) receptors. Presynaptic regulation of glutamate release occurs through metabotropic glutamate receptors (mGlu2 and mGlu3), which subserve the function of autoreceptors; however, these receptors are also located on the postsynaptic element. Glutamate has its action terminated in the synapse by reuptake mechanisms utilizing distinct glutamate transporters that exist on not only presynaptic nerve terminals but also astrocytes; indeed, current data suggest that astrocytic glutamate uptake may be more important for clearing excess glutamate, raising the possibility that astrocytic loss (as has been documented in mood disorders) may contribute to deleterious glutamate signaling, but more so by astrocytes. It is now known that a number of important intracellular proteins are able to alter the function of glutamate receptors (see diagram). Also, growth factors such as glial-derived neurotrophic factor (GDNF) and S100β secreted from glia have been demonstrated to exert a tremendous influence on glutamatergic neurons and synapse formation. Of note, serotonin_1A (5-HT_1A) receptors have been documented to be regulated by antidepressant agents; this receptor is also able to modulate the release of S100β. AKAP=A kinase anchoring protein; AMPA=α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid; CaMKII=Ca2+/calmodulin–dependent protein kinase II; ERK=extracellular response kinase; H-ras=Harvey rat sarcoma proto-oncogene; GKAP=guanylate kinase–associated protein; Glu=glutamate; Gly=glycine; GTg=glutamate transporter glial; GTn=glutamate transporter neuronal; Hsp70=heat shock protein 70; MEK=mitogen-activated protein kinase/ERK; mGluR=metabotropic glutamate receptor; MyoV=myosin V; NMDAR=NMDA receptor; nNOS=neuronal nitric oxide synthase; PKA=phosphokinase A; PKC=phosphokinase C; PP1, PP2A, PP2B=protein phosphatases; PSD-95=an abundant postsynaptic density (PSD) protein that forms a two-dimensional lattice immediately under the postsynaptic membrane; PTP1D=a protein tyrosine phosphatase; PYK2=protein tyrosine kinase 2; Rac1=Ras-related C3 botulinum toxin substrate 1; Raf=Raf-1 proto-oncogene, serine/threonine kinase; Rap2=related to AP2 domain protein; RSK=ribosomal S6 kinase; SHP2=src homology 2 (SH2) domain–containing tyrosine phosphatase; Src=SRC proto-oncogene, non–receptor tyrosine kinase; SynGAP=synaptic Ras-GTPase activating protein.

Source. Adapted from Cooper JR, Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology, 8th Edition. New York, Oxford University Press, 2001. Copyright 1970, 1974, 1978, 1982, 1986, 1991, 1996, 2001 by Oxford University Press, Inc. Used by permission of Oxford University Press, Inc. Table modified from Nestler et al. 2015.

The major glutamate transporter proteins found in the CNS, all of which clear the released glutamate from synapses, include the Na⁺-dependent excitatory amino acid transporters (EAATs): the “faster turnover” EAATs—EAAT1 (or GLAST), EAAT2 (or GLT-1), and EAAT3 (or EAAC1); and the “slower turnover” EAATs—EAAT4 and EAAT5 (Robinson and Jackson 2016). The corresponding human gene names are SLC1A3, 2, 1, 6, and 7, respectively (Robinson and Jackson 2016). Additionally, these transporters are differentially expressed in specific cell types, with EAAT1 and EAAT2 being found primarily in glial cells, responsible for most of the glutamate reuptake, and EAAT3 being localized on both excitatory and inhibitory neurons and oligodendrocytes (Robinson and Jackson 2016). EAAT2 is the most predominantly expressed form in the forebrain; EAAT4 is mainly localized in Purkinje cells of the cerebellum but also found elsewhere; EAAT5 is on the presynaptic termini of bipolar neuronal cells of the retina (Robinson and Jackson 2016; Squire 2013). Evidence indicates that phosphorylation of the transporters by protein kinases differentially regulates glutamate transporters and therefore glutamate reuptake (Casado et al. 1993; Conradt and Stoffel 1997; Pisano et al. 1996). Glutamate concentrations have been shown to rise to excitotoxic levels within minutes following traumatic or ischemic injury, and evidence indicates that the function of the glutamate transporters becomes impaired under these excitotoxic conditions (Faden et al. 1989). It is surprising that the glutamatergic system has only recently undergone extensive investigation with regard to its possible involvement in the pathophysiology of major mental illnesses, because it is the major excitatory neurotransmitter in the CNS and is known to play a role in regulating the threshold for excitation of most other neurotransmitter systems. After decades of research exploring the classic dopamine dysfunction hypothesis as the pathophysiological basis for schizophrenia, the glutamate dysfunction hypothesis became an equally vigorous research effort on the basis of initial observations of the psychotogenic properties of phencyclidine (PCP) and ketamine. Both are NMDA receptor antagonists that acutely increase glutamate levels in the synapses, causing many of the core features of schizophrenia (psychosis, thought disorder, negative symptoms, and executive cognitive deficits) to emerge from otherwise healthy human subjects (Moghaddam and Krystal 2012). However, the enthusiasm for the use of ketamine-induced psychosis as a model for schizophrenia has dampened because of the lack of efficacy of haloperidol (a D₂ receptor antagonist) and group II metabotropic glutamate receptor (mGlu2/3) agonists in ameliorating ketamine-induced psychosis and the lack of genetic and postmortem data to support NMDA receptor dysfunction in schizophrenia (Moghaddam and Krystal 2012).

It is now clear that modification of the levels of synaptic AMPA-type glutamate receptors—in particular, by receptor subunit trafficking, insertion, and internalization—is a critically important mechanism for regulating various forms of synaptic plasticity and behavior. Studies have identified region-specific alterations in expression levels of AMPA and NMDA glutamate receptor subunits in patients with mood disorders (Beneyto et al. 2007). Supporting the suggestion that abnormalities in glutamate signaling may be involved in mood pathophysiology, AMPA receptors have been shown to regulate affectivelike behaviors in rodents. AMPA receptor antagonists have been found to attenuate amphetamine- and cocaine-induced hyperactivity and psychostimulant-induced sensitization and hedonic behavior (Goodwin et al. 2007). Patients with treatment-resistant unipolar and bipolar depression given one intravenous dose of ketamine had strong relief of their depressive symptoms (Berman et al. 2000; Diazgranados et al. 2010; Zarate et al. 2006a). Ketamine antidepressant efficacy typically has been assumed to depend on direct glutamate NMDA receptor inhibition, which is how it works as an anesthetic (Leung and Baillie 1986). However, the results of human treatment trials indicate that alternative NMDA receptor antagonists lack the strong, rapid, or sustained antidepressant properties of ketamine (Newport et al. 2015). It has recently been shown that the metabolism of ketamine to one of its major metabolites, (2S,6S;2R,6R)-hydroxynorketamine, is essential for its antidepressant effects in mice (Zanos et al. 2016). These antidepressant actions are NMDA receptor inhibition independent, which requires activation of a different subtype of glutamate receptors, the AMPA receptors (Zanos et al. 2016).

Glutamatergic Receptors

The many subtypes of glutamatergic receptors in the CNS can be classified into two major subtypes: ionotropic and metabotropic receptors (see Figure 2–7).

Ionotropic glutamate receptors. The ionotropic glutamate receptor ion channels are assemblies of homo- or hetero-oligomeric subunits integrated into the neuron’s membrane. Every channel is assembled of (most likely) four subunits associated into a dimer of dimers, as has been observed in crystallographic studies (Ayalon and Stern-Bach 2001; Madden 2002; Nestler et al. 2015). Every subunit consists of an extracellular amino-terminal and ligand binding domain, three transmembrane domains, a reentrant pore loop (located between the first and the second transmembrane domains), and an intracellular carboxyl-terminal domain (Hollmann et al. 1994). The subunits associate through interactions between their amino-terminal domains, forming a dimer that undergoes a second dimerization mediated by interactions between the ligand binding domains and/or between the transmembrane domains (Ayalon and Stern-Bach 2001; Madden 2002). Three different subgroups of glutamatergic ion channels have been identified on the basis of their pharmacological ability to bind different synthetic ligands, each of which is composed of a different set of subunits. The three subgroups are the NMDA receptors, the AMPA receptors, and the kainate (KA) receptor. In the adult mammalian brain, NMDA and AMPA glutamatergic receptors are collocated in approximately 70% of the synapses (Bekkers and Stevens 1989). By contrast, at early stages of development, synapses are more likely to contain only NMDA receptors. Radioligand binding studies have shown that NMDA receptors and AMPA receptors are found at high densities in the cerebral cortex, hippocampus, striatum, septum, and amygdala.

NMDA receptors. The NMDA receptor is activated by glutamate and requires the presence of a co-agonist—namely, glycine or D-serine—to be activated, a process that likely varies in importance according to brain region (Panatier et al. 2006). However, the binding of both glutamate and glycine is still not sufficient for the NMDA receptor channel to open, because at resting membrane potential, the NMDA ion channel is blocked by Mg²⁺ ions. Only when the membrane is depolarized (e.g., by the activation of AMPA or KA receptors on the same postsynaptic neuron) is the Mg²⁺ blockade relieved. Under these conditions, the NMDA receptor channel will open and permit the entry of both Na⁺ and Ca²⁺ (see Figure 2–7).

The NMDA receptor channel is composed of a combination of GluN1, GluN2A, GluN2B, GluN2C, GluN2D, GluN3A, and GluN3B subunits (see Figure 2–7). The binding site for glutamate has been localized to the GluN2 subunit, and the site for the co-agonist glycine has been localized to the GluN1 subunit, which is required for receptor function. Two molecules of glutamate and two of glycine are thought to be necessary to activate the ion channel. Within the ion channel, two other sites have been identified—the sigma (σ) site and the PCP site. The hallucinogenic drugs PCP, ketamine, and the experimental drug dizocilpine (MK-801) all bind at the latter site and are considered noncompetitive receptor antagonists that inhibit NMDA receptor channel function.

In clinical psychiatric studies, ketamine has been shown to transiently induce psychotic symptoms in schizophrenic patients and to produce antidepressant effects in some depressed patients (Krystal et al. 2002). Building on these preclinical and preliminary clinical data, clinical trials have investigated the clinical effects of glutamatergic agents in patients with mood disorders. These and other data have led to the hypothesis that alterations in neural plasticity in critical limbic and reward circuits, mediated by AMPA and NMDA receptors, may represent a convergent mechanism for antidepressant action (Zarate et al. 2006b). Suicidal ideation also was rapidly improved with ketamine infusion (Ballard et al. 2015). This line of research holds considerable promise for developing new rapid-acting treatments for refractory MDD and refractory bipolar depression. NMDA receptors in the amygdala also may be of critical importance in the process of transforming a fixed and consolidated fear memory to labile states (Ben Mamou et al. 2006). The hope of cognitive improvements through NMDA modulation was dashed by a meta-analysis of 17 studies of NMDA receptor agonists glycine, D-serine, and D-cycloserine and other glutamate-enhancing agents that did not show improvement in cognition in patients with schizophrenia (Coyle et al. 2002; Iwata et al. 2015).

NMDA receptors play a critical role in regulating synaptic plasticity (Malenka and Nicoll 1999; Nestler et al. 2015). The best-studied forms of synaptic plasticity in the CNS are long-term potentiation (LTP) and long-term depression (LTD) of excitatory synaptic transmission. The molecular mechanisms of LTP and LTD have been extensively characterized and have been proposed to represent cellular models of learning and memory (Malenka and Nicoll 1999). Induction of LTP and LTD in the CA1 region of the hippocampus and in many regions of the brain has now clearly been shown to be dependent on NMDA receptor activation. During NMDA receptor–dependent synaptic plasticity, Ca²⁺ influx through NMDA receptors can activate a wide variety of kinases and/or phosphatases that modulate synaptic strength. An important development was the finding that two of the primary molecules involved—CaMKII and the NMDA subtype of glutamate receptor—form a tight complex with each other at the synapse (Lisman and McIntyre 2001). Interestingly, this binding appears to enhance both the autophosphorylation of the kinase and the ability of the entire holoenzyme, which has 12 subunits, to become hyperphosphorylated (Lisman and McIntyre 2001). This hyperphosphorylated state has been postulated to represent a “memory switch” that can lead to long-term strengthening of the synapse by multiple mechanisms. One important mechanism involves direct phosphorylation of the glutamate-activated AMPA receptors, which increases their conductance. Furthermore, CaMKII, once bound to the NMDA receptor, may organize additional anchoring sites for AMPA receptors at the synapse. Switching of synaptic NMDA receptor subunits, which bind CaMKII, for other NMDA receptor subunits having no affinity for this enzyme dramatically reduces LTP, highlighting glutamate and calcium signaling interactions critical for learning and memory (Barria and Malinow 2005).

Although the NMDA receptor clearly plays important roles in plasticity, abundant evidence has shown that excessive glutamatergic signaling is also involved in neuronal toxicity. With anoxia or hypoglycemia, the highly energy-dependent uptake mechanisms that keep glutamate compartmentalized in presynaptic terminals fail. Within minutes, glutamate is massively released into the synaptic space, resulting in activation of excitatory amino acid receptors. This leads to depolarization of target neurons via AMPA receptors and KA receptors and then to inappropriate and excessive activation of NMDA receptors (Farber et al. 2002). Considerable data suggest that the large excess of Ca²⁺ entering cells via the NMDA receptor channel may represent an important step in the rapid cell death that occurs via excitotoxicity (Nestler et al. 2015).

The hypothesis from the 1990s that NMDA receptor hypofunction is related to psychosis is supported by the observation of patients with autoimmune antibodies to the NMDA receptor (GluN1 subunit) who developed psychosis that reversed when the autoantibodies were removed by plasmapheresis and immunosuppression (Masdeu et al. 2016). Anti-NMDA receptor encephalitis was first described in young women who had psychosis and ovarian teratomas in 2005, but it has since been found in both men and women who have no known malignancy, and even with treatment, patients have a high recurrence rate of 12% 2 years after treatment (Pollak et al. 2014). In one review of the literature, 1.46% of 1,441 psychotic patients who were tested had the immunoglobulin G antibodies, and an increase in extracellular glutamate was reported in rats injected with CSF from affected patients (Manto et al. 2010; Pollak et al. 2014). Memantine, a noncompetitive NMDA receptor antagonist approved for the treatment of dementia, could be used as a possible treatment, especially in patients who proceed to develop catatonia or treatment-resistant schizophrenia (Pollak et al. 2014). Of note, patients with systemic lupus erythematosus also may develop antibodies to GluN2A and GluN2B subunits of the NMDA receptor treatment and have associated psychosis (Pollak et al. 2014). Encephalopathies have been found to be caused by an expanding list of autoimmune markers, including AMPA receptors, glycine receptors, and GABA_A and GABA_B receptors, all to be discussed later in this chapter (McKeon 2016).

In postmortem studies of patients with MDD who committed suicide, the PFC region had decreased levels of GluN2A and GluN2B subunits and the postsynaptic density protein-95 but no change in the level of the mandatory GluN1 subunits as compared with control subjects (Feyissa et al. 2009). Acamprosate (N-acetylhomotaurine) is an NMDA receptor antagonist approved for the treatment of alcohol use disorder, based on its success in decreasing cravings in clinical trials, but because it is short acting, patients must take this medication three times a day (Haass-Koffler et al. 2014).

AMPA receptors. The AMPA receptor is stimulated by the presence of glutamate and characteristically produces a fast excitatory synaptic signal that is responsible for the initial reaction to glutamate in the synapse. In fact, as discussed earlier, researchers generally believe that the activation of the AMPA receptor results in neuronal depolarization sufficient to liberate the Mg²⁺ cation from the NMDA receptor, thereby permitting its activation. The AMPA receptor channel is composed of the combination of the GluA1, GluA2, GluA3, and GluA4 subunits and requires two molecules of glutamate to be activated (see Figure 2–7). AMPA receptors have a lower affinity for glutamate than does the NMDA receptor, thereby allowing for more rapid dissociation of glutamate and, therefore, a rapid deactivation of the AMPA receptor (Henley and Wilkinson 2016).

Studies have indicated that AMPA receptor subunits are direct substrates of protein kinases and phosphatases. Phosphorylation of the receptor subunits regulates not only the intrinsic channel properties of the receptor but also the interaction of the receptor with associated proteins that modulate the membrane trafficking and synaptic targeting of the receptors (Malinow and Malenka 2002). Additionally, protein phosphorylation of other synaptic proteins has been proposed to indirectly modulate AMPA receptor function by affecting the macromolecular complexes that are important for the presence of AMPA receptors at the synaptic plasma membrane (Malinow and Malenka 2002; Nestler et al. 2015). Studies have been elucidating the cellular mechanisms by which AMPA receptor subunit insertion and trafficking occur and have identified two major mechanisms (Malinow and Malenka 2002; Nestler et al. 2015). The first mechanism is used for GluA1-containing AMPA receptor insertion and is regulated by activity. The second mechanism is governed by constitutive receptor recycling, mainly through GluA2/3 heteromers in response to activity-dependent signals. Data suggest that AMPA receptor subunit trafficking may play an important role in neuropsychiatric disorders and addictions. Thus, Nestler and associates have shown that the ability of drugs of abuse to elevate levels of the GluA1 subunit of AMPA glutamate receptors in the VTA of the midbrain is crucial for the development of sensitization (Carlezon and Nestler 2002). They have found that even transient increases in GluA1 levels within VTA neurons can trigger complex cascades of other molecular adaptations in these neurons and, within larger neural circuits, can cause enduring changes in the responses of the brain to drugs of abuse. Chronic lithium and valproate have been shown to reduce GluA1 expression in hippocampal synaptosomes, which may play a role in the delayed therapeutic effects of these agents (Du et al. 2003; Szabo et al. 2009).

Differential trafficking of AMPA receptor subunits as the sole, classic mechanism for regulation and induction of LTP was recently challenged when knockout mouse models demonstrated that LTP occurred even with no AMPA receptors but only KA receptors (Henley and Wilkinson 2016). One possible hypothesis that results from this observation would be that the subunits of the AMPA receptors do not dictate plasticity, but perhaps the numbers of “slot proteins” or “placeholder proteins” influence the development of LTP, and LTP is AMPA receptor subunit independent (Henley and Wilkinson 2016).

Investigators have sought to test the hypothesis that these “antidepressant anticonvulsants,” like traditional antidepressants, would enhance surface AMPA receptors (Du et al. 2007). It was found that lamotrigine and riluzole significantly enhanced the surface expression of GluA1 and GluA2 in a time- and dose-dependent manner in cultured hippocampal neurons. By contrast, the antimanic anticonvulsant valproate significantly reduced surface expression of GluA1 and GluA2. Concomitant with the GluA1 and GluA2 changes, the peak value of depolarized membrane potential evoked by AMPA was significantly higher in lamotrigine- and riluzole-treated neurons, supporting the surface receptor changes. In addition, lamotrigine and riluzole, as well as the traditional antidepressant imipramine, increased GluA1 phosphorylation at GluA1 (S845) in the hippocampus after chronic in vivo treatment.

Clinical studies have reported a consistent and rapid antidepressant effect of ketamine. Studies were therefore undertaken to test the hypothesis that ketamine brings about its rapid antidepressant effect via an AMPA receptor–dependent mechanism (Maeng et al. 2008). The AMPA receptor antagonist NBQX was without behavioral effects alone but blocked the antidepressant-like effects of ketamine. AMPA receptor antagonists also blocked the ketamine-induced changes in hippocampal GluA1 AMPA receptor phosphorylation (Maeng et al. 2008). It was also recently observed that these actions are observed with a metabolite of ketamine, (2R,6R)-hydroxynorketamine, which also causes strong and rapid potentiation of AMPA receptors (Zanos et al. 2016). Together, these results suggest that increased AMPA receptor activity in critical mood-relevant circuits may play an important role in antidepressant action.

Kainate receptors. The KA receptor has pre- and postsynaptic roles, sharing some properties with AMPA receptors. It is composed of the combination of the GluK1, GluK2, and GluK3 low-affinity subunits co-assembling with the GluK4 or GluK5 high-affinity subunits (formerly called the GluR5, GluR6, GluR7, KA1, and KA2 subunits, respectively) to form a dimer of dimers (tetrameric complex) (see Figure 2–7) (Møllerud et al. 2017). The crystal structures suggest that the pore remains closed even with glutamate bound to it, indicating that an additional mechanism is required to induce conformational change to open the pore (Møllerud et al. 2017). The precise role of KA receptors in the mature CNS is unknown, although the activity of the receptors clearly plays a role in synaptic function in many brain areas. Increasing data suggest the involvement of aberrant synaptic plasticity in the pathophysiology of bipolar disorder. KA receptors contribute to synaptic plasticity in different brain regions involved in mood regulation, including PFC, hippocampus, and amygdala. GluK2 (formerly called GluR6; the gene name continues to be GRIK2) is a subtype of KA receptor whose chromosomal loci of 6q16.3–q21 was identified as potentially harboring genetic polymorphism(s) contributing to the increased risk of mood disorders. The role of GluK2 in modulation of animal behaviors correlated with mood symptoms was investigated with GluK2 knockout mice and wild-type mice (Shaltiel et al. 2007). GluK2 knockout mice appeared to attain normal growth and lacked neurological abnormalities. The GluK2 knockout mice showed increased basal- or amphetamine-induced activity, were extremely aggressive, took more risks, and consumed more saccharin (a measure of hedonic drive). Notably, most of these aberrant behaviors responded to chronic lithium administration. These results suggest that abnormalities in KA receptor throughput generated by GluK2 gene disruption may lead to the concurrent appearance of a constellation of behaviors related to manic symptoms, including persistent hyperactivity, escalated irritability and aggression, risk taking, and hyperhedonia.

KA receptors play an important role in the hippocampus with place cell activity patterns and working memory (Sihra and Rodríguez-Moreno 2013). KA receptors modulate glutamate release between the mossy fibers of the granular cells of the dentate gyrus and the principal cells of the CA3 region, and there is high expression of KA receptors at those mossy fiber–CA3 synapses (Sihra and Rodríguez-Moreno 2013). KA receptors both prevent excessive glutamate release and facilitate increased glutamate release when the levels of glutamate are low in those same synapses (Sihra and Rodríguez-Moreno 2013). Recovery of the synapses with KA receptors is long and slow compared with the fast NMDA receptors, which may assist with memory formation (Sihra and Rodríguez-Moreno 2013). The CA3 may be more prone to develop seizures as a result of the KA receptors facilitating glutamate release (Sihra and Rodríguez-Moreno 2013).

Metabotropic glutamate receptors. The metabotropic glutamate (mGlu) receptors are GPCRs. The eight types of receptors that have been cloned can be organized into three different subgroups (groups I, II, and III) based largely on the signaling transduction pathways that they activate (see Figure 2–7). These receptors have a large extracellular N-terminal consisting of two lobes forming a “venus flytrap” binding pocket involved in glutamate recognition and a cysteine-rich extracellular domain that connects with seven transmembrane domains separated by short intra- and extracellular loops (see Figure 2–7). The intracellular loop plays an important role in the coupling with and selectivity of the G protein. The cytoplasmic carboxyl-terminal domain is variable in length and is involved with G protein activation and coupling efficacy (Bruno et al. 2001; Conn and Pin 1997; Nestler et al. 2015).

The mGlu receptor group I includes the mGlu1 (a, b, c, d) and mGlu5 (a, b) receptors (see Figure 2–7). They preferentially interact with the G_αq/11 subunit of G proteins, leading to activation of the IP₃/calcium and DAG/PKC cascades. The receptors are located on both pre- and postsynaptic neurons. Group II metabotropic receptors include mGlu2 and mGlu3, which have been best characterized as inhibiting adenylyl cyclase, but, like many receptors coupled to G_i/G_o, may also regulate ion channels. Group III receptors, which include mGlu4 (a, b), mGlu6, mGlu7 (a, b), and mGlu8 (a, b), are reported to produce inhibition of adenylyl cyclase as well but also to interact with the phosphodiesterase enzyme regulating guanosine monophosphate levels (Cooper et al. 2003; Squire 2013). The group II and III receptors are located in the presynaptic membrane and, because of their coupling with G_i/G_o proteins, appear to negatively modulate glutamate and GABA neurotransmission output when activated (i.e., they serve as inhibitory auto- and heteroreceptors). Preclinical studies suggest that mGlu group II and III receptors are “extrasynaptic” in their localization; that is, they are located some distance from the synaptic cleft and are thus activated only under conditions of excessive (pathological?) glutamate release, when glutamate is sufficient to diffuse out of the synapse to these receptors (Schoepp 2001). In preclinical studies, mGlu2/3 receptor agonists have been found to exert anxiolytic, antipsychotic, and neuroprotective properties (Schoepp 2001). However, subsequent clinical trials in humans have not sustained the promise of the preclinical studies (Moghaddam and Krystal 2012). mGlu receptors of all three groups appear to work as inhibitory autoreceptors, decreasing release of glutamate from the synapse, modulating glutamate release (Nestler et al. 2015).

Postmortem studies of the brains of patients with MDD who committed suicide reported several changes in expression of mGlu receptors. Patients with MDD have higher levels of RNA gene expression as measured by quantitative polymerase chain reaction for NMDA receptor subunits GluN2B and GluN2C, as well as for metabotropic receptors mGlu4 and mGlu5 in the neurons of the locus coeruleus, which together serve to stimulate noradrenergic neurons located there (Chandley et al. 2014). According to one study, Brodmann area 10 of the PFC of depressed and/or suicidal individuals had a 67% increase in (mostly presynaptic) mGlu2 and mGlu3 protein levels compared with healthy control subjects (Feyissa et al. 2010). Another study showed decreased mGlu5 protein expression in postmortem brain samples corresponding to decreased regional binding of a PET ligand to mGlu5 receptors in living depressed patients (Deschwanden et al. 2011). These observations suggest that there may be a relative glutamate “deficiency” in the synapse, leading to increased expression of glutamate receptors in depressed patients.

Glycine

Glycine is a nonessential amino acid that also functions as a neurotransmitter in the CNS, serving as a co-agonist for NMDA receptor activation. Glycine may be produced in the CNS by two distinct pathways. First, glycine is produced from serine by the enzyme serine-trans-hydroxymethylase in a reversible, folate-dependent reaction (Cooper et al. 2003; Nestler et al. 2015; Squire 2013). Second, a smaller proportion of glycine also may be produced from glyoxylate by the enzyme D-glycerate dehydrogenase. Glycine is found in higher concentrations in the spinal cord than in the rest of the CNS, and it acts as an inhibitory neurotransmitter predominantly in the brain stem and spinal cord (Nestler et al. 2015). As discussed earlier, a very important role that glycine also plays is to augment the NMDA-mediated frequency of NMDA receptor channel opening. This effect is strychnine-insensitive and pharmacologically suggests that the actions of glycine on NMDA receptor function are different from its effect on the spinal cord, where glycine’s inhibitory effect is blocked by strychnine (Cooper et al. 2003). The allosteric modulation of NMDA receptors via a glycine site is further underscored by receptor binding experiments yielding an anatomical distribution similar to that of NMDA receptors. Functionally, it has been postulated that glycine is able to augment the NMDA-mediated responses by speeding up the recovery process of the NMDA receptor (Cooper et al. 2003).

Glycine receptors, similar to GABA_A receptors, contain a chloride channel, and are composed of a combination of three α subunits containing the glycine binding site and two β subunits, which associate with gephyrin, a cytoplasmic protein (Nestler et al. 2015). Glycine receptors are bound by the compounds strychnine, a selective glycine receptor antagonist, and picrotoxin, a noncompetitive inhibitor, which block the chloride channel pore and cause seizures (Nestler et al. 2015). Synaptic glycine is removed mostly through the glycine transporter GlyT1. Ligands to GlyT1 are being explored as potential candidates for the treatment of schizophrenia, and a specific PET ligand, [¹⁸F]MK6577, recently has been developed (Xia et al. 2015). The endogenous ligands for the glycine receptor are actually D-serine and D-cycloserine (Labrie and Roder 2010). Glycine receptor agonism (D-serine) or GlyT antagonism (sarcosine) effectively increases glycine in the synapse, causing an increase in NMDA GluN1 receptor activation, which may ameliorate symptoms of schizophrenia, and clinical trials have suggested that these compounds may be helpful as augmentation agents together with other antipsychotics (Labrie and Roder 2010).

GABAergic System

GABA—the major inhibitory neurotransmitter system in the CNS—is one of the most abundant neurotransmitters, and GABA-containing neurons are located in virtually every area of the brain. Unlike the monoamines, GABA occurs in the brain in high concentrations on the order of micromoles per milligrams (about 1,000-fold higher than concentrations of monoamines) (Cooper et al. 2003; Nestler et al. 2015; Squire 2013). GABA is produced when glucose is converted to α-ketoglutarate, which is then transaminated to glutamate by GABA α-oxoglutarate transaminase (GABA-T). Glutamic acid is decarboxylated by glutamic acid decarboxylase, which leads to the formation of GABA (Figure 2–8). Indeed, the neurotransmitter and the rate-limiting enzyme are localized together in the brain and at approximately the same concentration. Catabolism of GABA occurs via GABA-T, which is also important in the synthesis of this transmitter.

FIGURE 2–8. The GABAergic system.

See Plate 11 to view this figure in color.

This figure depicts the various regulatory processes involved in GABAergic neurotransmission. The amino acid (and neurotransmitter) glutamate serves as the precursor for the biosynthesis of γ-aminobutyric acid (GABA). The rate-limiting enzyme for the process is glutamic acid decarboxylase (GAD), which utilizes pyridoxal phosphate as an important cofactor. Furthermore, agents such as L-glutamine-γ-hydrazide and allylglycine inhibit this enzyme and, thus, the production of GABA. Once released from the presynaptic terminal, GABA can interact with a variety of presynaptic and postsynaptic receptors. Presynaptic regulation of GABA neuron firing activity and release occurs through somatodendritic (not shown) and nerve-terminal GABA_B receptors, respectively. Baclofen is a GABA_B receptor agonist. The binding of GABA to ionotropic GABA_A receptors and metabotropic GABA_B receptors mediates the effects of this receptor. The GABA_B receptors are thought to mediate their actions by being coupled to Ca²⁺ or K⁺ channels via second-messenger systems. Many agents are able to modulate GABA_A receptor function. Benzodiazepines, such as diazepam, increase Cl^– permeability, and there are numerous available antagonists directed against this site. There is also a distinctive barbiturate binding site on GABA_A receptors, and many psychotropic agents are capable of influencing the function of this receptor (see blown-up diagram). GABA is taken back into presynaptic nerve endings by a high-affinity GABA uptake transporter (GABAT) similar to that of the monoamines. Once inside the neuron, GABA can be broken down by GABA transaminase (GABA-T), which is localized in the mitochondria; GABA that is not degraded is sequestered and stored in secretory vesicles by vesicular GABA transporters (VGATs), which differ from vesicular monoamine transporters (VMATs) in their bioenergetic dependence. The metabolic pathway that produces GABA, mostly from glucose, is referred to as the GABA shunt. The conversion of α-ketoglutarate into glutamate by the action of GABA-T and GAD catalyzes the decarboxylation of glutamic acid to produce GABA. GABA can undergo numerous transformations, of which the simplest is the reduction of succinic semialdehyde (SS) to γ-hydroxybutyrate (GHB). On the other hand, when SS is oxidized by succinic semialdehyde dehydrogenase (SSADH), the production of succinic acid (SA) occurs. GHB has received attention because it regulates narcoleptic episodes and may produce amnestic effects. The mood stabilizer and antiepileptic drug valproic acid is reported to inhibit SSADH and GABA-T. AC=adenylyl cyclase; TBPS=t-butylbicyclophosphorothionate.

Source. Adapted from Cooper JR, Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology, 8th Edition. New York, Oxford University Press, 2001. Copyright 1970, 1974, 1978, 1982, 1986, 1991, 1996, 2001 by Oxford University Press, Inc. Used by permission of Oxford University Press, Inc.

The contrasting functions of GABA-T, an enzyme able to synthesize both glutamate and the opposing GABA neurotransmitters as well as to catabolize GABA, become apparent when GABA-T is placed in the context of its role in the metabolic process. GABA-T converts GABA to succinic acid, and subsequent removal of the amino group yields α-ketoglutarate. Thus, α-ketoglutarate is able to be used by GABA-T in GABA biosynthesis as mentioned earlier (Cooper et al. 2003). This process, called the GABA shunt, maintains a steady GABA supply in the brain. As with the monoamines, the major mechanism by which the effects of GABA are terminated in the synaptic cleft is by reuptake through GABA transporters. The GABA transporters have a high affinity for GABA and mediate their reuptake via a Na⁺ and Cl^– gradient (Squire 2013).

Studies have measured the density and size of calbindin-immunoreactive neurons (presumed to be GABAergic) in layers II and III of the dorsolateral PFC, reporting a 43% reduction in the density of these neurons in patients with MDD compared with control subjects (Rajkowska 2002). Of particular note, in the rostral orbitofrontal cortex, there was a trend toward a negative correlation between the duration of depression and the size of neuronal cell bodies, suggesting changes associated with disease progression. Valproate also has been shown to have neurogenic effects in at least one study. In cultured embryonic rat cortical cells and striatal primordial stem cells, valproate markedly increased the number and percentage of primarily GABAergic neurons and increased neurite outgrowth (Laeng et al. 2004).

Low levels of GABA were observed in patients with MDD, particularly in the occipital and prefrontal cortex of living patients as well as in postmortem brain samples of people with MDD, some of whom committed suicide (Maciag et al. 2010). Specifically, the calbindin immunoreactive GABAergic neurons of the occipital lobe were decreased by 28% in depressed individuals as compared with control subjects, consistent with previous observations of differences in visual evoked potentials between depressed and healthy populations (Maciag et al. 2010). In addition, a PET study that used [¹¹C]flumazenil, a ligand for the benzodiazepine site of the GABA_A receptor, reported decreased GABA transmission in the medial temporal lobe of living antipsychotic-naïve patients with schizophrenia, consistent with postmortem studies showing lower GABA in the cortex (Frankle et al. 2015; Lewis et al. 2005).

GABA Receptors

The two major types of well-characterized GABA receptors are GABA_A and GABA_B, and most neurons in the CNS possess at least one of these types. The GABA_A receptor is the more prevalent of the two in the mammalian CNS, and as a result has been extensively studied and characterized. GABA_A contains an integral transmembrane chloride channel, which is opened on receptor activation, generally resulting in hyperpolarization of the neuron (i.e., suppressing excitability) (Nestler et al. 2015). The GABA receptor is a heteropentameric glycoprotein of approximately 275 kDa composed of a combination of multiple polypeptide subunits. GABA_A shows enormous heterogeneity, being composed of a combination of five classes of polypeptide subunits (α, β, γ, δ, ε), of which there are at least 18 total subtypes (Nestler et al. 2015). The various receptors have variation in functional pharmacology, hinting at the multiple finely tuned roles that inhibitory neurotransmission plays in brain function.

It is now well established that benzodiazepines function by binding to a potentiator site on the GABA_A receptor, increasing the amplitude and duration of inhibitory postsynaptic currents in response to GABA binding. Coexpression of additional γ subunits is believed to be necessary for the potentiation of GABA-mediated responses by benzodiazepines. In addition to benzodiazepines, barbiturates and ethanol are believed to exert many of their effects by potentiating the opening of the GABA_A receptor chloride channel (see Figure 2–8) (Nestler et al. 2015). As noted earlier, GABA_A receptors have a widespread distribution in the brain, and most of these receptors in the brain are targets of the currently available benzodiazepines. For this reason, there has been considerable interest in determining whether the desirable and undesirable effects of benzodiazepines can be differentiated on the basis of the presence of a different subunit composition. Much of the work has used gene knockout technology; thus, mutation of the benzodiazepine binding site of the α₁ subunit in mice blocks the sedative, anticonvulsive, and amnesic, but not the anxiolytic, effects of diazepam (Gould et al. 2003; Möhler et al. 2002). In contrast, the α₂ subunit (expressed highly in the cortex and hippocampus) is necessary for diazepam anxiolysis and myorelaxation. Thus, an α₂-selective ligand would provide effective acute treatment of anxiety disorders without the unfavorable side-effect profile of benzodiazepines. A compound with this preferential affinity for α₂ has been reported to exert fewer sedative and depressant effects than diazepam in rat behavioral studies (Gould et al. 2003; Möhler et al. 2002).

The phosphorylation of GABA_A receptors is another mechanism by which this receptor complex can be regulated in function and expression. In this context, it is noteworthy that studies have shown that knockout mice deficient in PKC ε isoforms show reduced anxiety and alcohol consumption and an enhanced response to effects of benzodiazepines (discussed in Gould et al. 2003). Furthermore, α/δ subunit assemblies are a novel neuronal GABA_A receptor subunit partnership present in hippocampal interneurons, which functions to mediate tonic inhibitory currents. Notably, this assembly results in a complex that is highly sensitive to low concentrations of ethanol (Glykys et al. 2007).

The GABA_B receptors are coupled to G_i and G_o and thereby regulate adenylyl cyclase activity (generally inhibit), K⁺ channels (open), and Ca²⁺ channels (close). GABA_B receptors can function as an autoreceptor but are also found abundantly postsynaptically on non-GABAergic neurons. Of interest, mounting evidence indicates that receptor dimerization may be required for the activation of GABA_B and possibly other GPCRs; although receptor dimerization has long been known to occur for growth factor and JAK (Janus tyrosine kinase)/STAT (signal transducers and activators of transcription) receptors, this was not expected for GPCRs. However, studies have reported that coexpression of two GABA_B receptor subunits—subunit 1 (GABA_B1) and subunit 2 (GABA_B2)—is necessary for the formation of a functional GABA_B receptor (Bouvier 2001). Data suggest that GABA_B2 receptor subunits may be necessary for proper protein folding of GABA_B1 receptor subunits (acting as a molecular chaperone) in the endoplasmic reticulum, but this remains to be definitively established. Support for the physiological relevance of this dimerization comes from studies showing that the GABA_B1 and GABA_B2 receptor subunits can be co-immunoprecipitated in rat cortical membrane preparations (Kaupmann et al. 1997); thus, the dimerization is not simply an in vitro phenomenon.

Relatively new GABAergic agents include zolpidem, a short-acting, positive allosteric modulator of the benzodiazepine site of the GABA_A receptor that is used for initial insomnia; gaboxadol, a δ-subunit-selective extrasynaptic GABA_A receptor agonist; and tiagabine, a GABA reuptake inhibitor (blocks the GABA membrane transporter GAT1) used as an anticonvulsant (Nutt et al. 2015). Alcohol and benzodiazepines both notably bind GABA_A receptors and are cross-reactive with each other, to the point that benzodiazepines are used routinely to detoxify patients from alcohol intoxication while simultaneously avoiding precipitating seizures. Baclofen is a GABA_B receptor agonist used to treat muscle spasms. Methaqualone, a sedative GABA_A receptor agonist widely prescribed in the 1960s as a safe alternative to barbiturates, was taken off the market when it was found to be highly addictive and was abused in combination with alcohol (Hammer et al. 2015). Propofol, an anesthetic agent that has been in the news for abuse potential and accidental overdoses among famous celebrities, also binds the GABA_A receptor in a manner similar to barbiturates and methaqualone, which all are highly addictive and keep the chloride channel open longer (Hammer et al. 2015).

Purinergic Neurotransmission: Focus on Adenosine

It has been known for quite some time that ATP is capable of exerting profound effects on the nervous system (Drury and Szent-Györgyi 1929). However, adenosine and adenosine nucleotides have become more widely accepted as neuroactive substances in the CNS (Cooper et al. 2003). Adenosine is released from neurons and glia, but many of the neurotransmitter criteria outlined in the beginning of this chapter are not met. Nonetheless, adenosine is able to activate many cellular functions that can produce changes in neuronal and behavioral states. For instance, adenosine stimulates cAMP in vitro in brain slices, and caffeine (which in addition to being a phosphodiesterase inhibitor is a well-known adenosine receptor antagonist) is able to block this response.

In the P1 (for purine) adenosine receptor class, four adenosine receptors have been cloned (A₁, A_2A, A_2B, and A₃), each of which has a unique tissue distribution, ligand binding affinity (nanomolar range), and signal transduction mechanisms (Cooper et al. 2003). Data suggest that the high-affinity adenosine receptors (A₁ and A_2A) may be activated under normal physiological conditions, whereas in pathological states such as hypoxia and inflammation (in which high adenosine concentrations [micromolar range] are present), low-affinity A_2B and A₃ receptors are also activated. A_2B receptors are expressed in low levels in the brain but are ubiquitous in the rest of the body, whereas A_2A receptors are found in high concentrations in areas of the brain that receive dopaminergic projections (i.e., striatum, nucleus accumbens, and olfactory tubercle) (Nestler et al. 2015). Given this receptor’s distribution, and the inverse relation between dopamine and adenosine, it has been postulated that A_2A antagonists may have some utility in the treatment of Parkinson’s disease (Nestler et al. 2015). The mood stabilizer and antiepileptic drug carbamazepine, which primarily works through blocking voltage-gated sodium channels, also acts as an antagonist of the A₁ subtype, which is epileptogenic in some vulnerable populations, such as children (Booker et al. 2015; Gould et al. 2002).

Adenosine is also able to alter the function (both pre- and postsynaptically) of numerous neurotransmitters and their receptors, including NMDA, mGlu receptors, ionotropic nicotinic receptors, norepinephrine, serotonin, dopamine, GABA, and various peptidergic receptors. Adenosine is widely regarded as a major component that regulates homeostasis of blood flow and metabolic demands in peripheral tissue physiology. Evidence suggests that adenosine is implicated as a fatigue factor to decrease cholinergic activity–arousal via presynaptic inhibition of glutamate release (Brambilla et al. 2005). In addition, the ionotropic ligand-gated P2X trimeric receptor class (containing seven subtypes: P2X₁ through P2X₇) and the metabotropic GPCR P2Y receptor class (containing subtypes P2Y₁, P2Y₂, P2Y₄, P2Y₆, P2Y₁₁, P2Y₁₂, P2Y₁₃, and P2Y₁₄) are purine receptors that can be activated by ATP (Nestler et al. 2015; Ortiz et al. 2015). It has been found that ATP is released from astrocytes (through an unknown mechanism) and that the release is accompanied by glutamate release (Ca²⁺-dependent) (Innocenti et al. 2000). However, more data suggest adenosine (which is derived from ATP) may serve as the true ligand for these purinergic receptors (Fields and Stevens-Graham 2002). The ATP/adenosine is then able to activate purine receptors (P2Y receptors) on neighboring astrocytes, and this stimulates Ca²⁺ influx and subsequent release of glutamate and ATP to then affect other astrocytes and neurons. This may be a critical component in the communication process between glial cells, as well as representing a signaling molecule from glia to neurons (Fields and Stevens-Graham 2002).

Caffeine in coffee, theophylline in tea, and theobromine in cocoa are all methylxanthine compounds that at moderate doses cause increased alertness through antagonizing the A₁—and, to a lesser extent, the A_2A—receptors, but in susceptible people at higher doses, they cause anxiety and even panic attacks (Nestler et al. 2015). Adenosine A₁ agonists may have neuroprotective effects in stroke both by inhibiting glutamate release (excitotoxicity) presynaptically and by inhibiting postsynaptic membrane depolarization and calcium influx (exacerbating excitotoxicity) (Nestler et al. 2015). Because the striatum has high levels of A_2A receptors, A_2A receptor antagonists are being explored for the treatment of Parkinson’s disease–related L-dopa-induced hyperkinesia due to the inverse relation between adenosine and dopamine in the striatum. P2X₃ and P2X_2/3 receptors are being explored for pain modulation (Nestler et al. 2015). The literature is conflicted as to whether adenosine analogues cause depression or treat depression, but low brain purine levels have been reported in depressed female patients (Krügel 2016). Evidence has also increased in the literature that schizophrenia may be a hypoadenosinergic state, such as reduced A_2A receptors in postmortem brain samples (Krügel 2016). Numerous adenosine PET ligands to adenosine receptors have been developed (Mishina and Ishiwata 2014).

Peptidergic Neurotransmission

Neuropeptides have garnered increasing attention as critical modulators of CNS function. In general, peptide transmitters are released from neurons when they are stimulated at higher frequencies than those required to facilitate release of traditional neurotransmitters but can also be co-localized and co-released together with other neurotransmitters (Cooper et al. 2003; Nestler et al. 2015). Modulation of the firing rate pattern of neurons and subsequent release of neurotransmitters and peptides in a circumscribed fashion are likely important in the basal functioning of the brain, as well as response to specific stimuli. Interestingly, cannabinoids, an example of a neuropeptide neurotransmitter, do not alter firing rates of hippocampal neurons but change temporal coordination, an effect that correlates with memory deficits in individuals (Soltesz and Staley 2006). Localization of brain stem cannabinoid receptors (CB₁ and CB₂) provided clues to how cannabinoids regulate brain function (Van Sickle et al. 2005). Virtually every known mammalian bioactive peptide is synthesized first as a precursor protein in which product peptides are flanked by cleavage sites. Neuropeptides are generally found in large dense-core vesicles, whereas other neurotransmitters, such as the monoamines, are packaged in small synaptic vesicles (approximately 50 nm) and are usually half the size of their peptidergic counterparts (Kandel 2013; Squire 2013).

Space limitations preclude an extensive discussion of the diverse array of neuropeptides known to exist in the mammalian brain. Table 2–2 highlights some of the major neuropeptides that may be of particular psychiatric relevance. In the remainder of this section, we highlight the basic aspects of peptidergic transmission vis-à-vis an overview of opioidergic neurotransmission. We briefly discuss selected neuropeptides here; a general review of neuropeptides as potential drug targets can be found elsewhere (Hoyer and Bartfai 2012).

Corticotropin-Releasing Factor

Hyperactivity of the hypothalamic-pituitary-adrenal (HPA) axis may relate to the pathophysiology of mental illness. Corticotropin-releasing factor (CRF) is a key neuropeptide controlling the HPA axis with a 41–amino acid sequence that is identical in humans and in rats (Montecucchi et al. 1980). CRF is produced in the paraventricular nucleus (PVN) of the hypothalamus, and once secreted into pituitary portal blood, it binds to G_s-coupled receptors on cells in the pituitary to increase adenylate cyclase, cAMP-dependent protein kinase, and cytosolic calcium concentrations (Lovejoy et al. 2014). CRF is also capable of activating G_q-, G_i-, G_o-, and G_z-coupled receptors (Grammatopoulos et al. 2001). CRF receptors are classified into CRF1 and CRF2 (Chen et al. 1993; Liaw et al. 1996; Lovejoy et al. 2014). The CRF1A receptor is the dominant subtype in the brain and peripheral tissue, whereas the CRF1D receptor is thought to play a competing role against CRF1A, with CRF1B and CRF1C also being identified (Hillhouse and Grammatopoulos 2006). Although CRF2 receptors and their spliced variants CRF2A, CRF2B, and CRF2C are localized in the brain, they exist to a greater degree in peripheral tissue (Hillhouse and Grammatopoulos 2006).

TABLE 2–2. Selected peptides and their presumed relevance to psychiatric disorders and treatment

Group	Potential clinical relevance
Opioid and related peptides
Endorphin	Analgesia for chronic pain
Enkephalin	Analgesia
Dynorphin	Analgesia
Nociceptin	Binds ORL1 receptor; increased pain perception
Gut-derived peptides
Vasoactive intestinal peptide	Sexual behavior
Cholecystokinin	Anxiety/panic
Secretin	Brief reports of help with autism and transient antipsychotic properties
Somatostatin	Mood disorders and treatment
Tachykinin peptides
Substance P (substance K)	Receptor is NK1, no efficacy shown for depression; inflammation
Neurotensin/neuromedin N	Analgesia, hypothermia; regulated by lithium, involved in dopamine signaling
Pituitary peptides
Oxytocin	Affiliative, prosocial, bonding behavior; decreases fear; may help increase sociability in autism
Vasopressin	Also called antidiuretic hormone; receptors are V1a, V1b, V2; antagonists have anxiolytic, anti-aggression, and anti-irritability properties
Adrenocorticotropic hormone	Dysregulated in mood disorders
Melanocyte-stimulating hormone	Antidepressant and anti-anxiety properties
Hypothalamic-releasing factors
Corticotropin-releasing factor	Anxiety and fear
Thyrotropin-releasing factor	Potential antidepressant effects
Others
Calcitonin gene–related peptide	Regulated by ECT and lithium
Angiotensin	Mood disorders, bipolar disorder
Leptin	Satiety signal; inhibits feeding drive
Cocaine- and amphetamine-related transcript	Drug addiction, eating disorders
Galanin	Potentially relevant for Alzheimer’s diagnosis and other cognitive disorders
Neuropeptide Y	Potential endogenous anxiolytic; regulated by antidepressants/lithium; reduced by early maternal separation; antagonists may decrease appetite
Orexin/hypocretin	Narcolepsy and other sleep abnormalities; eating disorders
Note. This table summarizes selected peptides and their presumed relevance for psychiatric disorders and their treatment; it is not meant to be an exhaustive listing of findings. It should also be noted that in some cases—for example, CRF (mood/anxiety), NPY and neurotensin (regulation by medications), oxytocin (affiliative behavior), and orexin (narcolepsy)—the data are quite convincing. In many of the other examples noted, the evidence must be considered preliminary but is, in our opinion, quite noteworthy and warrants further investigation. A discussion of these peptides is beyond the scope of this introductory chapter; nevertheless, readers are encouraged to explore the latest research in this rapidly evolving and exciting literature. ECT=electroconvulsive therapy.

Stressful life events and chronic stress can induce brain circuit changes that are thought to relate to the pathophysiology of some anxiety disorders and PTSD. Increased CRF secretion from the hypothalamus and adrenocorticotropic hormone (ACTH) from the anterior pituitary gland are associated with aberrant cortisol responses to stress. Depressed patients and suicide victims have elevations of CRF concentrations in the CSF, increased cerebrocortical CRF immunoreactivity, and decreased CRF1 receptor binding; furthermore, decreased CRF1 mRNA expression has been reported in postmortem brain tissue of suicide victims (Sanders and Nemeroff 2016). Although probes of HPA function, such as the dexamethasone/CRF suppression tests, indicate hyperactivity of the axis in depressed patients, it is largely a nonspecific finding and can occur in other psychiatric illnesses.

Oxytocin and Vasopressin

Other brain-related peptides of interest to psychiatric illnesses that can be co-secreted with CRF from the hypothalamus are oxytocin and vasopressin, but treatments in psychiatry directed at these receptors have not yet received clinical use. Oxytocin and vasopressin are both 9–amino acid peptide hormones. Oxytocin has a 6–amino acid ring and a 3–amino acid tail with the ability to exert disulfide bonds that may relate to its neurophysiological mode of action. PVN and supraoptic nuclei of the hypothalamus are known to contain high levels of oxytocin, whereas vasopressin is localized in other cells in the hypothalamus; however, both can extend cellular processes into the posterior pituitary. These neuropeptides do not function as traditional neurotransmitters, but rather permeate through neural tissue by volume transmission (Neumann and Landgraf 2012). It is noteworthy that oxytocin may modulate the amygdala and brain stem neurons to facilitate fear avoidance, whereas vasopressin may exert a role in PTSD (Carter 1998; Wentworth et al. 2013). Furthermore, oxytocin and vasopressin bind to distinct brain receptors that can be modulated by epigenetics (Ebstein et al. 2012). The oxytocin receptor is a GPCR located on 3p24–26, whereas three receptor subtypes have been identified for vasopressin (V1a, V1b, V2), with V1a and V1b currently associated with modulation of behaviors (Gimpl and Fahrenholz 2001).

Opiates

Opioids are a family of peptides that occur endogenously in the brain (endorphins), as botanicals, or as drugs. POMC is a precursor protein characterized in the 1980s that gives rise to ACTH and a class of endogenous opiates called endorphins. POMC, proenkephalin-derived peptides, and prodynorphin-derived peptides yield opioid peptides on cleavage. Three opioid peptide families exist: enkephalins, endorphins, and dynorphins. POMC gene expression occurs in various areas of the brain and in other tissues. POMC has tissue- and cell-specific regulatory factors at every step from gene transcription to its posttranslational processing. Opioid peptides are stored in large dense-core vesicles and are co-released from neurons that usually contain a classic neurotransmitter agent (e.g., glutamate and norepinephrine). Opiorphin, an endogenously derived enkephalin that inactivates zinc ectopeptidase, has been described as equal to morphine in the perception of pain (Wisner et al. 2006). Although opiates are widely associated with and used therapeutically in pain modulation, evidence indicates that dynorphin can actually activate bradykinin receptors and contribute to neuropathic pain (Altier and Zamponi 2006).

Opioids activate a variety of signal transduction processes, and different mechanisms in their regulation are in place for different cell types. The opioid receptors are GPCRs and exert their cellular effects by inhibiting adenylyl cyclase and regulating K⁺ and Ca²⁺ channels, via activation of G_i/G_o. There are three types of opioid receptors—μ, δ, and κ, each of which is further subclassified—in addition to opioid receptor like–1 (ORL1) receptor. These receptors are 7-transmembrane-spanning proteins that couple to inhibitory G-proteins or form homo- and heterodimeric complexes. They also alter calcium signaling through dissociation of G_βγ subunits and by reducing sensitivity to L-type, N-type, and P/Q-type channels. Also, numerous mechanisms have been described that allow opiates and synthetic opiate agents (i.e., morphine, fentanyl) to regulate receptor signaling, which can occur from the receptor being phosphorylated, desensitized, and internalized. Once the receptor is phosphorylated, recruitment of arrestins to the receptor occurs and can prime for sequestration. Interestingly, arrestin-3–deficient mice have tolerance to morphine and other μ opioid receptor agents, whereas polymorphisms in OPRM1 have been associated with opioid dependency. Intracellular cascades associated with opioid dependence and withdrawal have documented changes in mitogen-activated protein kinase (MAPK) and extracellular signal-regulated kinase 1/2 (ERK1/2) cascades, as well as changes in transcription factors such as phosphorylated cAMP response element-binding protein (pCREB) and DeltaFosB, which have been linked to changes in the reward system (Al-Hasani and Bruchas 2011).

The continued study of the opioid system and the second-messenger changes brought about by the chronic administration of opioids has greatly facilitated our understanding of the molecular and cellular effects of drugs of abuse and the potential to develop novel therapeutics (Nestler et al. 2015). In response to the worldwide heroin epidemic, the medication buprenorphine, a partial agonist of the μ opioid receptor and an antagonist of the δ and κ opioid receptors, has become one of the most widely prescribed medications in the world to treat opioid use disorders, perhaps because it can be prescribed from an outpatient office setting.

Conclusion

We have provided an overview of some fundamental aspects of neurotransmitters and brain receptor classes. For most psychiatrists, molecular and cellular biology have not traditionally played a major role in day-to-day clinical practice. However, new insights into the molecular and cellular basis of disease and drug action are being generated at an ever-increasing rate and will ultimately result in a transformation of our understanding and management of diseases. The “molecular medicine revolution” has used the power of sophisticated cellular and molecular biological methodologies to tackle many of society’s most devastating illnesses. The rate of progress has been exciting indeed, and hundreds of GPCRs and their effectors have now been identified and characterized at the molecular and cellular levels. These efforts have allowed the study of a variety of human diseases that are caused by abnormalities in cell-to-cell communication. Studies of such diseases are offering unique insights into the physiological and pathophysiological functioning of many cellular transmembrane signaling pathways.

Psychiatry, like much of the rest of medicine, has entered a new and exciting age demarcated by the rapid advances and the promise of molecular and cellular biology and neuroimaging. There is a growing appreciation that severe psychiatric disorders arise from abnormalities in cellular plasticity cascades, leading to aberrant information processing in synapses and circuits mediating affective, cognitive, motoric, and neurovegetative functions. Thus, these illnesses can be best conceptualized as genetically influenced disorders of synapses and circuits rather than simply as deficits or excesses in individual neurotransmitters. Furthermore, many of these pathways play critical roles not only in synaptic and behavioral plasticity but also in long-term atrophic processes. Targeting these pathways in treatment may stabilize the underlying disease process by reducing the frequency and severity of the profound mood cycling that contributes to morbidity and mortality.