Discover millions of ebooks, audiobooks, and so much more with a free trial

Only $11.99/month after trial. Cancel anytime.

Gap Junctions in the Brain: Physiological and Pathological Roles
Gap Junctions in the Brain: Physiological and Pathological Roles
Gap Junctions in the Brain: Physiological and Pathological Roles
Ebook943 pages26 hours

Gap Junctions in the Brain: Physiological and Pathological Roles

Rating: 0 out of 5 stars

()

Read preview

About this ebook

Gap junctions between glial cells or neurons are ubiquitously expressed in the mammalian brain and play a role in brain development including cell differentiation, cell migration and survival, and tissue homeostasis, as well as in human diseases including hearing loss, neuropathies, epilepsy, brain trauma, and cardiovascular disease. This volume provides neuroscience researchers and students with a single source for information covering the physiological, behavioral and pathophysiological roles of gap junctions in the brain. In addition, the book also discusses human disease conditions associated with mutations in single gap junction connexion genes, making it applicable to clinicians doing translational research. Finally, it includes reviews of pharmacological studies with gap junction blockers and openers, summarizing information obtained from phenotyping gap junctions mouse mutants.

  • Serves as the most current and comprehensive reference available covering the physiological, behavioral and pathophysiological roles of gap junctions in the brain
  • Chapters summarize knowledge of the basic physiology of gap junctions in the brain, as well as of human disease conditions associated with mutations in single gap junction connexin genes
  • Includes reviews of pharmacological studies with gap junction blockers and openers, summarizing information obtained from phenotyping gap junctions mouse mutants
LanguageEnglish
Release dateDec 12, 2012
ISBN9780124159273
Gap Junctions in the Brain: Physiological and Pathological Roles

Related to Gap Junctions in the Brain

Related ebooks

Biology For You

View More

Related articles

Related categories

Reviews for Gap Junctions in the Brain

Rating: 0 out of 5 stars
0 ratings

0 ratings0 reviews

What did you think?

Tap to rate

Review must be at least 10 words

    Book preview

    Gap Junctions in the Brain - Ekrem Dere

    Germany

    PART I

    The Physiology of Gap Junctions in the Brain

    Chapter 1 Gap Junctions in the Brain

    Chapter 2 Physiology and Function of Glial Gap Junctions in the Hippocampus

    Chapter 1

    Gap Junctions in the Brain

    Armin Zlomuzica∗, Sonja Binder† and Ekrem Dere∗∗

    ∗Mental Health Research and Treatment Center, Ruhr-University Bochum, Germany

    †Department of Neuroendocrinology, University of Lübeck, Germany

    ∗∗Neurobiologie des Processus Adaptatifs, Université Pierre et Marie Curie, Paris, France

    Outline

    Introduction

    Structural Aspects of Gap Junctions

    Types of Gap Junction

    Cellular and Brain Regional Expression Patterns of Specific Connexins

    Connexin30

    Connexin30.3

    Connexin31.1

    Connexin32

    Connexin36

    Connexin43

    Connexin45

    Gap Junction Pharmacology

    Synchronization and Rhythmic Oscillation of Neural Activity

    Theta and Gamma Oscillations

    High-Frequency Oscillations

    Connexin45 and Neuronal Network Oscillations

    Connexin36 and Olivocerebellar Network Oscillations

    Gap Junction-Related Human Diseases

    Conclusion

    Acknowledgments

    Introduction

    Gap junctions in the brain of mammals including rats and non-human primates were first detected by electron microscopy. Gap junctions were observed in the neocortex, between non-pyramidal cell dendrites in the CA1 and CA3 subregion of the rat hippocampus (Kosaka, 1983), between dendrites and somata of neurons in the sensorimotor cortex of non-human primates (Sloper and Powell, 1978), and in the brainstem and cerebellum (Sotelo and Korn, 1978). Gap junctions allowing intercellular electrotonic and metabolic coupling between adjoining cells are expressed in organs and tissues that depend on rapid intercellular signal transfer and communication, including the heart, skin, inner ear and brain. They play an important role in brain development and maturation and have been implicated in neural stem and progenitor cell proliferation, as well as in cell migration and differentiation (Bruzzone and Dermietzel, 2006). Gap junctions are intercellular channels with a diameter of 1.2 nm forming an aqueous pore penetrating the lipid bilayer of two neighboring cells. These channels are composed of integral transmembrane proteins named connexins. Recently, two more classes of proteins, the pannexins and innexins, have been identified as molecular subunits of gap junction channels in the phylum chordata and in invertebrates, respectively (see Zoidl et al., 2008, for review). Pannexins are large-pore ion channels that are widely expressed within the brain. In contrast to connexins, pannexins may not form gap junction channels but rather form hemichannels located on neurons and astrocytes releasing arachidonic acid derivatives, adenosine triphosphate (ATP) or neurotransmitters into the extracellular space (Stout et al., 2004). Pannexin hemichannels have been implicated in pathological conditions such as ischemia, excitotoxic and ATP-dependent neuronal cell death and brain inflammation (MacVicar and Thompson, 2010).

    A gap junction channel consists of two hemichannels (connexons), which are contributed by two neighboring cells. Each connexon is composed of six connexin proteins (Kumar and Gilula, 1996) (Figure 1.1). Gap junction plaques (consisting of tightly clustered channels) can be visualized ultrastructurally by means of freeze-cleaved replicas (Goodenough and Revel, 1970).

    FIGURE 1.1 Each connexon is composed of six connexion proteins. Top left: homomeric connexon that are composed of a single connexin. Top right: heteromeric connexon that is composed of two different connexins. Bottom: gap junctions are formed by two opposed connexons that can vary in terms of connexin composition ranging from bi-homomeric and homotypic to bi-tereromeric and heterotypic.

    This figure is reproduced in color in the color plate section.

    Gap junction channels are 12–16 Å wide and permit the intercellular bidirectional diffusion of nutrients, ions, metabolites, second messengers, such as potassium (K+), calcium (Ca²+), cyclic adenosine monophosphate (cAMP), inositol 1,4,5-trisphosphate (IP3), cyclic guanosine monophosphate (cGMP), glucose, and other small molecules of about 1000 Da or less than 16 Å in diameter (Dobrowolski and Willecke, 2009; Zoidl and Dermietzel, 2010).

    There are also hemichannels that remain unpaired after being transported to the membrane, until they align with another hemichannel located on an adjacent cell to form a gap junction channel. These hemichannels support autocrine and paracrine actions (Spray et al., 2006) and have important functions. For example, hemichannels located on astrocytes are involved in the control of glutamate homeostasis and can release glutathione, glutamate, taurine and ATP into the extracellular space (Froger et al., 2010; Rana and Dringen, 2007; Spray et al., 2006; Ye et al., 2003).

    Gap junctions help to coordinate cell firing in neuronal networks and adjust metabolic and transcriptional activities between coupled neurons and astrocytes. The ionic and metabolic coupling behavior of gap junctions are traditionally studied by means of double patch-clamp electrophysiological recordings in cell cultures or acute brain slices (Meme et al., 2009), and the uptake and redistribution or intercellular transfer of fluorescent dyes including lucifer yellow, biocytin and neurobiotin.

    Homomeric connexons feature connexins of a single type, whereas heteromeric ones are composed of different connexins. Gap junctions composed of identical connexons are called homotypic channels, whereas heterotypic channels are formed by different connexons (Willecke et al., 2002). Although the number of different combinations of connexins in heterotypic channels may be very large, specific connexins may be compatible with only a small number of other connexins. There are also connexins that are not able to form heterotypic junctions, no matter what other type of connexin is expressed in the partner cell. The particular connexin composition of the two hemichannels determines the channel’s biophysical properties including its conductance, which can range between 10 and 300 pS, general permeability, ion and metabolite selectivity, affinity and coupling asymmetry (Sosinsky, 1996). The gating and kinetic properties of gap junctions composed of specific connexons have been investigated after complementary DNA (cDNA) transfection of cultured cell lines or Xenopus oocytes. The gating of gap junction channels in the brain is dynamically regulated. The channel conductance changes depending on transjunctional voltage, intracellular Ca²+ levels, sodium and magnesium levels, phosphorylation, intracellular pH and cytokines (Dermietzel, 1998; Nicholson et al., 2000; Salameh and Dhein, 2005). Gap junction channels show activity-dependent plasticity, such as changes in gap junction conductance (Yang et al., 1990), subunit composition, number of coupled cells and changes due to posttranslational modifications.

    Intercellular communication via gap junctions allows much faster information transfer between cells and across the brain compared with chemical transmission between presynaptic and postsynaptic domains. Compared to chemical transmission, intercellular transmission mediated via gap junctions has been found to be either rectifying or bidirectional (Phelan et al., 2008). However, it is generally assumed that the signal transmission mediated by gap junctions is much more diffuse than that mediated by chemical neurotransmission. Nevertheless, the importance of gap junctions for neuronal function, communication and plasticity, as well as for behavioral processes, has been underestimated. In the past decade various gap junction mouse mutants have been generated, which are deficient in one or two connexins either in specific cell types or in all cells of the central nervous system. These gap junction mutant mice have enabled neuroscientists to examine the effects of connexin deficiency on various electrophysiological and behavioral measures.

    Structural Aspects of Gap Junctions

    In the human genome 21 and in the mouse genome 20 different connexin genes have been identified, which are coding for distinct connexin proteins (Söhl and Willecke, 2003; Willecke et al., 2002). In this chapter connexin genes and their corresponding proteins will be named according to the most widely used nomenclature based on the predicted molecular weights of different connexin proteins (e.g. Cx45 has a mass of Cx45 kDa). The connexin genes Cx25 and Cx59 are specific for the human genome and the Cx33 gene is found only in the mouse genome (Söhl and Willecke, 2004). Connexin genes are generally classified in terms of their specific molecular mass, which is expressed in kDa, and exhibit cell-type-, organ- and tissue-specific expression patterns. However, most of the cells forming gap junctions with other cells express connexins of more than one connexin type in a combination that is specific for each cell type (White and Paul, 1999). The molecular topography of connexin proteins includes four alpha-helical transmembrane domains, intracellular N- and C-termini, two extracellular loops and a cytoplasmic loop. The two or three cysteine residues are located in the two extracellular loops, which are required for the correct alignment of two hemichannels through H-bonds to form a continuous gap junction channel between apposed cells (Yeager, 1998). The cytoplasmic C-terminus carries several serine, threonine and tyrosine residues, which are targets of a number of protein kinases and phosphatases for posttranslational modifications (Solan and Lampe, 2005). The intracellular loop and the C-terminus constitute the least homologous regions across different connexins and are likely to confer different biophysical properties to gap junctions that are composed of different connexins, including the degree of unitary conductance, pH dependence, voltage dependence and permeability to small molecules up to 1 kDa. After six connexin proteins have been oligomerized to form a gap junction hemichannel they are enclosed in vesicles and transported to the cell membrane, where they are inserted and become an integral part of the cell membrane. Similarly to other membrane proteins such as neurotransmitter receptors, gap junction hemichannels are internalized into cytoplasmic vesicles and metabolized by proteasomes and/or lysosomes (Jordan et al., 2001; Laird, 2008). Recently, it has been demonstrated that connexin proteins can interact with adhesion molecules and other signaling elements of the cell membrane (Alev et al., 2008; Kardami et al., 2007).

    Types of Gap Junction

    Astrocytes in the brain are extensively coupled via gap junctions constituting a glial syncytium. The great majority of the gap junctions in the brain are formed between astrocytes composed of Cx43 and Cx30. Dye transfer coupling studies performed in doubly Cx43 and Cx30-deficient mice confirmed that astrocytic wiring via gap junctions is critically dependent on these two connexins (Wallraff et al., 2006).

    Coupled astrocytes regulate interstitial ion concentrations such as K+ and glutamate, as well as metabolic processes (Longuemare et al., 1999; Tsacopoulos and Magistretti, 1996), and thereby maintain a normal level of neuronal excitability. Furthermore, astrocytic gap junctions are critically involved in the redistribution of K+ ions after neuronal activity, glutamate homeostasis, long-distance propagation of Ca²+ waves in the brain (Giaume and Venance, 1998; Rottingen and Iversen, 2000), and transmitter release by neurons and astrocytes (Martin et al., 2001). The conditional astrocyte-specific deletion of Cx43-coding DNA in the mouse attenuated gap junctional coupling and impaired the propagation of calcium waves (Theis et al., 2003). The coupling of astrocytes via gap junctions in the brain serves important physiological functions including intercellular calcium signaling and metabolic trafficking within the brain. It is known that calcium waves in coupled astrocytes are transmitted to surrounding neurons, which respond with enduring increases in intracellular calcium levels. Prolonged increases in intracellular calcium levels in neurons can activate protein kinases and gene transcription factors, such as Ca²+/calmodulin-dependent protein kinase II (CaMKII), CaMKIV and protein kinase C (PKC), which in turn can change the excitability and morphology of synapses located on that neuron, similar to the changes seen after the induction of synaptic long-term potentiation (Ben Achour et al., 2010; Haydon, 2001; Wang et al., 2006). Given that coupled astrocytes can communicate with each other via the propagation of calcium waves and with surrounding neurons via the release of neurotransmitters (such as glutamate or D-serine), as well as other extracellular signaling molecules (such as ATP), it is conceivable that they play a much more active role in information processing and higher cognitive functions than previously assumed (Fields and Stevens-Graham, 2002; Nedergaard et al., 2003). Recently, astrocyte hemichannels composed of Cx43 have been implicated in the β-amyloid-induced neurodegeneration that is associated with Alzheimer’s disease. It has been found that low concentrations of the active fragment of β-amyloid increased hemichannel activity in cultured microglia, astrocytes or neurons. The ATP and glutamate released from activated neuronal hemichannels have been found to induce neuronal death, and β-amyloid-induced neurodegeneration was significantly reduced in acute slices from Cx43 knockout mice (Orellana et al., 2011).

    Another type of glial cell, which is important for the fast propagation of action potentials in myelinated axons, is the oligodendrocyte. Oligodendrocytes express the connexins Cx29, Cx32 and Cx47 (Menichella et al., 2003; Nagy et al., 2001). In vivo, oligodendrocytes form gap junctions with astrocytes, by heterotypic gap junctions consisting of Cx32 and Cx47 at the oligodendrocyte and Cx30 and Cx43 at the astrocytic side, but not with each other. In Cx43/Cx30 double-deficient mice these gap junctions are nearly abolished (Li et al., 2008; Maglione et al., 2010).

    In the adult brain, neurons form gap junctions with other neurons (Bennett and Zukin, 2004; Connors and Long, 2004; Hormuzdi et al., 2004), but not with astrocytes. However, there is some evidence for significant neuron–glia coupling in a few brain regions. By means of positive immunolabeling, gap junctions between astrocytes and neurons have been detected in the cortex (Bittman et al., 2002) and the noradrenergic locus coeruleus (Alvarez-Maubecin et al., 2000). Accordingly, the traditional view of an absolute functional dichotomy between gap junctions and chemical synapses has recently been corrected. Some neurons, such as Mauthner cells of goldfish, exhibit mixed synapses. Mauthner cells exhibit both gap junctions and chemical synapses in a single synaptic contact, therefore allowing electrotonic, metabolic and chemical transmission (Kandler and Thiels, 2005; Pereda et al., 2004).

    Neocortical and hippocampal gap junctions between neurons are primarily formed between a subset of inhibitory interneurons. These γ-aminobutyric acidergic (GABAergic), fast-spiking, parvalbumin-containing interneurons are coupled through Cx36. Accordingly, it has been reported that the electrical coupling between pairs of interneurons in the CA3 region and in the dentate gyrus observed in wild-type mice is abolished in Cx36 knockout mice (Hormuzdi et al., 2001).

    In respect of the subcellular localization of gap junctions between GABAergic neurons, it has been found that these interneurons form gap junctions between their dendrites, and between their dendrites and somata (Fukuda et al., 2006; Simon et al., 2005). However, there may also be gap junctions, which are formed between excitatory pyramidal neurons (Connors and Long, 2004; Fukuda, 2007). Indirect evidence for the existence of gap junctions between the axons of pyramidal neurons in the neocortex and CA1 region of the hippocampus has been obtained from intracellular recordings from pairs of cortical and CA1 pyramidal neurons (Mercer et al., 2006; Schmitz et al., 2001; reviewed in Traub et al., 2002; Y. Wang et al., 2010). Here, the observed interpyramidal electrotonic coupling was relatively strong, with on average 25% of the steady state and 10% of the peak action potential voltage change in one cell transferred to the other (Mercer et al., 2006).

    Cellular and Brain Regional Expression Patterns of Specific Connexins

    In the following, the cellular and brain regional messenger RNA (mRNA) and protein expression patterns of specific glial and neuronal connexins are described for which mouse mutants have been generated and phenotyped in behavioral and/or electrophysiological experiments. The connexins Cx26, Cx30.2, Cx36, Cx45 and Cx57 are expressed in neurons (Kreuzberg et al., 2008; Söhl et al., 2005; Vandecasteele et al., 2006; Venance et al., 2004). Cx57 is exclusively expressed by horizontal cells of the retina (Shelley et al., 2006). Cx31.1 may also be expressed in neurons of the brain. In the rat brain, Cx31.1 mRNA has been demonstrated in GABAergic striatal output neurons (Venance et al., 2004) and dopaminergic neurons of the substantia nigra pars compacta (Vandecasteele et al., 2006). Thus, there are probably five different connexins (Cx26, Cx30.2, Cx31.1, Cx36 and Cx45) that are expressed in neurons of the brain. The Cx30.3 is expressed in the progenitor cells of the olfactory epithelium and in cells of the vomeronasal organ (Zheng-Fischhöfer et al., 2007a) and cochlea (W.-H. Wang et al., 2010). The expression patterns of selected connexins that are expressed in the brain and some of their physiological and pathological functions in the brain are described in the following paragraphs.

    Connexin30

    Cx30 is expressed in astrocytes, ependymal and leptomeningeal cells. Cx30 expression has been demonstrated in various tissues including the skin, cochlea, uterus, lung and eye (Dahl et al., 1996; Kunzelmann et al., 1999). Cx30-deficient mice exhibited inner ear pathology in terms of increased apoptosis within the cochlear sensory epithelium, lacked endocochlear potential and showed progressive hearing loss (Teubner et al., 2003). The behavioral phenotypes of these and other gap junction mouse mutants are described in detail in Chapter 17.

    Connexin30.3

    Immunoblot analysis indicates that the Cx30.3 protein is expressed in the skin, heart, kidney and cochlea (W.-H. Wang et al., 2010). The expression pattern of Cx30.3 in the brain and periphery was also characterized by analyzing tissues from Cx30.3-deficient mice, which express a lacZ reporter gene instead of the Cx30.3 protein. The Cx30.3 protein is expressed in the epidermis, in the kidney, in progenitor cells of the olfactory epithelium and in the vomeronasal organ. However, no abnormalities in the skin or in the chemosensory systems were observed in Cx30.3-deficient mice.

    Connexin31.1

    The Cx31.1 gene was initially detected in the testes and the epidermis of the skin (Haefliger et al., 1992). Later on, it was demonstrated that Cx31.1 mRNA is also expressed in dopaminergic neurons of the substantia nigra pars compacta (Vandecasteele et al., 2006) and in rat striatal output neurons (Venance et al., 2004). These data suggest that Cx31.1 is likely to be expressed in neurons in other parts of the brain. To further characterize further the brain regional expression characteristics and behavioral functions of this connexin, Cx31.1-deficient mice were generated (Zheng-Fischhöfer et al., 2007b). In Cx31.1-deficient mice part of the coding sequence of the gene has been replaced with a LacZ reporter gene. Although Cx31.1 is strongly expressed in the skin, Cx31.1-deficient mice did not show significant morphological or functional defects of skin. LacZ staining revealed that Cx31.1 is expressed in the olfactory epithelium (Zheng-Fischhöfer et al., 2007b).

    Connexin32

    Cx32 is expressed in myelinating Schwann cells of the peripheral nervous system, where it connects the Schwann cell body and the cytoplasmic collar of the myelin sheath, and in oligodendrocytes and their processes (Scherer et al., 1995, 1998). As described in some detail below, missense mutations in the Cx32 gene cause Charcot–Marie–Tooth disease (Abrams et al., 2001).

    Connexin36

    The major neuronal connexin, Cx36, is exclusively expressed in GABAergic, fast-spiking, parvalbumin-positive interneurons throughout the mammalian brain. While Cx36 mRNA has been identified in pyramidal cells of the CA3 subregion in the hippocampus (Condorelli et al., 1998; Venance et al., 2000), its corresponding Cx36 protein could not be detected in these cells. Cx36 forms only homomeric connexons (Teubner et al., 2000) and gap junction channels with a very low main state conductance of 10–15 pS (Srinivas et al., 1999). Brain Cx36 expression has been determined by in situ hybridization RNA, lacZ reporter gene and Western-blot protein analysis. Cx36-positive neurons have been demonstrated in the retina, dentate gyrus, CA1, CA3 and CA4 regions of the hippocampus, cerebral and piriform cortex, amygdala, cerebellum, mesencephalon, suprachiasmatic nucleus, thalamus, hypothalamus and various brainstem nuclei (Condorelli et al., 1998, 2000, 2003; Degen et al., 2004; Helbig et al., 2010; Rash et al., 2007; Söhl et al., 1998). Gap junctions featuring the Cx36 protein have been described among various types of neurons in the olivocerebellar system (De Zeeuw et al., 1998; Kistler et al., 2000). The neurons in the nucleus olivaris inferior send the climbing fiber input to the Purkinje cells. Cx36 mRNA and its corresponding protein were detected in cerebellar basket cell interneurons of the molecular layer (Belluardo et al., 1999; Condorelli et al., 2000), and in GABAergic Purkinje cells (Meller et al., 2005). Furthermore, Cx36 has been detected in the neostriatum and ventral striatum, including the nucleus accumbens (Condorelli et al., 1998, 2000), in dopaminergic neurons of the substantia nigra (Vandecasteele et al., 2006) which project to the neostriatum, and in GABAergic interneurons of the area tegmentalis ventralis (Allison et al., 2006). Cx36-deficient mice have been generated in three different laboratories (Deans et al., 2001; Güldenagel et al., 2001; Hormuzdi et al., 2001). In Cx36-deficient mice electrical coupling between subpopulations of neurons in the dentate gyrus and CA3 region of the hippocampus (Hormuzdi et al., 2001), neocortex (Deans et al., 2001), thalamic reticular nucleus (Landisman et al., 2002) and inferior olivary nucleus (Long et al., 2002) is either completely absent or strongly decreased.

    Connexin43

    Cx43 forms gap junction channels with a moderate conductance of 90–115 pS (Brink et al., 1996). Cx43 channels are rather insensitive to differences in transjunctional voltage and close upon membrane depolarization (González et al., 2007). In the developing embryo neuronal precursor cells are extensively coupled via gap junction channels containing Cx43 (Bruzzone and Dermietzel, 2006). In the adult brain Cx43 is the main constituent of the brain-spanning astrocytic gap junction network (Yamamoto et al., 1992). Other connexins expressed in astrocytes are Cx30 and Cx26 (Koulakoff et al., 2008; Nagy and Rash, 2000). This astrocytic network exerts several important functions including the control of extracellular ion and neurotransmitter concentrations and metabolic processes. The astrocytic gap junction network is dynamically regulated through phosphorylation of Cx43 by protein kinases including mitogen-associated protein kinase (Warn-Cramer, 1998), PKC (Lampe, 1994) and tyrosine kinase (Loo et al., 1995). Phosphorylation of Cx43 induces the uncoupling of cells and suppresses gap junction-mediated intercellular signal transfer. Astrocytes that are deficient in Cx43 exhibit impaired gap junction coupling and propagation of calcium waves (Scemes et al., 1998; Naus et al., 1997). The latter are known to influence neuronal activity. Given that the knockout of Cx43 in the postnatal mouse is lethal, owing to heart malfunction (Reaume et al., 1995), conditional, for example astrocyte-specific Cx43-deficient mice (Theis et al., 2003) have been generated for electrophysiological and behavioral studies.

    Connexin45

    Cx45 forms gap junction channels with a rather low main state conductance of 30 pS. Gap junction channels composed of Cx45 are highly sensitive to differences in transjunctional voltage channels and close upon membrane hyperpolarization (González et al., 2007). Cx45 protein expression has been detected in various neuronal marker protein-positive neurons in the mouse brain. Cx45-positive pyramidal cells have been identified in the following brain regions: neocortex and perirhinal cortex, hippocampus (regions CA1–CA4) and thalamus (Krüger et al., 2000; Maxeiner et al., 2003). Furthermore, Cx45 is expressed in neurons of the olfactory bulb (Zhang and Restrepo, 2002) and in subpopulations of neurons in the olivocerebellar system (Van der Giessen, 2006; Weickert et al., 2005). Cx45 is not expressed in oligodendrocytes or astrocytes (Maxeiner et al., 2003), but may be expressed in the pyramidal cells of the hippocampus, because the latter are known to be electrically coupled, but do not express Cx36 (Schmitz et al., 2001).

    Whereas homozygous Cx45-deficient mice showed cardiovascular defects and embryonic lethality (Krüger et al., 2000), conditional neuron-specific Cx45-deficient mice were viable and fertile (Maxeiner et al., 2003).

    Gap Junction Pharmacology

    Several compounds are able, among other effects, to suppress intercellular communication via gap junctions, as inferred from the reduction or cessation of electrotonic or dye-transfer coupling between cells. The most widely used gap junction blockers for in vitro studies are carbenoxolone, quinidine, mefloquine, heptanol, octanol, anandamide and oleamide (Juszczak and Swiergiel, 2009). Carbenoxolone is the most popular gap junction blocker. It is a water-soluble glycyrrhetinic acid derivative that rapidly and reversibly blocks gap junctions in cell cultures and acute brain slices (Blomstrand et al., 2004).

    Regarding the pharmacological blockade of gap junction-mediated cell-to-cell communication in the electrophysiological studies reviewed below, one should take into account that pharmacological research into the question of how electrical coupling is involved in neuronal network oscillations suffers from the poor specificity of the available drugs. Carbenoxolone has been reported to inhibit both α-amino-3-hydroxy-5-methyl-4-isoxazole-proprionic acid (AMPA)- and GABA-receptor mediated synaptic transmission (Rouach et al., 2003; Tovar et al., 2009). Given that neuronal network oscillations are the consequence of complex cellular interactions involving both AMPA- and GABA-mediated synaptic transmission, lack of specificity of gap junction uncoupling agents complicates the interpretation of pharmacological studies on the role of gap junctions in the generation of neuronal network oscillations.

    The functional role of neuronal or glial connexins at the behavioral level has been difficult to assess by means of conventional behavioral pharmacology, since the drugs available are chemically diverse, tend to have only partial efficacy, affect vital peripheral organs, such as the heart (when administered systemically), have poor selectivity for different connexins and have significant effects on other cellular processes in addition to blockade of gap junction transmission (Hervé and Sarrouilhe, 2005). For example, the gap junction blocker, carbenoxolone, has only poor blood–brain barrier permeability (Leshchenko et al., 2006) and has, in addition to the blockade of intercellular communication, several other effects including the inhibition of voltage-gated Ca²+ channels (Vessey et al., 2004), p2x7 receptors (Suadicani et al., 2006) and 11β-hydroxysteroid dehydrogenase (Bujalska et al., 1997). The use of carbenoxolone in behavioral studies, e.g. via intracerebral microinfusion, is further complicated by the fact that it has toxic effects on mitochondria (Pivato et al., 2006).

    Current tools for the pharmacological modulation of gap junctions also include gap junction channel openers and connexin mimetic peptides. For example, trimethylamine is a gap junction opener, which has been shown to transiently enhance the amplitude, power and duration of theta oscillations in anesthetized rats (Bocian et al., 2011), and to reverse the anticonvulsant effects of the gap junction blocker quinine in the pentylenetetrazole model of epilepsy in rats (Nassiri-Asl et al., 2008). Recently, connexin mimetic peptides have been developed to enable the rapid and reversible inhibition of connexin channels (Evans and Leybaert, 2007). The most widely used connexin mimetic peptides are gap26 and gap27. They bind to the extracellular loops of Cx37, Cx40 and Cx43 hemichannels and decrease the conductance of the channels (Evans and Leybaert, 2007).

    Within the limits outlined above, gap junction blockers including carbenoxolone and gap junction openers such as trimethylamine have proved to be useful tools to investigate the role of gap junctions in the synchronization of neuronal network activity, in epileptiform activity and recently for synaptic plasticity in vitro (Chepkova et al., 2008). In terms of electrophysiological studies in acute slice preparations reviewed in the remainder of this chapter, quinine and its derivatives are especially interesting since they are probably able to selectively block neuronal gap junctions featuring Cx36 or Cx45, while having no inhibitory effect on gap junctions composed of Cx26, Cx32 and Cx43 (Srinivas et al., 2001).

    Synchronization and Rhythmic Oscillation of Neural Activity

    There is in vivo and in vitro pharmacological evidence indicating that synchronization and rhythmic oscillation of large neuronal ensembles in the hippocampus and neocortex at theta, gamma and high-frequency oscillations are mediated or can be modulated by intercellular electrotonic and metabolic communication via gap junctions (Draguhn et al., 2000; Hormuzdi et al., 2004; Ylinen et al., 1995). Some of these rhythmic oscillations have been proposed to be involved in processes of perception, attention and memory consolidation at both the cellular and systems level (Buzsáki and Chrobak, 1995; Singer and Gray, 1995). Gamma (40–100 Hz), theta (4–12 Hz) and high-frequency oscillations (in the range of 80–300 Hz) of neuronal ensembles in the hippocampus formation correlate with distinct behavioral states in the rat (Buzsáki, 1989, 2002; Draguhn et al., 2000). During exploratory behaviors or during rapid eye movement sleep, mainly theta and gamma oscillations are recorded from neuronal ensembles in the hippocampus. These frequency bands have been related to the read-in of sensory information. During immobile waking, consummatory behaviors or slow-wave sleep, high-frequency oscillations have been recorded from neuronal ensembles in the hippocampus. High-frequency oscillations have been related to the read-out of the results of intrahippocampal computations for their permanent storage in the neocortex (Buzsáki, 1989, 2002).

    Theta and Gamma Oscillations

    The gap junction blocker carbenoxolone reversibly suppressed synchronized theta activity both in hippocampal slice preparations and in vivo, where theta activity was recorded from the hippocampus of urethane-anesthetized rats (Bocian et al., 2009; Konopacki et al., 2004). The gap junction blocker mefloquine disrupted synchronized activity at the frequency range of 8–20 Hz in the CA3 subfield of the hippocampus (Gee et al., 2010). In addition to carbenoxolone, quinine, another gap junction blocker, was capable of suppressing theta activity recorded from the hippocampus of rats (Konopacki et al., 2004). In line with these findings, the gap junction opener trimethylamine enhanced the amplitude and power of theta oscillations and increased the duration of theta epochs in anesthetized rats (Bocian et al., 2011). In freely moving rats, intracerebroventricular infusion of carbenoxolone attenuated the power of theta rhythms in the CA1 region and increased their frequency, but did not change the positive correlation between theta power and running speed (Bissiere et al., 2011). Similarly, in freely moving cats the gap junction blockers carbenoxolone and quinine, administered either intraperitoneally or intrahippocampally, abolished and diminished, respectively, theta activity in the hippocampus (Gołebiewski et al., 2006).

    It has been shown that gap junctions between interneurons in the neocortex and neostriatum contribute to subthreshold and suprathreshold synchronization of network activity (Gibson et al., 1999; Koos and Tepper, 1999). Modeling studies have indicated that both theta and gamma oscillations could be maintained by the activity of inhibitory interneurons that are wired via gap junctions (Traub et al., 2000). Gap junctions between the dendrites of interneurons increase the synchrony of gamma oscillations in isolated interneuron and mixed interneuron–pyramid cell networks (Csicsvari et al., 2003; Traub et al., 2001, 2003). The knockout of the Cx36 gene in the mouse (Güldenagel et al., 2001) induces a loss of gap junction coupling between fast-spiking interneurons in the hippocampus and the cortex (Buhl et al., 2003; Deans et al., 2001; Hormuzdi et al., 2001). Surprisingly, hippocampal theta or gamma oscillations were not completely abolished in the Cx36 knockout mice. However, hippocampal gamma oscillations in vitro (Hormuzdi et al., 2001) and in vivo (Buhl et al., 2003) were reduced in terms of overall power and synchrony in the Cx36-deficient mice. Recently, it has been reported that Cx36 knockout mice also exhibited slower hippocampal theta oscillations, decreased spatial and temporal coding of pyramidal place cells (Allen et al., 2011).

    Furthermore, it has been reported that gap junction blockers suppress gamma oscillations in Cx36-deficient mice (Traub et al., 2001, 2003). Given that gamma oscillations in the hippocampus of Cx36-deficient mice are preserved and can be suppressed by pharmacological gap junction inhibition, these oscillations may be maintained by possible axoaxonal gap junctions between CA1 pyramidal neurons (Schmitz et al., 2001), which may be composed of Cx45. In conclusion, gamma frequency oscillations in the hippocampus may therefore depend on axonal gap junctions between pyramidal cells (which, however, remains to be demonstrated), while dendritic gap junctions between interneurons may have only a modulatory effect on these oscillations (Traub et al., 2000, 2003, 2004). However, it should be noted that to date there is only one study demonstrating the existence of axoaxonal gap junctions between axons of dentate granule cells (Hamzei-Sichani et al., 2007).

    High-Frequency Oscillations

    The first demonstration that gap junctions are involved in the generation of high-frequency oscillations in the hippocampus has been provided by Ylinen et al. (1995). They showed that the anesthetic drug halothane, which also blocks gap junctions, prevented the occurrence of high-frequency oscillations in the CA1 region of rats. Later, it was shown that the gap junction blocker carbenoxolone inhibits high-frequency oscillations also in the CA3 region of the rat hippocampus (Draguhn et al., 1998; Papatheodoropoulos, 2007). Moreover, it has been demonstrated that high-frequency oscillations in the hippocampus persist when slices are perfused with the GABAA receptor antagonist bicuculline or the ionotropic glutamate receptor antagonist NBQX, or even when chemical synaptic transmission is blocked completely by a nominally calcium-free solution (Draguhn et al., 1998). However, it should be noted that high-frequency oscillations induced by disinhibition through GABAA receptor blockade (D’Antuono et al., 2005) or by tetanic stimulation (Poschel et al., 2003) are not inhibited by pharmacological gap junction blockade.

    In Cx36-deficient mice theta- and high-frequency oscillations analyzed in terms of power, intraepisode frequency or probability of occurrence in vivo remained largely intact (Buhl et al., 2003; Hormuzdi et al., 2001). However, there is evidence (although only in vitro) that high-frequency oscillations occurred less frequently and were slightly slower in Cx36-deficient mice (Maier et al., 2002).

    Connexin45 and Neuronal Network Oscillations

    Gap junction channels composed of Cx45 exhibit large voltage sensitivity and show a rather low single channel conductance of about 30 pS. Cx45-containing channels show even lower conductance when differences between membrane potentials of coupled cells increase (Schubert et al., 2005). In contrast, when coupled cells are firing simultaneously, these gap junction channels may increase their conductance and therefore support the synchronization of cell firing. Conditional neuron-specific Cx45-deficient mice showed normal general excitability, synaptic short-term plasticity and spontaneous high-frequency oscillations in the hippocampus. Stimulation with kainate of hippocampus slices derived from neuronal Cx45-deficient mice induced decreases in gamma-oscillation amplitudes in the CA3, but not in the CA1 subfield. CA1 subfield gamma frequency in neuronal Cx45-deficient mice was more variable, suggesting a role of Cx45 in neuronal synchronization at gamma frequency bands (Zlomuzica et al., 2010).

    Connexin36 and Olivocerebellar Network Oscillations

    The olivocerebellar system has been implicated in the temporal sequencing of motor programs and the detection of errors during their execution. Neurons in the inferior olive send climbing fibers to the cerebellar cortex and exhibit low-frequency (1–10 Hz) subthreshold oscillations, which are mediated by intercellular electrotonic coupling via gap junctions (Llinás, 2009). In Cx36-deficient mice morphological and electrophysiological changes in neurons of the inferior olive have been observed. Neurons in the inferior olive had altered membrane properties and displayed abnormally thick dendrites with non-functional gap junction-like structures (De Zeeuw et al., 2003). However, these changes had no suppressive effect on low-frequency subthreshold oscillations, which have been proposed to be mediated by gap junctions containing Cx36 (De Zeeuw et al., 2003). Instead, these low-frequency subthreshold oscillations were, although reduced, still recordable and appeared to be normal in terms of their shape and frequency (Long et al., 2002). However, the low-frequency subthreshold oscillations as well as the spikes they evoked were not synchronized among neighboring cells in the inferior olive of the Cx36-deficient mice (De Zeeuw et al., 2003). It has been concluded that gap junctions are not critical for the generation but are necessary for the synchronization of low-frequency subthreshold oscillations in the inferior olive.

    Gap Junction-Related Human Diseases

    It is increasingly recognized that dysfunctional intercellular communication between glia cells or neurons mediated by gap junctions can contribute to or even cause a variety of human diseases, including inherited diseases based on gene defects or mutations. In this regard it has been proposed that changes to gap junctional intercellular communication may play a role in the establishment of epileptiform activity, induce brain damage and trigger neurodegenerative diseases. For example, it has been reported that the gap junction blocker carbenoxolone inhibited glutamate release from activated microglia and decreased neuronal cell death in vitro and protected against transient global ischemia-induced neurodegeneration in the hippocampus of gerbils (Takeuchi et al., 2008). Furthermore, carbenoxolone inhibited both spontaneous and evoked epileptiform activity in organotypic hippocampus slice cultures (Samoilova et al., 2008). Mutations in single gap junction genes or their aberrant expression have been implicated in a variety of human diseases (Spray and Dermietzel, 1995). On the other side, it has been demonstrated that changes in intercellular gap junction-mediated communication in astrocytic networks in response to brain injury can have neuroprotective effects (Leite et al., 2009).

    Several human skin disorders are correlated with mutations in connexin genes. Mutations in the DNA encoding for Cx30.3 or Cx31 have been linked to the skin disease erythrokeratodermia variabilis (Macari et al., 2000; Richard et al., 1998). However, humans deficient in functional Cx30.3 (van Geel et al., 2002) and Cx30.3-deficient mice have no epidermal pathology (Zheng-Fischhöfer et al., 2007a). It seems that mutations in the genes coding for Cx31.1 or Cx30.3 in which no connexin protein is expressed can be compensated by the remaining functional connexin gene and do not lead to skin pathology. However, such compensation is either only initiated or successful in the case of the complete absence of one of these connexins, and this compensation fails if a mutated and non-functional form of a connexin is expressed at the protein level. Other inherited skin diseases, including Vohwinkel’s syndrome, coincide with mutations in the gene coding for the Cx26 protein.

    Mutations in connexin genes which are expressed in myelin-producing cells have been implicated in demyelinating diseases. Cx32 is expressed in myelinating Schwann cells in the peripheral nervous system and in oligodendrocytes. Mutations in the DNA sequence encoding for the human Cx32 protein cause a peripheral neuropathy known as Charcot–Marie–Tooth disease, which is characterized by a demyelinating phenotype with reductions in motor nerve conduction velocities and distal muscle weakness (Garcia, 1999). Cx32 knockout mice exhibit the expected progressive demyelinating peripheral, but not central neuropathy. The peripheral neuropathy of these mice affects both motor and sensory nerves (Scherer et al., 1998).

    Furthermore, Pelizaeus–Merzbacher-like disease is an immedicable disease caused by homozygous mutations in the human DNA encoding for the Cx47 gene, and is characterized by hypomyelination, progressive degeneration of white matter in the brain, nystagmus, dysarthria, impaired mental and psychomotor development, and progressive spasticity. Individuals with corresponding heterozygous mutations of the Cx47 gene do not exhibit these neurological symptoms (Uhlenberg et al., 2004). Recently, transgenic mouse models of Pelizaeus–Merzbacher-like disease have been generated, which can be used to investigate the neuropathological mechanisms of this disease and to search for effective treatments (Tress et al., 2011).

    Gap junctions between astrocytes have been implicated in traumatic brain injury, brain infarcts, epilepsy and migraine, and recently in Alzheimer’s disease (Giaume et al., 2007; Koulakoff et al., 2012). Mice that are expressing only a truncated form of the Cx43 protein exhibit increased cerebral infarct size and inflammatory cell invasion in the peri-infarct region together with reduced astrogliosis in an animal model of stroke (middle cerebral artery occlusion) (Kozoriz et al., 2010). Increased Cx43 expression has also been reported in response to ischemic brain lesions and Cx43-deficient mice exhibit greater infarct sizes after experimental stroke (Siushansian et al., 2001), suggesting that Cx43 deficiency decreases astrocyte-mediated neuroprotection in experimental ischemia. Haupt et al. (2007a) analyzed the expression of Cx43 in the postischemic rat brain induced by photothrombosis. In the hippocampal formation and somatosensory cortex of rats subjected to photothrombosis, the number of Cx43 mRNA-positive astrocytes was increased. In addition, the expression of the Cx43 protein was found to be selectively increased in the ipsilateral stratum oriens, but reduced in the somatosensory cortex of the injured hemispheres. The same group has also reported that after brain injury, reactive proliferating astrocytes at the lesion site exhibited upregulated Cx43 expression at both the mRNA and protein level (Haupt et al., 2007b). From these results it was concluded that Cx43 plays an important role in glial scar formation and in the proliferation of astrocytes in response to brain injury. In addition to Cx43, Cx30 may play a role during or in response to epileptoform activity in the brain. It has been reported that after the induction of seizures Cx30 transcripts are upregulated in astrocytes and neurons (Condorelli et al., 2002).

    The neuroprotective effect of the corticotropin-releasing hormone may be associated with increased levels of astrocytic Cx43 expression (Hanstein et al., 2009). It is therefore possible that the increased expression of Cx43 after brain injury or ischemia has a neuroprotective role (Nakase et al., 2003).

    The term cortical spreading depression refers to a propagating wave of excessive neuronal and astrocytic depolarization due to elevated levels of extracellular K+, which is followed by short periods of complete cessation of neuronal activity. There is evidence indicating that cortical spreading depression is involved in the generation of migraine aura and pain. Recently, it has been reported that gap junction blockers may have therapeutic potential in preventing migraine attacks (Durham and Garrett, 2009; Silberstein, 2009).

    It is known that cells coupled via gap junctions containing the astrocytic Cx43 are involved in the mediation of cortical spreading depression. Conditional astrocyte-directed Cx43-deficient mice exhibited accelerated potassium-induced spreading depression in hippocampal slices (Theis et al., 2003), suggesting that astrocytic gap junctions consisting of Cx43 may be implicated in the spatial buffering of extracellular K+ after physiological and pathophysiological forms of neuronal activity including epileptoform activity (Theis et al., 2005). It is also known that gap junctions are involved in the synchronization of neural activity in various parts of the brain including the hippocampus and amygdala. These brain structures tend to generate epileptoform activity (Löscher and Ebert, 1996; Sinfield and Collins, 2006). It has been proposed that gap junctions may play a role in the generation of epileptic seizures (Carlen et al., 2000). In line with this proposal, anticonvulsant effects of gap junction blockade have been reported (Carlen et al., 2000). However, Tsc1-deficient mice, an animal model of tuberous sclerosis complex, which is associated with astrocytic dysfunction in affected individuals, exhibited decreased levels of Cx43 expression together with epileptiform activity and reduced astrocytic coupling via gap junctions (Xu et al., 2009).

    There is also evidence that gap junctions may be involved in the pathophysiological processes seen after traumatic brain injury. In rats, traumatic brain injury increases the expression of phosphorylated Cx43 and extracellular signal-regulated kinase (ERK) expression in astrocytes of the ipsilateral hippocampus, but not in the cortex. This increase is observed as soon as 1 h after the traumatic brain injury, reaches its plateau after 6 h, and is still elevated 24 h after the lesion (Ohsumi et al., 2010).

    Recent research showed that adult neurogenesis generated by radial glia-like cells in the dentate gyrus of the hippocampus involves gap junctions or is at least dependent on the expression of connexins (Kunze et al., 2009). It was shown that the majority of radial glia-like cells are coupled via gap junctions made of Cx43 and that the deletion of both Cx43 and Cx30 in the mouse (Cx30−/−/Cx43fl/fl/hGFAP-Cre/dKO mice) led to a near complete block of cell proliferation and decreased the number of both radial glia-like and granule cells (Kunze et al., 2009). Hippocampal neurogenesis was also significantly decreased by the blockade of connexin expression by means of injections of a Cre-expressing retrovirus into Cx30−/−/Cx43fl/fl mice that had a selective toxicity for proliferative cells (Kunze et al., 2009). The implication of gap junctions in adult neurogenesis is especially interesting in respect of several age-related neurodegenerative diseases including Alzheimer’s disease. Future research on this finding may reveal novel molecular targets to induce or facilitate neurogenesis in the aged rodent brain and, it is hoped, in individuals suffering from Alzheimer’s disease.

    Conclusion

    The evidence reviewed above clearly indicates that intercellular communication via gap junctions in the brain plays important physiological and pathological roles. Disruption of gap junctional coupling between cells in the brain is associated with a variety of disease conditions. Gap junctions in the brain play a generic role in neuronal oscillation and synchronization of cell firing in the neocortex and hippocampus, which may be important for basic cognitive processes as well as for disease conditions such as epilepsy.

    A PubMed search with the terms gap junctions and brain yielded only 23 articles for the year 1990, while the same search for the year 2010 resulted in 193 publications. It can be envisaged that the number of publications in this still emerging research field will increase further in the future. The accumulating literature on the physiological and pathological functions of gap junctions in the brain indicates that it is increasingly recognized that chemical synapses are only one half of the story on how neurons and other cells in the brain communicate and perform their tasks.

    Acknowledgments

    Supported by the German Science Foundation (Deutsche Forschungsgemeinschaft) through grant nos DE1149/4-1 and DE1149/5-1.

    REFERENCES

    1. Abrams CK, Freidin MM, Verselis VK, Bennett MV, Bargiello TA. Functional alterations in gap junction channels formed by mutant forms of connexin 32: evidence for loss of function as a pathogenic mechanism in the X-linked form of Charcot–Marie–Tooth disease. Brain Res.

    Enjoying the preview?
    Page 1 of 1