Development, organization and function of electrical synapses in the thalamus
TABLE OF CONTENTS
Chapter 1: Introduction 1
Electrical synapses and neuronal gap junctions 2 History 2 Molecular basis of electrical synapses 4 Electrical synapses in the developing brain 6 Electrical synapses in the mature brain 8 The synaptic organization of the thalamus 11 Electrical synapses in the thalamus 11 Thalamic circuitry and the TRN 13 Spatial organization of the TRN 18 Connectivity between thalamic relay nuclei and TRN 20 Hypothesis and goals of the thesis research 21
Chapter 2: Electrical and chemical synapses between relay neurons in developing thalamus 31
Abstract 32 Introduction 33 Methods 35 Results 38 Electrical synapses between rodent VB relay neurons 38 Developmental regulation of electrical synapses in VB 38 Connexin36 is a major component of VB electrical synapses 40 Cx36 reporter expression in VB 41 Chemical synaptic interactions between VB relay neurons 42 Discussion 45 Strength of VB electrical synapses and their possible roles 45 Mechanism of down-regulation of electrical synapses 46 Role of Cx36 and Cx45 in gap junction coupling in thalamus 47 Thalamic relay cell development and electrical synapses 48
Chapter 3: Spatial organization of gap junction-coupled networks of GABAergic neurons in the thalamic reticular nucleus 63
Abstract 64 Introduction 66 Methods 72 Results 79 Novel dye injection method for dye-coupling 79
Dye coupling between TRN neurons 81 Distances between dye-coupled neurons 84 Cx36 and dye-coupling 85 Spatial pattern of dye-coupled neuronal clusters 87 Axonal projections, somatic locations, and cluster patterns 90 Discussion 93
Chapter 4: Synchrony of inhibitory synaptic inputs onto thalamic relay neurons from electrically coupled neurons of the thalamic reticular nucleus 122
Abstract 123 Introduction 125 Methods 129 Results 131 Tetanus-induced, long-lasting IPSC barrages in VB neurons 131 Synchrony of tetanus-induced IPSCs between VB neurons from wild-type and Cx36 KO mice 133 Discussion Tetanus-induced barrage of IPSCs onto VB neurons 136 Technical considerations 136 Synchrony of IPSCs between VB neurons 138
Chapter 5: Conclusions 154
Electrical synapses between thalamic relay neurons 155 Spatial organization of electrical synapses in the TRN 158 Functional role of TRN electrical synapses in thalamic circuitry 160 Future directions 161
LIST OF FIGURES
1-1 Molecular organization of gap junction channels 26 1-2 Simplified schematic diagram of thalamocortical circuit 27 1-3 Schematic organization of the lemniscal and paralemniscal streams of vibrissal information in rodents 28 1-4 Spatial organization of TRN neurons 29 1-5 Schematic diagram of the organization of TRN axonal projections 30 2-1 Electrical synapses between rat VB neurons 56 2-2 Developmental change in electrical coupling between rat VB neurons 57 2-3 Electrical synapses between VB neurons of WT mice 58 2-4 Most electrical coupling between young mouse VB neurons depends on Cx36 59 2-5 β-galactosidase reporter expression for Cx36 in mouse VB at different postnatal ages. 60 2-6 Excitatory chemical synaptic transmission between VB neurons 61 2-7 Disynaptic inhibitory transmission between VB neurons 62 3-1 Dye injection into neurons 107 3-2 Three examples of dye-coupled clusters from NB-injected TRN neurons 108 3-3 Number of coupled neurons per cluster under various conditions 109 3-4 Distances between coupled neurons. 110 3-5 Dye-coupling among TRN neurons from Cx36 KO mice 111 3-6 An example of a ‘rod-like’ dye-coupled cluster from a mouse TRN 112 3-7 An example of a ‘disc-like’ dye-coupled cluster from mouse TRN 113 3-8 An example of a ‘disc-like’ dye-coupled cluster from mouse TRN 114 3-9 An example of a ‘sphere-like’ dye-coupled cluster from mouse TRN 115 3-10 An example of a ‘sphere-like’ dye-coupled cluster from mouse TRN 116 3-11 Ratios of different spatial patterns of coupled clusters. 117 3-12 Axonal projections and cluster patterns. 118 3-13 Relative locations of primary injected neurons within TRN with color- coded properties of cluster 119 3-14 Plot of input resistance of primary injected neurons vs. number of coupled neurons within each dye-coupled cluster 120 3-15 Summary diagram of the shapes of clusters in TRN 121 4-1 Tetanus-induced barrages of IPSCs in VB neurons from a wild-type mouse (example 1) 146 4-2 Tetanus-induced barrages of IPSCs in VB neurons from a wild-type mouse (example 2) 147 4-3 Tetanus-induced barrages of IPSCs in VB neurons from a Cx36 KO mouse (example 1) 148 4-4 Tetanus-induced barrage of IPSCs in VB neurons from a Cx36 KO mouse (example 2) 149 4-5 Dominant oscillation frequencies of IPSCs in VB neurons 150 4-6 Cross correlograms from IPSC data collected from VB neuron pairs 151
4-7 Maximum mean cross-correlation coefficients from tetanus-induced IPSCs of VB neuron pairs from WT and Cx36 KO mice 152 4-8 Diagrams of three possible connectivity schemes between TRN neurons and VB neurons 153 4-9 Diagrams of two possible spatial patterns of connection from electrically coupled TRN neurons to VB neurons 154
LIST OF TABLES
2-1 Postnatal development of input resistance and resting membrane potential of VB cells
I. ELECTRICAL SYNAPSES AND NEURONAL GAP JUNCTIONS
History A synapse can be defined as a morphologically specialized structure mediating functional interaction between neurons or between neurons and follower cells (Bennett, 2000). Communication between neurons by synaptic transmission forms a fundamental functional unit of the nervous system. In the early days of neuroscience, a prominent controversy was whether synaptic transmission was chemical or electrical, a dichotomy sometimes characterized as ‘soup or sparks’ (Bennett, 1997). Neurophysiologists favored the notion that synaptic transmission is purely electrical, believing chemical transmission to be both slow and unnecessary. Pharmacologists favored the idea of chemical communication. After a period of research and debate, accumulating evidence indicated that chemical transmission clearly occurs at the neuromuscular junction and many excitatory and inhibitory synapses in the central nervous system at fast time scales (Brock et al., 1952; Coombs et al., 1955; Katz, 1966). Followed by more demonstrations of chemical transmission using intracellular recording technique with glass microelectrodes, the possibility that electrical synapses exist fell from favor, and chemical transmission prevailed as the sole explanation for synaptic transmission in the central nervous system. However, soon electrical synapses were also discovered in invertebrates such as mantid shrimp (cardiac ganglion neurons) (Watanabe, 1958) and the crayfish (giant fibre system) (Furshpan and Potter, 1957), as well as in lower vertebrates (spinal cord neurons of pufferfish) (Bennett et al., 1959).
Following these early findings, researchers slowly revealed electrical coupling in a number of brain regions. In mammals, evidence of electrical synapses was reported in the mesencephalic nucleus of cranial nerve V (Hinrichsen, 1970; Baker and Llinas, 1971), the vestibular nucleus (Korn et al., 1973; Wylie, 1973), and the inferior olivary nucleus (Llinas et al., 1974). However, the incidence of electrical synapses still appeared to be very low compared to chemical synapses, and the functional significance of electrical synapses in mammals, although not in cold-blooded animals, was long underestimated.
In spite of their potential importance, the distribution and function of electrical synaps es in the cerebral cortex and other central regions remained elusive because of technical difficulties in directly assessing their existence. After decades of inattention, investigators slowly overcame the difficulties with technical developments in electrophysiology and cell visualization in living brain slices. By recording intracellularly from identified neighboring cells simultaneously, electrical synapses between inhibitory interneurons in the mammalian neocortex were discovered (Galarreta and Hestrin, 1999; Gibson et al., 1999). After this finding, a plethora of studies revealed that electrical synapses interconnect a wide variety of types of neurons throughout the mammalian brain. Some examples of electrically coupled neurons include GABAergic neurons in neocortex, hippocampus, striatum, and thalamus (Galarreta and Hestrin, 1999; Gibson et al., 1999; Koos and Tepper, 1999; Hormuzdi et al., 2001; Landisman et al., 2002), excitatory neurons such as the mitral cells in the olfactory bulb (Christie et al., 2005), AII amacrine cells and many other neurons in retina (Bloomfield and Volgyi, 2009) and inferior olivary
neurons (Long et al., 2002), as suggested earlier, and some neurons of the suprachiasmatic nucleus (Long et al., 2005).
Molecular basis of electrical synapses Electrical synapses are thought to be mediated by neuronal gap junctions, although other nonsynaptic mechanisms (ephatic mechanisms, field potential effects) have been suggested as alternative ways coupling neurons electrically (Jefferys, 1995). Gap junctions are specialized cell–cell contacts found only between eukaryotic cells, which can mediate electrical and metabolic coupling by allowing the passive movement of ions as well as small molecules such as nutrients, metabolites, and second messengers.
Gap junctions are composed of clusters of intercellular channels that are coextensive across regions of apposing membranes of coupled cells (Unger et al., 1999). The intercellular channel is formed by the end-to-end binding
of two oligomers (hemichannels) termed “connexons”, each of which is a hexameric complex of connexin subunit proteins (Fig. 1). Each connexin subunit is thought to traverse
the membrane bilayer four times, placing the NH 2 - and COOH- termini
on the cytoplasmic membrane surface. So far, 20 connexin genes have been described in the mouse and 21 in the human genome (Sohl et al., 2005). Connexin subtypes are named according to their predicted molecular weights (e.g. Cx36). Recently a new family of gap junction-forming proteins called pannexins were discovered in mammals (Bruzzone et al., 2003). Pannexins have strong sequence similarity to innexins, the intervertebrate gap junction proteins; curiously, connexins have no sequence similarity to either pannexins or innexins. Connexin-based gap junction
channels have unitary conductances, gating properties, and permeabilities as other ion channels. These properties vary widely between connexin subtypes, providing functional diversity between different gap junctions.
Gap junctions connecting neurons can function as electrical synapses. Gap junction- m ediated exchange of intracellular molecules is a common phenomenon found in many organs throughout the body. Several connexin subtypes are reported to exist in the brain, but only 5 connexins, Cx36, Cx45, Cx57, Cx30.2, Cx50, have consistently been suggested to be expressed in mammalian neurons. Among the five neuronal connexins, Cx57 and Cx50 expression appears to be limited only to horizontal cells in retina. Cx45 is expressed in many non-neuronal tissues and in neurons of the brain and retina. Although it has been suggested that Cx45 forms heterotypic gap junctions (i.e. gap junctions in which each connexon is comprised of different connexins; Fig. 1) between AII amacrine cells and bipolar cells, whether Cx45 forms neuronal homotypic gap junctions has not been clearly demonstrated yet (Maxeiner et al., 2005). Recently Cx30.2 was also suggested to be a neuronal connexin (Kreuzberg et al., 2008). Gap junction channels formed from Cx30.2 and Cx36 appear to have uniquely small unitary conductances.
Cx36 is the most extensively studied neuronal connexin, and evidence suggests that it is the m ajor neuronal gap junction protein in mammals. Cx36 is widely expressed throughout the brain, and its expression is almost exclusively neuronal (Condorelli et al., 1998). Single Cx36 channels have the lowest unitary conductance (~15pS) and the least voltage dependence of all the connexin channel subtypes that have been tested (Srinivas
et al., 1999). Studies of Cx36 knockout mice have implied that Cx36 underlies many electrical synapses, including those in neocortical inhibitory interneurons (Deans et al., 2001; Blatow et al., 2003), thalamic reticular neurons (Landisman et al., 2002), hippocampal interneurons (Hormuzdi et al., 2001), inferior olive neurons (Long et al., 2002), olfactory bulb mitral cells (Christie et al., 2005), retinal AII amacrine cells(Deans et al., 2002) and neurons of the suprachiasmatic nucleus (Long et al., 2005). So far, there is no strong evidence implicating any connexins except Cx36 in the electrical synapses of the mammalian brain.
Gap junction-mediated electrical synapses provide a fast, simple means of cell-to-cell communication, compared to chemical synapses. In early studies of invertebrates, electrical transmission seemed to mediate two different functions: 1) transmitting excitation from an active axon to a postsynaptic cell by monodirectional transmission (due to rectification in the junctional membrane), which could be quite similar at electrical and chemical synapses, and 2) synchronizing the activity of cell bodies by bidirectional transmission (Bennett, 1997). All electrical synapses observed so far in the mammalian brain provide bidirectional transmission, and may primarily serve to synchronize neurons (Connors and Long, 2004). In addition to their function as electrical synapses, neuronal gap junctions may provide metabolic biochemical coupling.
Electrical synapses in the developing brain In the context of development, there is a long-standing hypothesis that neuronal gap junctions are important for the maturation of neural circuits (Westerfield and Frank, 1982;
Kandler and Katz, 1995; Roerig and Feller, 2000), including the formation of neural circuits involving chemical synapses. In the vertebrate central nervous system there is substantial evidence for early gap junctional coupling among neuronal types, such as pyramidal neurons of the cerebral cortex, cortical subplate neurons, and spinal cord motor neurons (Fulton et al., 1980; Westerfield and Frank, 1982; Connors et al., 1983; Lo Turco and Kriegstein, 1991; Peinado et al., 1993; Kandler and Katz, 1995; Chang and Balice- Gordon, 2000; Mentis et al., 2002; Montoro and Yuste, 2004; Dupont et al., 2006). In those regions there is extensive dye-coupling between neurons during early circuit formation periods, and neurons generally uncouple as animals mature. The role of this coupling in the formation of spontaneously active domains in developing neural tissue has been suggested. In brain slices prepared from the early postnatal rat neocortex, for example, neuronal domains occur as spontaneous, locally restricted intercellular calcium waves (Yuste et al., 1992, 1995). Propagation of these calcium waves appeared to be mediated by gap junctions and not by chemical synaptic transmission. However, the molecular basis of this coupling in immature cortex has not been studied, and coupling has not been demonstrated by paired recording. In some cases, disruption of neuronal gap junctions appears to induce developmental abnormalities. For example, reducing electrical coupling between motor neurons accelerated neuromuscular synapse elimination in mice (Personius et al., 2007). Recent work on the olfactory bulb showed that Cx36-dependent electrical synapses between mitral cells were necessary for proper development of a glutamatergic synaptic circuit (Maher et al., 2009). However, in most cases it is still not clear why neuronal gap junctions are necessary and how they contribute to development or maturation of neural circuits.
Electrical synapses in the mature brain While some neuronal gap junctions are down-regulated after the early postnatal period, many neuronal types remained electrically coupled throughout juvenile and adult life. Here I describe a few important examples.
Retina: Electrical synapses are very common and important in retina. Electrical synapses have been found between most neuronal types in retina. In some cases they connect cells of the same type: cone photoreceptor cells (DeVries et al., 2002), AII amacrine cells (AII-AII amacrine cell) (Veruki and Hartveit, 2002a), horizontal cells (Hombach et al., 2004), and many types of ganglion cells (Hidaka et al., 2004); in other cases they connect cells of different types: AII amacrine cell-to-ON cone bipolar cell (Veruki and Hartveit, 2002b). The electrical synapses in retina usually have very high coupling strength (Veruki and Hartveit, 2002b).
Among many functions, electrical synapses play an essential part in the rod pathways of the retina (Deans et al., 2002). In the primary rod pathway, rod photoreceptors form glutamatergic synapses onto rod bipolar cells, which carry the signals radially to the inner retina and contact AII amacrine cells. The AII amacrine cells, in turn, form inhibitory synapses with OFF cone bipolar cell axons, and electrical synapses with ON cone bipolar cell axons. In addition, gap junctions between rod and cone photoreceptors provide an alternative ‘secondary’ rod pathway for the transmission of scotopic signals. Deletion of Cx36 disrupts both AII-bipolar and rod-cone gap junctions, resulting in the loss of
signaling in the primary and secondary rod pathways, respectively (Deans et al., 2002; Bloomfield and Volgyi, 2009). Neuronal gap junctions between horizontal cells, which depend on Cx57 (Hombach et al., 2004) or Cx50, provide extremely efficient lateral spread of visual signals. Regulation of the conductance of this electrical synapse by dopamine or nitric oxide is thought to be involved in controlling receptive field size and local contrast detection (Bloomfield and Volgyi, 2009). Cx36-dependent AII-AII amacrine cell gap junctions act in a similar way. AII-AII cell coupling serves to sum synchronous signals and subtract asynchronous noise, thereby preserving the high sensitivity of signals carried by the primary rod pathway with its modulation by light intensity (Bloomfield and Volgyi, 2004). The functions of electrical synapses between many other retinal neuron types, such as ganglion cells, are unknown (Bloomfield and Volgyi, 2009).
GABAergic inhibitory interneurons: GABAergic interneurons in the cereberal cortex, thalamus, striatum and cerebellum are extensively interconnected by electrical synapses. Many different types of cortical inhibitory interneurons have been shown to be electrically coupled, including fast-spiking (FS) neurons in barrel cortex layer 4 (Gibson et al., 1999), low threshold-spiking (LTS) cells in barrel cortex layer 4 (Gibson et al., 1999), FS cells in visual cortex layer 2/3 (Galarreta and Hestrin, 1999), CB1 receptor- expressing interneurons in cortical layer 2/3 (Galarreta et al., 2004), late-spiking (LS) neurons in layer 1 (Chu et al., 2003), multipolar bursting (MB) interneurons (Blatow et al., 2003) and hippocampal interneurons in stratum lacunosum-moleculare of CA1 (Zsiros and Maccaferri, 2005). The strengths of coupling between these cortical
interneurons is usually fairly strong, and coupling persists until adult ages (Galarreta and Hestrin, 2002). FS cells, LTS cells in barrel cortex, and MB cells were demonstrated to be Cx36-dependent (Deans et al., 2001; Blatow et al., 2003). In immunostaining studies of cerebral cortex, scattered expression of Cx36 protein is selectively localized to interneurons (Degen et al., 2004; Fukuda et al., 2006). Electrical synapses were found between GABAergic projection neurons in thalamic reticular nucleus, and between GABAergic local interneurons of striatum and cerebellum (Koos and Tepper, 1999; Mann-Metzer and Yarom, 1999).
In neocortex and thalamus, it has been shown that electrical synapses can m ediate strongly correlated spiking and subthreshold membrane potential fluctuations between inhibitory interneurons (Beierlein et al., 2000; Long et al., 2004). Electrical coupling is a major mechanism of synchronized oscillations among many types of neurons (Draguhn et al., 1998; Connors and Long, 2004; Traub et al., 2004). In Cx36 knockout mice, gamma rhythms are reduced in hippocampus (Hormuzdi et al., 2001; Buhl et al., 2003), and synchrony of theta range (4-7 Hz) oscillations mediated by LTS neurons in barrel cortex is impaired (Beierlein et al., 2000; Deans et al., 2001). However, the roles of electrical synapses in information processing and in behavior itself are very poorly understood. Interestingly, Cx36 knockout mice were reported to have memory deficits for novel object recognition (Frisch et al., 2005), suggesting an important role of electrical synapses in this cortical function.
Inferior olive: The inferior olive (IO) is another important area that has abundant electrical synapses. Many IO neurons generate large, spontaneous, subthreshold fluctuations of membrane potential. In Cx36 KO mice, such spontaneous rhythms still exist but synchrony of the rhythms across neurons is abolished (Long et al., 2002). It has been suggested that the synchrony of IO neurons is important for fine motor coordination and motor learning (Placantonakis et al., 2004; Van Der Giessen et al., 2008).
One of the most consistent conclusions from studies of electrically coupled neurons is that electrical synapses are an efficient mechanism for synchronizing neural activity, or transmitting or spreading signals. The functions of electrical synapse-mediated synchronization should be considered within the context of each neuronal circuit, based on the role, connectivity, and properties of each electrically coupled neural population.
II. THE SYNAPTIC ORGANIZATION OF THE THALAMUS My goals in this dissertation were to characterize the organization and functions of electrical synapses in the thalamus. Here I review previous work about electrical synapses in the thalamus, and thalamic circuitry and organization.
Electrical synapses in the thalamus Because Cx36 is the dominant neuronal gap junction protein in the brain, its expression levels provide strong clues about the distribution of electrical synapses. Most thalamic relay nuclei show relatively low Cx36 mRNA expression, suggesting that electrical synapses are weak or absent in those nuclei (Condorelli et al., 2000; Deans et al., 2001;
Landisman et al., 2002; Liu and Jones, 2003; Degen et al., 2004). For that reason, there has been relatively little work testing whether thalamic relay neurons are electrically coupled. However, several intriguing studies have suggested the existence of electrical coupling among thalamic relay neurons. First, anatomical data imply that another neuronal connexin, Cx45, is expressed in thalamic relay neurons and that this expression persists into adulthood (Maxeiner et al., 2003; Sohl et al., 2005). Second, intracellular recordings of spikelets and oscillatory activity resistant to blockade of chemical synapses have been interpreted as evidence for gap junction-dependent synchronization of thalamic relay neurons (Hughes et al., 2002; Hughes et al., 2004). Because relay neurons rarely make excitatory chemical synapses with other relay neurons, the possibility that they may interact via electrical coupling is quite interesting. However, the evidence for electrical synapses among relay cells is so far indirect, and paired intracellular recordings have not been used to test this possibility more definitively.
In contrast to the dearth of Cx36 expression in relay nuclei, it appears to be strongly expressed in the thalam ic reticular nucleus (TRN), as shown by situ hybridization, immunocytochemistry, and reporter gene patterns (Condorelli et al., 2000; Deans et al., 2001; Landisman et al., 2002; Liu and Jones, 2003; Degen et al., 2004). Consistent with these molecular and anatomical data, electrical synapses have been directly observed between neurons of the TRN using paired-cell recording (Landisman et al., 2002; Long et al., 2004). The strength of coupling between subpopulations of TRN neurons is high enough to induce correlated subthreshold membrane potential fluctuations and spiking activity in many cases. The strength and prevalence of TRN coupling is also remarkably
stable across postnatal development (Blethyn et al., 2008; Parker et al., 2009). However, the functional role of electrical coupling in TRN is poorly understood. I will discuss more about studies of the properties and functions of electrical synapses in TRN in the context of thalamic circuit and organization.
Thalamic circuitry and the TRN The thalamocortical network: The thalamus is a major subcortical structure that relays afferent information to neocortex, provides a pathway for corticocortical interaction, participates in state-dependent modulation of forebrain activity, and plays key roles in thalamocortical rhythm generation. It relays information from the sensory periphery (visual, somatosensory, auditory), and from lower brain centers such as the cerebellum (motor) and the mammillary bodies (limbic). Each sector is usually divided into two nuclei, based on the origin of the main driver inputs. ‘First-order’ relay nuclei receive driver inputs from ascending pathways, such as information from the sensory periphery, cerebellum, and mammillary bodies. Then it relays information to primary sensory areas, motor cortex, and cingulate cortex. ‘Higher-order’ relay nuclei receive their main driver inputs from layer 5 neurons of primary cortices as well as some ascending input (Guillery and Sherman, 2002). The functions of higher-order thalamic nuclei are not well understood. However, higher-order nuclei are usually connected to multiple cortical areas, and thus it has been suggested that they mediate corticocortical interactions (Guillery and Sherman, 2002).