Ординатура / Офтальмология / Английские материалы / Eye, Retina, and Visual System of the Mouse_Chalupa, Williams_2008

show clear signs of segregation, and by natural eye opening (P12–P14), they are well segregated and resemble the pattern found in the adult.

In recent years, the cholera toxin β subunit (CTB) has been used extensively as an effective and reliable anterograde tracer (Angelucci et al., 1996; Jaubert-Miazza et al., 2005; Muir-Robinson et al., 2002; Torborg and Feller, 2004; Torborg et al., 2005). When CTB conjugated to different fluorescent dyes is injected into the eyes, retinal projections from both eyes can be visualized simultaneously in a single section of the LGN. This allows the spatial extent for crossed and uncrossed projections to be estimated and the degree of overlap that exists between them to be captured (see figure 34.2). For these types of labeling experiments, estimates of spatial extent are based on the density of pixels that express a given fluorescence intensity above some threshold value (figure 34.2C). At P3, uncrossed projections occupy about 60% of the LGN and overlap with crossed ones by as much as 57%. By P7, the uncrossed projections are receding but still occupy about 25% and share 20% of LGN with crossed projections. By P12–P14, the time of natural eye opening, uncrossed projections take on the adult profile, occupying around 12% of LGN and sharing little (<2%), if any, territory with crossed projections.

The difference between unsegregated and segregated retinal projections is better understood by analyzing the fluorescence intensity of individual pixels (figure 34.2B and C; see chapter 28). These measures have advantages over those that rely on spatial extent because they provide an unbiased or threshold-independent index of segregation (Torborg and Feller, 2004). Scatterplots of pixel intensity underscore a major difference between unsegregated and segregated patterns (see figure 34.2B). For these functions, each point represents a pixel in which the fluorescence intensity of the uncrossed projection is plotted against the intensity of the crossed one. When uncrossed and crossed projections overlap, pixel intensities are high for both and show a positive correlation. However, when projections from the two eyes segregate, pixel intensities show an inverse relationship and a negative correlation. Pixel intensity can also be expressed as the logarithm of the ratio fluorescence intensities (e.g., R = log10 FI/FC) (Torborg and Feller, 2004). These ratio or R values can be plotted as a frequency distribution (see figure 34.2B). At young postnatal ages, R distributions tend to be narrow and unimodal because the majority of pixels have intensity values that are not dominated by one projection or the other. In contrast, at older ages, when eyespecific patterns are well developed, distributions become skewed, wide, and bimodal. The variance of R distributions can be used statistically to compare the relative widths of distributions across mice of different ages (see figure 34.2C; Jaubert-Miazza et al., 2005) or to assess differences between wild-type mice and transgenic strains (Torborg and Feller,

2004; Torborg et al., 2005; see chapter 28). In the case of retinogeniculate development, a marked increase in variance occurs between P3 and P10–P12. Interestingly, this time course parallels the changes that occur in spatial extent (see figure 34.2C). Taken together, they indicate that axon segregation stabilizes just prior to the period of natural eye opening. In the case of transgenic mouse models, application of these techniques has proved to be extremely powerful because it allows one to assess segregation independent of other aspects of visual map formation and refinement (Torborg and Feller, 2004; Torborg et al., 2005; see chapter 28).

The mechanisms underlying eye-specific segregation have been the topic of intense study, and transgenic mouse models have been particularly fruitful in this area (see chapters 28 and 35). Briefly, what has emerged is that molecular cues, in the form of the ephrin family of receptor tyrosine kinases and their cell surface–bound ligands, the ephrin As, play an essential role in the establishment of topographic maps and the location of eye-specific modules (Pfeiffenberger et al., 2005). However, the refinement of eye-specific segregation into nonoverlapping domains is due to the high-frequency burst discharges (Torborg and Feller, 2005; Torborg et al., 2005; see also chapter 28) associated with the wavelike patterns of spontaneous retinal activity that prevail during early postnatal life (Demas et al., 2003). At least two types of spontaneous retinal activity seem involved, an early phase (P0–P8) of cholinergic transmission that contributes to largescale rearrangements of eye-specific territories and a late phase (P10–P14) involving glutamate signaling that drives local patterns of segregation (Muir-Robinson et al., 2002). Moreover, it appears that late phases of spontaneous retinal activity are important to maintain and stabilize newly established eye-specific patterns (Demas et al., 2006). In nob mutant mice, known for a defect at the rod–bipolar cell synapse, retinal axons actually desegregate following the onset of abnormal spontaneous retinal activity close to the time of natural eye opening. Finally, a recent electrophysiological report shows that newly established adultlike connections continue to remain labile well after the period of anatomical segregation. Indeed, they show a remarkable sensitivity to patterned visual activity (Hooks and Chen, 2006; see also chapter 35). Visual deprivation 1 week after natural eye opening leads to synaptic weakening and the recruitment of additional retinal inputs. Whether these finescale changes in connectivity can be detected with anatomical labeling techniques remains to be tested.

Structural composition of the developing retinogeniculate pathway

As in most rodent species, the LGN in the mouse is a beanshaped nucleus that resides in the dorsolateral aspect of the

thalamus. In Nissl-stained material, the nucleus is a homogeneous structure with cytoarchitectural boundaries that separate it from the ventral basal complex, medial geniculate nucleus, the intrageniculate leaflet, and the ventral geniculate nuclei (see figure 34.1A), although “hidden laminae” do exist in the form of eye-specific patterns of retinal innervation and as a biochemically distinct dorsolateral shell that contains small calbindin-positive cells presumed to receive input from the superior colliculus (Grubb and Thompson, 2004; Reese, 1988).

In the developing LGN, neurons differentiate between E10 and E13 (Angevine, 1970). By E16, the nucleus is visible with a Nissl stain and composed of a narrow strip of cells along the dorsolateral edge of the thalamus (Godement et al., 1984). Between E18 and early postnatal days, the LGN is readily distinguished from other adjacent nuclei (Godement et al., 1984; Jaubert-Miazza et al., 2005) and resembles the adult profile (see figure 34.1A).

The cellular composition of mouse LGN is similar to that of higher mammals. There are two basic cell types, thalamocortical relay cells and interneurons. In rodents, the LGN is the only sensory nucleus of thalamus that contains intrinsic interneurons. Both cell types receive retinal input, but only relay cells have axons that exit LGN and project to the visual areas of cortex. Relay cells are excitatory and utilize glutamate as a neurotransmitter, while intrinsic interneurons are inhibitory and use γ-aminobutyric acid (GABA). The morphology of neurons in the rodent LGN has been examined from Golgi-impregnated material (Grossman et al., 1973; Parnavelas et al., 1977; Rafols and Valverde, 1973) and more recently from intracellular fills performed during in vitro recording experiments ( Jaubert-Miazza et al., 2005; Ziburkus and Guido, 2006). Overall, these studies show that relay cells have type I or Class A morphology, which consists of a thick unbranched axon, large, spherical somata, and complex multipolar dendritic arbors that are studded with spinelike protrusions (see figure 34.1C). Although relay cells display some morphological variation in soma size and dendritic complexity and orientation, it is not clear whether they can be grouped into distinct subclasses, as observed in other species such as the cat (X, Y, and W cells) or primate (M, P, and K cells). Interestingly, most relay cells in the mouse LGN show linear spatial summation in their visual response, a feature that is consistent with X cells (Grubb and Thompson, 2003). Interneurons, with type II or Class B morphology, have a smaller spindle-shaped soma and just a few long sinuous dendrites oriented in a dorsoventral plane that follows the boundaries of the LGN (see figure 34.1C). These cells lack a conventional axonal projection, but embedded in their dendritic trees are thin axonlike processes with terminal swellings. Relay cells and interneurons can be distinguished at early postnatal ages by using these morphological criteria or with immunocytochemical markers

( Jaubert-Miazza et al., 2005) that stain for the neurofilament protein SMI-32 (relay cells) or the synthesizing enzyme for GABA (interneurons). Compared to their adult counterparts, developing cells have smaller somata and a simpler dendritic architecture comprised of fewer arbors with short branches (see figure 34.1C; Jaubert-Miazza et al., 2005; Rafols and Valverde, 1973). A detailed morphometric analysis in the developing mouse is presently lacking. However, in the rat, relay cells undergo two notable “growth spurts” characterized by marked changes in soma size and dendritic complexity (Parnavelas et al., 1977). One occurs at P4–P6, and the other occurs close to the time of eye natural opening (P12–P14). Interneurons seem to develop more slowly and gradually, and do not reach their mature state until P18. What remains to be examined is how the dynamic changes in dendritic complexity are coordinated with the retraction of retinal projections to form adult patterns of connectivity.

In adult mouse, the synaptic arrangements of retinal axons, relay cells, and interneurons have a highly conserved and stereotypic ultrastructure (see figure 34.1D; Rafols and Valverde, 1973). At the EM level, retinal (RLP) terminals make synaptic contacts with the proximal portions of relay cell dendrites. Interneurons have a conventional synaptic output via an axon that makes contact (F1 profile) with a dendrite of a relay cell. They also have presynaptic terminals (F2) that arise from their dendrites. The latter form part of a specialized triadic arrangement that involves three synapses: one between a retinal terminal and an F2 terminal, one between the same retinal terminal and the dendrite of a relay cell, and one between the F2 terminal and same dendritic appendage of a relay cell. This triadic arrangement is found within a glomerulus and provides the basis for excitatory responses in relay cells and interneurons, as well as feedforward inhibition between interneurons and relay cells. Although synapses can be readily identified at early ages (P7), it remains difficult to categorize them on the basis of their ultrastructure (see figure 34.1D; Slusarczyk et al., 2006). Pale mitochondria characteristic of RLP terminals can be identified, but vesicles are sparsely distributed. Further, it becomes even more difficult to distinguish inhibitory profiles even when GABA-labeled gold particles are used to identify inhibitory elements. By P14, all classes of terminals become recognizable and resemble those found in the adult (see figure 34.1D).

Functional composition of the developing retinogeniculate pathway

The morphological changes in cellular composition and synaptic morphology noted in the previous section are accompanied by several functional changes that occur in the membrane properties, the receptor composition of synaptic

responses, and patterns of connectivity. The membrane properties of rodent LGN cells have been studied extensively using in vitro intracellular recordings in thalamic slices (Crunelli et al., 1988; Jaubert-Miazza et al., 2005; MacLeod et al., 1997). Adult LGN cells are equipped with several voltage-gated ion conductances. When active, they give rise to complex firing patterns that greatly affect the gain of retinogeniculate signal transmission (figure 34.3A). The most notable ones include a mixed cation conductance (IH) that results in a strong inward rectification during membrane hyperpolarization; a low-threshold, T-type Ca2+ conductance that produces a large triangular depolarization (LT Ca2+ spike) and burst firing when the membrane is repolarized from a hyperpolarized state; a transient A-type K+ conductance that leads to an outward rectification during membrane depolarization, thereby introducing a significant delay in spike firing; and a family of other K+ conductances that become active during repetitive spike firing, producing after-hyperpolarizing (AHPs) responses between spikes and frequency accommodation.

Recordings done at prenatal ages indicate that thalamic cells exhibit inward and outward rectifying responses to current injection as early as E17, but rudimentary action potentials are not evident until P0 (MacLeod et al., 1997). Between P0 and P7, action potentials continue to mature, showing decreases in spike width and increases in amplitude. Sustained levels of depolarization can evoke spike trains that exhibit delays in spike firing and spike frequency accommodation. Membrane depolarization from hyperpolarized states also evokes large low-threshold Ca2+ spikes and burst firing. Additionally, high-threshold Ca2+-mediated “spikes” can be elicited at depolarized levels. This response is transient and is observed only at early postnatal ages. By P7–P14, the full complement of active membrane properties is evident, and the voltage responses resemble those seen in the adult (Jaubert-Miazza et al., 2005; MacLeod et al., 1997). Thus, during early postnatal life, relay neurons possess many of the requisite membrane properties needed to receive, respond, and transmit afferent patterns of stimulation. In fact, developing cells are remarkably efficient at doing this. In a novel in vitro preparation in which the eyes, optic pathways, and part of the diencephalon remained intact, it was shown that spontaneous retinal activity generates periodic bursts of action potentials that are temporally coupled to the frequency of the retinal waves (Mooney et al., 1996). Thus, functional if morphologically immature synapses are present even at the earliest stages of postnatal development. However, despite the early presence of synaptic responses, the receptor composition of postsynaptic activity undergoes marked changes during the period of retinogeniculate axon segregation.

In vitro intracellular recordings made in mature LGN cells reveal that retinal stimulation evokes an excitatory post-

synaptic potential (EPSP) that is typically followed by inhibitory postsynaptic (IPSP) activity (figure 34.3B; Blitz and Regehr, 2005; Jaubert-Miazza et al., 2005; Ziburkus et al., 2003). This EPSP/IPSP sequence reflects the pattern of triadic synaptic arrangements described earlier. Retinal axons make excitatory connections with relay cells. They also have collateral branches that make excitatory connections with interneurons, which in turn form feedforward inhibitory connections with relay cells. IPSP activity serves many functions in LGN, including shaping the centersurround antagonistic receptive field (RF) structure of relay cells, regulating tonic and burst firing patterns, and establishing the overall gain of signal transmission.

As the ultrastructure suggests, these inhibitory aspects of synaptic circuitry are not fully developed at birth. At this time, the majority of synaptic responses are purely excitatory and involve the coactivation of two ionotropic glutamate receptor subtypes, AMPA and NMDA (Chen and Regehr, 2000; Ziburkus et al., 2003). AMPA receptors mediate an early fast-rising, brief depolarization, whereas NMDA receptors contribute to a slower and longer form of excitation. The unique aspect of NMDA activity is its voltage dependency. That is, at resting membrane potentials, NMDA receptors are inactive due to a voltage-dependent blockade of the channel pore by Mg2+ ions. However, in the presence of strong postsynaptic depolarization, one that is likely brought about by the activation of AMPA receptors, the Mg2+ blockade is relieved and NMDA activation ensues. In many developing sensory structures, the voltage dependency of NMDA receptor activation serves as a “co-incident detector,” allowing for an influx of Ca2+ ions during periods of strong postsynaptic activation. The activity-depen- dent influx of Ca2+ triggers long-term changes in synaptic strength and a series of signaling cascades that leads to the eventual consolidation of active connections, as well as the elimination of less active ones. At early ages, the excitatory responses of LGN cells are composed largely of NMDA activity (Chen and Regehr, 2000). However, during postnatal development this prevalence gradually wanes, and AMPA responses become more prominent. In fact, by P23 there is a fourfold increase in the AMPA to NMDA ratio, which contributes to the excitatory response. During this transition, NMDA responses also show reduced activation times, possibly reflecting a shift in subunit composition from NR2B to NR2A. During retinogeniculate development, NMDA responses may facilitate synapse strengthening (Chen and Regehr, 2000; Mooney et al., 1993) or contribute to the activation of other nonsynaptic, voltagedependent events.

Although it has yet to be tested in the mouse LGN, it seems unlikely that NMDA activation leads directly to the activity-dependent refinement of eye-specific connections. Correlated firing between RGCs and LGN relay cells per-

Figure 34.3 Electrophysiological and synaptic properties of developing LGN cells. Neonatal responses are shown on the left, and adult responses are shown on the right. A, Examples are shown of the voltage responses (bottom traces) to current pulse injections (top traces). Even at early postnatal ages, the voltage responses of LGN reflect the activation of a number of voltage conductances. Membrane hyperpolarization activates a mixed cation conductance (H) that results in inward rectification. Cessation of the hyperpolarizing pulses leads to the activation of a T-type Ca2+ conductance that produces a “rebound” triangular-shaped, low-threshold Ca2+ spike

(T) and burst firing (B). Membrane depolarization activates a transient A-type K+ conductance that results in outward rectification and delays spike firing. Large and sustained depolarizing pulses elicit repetitive spiking with frequency accommodation. The repolarization following between each spike activates additional K+ conductances (AHP). B. Examples of postsynaptic activity (EPSP

and IPSP) evoked by electrical stimulation of the optic tract. Initially evoked responses are excitatory and take the form of a longduration EPSP that contains a large NMDA component. At later ages, the EPSP is shorter in duration and typically followed by a pair of IPSPs. The short one occurs just after the EPSP and is mediated by GABAA receptors. The longer slower one follows the early IPSP and is mediated by GABAB receptors. C, Synaptically evoked plateau potentials in the developing LGN. Shown are responses evoked by single or repetitive stimulation of optic tract. At early ages, suprathreshold stimulation evokes a high-amplitude, long-duration plateau potential. This response is transient and rarely observed after P14. Even strong levels of stimulation fail to evoke such a response. Single and repetitive (25–100 Hz) are shown separately. All responses were recorded between −55 and −67 mV. (Adapted from Jaubert-Miazza et al., 2005.)

sists in the absence of NMDA receptor function (Mooney et al., 1996), and in other mammals, NMDA receptor blockade does not seem to interfere with eye-specific segregation (Smetters et al., 1994), although it may be involved in more subtle forms of anatomical remodeling, such as the refinement of ON and OFF sublaminae (Hahm et al., 1991).

During the time when the receptor composition of excitatory responses is changing, inhibitory responses begin to emerge, but such activity does not fully mature until P10 (Ziburkus et al., 2003). Inhibitory responses in LGN involve two types of GABA receptors (Crunelli et al., 1988; Ziburkus et al., 2003). The first to appear is an early, fast hyperpolarizing response that involves a Cl− conductance acting through the GABAA receptor subtype. These early fast GABAA- mediated IPSPs immediately follow EPSP activity and often curtail the late NMDA component of the excitatory response. A second type of IPSP appears near the end of the first week and involves a G protein-activated K+ conductance acting through a GABAB receptor subtype. It follows the GABAA IPSP and is slower and longlasting. Often, the slow decay of the GABAB response gives rise to a rebound LT spike and bursting.

Thus, the balance of excitatory and inhibitory responses develops at different rates, with excitatory ones maturing faster than inhibitory ones. The functional significance of this sequence is not clear, but the delayed onset of inhibitory activity may promote an increased level of excitatory postsynaptic events implicated in synaptic remodeling (e.g., NMDA and high-threshold voltage-gated Ca2+ channel activity) or help to mediate the transmission of strong and sustained levels of depolarization brought about by retinal wave activity (Lo et al., 2002; Mooney et al., 1996).

Finally, of notable significance is the discovery of a transient synaptic event that prevails during early postnatal life. Strong or repetitive retinal stimulation of retinal fibers evokes EPSPs that activate a high-amplitude, longlasting, slow-decaying depolarization (figure 34.3C). These “plateau potentials” have a voltage dependency and pharmacology consistent with the activation of high-threshold L-type (longlasting) Ca2+ channels (Kammermeier and Jones, 1998; Lo et al., 2002). L-type Ca2+ channels are found throughout the nervous system and have been implicated in a variety of cellular events, including the presynaptic release of transmitter substance, activity-dependent gene expression, cell excitability, synaptic plasticity, and cell survival (Lipscombe et al., 2004). In the rodent LGN, L-type channels are located primarily on somata and proximal dendrites (Budde et al., 1998; Jaubert-Miazza et al., 2005). Indeed, these channels reside in the same vicinity as retinal terminals (Rafols and Valverde, 1973; Slusarczyk et al., 2006), thus placing them in an ideal location to amplify retinally evoked events. Plateau potentials in LGN are encountered far more frequently between P0 and P7 (90%), then decline gradually

with age so, that by P18–P21 they are rarely observed (Jaubert-Miazza et al., 2005).

At least two factors have been identified that contribute to the developmental regulation of plateau potentials (Ziburkus et al., 2003). First, the high degree of retinal convergence, coupled with heightened NMDA activity, favors the spatial and temporal summation of EPSPs. These synaptic events lead to sustained levels of depolarization and thereby greatly increase the likelihood that high-threshold L-type channels are activated. Interestingly, the episodic barrages of retinal EPSPs associated with the transmission of spontaneous retinal waves (Mooney et al., 1996) are well suited to activate L-type-mediated plateau potentials. Indeed, repetitive stimulation of retinal afferents in a manner that approximates the high-frequency discharge of retinal waves triggers robust plateau-like activity (Lo et al., 2002). Second, the density of L-type Ca2+ channels found among LGN cells changes with postnatal age. Using an antibody that recognizes the pore forming α1C subunit of the L-type channel, Jaubert-Miazza et al. (2005) showed that expression peaks between P0 and P7 and then gradually declines, so that by P28 there is a fourfold reduction in the density of labeled cells. Finally, it is worth noting that L-type activity recorded in LGN is identical to the synaptically evoked plateau potentials recorded in developing neurons of the rodent superior colliculus (Lo and Mize, 2000) and brainstem trigeminal complex (Lo and Erzurumlu, 2002). It also bears a similarity to the high-threshold Ca2+ activity reported in some brainstem nuclei (Rekling and Feldman, 1997), spinal cord (Kien and Eken, 1998), and invertebrate motor neurons (Dicaprio, 1997). Thus, this event may reflect a highly conserved mechanism by which cells can acquire large amounts of Ca2+ in an activity-dependent manner.

Patterns of synaptic connectivity in the developing lateral geniculate nucleus

The age-related changes in postsynaptic receptor and channel function are accompanied by major modifications in the pattern of synaptic connectivity. In vivo recordings in mature mice indicate that relay cells receive monocular input (Grubb and Thompson, 2003). However, at early postnatal ages, when the inputs from the two eyes have overlapping projections, a high incidence of binocular responses is observed (Jaubert-Miazza et al., 2005; Ziburkus and Guido, 2006). In an in vitro explant preparation in which LGN circuitry and large segments of each optic nerve remained intact, separate and distinct EPSPs could be evoked by stimulating either optic nerve (figure 34.4A). In addition to binocular responses, another transient feature in synaptic connectivity is the high incidence of synaptic responses, reflecting the convergence of multiple RGC inputs onto a single LGN cell (Chen and Regehr, 2000; Jaubert-Miazza

Figure 34.4 Synaptic connectivity in the developing LGN. A, Examples of EPSPs evoked by the separate stimulation of either the contralateral (left) or ipsilateral (right) optic nerve. Shown is an example of a binocular response at P8 and a monocular contralateral eye response at P19. Responses at P8 have a spike riding the peak of the EPSP. Binocular responses are frequently encountered before the period of natural eye opening. B, Examples of synaptic responses at P8 and P19 evoked by stimulating optic tract (OT) fibers at progressively higher levels of stimulation. Below each response is the corresponding amplitude by stimulus intensity plot.

Each step is numbered and corresponds to the recruitment of a separate input. At P8, a progressive increase in stimulus intensity leads to nine incremental gradations in amplitude. At P19, increases in stimulus intensity produce only two steps. The scatterplot below plots estimates of the number of retinal inputs LGN cells receive as a function of age. Each point represents a single cell. There is an age-related decrease in retinal convergence. All responses were recorded between −60 and −67 mV. (Adapted from Jaubert-Miazza et al., 2005.)

et al., 2005; Ziburkus and Guido, 2006; see also chapter 35). Estimates of the number of synaptic inputs a relay cell receives can be obtained by gradually increasing the stimulus intensity applied to retinal fibers and then measuring the amplitude of the evoked response (figure 34.4B). In developing LGN cells, progressive increases in stimulus intensity often give rise to multiple stepwise increases in EPSPs (see figure 34.4B). These graded changes reflect the successive recruitment of active inputs innervating a single cell. In mature LGN cells tested with an identical protocol, EPSP amplitude shows far fewer steps. In general, these studies reveal a gradual reduction in retinal convergence with age (see figure 34.4B). During the first postnatal weeks, cells are reported to receive as many as 12–20 inputs (Chen and Regehr, 2000; Hooks and Chen, 2006; Jaubert-Miazza et al., 2005). Studies in the rat reveal that many developing cells receive up to 4–6 inputs from each eye (Ziburkus and Guido, 2006). Sometime between the second and third postnatal week, monocular responses prevail, and the number of inputs declines to 1–3, thus resembling the adult pattern of connectivity. During this pruning process, there are corresponding changes in synaptic strength. As weak inputs are eliminated, the remaining ones show as much as a 50-fold increase in synaptic strength (Chen and Regehr, 2000). Although not yet demonstrated in the developing mouse, these reported changes in retinal convergence may contribute to the postnatal maturation of RF properties. In the cat and ferret, immature RFs are binocular, quite large, irregularly shaped, and lack distinct ON and OFF subregions (Shatz and Kirkwood, 1984; Tavazoie and Reid, 2000). In contrast, mature RFs are monocular, much smaller, and have well-defined concentric center-surround organization.

The changes in synaptic connectivity described earlier persist after anatomical segregation is complete. Thus, there may be two phases of axonal remodeling: an early coarse phase, best captured by bulk anterograde labeling techniques, that involves eye-specific segregation, and another fine-scale functional form of refinement that involves the physiological strengthening of some synapses and elimination of others.

Potential mechanisms underlying the remodeling of retinogeniculate connections

While retinal waves are thought to play a prominent role in driving the refinement of retinogeniculate connections, the postsynaptic mechanisms responsible for the implementation of these changes remains unresolved. One likely mechanism is based on a Hebbian model of synaptic modification in which temporally correlated activity between preand postsynaptic elements leads to a strengthening (i.e., longterm potentiation, LTP) and consolidation of synapses, whereas asynchronous or absent activity results in synapse

weakening (i.e., long-term depression, LTD) and synapse elimination. Hebbian-based changes in synaptic strength have been noted in several developing sensory structures, including the mammalian neocortex and the optic tectum of Xenopus. In some instances they also seem to contribute to the structural refinement of connections (Malenka and Bear, 2004; Ruthazer and Cline, 2004). Long-term changes in synaptic strength may also be involved in stabilizing retinogeniculate connections. Clearly, retinal waves are sufficient to generate robust postsynaptic activity in LGN (Lo et al., 2002; Mooney et al., 1996), and the “nearest-neighbor, same-eye relations” underlying the spatiotemporal patterning of wave activity seems well suited for promoting Hebbian synaptic plasticity (Torborg and Feller, 2005; see also chapters 28 and 35). Developing LGN cells also exhibit long-term changes in synaptic strength (Mooney et al., 1993) although the exact nature and polarity of these changes await further testing.

Most models of activity-dependent synaptic plasticity rely on coincident detection via NMDA receptor activation. However, a role for NMDA in driving eye-specific refinement seems unlikely. Instead, a more plausible substrate is the L-type Ca2+ channel. Patterned retinal activity similar to retinal waves evokes longlasting L-type-mediated plateau potentials in developing LGN cells (Lo et al., 2002). L-type activity can induce long-term changes in synaptic strength in a number of structures, including the hippocampus (Magee and Johnston, 1997), developing superior colliculus (Lo and Mize, 2000), and trigeminal principal nucleus of the brainstem (Guido et al., 2001). Additionally, synaptically evoked Ca2+ influx through L-type channels is particularly effective at activating transcription factors that lead to the expression of genes linked to synaptic plasticity (Lonze and Ginty, 2002; West et al., 2002). One in particular involves the cAMP response element (CRE/CREB) transcription pathway (Dolmetsch et al., 2001; Mermelstein et al., 2000). In mouse LGN, CRE-mediated gene expression peaks during early postnatal life (Pham et al., 2001). Moreover, retinofugal projections of mutant mice that show reduced levels of L-type Ca2+ channel activity or decreased levels of CRE expression fail to segregate properly (Cork et al., 2001; Green et al., 2004; Pham et al., 2001).

Conclusion

The structural and functional composition of the retinogeniculate pathway undergoes a rapid and dramatic period of remodeling during the first 2 weeks of life (figure 34.5). At or near the time of birth, retinal axons have already innervated the LGN, and within the first few days of postnatal life they establish a coarse topographic map. However, eyespecific domains are not yet present, and retinal projections from the two eyes are diffusely organized, having overlap-

ping terminal fields in LGN. During the first postnatal week, LGN cells are growing in size and complexity and the ultrastructural features of synapses are taking shape. Functional synapses are present, and LGN cells possess the rudimentary membrane properties sufficient to fire bursts of action potentials at frequencies that match the temporal structure of retinal waves. Postsynaptic activity is largely excitatory and comprised of NMDA receptor activation. LGN cells are binocularly responsive, receiving multiple inputs from the two eyes. This high degree of retinal convergence, coupled with an elevated density of L-type Ca2+ channels, gives rise to plateau-like depolarizations. The large Ca2+ influx associated with L-type channel activation and plateau potentials may contribute to long-term changes in synaptic strength and signaling events involved in the

remodeling and stabilization of developing retinogeniculate connections.

After the first postnatal week, retinal inputs from the two eyes show clear signs of segregation, while the membrane properties and spike firing continue to mature. NMDA responses, the density of L-type Ca2+ channels, and plateau activity also begin to subside. Inhibitory activity emerges and grows stronger through the formation of a feedforward circuit involving intrinsic interneurons. Finally, before natural eye opening (P12–P14), retinogeniculate axon segregation is complete; there occur a loss of binocular responsiveness, a further reduction in retinal convergence, and a decline in the expression of L-type Ca2+ channels, which ultimately contributes to elimination of plateau potential activity. After 2 weeks of age, retinal waves disappear, and

visual activity provides the primary drive for newly established connections.

Although a great deal is known on the development of the mouse retinogeniculate pathway, a number of issues remained unresolved. For example, additional morphological information is needed on the growth and maturation of retinogeniculate synapses and the dendrites of relay cells. How are such changes regulated, how are they coordinated with retinal axon segregation, and to what extent do they contribute to the remodeling process? What factors govern whether certain retinal inputs are preserved and strengthened while others are weakened and eventually eliminated? Do such changes rely on retinal activity and a Hebbian form of synaptic plasticity? If so, what neural elements and signaling cascades are involved? These are but a few areas of inquiry, and further investigation is expected to shed light on the mechanisms underlying mammalian sensory development.

acknowledgments Work was supported by the National Eye Institute (grant no. EY12716). I thank Fu-Sun Lo, Martha Bickford, Erick Green, Lisa Jaubert-Miazza, Jeremy Mills, and Kim Bui for their contributions, as well as Tania Seabrook for editorial assistance.

REFERENCES

Angelucci, A., Clasca, F., and Sur, M. (1996). Anterograde axonal tracing with the subunit B of cholera toxin: A highly sensitive immunohistochemical protocol for revealing fine axonal morphology in adult and neonatal brains. J. Neurosci. Methods 65(1):101–112.

Angevine, J. B., Jr. (1970). Time of neuron origin in the diencephalon of the mouse: An autoradiographic study. J. Comp. Neurol. 139(2):129–187.

Blitz, D. M., and Regehr, W. G. (2005). Timing and specificity of feed-forward inhibition within the LGN. Neuron 45(6): 917–928.

Budde, T., Munsch, T., and Pape, H. C. (1998). Distribution of L-type calcium channels in rat thalamic neurones. Eur. J. Neurosci. 10(2):586–597.

Chen, C., and Regehr, W. G. (2000). Developmental remodeling of the retinogeniculate synapse. Neuron 28(3):955–966.

Cork, R. J., Namkung, Y., Shin, H. S., and Mize, R. R. (2001). Development of the visual pathway is disrupted in mice with a targeted disruption of the calcium channel beta(3)-subunit gene. J. Comp. Neurol. 440(2):177–191.

Crunelli, V., Haby, M., Jassik-Gerschenfeld, D., Leresche, N., and Pirchio, M. (1988). Cl− and K+ dependent inhibitory postsynaptic potentials evoked by interneurones of the rat lateral geniculate nucleus. J. Physiol. 399:153–176.

Demas, J., Eglen, S. J., and Wong, R. O. (2003). Developmental loss of synchronous spontaneous activity in the mouse retina is independent of visual experience. J. Neurosci. 23(7):2851– 2860.

Demas, J., Sagdullaev, B. T., Green, E., Jaubert-Miazza, L., McCall, M. A., Gregg, R. G., Wong, R. O., and Guido, W. (2006). Failure to maintain eye-specific segregation in nob, a

mutant with abnormally patterned retinal activity. Neuron 50 (2):247–259.

Dicaprio, R. (1997). Plateau potentials in motor neurons in the ventilatory system of the crab. J. Exp. Biol. 200(Pt. 12):1725– 1736.

Dolmetsch, R. E., Pajvani, U., Fife, K., Spotts, J. M., and Greenberg, M. E. (2001). Signaling to the nucleus by an L-type calcium channel-calmodulin complex through the MAP kinase pathway. Science 294(5541):333–339.

Dräger, U. C., and Olsen, J. F. (1980). Origins of crossed and uncrossed retinal projections in pigmented and albino mice. J. Comp. Neurol. 191(3):383–412.

Godement, P., Salaun, J., and Imbert, M. (1984). Prenatal and postnatal development of retinogeniculate and retinocollicular projections in the mouse. J. Comp. Neurol. 230(4):552–575.

Gomez-Ospina, N., Tsuruta, F., Barreto-Chang, O., Hu, L., and Dolmetsch, R. (2006). The C terminus of the L-type voltagegated calcium channel Ca(V)1.2 encodes a transcription factor. Cell 127(3):591–606.

Green, E., Bui, K., Mills, J., Shin, H., Gregg, R. G., and Guido, W. (2004). Anomalous retinal projections and altered L-type calcium channel expression in the LGN of calcium channel β3 subunit deficient mice. Soc. Neurosci. Abstr. 613.12.

Grossman, A., Lieberman, A. R., and Webster, K. E. (1973). A Golgi study of the rat dorsal lateral geniculate nucleus. J. Comp. Neurol. 150(4):441–466.

Grubb, M. S., Rossi, F. M. Changeux, J. P., and Thompson, I. D. (2003). Abnormal functional organization in the dorsal lateral geniculate nucleus of mice lacking the beta 2 subunit of the nicotinic acetylcholine receptor. Neuron 40(6):1161–1172.

Grubb, M. S., and Thompson, I. D. (2003). Quantitative characterization of visual response properties in the mouse dorsal lateral geniculate nucleus. J. Neurophysiol. 90(6):3594–3607.

Grubb, M. S., and Thompson, I. D. (2004). Biochemical and anatomical subdivision of the dorsal lateral geniculate nucleus in normal mice and in mice lacking the beta2 subunit of the nicotinic acetylcholine receptor. Vision Res. 44(28):3365– 3376.

Guido, W., Lo, F. S., and Erzurumlu, R. S. (2001). Synaptic plasticity in the trigeminal principal nucleus during the period of barrelette formation and consolidation. Brain Res. Dev. Brain Res. 132(1):97–102.

Hahm, J. O., Langdon, R. B., and Sur, M. (1991) Disruption of retinogeniculate afferent segregation by antagonists to NMDA receptors. Nature 351:568–570.

Hooks, B. M., and Chen, C. (2006). Distinct roles for spontaneous and visual activity in remodeling of the retinogeniculate synapse. Neuron 52(2):281–291.

Jaubert-Miazza, L., Green, E., Lo, F. S., Bui, K., Mills, J., and Guido, W. (2005). Structural and functional composition of the developing retinogeniculate pathway in the mouse. Vis. Neurosci. 22(5):661–676.

Kammermeier, P. J., and Jones, S. W. (1998). Facilitation of L-type calcium current in thalamic neurons. J. Neurophysiol. 79(1):410 – 417.

Kiehn, O., and Eken, T. (1998). Functional role of plateau potentials in vertebrate motor neurons. Curr. Opin. Neurobiol. 8(6):746– 752.

Lipscombe, D., Helton, T. D., and Xu, W. (2004). L-type calcium channels: The low down. J. Neurophysiol. 92(5):2633–2641.

Lo, F. S., and Erzurumlu, R. S. (2002). L-type calcium channelmediated plateau potentials in barrelette cells during structural plasticity. J. Neurophysiol. 88(2):794–801.

Lo, F. S., and Mize, R. R. (2000). Synaptic regulation of L-type Ca(2+) channel activity and long-term depression during refinement of the retinocollicular pathway in developing rodent superior colliculus. J. Neurosci. 20(3):RC58.

Lo, F. S., Ziburkus, J., and Guido, W. (2002). Synaptic mechanisms regulating the activation of a Ca2+ mediated plateau potential in developing relay cells of the LGN. J. Neurophysiol. 87(3):1175–1185.

Lonze, B. E., and Ginty, D. D. (2002). Function and regulation of CREB family transcription factors in the nervous system. Neuron 35(4):605–623.

MacLeod, N., Turner, C., and Edgar, J. (1997). Properties of developing lateral geniculate neurones in the mouse. Int. J. Dev. Neurosci. 15(2):205–224.

Magee, J. C., and Johnston, D. (1997). A synaptically controlled, associative signal for Hebbian plasticity in hippocampal neurons. Science 275(5297):209–213.

Malenka, R. C., and Bear, M. F. (2004). LTP and LTD: An embarrassment of riches. Neuron 44(1):5–21.

Mermelstein, P. G., Bito, H., Deisseroth, K., and Tsien, R. W. (2000). Critical dependence of cAMP response element-binding protein phosphorylation on L-type calcium channels supports a selective response to EPSPs in preference to action potentials. J. Neurosci. 20(1):266–273.

Mooney, R., Madison, D. V., and Shatz, C. J. (1993). Enhancement of transmission at the developing retinogeniculate synapse. Neuron 10(5):815–825.

Mooney, R., Penn, A. A., Gallego, R., and Shatz, C. J. (1996). Thalamic relay of spontaneous retinal activity prior to vision. Neuron 17(5):863–874.

Muir-Robinson, G., Hwang, B. J., and Feller, M. B. (2002). Retinogeniculate axons undergo eye-specific segregation in the absence of eye-specific layers. J. Neurosci. 22(13):5259–5264.

Parnavelas, J. G., Mounty, E. J., Bradford, R., and Lieberman, A. R. (1977). The postnatal development of neurons in the dorsal lateral geniculate nucleus of the rat: A Golgi study. J. Comp. Neurol. 171(4):481–499.

Pfeiffenberger, C., Cutforth, T., Woods, G., Yamada, J., Renteria, R. C., Copenhagen, D. R., Flanagan, J. G., and Feldheim, D. A. (2005). Ephrin-As and neural activity are required for eye-specific patterning during retinogeniculate mapping. Nat. Neurosci. 8(8):1022–1027.

Pham, T. A., Rubenstein, J. L., Silva, A. J., Storm, D. R., and Stryker, M. P. (2001). The CRE/CREB pathway is transiently expressed in thalamic circuit development and contributes to refinement of retinogeniculate axons. Neuron 31(3):409– 420.

Rafols, J. A., and Valverde, F. (1973). The structure of the dorsal lateral geniculate nucleus in the mouse: A Golgi and electron microscopic study. J. Comp. Neurol. 150(3):303–332.

Reese, B. E. (1988). “Hidden lamination” in the dorsal lateral geniculate nucleus: The functional organization of this thalamic region in the rat. Brain Res. 472(2):119–137.

Rekling, J. C., and Feldman, J. L. (1997). Calcium-dependent plateau potentials in rostral ambiguus neurons in the newborn mouse brain stem in vitro. J. Neurophysiol. 78(5):2483–2492.

Ruthazer, E. S., and Cline, H. T. (2004). Insights into activitydependent map formation from the retinotectal system: A middle of the brain perspective. J. Neurobiol. 59(1):134–146.

Shatz, C. J., and Kirkwood, P. A. (1984). Prenatal development of functional connections in the cat’s retinogeniculate pathway. J. Neurosci. 4:1378–1397.

Slusarczyk, A. S., Kucuk, C., Chomsung, R., Eisenback, M. A., Guido, W., and Bickford, M. E. (2006). Synaptic organization of the adult and neonatal mouse dorsal lateral geniculate nucleus.

Soc. Neurosci. Abstr. 241.3.

Smetters, D. K., Hahm, J., and Sur, M. (1994). An N-methyl-d aspartate receptor antagonist does not prevent eye-specific segregation in the ferret retinogeniculate pathway. Brain Res. 658: 168–178.

Tavazoie, S. F., and Reid, R. C. (2000). Diverse receptive fields in the lateral geniculate nucleus during thalamocortical development. Nat. Neurosci. 3(6):608–616.

Torborg, C. L., and Feller, M. B. (2004). Unbiased analysis of bulk axonal segregation patterns. J. Neurosci. Methods 135(1–2): 17–26.

Torborg, C. L., and Feller, M. B. (2005). Spontaneous patterned retinal activity and the refinement of retinal projections. Prog. Neurobiol. 76(4):213–235.

Torborg, C. L., Hansen, K. A., and Feller, M. B. (2005). High frequency, synchronized bursting drives eye-specific segregation of retinogeniculate projections. Nat. Neurosci. 8(1):72–78.

Van Hooser, S. D., and Nelson, S. B. (2006). The squirrel as a rodent model of the human visual system. Vis. Neurosci. 23(5): 765–778.

Wagner, E., McCaffery, P., and Dräger, U. C. (2000). Retinoic acid in the formation of the dorsoventral retina and its central projections. Dev. Biol. 222(2):460–470.

Webster, M. J., and Rowe, M. H. (1984). Morphology of identified relay cells and interneurons in the dorsal lateral geniculate nucleus of the rat. Exp. Brain Res. 56(3):468–474.

West, A. E., Griffith, E. C., and Greenberg, M. E. (2002). Regulation of transcription factors by neuronal activity. Nat. Rev. Neurosci. 3(12):921–931.

Ziburkus, J., and Guido, W. (2006). Loss of binocular responses and reduced retinal convergence during the period of retinogeniculate axon segregation. J. Neurophysiol. 96(5):2775–2784.

Ziburkus, J., Lo, F. S., and Guido, W. (2003). Nature of inhibitory postsynaptic activity in developing relay cells of the lateral geniculate nucleus. J. Neurophysiol. 90(2):1063–1070.