Spontaneous patterned retinal activity and the refinement of retinal projections

https://doi.org/10.1016/j.pneurobio.2005.09.002Get rights and content

Abstract

A characteristic feature of sensory circuits is the existence of orderly connections that represent maps of sensory space. A major research focus in developmental neurobiology is to elucidate the relative contributions of neural activity and guidance molecules in sensory map formation. Two model systems for addressing map formation are the retinotopic map formed by retinal projections to the superior colliculus (SC) (or its non-mammalian homolog, the optic tectum (OT)), and the eye-specific map formed by retinal projections to the lateral geniculate nucleus of the thalamus. In mammals, a substantial portion of retinotopic and eye-specific refinement of retinal axons occurs before vision is possible, but at a time when there is a robust, patterned spontaneous retinal activity called retinal waves. Though complete blockade of retinal activity disrupts normal map refinement, attempts at more refined perturbations, such as pharmacological and genetic manipulations that alter features of retinal waves critical for map refinement, remain controversial. Here we review: (1) the mechanisms that underlie the generation of retinal waves; (2) recent experiments that have investigated a role for guidance molecules and retinal activity in map refinement; and (3) experiments that have implicated various signaling cascades, both in retinal ganglion cells (RGCs) and their post-synaptic targets, in map refinement. It is likely that an understanding of retinal activity, guidance molecules, downstream signaling cascades, and the interactions between these biological systems will be critical to elucidating the mechanisms of sensory map formation.

Introduction

The brain contains highly ordered circuitry in which sensory inputs are organized into maps that represent different features of sensory space. One dogma of developmental neurobiology is that guidance molecules mediate initial map development, while refinement of these maps requires neural activity. This underlies many hypotheses regarding the relative contribution of activity-dependent and activity-independent factors in map formation.

In the visual system, two well-studied organizational schemes are retinotopic and eye-specific maps (see Fig. 1). In this section, we introduce these maps and provide overviews of the hypotheses regarding the relative role of activity-dependent processes and guidance molecules in their development.

Mature retinotopic maps are organized such that a visual stimulus activates neighboring retinal neurons, which in turn project to and stimulate neighboring neurons in the corresponding target structure in the brain. For example, in the superior colliculus (SC), and its non-mammalian homolog the optic tectum (OT), the nasal–temporal (N–T) axis of the retina maps to the posterior–anterior (P–A) axis of the SC, such that stimulation of the nasal retina elicits responses in posterior SC neurons, while anterior SC neurons respond to stimulation of temporal retina. The dorsal–ventral (D–V) axis of the retina maps along the lateral–medial (L–M) axis of the SC in a similar fashion. In the dorsal lateral geniculate nucleus of the thalamus (dLGN), the N–T axis of retina maps to the D–V axis of the dLGN and the D–V axis of retina maps project to the L–M axis of the dLGN (Fig. 1).

Eye-specific maps are organized such that neurons located in distinct regions in the target structure respond to visual stimuli that activate neurons in one eye or the other. For example, axons from contralateral and ipsilateral retinas make synaptic connections with neurons in separate layers within the dorsal lateral geniculate nucleus of the thalamus, each layer receiving only ipsilateral or only contralateral input. Similarly, axons from the distinct eye-specific regions in the dLGN project to distinct alternating right-eye/left-eye columns, called ocular dominance columns, in primary visual cortex. Though the retinal projections to the SC are primarily contralateral, there is an ipsilateral projection to a distinct cellular layer in the rostral part of the nucleus (Fawcett et al., 1984, O’Leary et al., 1986, Thompson and Holt, 1989).

In mammals, the precise targeting of retinal ganglion cell (RGC) axons necessary for retinotopy and eye-specific layers emerges from initially unordered projections of RGC axons within the SC and dLGN. Within the SC, all RGC axons initially extend beyond their correct topographic position, growing toward the posterior pole (Simon and O’Leary, 1992). As the axons retract, side branches form in the region that will be the topographically correct termination zone (TZ). These side branches continue to elaborate into wide, diffuse axon arbors and must further refine before becoming the tight, dense TZ seen in adults. Retinal projections to the dLGN are not thought to undergo dramatic retinotopic refinement (Huberman et al., 2005b, Sretavan et al., 1988).

Within the dLGN, RGC axons from both eyes are initially intermixed (Godement et al., 1984, Linden et al., 1981, Penn et al., 1998, Shatz and Kirkwood, 1984). Shortly after RGC axons have reached the dLGN, intraocular injection of an anterograde tracer into the ipsilateral eye labels a large region of the dLGN, while tracer injection into the contralateral eye results in a uniform labeling of the entire dLGN. Later, contralaterally projecting axons are excluded from the region occupied by ipsilaterally projecting axons. The changes in axonal morphology that may underlie this segregation have been studied in the cat, where, after arriving in the dLGN, small side branches of retinal axons begin to form off the main axon shaft (Sretavan and Shatz, 1986). Throughout development, a dense axonal arbor begins to form in the correct eye-specific layer, while incorrect branches are eliminated. This process of elaboration and retraction eventually results in the segregation of afferents from each eye into separate layers within the dLGN. A similar process underlies eye-specific segregation of retinal projections to the SC (Fawcett et al., 1984, O’Leary et al., 1986, Thompson and Holt, 1989). A recent study suggests that in fetal primates, eye-specific layers form before the analogous period of synaptic refinement described above (Huberman et al., 2005a).

Two types of mechanisms have been proposed to underlie map formation: matching molecular cues; and activity-dependent synaptic rearrangements. The concept that gradients of molecular cues specify the correct topographic location of axonal termination within a target structure was first proposed by Roger Sperry as a chemoaffinity hypothesis (Sperry, 1963). Sperry hypothesized that molecular gradients within the retina and tectum give each cell a unique chemical tag or identification. These chemical tags specify which retinal and tectal cells should form synapses with each other. This hypothesis can be applied to eye-specific layers, whereby each RGC expresses a chemical tag corresponding to its eye of origin that must match with a chemical tag for the correct eye layer within the dLGN.

Most activity-dependent mechanisms of map formation have been based on Hebb's postulate (1949), which states that synapses are strengthened when the pre- and post-synaptic cells are simultaneously active (Zhang and Poo, 2001). Hebb's postulate has been expanded to include the converse: asynchronously active synapses are weakened and eventually eliminated. Models based on Hebb's postulate predict that activity plays an “instructive role,” meaning that all of the information required to drive the refinement is contained in the activity patterns. In this scenario, activity that drives the neural rearrangements that underlie retinotopic refinement would contain information about the distance between two RGCs. Neighboring neurons are characterized by temporally correlated firing of action potentials, with the amount of correlation decreasing as a function of distance between the RGCs. Furthermore, activity that drives eye-specific refinement would contain information about the eye of origin, e.g., activity from neighboring RGCs originating in one retina is more correlated than activity between retinas.

Does the visual system generate the types of activity that could drive activity-dependent refinement? In amphibians and fish, the retina is responsive to light before its projections reach their target structures in the brain, suggesting that vision could provide the activity necessary for refinement. In mammals, the refinement of retinotopic retinocollicular maps and eye-specific retinogeniculate maps, however, occurs prior to the presence of functional photoreceptors. Though there are no visually evoked responses during this time, spontaneous waves of depolarization, termed retinal waves, traverse the retina (for recent review, see Firth et al., 2005). Retinal waves are composed of bursts of action potentials that correlate the firing of neighboring RGCs, while the firing of more distant RGCs is less correlated (Demas et al., 2003a, McLaughlin et al., 2003b, Meister et al., 1991, Wong et al., 1993). This change in the amount of correlation as a function of distance between cells may be appropriate for establishing retinotopic maps. In addition, individual RGCs participate in waves infrequently (firing for approximately 5 s and then silent for 1–2 min). Thus, there is a low probability that RGCs from each eye will fire simultaneously. These strong intra-retinal correlations combined with weak inter-retinal correlations may be appropriate for driving eye-specific segregation.

Here we review evidence that elucidates the relative role of molecular cues and patterned spontaneous retinal activity in the refinement of retinotopic maps in retinocollicular projections and eye-specific refinement of retinogeniculate projections. First, we review the mechanisms that underlie the generation of retinal waves. Second, we explore how manipulations that alter retinal wave activity influence the development of retinal projections. Third, we will review the role of guidance molecules and signaling cascade molecules in the refinement of those projections. Last, we propose that there are several mechanisms by which patterned activity and guidance molecules interact and that a synergism of mechanisms may be the key to understanding the refinement of connections within the developing visual system.

Section snippets

The cellular basis of spatiotemporal patterns of retinal waves

To test the hypothesis that retinal waves contain all the information to drive map refinement, there is an ongoing search to develop manipulations that alter specific features of the firing patterns. Hence, identifying the cell classes and mechanisms involved in generating waves is critical. The synaptic circuits that generate retinal waves have been described in several vertebrate species, including ferret, cat, mouse, rabbit, turtle, and chick (for reviews see Catsicas et al., 1998, Feller,

Role for neural activity in the formation of retinotopic and eye-specific maps

Prior to eye-opening, retinal axons establish synaptic connections with their targets in the brain. These connections are organized within the target tissue in a fashion that is dictated by the relative spatial location of their origin in the retina. The temporal coincidence of retinal waves and establishment of both retinotopic and eye-specific maps has led to the hypothesis that the activity induced by retinal waves is essential for correctly establishing these connections. In this section we

Markers that specify retinotopic location

Retinotopic refinement within the SC, or its non-mammalian homolog the optic tectum, is often used as a classic example of the ability of gradients of guidance molecules to bring about the precise targeting of axons (for reviews, see Flanagan and Vanderhaeghen, 1998, Kullander and Klein, 2002, McLaughlin et al., 2003a, Ruthazer and Cline, 2004). In fact, it was based upon experiments in this system that Sperry first proposed his chemoaffinity hypothesis, which states that each topographic

Interactions between activity and guidance molecules

Whether activity or guidance molecules are responsible the formation of maps within the visual system has been debated for many years. While in the past many scientists have taken either one side or the other in the debate, recently the two sides have begun to reach a middle ground, acknowledging that both activity and molecular cues may be responsible for establishing maps. Now the debate is changing from whether activity or molecular cues establish maps to what the relative contributions of

Linking patterned activity to axonal rearrangements

We have reviewed evidence that some features of retinal waves are necessary for axonal rearrangements during retinotopic and eye-specific refinement. However, the mechanisms by which action potentials result in these axonal rearrangements have not been elucidated. We now discuss evidence for three classes of mechanisms involved in activity-dependent synaptic refinement: (1) Hebbian-based synaptic competition; (2) activity-dependent gene transcription; and (3) intracellular signaling cascades.

Conclusions

How the connectivity of the visual system emerges during development remains one of the great mysteries of developmental neuroscience. Classically, the mechanisms that underlie the establishment of sensory maps have been categorized as activity-dependent or activity-independent. If any progress has been made in the last few years, it is that this distinction may no longer be appropriate. There is ample evidence that both guidance molecules and spontaneous activity are critical for driving map

Acknowledgment

This work was supported in part by a McKnight Scholars Fund, and the National Institute of Health (grant no. NS13528).

References (176)

  • M.B. Feller et al.

    A precritical period for plasticity in visual cortex

    Curr. Opin. Neurobiol.

    (2005)
  • M.B. Feller et al.

    Dynamic processes shape spatiotemporal properties of retinal waves

    Neuron

    (1997)
  • R.D. Fields et al.

    Temporal integration of intracellular Ca2+ signaling networks in regulating gene expression by action potentials

    Cell Calcium

    (2005)
  • S.I. Firth et al.

    Retinal waves: mechanisms and function in visual system development

    Cell Calcium

    (2005)
  • J. Frisen et al.

    Ephrin-A5 (AL-1/RAGS) is essential for proper retinal axon guidance and topographic mapping in the mammalian visual system

    Neuron

    (1998)
  • J.L. Goldberg et al.

    Retinal ganglion cells do not extend axons by default: promotion by neurotrophic signaling and electrical activity

    Neuron

    (2002)
  • M.S. Grubb et al.

    Abnormal functional organization in the dorsal lateral geniculate nucleus of mice lacking the beta 2 subunit of the nicotinic acetylcholine receptor

    Neuron

    (2003)
  • M.S. Grubb et al.

    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.

    (2004)
  • M.S. Grubb et al.

    The influence of early experience on the development of sensory systems

    Curr. Opin. Neurobiol.

    (2004)
  • M.J. Hansen et al.

    Retinal axon response to ephrin-as shows a graded, concentration-dependent transition from growth promotion to inhibition

    Neuron

    (2004)
  • M.G. Hanson et al.

    Normal patterns of spontaneous activity are required for correct motor axon guidance and the expression of specific guidance molecules

    Neuron

    (2004)
  • E. Herrera et al.

    Zic2 patterns binocular vision by specifying the uncrossed retinal projection

    Cell

    (2003)
  • R. Hindges et al.

    EphB forward signaling controls directional branch extension and arborization required for dorsal–ventral retinotopic mapping

    Neuron

    (2002)
  • R. Juttner et al.

    Early onset of glutamatergic and GABAergic synaptic activity in the visual layers of the rodent superior colliculus

    Int. J. Dev. Neurosci.

    (2001)
  • M. Klinger et al.

    Adenosine receptors: G protein-mediated signalling and the role of accessory proteins

    Cell Signal.

    (2002)
  • B.E. Lonze et al.

    Function and regulation of CREB family transcription factors in the nervous system

    Neuron

    (2002)
  • N. Matsumoto et al.

    Regenerating retinal fibers of the goldfish make temporary and unspecific but functional synapses before forming the final retinotopic projection

    Neuroscience

    (1987)
  • T. McLaughlin et al.

    Regulation of axial patterning of the retina and its topographic mapping in the brain

    Curr. Opin. Neurobiol.

    (2003)
  • T. McLaughlin et al.

    Retinotopic map refinement requires spontaneous retinal waves during a brief critical period of development

    Neuron

    (2003)
  • E. Menna et al.

    The anterogradely transported BDNF promotes retinal axon remodeling during eye specific segregation within the LGN

    Mol. Cell Neurosci.

    (2003)
  • R.L. Meyer

    Tetrodotoxin inhibits the formation of refined retinotopography in goldfish

    Brain Res.

    (1983)
  • R. Mooney et al.

    Enhancement of transmission at the developing retinogeniculate synapse

    Neuron

    (1993)
  • R. Mooney et al.

    Thalamic relay of spontaneous retinal activity prior to vision

    Neuron

    (1996)
  • M. Nakamoto et al.

    Topographically specific effects of ELF-1 on retinal axon guidance in vitro and retinal axon mapping in vivo

    Cell

    (1996)
  • H. Adelsberger et al.

    Cortical calcium waves in resting newborn mice

    Nat. Neurosci.

    (2005)
  • A. Angelucci et al.

    Experimentally induced retinal projections to the ferret auditory thalamus: development of clustered eye-specific patterns in a novel target

    J. Neurosci.

    (1997)
  • D.W. Arnett

    Statistical dependence between neighboring retinal ganglion cells in goldfish

    Exp. Brain Res.

    (1978)
  • A. Bansal et al.

    Mice lacking specific nAChR subunits exhibit dramatically altered spontaneous activity patterns and reveal a limited role for retinal waves in forming ON/OFF circuits in the inner retina

    J. Neurosci.

    (2000)
  • C. Blazynski

    Displaced cholinergic, GABAergic amacrine cells in the rabbit retina also contain adenosine

    Vis. Neurosci.

    (1989)
  • L.N. Borodinsky et al.

    Activity-dependent homeostatic specification of transmitter expression in embryonic neurons

    Nature

    (2004)
  • L.M. Boulanger et al.

    Immune signalling in neural development, synaptic plasticity and disease

    Nat. Rev. Neurosci.

    (2004)
  • D.A. Butts et al.

    Retinal waves are governed by collective network properties

    J. Neurosci.

    (1999)
  • D.A. Butts et al.

    The information content of spontaneous retinal waves

    J. Neurosci.

    (2001)
  • G. Campbell et al.

    Synapses formed by identified retinogeniculate axons during the segregation of eye input

    J. Neurosci.

    (1992)
  • M. Catsicas et al.

    Spontaneous Ca2+ transients and their transmission in the developing chick retina

    Curr. Biol.

    (1998)
  • L.M. Chalupa et al.

    Development and Organization of the Retina: From Molecules to Function

    (1997)
  • A.R. Chandrasekaran et al.

    Evidence for an instructive role of retinal activity in retinotopic map refinement in the superior colliculus of the mouse

    J. Neurosci.

    (2005)
  • B. Chapman

    Necessity for afferent activity to maintain eye-specific segregation in ferret lateral geniculate nucleus

    Science

    (2000)
  • H.T. Cline et al.

    N-Methyl-d-aspartate receptor antagonist desegregates eye-specific stripes

    Proc. Natl. Acad. Sci. U.S.A.

    (1987)
  • M. Constantine-Paton et al.

    Eye-specific termination bands in tecta of three-eyed frogs

    Science

    (1978)
  • Cited by (201)

    View all citing articles on Scopus
    View full text