Molecular Gradients andDevelopment of RetinotopicMapsTodd McLaughlin and Dennis D.M. O’LearyMolecular Neurobiology Laboratory, The Salk Institute, La Jolla, California 92037;email: [email protected], [email protected]. Rev. Neurosci.2005. 28:327–55doi: 10.1146/annurev.neuro.28.061604.135714Copyrightc 2005 byAnnual Reviews. All rightsreserved0147-006X/05/0721-0327$20.00Key Wordsaxon branching, axon guidance, bidirectional signaling, Ephs,ephrins, visual system developmentAbstractGradients of axon guidance molecules have long been postulated tocontrol the development of the organization of neural connectionsinto topographic maps. We review progress in identifying moleculesrequired for mapping and the mechanisms by which they act, fo-cusing on the visual system, the predominant model for map de-velopment. The Eph family of receptor tyrosine kinases and theirligands, the ephrins, remain the only molecules that meet all cri-teria for graded topographic guidance molecules, although othersfulfill some criteria. Recent reports further define their modes ofaction and new roles for them, including EphB/ephrin-B controlof dorsal-ventral mapping, bidirectional signaling of EphAs/ephrin-As, bifunctional action of ephrins as attractants or repellents in acontext-dependent manner, and complex interactions between mul-tiple guidance molecules. In addition, spontaneous patterned neuralactivity has recently been shown to be required for map refinementduring a brief critical period. We speculate on additional activitiesrequired for map development and suggest a synthesis of molecularand cellular mechanisms within the context of the complexities ofmap development.327ContentsINTRODUCTION................. 328Toward Discovering GradedTopographic GuidanceMolecules ..................... 330MECHANISMS OF MAPFORMATION ................... 331Axon Extension and TargetOvershoot During APMapping ...................... 333TN Retinotopic MappingAchieved Through AP-SpecificInterstitial Branching .......... 335Mechanisms for AP BranchSpecificity..................... 337Parallel AP Gradients of Promotersand Inhibitors of Branching .... 337Opposing AP Gradients of BranchInhibitors ..................... 338Lateral-Medial MappingAccomplished by DirectedGrowth of Interstitial Branches 338EphBs and Ephrin-Bs ControlDV Retinotopic Mapping ...... 339Distinctions in Guidance ofPrimary Axons and InterstitialBranches Require UniqueMechanisms................... 340Multiple Actions and Models ofEphBs and Ephrin-Bsin DV Map Development ...... 342Accounting for Species Differencesin Development of TopographicMaps ......................... 343Refinement of the RetinotopicMap: Patterned Activity andAxon Repellents ............... 344Additional Activities andInteractions PotentiallyRequiredfor Map Development ......... 346Genetic Screens: Mapping theFuture? ....................... 346FUTURE DIRECTIONS........... 347INTRODUCTIONA critical function of the nervous system is tointerpret the environment through the con-nections of various sensory organs to thebrain. To accomplish this task, incoming in-formation must be organized in an efficientmanner. Perhaps the most efficient organi-zation is achieved through the use of topo-graphic maps, which are present through-out the brain, to process sensory information(Kaas 1997). In general a topographic map is aprojection from one set of neurons to anotherwherein the receiving set of cells reflects theneighbor relationships of the projecting set.In the nervous system of higher vertebratestopographic maps are common and includesensory maps of the body, tonotopic mapsfor auditory stimuli, and maps of the visualfield. Furthermore, topographic maps persistin some form throughout the circuitry fromfirst-order to higher-order connections.Map development has been studied in sev-eral vertebrate projection systems, includ-ing thalamocortical (Dufour et al. 2003,Vanderhaeghen & Polleux 2004), hippocam-poseptal (Gao et al. 1996, Yue et al. 2002),olfactory/vomeronasal (Sidebar 1), motor ax-ons to muscles (Feng et al. 2000, Nguyen et al.2002), and retina to its targets in the brain(see below). However, this latter system, theprimary visual projection formed by the ax-ons of retinal ganglion cells (RGCs) to theirmost prominent midbrain target—the optictectum (OT) of fish, amphibians, and chick,or the superior colliculus (SC) of mammals—has been far and away the predominant modelfor studying the development of topographicmaps and the gradients of guidance moleculesthat control their formation. Therefore, wefocus on the visual system and primarily on themechanisms of mapping in the target, with thegoal of providing a detailed account of the de-velopment of a vertebrate neural map and themolecular mechanisms that control it, thoughwe recognize the importance of growth coneguidance to the target and the intricaciesof multiple interacting signaling pathways328 McLaughlin·O’Leary(see recent reviews Huber et al. 2003, vanHorck et al. 2004). We devote most atten-tion to mammals and chickens in which ax-onal mechanisms of map development re-quire unique actions of topographic guidancemolecules in a specific temporal sequence,but we do provide examples of mechanismsand molecules involved in developing mapsin lower vertebrates such as frogs and fish.The representation of the retina onto theOT or SC can be simplified to the map-ping of two sets of orthogonally orientedaxes: the temporal-nasal (TN) axis of theretina along the anterior-posterior (AP) axisof the OT/SC, and the dorsal-ventral (DV)axis of the retina along the lateral-medial(LM) axis of the OT/SC (corresponding tothe ventral-dorsal OT axis in nonmammalianvertebrates). Criteria for a topographic guid-ance molecule in the retinotectal projectionare that it is expressed in a graded or restrictedmanner in the retina or OT/SC, that RGCaxons from different parts of the retina ex-hibit distinct responses to it, and that it af-fects RGC mapping in vivo; an additional cri-terion to provide a stricter definition statesthat the molecule is required for the develop-ment of a proper topographic map, althoughsome molecules can be involved in map devel-opment but their role is masked by functionalredundancy.Being well into the molecular era with en-tire genomes becoming available, one mightpresume that most major players in topo-graphic guidance are known, although theirprecise roles and interactions are not. Surpris-ingly, though, to