Chapter 6 Target Selection

© 2019 Elsevier Inc. All rights reserved. Fig. 6.1 Conceptual, not necessarily successive, stages of targeting. From top to bottom, an axon defasciculates from the main nerve in the region of the target. It enters the target and begins to branch, but is prevented from exiting by a repulsive border. The axon responds to a topographic gradient that promotes branching at the correct location. It then selects a particular layer and finally homes in on particular target cells.

© 2019 Elsevier Inc. All rights reserved. 2 Fig. 6.2 Defasciculation is regulated by Beat proteins. (A) Two intersegmental nerve (ISN) motor axons are shown, one that branches off the nerve in the region of its target. (B) In a beat-1a mutant, the motor axon on the right does not defasciculate. (C) The beat-1a mutant phenotype is rescued in a fas II mutant background.

© 2019 Elsevier Inc. All rights reserved. 3 Fig. 6.3 Some sympathetic neurons use a change in NT-3 expression to innervate their targets in the ear. (A) Some SCG neurons project to and arborize in the pinna of a normal mouse. (B) In NT-3 knockout mice, these fibers do not invade the pinna. (C) Restoration of targeting by injection of NT-3 into the ear.

© 2019 Elsevier Inc. All rights reserved. 4 Fig. 6.4 Innervation of the inner ear is regulated by BDNF (purple) and NT-3 (green). (A) In the wild-type animal, the vestibulo-cochlear ganglion, all of whose neurons express TrkB and TrkC, grow toward the developing inner ear, which has a vestibular and a cochlear primordium. As the system develops and the primordia develop into semicircular canals and a cochlea, the ingrowing axons innervate both parts of the inner ear. (B) In BDNF, NT-3, or TrkB, TrkC double mutants, the inner ear remains uninnervated. (C) In transgenic mice in which BDNF has been knocked into the NT-3 coding region, the cochlear region becomes innervated by the vestibular part of the ganglion. © 2019 Elsevier Inc. All rights reserved. 5 Fig. 6.5 Growth cones change when they enter their target zones. (A) Images from a time-lapse movie of a retinal ganglion cell growing in the optic tract and then crossing (at the dotted line) into the tectum. The simple growth cone becomes much more complex and slows down dramatically as it enters the target. (B) Retinal axons grow on regions high in FGF and slow within the tectum, which has low FGF levels (top); tectum is avoided by retinal axons that misexpress a dominant negative FGF receptor (bottom). (C) Retinal axons slow down and branch when they reach the tectum in control animals (top), but when the pathway is exposed to high levels of FGF-2, the axons keep going and do not innervate the tectum (bottom).

© 2019 Elsevier Inc. All rights reserved. 6 Fig. 6.6 Axons branch at pause points. (A) A growth cone pauses, microtubules splayed. (B) The growth cone moves on but leaves behind it a zone where the cytoskeleton remains somewhat disorganized. (C) A branch forms at this zone.

© 2019 Elsevier Inc. All rights reserved. 7 Fig. 6.7 Cross wiring of visual signals to the somatosensory and auditory cortex. (A) In a normal control animal, the somatosensory input goes to the ventrobasal nucleus (VBN) of the thalamus and from there to the somatosensory cortex. Retinal input is to the lateral geniculate nucleus (LGN) and the superior colliculus (SC), although neonatally, there is a transient projection to the VBN that is normally eliminated. The visual input to the LGN then goes on to the visual cortex. In an experimental animal in which the superior colliculus, the visual cortex, and the input to the VBM are removed neonatally, the transient retinal projection to the VBN is stabilized and the visual information is thus provided to the somatosensory cortex via the VBN. (B) When similar visual cortical and SC ablations are done and the auditory pathway is cut, the retinal input sprouts into the medial geniculate nucleus (MGN, the normal thalamic target of the auditory input), which projects as usual giving visual physiological properties to the auditory cortex. (C) EphrinA2 and EphrinA5 are expressed at the border between the LGN and the MGN. Even though the auditory pathway to the MGN has been cut, the retinal fibers do not invade the MGN, but they do so in EphrinA2, EphrinA5 double knockouts. © 2019 Elsevier Inc. All rights reserved. 8 Fig. 6.8 Regeneration of topographic specificity. (A) Langley’s classic study showed that stimulation of preganglionic root T4 relayed through ganglion cells in the SCG that caused vasoconstriction of the ear pinna vessels, whereas stimulation of root T1 excited other SCG neurons that caused dilation of the pupil. (B) When Langley cut the sympathetic tract above T1, all these sympathetic reflexes were abolished, but with time they recovered. (C) Shows the specificity associated with this regeneration, such that the axons that enter the chain at T4 reinnervate the SCG cells that cause ear vasoconstriction, and the axons that enter at T1 reinnervate the cells that cause pupil dilation. © 2019 Elsevier Inc. All rights reserved. 9 Fig. 6.9 Topographic input into the sympathetic chain. (A) Electrophysiological studies show that the SCG receives input from many roots but primarily the more anterior ones. The ganglion at T5 receives its primary inputs from more posterior roots. (B) When the SCG is removed and replaced with another SCG, the axons that reinnervate it tend to be from more anterior roots. (C) When a T5 ganglion is put in place of the SCG, its neurons still tend to get innervated by more posterior roots even though the ganglion is in an anterior position. (D) This topographic specificity of reinnervation is reflected in the shape of the histogram of EpSp amplitudes as a function of nerve root stimulation for the homototopic and heterotopic transplants.

© 2019 Elsevier Inc. All rights reserved. 10 Fig. 6.10 Maladaptive topography implies chemospecificity. (A) Retinal ganglion cell axons map topographically onto the tectum in fish and amphibians, and regenerate following nerve crush to reinnervate their original topographic locations. (B) A normal frog sees a fly above on its ventral retina, which projects to the dorsal tectum, leading to a snap in the appropriate upward direction. (C) A frog with a rotated eye sees the same fly on what used to be the dorsal retina, which projects as ever to the ventral tectum, leading to a snap in the wrong downward direction.

© 2019 Elsevier Inc. All rights reserved. 11 Fig. 6.11 The striped carpet assay. (A) An equatorial strip of retina spanning the nasal (N) temporal (T) extent is positioned on a striped carpet of alternating anterior (A) and posterior (P) tectal membranes. The nasal fibers from the retinal explant grow on both A and P tectal membranes, but the temporal fibers grow only on the A membranes. (B) If the tectal membranes are denatured or treated with PI-PLC, which releases PI-linked membrane proteins, the temporal axons also grow on both types of membranes, suggesting that the P membranes normally have a PI-linked repulsive guidance molecule.

© 2019 Elsevier Inc. All rights reserved. 12 Fig. 6.12 Nasotemporal to anteroposterior retinotopic guidance system. (A) There is a gradient of Ephrins in the tectum, high in the posterior or caudal (C) pole and low in the anterior or rostral (R) pole. (B) A retinal ganglion cell in the temporal retina expresses active receptors for these tectal Ephrins and avoids the posterior tectum. (C) and (D) The opposing gradients of active Eph-A receptor expressed in the retina and the gradients of A-type Ephrins in the tectum. This system can at least partially account for topographic mapping in this axis. (E) The Ephrin-A gradient is shown in the tectum of a chick that uses a soluble Eph-A receptor fused to alkaline phosphatase to reveal the distribution of the Ephrin-A ligands in the tectum. © 2019 Elsevier Inc. All rights reserved. 13 Fig. 6.13 Topographic mapping in Ephrin-A2 and Ephrin-A5 double knockouts and with mosaic Eph-A3 misexpression. (A) Label is injected into the temporal (red, left) or nasal (green, right) retinas. (B) The result in the normal mice is a projection to the anterior (left) or posterior (right) colliculus. (C) In the Ephrin-A2, Ephrin-A5 double knockouts (tectum is gray), termination zones are all over the A-P extent of the colliculus for both nasal and temporal injections. (D) Two separate overlapping maps form from the subset of RGCs that express Isl2 and thus extra Eph-A3 (blue), and those that express the normal amount of Eph-A3 (red)—see graph below and micrograph of retinal axons from neighboring dye injection sites in the retina projecting to two distinct locations in the tectum, while maintaining their nearest neighbor topography. © 2019 Elsevier Inc. All rights reserved. 14 Fig. 6.14 Retinal axon outgrowth with variation in both retinal position and Ephrin-A2 concentration. Representative photographs showing outgrowth from the eight contiguous explant positions (numbered 1–8) across the nasal-temporal axis of the retina grown on substrates containing different proportions of membranes from Ephrin-A2 DNA-transfected and untransfected cells. Outgrowth varies with both retinal position and Ephrin concentration. Responses to Ephrin-A2 membranes vary from total outgrowth inhibition (at higher concentrations and more temporal positions) to severalfold outgrowth promotion (at lower concentrations and nasal positions).

© 2019 Elsevier Inc. All rights reserved. 15 Fig. 6.15 Topographic mapping of the dorsal-ventral retina onto the mediolateral tectum. In Xenopus, Ephrin-B expressing retinal ganglion cells from the dorsal retina are attracted to the high levels of Eph-B in the lateral tectum via reverse signaling, while in chicks Eph-B expressing retinal ganglion cells are attracted via forward signaling to the high levels of Ephrin-B in the medial tectum. In chicks (bottom), retinal ganglion cells additionally express Wnt receptors Frizzled (Fz), mediating attraction, and Ryk, mediating repulsion, responding to graded Wnt3 expression along the tectal D-V axis. D, dorsal; V, ventral; M, medial; L, lateral.

© 2019 Elsevier Inc. All rights reserved. 16 Fig. 6.16 Countergradients of EphA/ephrin-A guidance cues are found throughout the developing visual system. Topographic information flow is ensured by maps that use the same guidance systems and are aligned from one processing center to the next: retina, superior colliculus (SC), lateral geniculate nucleus (LGN), primary visual cortex (V1). N, nasal; T, temporal; A, anterior; P, posterior; D, dorsal; V, ventral; L, lateral; M, medial.

© 2019 Elsevier Inc. All rights reserved. 17 Fig. 6.17 The role of Sema3A in keeping some axons out of a target region. (A) Stretch receptors project their central axons into the ventral horn of the spinal cord where they synapse on motor neuron dendrites. Nocioreceptive and thermoreceptive axons, which are NGF-dependent, however, terminate in the dorsal gray matter of the spinal cord. Sema3A is expressed only in the ventral cord. (B) COS cells in culture repel axons from DRG cultures consisting of nocioceptive and thermoreceptive axons, generated by providing NGF but not NT-3, which would be required for survival of stretch receptors.

© 2019 Elsevier Inc. All rights reserved. 18 Fig. 6.18 Matching sensory motor connectivity determined by muscles. (A, B) Spindle afferents terminate on homonymous motor neurons. (C) If the sensory fibers that normally innervate ventral muscles are forced to innervate dorsal muscles, they switch synaptic partners to the motor neurons that normally innervate dorsal muscles. (D) Different muscles seem to induce or maintain ETS molecules on both the motor and sensory neurons that innervate it. Thus, the motor and sensory neurons that innervate the adductor muscle express the ETS ER81 molecule. (E) If the peripheral muscles are removed, the motor and sensory neurons no longer express ETS molecules. © 2019 Elsevier Inc. All rights reserved. 19 Fig. 6.19 Visual pathway in zebrafish. (A) Light is projected onto photoreceptors (PhRs), which transmit this information to bipolar cells (BCs) via synaptic connections in the outer plexiform layer (OPL). BCs forward visual information to amacrine cells (ACs) and the dendrites of ganglion cells (GCs) in the inner plexiform layer (IPL), composed of 10 discrete synaptic sublaminae (colored layers), each computing different types of visual information. Information from several IPL sublaminae is passed on to distinct layers in the tectum via ganglion cell axons. The tectum integrates multiple sensory modalities and issues appropriate motor commands. INL, inner nuclear layer. (B, C) Dorsal and side views of several retinal ganglion cells, distinctly labeled with the “Brainbow” system of fluorophores, terminating in separate synaptic laminae of the tectum. Scale bar: 20 μm.

© 2019 Elsevier Inc. All rights reserved. 20 Fig. 6.20 (A) The CAM code of sublaminar targeting. In the inner plexiform layer in the retina, the principal segregation of ON and OFF circuitry is in part mediated by repulsive Sema6A-PlexinA4 interactions. Neurons projecting into the OFF sublaminae S1 and S2 express PlexinA4, which mediates repulsion from Sema6A expressing processes in the ON sublaminae S3-S5. The cell shown in yellow is Tyrosine Hydroxylase (TH)-positive with processes confined to layer S1 (this panel is a superimposition of separate images from Matsuoka et al., 2011). (B, C). Each synaptic lamina has a unique combination of homophilic cell adhesion molecules, encoded by related genes of the Immunoglobulin Superfamily. Misexpression and knockdown experiments show that these cell adhesion molecules can be sufficient and necessary for targeting synaptic terminals to specific sublaminae. (After Yamagata et al., 2003; Yamagata and Sanes, 2008; Matsuoka et al., 2011; Baier, 2013.) © 2019 Elsevier Inc. All rights reserved. 21 Fig. 6.21 Multiple forms of Dscam are generated by alternative splicing. (A) The Dscam gene spans 61.2 kb of genomic DNA. Dscam mRNA extends 7.8 kb and comprises 24 exons. Mutually exclusive alternative splicing occurs for exons 4, 6, 9, and 17: 1 of 12 exon 4 alternatives (red), 1 of 48 exon 6 alternatives (blue), 1 of 33 exon 9 alternatives (green), and 1 of 2 exon 17 alternatives (yellow) are retained in each mRNA. Constant exons are represented by gray lines in genomic DNA and white boxes in mRNA. The splicing pattern shown (4.1, 6.28, 9.9, 17.1) corresponds to that obtained in the initial cDNA clone. The alternatively spliced exons 4, 6, 9, and 17 encode the N-terminal half of Ig2 (red), the N-terminal half of Ig3 (blue), the entire Ig7 (green), and the transmembrane domain (yellow), respectively (from Schmucker et al., 2000). (B) Schematic of mechanosensory projections into the adult fly CNS. The first panel shows the wild-type trajectory. The remaining panels show the same mechanosensory neuron’s projection in Dscam deletion flies with exon 4 isoform diversity indicated by an expanded view of the exons present in each mutant. The blue boxes indicate deletions. Branches prevalent in either Dscam deletion mutant (but not both) are highlighted by blue and green arrowheads, and the blue line denotes the CNS midline (from Bharadwaj and Kolodkin, 2006). (C) The molecular mechanism of self-avoidance involves self-recognition mediated repulsion. As a neurite branches, each sister branch expresses the same set of Dscam1 isoforms. These isoforms bind homophilically and this transduces a signal to dissociate the receptor complex and initiate a repulsive response.

© 2019 Elsevier Inc. All rights reserved. 22 Fig. 6.22 Connectivity in the Drosophila larval motor system is specified by combinations of guidance cues. Abdominal half segments have 30 muscles, each innervated by a unique motor axon. (A) Choice of nerves (ISN versus SN) directs delivery of growth cones to either set of antagonistic muscles, internal versus external (blue, expressing Connectin). Nerve subbranches direct growth cones to proximal versus distal targets within a set. (A–C) Some unique cell adhesion molecules (A), secreted guidance factors (B), and heterophilic ligands (C) are shown, illustrating that specificity of synaptic partner matching within a target region can be achieved by each muscle being distinct by virtue of the combinatorial action of multiple cues, both contact-mediated and secreted. © 2019 Elsevier Inc. All rights reserved. 23 Fig. 6.23 Synaptic specificity at subcellular resolution. (A) Cerebellar Purkinje cells attract GABAergic innervation at their axon initial segments (AIS) by virtue of a localized enrichment of the cortical protein Ankyrin G. This concentrates the transmembrane cell adhesion molecule Neurofascin186 (NF186), which in turn mediates synapse formation with basket cell axons. (B) Micrograph of a Purkinje neuron doubly labeled to visualize Calbindin (green) and Neurofascin 186 (red).

© 2019 Elsevier Inc. All rights reserved. 24 Fig. 6.24 Comparison of topographic mapping in the visual system (B, top), where neighboring cells project to neighboring targets, creating a central representation of visual space; and the olfactory system (A) where cells of the same type are intermingled and yet their axons sort out and converge forming an odor representation map (B, bottom).

© 2019 Elsevier Inc. All rights reserved. 25 Fig. 6.25 Formation of a discrete neural map in the mouse olfactory system. (A, top) Thousands of olfactory receptor neurons (ORNs) in the olfactory epithelium (OE) that express a common odorant receptor (OR) converge their axonal projections onto the same glomerulus (arrow) in the olfactory bulb (OB). (Bottom) ORNs that express a given OR are distributed within a band along the dorsomedial (DM) to ventrolateral (VL) axis in the OE and project their axons to corresponding color-matched positions along the dorsal-ventral (D-V) axis of the OB. According to data from Miyamichi et al. (2005). Dotted rectangles correspond to OB schematics in (B) and (C) as indicated. A, anterior; P, posterior; D, dorsal; V, ventral. The A-P axis at bottom left (nasal epithelium) is orthogonal to the plane shown. (B) At early stages, basal level G protein (three purple subunits) coupling with individual ORs leads to activation of adenylate cyclase independent of agonists binding to ORs. Thus each OR mediates distinct levels of cAMP and CREB activation, which induces gene expression of guidance cue receptors, PlexinA1 (PlxnA1) and Neuropilin1 (Nrp1). These contribute to ORN axon-axon sorting and thus targeting within the anterior-posterior axis. According to data from Imai et al. (2006). (C) Activation of ORs by odorants/agonists (cyan) leads to correlated and activity-dependent expression of homophilic cell adhesion molecules Kirrel2 and Kirrel3, mediating cohesion within glomeruli, and repulsive axon guidance ligand/receptor pair Ephrin-A and EphA, implementing segregation by repulsion between neighboring glomeruli. Together these mechanisms contribute to local axonal convergence and sorting by segregation. According to data from Serizawa et al. (2006) and Nakashima et al. (2013). © 2019 Elsevier Inc. All rights reserved. 26 Fig. 6.26 Olfactory receptors are involved in central targeting. (A) Axons from olfactory neurons expressing the P2 odorant receptor converge on the P2 glomerulus. (B) If the P2 receptor is deleted, these axons do not converge on any glomerulus. (C) If these neurons are made to express the M71 odorant receptor instead of the P2 receptor, they converge on a glomerulus somewhere inbetween the normal P2 and the normal M71 glomerulus. (D) If they are forced to express the P3 odorant receptor which is relatively near the P2 glomerulus, they converge on a glomerulus right next to P3.

© 2019 Elsevier Inc. All rights reserved. 27 Fig. 6.27 Agonist-independent baseline activity of Odorant Receptors (ORs) mediates axon targeting along the anterior-posterior axis of the olfactory bulb (OB). (A) A specific OR was replaced by another G-protein coupled receptor, the β2-Adrenergic Receptor (β2-AR). Characterized mutant forms of the β2-AR with different levels of baseline G-protein activation directed olfactory neuron axons to distinct locations along the anterior–posterior (A-P) axis. Micrographs show termination zones in the OB established by these mutant forms of β2- AR with different levels of baseline G-protein activity relative to control, wild type β2-AR (white). Scale bars, 500 μm. (B) Quantification of G-protein baseline activities by wild type (standard) and mutant β2-ARs. (C) Diagrammatic summary of data from A.

© 2019 Elsevier Inc. All rights reserved. 28 Fig. 6.28 RGC axons overshoot their correct termination zone in a position-dependent manner in chick embryos. (A) RGC axons in tectal whole mounts labeled by a small focal DiI injection into peripheral temporal retina (top), central retina (middle), or peripheral nasal retina (bottom) on E11. The injection sites are shown in drawings of the retinal whole mounts and marked by arrows. The relative positioning of the labeled axons and branches within the tectum is shown to the right, with drawings of the outline of each tectum on which the labeled axons and branches are traced. Axons overshoot their topographically correct TZ (the predicted locations of the TZs are marked with black arrowheads), but the distribution of interstitial branches along the axon shafts (white arrowheads) is strongly biased for the location of the future TZ along the anterior (A)-posterior (P) tectal axis. Peripheral temporal axons exhibit the greatest overshoot and peripheral nasal axons the least. (B) Distribution of interstitial branches along the axon shaft expressed in percentages. The A-P tectal axis was divided into 500 mm bins, and the number of branches in each bin is graphed as the percentage of total branches for each of the three groups of injections.

© 2019 Elsevier Inc. All rights reserved. 29 Fig. 6.29 Shifting connections. (A) During the lifetime of a frog or fish, its eye and brain continue to grow. The retina grows circumferentially like a tree, but the tectum grows in expanding posterior crescents. As a result, new retina that is added temporally must send axons to the anterior primordial tectum while fibers from the central primordial retina must shift posteriorly, and new nasal fibers map to the new posterior tectum in order to keep the map in topographic order. (B) If half the tectum is removed from a fish, after about a month the retinotopic map will regulate and compress, mapping out evenly over the remaining half tectum. (C) Similar regulation occurs when half the retina is removed. The remaining projection eventually expands over the whole tectum.

© 2019 Elsevier Inc. All rights reserved. 30 Fig. 6.30 Somatotopic representation in the cortex. (A) The motor area of the precentral gyrus of the cerebral cortex was stimulated electrically in human patients during neurosurgery. (B) A “homunculus” of the body on the motor cortex illustrates the sequence of representation as well as the disproportionate representation given to the various muscles involved in skilled movements.

© 2019 Elsevier Inc. All rights reserved. 31 Fig. 6.31 Large-scale plasticity in the somatosensory cortex. (A) The normal organization of the cortical topography in the somatosensory system. (B) Damage to the arm may result in large-scale reorganization of the cortical map. (C) There is an isomorphic mapping of the somatosensory thalamus onto the primary cortex. (D) When the arm is damaged peripherally and the sensory neurons in the DRG survive, the reorganization is cortical rather than thalamic. (E) When the damage is more central causing the DRG cells to degenerate, both the thalamus and cortex get reorganized.

© 2019 Elsevier Inc. All rights reserved. 32 Fig. 6.32 Plasticity of the mouse barrel field in the somatosensory cortex. (A) The correspondence between bristles and barrel field in the cortex of a normal mouse. (B) An extra whisker in row C leads to the formation of an extra barrel in the appropriate location in the cortex. (C) Neonatal damage to the B3 whisker causes the shrinkage of this barrel and the expansion of neighboring ones. (D) Tying two whiskers together causes their barrel field to coalesce.

© 2019 Elsevier Inc. All rights reserved. 33