Current Biology Vol 15 No 9R332

Cary Lai

Schwann cells —the glial cellsthat wrap peripheral axons—arise from trunk neural crest cellsthat find their way to theemerging axons of sensory andmotor neurons. Theseprogenitors are then thought tomigrate along the outgrowingaxons and to proliferate in orderto produce a sufficient number ofcells for the myelination of theaxons. The number of these pre-myelinating Schwann cells is

believed to be regulated by axon-derived survival signals. EachSchwann cell typically firstenvelops multiple axonalsegments but ultimatelysurrounds a segment of a singleaxon [1]. Only large axons(diameter > 1 µµm) are myelinatedwhile smaller axons are not

Neuregulin as a Regulator ofSchwann Cell Number andMyelin ThicknessThe growth factor neuregulin-1plays a pervasive role in the life of

a Schwann cell [2]. Neuregulin-1was identified over 25 years agoas two distinct biologicalactivities. It was first described asglial growth factor [3], for itsability to serve as a potentSchwann cell mitogen; separately,it was shown to regulateacetylcholine receptors in musclecells in vitro and described as theacetylcholine receptor-inducingactivity (ARIA) [4]. It was alsorecognized that axonalmembranes contained asubstance, now known to beneuregulin-1, that promotedSchwann cell proliferation [5].Subsequently, neuregulin-1 wasrecognized for its ability tosupport Schwann cell survival invitro [6,7], a finding thatsuggested it may also function toregulate the number of pre-myelinating glia by serving as alimiting survival factor.

The primary receptor forneuregulin-1 in Schwann cells is a

complete accuracy. A gray objectin shadow appears slightly darkerthan it would appear in sunlight. Agray paper looks lighter on a blackbackground than on a whitebackground. Many of these errorsare captured in delightful illusions.These errors must come from thevisual system itself. And theseerrors are systematic, not random.They constitute a sort of signatureof the visual software employed bythe brain [11]. Thus the overallpattern of lightness errors shownby humans provides a powerfulconstraint on theories of lightness.

Inverse optics models are greatfor computing gray shadescorrectly. But what about theerrors? In principle, the errorscould be accounted for by partialfailures in the scission process.But such efforts to model theerrors [12,13] have not proven veryeffective.

For this reason, several theoristshave resurrected the olderframeworks concept in a modifiedform that can explain the errors[14,15]. Combining the concept offrameworks with a process ofcrosstalk between frameworks,seems to provide an impressiveaccount of lightness errors.

Anderson and Winawer [1]acknowledge these claims of theframeworks approach. And yet,their chess-piece demonstrationoffers compelling evidence ofperceptual scission.

Layer proponents, like Andersonand Winawer [1], argue that failuresin the scission process couldpotentially account for the errorspattern. Likewise frameworkproponents suggest thatframework-based models couldpotentially be expanded to includethe perception of illumination. Bothsides are open to an integration ofthe two approaches. Stay tuned.

References1. Anderson, B., and Winawer, J. (2005).

Image segmentation and lightnessperception. Nature 434, 79–83.

2. Gilchrist, A. (1999). Lightness Perception.In MIT Encyclopedia of CognitiveScience, R.W.F. Keil, ed. (Cambridge:MIT press), pp. 471–472.

3. Cornsweet, T.N. (1970). VisualPerception (New York: Academic Press).

4. Blakeslee, B., and McCourt, M.E. (1999).A multiscale spatial filtering account ofthe White effect, simultaneousbrightness contrast and gratinginduction. Vision Res. 39, 4361–4377.

5. Koffka, K. (1935). Principles of GestaltPsychology (New York: Harcourt, Brace,and World).

6. Barrow, H.G., and Tenenbaum, J. (1978).Recovering intrinsic scenecharacteristics from images. InComputer Vision Systems, A.R. Hansonand E.M. Riseman, eds. (Orlando:

Academic Press), pp. 3–26.7. Gilchrist, A. (1979). The perception of

surface blacks and whites. Sci. Am. 240,112–123.

8. Singh, M., and Anderson, B.L. (2002).Toward a perceptual theory oftransparency. Psychol. Rev. 109,492–519.

9. Marr, D. (1982). Vision (San Francisco:Freeman).

10. Metelli, F. (1975). Shadows withoutpenumbra. In Gestalttheorie in dermodernen Psychologie, S. Ertel, L.Kemmler and M. Stadler, eds.(Darmstadt: Steinkopff), pp. 200–209.

11. Gilchrist, A. (2003). The importance oferrors in perception. In ColourPerception: Mind and the PhysicalWorld, R.M.D. Heyer, ed. (Oxford: OxfordUniversity Press.), pp. 437–452.

12. Gilchrist, A. (1988). Lightness contrastand failures of constancy: a commonexplanation. Percept. Psychophys. 43,415–424.

13. Ross, W., and Pessoa, L. (2000).Lightness from contrast: A selectiveintegration model. Percept. Psychophys.62, 1160–1181.

14. Gilchrist, A., Kossyfidis, C., Bonato, F.,Agostini, T., Cataliotti, J., Li, X., Spehar,B., Annan, V., and Economou, E. (1999).An anchoring theory of lightnessperception. Psychol. Rev. 106, 795–834.

15. Adelson, E.H. (2000). Lightnessperception and lightness illusions. In TheNew Cognitive Neuroscience, 2nd Ed.,M. Gazzaniga, ed. (Cambridge, MA: MITPress), pp. 339–351.

Psychology Department, RutgersUniversity, Newark, New Jersey 07102,USA. E-mail: alan@psychology.rutgers.edu

DOI: 10.1016/j.cub.2005.04.028

Peripheral Glia: Schwann Cells inMotion

Neuregulin signaling through ErbB receptors is known to play anessential role in Schwann cell proliferation, survival and myelination.Recent studies in zebrafish provide a peek at living Schwann cellsmigrating along axons in vivo and suggest that ErbB signaling, whilenot required for cell movement per se, is required to maintain thedirected migration of these cells.

heterodimer composed of thetransmembrane tyrosine kinasesErbB2 and ErbB3 [6,8,9].Knockout mice lacking eitherneuregulin-1 or ErbB2 show asubstantial loss of Schwann cellprogenitors and at E10.5 theseembryos died from a failure inheart development. ErbB3-deficient mice, which can surviveuntil birth, exhibit a total loss ofSchwann cells along theirperipheral axons [10]. Thisphenotype is shared by ErbB2mutant mice whose cardiac defecthas been rescued by theexpression of this receptor in theheart [11,12]. In both of thesemouse lines, pre-migratorySchwann cells are present nearthe dorsal root ganglia, but theyfail to move away from thislocation and onto the axons.Together, the ErbB2 and ErbB3mutant mice revealed an essentialrole for neuregulin-1 signaling ingenerating Schwann cells and apossible role in the ability of thesecells to migrate down the axon.

In the postnatal period,neuregulin-1 also appears to berequired for the myelinationprocess itself. The first hint of thiswas that the induced loss ofErbB2 in Schwann cells after theinitiation of myelination resulted inthe production of thinlymyelinated fibers [13].Furthermore, constitutiveoverexpression of the type IIIisoform of neuregulin results in anincreased thickness of the myelinsheath [14]. Thus, the current viewis that neuregulin-1 serves toregulate the number of pre-myelinating Schwann cells byacting as both a mitogen and asurvival factor. Once myelinationhas been initiated, neuregulin-1also appears to play a role inregulating myelin sheaththickness.

Zebrafish ErbB MutantsWith the efforts now reported byLyons et al. [15], ourunderstanding of the roles ofErbB2 and ErbB3 in developingzebrafish may have, in somerespects, exceeded that of whatwe know from mice. The authorsperformed a genetic screen tosearch for mutants in myelinformation, which led them to

identify the zebrafishhomologues of ErbB2 and ErbB3.Similar to the phenotype seen inthe knockout mice, the mostseverely affected ErbB2 andErbB3 mutants lack Schwanncells throughout the peripheralnervous system.

By using transgenic zebrafishexpressing green fluorescentprotein (GFP) under the control ofthe foxd3 promoter (foxd3::GFP),which marks neural crest-derivedcells, the authors were able tofollow the migration of earlySchwann cells in vivo (Figure 1).Neural crest cells could bedetected as they migrated from aregion near the dorsal neural tubeto the posterior lateral lineganglion. Some of theseSchwann cell progenitors couldbe observed as they moved awayfrom the ganglion and onto theextending axons of the posteriorlateral line nerve, which containsafferent sensory axons. Whenfoxd3::GFP fish were crossed tothe most severely affected ErbB3mutant, GFP-expressingSchwann cell progenitors wereobserved to reach the posteriorlateral line ganglion but they didnot migrate out onto the axons.In addition, when cellproliferation was assayed, areduction in the number of GFP-expressing cells in early stageErbB3 mutant fish embryos wasobserved. Thus, the ErbB3mutants in fish and mice bearstriking similarities.

When foxd3::GFP fish embryoswere immersed in a solutioncontaining a pan-specificinhibitor of ErbB tyrosine kinaseactivity, cell migration wasreduced, but many cellsremained axon-associated andmotile. These cells were oftenobserved to migrate in the wrongdirection, i.e., opposite to thedirection of axonal extension.These findings suggested thatErbB signaling promotes thedirected migration of Schwanncell precursors.

To investigate the role of ErbBsignaling after Schwann cellmigration has been completed,foxd3::GFP fish embryos weretreated with BrdU and the ErbBinhibitor. When compared tountreated controls, there was a

nearly two-fold difference in cellnumber, suggesting that the post-migratory cells undergo a finalround of cell division. When thisphase of proliferation was blocked,myelination could not be initiated.The kinase inhibitor was no longerable to block the initiation ofmyelination when applied after thisround of proliferation wascompleted. Intriguingly, theauthors did not observe significantlevels of post-migratory cell deathprior to myelination, which ledthem to propose that proliferationis the primary regulatory step incontrolling the number of pre-myelinating Schwann cells.

This hypothesis differs from thepresumed situation in mice, whereproliferation in the early postnatalperiod is followed by cell deathbefore the onset of myelination[16]. In chick, two waves ofSchwann cell death have beenobserved in vivo. The first wave isthought to correlate with a failureto successfully compete foraxonally derived trophic support,presumably neuregulin-1; whilethe second one is due to the lossof Schwann cells due to thepruning of the neuronal population(and the associated loss of trophicsupport; [17]). A late phase ofproliferation prior to myelinationhas not been reported in eitherrodents or chick.

Initiating Myelination — Are FishDifferent?The findings of Lyons et al. [15]have built upon the efforts of

Dispatch R333

Figure 1. Zebrafish Schwann cellprecursors on the move.

Schwann cell precursors are labelled byfox3::GFP expression (green) and migratealong the posterior lateral line axon,labelled by acetylated tubulin expression(red). Figure courtesy of D. Lyons.

Current Biology Vol 15 No 9R334

Gilmour and colleagues [18], whodemonstrated that in zebrafish,axons provide instructiveguidance cues for the migration ofperipheral glia. Through theanalysis of mutations in thezebrafish ErbB2 and ErbB3 genes,the authors have demonstratedthat ErbB signaling provides oneof the cues for Schwann cellmigration. One of the novelfindings presented is that ErbB2/3signaling may not be required formotility per se, but for the directedmigration of Schwann cells alongtheir substrate axons. NRG1 hasbeen previously shown to promotethe motility and directed migrationof rat Schwann cells in vitro [19]and it would not be too surprisingif this role was conserved betweenfish and mammals in vivo. Theability of ErbB2/3 signaling toinfluence directed migration raisesthe issue of how proximal anddistal regions of the axon aredistinguished and whether it is thedistribution of ErbB ligands or ofother molecules that provide cuesfor directed guidance.

The most provocative claim putforth is the concept that it isproliferation and not survival that isthe primary regulator of pre-myelinating Schwann cell number.This hypothesis is at odds with theprevailing dogma that the finalnumber of pre-myelinatingSchwann cells is regulated bycompetition for a limiting amountof axonally supplied NRG1.Although this proposal is intriguing,future studies to carefully assesscell death in the developingzebrafish embryos will be requiredto determine if fish truly differ frommammals in how they regulateSchwann cell number. However,regardless of the outcome, theimpressive strides made by Lyonset al. [15] in understanding howSchwann cells develop in the fishhave underscored the advantagesprovided by the visualtransparency of zebrafish embryos.

References1. Lobsiger, C.S., Taylor, V., and Suter, U.

(2002). The early life of a Schwann cell.Biol. Chem. 383, 245–253.

2. Garratt, A.N., Britsch, S., and Birchmeier,C. (2000). Neuregulin, a factor with manyfunctions in the life of a Schwann cell.Bioessays 22, 987–996.

3. Lemke, G. (1996). Neuregulins indevelopment. Mol. Cell. Neurosci. 7,247–262.

4. Fischbach, G.D., and Rosen, K.M. (1997).ARIA: a neuromuscular junctionneuregulin. Annu. Rev. Neurosci. 20,429–458.

5. Salzer, J.L., Bunge, R.P., and Glaser, L.(1980). Studies of Schwann cellproliferation. III. Evidence for the surfacelocalization of the neurite mitogen. J.Cell Biol. 84, 767–778.

6. Grinspan, J.B., Marchionni, M.A.,Reeves, M., Coulaloglou, M., andScherer, S.S. (1996). Axonal interactionsregulate Schwann cell apoptosis indeveloping peripheral nerve: neuregulinreceptors and the role of neuregulins. J.Neurosci. 16, 6107–6118.

7. Syroid, D.E., Maycox, P.R., Burrola, P.G.,Liu, N., Wen, D., Lee, K.F., Lemke, G.,and Kilpatrick, T.J. (1996). Cell death inthe Schwann cell lineage and itsregulation by neuregulin. Proc. Natl.Acad. Sci. USA 93, 9229–9234.

8. Vartanian, T., Goodearl, A., Viehover, A.,and Fischbach, G. (1997). Axonalneuregulin signals cells of theoligodendrocyte lineage throughactivation of HER4 and Schwann cellsthrough HER2 and HER3. J. Cell Biol.137, 211–220.

9. Buonanno, A., and Fischbach, G.D.(2001). Neuregulin and ErbB receptorsignaling pathways in the nervoussystem. Curr. Opin. Neurobiol. 11,287–296.

10. Riethmacher, D., Sonnenberg-Riethmacher, E., Brinkmann, V., Yamaai,T., Lewin, G.R., and Birchmeier, C.(1997). Severe neuropathies in mice withtargeted mutations in the ErbB3receptor. Nature 389, 725–730.

11. Morris, J.K., Lin, W., Hauser, C.,Marchuk, Y., Getman, D., and Lee, K.-F.(1999). Rescue of the cardiac defect inErbB2 mutant mice reveals essentialroles of ErbB2 in peripheral nervoussystem development. Neuron 23,273–283.

12. Woldeyesus, M.T., Britsch, S.,Riethmacher, D., Xu, L., Sonnenberg-Riethmacher, E., Abou-Rebyeh, F.,Harvey, R., Caroni, P., and Birchmeier, C.(1999). Peripheral nervous systemdefects in erbB2 mutants following

genetic rescue of heart development.Genes Dev. 13, 2538–2548.

13. Garratt, A.N., Voiculescu, O., Topilko, P.,Charnay, P., and Birchmeier, C. (2000). Adual role of erbB2 in myelination and inexpansion of the Schwann cell precursorpool. J. Cell Biol. 148, 1035–1046.

14. Michailov, G.V., Sereda, M.W.,Brinkmann, B.G., Fischer, T.M., Haug, B.,Birchmeier, C., Role, L., Lai, C., Schwab,M.H., and Nave, K.-A. (2004). Axonalneuregulin-1 regulates myelin sheaththickness. Science 304, 700–703.

15. Lyons, D.A., Pogoda, H.-M., Voas, M.G.,Woods, I.G., Diamond, B., Nix, R., Arana,N., Jacobs, J., and Talbot, W.S. (2005).erbB3 and erbB2 are essential forSchwann cell migration and myelinationin zebrafish. Curr. Biol. 15, 513–524.

16. Nakao, J., Shinoda, J., Nakai, Y., Murase,S.-i., and Uyemura, K. (1997). Apoptosisregulates the number of Schwann cellsat the pre-myelinating stage. J.Neurochem. 68, 1853–1862.

17. Winseck, A.K., Caldero, J., Ciutat, D.,Prevette, D., Scott, S.A., Wang, G.,Esquerda, J.E., and Oppenheim, R.W.(2002). In vivo analysis of Schwann cellprogrammed cell death in the embryonicchick: regulation by axons and glialgrowth factor. J. Neurosci. 22,4509–4521.

18. Gilmour, D.T., Maischein, H.-M., andNusslein-Volhard (2002). Migration andfunction of a glial subtype in thevertebrate peripheral nervous system.Neuron 34, 577–588.

19. Mahanthappa, N.K., Anton, E.S., andMatthew, W.D. (1996). Glial growth factor2, a soluble neuregulin, directlyincreases Schwann cell motility andindirectly promotes neurite outgrowth. J.Neurosci. 16, 4673–4683.

The Scripps Research Institute, 10550N. Torrey Pines Road, La Jolla, CA92037, USA. E-mail: carylai@scripps.edu

DOI: 10.1016/j.cub.2005.04.024

Samuel Cotton, David W. Rogersand Andrew Pomiankowski

Models of sexual selection showthat female mate preferences forelaborate male sexual ornamentsand displays can evolve when theexpression of courtship traitsreflects male fitness, throughvariation in attractiveness or

viability, or both [1,2]. Althoughnumerous studies claim to findevidence supporting thesehypotheses, the evidence isusually based on measures of oneor a few components of fitness[3]. Although indicative, this is notconvincing, because thecorrelation between componentsand fitness itself is variable and

Sexual Selection: The Importanceof Long-Term Fitness Measures

New results from a 20-year study of free-living song sparrows confirmthat attractive males contribute more offspring than less attractivemales. They also reveal that the offspring of preferred males producemore descendents themselves. Females prefer males with a large songrepertoire, which further work shows is a condition-dependentindicator of male quality.