in vitro analyses of interactions between olfactory receptor growth cones and glial cells that...

In Vitro Analyses of Interactionsbetween Olfactory Receptor GrowthCones and Glial Cells That Mediate

Axon Sorting and Glomerulus Formation

ERIC S. TUCKER,1,2 LYNNE A. OLAND,2AND LESLIE P. TOLBERT1,2*

1Department of Cell Biology and Anatomy, University of Arizona,Tucson, Arizona 85724-5044

2Arizona Research Laboratories Division of Neurobiology, University of Arizona,Tucson, Arizona 85721-0077

ABSTRACTDuring development, the axons of olfactory receptor neurons project to the CNS and

converge on glomerular targets. For vertebrate and invertebrate olfactory systems, neuron–glia interactions have been hypothesized to regulate the sorting and targeting of olfactoryreceptor axons and the development of glomeruli. In the moth Manduca sexta, glial reductionexperiments have directly implicated two types of central olfactory glia, the sorting zone- andneuropil-associated glia, in key events in olfactory development, including axon sorting andglomerulus stabilization. By using cocultures containing central olfactory glial cells andexplants of olfactory receptor epithelium, we show that olfactory receptor growth coneselaborate extensively and cease advancement following contact with sorting zone- andneuropil-associated glial cells. These effects on growth cone behavior were specific to centralolfactory glia; peripheral glial cells of the olfactory nerve failed to elicit similar responses inolfactory receptor growth cones. We propose that sorting zone- and neuropil-associated glialcells similarly modify axon behavior in vitro by altering the adhesive properties and cytoskel-eton of olfactory receptor growth cones and that these in vitro changes may underlie func-tionally relevant changes in growth cone behavior in vivo. J. Comp. Neurol. 472:478–495,2004. © 2004 Wiley-Liss, Inc.

Indexing terms: Manduca sexta; olfaction; development; neuron–glia interaction

Developing olfactory pathways provide models forstudying growth cone guidance in systems in which thetargeting of sensory axons does not depend absolutely onthe location of sensory neurons in the periphery. The cellbodies of olfactory receptor neurons (ORNs) that expressthe same odorant receptor are widely distributed withinbroad zones of olfactory epithelia (Ressler et al., 1993;Vassar et al., 1993; Clyne et al., 1999; Vosshall et al.,1999), yet their axons converge on the same glomerulartargets in the brain (Vassar et al., 1994; Ressler et al.,1994; Mombaerts et al., 1996; Vosshall et al., 2000; Gao etal., 2000). The initial segregation of olfactory informationfor further processing is, therefore, dependent on theproper sorting of ORN axons. Genetic swapping of odorantreceptors in mice leads to the mistargeting of ORN axonsand demonstrates that odorant receptors are involved inaxon pathfinding (Mombaerts et al., 1996; Wang et al.,1998; Bozza et al., 2002). However, axon targeting in

olfactory systems also appears to depend on the coordi-nated expression of many different cell-surface (Puche etal., 1996; Treloar et al., 1997; Yoshihara et al., 1997; St.John and Key, 1999, 2001; Mori et al., 1999; Walz et al.,2002) and extracellular (Gong and Shipley, 1996; Treloaret al., 1996; Crandall et al., 2000) guidance molecules.

Grant sponsor: National Institutes of Health; Grant number: DC04598(L.P.T.).

*Correspondence to: Leslie P. Tolbert, Arizona Research LaboratoriesDivision of Neurobiology, University of Arizona, P.O. Box 210077, Tucson,AZ 85721-0077. E-mail: [email protected]

Received 28 August 2003; Revised 22 December 2003; Accepted 26 De-cember 2003

DOI 10.1002/cne.20058Published online the week of March 29, 2004 in Wiley InterScience

(www.interscience.wiley.com).

THE JOURNAL OF COMPARATIVE NEUROLOGY 472:478–495 (2004)

© 2004 WILEY-LISS, INC.

Although our understanding of olfactory axon guidancehas grown, the intercellular interactions that regulateaxon guidance and target selection remain poorly under-stood. The olfactory system of the moth Manduca sextashares many neuroanatomical and physiological similari-ties with vertebrate olfactory systems (Hildebrand andShepherd, 1997) and is well suited for studies that exam-ine the cellular mechanisms underlying critical aspects ofolfactory pathway development (Oland and Tolbert, 1996;Hildebrand et al., 1997; Tolbert, 1998). In Manduca, glialcells are required for sorting ORN axons into fascicles thatterminate in particular glomeruli (Rossler et al., 1999)and for stabilizing ORN axon terminals during the forma-tion of glomeruli (Oland and Tolbert, 1988; Oland et al.,1988; Baumann et al., 1996).

The current study examines interactions between cul-tured ORN axons and glial cells from two regions of theManduca olfactory pathway in which previous studieshave demonstrated important axon–glia interactions. Inother systems, growth cone form provides insights intogrowth cone behavior. In the mouse retinofugal pathway,for instance, elaborations in growth cone morphology oc-cur most frequently at the chaismatic midline, where ret-inal axons “decide” whether to cross or turn after encoun-tering a palisade of radial glia (Mason and Erskine, 2000).We hypothesized that glial cells of the sorting zone (SZ)region of the Manduca antennal nerve (AN) and the neu-ropil (NP) region of the antennal lobe would directly in-fluence the morphology and behavior of the growth conesof ORN axons from the antenna. To test for such an effect,we developed a coculture assay that allowed us to charac-terize glial influences on ORN axon behavior. Our resultsindicate that contact with SZ and NP glial cells leads toelaboration of growth cone morphology and reduction ingrowth cone motility. We speculate that contact with glialcells alters growth cone motility by changing growth coneadhesive properties and/or cytoskeletal dynamics and thatthese changes may underlie growth cone guidance behav-iors in situ. Some of this work has appeared previously inabstract form (Tucker et al., 2000, 2001).

MATERIALS AND METHODS

Animals

Manduca sexta (Lepidoptera: Sphingidae) were rearedon an artificial diet under a long-day photoperiod (17hours light, 7 hours dark) in environmental chambersmaintained at 25°C and 50–60% relative humidity. Underthese conditions, adult metamorphic development occursover 18 stages, each lasting for 1–4 days, starting atpupation and ending at eclosion of the moth. Pupae werestaged by examining morphological changes in externaladult structures visible beneath the pupal cuticle afterfiber optic illumination (Tolbert et al., 1983; Oland andTolbert, 1987; Dubuque et al., 2001).

Removal of antennal input

In some animals, one antennal anlage was removedduring the first stage of adult metamorphic development,before the birth of olfactory receptor neurons that residein the antennal receptor epithelium (Sanes and Hilde-brand, 1976). An opening was made in the cuticle coveringthe base of the antenna and the exposed antennal anlagewas excised with forceps. The inner surface of the anten-

nal trough was then scraped clean and the opening to thehead plugged with melted wax, preventing axons from thesurviving distal antennal segments from reaching thebrain. Surgically treated pupae were allowed to develop inan environmental chamber until they reached early stage7 of adult metamorphic development. In Manduca, anten-nal innervation of the primary olfactory system is strictlyipsilateral, and the removal of one antennal anlage doesnot lead to aberrant innervation from the contralateralantenna (Sanes et al., 1977; Kent, 1985). Therefore, thissurgical procedure completely deprives the antennal lobeon the operated side of its normal antennal (olfactory)input, leaving only minor sensory inputs from the labialpalp unperturbed (Kent et al., 1986, 1999).

Preparation of cultures

Explants of olfactory receptor epithelium. Wholeantennae were removed from the antennal troughs ofstage 4 female pupae and placed into a 35-mm dish con-taining sterile phosphate-buffered saline (PBS), pH 7.4.Antennae were filleted along a line that visibly markedthe border between the olfactory receptor and nonreceptorepithelia. Dissected olfactory receptor epithelia weretransferred to a sterile polystyrene test tube containingapproximately 400 �l of PBS on ice. Receptor epitheliumwas subjected to mild enzymatic digestion in a Ca2�- andMg2�-free Hanks’ balanced salt solution (21250-014;Gibco, Grand Island, NY) containing 0.05 mg/ml collage-nase (LS004196; Worthington, Freehold, NJ) and 0.2mg/ml dispase II (165859; Boehringer Mannheim, Mann-heim, Germany). The enzyme-treated antennal tissue wasgently triturated with a fire-polished Pasteur pipette, lay-ered onto 6 ml of culture saline, and allowed to settle bygravity. Large bits of antennal tissue were removed, leav-ing behind pieces approximately 100–200 �m in diameter.Culture saline was then aspirated, leaving aggregatedtissue behind, and the step was repeated first with culturesaline and again with culture medium. Explants wereevenly suspended in fresh culture medium and plated in100-�l aliquots into the wells of premade culture dishes.One antenna provided sufficient explants for three wells.Culture dish wells were made by attaching coverslipsbeneath 8-mm-diameter holes drilled into the bottoms of35-mm Falcon dishes. Dish wells were coated with a solu-tion containing 400 �g/ml concanavalin A (C2010; Sigma,St. Louis, MO) and 4 �g/ml laminin (40232; CollaborativeResearch, Bedford, MA) and rinsed thoroughly with astream of sterile water prior to cell plating. After platingof explants, culture dishes were sealed with Parafilm toprevent evaporation and incubated in a 26°C humidifiedincubator with room air.

Glial cell cultures and cocultures. Acutely isolatedglial cells were prepared by using essentially the methodsdescribed previously by Lohr et al. (2002). Briefly, femalepupae from stages late 6 to early 7 of metamorphic adultdevelopment were cooled on ice, and brains were dissectedinto 35-mm culture dishes containing sterile dissectingmedium. Antennal lobes were desheathed, the neuronalcell-body packets were removed, and the antennal lobeswere separated from the AN, leaving only glia and pre-sumably a small number of tracheolar cells intact. Differ-ential dissection was then used to yield one or more of thethree distinct populations of olfactory glia (see Fig. 1): 1)SZ glia, 2) antennal lobe NP glia, or 3) AN glia. Tissue wasdigested with 0.1 mg/ml papain (5125; Calbiochem, La

479GLIA INFLUENCE OLFACTORY GROWTH CONE BEHAVIOR

Jolla, CA) in simple salt solution for 4 minutes at 37°Cprior to trituration. Dissociated cells in suspension werethen layered onto recovery solution in a 15-ml Falcontube, and 200 U of DNase (D4263; Sigma) in simple saltsolution were added to the top layer of suspended cells.Cells were centrifuged at 500g in a tabletop centrifuge for4 minutes. The resultant pellet was resuspended in cul-ture medium and centrifuged as described above. Cellswere again resuspended in culture medium and platedinto 8-mm wells (100 �l/well) constructed in 35-mm petridishes. One animal was used per dish of glial cells plated.

For cocultures, glial cells were resuspended in culturemedium and added to explant cultures previously grownfor 1 day in vitro (1 DIV). For each dish, 50 �l of culturemedium were removed from the 100-�l bubble of mediumoverlying the cultured explants and then gently replacedwith 50 �l of the glial suspension. After cell plating, theculture dishes were sealed with Parafilm and incubatedfor 2 hours. Cultures were then flooded with at least 1 mlof culture medium.

Unafferented antennal lobes, from animals that hadhad antennal anlagen surgically removed at stage 1, weredissected as described above at early stage 7 on the day ofthe experiment. Afferented antennal lobes were removedfrom the unoperated side of experimental animals, differ-entially dissected to obtain NP tissue, and dissociated andplated separately from the unafferented lobes.

Tissue culture solutions

Culture saline (Oland et al., 1996). The followingingredients were used: 149.9 mM NaCl, 3 mM KCl, 3 mMCaCl2, 0.5 mM MgCl2, 10 mM TES, 11 mM D-glucose, 3g/liter lactalbumin hydrosylate (11800-026; Gibco), 2.5g/liter TC yeastolate (255772; Difco, Detroit, MI), 10%fetal bovine serum (FBS; Hyclone, Logan, UT), 100 U/mlpenicillin, and 100 �g/ml streptomycin, pH 7.0, 360mOsm.

Culture medium (supplemented Leibovitz’s L-15 cul-

ture medium; Lohr et al., 2002). The following ingredi-ents were added to 500 ml of L-15: 50 ml fetal bovineserum (FBS), 185 mg �-ketoglutaric acid, 200 mg D-(–)-fructose, 350 mg D-glucose, 335 mg DL-malic acid, 30 mgsuccinic acid, 1.4 g lactalbumin hydrosylate, 1.4 g TCyeastolate, 0.1 mg niacin, 30 mg imidazole, 500 �g 20-hydroxyecdysone (H5142; Sigma), 100 U/ml penicillin, 100�g/ml streptomycin, and 2.5 ml stable vitamin mix (SVM).A 5-ml stock solution of SVM consisted of 15 mg asparticacid, 15 mg cystine, 5 mg �-alanine, 0.02 mg biotin, 2 mgvitamin B12, 10 mg inositol, 10 mg choline chloride, 0.5mg lipoic acid, 5 mg p-aminobenzoic acid, 25 mg fumaricacid, 0.4 mg coenzyme A, 15 mg glutamic acid, and 0.5 mgphenol red. The pH was adjusted to 7.0, and the osmolar-ity was raised to 390 mOsm with D-glucose prior to sterilefiltration.

Simple salt solution. The following ingredients wereused: 160 mM NaCl, 6 mM KCl, 78.8 mM D-glucose, 10mM HEPES, 100 U/ml penicillin, and 100 �g/ml strepto-mycin, pH 7.0, 420 mOsm.

Recovery solution. The following ingredients wereused: 50% (v/v) culture saline and 50% (v/v) simple saltsolution, pH 7.0, 380 mOsm.

Dissecting medium. The following ingredients wereused: 50% Leibovitz’s L-15 (41300-039; Gibco), 25% (v/v)culture saline, 25% simple salt solution with 5 mM EDTA,and 18 mM D-glucose, pH 7.0, 360 mOsm.

Live-cell microscopy

Explant cultures grown for 1 DIV or cocultures grownfor 2 hours after the plating of glial cells were used fortime-lapse differential interference contrast (DIC) imag-ing experiments. Before being imaged, cultures wereflooded with 3 ml of culture medium and placed on atemperature-controlled microscope stage to equilibrate.The imaging system included a fixed-stage microscope(BX50WI; Olympus, Tokyo, Japan) equipped with DICand epifluorescence optics, long-working-distance water-immersion objectives, Uniblitz (Vincent Associates, Roch-ester, NY) shutters, a Ludl motorized z-drive, a cooledCCD camera (KAF1400; Photometrics, Tucson, AZ), acomputer with dual 17-inch monitors, and SimplePCI(Compix Inc., Cranberry Township, PA) acquisition andanalysis software. A 12-V/100-W halogen bulb filtered by agreen optical lens (543 nm) was used for brightfield illu-mination during DIC imaging. A constant dish tempera-ture of 25°C was maintained with the aid of a temperaturecontroller (TC202A; Harvard Apparatus, Holliston, MA)combined with an open perfusion microincubator(PDMI-2; Harvard Apparatus). An insulated chamber en-closed the microscope stage and provided a humidifiedinternal environment that limited disturbances caused byambient temperature fluctuations and air currents.

After an area within the dish was selected for imaging,a thin layer of canola oil was applied over the culturemedium surface to prevent evaporation of the culture me-dium and to provide thermal insulation throughout theimaging experiment. High-magnification images were col-lected with a �60 objective at 20-minute intervals for upto 24 hours. To overcome the small drift in focus thatoccurred over the imaging period, series of five images atsuccessive focal planes 1 �m apart were collected at eachtime point. Only the in-focus images at each time pointwere used.

Rate analysis

Time-lapse movie sequences were used to measure thedistances that individual axons grew. Five axons fromeach coculture condition were selected for measurement,provided that they met the following criteria: 1) Each axoncould clearly be identified as an individual, 2) the growthof the axon could be monitored before and after contactwith glial cells, and 3) at least one image from eachz-series was in focus for every time point of the moviesequence. Reference markers were placed at the distal tipof growing axons for each time point of the collection, andthe distances between consecutive markers were mea-sured in micrometers using SimplePCI (see Fig. 5). Dis-tances, positive for growth and negative for retraction,were recorded between adjacent frames of the selectedmovie sequence. At each point in time, the net distancegrown by each individual axon was determined by sum-ming all distances recorded prior to that time point. Netdistances were plotted as a function of time for each mea-sured axon. Growth rates were determined by calculatingthe slope of regression lines fit to precontact and postcon-tact curves for each plot.

Immunocytochemistry, phalloidin staining,and confocal microscopy

Labeling of microtubules and F-actin was performed asdescribed previously (Tucker and Tolbert, 2003). Briefly,

480 E.S. TUCKER ET AL.

cultures were rinsed with microtubule-stabilizing buffer(K-PIPES buffer; 80 mM PIPES-KOH, pH 6.8, 5 mMEGTA, 2 mM MgCl2) and then fixed and extracted for 30minutes in K-PIPES buffer containing 0.5% glutaralde-hyde and 0.1% Triton X-100. Autofluorescence wasquenched by rinsing cultures for 3 � 5 minutes in PBScontaining 1 mg/ml NaBH4. Cultures were then rinsedand blocked for 1 hour in PBS containing 0.2% fish skingelatin and 0.1% Triton X-100 (blocking buffer) and incu-bated with primary antibody (anti-�-tubulin; T9028;Sigma) diluted 1:800 in blocking buffer for 2 hours at roomtemperature. After being rinsed, cultures were incubatedfor 1.5 hours in a 1:1,000 dilution of secondary antibody(goat anti-mouse Cy3; Jackson Immunoresearch, WestGrove, PA) in blocking buffer containing approximately2.5 U/ml, or 83 nM, Alexa-488 conjugated phalloidin (A-12379; Molecular Probes, Eugene, OR), to label microtu-bules and F-actin simultaneously. Cultures were thenrinsed and mounted with coverslips in an aqueous polyvi-nyl alcohol (PVA)-based medium that included 1,4-diazobicyclo[2,2,2]octane (DABCO) to limit photobleach-ing. Controls for nonspecific labeling were performed byomitting the primary antibody.

Laser scanning confocal microscopy was performed witha Nikon (Tokyo, Japan) PCM 2000 system equipped witha Nikon E800 microscope, a 50-mW argon laser, 4-mWgreen and 10-mW red HeNe lasers, and a computer run-ning SimplePCI acquisition and analysis software. Prep-arations were simultaneously illuminated with argon andgreen HeNe lasers, using appropriate dichromatic filtersfor multichannel collection. Serial optical sections weretaken at sequential depths, 0.5 �m apart, and stored as astack of optical images. If needed, confocal images weremanipulated for brightness, contrast, and intensity withCorel Photo Paint 9 or SimplePCI and prepared in figureformat with Corel Draw 9.

Growth cone sampling and statisticalanalyses

Cultures stained for tubulin and F-actin were used forlarge-scale comparisons of growth cone morphology acrossmultiple dishes. Three dishes were prepared for everycondition within each experiment. Dishes were codedprior to collection of confocal images, so that the observerwould not know which type of glial cells they contained. Atleast three explants displaying radial growth and a highnumber of surrounding glial cells were randomly selectedfrom each dish. The perimeters of the selected explantswere sampled with a �60 objective, and images (�30/dish)were saved. Unmanipulated confocal images were printed,and individual growth cones were scored based on a qual-itative scale of morphological diversity. Scoring accuracywas confirmed by comparing independent results from twoseparate observers. After they were scored, dishes wererecoded according to the original experimental condition.Results from similarly treated dishes were combined, be-cause in all cases they were found to be internally consis-tent (i.e., distributions of growth cone scores were statis-tically similar to one another). Two-by-two contingencytables containing total numbers of simple and complexgrowth cones from each experimental condition were con-structed and tested for statistical differences by Fisher’sexact test (Fisher, 1925). Qualitative variables from testedconditions were defined as being statistically different atP � .05.

Data from preliminary experiments were used to deter-mine that approximately 30 images/dish provided an ad-equate sample size (number of growth cones) for studies ofappropriate power. Five independent experiments thatincluded at least two SZ coculture dishes gave resultsidentical to those of the experiment analyzed in full detailin Figure 8.

RESULTS

Development of SZ glia and the generationof cocultures containing antennal explants

and glial cells

In Manduca, the first cohort of ORN axons extends intothe antennal (olfactory) lobe of the brain during stage 3 ofadult metamorphic development (Sanes and Hildebrand,1976; Oland and Tolbert, 1987) and triggers the localizedproliferation and migration of a group of central glia awayfrom the antennal lobe neuropil to the most proximalportion of the AN, called the SZ (Rossler et al., 1999). Glialcells rapidly fill the SZ during the ensuing developmentalstages (Fig. 1A) and subsequently influence the fascicula-tion patterns of ORN axons en route to developing glomer-uli within the antennal lobe (Rossler et al., 1999). NP glia,apparently derived from the same set of central precursorsin the antennal lobe, are initially confined to a rind sur-rounding the antennal lobe NP. They respond to ORNaxon ingrowth by migrating and extending processes tosurround “protoglomeruli” formed by the ORN axons (Fig.1A; Oland and Tolbert, 1987).

The anatomical positioning of glial cells within the SZand the antennal lobe, together with the absence of neu-ronal cell bodies in the NP, makes it possible to establishhighly purified glial cultures by differential dissection anddissociation (Fig. 1B). Glial cells were harvested for cul-ture at early stage 7, when glial density within the SZnears its peak and before peripheral glial cells migratingdown the antennal nerve from the antenna reach thedistal edge of the SZ (Fig. 1A; Rossler et al., 1999). Forcocultures, dissociated glial cells were introduced to pre-viously plated explants of olfactory receptor epitheliumthat had been isolated from stage 4 antennae and grownfor 1 DIV (Fig. 1B).

Behavior and morphology of ORN axonsin vitro

In vitro, hundreds of olfactory receptor axons extendradially from cultured explants of developing olfactoryreceptor epithelium (Fig. 1C). The robust nature of axonaloutgrowth, particularly within the first 48 hours of cul-ture, allowed us to study dynamic changes in axon behav-ior. Because a dense meshwork of axonal processes sur-rounded explants, individual axons were distinguishableonly at the outermost fringe of axon outgrowth (Fig. 1C,box). Time-lapse imaging showed that axons elongatedwhile extending and retracting fine branches from theirgrowth cones (not shown). Thin filopodial processes be-came engorged with cytoplasm during the advancement ofmotile branches, whereas nonmotile branches were re-tained but left with little remaining cytoplasm. Most ax-ons had simple, bullet-shaped growth cones (Fig. 1D, ar-rows). Infrequently, large, flattened growth cones wereseen at the tips of growing axons (Fig. 1D, arrowheads).Multiple branches often arose from the periphery of flat-


Figure 1

tened growth cones. One or two nascent branches wouldsubsequently extend away from the growth cone, leavingthe lamellar region behind.

Glial cells from the SZ and antennal lobe NP had verysimilar morphologies in vitro. Glial cell bodies were rela-tively small (�15 �m) and had thin processes with a rangeof branching patterns (Fig. 1E–G). Their small size anddistinctive morphologies allowed for easy identificationand distinction from ORNs and ORN axons.

Axon behavior after contact with SZ-derivedglial cells

Cocultures containing explants of olfactory receptor ep-ithelium and either SZ or NP glial cells were prepared bydifferentially dissecting antennal lobes, separately collect-ing and dissociating SZ and NP tissue, and separatelyadding SZ or NP glial cells to preplated explant cultures.Photographs were collected in 20-minute intervals to cap-ture global changes in axon behavior that occurred overmany hours of time-lapse imaging. (Full movie can beviewed at www.mrw.interscience.wiley.com/suppmat/0021-9967/suppmat/index.html. Although this strategydid not permit the observation of rapid changes in growthcone dynamics, it was chosen to allow long-term observa-tion of the axon growth and growth cone morphology thattranspired both before and after contact with glial cells. Incocultures, receptor axons that did not encounter glialcells grew in a pattern indistinguishable from axons ofexplants grown alone. Single filopodial contact with iso-lated SZ glial cells, however, was sufficient to influencegrowth cone morphology and axon behavior markedly.Growth cones contacting SZ glial cells often developedlarge, lamellar profiles at the point of axon contact (Fig.2A). Elaboration of growth cones began within minutes ofglial contact and continued for many hours (Fig. 2A, ar-rowheads). Small branches often extended from glial cell-contacting growth cones (Fig. 2A, open arrowheads inframes 5–7), but the flattened growth cone morphologieswere usually retained. Lamellar processes extended on,but never past, glial processes or cell bodies (Fig. 2A,arrowheads). Although growth cone flattening was the pre-dominant response to glial cell contact, glia-mediated alter-ations in growth cone morphology were variable. Somegrowth cones branched near their base, whereas others re-mained simple after contacting SZ glial cells (Table 1).

Contact-mediated branching often resulted in the formationof highly branched growth cones (Fig. 2B). Axon branchingusually occurred at the point of glial cell contact, but it wasnot restricted to the growth cone tip (Fig. 2B, arrowhead).Axon branching continued for several hours after contact,and growth cones usually remained closely associated withthe glial cells that they encountered (Fig. 2B).

Regardless of the morphological response, the majorityof encounters (88%) between individual axons and SZ glialcells resulted in the cessation of growth cone advancement(Fig. 2A, Table 1). Growth cones often remained activeafter glial contact by extending and retracting filopodia,lamellar processes, and/or growth cone branches, yet theyroutinely failed to advance. A relatively small percentageof axons (12%) did continue to elongate past SZ glial cells(Fig. 2B, Table 1), but those axons never had flattenedgrowth cones. Thus, after contact with glial cells, growthcones that stopped flattened, branched, or remained sim-ple, whereas growth cones that advanced remained simpleor branched but never flattened. Elongating axons occa-sionally branched after glial contact before extending overor around the impeding glial cells (Fig. 2B). Axons thatbranched but continued to elongate were counted as “elon-gating” axons in Table 1. In all recordings analyzed, axonsthat did not encounter glial cells continued to elongate inclose proximity to axons that stalled following glial con-tact. This finding precludes the possibility that growthcone motility was unintentionally compromised duringthe process of time-lapse imaging and underscores theimportance of glial cell contact in the alteration of growthcone behavior.

Axon behavior after contact with NP-derived glial cells

NP glial cells evoked the same range of morphological andbehavioral changes in ORN axons as did SZ glial cells (Table1). Again, ORN growth cones flattened (Fig. 3A), branched(Fig. 3B; see also Fig. 3C, open arrows and open arrow-heads), or remained simple in morphology (Fig. 3C, solidarrows and solid arrowheads) following contact with NP glialcells. The distributions of growth cone responses to SZ andNP glial cells were not significantly different (P � .132) by 2

analysis. The contact-mediated formation of flattenedgrowth cones occurred identically to such formation follow-ing SZ glial cell contact, with lamellar processes extendingrapidly from stabilized sites of axon–glial cell contacts. Elab-oration again occurred in direct apposition to the processesor the cell bodies of NP glial cells, often resulting in exten-sion onto but not past NP glial cell processes (Fig. 3A, ar-rowheads). Axons that branched following NP glial contactappeared similar in morphology to those branching after SZglial contact. As with SZ contacts, growth cone branchesextended both on and away from the surfaces of NP glialcells (Fig. 3B). Axons typically remained closely associatedwith glial processes while branching (Fig. 3B; see also solidarrowheads; Fig. 3C, open arrowheads). Some axons re-mained tipped with simple, bullet-shaped growth cones aftercontacting SZ (not shown) and NP (Fig. 3C, solid arrow-heads) glial cells. These simply tipped axons maintainedtheir precontact morphology but occasionally were engorgedwith cytoplasm, becoming thicker in appearance followingglial cell contact.

Encounters between growth cones and glial cells werelong-lived, often lasting many hours after the initialfilopodial contact. Most axons (87%) that contacted NP

Fig. 1. Generation of cultures containing explants of olfactory recep-tor epithelium and glial cells from the olfactory pathway of Manducasexta. A: Schematic diagram illustrating changes in glial number andposition that occur during development. After the first ORN axons ar-rive, sorting zone (SZ) glia (medium gray) proliferate and migrate to fillthe base of the antennal nerve. Glia surrounding the antennal lobeneuropil (NP; light gray) proliferate, migrate, and extend processes toenvelop nascent glomeruli. Antennal nerve (AN) glia (light gray), born inthe antenna, migrate toward the antennal lobe (AL) and fill the AN.B: Summary diagram of coculture methods. On day 1, stage 4 antennaeare dissociated to yield explants of olfactory receptor epithelium. On day2, tissue from the SZ, AL, or AN, is dissociated from early stage 7 brains,and the resulting glial cells are plated onto explant cultures. C: Low-magnification view of an explant grown for 1 DIV. D: High-magnificationview of the boxed region in C showing the distal fringe of ORN axonoutgrowth. Arrows indicate simple growth cones; arrowheads indicateflattened growth cones. E–G: Freshly cultured SZ glial cells. Scale bar inC � 100 �m; bar in D � 20 �m; bar in E � 20 �m for E–G. [Color figurecan be viewed in the online issue, which is available at http://www.interscience.wiley.com]


glial cells failed to elongate beyond the contacted glialcells during our recordings. As for contacts with SZ glialcells, the cessation of growth cone advancement follow-ing contact with NP glial cells was not strictly corre-lated with specific changes in growth cone morphology

(Fig. 3A–C). The overall percentage of axons stoppingafter contact with glial cells was nearly identical for SZ(88%) and NP (87%) encounters. Thus, contact with SZand NP glial cells similarly influenced growth cone mo-tility.

Fig. 2. ORN growth cone encounters with SZ glial cells. A: Framesfrom a movie sequence showing a typical growth cone response tocontact with SZ glial cells. (Full movie can be viewed atwww.mrw.interscience.wiley.com/suppmat/0021-9967/suppmat/index.html.) Shortly after contact (3:20), the growth cone flattens, elabo-rates, and stops advancing. B: Movie frames showing a less frequent

type of response following contact with SZ glial cells. Growth conebranches after contact (2:20), pauses, then continues to advance overand around the pair of glial cells. Asterisks indicate glial cell bodies.Arrows mark axon shaft, and solid arrowheads mark axon tip. Openarrowheads in A mark the tip of a growth cone branch. Time stampsare in hours and minutes. Scale bars � 10 �m.


Receptor axons respond differently to ANglial cells

The axonal responses to SZ and NP glial cells were notgeneralized reactions to cell contact but rather specificaxon behaviors triggered by contact with centrally derivedglial cells. Peripheral glial cells from the antenna elicitedan entirely different behavior in ORN axons after growthcone contact (Fig. 4; Tucker and Tolbert, 2003). AN glialcells were separated from SZ and NP glia by isolating andculturing the distal portion of early stage 7 antennalnerves. AN glial cells had distinctive morphologies, in-cluding oblong cell bodies and long, stout processes (Fig.4A, asterisk). Unlike SZ and NP glial cells, AN glial cellsoften initiated encounters with ORN axons by extendingtheir processes to meet advancing axonal growth cones(Fig. 4B, double arrowhead). After contact with AN glialcells, ORN axons elongated in direct apposition with glialcell processes (Fig. 4C–F, solid arrowheads) or elongatedon the substrate adjacent to AN glial cells (Fig. 4C–F,open arrowheads), without significant changes in growthcone morphology. The ORN axon shown in the inset inFigure 4F extended for 40 �m in close apposition to theglial process. In contrast, growth cones that advancedafter contact with SZ and NP glial cells never maintaineddirect associations with glial processes. Therefore, growthcone contact with AN glial cells affected growth cone mo-tility and morphology differently from contact with SZ andNP glial cells.

Rate analysis

To analyze contact-induced changes in axon elongation,sample recordings were selected from SZ and NP cocul-tures, and the net distances that individual axons grewwere measured and plotted as a function of time. Tenconsecutive frames from an NP coculture recording illus-trate how axon elongation was analyzed in the currentstudy (Fig. 5). Axon growth was measured by summingthe distances between markers placed at the growth conetip in consecutive movie frames (Fig. 5, black dots). Netdistances were determined for each point in time by sum-ming all distances recorded up to that point. In Figure 5,a reference marker corresponding to the original positionof the axon tip was placed in each frame to aid in visual-ization of axon growth. The same strategy was used tomeasure growth cone branches (Fig. 5, white dots), pro-vided that they unambiguously arose from their parentaxon and persisted for longer than 1 hour after formation.

Contact with SZ (Fig. 6A) and NP (Fig. 6B) glial cellsusually prevented forward axon progression past the pointof contact (Fig. 6, arrows), making the rate of axon elon-gation markedly different before and after encounterswith glial cells (Fig. 6A,B, tables). Axons behaved simi-

larly whether growth cones flattened (Fig. 6A, axons 1–4;Fig. 6B, axons 2, 3), branched (Fig. 6A, axon 5; Fig. 6B,axon 3), or remained simple (Fig. 6A, axon 4b; Fig. 6B,axons 1, 4, 5) after contacting SZ and NP glial cells. Theinitial growth seen immediately following glial cell contactusually represented the expansion of lamellar processes orthe formation of growth cone branches. About half of theanalyzed axons branched during the recording, some be-fore and some after contact with glial cells. Growth rateswere reported only for those branches that formed prior toglial cell contact (Fig. 6A, axons 4, 5), so that before andafter comparisons could be made. Axon 3 of the NP plots(Fig. 6B) represents the complete measurement of theaxon–glial cell encounter depicted in Figure 5. The saw-toothed pattern appearing in the plot after glial contact(Fig. 6B, axon 3, arrows) corresponded to the extensionand retraction of the axon tip seen following contact (Fig.5, arrowheads). The branch that formed after contact (Fig.5, white dots) temporarily continued to advance, while theoriginal axon tip stalled (Fig. 5, black dots; Fig. 6B, axon3). On average, the rate of axon elongation after SZ glialcontact was 2.8% of the average precontact rate (Fig. 6A,table). Similarly, axon elongation slowed to an average of9.1% of the original rate following contact with NP glialcells (Fig. 6B, table).

In stark contrast, growth cone motility was essentiallyunchanged following contact with AN glial cells (Fig. 6C;Tucker and Tolbert, 2003). Axon branches were measuredfor the encounter depicted in Figure 4. In this recording,glial contact occurred at the 1-hour time point (Fig. 4B,double arrowhead; Fig. 6C, downward arrow). After con-tact, the axon branched and continued to elongate. Onebranch contacted a glial cell process (Fig. 4C, arrowhead;Fig. 6C, upward arrow) and subsequently extended incontact with the process (Figs. 4C–F, solid arrowheads;Fig. 6C, branch b), whereas the second branch elongatedadjacent to the glial cell (Figs. 4C–F, open arrowheads;Fig. 6C, branch a). Axon branch b became obscured by theglial cell after the 5-hour time point (Fig. 4F) and, there-fore, was not measured further. Branch a was measureduntil the end of the recording. The average rate of axonelongation following glial contact for branches a and b was61% of their precontact rate. Branch b actually advancedat a faster rate while growing on the glial process thanbefore it had contacted the glial cell. The rate of axonelongation after contact with an AN glial cell was mark-edly different from the elongation rates seen followingcontact with SZ and NP glia.

Morphological diversity of ORNgrowth cones

To analyze many more growth cones than was possiblewith live-cell imaging, we used cytoskeletal staining infixed cells to evaluate growth cone morphology. Images ofthe entire perimeter of randomly selected explants werecollected on the confocal microscope, and all isolatedgrowth cones were qualitatively scored according to theirmorphological appearance. Axons grown without glia weretipped with growth cones that exhibited a range in mor-phological diversity (Fig. 7). All growth cones could begrouped into two broad categories, those with “simple” andthose with “complex” morphologies. Simple growth coneshad either unbranched or branched microtubule-basedprocesses and were tipped by filopodial spikes (Fig. 7A).Complex growth cones had lamellar regions containing

TABLE 1. ORN Growth Cone Responses (%) to Contact With Sorting Zone-and Neuropil-Associated Glial Cells1

Growth conebehavior SZ glia (n � 50) NP glia (n � 40)

Flatten and stop 60 40Branch and stop 18 20Stay simple and stop 10 27Elongate 12 13

1Time-lapse microscopy revealed four categories of growth cone behaviors that followedcontact with sorting zone (SZ) and neuropil-associated (NP) glial cells. The totalnumber of growth cone encounters (n) is indicated for each glial cell type.


splayed microtubules and were surrounded by a densefringe of short, F actin-based filopodia (Fig. 7B). Size wasnot a determinant for categorization; complex growthcones in particular varied in both their length and theirwidth. The vast majority of growth cones (�85%) in glia-free cultures had simple morphologies (Figs. 8I, 9I).

Glial contact-mediated change in growthcone morphology

Cocultures containing either SZ or NP glial cells weresimultaneously prepared by using differentially dissectedtissue from the same experimental animals. After immu-

Fig. 3. ORN growth cone encounters with NP glial cells. A: Movieframes showing an ORN growth cone that flattened and stoppedadvancing after contacting an NP glial cell (6:00). Arrows, axon shaft;arrowheads, axon tip. B: A growth cone extends toward and contactsa glial cell process (8:40), branches, and remains associated with thesurface of the glial process. Arrows, axon shaft; arrowheads, tips of

growth cone branches. C: Two growth cones extending from separateaxon shafts (solid arrows; open arrows) contact the same glial cell.First growth cone (solid arrowheads) contacts the glial cell (9:20) andremains simple in morphology for the duration of the recording.Second growth cone (open arrowhead) contacts the glial cell (15:20)and branches. Scale bar � 10 �m.


nocytochemistry, dish identity was coded so that observ-ers were blind to glial cell origin during confocal micros-copy and growth cone scoring. In the experimentrepresented in Figure 8, three experimental dishes wereprepared and analyzed for each condition. Dishes from thesame experimental group were summed after confirma-tion of their statistical similarity.

In SZ and NP cocultures, growth cones not in contactwith glial cells were predominantly simple and had adistribution of growth cone morphologies that closelymatched those of growth cones grown without glial cells

(Fig. 8I). The distributions of growth cone morphologieswere not statistically different between any of the noncon-tacting conditions. Figure 8 shows examples of growthcones contacting SZ and NP glial cells falling into thesimple (Fig. 8A,B,E,F) and complex (Fig. 8C,D,G,H) cate-gories. The frequency of complex growth cones was statis-tically significantly higher among growth cones that con-tacted either SZ (Fig. 8J) or NP (Fig. 8J) glial cellscompared with those that did not contact glia (P � .004 forSZ and P � .001 for NP cocultures). Growth cone morphol-ogy was equally affected by contact with SZ and with NP

Fig. 4. ORN growth cone encounters with peripherally derivedantennal nerve glial cells do not halt axon elongation. A–F: Framesfrom a movie sequence showing typical growth cone responses tocontact with antennal nerve glial cells. The glial cell (asterisk) islocated at bottom left; axons extend from an explant located out of thefield (upper right). A,B: A glial process (double arrowheads) extends tomeet an elongating axon (arrows). C: The contacted axon branchesand two growth cones emerge. One growth cone (solid arrowhead)contacts a glial process, as the other remains isolated on the tissueculture substrate (open arrowhead). D,E: Both growth cones advance,either directly in contact with the glial process (solid arrowheads) oron the substrate adjacent to the glial cell (open arrowheads). F: Bothgrowth cones have continued to advance on their original trajectories.Inset: Digital enlargement of the boxed region in F, showing closeapposition between the axon (arrowheads) and the glial process uponwhich it elongated. Scale bar � 10 �m.

Fig. 5. Sequential frames from a time-lapse recording illustratinghow distance measurements were obtained. In the first frame (600minutes), a reference mark (black dots) was placed at the tip of theadvancing growth cone. As the growth cone advanced or retracted, anew marker was placed at its tip in every frame. The distancesbetween markers in adjacent movie frames were measured in mi-crometers and summed across all frames to yield the net distance thatthe axon grew. This growth cone branched (720 minutes), and thebranch was marked (white dots) and measured using the same strat-egy. Arrowhead denotes growth cone contact with the glial cell. Scalebar � 10 �m.


Figure 6

glial cells. Because growth cone morphology changed onlyin axons contacting glial cells, the possibility that a long-range soluble factor influenced growth cone morphologywas discounted.

In summary, the fixed-cell cytoskeletal staining ap-proach used to examine growth cone morphology recapit-ulated results obtained with live-cell imaging. Growthcones that contacted both SZ and NP glial cells showed a

greater degree of morphological complexity than growthcones that did not contact glial cells.

Glial cells deprived of ORN axon ingrowthcan influence growth cone morphology

Reciprocal communication between ORN axons andglial cells has been hypothesized to underlie several pro-cesses that are critical to the formation of the moth pri-mary olfactory system. First, early-arriving ORN axonstrigger the migration of glial cells to the SZ. SZ glial cellsthen influence the behavior of subsequently arriving re-ceptor axons (Rossler et al., 1999). Second, the arrival ofORN axons influences the morphogenesis of NP glia bytriggering the extension of glial processes into the NP andthe migration of glial cell bodies to surround the develop-ing glomeruli (Oland and Tolbert, 1987). The normal pro-liferation and migration of NP glial cells subsequentlyserve to stabilize clusters of receptor axon terminals,called protoglomeruli, and to partition the antennal lobeNP into functional glomerular units (Oland and Tolbert,1988; Oland et al., 1988; Baumann et al., 1996). Thus,communication between ORN axons and glial cells couldconceivably lead to changes in gene expression that inturn modify subsequent cellular behavior.

To test whether interactions with ORN axons were re-quired for SZ and NP glial cells to develop the ability toinfluence growth cone morphology in vitro, glial cells wereallowed to develop without olfactory input before beingharvested for coculture. Unilaterally unafferented ani-mals were produced and allowed to develop until earlystage 7 (see Materials and Methods). Each experimentalanimal contained one normally afferented and one unaf-ferented antennal lobe. Although SZ glia are not produced,NP glia continue to proliferate normally in the absence ofreceptor axon innervation (Oland and Tolbert, 1989).Therefore, a direct comparison between NP glial cells fromnormal and unafferented antennal lobes could be made.The neuropil regions of unafferented and control antennallobes were dissected, dissociated, and introduced to ex-plant cultures separately. The fixed-cell cytoskeletalstaining approach was used to permit statistical analysisof large numbers of growth cones across multiple experi-mental dishes. The observers were again blind to glial cellorigin during confocal microscopy and scoring of growthcone morphologies.

Growth cones from cocultures containing NP glial cellsisolated from afferented (NP/Aff-AL) and unafferented an-tennal lobes (NP/Unaff-AL) displayed the same range ofmorphological diversity as growth cones in explant only,normal SZ, and normal NP cocultures (Figs. 7, 8I). Thenoncontacting axons from NP/Aff-AL and NP/Unaff-ALcocultures had distributions of growth cone morphologiesnearly identical to those of axons grown without glial cells(Fig. 9I). Contacting axons in NP/Aff-AL or NP/Unaff-ALcocultures were categorized as having either simple (Fig.9A,B,E,F) or complex (Fig. 9C,D,G,H) growth cone mor-phologies. Regardless of whether their parent antennallobes had been exposed to sensory axons, growth conescontacting glial cells were less likely to be simple andmore likely to be complex. Axons that contacted NP glialcells from unafferented antennal lobes behaved identicallyto the axons that contacted NP glial cells from afferentedlobes (Fig. 9J); the distributions of growth cone morphol-ogies were statistically equivalent between contacting ax-ons in NP/Aff-AL and NP/Unaff-AL cocultures (P .999).

Fig. 6. ORN growth cones cease advancement following contactwith SZ and NP glial cells. Plots represent total distances that axonsgrew before and after contact with glial cells. Arrows mark the mo-ment of glial cell contact. Dotted lines indicate distances that axonbranches grew. Tables include rates of axon elongation before andafter contact. A: Growth cone contact with SZ glial cells halts axonelongation in five independent cases. B: Similar growth arrest is seenin cases of encounters with NP glial cells. C: Single plot of an encoun-ter with a peripherally derived antennal nerve glial cell. Branch aextended on the substrate adjacent to the glial cell, whereas branch bextended directly in contact with a glial process.

Fig. 7. Confocal micrographs of isolated ORN growth conesstained to reveal microtubules (red) and actin filaments (green). Im-ages represent the normal range of ORN growth cone morphologies invitro. A: Simple growth cones, having either unbranched or branchedmicrotubule domains tipped by actin-rich filopodia. B: Complexgrowth cones, having flattened regions with splayed microtubulessurrounded by a dense fringe of actin-based filopodia. Scale bar � 10�m.


The distributions of growth cone morphologies were sig-nificantly different between contacting and noncontactingaxons, both in NP/Aff-AL (P � .001) and in NP/Unaff-AL(P � .001) cocultures, indicating that previous interactionwith ORN axons was not necessary for glia to develop theability to affect growth cone behavior.

DISCUSSION

Previous studies have indicated that SZ and NP gliacritically influence the development of ORN axons inthe olfactory system of Manduca sexta. In the presentstudy, we have demonstrated that direct contact withindividual SZ and NP glial cells alters the behavior ofindividual ORN axons in vitro. Live-cell imagingshowed that filopodial contact with SZ and NP glial cells

typically caused ORN growth cones to increase in mor-phological complexity and cease advancement withinminutes. Analysis of growth cones in fixed cultures dem-onstrated that contact with SZ and NP glial cells had astatistically significant effect on growth cone complex-ity, regardless of whether glial cells had previously beenexposed to ingrowing ORN axons in vivo. Finally, ORNaxons that contacted peripherally derived AN glial cellsdid not stop but continued to elongate (see also Tuckerand Tolbert, 2003), suggesting that contact-mediatedgrowth cone responses to SZ and NP glial cells werespecific behaviors and not generalized reactions to con-tact with glial cells. We anticipate that the in vitrococulture system described in the current paper will beused in future experiments to identify factors that me-diate growth cone– glial cell interactions in vivo.

Fig. 8. Contact with SZ and NP glial cells leads to an increase ingrowth cone complexity, in cultures stained to reveal tubulin (red) andactin (green). A–H: Isolated growth cones (arrows) in contact withglial cells (asterisks). A–D: Growth cones contacting SZ glial cells.A,B, simple; C,D, complex. E–H: Growth cones contacting NP glialcells. E,F, simple; G,H, complex. I: Distribution of growth cone mor-phologies. Whether grown alone (black bars), with SZ glia (medium

gray bars), or with NP glia (light gray bars), growth cones not con-tacting glial cells have predominantly simple morphologies. J: Growthcones contacting SZ glial cells (yellow bars) and NP glial cells (redbars) have statistically significantly different proportions of simpleand complex growth cones (SZ: P � .004; NP: P � .001) than growthcones not contacting glial cells (gray bars) in the same cultures. Scalebar � 10 �m.


Growth cone responses to SZ and NPglial cells

In vivo, axon behavior dramatically changes in the glia-rich SZ, where axons sort into fascicles (Rossler et al.,1999), and in the antennal lobe, where axons branch andform protoglomeruli that are stabilized by NP glial cells(Baumann et al., 1996). Proper axon behavior in the SZand antennal lobe depends on the presence of adequatenumbers of glial cells (Oland and Tolbert, 1988; Oland etal., 1988; Baumann et al., 1996; Rossler et al., 1999),suggesting that interactions with glia are responsible forregion-specific changes in axon behavior. Here we reportthe first evidence that contact with SZ and NP glial cells

directly alters the behavior and morphology of ORNgrowth cones in vitro, suggesting a potential contact-dependent role for olfactory glia in the orchestration ofaxon growth and guidance in vivo. Many growth cones ofingrowing ORN axons travel in close association with glialprocesses in vivo (Oland et al., 1998), yet it remains un-known whether all ORN growth cones directly contact gliain the SZ or antennal lobe NP or whether a subset ofgrowth cones contacts olfactory glia and subsequentlyevokes region-specific changes in axon behavior throughaxon–axon interactions.

Despite their ability to promote distinctly different in-fluences on ORN axons in vivo, SZ and NP glial cells

Fig. 9. Glial cells do not require exposure to afferent axons to elicitORN growth cone elaboration in vitro. Red, tubulin; green, actin.A–H: Isolated growth cones (arrows) in contact with glial cells (aster-isks). A–D: Growth cones contacting glial cells from normally affer-ented antennal lobes. A,B, simple; C,D, complex. E–H: Growth conescontacting glial cells from unafferented antennal lobes (see Materialsand Methods). E,F, simple; G,H, complex. I: Distribution of growthcone morphologies. The proportions of simple and complex growthcones are nearly equivalent between growth cones cultured without

glia (black bars) and with glia from normally afferented (mediumgray) and unafferented (light gray) antennal lobes. J: Axons contact-ing glial cells from normally afferented lobes (red bars) and unaffer-ented lobes (blue bars) have statistically significantly different distri-butions of growth cone morphologies (Aff-AL: P � .001; Unaff-AL: P �.001) from growth cones not contacting glial cells (gray bars) in thesame cultures. The proportions of complex growth cones contactingglial cells are statistically identical between Aff-AL and UnAff-ALcocultures (P .999). Scale bar � 10 �m.


surprisingly cause nearly identical changes in the behav-ior and morphology of ORN growth cones in vitro. Al-though the behavioral consequences of contacting SZ andNP glia differ in vivo, the current in vitro findings raisethe possibility that similar mechanisms mediate ORNgrowth cone responses to both SZ and NP glial cells.Similarities in growth cone behavior may reflect the com-mon central origin of SZ and NP glia from a set of cellsoutlining the neuropil of the young antennal lobe (Rossleret al., 1999). SZ and NP glia may continue to share manymolecular properties after terminal differentiation. Theseshared properties could govern what occurs in the simpli-fied environment of tissue culture, where extrinsic cuespresent in native tissue are removed. In addition, thehistory of glial encounters in vivo might be critical, andprogression of ORN axons through the SZ might “prime”growth cones for later encounters with NP glia. Currently,we are challenging ORN axons to grow into antennal lobeslices from abnormal trajectories to determine whetheraxon behavior or growth cone morphology is altered if thesequence of glial encounters is varied such that ORNaxons encounter NP glia without first encountering glia inthe SZ.

We report in the current paper that centrally derivedglial cells did not require interactions with ORN axonsin vivo in order to induce contact-dependent changes inORN growth cone morphology in vitro. This findingsuggests that SZ and NP glia have an axon-independentability to influence ORN growth cone morphology andbehavior. An interesting question is whether the heter-ogeneity in growth cone behavior observed after contactwith SZ and NP glial cells is due to intrinsic differencesamong individual ORN axons or, alternatively, intrinsicdifferences among individual glial cells. Although exam-ples are limited, individual glial cells do not appear toimpart identical changes in growth cone behavior, asexemplified in Figure 3C, where two ORN axons re-spond differently to contact with the same glial cell.Another possibility is that the local or overall density ofglial cells could have an influence on the behavior ofORN axons. Our in vitro experiments were not designedto assess effects of glial density on ORN growth conebehavior or morphology, but several observations doaddress this issue. The results shown in Figure 3C andthe finding that growth cones may display similar com-plexities whether they are contacting individual SZ andNP glial cells or small groups of these cells (e.g., in Fig.2B) suggest that glial density does not have a simple,direct effect.

The precise functional significance of glia-induced elab-oration of growth cone morphology and cessation of axonadvancement remains to be elucidated through more dif-ficult in situ imaging of living axons, where the rates ofaxon elongation and the morphology of individual ORNgrowth cones will be monitored with respect to olfactoryglia in their normal environments. By combining tissueculture experiments, which provide a means to perturbexperimentally interactions between ORN axons and SZ,NP, and AN glial cells, with in vivo studies of axon behav-ior, we aim to identify in the future molecules and signal-ing pathways that regulate growth cone responses to ol-factory glia and relate those responses to the sorting andtargeting of ORN axons in vivo.

Changes in growth cone adhesive propertiesand the cytoskeleton may alter growth cone

behavior after glial cell contact

Our findings that ORN growth cones elaborate and stopadvancing after contacting SZ and NP glial cells in vitroare consistent with the possibility that growth cone adhe-sive properties are altered following contact with glialcells. Classic in vitro studies demonstrated that growthcones with flattened, lamellar morphologies were stronglyattached to the underlying substrate (Letourneau, 1975).Subsequent work indicated that growth cones were large,flat, more adherent, and slower growing when extendingon cell adhesion molecules and were small, filopodial, lessadherent, and faster growing when extending on extracel-lular matrix proteins (Payne et al., 1992; Lemmon et al.,1992; Drazba et al., 1997). In the present studies, nearly90% of ORN growth cones stopped advancing after en-countering SZ and NP glial cells. These findings suggestthe interesting possibility that glial contact increasedgrowth cone adhesion to the underlying concanavalinA/laminin substrate. Growth cones that flattened aftercontacting SZ or NP glial cells (over 50% of all encounters)never advanced, perhaps indicating that adhesion wasmaximal in flattened growth cones.

Although differential adhesion between growth conefilopodia and the extracellular substrate may not directlyinfluence steering decisions made by navigating axons(Isbister and O’Connor, 1999), changes in axon adhesioncould regulate the behavior of ORN axons at specificchoice points in vivo. After encountering SZ glia in vivo,ORN axons change associations by sorting into new axonfascicles (Oland et al., 1998). Adhesive properties mustchange as axons become less adherent to their neighborsand more adherent to axons of similar olfactory specificity.Support for this hypothesis comes from observations thata subset of ORN axons expressing the homophilic celladhesion molecule Manduca fasciclin II are initially dis-persed in the AN and undergo glia-dependent sorting intofasciclin II-positive bundles in the SZ before targeting asubset of glomeruli (Rossler et al., 1999; Higgins et al.,2002). Adhesive interactions among similar ORN axonsmight also regulate axon behavior in the antennal lobe.After ORN axons grow through glia surrounding the an-tennal lobe NP, they abruptly branch and spread out toform a fringe of terminal processes (Oland et al., 1998).Eventually, terminal arbors of ORN axons coalesce intodiscrete nodules called protoglomeruli (Oland et al., 1990,1998). Adhesion among axons of similar specificity couldprevent the intermingling of dissimilar axon terminalsand confine the arborization of later-arriving axons toterritories that are occupied by terminal branches of likeidentity.

Contact with glial cells may, therefore, increase thequantity, the quality, or the availability of molecules thataffect cell adhesion at growth cone surfaces. Whereaschanges in cell surface adhesion molecules in vitro gener-ally might lead to growth cone flattening and adherence tothe underlying substrate, changes in the same moleculesin vivo might lead to other region-specific changes ingrowth cone behavior. Future in vitro experiments willdirectly test the role that adhesion plays in modulatingORN axon outgrowth, growth cone morphology, and axonbehavior.


Another possibility is that contact-dependent signalingcascades act to regulate cytoskeletal dynamics in glialcell-contacting growth cones. The Rho family of small GTP-ases could mediate shape change and loss of motility; theseGTPases directly affect growth cone behavior by regulatingthe actin cytoskeleton through associated kinases (Nikolic,2002), and they are downstream targets of activated cellsurface guidance receptors (Dickson, 2001; Grunwald andKlein, 2002). Ephrins and semaphorins can alter the growthcone cytoskeleton by differentially regulating small GTPases(Shamah et al., 2001; Hu et al., 2001) and lead to the collapseof growth cones that express their cognate receptors (Luo etal., 1993; Drescher et al., 1995; Xu et al., 2000).

Microtubule reorganization also changes growth coneform and behavior. For instance, microtubule looping hasbeen described for a wide variety of neuronal growth conesand appears to be correlated with extended periods ofdecreased neuritic outgrowth (Tsui et al., 1984; Lankfordand Klein, 1990; Tanaka and Kirschner, 1991; Sabry etal., 1991; Roos et al., 2000) and growth cone branching(Dent et al., 1999). In the present study, fixed ORN growthcones from Manduca are large and preferentially containsplayed microtubules when contacting SZ and NP glialcells; this morphology is correlated with extensive periodsof growth cone stalling. Microtubule rearrangements un-derlie turning behaviors in growth cones (Sabry et al.,1991; Lin and Forscher, 1993; Tanaka et al., 1995) andmay cause Manduca ORN growth cones to turn aftercontacting SZ glia in vivo. In addition, axon contact withNP glia within the antennal lobe may induce microtubulereorganization and initiate terminal branching. Ratherthan occurring as mutually exclusive events, changes ingrowth cone adhesion and the cytoskeleton might occursimultaneously to produce region-specific changes ingrowth cone behavior.

Comparisons with neuron–glia interactionsin the mammalian olfactory system

Increasing evidence suggests that mammalian glia, likeglia in Manduca, participate with neurons to fashion thedeveloping primary olfactory center. Olfactory ensheath-ing cells migrate from the olfactory placode and enwrapbundles of ORN axons in the olfactory nerve and in thenerve layer of rodent olfactory bulbs (Marin-Padilla andAmieva, 1989; Doucette, 1989, 1991). Olfactory ensheath-ing cells display a blend of Schwann cell and astrocyteproperties (Raisman, 1985; Ramon-Cueto and Avila, 1998;Bartolomei and Greer, 2000) and promote neurite growth(Ramon-Cueto and Valverde, 1995; Kafitz and Greer,1999; Tisay and Key, 1999). Ensheathing cells also ex-press molecules that can influence the growth, sorting,and targeting of ORN axons (Puche et al., 1996; St. Johnand Key, 1999; Tisay et al., 2000; Crandall et al., 2000;Schwarting et al., 2000; Gilbert et al., 2001). Despite theirdistinctly different origins, Manduca SZ glia and mamma-lian olfactory ensheathing cells may play functionallyequivalent roles (Valverde, 1999; Key and St. John, 2002).Likewise, mammalian astrocytes delineate glomerularboundaries and perhaps play a role similar to ManducaNP glia in glomerular stabilization (Valverde et al., 1992;Gonzalez and Silver, 1994; Treloar et al., 1999; Valverde,1999).

In conclusion, glial cells isolated from the developing SZand antennal lobe NP of Manduca sexta directly influencethe motility and the morphology of ORN growth cones in a

well-defined culture system. We suggest that glia-mediated alterations in growth cone behavior in vitro re-flect changes in growth cone adhesive properties and thecytoskeleton and that those changes could enable ORNaxons to sort into fascicles, reorient during steering, slowgrowth, form branches, and segregate into protoglomeruliin vivo. Future studies will examine the molecular basesof growth cone–glial cell interactions, directly test theinfluence of adhesion on growth cone morphology andbehavior, and relate in vitro findings to the behavior ofindividually labeled ORN axons in situ.

ACKNOWLEDGMENTS

The authors thank Drs. Brian Lipscomb, Nicholas Gib-son, Richard Levine, and Paul St. John for critical reviewof the article. We also extend enormous gratitude to Ms.Patricia Jansma for technical assistance in the Division’sMicroscopy Facility, to Ms. Carole Turner for maintainingthe Division’s Tissue Culture Facility, to Dr. A.A. Osmanfor rearing Manduca sexta, to Mr. Mark Higgins for as-sistance in pupal surgery and in figure preparation, to Dr.Duane Sherrill for aid in statistical analyses, and to Mr.Rachit Kumar in the Undergraduate Biology ResearchProgram for assistance in growth cone scoring.

LITERATURE CITED

Bartolomei JC, Greer CA. 2000. Olfactory ensheathing cells: bridging thegap in spinal cord injury. Neurosurgery 47:1057–1069.

Baumann PM, Oland LA, Tolbert LP. 1996. Glial cells stabilize axonalprotoglomeruli in the developing olfactory lobe of the moth Manducasexta. J Comp Neurol 373:118–128.

Bozza T, Feinstein P, Zheng C, Mombaerts P. 2002. Odorant receptorexpression defines functional units in the mouse olfactory system.J Neurosci 22:3033–3043.

Clyne PJ, Warr CG, Freeman MR, Lessing D, Kim J, Carlson JR. 1999. Anovel family of divergent seven-transmembrane proteins: candidateodorant receptors in Drosophila. Neuron 22:327–338.

Crandall JE, Dibble C, Butler D, Pays L, Ahmad N., Kostek C, Puschel AW,Schwarting GA. 2000. Patterning of olfactory sensory connections ismediated by extracellular matrix proteins in the nerve layer of theolfactory bulb. J Neurobiol 45:195–206.

Dent EW, Callaway JL, Szebenyi G, Baas PW, Kalil K. 1999. Reorganiza-tion and movement of microtubules in axonal growth cones and devel-oping interstitial branches. J Neurosci 19:8894–8908.

Dickson BJ. 2001. Rho GTPases in growth cone guidance. Curr OpinNeurobiol 11:103–110.

Doucette R. 1989. Development of the nerve fiber layer in the olfactory bulbof mouse embryos. J Comp Neurol 285:514–527.

Doucette R. 1991. PNS-CNS transitional zone of the first cranial nerve.J Comp Neurol 312:451–466.

Drazba J, Liljelund P, Smith C, Payne R, Lemmon V. 1997. Growth coneinteractions with purified cell and substrate adhesion molecules visu-alized by interference reflection microscopy. Brain Res Dev Brain Res100:183–197.

Drescher U, Kremoser C, Handwerker C, Loschinger J, Noda M, Bonhoef-fer F. 1995. In vitro guidance of retinal ganglion cell axons by RAGS, a25 kDa tectal protein related to ligands for Eph receptor tyrosinekinases. Cell 82:359–370.

Dubuque SH, Schachtner J, Nighorn AJ, Menon KP, Zinn K, Tolbert LP.2001. Immunolocalization of synaptotagmin for the study of synapsesin the developing antennal lobe of Manduca sexta. J Comp Neurol441:277–287.

Fischer RA. 1925. Statistical methods for research workers. Edinburgh:Oliver and Boyd.

Gao Q, Yuan B, Chess A. 2000. Convergent projections of Drosophilaolfactory neurons to specific glomeruli in the antennal lobe. Nat Neu-rosci 3:780–785.

Gilbert M, Smith J, Roskams AJ, Auld VJ. 2001. Neuroligin 3 is a verte-


brate gliotactin expressed in the olfactory ensheathing glia, a growth-promoting class of macroglia. Glia 34:151–164.

Gong Q, Shipley MT. 1996. Expression of extracellular matrix moleculesand cell surface molecules in the olfactory nerve pathway during earlydevelopment. J Comp Neurol 366:1–14.

Gonzalez ML, Silver J. 1994. Axon–glia interactions regulate ECM pat-terning in the postnatal rat olfactory bulb. J Neurosci 14:6121–6131.

Grunwald IC, Klein R. 2002. Axon guidance: receptor complexes and sig-naling mechanisms. Curr Opin Neurobiol 12:250–259.

Higgins MR, Gibson NJ, Eckholdt PA, Nighorn A, Copenhaver PF, Nardi J,Tolbert LP. 2002. Different isoforms of fasciclin II are expressed by asubset of developing olfactory receptor neurons and by olfactory-nerveglial cells during formation of glomeruli in the moth Manduca sexta.Dev Biol 244:134–154.

Hildebrand JG, Shepherd GM. 1997. Mechanisms of olfactory discrimina-tion: converging evidence for common principles across phyla. AnnuRev Neurosci 20:595–631.

Hildebrand JG, Rossler WR, Tolbert LP. 1997. Postembryonic developmentof the olfactory system in the moth Manduca sexta: primary-afferentcontrol of glomerular development. Semin Cell Dev Biol 8:163–170.

Hu H, Marton TF, Goodman CS. 2001. Plexin B mediates axon guidance inDrosophila by simultaneously inhibiting active Rac and enhancingRhoA signaling. Neuron 32:39–51.

Isbister CM, O’Connor TP. 1999. Filopodial adhesion does not predictgrowth cone steering events in vivo. J Neurosci 19:2589–2600.

Kafitz KW, Greer CA. 1999. Olfactory ensheathing cells promote neuriteextension from embryonic olfactory receptor cells in vitro. Glia 25:99–110.

Kent K. 1985. Metamorphosis of the antennal center and the influence ofsensory innervation on the formation of glomeruli in the hawkmothManduca sexta. PhD dissertation, Harvard University.

Kent KS, Harrow ID, Quartararo P, Hildebrand JG. 1986. An accessoryolfactory pathway in Lepidoptera: the labial pit organ and its centralprojections in Manduca sexta and certain other sphinx moths and silkmoths. Cell Tissue Res 245:237–245.

Kent KS, Oland LA, Hildebrand JG. 1999. Development of the labial pitorgan glomerulus in the antennal lobe of the moth Manduca sexta: therole of afferent projections in the formation of identifiable olfactoryglomeruli. J Neurobiol 40:28–44.

Key B, St. John J. 2002. Axon navigation in the mammalian primaryolfactory pathway: where to next? Chem Senses 27:245–260.

Lankford KL, Klein WL. 1990. Ultrastructure of individual neurons iso-lated from avian retina: occurrence of microtubule loops in dendrites.Brain Res Dev Brain Res 51:217–224.

Lemmon V, Burden SM, Payne HR, Elmslie GJ, Hlavin ML. 1992. Neuritegrowth on different substrates: permissive vs. instructive influencesand the role of adhesive strength. J Neurosci 12:818–826.

Letourneau PC. 1975. Possible roles for cell-to-substratum adhesion inneuronal morphogenesis. Dev Biol 44:77–91.

Lin CH, Forscher P. 1993. Cytoskeletal remodeling during growth cone-target interactions. J Cell Biol 121:1369–1383.

Lohr C, Tucker E, Oland LA, Tolbert LP. 2002. Development ofdepolarization-induced calcium transients in insect glial cells is depen-dent on the presence of afferent axons. J Neurobiol 52:85–98.

Luo Y, Raible D, Raper JA. 1993. Collapsin: a protein in brain that inducesthe collapse and paralysis of neuronal growth cones. Cell 75:217–227.

Marin-Padilla M, Amieva MR. 1989. Early neurogenesis of the mouseolfactory nerve: Golgi and electron microscopic studies. J Comp Neurol288:339–352.

Mason C, Erskine L. 2000. Growth cone form, behavior, and interactions invivo: retinal axon pathfinding as a model. J Neurobiol 44:260–270.

Matheson SF, Levine RB. 1999. Steroid hormone enhancement of neuriteoutgrowth in identified insect motor neurons involves specific effects ongrowth cone form and function. J Neurobiol 38:27–45.

Mombaerts P, Wang F, Dulac C, Chao SK, Nemes A, Mendelsohn M,Edmondson J, Axel R. 1996. Visualizing an olfactory sensory map. Cell87:675–686.

Mori K, Nagao H, Yoshihara Y. 1999. The olfactory bulb: coding andprocessing of odor molecule information. Science 286:711–715.

Nikolic M. 2002. The role of Rho GTPases and associated kinases inregulating neurite outgrowth. Int J Biochem Cell Biol 34:731–745.

Oland LA, Tolbert LP. 1987. Glial patterns during early development ofantennal lobes of Manduca sexta: a comparison between normal lobesand lobes deprived of antennal axons. J Comp Neurol 255:196–207.

Oland LA, Tolbert LP. 1988. Effects of hydroxyurea parallel the effects ofradiation in developing olfactory glomeruli in insects. J Comp Neurol278:377–387.

Oland LA, Tolbert LP. 1989. Patterns of glial proliferation during forma-tion of olfactory glomeruli in an insect. Glia 2:10–24.

Oland LA, Tolbert LP. 1996. Multiple factors shape development of olfac-tory glomeruli: insights from an insect model system. J Neurobiol30:92–109.

Oland LA, Tolbert LP, Mossman KL. 1988. Radiation-induced reduction ofthe glial population during development disrupts the formation ofolfactory glomeruli in an insect. J Neurosci 8:353–367.

Oland LA, Orr G, Tolbert LP. 1990. Construction of a protoglomerulartemplate by olfactory axons initiates the formation of olfactory glomer-uli in the insect brain. J Neurosci 10:2096–2112.

Oland LA, Muller T, Kettenmann H, Hayashi J. 1996. Preparation ofprimary cultures and acute slices of the nervous system of the mothManduca sexta. J Neurosci Methods 69:103–112.

Oland LA, Pott WM, Higgins MR, Tolbert LP. 1998. Targeted ingrowth andglial relationships of olfactory receptor axons in the primary olfactorypathway of an insect. J Comp Neurol 398:119–138.

Payne HR, Burden SM, Lemmon V. 1992. Modulation of growth conemorphology by substrate-bound adhesion molecules. Cell Motil Cy-toskeleton 21:65–73.

Puche AC, Shipley MT. 2001. Radial glia development in the mouse olfac-tory bulb. J Comp Neurol 434:1–12.

Puche AC, Poirier F, Hair M, Bartlett PF, Key B. 1996. Role of galectin-1in the developing mouse olfactory system. Dev Biol 179:274–287.

Raisman G. 1985. Specialized neuroglial arrangement may explain thecapacity of vomeronasal axons to reinnervate central neurons. Neuro-science 14:237–254.

Ramon-Cueto A, Avila J. 1998. Olfactory ensheathing glia: properties andfunction. Brain Res Bull 46:175–187.

Ramon-Cueto A, Valverde F. 1995. Olfactory bulb ensheathing glia: aunique cell type with axonal growth-promoting properties. Glia 14:163–173.

Ressler KJ, Sullivan SL, Buck LB. 1993. A zonal organization of odorantreceptor gene expression in the olfactory epithelium. Cell 73:597–609.

Ressler KJ, Sullivan SL, Buck LB. 1994. Information coding in the olfac-tory system: evidence for a stereotyped and highly organized epitopemap in the olfactory bulb. Cell 79:1245–1255.

Roos J, Hummel T, Ng N, Klambt C, Davis GW. 2000. Drosophila Futschregulates synaptic microtubule organization and is necessary for syn-aptic growth. Neuron 26:371–382.

Rossler WR, Oland LA, Higgins MR, Hildebrand JG, Tolbert LP. 1999.Development of a glia-rich axon-sorting zone in the olfactory pathwayof the moth Manduca sexta. J Neurosci 19:9865–9877.

Sabry JH, O’Connor TP, Evans L, Toroian-Raymond A, Kirschner M,Bentley D. 1991. Microtubule behavior during guidance of pioneerneuron growth cones in situ. J Cell Biol 115:381–395.

Sanes JR, Hildebrand JG. 1976. Origin and morphogenesis of sensoryneurons in an insect antenna. Dev Biol 51:300–319.

Sanes JR, Prescott DJ, Hildebrand JG. 1977. Cholinergic neurochemicaldevelopment of normal and deafferented antennal lobes during meta-morphosis of the moth, Manduca sexta. Brain Res 119:389–402.

Schwarting GA, Kostek C, Ahmad N, Dibble C, Pays L, Puschel AW. 2000.Semaphorin 3A is required for guidance of olfactory axons in mice.J Neurosci 20:7691–7697.

Shamah SM, Lin MZ, Goldberg JL, Estrach S, Sahin M, Hu L, BazalakovaM, Neve RL, Corfas G, Debant A, Greenberg ME. 2001. EphA receptorsregulate growth cone dynamics through the novel guanine nucleotideexchange factor ephexin. Cell 105:233–244.

St. John JA, Key B. 1999. Expression of galectin-1 in the olfactory nervepathway of rat. Brain Res Dev Brain Res 117:171–178.

St. John JA, Key B. 2001. EphB2 and two of its ligands have dynamicprotein expression patterns in the developing olfactory system. BrainRes Dev Brain Res 126:43–56.

Tanaka EM, Kirschner MW. 1991. Microtubule behavior in the growthcones of living neurons during axon elongation. J Cell Biol 115:345–363.

Tanaka EM, Kirschner MW. 1995. The role of microtubules in growth coneturning at substrate boundaries. J Cell Biol 128:127–137.

Tisay KT, Key B. 1999. The extracellular matrix modulates olfactoryneurite outgrowth on ensheathing cells. J Neurosci 19:9890–9899.


Tisay KT, Bartlett PF, Key B. 2000. Primary olfactory axons form ectopicglomeruli in mice lacking p75NTR. J Comp Neurol 428:656–670.

Tolbert LP. 1998. Olfactory development in invertebrates: on the scent ofcentral development issues. Ann N Y Acad Sci 855:95–103.

Tolbert LP, Matsumoto SG, Hildebrand JG. 1983. Development of syn-apses in the antennal lobes of the moth Manduca sexta during meta-morphosis. J Neurosci 3:1158–1175.

Treloar HB, Nurcombe V, Key B. 1996. Expression of extracellular matrixmolecules in the embryonic rat olfactory pathway. J Neurobiol 31:41–55.

Treloar H, Tomasiewicz H, Magnuson T, Key B. 1997. The central pathwayof primary olfactory axons is abnormal in mice lacking the N-CAM-180isoform. J Neurobiol 32:643–658.

Treloar HB, Purcell AL, Greer CA. 1999. Glomerular formation in thedeveloping rat olfactory bulb. J Comp Neurol 413:289–304.

Tsui HT, Lankford KL, Ris H, Klein WL. 1984. Novel organization ofmicrotubules in cultured central nervous system neurons: formation ofhairpin loops at ends of maturing neurites. J Neurosci 4:3002–3013.

Tucker ES, Tolbert LP. 2003. Reciprocal interactions between olfactoryreceptor axons and olfactory nerve glia cultured from the developingmoth Manduca sexta. Dev Biol 260:9–30.

Tucker ES, Oland LA, Tolbert LP. 2000. In vitro study of interactionsbetween olfactory receptor growth cones and glial cells of the axonalsorting zone. Soc Neurosci Abstr 26:1611.

Tucker ES, Oland LA, Tolbert LP. 2001. In vitro analysis of interactionsbetween olfactory receptor growth cones and centrally derived glia.Assoc Chemorecept Sci Abstr 23:72.

Valverde F. 1999. Building an olfactory glomerulus. J Comp Neurol 415:419–422.

Valverde F, Santacana M, Heredia M. 1992. Formation of an olfactoryglomerulus: morphological aspects of development and organization.Neuroscience 49:255–275.

Vassar R, Ngai J, Axel R. 1993. Spatial segregation of odorant receptorexpression in the mammalian olfactory epithelium. Cell 74:309–318.

Vassar R, Chao SK, Sitcheran R, Nunez JM, Vosshall JB, Axel R. 1994.Topographic organization of sensory projections to the olfactory bulb.Cell 79:981–991.

Vosshall LB, Amrein H, Morozov PS, Rzhetsky A, Axel R. 1999. A spatialmap of olfactory receptor expression in the Drosophila antenna. Cell96:725–736.

Vosshall LB, Wong AM, Axel R. 2000. An olfactory sensory map in the flybrain. Cell 102:147–159.

Walz A, Rodriguez I, Mombaerts P. 2002. Aberrant sensory innervation ofthe olfactory bulb in neuropilin-2 mutant mice. J Neurosci 22:4025–4035.

Wang F, Nemes A, Mendelsohn M, Axel R. 1998. Odorant receptors governthe formation of a precise topographic map. Cell 93:47–60.

Xu XM, Fisher DA, Zhou L, White FA, Ng S, Snider WD, Luo Y. 2000. Thetransmembrane protein semaphorin 6A repels embryonic sympatheticaxons. J Neurosci 20:2638–2648.

Yoshihara Y, Kawasaki M, Tamada A, Fujita H, Hayashi H, KagamiyamaH, Mori K. 1997. OCAM: a new member of the neural cell adhesionmolecule family related to zone-to-zone projection of olfactory andvomeronasal axons. J Neurosci 17:5830–5842.


in vitro analyses of interactions between olfactory receptor growth cones and glial cells that...

Documents