Wnt/Calcium Signaling Mediates Axon Growth and

Guidance in the Developing Corpus Callosum

B. Ian Hutchins,1* Li Li,1 Katherine Kalil1,2

1
Neuroscience Training Program, University of Wisconsin-Madison, Madison, Wisconsin
2
Department of Anatomy, University of Wisconsin-Madison, Madison, Wisconsin

Received 8 July 2010; revised 26 August 2010; accepted 30 September 2010

ABSTRACT: It has been shown in vivo that Wnt5a regulate post-crossing axon outgrowth and guidance.
gradients surround the corpus callosum and guide Co-electroporation of Ryk siRNA and DsRed revealed
callosal axons after the midline (postcrossing) by Wnt5a- that knock down of the Ryk receptor reduced outgrowth
induced repulsion via Ryk receptors. In dissociated rates of postcrossing but not precrossing axons by 50%
cortical cultures we showed that Wnt5a simultaneously and caused axon misrouting. Guidance errors in axons
promotes axon outgrowth and repulsion by calcium sig- with Ryk knockdown resulted from reduced calcium ac-
naling. Here to test the role of Wnt5a/calcium signaling tivity. In the corpus callosum CaMKII inhibition reduced
in a complex in vivo environment we used sensorimotor the outgrowth rate of postcrossing (but not precrossing)
cortical slices containing the developing corpus callosum. axons and caused severe guidance errors which resulted
Plasmids encoding the cytoplasmic marker DsRed and from reduced CaMKII-dependent repulsion downstream
the genetically encoded calcium indicator GCaMP2 were of Wnt/calcium. We show for the first time that Wnt/Ryk
electroporated into one cortical hemisphere. Postcrossing calcium signaling mechanisms regulating axon outgrowth
callosal axons grew 50% faster than pre-crossing axons and repulsion in cortical cultures are also essential for the
and higher frequencies of calcium transients in axons and proper growth and guidance of postcrossing callosal axons
growth cones correlated well with outgrowth. Application which involve axon repulsion through CaMKII. ' 2010
of pharmacological inhibitors to the slices showed that Wiley Periodicals, Inc. Develop Neurobiol 71: 269–283, 2011
signaling pathways involving calcium release through IP3 Keywords: corpus callosum; calcium signaling; Wnt;
receptors and calcium entry through TRP channels CaMKII; growth cone

INTRODUCTION
Additional Supporting Information may be found in the online
version of this article. Wnt proteins are a family of morphogens that have
B.I.H. designed and performed the cortical slice experiments. recently been shown to function as axon guidance
L.L. and B.I.H. designed and performed Dunn chamber turning
assays. L.L. performed the dissociated outgrowth assays and West- cues (Salinas and Zou, 2008). Wnt5a, through the
ern blots. K.K. helped with experimental design and wrote the Ryk receptor, mediates the guidance of efferent corti-
manuscript with assistance from B.I.H. and L.L. cospinal and callosal axons (Liu et al., 2005; Keeble
*Present address: National Institute of Neurological Disorders
and Stroke, National Institutes of Health, 35 Convent Dr., et al., 2006; Zou and Lyuksyutova, 2007). Knockout
Bethesda, MD 20892. of the Ryk receptor causes misrouting of corpus cal-
Correspondence to: K. Kalil ([email protected]). losal axons in vivo after axons have crossed the mid-
Contract grant sponsor: National Institutes of Health; contract
grant numbers: NS14428, GM801642. line (Keeble et al., 2006). Gradients of Wnt5a sur-
Contract grant sponsors: Herman I. Shapiro Fellowship, Univer- round the callosum and corticospinal tract and Wnt5a
sity of Wisconsin Graduate School. repels cortical axons in explant cultures. Thus in the
' 2010 Wiley Periodicals, Inc. callosum of knockout mice lacking Ryk receptors
Published online 8 October 2010 in Wiley Online Library
(wileyonlinelibrary.com) guidance errors were attributed to disruption of
DOI 10.1002/dneu.20846 Wnt5a/Ryk-mediated axon repulsion. However, the
269
270 Hutchins et al.

signaling mechanisms downstream of Ryk in the con- inserts (Millipore) in plating medium containing 5% fetal
text of axon growth and guidance were completely bovine serum (Invitrogen), 2% B27 supplement (Invi-
unknown (Liu et al., 2005; Keeble et al., 2006). trogen), and 1% liquid glutamine-penicillin-streptomycin
Recently we found that Wnt5a gradients not only (Invitrogen) in Neurobasal medium (Invitrogen) and were
repel cortical axons in an in vitro turning assay but at maintained at 378C at 5% CO2. After recovering for up to 1
day in vitro, slices containing the corpus callosum were
the same time increase their rates of outgrowth (Li et
placed into the well of an open chamber fitted with a plati-
al., 2009), consistent with the propulsive model of num electrode bottom (CUY700P10E, Nepagene). Plasmids
Wnt5a signaling (Zou and Lyuksyutova, 2007). Fur- (1 lg lL
1) encoding DsRed2, a cytoplasmic fluorescent
ther, we found that Ryk receptors are essential for the protein, were pressure injected (from a glass pipette with a
growth promoting and repulsive guidance effects of 25 lm tip for 20 ms at 12 PSI) alone into several sites
Wnt5a gradients and that these effects are mediated within a single cortical hemisphere or were coinjected with
by calcium signaling pathways. Ryk siRNA (diluted to 5 lg lL
1) to knock down Ryk
We considered it important to test the in vivo rele- receptors. Alternatively, plasmids encoding GCaMP2
vance of the Wnt/calcium signaling mechanisms that (Addgene plasmid 18927) or EGFP-CaMKIIN were used to
we previously identified in dissociated cortical cultures visualize calcium activity or inhibit CaMKII, respectively.
(Li et al., 2009). In dissociated cultures neurons are For ratiometric imaging experiments, DsRed2 and
GCaMP2 were coinjected into slices with or without Ryk
maintained in a simplified environment and effects of
siRNA. About 88% of axons expressing GCaMP2 also
molecular cues on axons are tested one at a time. expressed DsRed2, indicating a high cotransfection effi-
In vivo, axons encountering a complex environment ciency. Electroporation was carried out with a square wave
must respond to a multitude of signals. Thus responses pulse generator (CUY-21, Nepagene) which delivered 20
of axons in culture may not reflect how they behave in pulses of 10-ms duration at 4 Hz and 50 V. Slices were then
a complex neural pathway in vivo (Gomez and Zheng, allowed to recover for 48 h before imaging. At P2 efferent
2006). For example, knocking down calcium/calmodu- cortical axons are extending toward and into the corpus cal-
lin-dependent protein kinase I (CaMKI) in dissociated losum but have not projected across the midline. Thus ex-
cultures decreases axon elongation (Ageta-Ishihara et amination of axons 48 h after electroporation allowed us to
al., 2009; Davare et al., 2009; Neal et al., 2010). In con- follow callosal axons across the midline and contralaterally.
trast, knocking down CaMKI in vivo decreases callosal
axon branching into cortex without affecting rates of Experimental Reagents
axon elongation (Ageta-Ishihara et al., 2009). We there-
fore used developing cortical slices that contained the Stock solutions were prepared by dissolving drugs in water
entire callosal pathway through the sensorimotor cortex, or dimethyl sulfoxide (DMSO) according to the recommen-
which permitted imaging of intact callosal axons dations of the manufacturer. Stock solutions were then
diluted into ACSF (described below) and perfused over
extending along their entire trajectory (Halloran and
slice cultures. The following reagents were used: 2-aminoe-
Kalil, 1994). Another important advantage of the slice
thoxydiphenyl borate (2-APB, Calbiochem), SKF96365
preparation is that experimental manipulations of mo- (Alexis Biochemicals), bovine serum albumin (BSA,
lecular signaling pathways can be carried out at specific Sigma), recombinant protein Wnt5a (R&D systems), ON-
locations and at specific times in development. In TARGETplus SMARTpool mouse Ryk siRNA (Dhar-
the present study we identified Wnt/calcium signaling macon), and a second, independent Ryk siRNA pool (Santa
mechanisms that mediate growth and guidance of Cruz Biotechnology).
callosal axons.
Imaging of Callosal Axons
MATERIALS AND METHODS Slices were placed in an open perfusible chamber (Warner
Instruments) and viewed either with an Olympus (Center
Slice Preparation and Electroporation Valley) Fluoview 500 laser-confocal system mounted on an
AX-70 upright microscope with a 403 plan fluor water
Cortical slice injection and electroporation methods were immersion objective (outgrowth and calcium imaging
adapted from (Uesaka et al., 2005). Briefly, slices were experiments) or a Nikon TE300 inverted microscope with a
obtained from P0 hamster brains. Pups were anesthetized 203 objective (outgrowth experiments only). Temperature
on ice and the brains are rapidly removed into ice-cold was maintained at 378C with a temperature controller
Hank’s Balanced Salt Solution (HBSS, Invitrogen). The (Warner Instruments). A perfusion system was used for
brains were encased in 4% agar and solidified on ice. Coro- continuous oxygenation of the heated artificial cerebrospi-
nal slices (400 lm) through the forebrain are cut on a vibra- nal fluid (ACSF, containing 124 mM NaCl, 24 mM
tome and collected in cold HBSS (Halloran and Kalil, NaHCO3, 3 mM KCl, 1.25 mM NaH2PO4, 2 mM CaCl2, 1.5
1994). Slices were then cultured on 0.4 lM membrane mM MgCl2, 10 mM glucose, and 20 mM HEPES) to which
Developmental Neurobiology
 Wnt/Calcium in Callosal Axons 271

pharmacological reagents (2-APB, 50 lM; SKF96365, 3 Quantification of Axon Outgrowth

lM) were added. Perfusion of the slices with medium was and Trajectories
carried out at a flow rate of 2 mL min
1. Time lapse images
were obtained every 5–15 s for measurements of axon out- Outgrowth was measured as the displacement in lm of the
growth for up to 90 min. For calcium imaging, images were distal tip of the growth cone between the first and last
obtained twice a second on the Fluoview 500 system during frames of an imaging session divided by the duration of that
free-scan mode for up to 30 min. Long term (3 h) treat- session. Overexpression of multiple constructs (DsRed and
ments with 2-APB or SKF96365 were returned to the incu- GCaMP2) had no deleterious effect on rates of postcrossing
bator and imaged at the beginning and end of this treatment axon outgrowth, which grew at 114% of the rate of controls
to assess effects on axon trajectories. expressing only one construct (a nonsignificant increase).
Trajectories were measured as the angle between the hori-
zontal axis of the slice and the distal 20 lm of callosal
axons, plotted versus the horizontal distance from the mid-
Measurements of Calcium Activity line. These data were best fit by a quadratic regression curve
Calcium activity was measured as the average fluorescence which we used to describe the standard trajectory taken by
pixel intensity (F) in an axon region divided by the baseline control axons in our control experiments. Deviation away
fluorescence in that region (F0). Background fluorescence from the standard trajectory of control axons was measured
was measured frame-by-frame and was subtracted from as the difference in degrees between the measured angle of
measurements of fluorescence intensity. To minimize the an axon and the angle predicted by the regression curve for
effects of any morphological changes that could affect fluo- an axon at that distance from the midline. Plots of the trajec-
rescence measurements through changes in volume, the tories of axons from this study are shown in Figures 3–5
baseline (F0) was calculated as a shifting average of the flu- alongside the best-fit regression curve and 90% prediction
orescence intensity over a 30-frame window. intervals describing the trajectories of control axons. Indi-
To choose a threshold that defined a calcium transient, vidual axons in our experimental manipulation groups were
we first simulated the number of false positive readings we considered to be significantly deviating from the standard
would measure in a signal that was derived from Gaussian trajectory if they fell outside the 90% prediction intervals
noise with a similar mean and standard deviation as our [Fig. 3(A)]. These axons are shown as deviating from the
measured calcium signals. The number of false positive corpus callosum in our tracings (Figs. 3–5) and are marked
readings measured from our simulation of 50 calcium imag- with arrowheads. Unless otherwise noted, n is the number
ing experiments was acceptably low at a threshold of 3.5 of axons from at least three independent experiments.
standard deviations above baseline (corresponding to 1.8
false positive transients h
1). Thus, calcium transients were
defined as fluorescence signals (F/F0) that exceed 3.5 stand- Dissociated Cortical Neuron Cultures
ard deviations above baseline, which were confirmed by and Wnt5a Experiments
frame-by-frame analysis of the time-lapse images. For Culture of dissociated cortical neurons and bath application
ratiometric experiments, slices were co-electroporated with experiments with Wnt5a were performed as previously
DsRed2 and GCaMP2. Fluorescence images of DsRed2 described (Li et al., 2009). Briefly, cortical neurons were
acquired simultaneously with each frame of GCaMP2 fluo- dissociated from P0 hamster sensorimotor cortex and elec-
rescence. Ratiometric measurements (R) were obtained by troporated with EGFP-CaMKIIN plasmids with an Amaxa
dividing the GCaMP2 fluorescence value by the fluores- Nucleofector. These neurons were plated onto coverslips
cence value of DsRed2. Frame-by-frame background sub- coated with 0.5 mg mL
1 poly-D-lysine (Sigma) and 20 lg
traction was performed for each indicator as described mL
1 laminin (Sigma/Invitrogen) at a density of 2000–
above. Calcium signals (R/R0) were then measured as the 7000 per cm2 and were incubated in 5% CO2 and 9% O2 at
percent change from a shifting average baseline (R0) of the 378C for 2 days. For long term axon outgrowth assays, 400
ratiometric measurements as described above for nonratio- ng mL
1 Wnt5a in 0.5% BSA is PBS, or BSA alone, was
metric measurements. Although expression levels of then added to the cultures. Cultures were then incubated for
GCaMP2 varied from cell to cell, this did not affect the 72 h before fixation. Axon lengths were measured in neu-
frequency of calcium transients reported. Raw baseline rons expressing EGFP-CaMKIIN or in untransfected neu-
fluorescence did not correlate with frequency (Spearman rons from the same dish as a control.
correlation coefficient r ¼ 0.09, p ¼ 0.69). We validated
our calcium transient measurements with additional power
spectral density analysis (Uhlen, 2004; Bortone and Dunn Chamber Axon Guidance Assay
Polleux, 2009), which measures periodicity in a time series and Analysis
signal without an arbitrary definition of a transient. This
analysis (our unpublished observations) confirmed higher For Dunn chamber axon guidance assays, P0 hamster corti-
periodicity as measured by average relative power in cal neurons were grown on appropriately coated (see above)
calcium signals in contralateral vs. ipsilateral axons at a fre- 22-mm2 No. 1.5 coverslips (Corning) at a low density (10 k
quency of 15 per hour, the frequency of calcium transients cells/well in a six well plate (Falcon). Assembly of the
evoked by Wnt5a in vitro (Li et al., 2009). Dunn chamber (Hawksley, UK) was modified from previ-
Developmental Neurobiology
272 Hutchins et al.

ous studies (Yam et al., 2009). Dunn chambers were rinsed noted, comparisons between two groups were made with
by serum-free medium once and then both inner and outer Student’s t test and comparisons between multiple groups
wells were filled by serum-free medium. To secure cover- were made with a one-way ANOVA with Dunnett’s postt-
slips with neurons on the chamber, silicon sealant (Dow est. Measurements are given in mean 6 SEM unless other-
Corning) was applied at *0.5 cm from the border of outer wise noted. Images were modified with a low-pass filter in
well but omitted at one side to form a slit later for draining MetaMorph to reduce single-pixel noise. The images pre-
and refilling the outer well. A coverslip with neurons was sented in figures were enhanced with brightness-contrast
inverted over the Dunn chamber leaving a narrow slit at the adjustments in Adobe (Mountain View, CA) Photoshop,
edge without the sealant. Media at the outer well was aspi- and with Flatten Background and Sharpen adjustments in
rated and then medium with 400 ng mL
1 Wnt5a was MetaMorph for slice images taken from the Nikon epifluo-
added to the outer well. The narrow slit was sealed by fix- rescence system [Fig. 3(C)].
ing a small piece of parafilm (American National Can) to
the chamber with sealant. Images were acquired immedi-
ately after Dunn chamber assembly and 2 h later with a 20
3 0.5 numerical aperture (NA) Plan Fluor objective RESULTS
mounted on a Nikon TE300 Quantum inverted microscope
equipped with a Photometrics Cascade II: 512. All dissoci- Calcium Signaling Regulates Growth and
ated neurons in the bridge region of the Dunn chamber Guidance of Callosal Axons
were imaged in each experiment. Dunn chambers were kept
at 378C during imaging. As a first step we labeled small numbers of cortical
We followed similar criteria as (Yam et al., 2009) for neurons by electroporation with plasmids encoding
data analysis. Briefly, we only included individual axons the cytoplasmic marker DsRed2 into one hemisphere
with >10 lm straight distal end and with net outgrowth >5 of cortical slices from P0 hamster brains and used
lm h
1. The outgrowth of axons was measured by tracing live imaging to follow the extension of individual
axons from the original position of growth cones to their axonal growth cones in the corpus callosum. As
final position. The original direction of axon outgrowth was shown in Figure 1(A), electroporation labeled a small
defined by the direction of the 10-lm distal axon segment. region of the sensorimotor cortex including Layers
The initial angle (08–1808) was defined by the angle
2–3 and 5 that give rise to subcortical, horizontal, and
between the original direction of axon outgrowth and the
callosal axons. Time lapse imaging showed that
direction of the gradient. Final angles were calculated as
the angles between the original direction of axon outgrowth callosal growth cones (n ¼ 46) exhibit continual
and a line connecting the original position of the growth motility, forward advance, and turning behaviors
cone and its final position. When axons turned towards the (Supporting Information, Movie 1). In sequences
gradient, the angle turned was assigned as positive and [Fig. 1(B)] lasting up to 1.5 h we measured rates of
when axons turned away from the gradient, the angle turned growth cone advance in different regions of the cal-
was negative. Images presented in figures were rotated to so losum. Growth cones that have not yet crossed the
that the gradient increases along the y-axis to more easily midline (precrossing) consistently advance more
compare responses of axons. slowly (36.9 6 4.3 lm h
1, n ¼ 19) than those that
have crossed the midline (post-crossing) (54.6 6 2.9
Western Blotting lm h
1, n ¼ 27) Labeled axons followed the trajec-
tory of the callosal pathway as a whole [Fig. 3 (A)].
ON-TARGETplus nontargeting pool siRNA was purchased We next asked whether growth cones extending in
from Dharmacon. Neurons obtained from P0 hamsters were the corpus callosum express spontaneous calcium ac-
electroporated with 100 pmol control siRNAs and after tivity as observed in dissociated cortical cultures (Tang
4DIV, proteins were extracted by RIPA buffer (Thermo
et al., 2003; Tang and Kalil, 2005; Hutchins and Kalil,
Fisher Scientific) supplied with protease inhibitor cocktail
2008). Calcium activity was measured in post-crossing
tablets (Roche). Blots were blocked with 3% milk (Lab Sci-
entific) and 3% BSA (Sigma) for 2 h and then incubated with axons and growth cones by electroporating GCaMP2, a
mouse anti-human bIII tubulin (1:500, Millipor Bioscience genetically encoded calcium indicator (Tallini et al.,
Research Reagents) at 48C overnight and goat anti-mouse- 2006; Mao et al., 2008), into one hemisphere of the
HRP (1:10,000, Jackson ImmunoResearch) for 1 h. ECL plus sensorimotor cortex [Fig. 2(A)]. Growth cones showed
(GE health) was used to stain tubulin and Ryk receptors. continuous motility and forward advance while
expressing repetitive calcium transients [Fig. 2(D,E)]
averaging 8.6 6 1.7 per hour, a frequency somewhat
Statistical Analysis and Image lower than spontaneous activity of cortical growth
Processing cones in dissociated culture (11.1 transients h
1),
Graphs and statistical analysis were performed with Prism (Hutchins and Kalil, 2008). In some cases transients
(GraphPad) statistical analysis software. Unless otherwise were detected in the axon as well as the growth cone
Developmental Neurobiology
 Wnt/Calcium in Callosal Axons 273

Figure 1 Visualization of individual callosal axons and their growth cones as they extend through
the callosum. (A) A low power confocal image of a cortical slice at 3DIV, after electroporation of
cortical neurons with DsRed2 performed on the slice from a P0 hamster. Note that individual efferent
axons can be clearly visualized. Arrow indicates location of the cortical growth cone imaged at higher
power in the time lapse sequence in (B). (B) Turning behaviors in images at bottom are clearly visible
as are filopodia and lammellipodia. Scale bar, 10 lm. n, +, X, reference points.

[Fig. 2(D), Supporting Information, Movie 2] but in response to environmental cues (Li et al., 2005, 2009;
other cases changes in calcium activity were confined Shim et al., 2005). Previously in dissociated cortical
to a localized region of the growth cone [Fig. 2(F)] sug- cultures we found that calcium influx through TRP
gesting the expression of both global and localized cal- channels mediates axon outgrowth and repulsive
cium activity such as we had previously observed growth cone turning evoked by Wnt5a while calcium
(Hutchins and Kalil, 2008; Hutchins, 2010). release from stores through IP3 receptors mediates
We then asked whether the frequencies of calcium axon outgrowth but not turning. To determine
transients in callosal growth cones were related to whether these calcium signaling mechanisms regulate
axon growth rates. Since we found that the callosal axon outgrowth and guidance in the developing cor-
axons extended significantly more slowly before vs. pus callosum, we bath-applied 2-APB which is
after the midline, we measured the frequencies of cal- known to block calcium release from stores through
cium transients in callosal growth cones in these two IP3 receptors (Li et al., 2005, 2009) and SKF96365
locations. Since GCaMP2 has a lower signal-to-noise which is known to block TRP channels (Li et al.,
ratio than small molecule calcium indicators such as 2005, 2009; Shim et al., 2005). In vivo suppression of
Fluo-4, we included in our counts of calcium transi- spontaneous electrical activity in callosal axons was
ents only those events that exceeded 3.5 standard shown to decrease rates of axon outgrowth on the
deviations above baseline (see Methods). We found postcrossing but not the precrossing side of the cal-
that precrossing axons growing at an average rate of losum (Wang et al., 2007). Therefore in manipulating
36.9 6 4.3 lm h
1 had an average frequency of 2.99 calcium activity, we focused on axon growth and
6 1.36 transient h
1 whereas postcrossing axons with guidance of postcrossing axons. In slices electropo-
an average growth rate of 54.6 6 2.9 lm h
1 had an rated with plasmids encoding DsRed2, individual
average frequency of 12.6 6 2.12 transients h
1 [Fig. postcrossing callosal axons and their growth cones
2(G)]. Thus higher frequencies of calcium transients were imaged for 20 min in the presence of pharmaco-
are well correlated with higher rates of callosal axon logical inhibitors (see Fig. 3). Treatment with 2-APB
outgrowth [Fig. 2(H)]. Amplitudes and durations of caused no overt defects in the morphology or motility
calcium transients were unrelated to rates of growth, of the growth cones [Fig. 3(C)] but slowed the rate of
indicating that frequency-dependent mechanisms in axon outgrowth to 31 6 5.6 lm h
1 (n ¼ 12 axons in
particular could regulate rates of axon advance five slices) an almost 50% reduction of control growth
through the corpus callosum. rate [Fig. 3(D)]. However, trajectories of individual
Calcium release from internal stores and entry callosal axons were similar to those of untreated con-
through TRP channels are important sources of cal- trols [Fig. 3(B,E)]. Importantly, a 30-min washout of
cium for regulating axon growth and guidance in the 2-ABP restored the rates of axon outgrowth. Treat-
Developmental Neurobiology
Figure 2 Callosal axons express spontaneous calcium transients that are correlated with rates of
axon outgrowth. (A) A coronal cortical slice in which plasmids encoding GCaMP2 were injected
and electroporated into the left cortex (ipsi). The arrow indicates the position of the growth cone
imaged in B–D, which had crossed the midline. Red curves indicate the borders of the corpus cal-
losum (cc) and the midline. The white line is autofluorescence from the slice holder used in live
cell imaging. (B) Tracing of calcium activity measured by the change in GCaMP2 fluorescence
over baseline. Calcium activity increases after a few minutes of imaging. (C) Tracing of calcium
activity from (B) zoomed in to the time period indicated by the bracket (B, bottom). (D) Fluores-
cence images of the growth cone from (B–C) at the time points indicated by arrowheads in (C). (E)
Within 20 min of the onset of calcium activity shown in (B) the axon begins to rapidly advance
through the contralateral callosum. (F) Examples of single calcium transients measured by ratio-
metric imaging in growth cones coexpressing DsRed2 and GCaMP2. (G) Plot of frequencies of cal-
cium transients in pre-crossing or post-crossing callosal axons. **p < 0.01, t test. All frequencies
in units of transients h
1. (H) Scatter plot of the frequency of calcium transients versus the rate of
axon outgrowth in individual callosal axons. The line represents the least-squares linear regression
(slope significantly non-zero, p < 0.01). (I) An example of spontaneous calcium transients (top
row) which are attenuated by application of SKF (time 0:00, bottom rows). (J) Tracing of calcium
activity in the growth cone shown in (I) before and after application of SKF. Scale bars, 10 lm
except I, which is 5 lm. Pseudocolor calibration bars indicate fluorescence intensity (D) or ratio of
GCaMP2 to DsRed2 fluorescence intensities (F) in arbitrary units.
 Wnt/Calcium in Callosal Axons 275

Figure 3 Blocking IP3 receptors and TRP channels reduces rates of postcrossing axon outgrowth
and blocking TRP channels leads to axon guidance defects. (A) Tracings of cortical axons express-
ing DsRed2 in the contralateral corpus callosum. Axons from different experiments were traced
and overlaid on a single outline of the corpus callosum. Curved lines, border of the corpus cal-
losum; vertical line, midline. (A, inset) Plot of growth cone distance from the midline versus axon
trajectory (see methods) in control experiments. The solid line represents a quadratic regression
curve which describes the standard trajectory taken by axons in control experiments; the dashed
lines represent the 90% prediction interval of the regression curve. (B) Tracings of cortical axons
in slices treated with 2-APB (blue) conformed to the standard trajectory of callosal axons without
deviating significantly (see Methods) while axons in slices treated with SKF96365 (red) deviated
dorsally toward the induseum griseum or ventrally toward the septum or lateral ventricle or cortical
plate in many cases (5 of 12 axons, arrowheads). (B, inset) Plot of growth cone distance from the
midline versus axon trajectory in axons in slices treated with SKF96365 (red) or 2-APB (blue). The
solid line indicates the standard trajectory derived from control axons and the dashed lines are the
90% prediction interval. (C) Time lapse images of a growth cone expressing DSRed2 extending
through the callosum after crossing the midline, during treatment with 2-APB. Scale bar, 10 lm.
(D) Rates of outgrowth of callosal axons under control conditions, during bath application of
2-APB or SKF96365, or after washout. n ¼ number of axons. (E) Measurement of the average
deviation of axons treated with 2-APB (n ¼ 10), SKF96365 (n ¼ 12) or medium (control, n ¼ 27)
from the standard trajectory. ***p < 0.001, One way ANOVA with Dunnett’s posttest. **p < 0.01,
*p < 0.05 One way ANOVA with Newman-Kewls posttest.

ment with SKF96365 (n ¼ 13 axons in five slices) also septum or the ventricle. In several cases [one example
reduced rates of axon outgrowth by about 50% (24.9 shown in Fig. 2(I,J) and Supporting Information,
6 3.8 lm h
1) which were restored close to control Movie 3] we were able to apply SKF to cortical slices
levels after washout. Remarkably blocking TRP chan- after imaging calcium activity in a postcrossing axon.
nels with SKF96365 caused severe misrouting of indi- In each case application of SKF attenuated ongoing
vidual callosal axons [5 of 12, Fig. 3(B,E)]. As shown calcium transients. Postcrossing axons treated with
in Figure 3(B), tracing of axon trajectories showed that SKF had a frequency of calcium transients similar to
some axons turned prematurely toward the cortical that of precrossing axons (2.99 6 1.36 per hour, n ¼
plate while others turned inappropriately toward the 10 for precrossing control axons vs. 3.2 6 2.33 per
Developmental Neurobiology
276 Hutchins et al.

hour, n ¼ 5 for SKF-treated postcrossing axons). This crossed the midline [Fig. 4(E)]. Ryk knockdown had
provides direct evidence that in callosal axons the no effect on precrossing growth rates [Fig. 4(F)]
growth and guidance defects observed after pharmaco- where Ryk is known to be inactive (Keeble et al.,
logical treatment with SKF were the result of 2006), demonstrating that electroporation with Ryk
decreased calcium activity. siRNA does not reduce rates of outgrowth in general
To quantify the deviation from the standard trajec- but rather selectively reduces rates of growth in the
tory of axons in the contralateral callosum, we first regions where Ryk is active. To further test for off
plotted the distance from the midline of DsRed target effects of siRNA we compared Ryk expression
expressing growth cones in control slices versus axon levels in cortical neurons electroporated with a con-
trajectory (the angle between the line formed by the trol pool of siRNA vs. mock transfection. Ryk
distal 20 lm of the axon and the horizontal axis of expression levels were the same in these two groups
the slice). These angles [Fig. 3(A), inset] increased as (Supporting Information Fig. S1), arguing against off
axons grew away from the midline reflecting the fact target effects of electroporation with siRNA. To
that axons turn dorsally after descending into the cal- assess whether Ryk knockdown disrupted the guid-
losum and crossing the midline. We then fit these ance of callosal axons we compared the trajectories
data with a nonlinear regression curve which of DsRed-labeled axons in control slices with axons
describes the standard trajectory of these axons. This in slices electroporated with Ryk siRNA [Fig. 4(A–
allowed us to compare the actual angle of an axon at C)]. We found that Ryk knockdown caused severe
a given distance from the midline versus the angle guidance errors in about a third of axons (n ¼ 7 out
predicted by the regression curve. As shown in Figure of 23) analyzed [Fig. 4(A,B)]. The variable effect on
3, axons in control and 2-APB-treated slices deviated axon guidance in siRNA-treated axons could be due
very little from the standard trajectory (14.78 6 2.28 to uneven knockdown of the Ryk receptor among
and 13.68 6 2.38, respectively) while axons in SKF axons. However, we were unable to test this possibil-
treated slices deviated significantly more (31.48 6 ity due to the ubiquitous expression of Ryk in the cor-
7.58, p < 0.01, One way ANOVA with Newman- tex (Keeble et al., 2006), which makes the detection
Kewls posttest). of Ryk expression on single axons against this back-
ground unfeasible. Similar results were obtained with
a second, independent pool of Ryk siRNA (Sup-
porting Information Fig. S1). As shown in the axon
Ryk Knockdown Disrupts Post-Crossing
tracings guidance errors of postcrossing callosal
Axonal Calcium Signaling, Rates of
axons involved premature dorsal turning toward the
Growth and Trajectories
overlying cortex or inappropriate ventral turning to-
Taken together, results thus far demonstrate the ward the septum.
requirement of calcium signaling mechanisms in cal- Results obtained in dissociated culture (Li et al.,
losal axon outgrowth and guidance but not the spe- 2009) showed that knocking down Ryk reduced the
cific involvement of Wnt5a signaling. In dissociated proportion of neurons that expressed calcium transi-
cortical cultures (Li et al., 2009) we found that ents in response to application of Wnt5a. Are the out-
knockdown of the Ryk receptor to Wnt5a prevented growth and guidance defects in the callosum of corti-
increased rates of axon outgrowth and repulsive cal slices in which Ryk was knocked down due to in-
growth cone turning evoked by Wnt5a. In vivo Ryk terference with Wnt evoked calcium signaling? To
knockout mice were found to have guidance errors in address this question we coelectroporated GCaMP2
callosal axons but the use of fixed material prevented with Ryk siRNA to monitor calcium activity in cal-
studies of signaling mechanisms downstream of Ryk losal growth cones in which Ryk/Wnt signaling has
(Keeble et al., 2006). We used electroporation of Ryk been disrupted. In these co-electroporated neurons
siRNA to knock down Ryk in a small number of cort- [Fig. 4(D,E)] frequencies of calcium transients were
ical axons to analyze cell autonomous functions of reduced to 3.4 6 2.2 transients h
1 compared to 12.6
Ryk in a wild type background; to visualize these transients h
1 for controls, a similar reduction in fre-
neurons and their axons, we co-electroporated DsRed. quency to that caused by treatment with SKF. Remark-
We used two pools of Ryk siRNA that we have exten- ably, in several cases we found that in growth cones
sively characterized in hamster cortical neurons (Li et projecting inappropriately toward the septum, calcium
al., 2009). Measurements of growth rates of fluores- transients were undetectable [Fig. 4(D)]. Taken to-
cently labeled axons revealed that postcrossing axons gether these results suggest that axon growth and guid-
slowed their growth rates to 28.4 6 3.2 lm h
1, ance errors caused by Ryk knockdown result from
about half the normal growth rate for axons that have attenuated calcium activity in callosal growth cones.
Developmental Neurobiology
 Wnt/Calcium in Callosal Axons 277

Figure 4 Ryk knockdown reduces frequencies of calcium transients, slows rates of axon exten-
sion, and causes axon guidance defects in post-crossing callosal axons. (A) Tracings of control
cortical axons expressing DsRed2 [also shown in Fig. 3(A)] in the contralateral corpus callosum.
(A, inset) Plot of growth cone distance from the midline versus axon trajectory in control experi-
ments. The solid line represents a quadratic regression curve which describes the standard trajec-
tory taken by axons in control experiments; the dashed lines represent the 90% prediction interval
of the regression curve. (B) Tracings of cortical axons in slices electroporated with DsRed2 and
anti-Ryk siRNA. Many of these axons with Ryk expression knocked down deviated dorsally toward
the induseum griseum or cortical plate or ventrally toward the septum (arrowheads; anti-Ryk
siRNA: 7 of 23 axons). (B, inset) Plot of growth cone distance from the midline versus axon trajec-
tory in Ryk knockdown experiments. The solid line indicates the standard trajectory derived from
control axons and the dashed lines are the 90% prediction interval. (C) Measurement of the average
deviation of axons expressing with DSRed2 plus anti-Ryk siRNA (n ¼ 23) or DsRed2 alone (con-
trol, n ¼ 27) from the standard axon trajectory. (D, left) Growth cones electroporated with Ryk
siRNA, also co-expressing DsRed2 (shown in left panels) and GCaMP2 that are extending toward
the septum (shown in (B) with hollow arrowheads). Scale bars, 10 lm. (D, right) Tracings of cal-
cium signals measured by ratiometric imaging showing that neither of these neurons express cal-
cium transients. (E) Quantifications of rates of axon outgrowth (left, black; n ¼ 27 for controls and
22 for Ryk siRNA experiments) and frequencies of calcium transients (right, white; n ¼ 14 for con-
trols and 10 for Ryk siRNA experiments) in post-crossing callosal axons. Units are transients h
1.
(F) Quantification of precrossing axon outgrowth in slices electroporated with DsRed or DsRed
plus Ryk siRNA (n > 6 axons from at least two slices). ***p < 0.001, **p < 0.01, t test.

CaMKII Regulates Repulsive Axon We reduced the activity of CaMKII by transfection of

Guidance plasmids encoding a specific CaMKII inhibitor pro-
tein, EGFP-CaMKIIN (Chang et al., 1998; Tang and
Since we found previously that CaMKII is also a
component of the Wnt/calcium signaling pathway (Li Kalil, 2005). For postcrossing but not precrossing axons
et al., 2009), (Supporting Information Fig. S2), we this treatment slowed the growth of callosal axons and
asked whether inhibiting CaMKII activity would caused guidance errors similar to those observed after
cause growth or guidance defects of callosal axons. Ryk knockdown. As shown in Figure 5(A,C) some
Developmental Neurobiology
278 Hutchins et al.

Since guidance errors in the callosum by Ryk

knockout were caused by interfering with Wnt5a
induced cortical axon repulsion (Keeble et al., 2006),
we asked whether CaMKII is also required for corti-
cal axon repulsion. To address this question we used
a Dunn chamber turning assay (Yam et al., 2009) in
which cortical neurons were exposed to a Wnt5a gra-
dient (Supporting Information Fig. S3) and their
growth cone turning angles measured over 2 h. As
shown in Figure 6(B), measurement of the Wnt5a
gradient in the Dunn chamber, as measured with a
fluorescent dextran conjugate similar in molecular
weight to Wnt5a, showed that a high to low Wnt5a
gradient was established in the bridge region of the
chamber that persisted for the 2-h duration of the
experiments. As we found previously in a pipette
turning assay (Li et al., 2009), growth cones of neu-
rons in the bridge region of the Dunn chamber consis-
tently turned away from Wnt5a gradients and
increased their growth rates by 50% [Figs. 6(C–E)
and S4]. In contrast when cortical neurons were trans-
fected with CaMKIIN they failed to increase their
rates of axon growth [Fig. 6(C)]. Importantly inhibi-
tion of CaMKII prevented axons from repulsive turn-
ing in response to Wnt5a and these axons continued
extending in their original trajectories [Fig. 6(D,E)].
Figure 5 CaMKII regulates cortical axon outgrowth and
guidance in the corpus callosum. (A) Tracings of cortical
These results suggest that, as with inhibition of Ryk
axons in slices electroporated with EGFP-CaMKIIN, which receptors (Li et al., 2009), reducing CaMKII activity
deviated dorsally toward the induseum griseum or cortical slows axon outgrowth and prevents Wnt5a growth
plate or ventrally toward the lateral ventricle in many cases cone repulsion.
(arrowheads; 7 of 16 axons). (A, inset) Plot of growth cone
distance from the midline versus axon trajectory in axons in
slices electroporated with EGFP-CaMKIIN.The solid line DISCUSSION
indicates the standard trajectory derived from control axons
and the dashed lines are the 90% prediction interval. (B) Taken together these results show that in a cortical
Rates of axon outgrowth in cortical neurons expressing slice model of the developing corpus callosum Wnt/
DSRed2 (control) or EGFP-CaMKIIN in pre- or postcross-
calcium signaling pathways, that we previously iden-
ing callosal axons. n ¼ number of axons. **p < 0.01, One
way ANOVA with Bonferroni’s posttest. (C) Measurement tified in dissociated cortical cultures (Li et al., 2009),
of the average deviation of axons expressing with EGFP- are essential for regulating callosal axon growth and
CaMKIIN (n ¼ 16) or DsRed2 (control, n ¼ 27) from the guidance. First we show that rates of callosal axon
standard trajectory. **p < 0.01, t test. outgrowth are almost 50% higher on the contralateral
side of the callosum. Second we find that higher fre-
quencies of calcium transients in postcrossing growth
transfected axons showed dramatic misrouting in which cones are strongly correlated with higher rates of out-
axons looped backwards in the callosum, prematurely growth in contrast to precrossing growth cones. Third
extended dorsally toward the cortical plate or grew we show that blocking IP3 receptors with 2-APB
abnormally toward the ventricle. In these experiments slows the rate of postcrossing axon growth rates but
we compared rates of precrossing (n ¼ 12 axons in does not affect axon guidance. In contrast blocking
four slices) vs. postcrossing (n ¼ 12 axons in five TRP channels not only reduces axon growth rates but
slices) callosal axons [Fig. 5(B)] and found that rates causes misrouting of postcrossing callosal axons.
of postcrossing axon outgrowth were reduced by Downstream of calcium, we found that CaMKII is
about 50% (36.2 6 4.0 vs. 54.6 6 2.9 lm h
1 essential for normal axon growth and guidance, dem-
for control axons) but rates of precrossing axon onstrating the importance of calcium signaling for de-
outgrowth were unaffected [Fig. 5(B)]. velopment of the corpus callosum. Finally, we dis-
Developmental Neurobiology
 Wnt/Calcium in Callosal Axons 279

Figure 6 CaMKII activity is required for repulsive growth cone turning away from a gradient of
Wnt5a. (A) At left, cortical growth cones responding to Wnt5a gradients in Dunn chambers over 2
h. Images have been oriented such that high-to-low concentration gradients of BSA (vehicle con-
trol) or Wnt5a are highest at the top of the images. (Top panel) Control growth cones in BSA con-
tinue straight trajectories. (Middle panels) Three different growth cones show marked repulsive
turning in Wnt5a gradients. (Bottom panel) Transfection with CaMKIIN abolishes Wnt5a induced
repulsion. Scale bars, 10 lm. (B) A graph of fluorescence intensity (Z axis) of a gradient of 40 kDa
Texas Red dextran at different positions in the bridge region of the Dunn chamber. A high-to-low
gradient (along the X axis) is formed from the edge of the bridge region facing the outer chamber
containing Texas Red dextran (0 lm) to the edge facing the inner chamber lacking Texas Red dex-
tran. This gradient persists for at least 2 h (Y axis). (C) Rates of outgrowth of control- or CaM-
KIIN-transfected axons in Dunn chambers treated with gradients of BSA or Wnt5a. (D) Cumulative
distribution graph of turning angles of control- or CaMKIIN-transfected axons in Dunn chambers
treated with gradients of BSA or Wnt5a. **p < 0.01, Wilcoxon signed rank test. (E) Graph of turn-
ing angles of control- or CaMKIIN-transfected axons in Dunn chambers treated with gradients of
BSA or Wnt5a. **p < 0.01, ANOVA on Ranks with Dunn’s posttest.

covered that knocking down Ryk expression reduces frequencies of growth cone calcium transients
postcrossing axon outgrowth and induces aberrant (Gomez and Spitzer, 1999). Here we show that cal-
trajectories. Importantly we show that these defects losal growth cones express repetitive calcium transi-
in axons treated with Ryk siRNA correspond with ents as they navigate across the callosum. In contrast
reduced calcium activity. These results suggest a to results in the Xenopus spinal cord, higher levels of
direct link between calcium regulation of callosal calcium activity are correlated with faster rates of
axon growth and guidance and Wnt/Ryk signaling. outgrowth. One possibility to account for these differ-
Although calcium transients in growth cones of ences is that in callosal growth cones calcium transi-
dissociated neurons have been extensively docu- ents were brief, lasting *1 s, whereas in Xenopus spi-
mented in regulating axon outgrowth and guidance nal growth cones calcium transients were long last-
(Henley and Poo, 2004; Gomez and Zheng, 2006; ing, averaging almost 1 min (Gomez and Spitzer,
Wen and Zheng, 2006), the role of axonal calcium 1999; Lautermilch and Spitzer, 2000). Thus calcium
transients has been little studied in vivo. A previous transients in Xenopus that slow axon outgrowth could
live cell imaging study of calcium transients in vivo represent a different kind of calcium activity, consist-
in the developing Xenopus spinal cord demonstrated ent with the finding that rates of axon outgrowth in
that rates of axon outgrowth are inversely related to dissociated spinal neuronal cultures were insensitive
Developmental Neurobiology
280 Hutchins et al.

to inhibitors of CaMKII (Zheng et al., 1994; Lauter- and their role in callosal axon guidance across the
milch and Spitzer, 2000). In dissociated cortical cul- midline has been well characterized (Serafini et al.,
tures calcium activity in growing axons was similar 1996; Shu and Richards, 2001; Shu et al., 2003; Lind-
in frequency and duration to callosal growth cones wall et al., 2007; Niquille et al., 2009; Piper et al.,
extending in slices (Hutchins and Kalil, 2008). Some 2009). However, our finding that inhibiting calcium
callosal growth cones exhibit calcium activity local- signaling only affected growth and guidance of axons
ized to the growth cone or even small regions of the after but not before the callosal midline suggested
growth cone, raising the possibility that asymmetries that these effects were due to axonal responses only
in levels of calcium could play a role in growth cone after they had crossed the midline. This points to the
steering in vivo as they do in isolated growth cones possible involvement of Wnt5a signaling, because,
(Henley and Poo, 2004). Thus the present study is the cortical axons do not respond to Wnt5a until the age
first to demonstrate the importance of repetitive cal- at which they cross the midline (Keeble et al., 2006).
cium transients for axon outgrowth and guidance in a Although Slit2 affects both pre- and postcrossing cal-
developing mammalian CNS pathway. losal axons (Shu and Richards, 2001; Shu et al.,
Previous studies have shown the importance of the 2003), our results showed that interfering with cal-
source of calcium activity for effects on axon growth cium/CaMKII affected only postcrossing axon out-
and guidance (Ooashi et al., 2005; Jacques-Fricke et growth. This result points to a mechanism that is spe-
al., 2006). For example, transients resulting from cal- cific to the contralateral hemisphere such as Wnt5a
cium entry through L-type channels was found to in- signaling. This interpretation is supported by our sim-
hibit axon outgrowth in dissociated cortical cultures ilar results on postcrossing axon outgrowth after
(Tang et al., 2003; Hutchins and Kalil, 2008). In con- knocking down the Ryk receptor. We recently
trast calcium release from stores through IP3 recep- showed that Wnt/Ryk signaling evokes repetitive cal-
tors promotes axon outgrowth (Takei et al., 1998; cium transients in cortical axons in dissociated cul-
Jacques-Fricke et al., 2006; Li et al., 2009). In the tures (Li et al., 2009). Although Wnt5a is thought to
present study blocking IP3 receptors reduced rates of be the ligand for Ryk in the developing corpus cal-
axon outgrowth by about 50% on the postcrossing losum (Keeble et al., 2006), it is possible that other
side of the callosum, showing for the first time that Wnt proteins contribute to Ryk-dependent outgrowth
axons growing in developing mammalian pathways and guidance in the contralateral callosum. We dem-
use similar calcium signaling mechanisms to regulate onstrate here that knocking down the Ryk receptor
their growth rates. Recent in vitro studies of axon reduces rates of callosal axon outgrowth and that, in
guidance in response to application of netrin-1 or axons lacking Ryk receptors, growth, and guidance
BDNF have shown the importance of calcium entry errors coincide with reduced calcium activity. Since
via TRP channels to induce attractive or repulsive Wnts are currently the only known ligands for the
growth cone turning (Li et al., 2005; Shim et al., Ryk receptor (Fradkin et al., 2010), this remarkable
2005; Wang and Poo, 2005). Similarly we found that result provides compelling evidence that calcium reg-
in dissociated cortical cultures repulsive turning of ulation of postcrossing axon growth and guidance in
cortical growth cones in Wnt5a gradients were inhib- the callosum is specifically evoked by Wnt signaling.
ited when TRP channels were blocked (Li et al., Although it is possible that signaling pathways
2009) although this also reduced rates of axon out- other than Wnt/Ryk activate downstream calcium/
growth. This result is consistent with the recent find- CaMKII signaling in callosal axons, the selective
ing that pharmacologically blocking TRP channels or effects of inhibiting CaMKII on postcrossing axon
knocking down TRPC5 reduces rates of hippocampal outgrowth suggest that Ryk is a primary activator of
axon outgrowth (Davare et al., 2009). Here we find the CaMKII signaling cascade. Our work in vitro
that application of TRP channel blockers to cortical indicates that CaMKII is specifically phosphorylated
slices blocks calcium transients and reduces rates of by Wnt5a signaling (our unpublished observations).
callosal axon outgrowth but also causes severe mis- Frequencies of calcium transients were 75% lower in
routing of callosal axons. This demonstrates the axons on the precrossing side of the callosum, where
requirement of TRP channels for axon guidance in Ryk is not active (Keeble et al., 2006). This result
the mammalian CNS. suggests that guidance cues affecting callosal axons
Although these results show the importance of cal- in that region do not promote outgrowth through fre-
cium signaling in regulating callosal growth and quency dependent calcium signaling, which is con-
guidance, calcium activity could be evoked by multi- sistent with our finding that inhibiting CaMKII had
ple guidance cues. For example, sources of netrins, no effect on the growth of precrossing axons. Further-
semaphorins, and Slit2 surround the corpus callosum more, our results which show that CaMKII promotes
Developmental Neurobiology
 Wnt/Calcium in Callosal Axons 281

Figure 7 Wnt/calcium signaling guiding axon growth through the corpus callosum. (A) A sche-
matic of a cortical slice showing cortical axons extending through the corpus callosum (CC) around
which Wnt5a is expressed in the induseum griesium (IG) and glial wedge (GW). Axons in the con-
tralateral callosum have high frequencies of calcium transients caused by Wnt5a signaling which
accelerate rates of axon outgrowth. In contrast, ipsilateral axons, which are not yet responsive to
Wnt5a, have little calcium activity and therefore grow more slowly. LV, lateral ventricle; CP, corti-
cal plate. (B) A schematic of intracellular signaling in the post-crossing axon from (A). A Wnt5a
gradient from the glial wedge activates Ryk receptors to open IP3 receptors and TRP channels,
causing release of calcium from intracellular stores and calcium influx through the plasma mem-
brane. These transients activate CaMKII to accelerate axon extension and repel growth cones away
from the glial wedge toward the contralateral cortical plate.

axonal growth selectively in regions of high-fre- et al., 2005; Hutchins and Li, 2009) and netrin-1
quency calcium activity suggests an instructive rather (Hong et al., 2000; Henley and Poo, 2004; Wang and
than permissive role for CaMKII in regulating growth Poo, 2005) or turning away from repulsive cues such
cone advance. This interpretation is supported by our as myelin-associated glycoprotein (MAG), (Henley
finding that experimentally reducing frequencies of et al., 2004) involve Ca2+ gradients in growth cones
calcium transients also reduces rates of axon out- with the elevated side facing toward the source of the
growth to levels seen in precrossing axons with natu- guidance cue (Zheng et al., 1994; Henley and Poo,
rally low calcium activity. The lack of any additive 2004; Wen et al., 2004; Jin et al., 2005; Gomez and
effects when calcium transients are pharmacologi- Zheng, 2006). One model of calcium signaling in
cally suppressed in axons expressing the CaMKII in- growth cone turning proposed that the amplitude of
hibitor CaMKIIN (Supporting Information Fig. S5) calcium gradients was higher in attractive growth cone
indicates that CaMKII does not have any calcium fre- turning but lower in repulsion (Wen et al., 2004).
quency-independent effects in callosal axons, further These different calcium gradients are detected by dif-
demonstrating an instructive role for CaMKII in cal- ferent calcium sensors such that high amplitude cal-
losal axon outgrowth. cium signals in attraction are detected by CaMKII and
Taken together, our results from dissociated cortical low amplitude signals in repulsion are detected by cal-
cultures (Li et al., 2009) and the present findings in cineurin. Thus our finding that CaMKII is involved in
cortical slices support a repulsive guidance function growth cone repulsion is surprising given that a role
for Wnt5a on cortical axons (see Fig. 7) in agreement for CaMKII has only been described for chemoattrac-
with previous studies (Liu et al., 2005; Keeble et al., tion (Wen et al., 2004; Wen and Zheng, 2006).
2006; Zou and Lyuksyutova, 2007). However, calcium Furthermore, the finding that CaMKII is required for
signaling mechanisms underlying growth cone turning axon guidance in the callosum emphasizes the impor-
in response to guidance cues remain poorly under- tance of these calcium-dependent guidance behaviors
stood. One recent study, on the basis of asymmetric in vivo. A previous study of calcium signaling path-
membrane trafficking in growth cones with calcium ways activating CaMKK and CaMKI reported no axon
asymmetries, suggested that attraction and repulsion guidance or extension defects during midline crossing,
are not simply opposite polarities of the same mecha- but rather showed reduced axon branching into cortical
nism but distinct mechanisms (Tojima et al., 2007). target regions (Ageta-Ishihara et al., 2009). This result
Axon growth and turning behaviors in response to suggests that distinct calcium/calmodulin kinases may
attractive cues such as BDNF (Song et al., 1997; Li be activated during different stages of callosal devel-
Developmental Neurobiology
282 Hutchins et al.

opment. Taken together, these findings indicate that Jacques-Fricke BT, Seow Y, Gottlieb PA, Sachs F, Gomez
Wnt5a-evoked calcium/CaMKII signaling instructs TM. 2006. Ca2+ influx through mechanosensitive chan-
specific growth and guidance behaviors in the corpus nels inhibits neurite outgrowth in opposition to other
callosum rather than reflecting a more general require- influx pathways and release from intracellular stores. J
Neurosci 26:5656–5664.
ment for calcium signaling during development.
Jin M, Guan CB, Jiang YA, Chen G, Zhao CT, Cui K, Song
YQ, et al. 2005. Ca2+-dependent regulation of rho GTPases
The authors thank Dr. Erik Dent for helpful comments
triggers turning of nerve growth cones. J Neurosci 25:
on the manuscript. They also thank Dr. Junichi Nakai and
2338–2347.
Addgene.org for providing the GCaMP2 plasmid.
Keeble TR, Halford MM, Seaman C, Kee N, Macheda M,
Anderson RB, Stacker SA, et al. 2006. The Wnt receptor
Ryk is required for Wnt5a-mediated axon guidance on
REFERENCES the contralateral side of the corpus callosum. J Neurosci
26:5840–5848.
Ageta-Ishihara N, Takemoto-Kimura S, Nonaka M, Ada- Lautermilch NJ, Spitzer NC. 2000. Regulation of calci-
chi-Morishima A, Suzuki K, Kamijo S, Fujii H, et al. neurin by growth cone calcium waves controls neurite
2009. Control of cortical axon elongation by a GABA- extension. J Neurosci 20:315–325.
driven Ca2+/calmodulin-dependent protein kinase cas- Li L, Hutchins BI, Kalil K. 2009. Wnt5a induces simultane-
cade. J Neurosci 29:13720–13729. ous cortical axon outgrowth and repulsive axon guidance
Bortone D, Polleux F. 2009. KCC2 expression promotes the through distinct signaling mechanisms. J Neurosci
termination of cortical interneuron migration in a voltage- 29:5873–5883.
sensitive calcium-dependent manner. Neuron 62:53–71. Li Y, Jia YC, Cui K, Li N, Zheng ZY, Wang YZ, Yuan XB.
Chang BH, Mukherji S, Soderling TR. 1998. Characteriza- 2005. Essential role of TRPC channels in the guidance of
tion of a calmodulin kinase II inhibitor protein in brain. nerve growth cones by brain-derived neurotrophic factor.
Proc Natl Acad Sci USA 95:10890–10895. Nature 434:894–898.
Davare MA, Fortin DA, Saneyoshi T, Nygaard S, Kaech S, Lindwall C, Fothergill T, Richards LJ. 2007. Commissure
Banker G, Soderling TR, et al. 2009. Transient receptor formation in the mammalian forebrain. Curr Opin Neuro-
potential canonical 5 channels activate Ca2+/calmodulin biol 17:3–14.
kinase Igamma to promote axon formation in hippocam- Liu Y, Shi J, Lu CC, Wang ZB, Lyuksyutova AI, Song XJ,
pal neurons. J Neurosci 29:9794–9808. Zou Y. 2005. Ryk-mediated Wnt repulsion regulates pos-
Fradkin LG, Dura JM, Noordermeer JN. 2010. Ryks: New terior-directed growth of corticospinal tract. Nat Neuro-
partners for Wnts in the developing and regenerating sci 8:1151–1159.
nervous system. Trends Neurosci 33:84–92. Mao T, O’Connor DH, Scheuss V, Nakai J, Svoboda K.
Gomez TM, Spitzer NC. 1999. In vivo regulation of axon 2008. Characterization and subcellular targeting of
extension and pathfinding by growth-cone calcium tran- GCaMP-type genetically-encoded calcium indicators.
sients. Nature 397:350–355. PLo S One 3:e1796.
Gomez TM, Zheng JQ. 2006. The molecular basis for cal- Neal AP, Molina-Campos E, Marrero-Rosado B, Bradford
cium-dependent axon pathfinding. Nat Rev Neurosci AB, Fox SM, Kovalova N, Hannon HE. 2010. CaMKK-
7:115–125. CaMKI signaling pathways differentially control axon
Halloran MC, Kalil K. 1994. Dynamic behaviors of growth and dendrite elongation in cortical neurons. J Neurosci
cones extending in the corpus callosum of living cortical 30:2807–2809.
brain slices observed with video microscopy. J Neurosci Niquille M, Garel S, Mann F, Hornung JP, Otsmane B,
14:2161–2177. Chevalley S, Parras C, et al. 2009. Transient neuronal
Henley JR, Huang KH, Wang D, Poo MM. 2004. Calcium populations are required to guide callosal axons: A role
mediates bidirectional growth cone turning induced by for semaphorin 3C. PLoS Biol 7:e1000230.
myelin-associated glycoprotein. Neuron 44:909–916. Ooashi N, Futatsugi A, Yoshihara F, Mikoshiba K, Kami-
Henley JR, Poo MM. 2004. Guiding neuronal growth cones guchi H. 2005. Cell adhesion molecules regulate Ca2+-
using Ca2+ signals. Trends Cell Biol 14:320–330. mediated steering of growth cones via cyclic AMP and
Hong K, Nishiyama M, Henley J, Tessier-Lavigne M, Poo ryanodine receptor type 3. J Cell Biol 170:1159–1167.
M. 2000. Calcium signaling in the guidance of nerve Piper M, Plachez C, Zalucki O, Fothergill T, Goudreau G,
growth by netrin-1. Nature 403:93–98. Erzurumlu R, Gu C, et al. 2009. Neuropilin 1-Sema sig-
Hutchins BI. 2010. Competitive outgrowth of neural proc- naling regulates crossing of cingulate pioneering axons
esses arising from long-distance cAMP signaling. Sci during development of the corpus callosum. Cereb Cor-
Signal 3:jc1. tex 19 (Suppl 1):i11–i21.
Hutchins BI, Kalil K. 2008. Differential outgrowth of axons Salinas PC, Zou Y. 2008. Wnt signaling in neural circuit as-
and their branches is regulated by localized calcium tran- sembly. Annu Rev Neurosci 31:339–358.
sients. J Neurosci 28:143–153. Serafini T, Colamarino SA, Leonardo ED, Wang H, Bed-
Hutchins BI, Li L. 2009. EphrinA and TrkB interact to pro- dington R, Skarnes WC, Tessier-Lavigne M. 1996.
mote axon branching. J Neurosci 29:4329–4331. Netrin-1 is required for commissural axon guidance in

Developmental Neurobiology
 Wnt/Calcium in Callosal Axons 283

the developing vertebrate nervous system. Cell 87:1001– involves asymmetric membrane transport and exocytosis
1014. in the growth cone. Nat Neurosci 10:58–66.
Shim S, Goh EL, Ge S, Sailor K, Yuan JP, Roderick HL, Boot- Uesaka N, Hirai S, Maruyama T, Ruthazer ES, Yama-
man MD, et al. 2005. XTRPC1-dependent chemotropic guid- moto N. 2005. Activity dependence of cortical axon
ance of neuronal growth cones. Nat Neurosci 8:730–735. branch formation: A morphological and electrophysio-
Shu T, Richards LJ. 2001. Cortical axon guidance by the logical study using organotypic slice cultures. J Neuro-
glial wedge during the development of the corpus cal- sci 25:1–9.
losum. J Neurosci 21:2749–2758. Uhlen P. 2004. Spectral analysis of calcium oscillations.
Shu T, Sundaresan V, McCarthy MM, Richards LJ. 2003. Sci STKE 2004:pl15.
Slit2 guides both precrossing and postcrossing callosal Wang CL, Zhang L, Zhou Y, Zhou J, Yang XJ, Duan SM,
axons at the midline in vivo. J Neurosci 23:8176–8184. Xiong ZQ, et al. 2007. Activity-dependent development
Song HJ, Ming GL, Poo MM. 1997. cAMP-induced switch- of callosal projections in the somatosensory cortex. J
ing in turning direction of nerve growth cones. Nature Neurosci 27:11334–11342.
388:275–279. Wang GX, Poo MM. 2005. Requirement of TRPC channels
Takei K, Shin RM, Inoue T, Kato K, Mikoshiba K. 1998. in netrin-1-induced chemotropic turning of nerve growth
Regulation of nerve growth mediated by inositol 1,4,5- cones. Nature 434:898–904.
trisphosphate receptors in growth cones. Science Wen Z, Guirland C, Ming GL, Zheng JQ. 2004. A CaMKII/
282:1705–1708. calcineurin switch controls the direction of Ca(2+)-
Tallini YN, Ohkura M, Choi BR, Ji G, Imoto K, Doran R, dependent growth cone guidance. Neuron 43:835–846.
Lee J, et al. 2006. Imaging cellular signals in the heart in Wen Z, Zheng JQ. 2006. Directional guidance of nerve
vivo: Cardiac expression of the high-signal Ca2+ indica- growth cones. Curr Opin Neurobiol 16:52–58.
tor GCaMP2. Proc Natl Acad Sci USA 103:4753–4758. Yam PT, Langlois SD, Morin S, Charron F. 2009. Sonic
Tang F, Dent EW, Kalil K. 2003. Spontaneous calcium hedgehog guides axons through a noncanonical, Src-
transients in developing cortical neurons regulate axon family-kinase-dependent signaling pathway. Neuron 62:
outgrowth. J Neurosci 23:927–936. 349–362.
Tang F, Kalil K. 2005. Netrin-1 induces axon branching in Zheng JQ, Felder M, Connor JA, Poo MM. 1994. Turning
developing cortical neurons by frequency-dependent cal- of nerve growth cones induced by neurotransmitters.
cium signaling pathways. J Neurosci 25:6702–6715. Nature 368:140–144.
Tojima T, Akiyama H, Itofusa R, Li Y, Katayama H, Miya- Zou Y, Lyuksyutova AI. 2007. Morphogens as conserved
waki A, Kamiguchi H. 2007. Attractive axon guidance axon guidance cues. Curr Opin Neurobiol 17:22–28.

Developmental Neurobiology