Research ArticleAxon Guidance

Control of Neuronal Growth Cone Navigation by Asymmetric Inositol 1,4,5-Trisphosphate Signals

See allHide authors and affiliations

Science Signaling  14 Jul 2009:
Vol. 2, Issue 79, pp. ra34
DOI: 10.1126/scisignal.2000196

Abstract

Inositol 1,4,5-trisphosphate (IP3) is generally viewed as a global messenger that increases cytosolic calcium ion (Ca2+) concentration. However, the spatiotemporal dynamics of IP3 and the functional significance of localized IP3 production in cell polarity remain largely unknown. Here, we demonstrate the critical role of spatially restricted IP3 signals in axon guidance. We found that IP3 and ensuing Ca2+ signals were produced asymmetrically across growth cones exposed to an extracellular gradient of nerve growth factor (NGF) and mediated growth cone turning responses to NGF. Moreover, photolysis-induced production of IP3 on one side of a growth cone was sufficient to initiate growth cone turning toward the side with the higher concentration of IP3. Thus, locally produced IP3 encodes spatial information that polarizes the growth cone for guided migration.

Introduction

During development, neurons extend axons to create neuronal networks. The growth cone, a highly motile structure at the tip of the elongating axon that senses guidance cues in the extracellular environment, plays a central role in establishing the proper connections (1, 2). Various guidance molecules attract or repel the developing axon by means of asymmetric Ca2+ signals in the growth cone (35): It turns either toward the side with higher Ca2+ concentration (attraction) or toward the side with lower Ca2+ concentration (repulsion). The directional polarity of growth cone turning with regard to the localization of Ca2+ signals depends on the source of Ca2+ signals: Ca2+ release from the endoplasmic reticulum (ER) through ryanodine receptors (RyRs) triggers attraction (6), whereas Ca2+ influx from the extracellular space through transient receptor potential channels (7) or cyclic nucleotide–gated channels (8) is implicated in repulsion. Therefore, understanding navigation depends critically on elucidating how specific Ca2+ channels are activated by guidance molecules that bind to growth cones. One possibility is that asymmetric Ca2+ signals are generated through localized production of inositol 1,4,5-trisphosphate (IP3). However, no direct evidence exists to support the notion that such local IP3 signaling controls growth cone navigation; rather, IP3 is viewed as a global messenger.

When a growth cone encounters an extracellular gradient of nerve growth factor (NGF), it turns toward the higher concentration of NGF (9). Growth cone attraction to NGF depends on the NGF receptor TrkA and on its downstream effectors, phospholipase C (PLC) and phosphatidylinositol 3-kinase (PI3K) (10, 11). Two cytosolic second messengers, Ca2+ and adenosine 3′,5′-monophosphate (cAMP), also play key roles in NGF-mediated turning responses: Eliminating Ca2+ signals abolishes growth cone turning, and blocking cAMP signaling converts attraction into repulsion (11). PLC activation can increase cytosolic Ca2+ concentration ([Ca2+]c) through IP3 production and subsequent IP3-induced Ca2+ release (IICR) from the ER (12), suggesting the involvement of the IP3 signaling cascade—and IICR—in NGF-induced growth cone turning. However, despite the importance of localized signals in polarized cell motility, including growth cone navigation, the spatial profile of IP3 signaling has not been well characterized.

Here, we use a genetically encoded fluorescent IP3 sensor and show that NGF elicits asymmetric IP3 production across the growth cone, with higher cytosolic IP3 concentration ([IP3]c) on the side facing the NGF source. This increase in [IP3]c causes an asymmetric [Ca2+]c increase through IICR, mediating attractive turning of the growth cone in response to NGF. We also show that asymmetric IP3 production followed by IICR is sufficient for growth cone attraction. By identifying the signal transduction cascade underlying NGF-mediated axon attraction, we show that locally produced IP3 encodes spatial information that polarizes the growth cone for guided migration.

Results

PLC-IP3 signaling pathway is required for attractive turning of growth cones toward NGF

We used an in vitro turning assay (13, 14) to investigate the chemoattractive effect of NGF on embryonic chick dorsal root ganglion (DRG) neurons grown on L1 or laminin (Fig. 1). Although the concentration of cAMP in growth cones grown on laminin is low (6, 15), and NGF repels growth cones when the cAMP signaling pathway is inhibited (11), an extracellular NGF gradient produced by pulsatile pressure ejection of NGF from a micropipette caused attractive turning of DRG growth cones on both substrata (Fig. 1, A, top panels, and B).

Fig. 1

The PLC-IP3 signaling pathway is required for NGF-mediated growth cone attraction. (A) DIC images of chick DRG growth cones cultured on an L1 substrate in the absence (top) or presence of bath-applied U-73122 (middle) or xestospongin C (bottom). (Left) Immediately and (right) 60 min after the start of repetitive NGF ejection from the direction indicated by arrows. Scale bar, 10 μm. (B) Angles of growth cone turning responses to NGF gradients on L1 or laminin. “PBS gradient” represents repetitive PBS ejection as a control. Positive and negative values of the abscissa indicate attraction and repulsion, respectively. Numbers in parentheses indicate the number of growth cones examined. *P < 0.05, **P < 0.01; ns, not significant versus control on the same substrate, Dunnett’s multiple comparison test. (C) The targets of pharmacological agents used in this study: K-252a, a TrkA inhibitor; U-73122, a PLC inhibitor; wortmannin, a PI3K inhibitor; xestospongin C, an IICR inhibitor; 2-APB, an IICR inhibitor; ryanodine, a CICR inhibitor when used at a high concentration; thapsigargin, a Ca2+-ATPase inhibitor.

Pharmacological inhibition of TrkA, or of its downstream effectors PLC or PI3K, blocked NGF-mediated attraction (Fig. 1, A, middle panels, and B), as previously reported (10, 11). Thapsigargin, which causes Ca2+-store depletion by inhibiting Ca2+-ATPase (adenosine triphosphatase) on the ER membrane, also blocked NGF-induced attraction, implicating Ca2+ release from the ER in the response, as did the IICR inhibitors xestospongin C (16) and 2-aminoethoxydiphenyl borate [2-APB (17)] (Fig. 1, A, bottom panels, and B). The presence of IP3 receptors (IP3Rs) throughout DRG growth cones was confirmed by immunocytochemistry (fig. S1). These data indicate that IICR, presumably triggered by TrkA-mediated activation of PLC, is required for attractive turning to NGF.

The possible involvement of Ca2+-induced Ca2+ release (CICR) from the ER through RyRs was tested by treating neurons with a high concentration of ryanodine (100 μM) to trap RyRs in the closed state and thereby inhibit CICR (18). Ryanodine had no effect on NGF-induced growth cone attraction (Fig. 1B), indicating that this response did not require CICR.

NGF gradients elicit asymmetric IP3 production in growth cones

The intracellular signals responsible for polarizing the growth cone in preparation for turning must be generated asymmetrically. Therefore, we investigated the spatial profile of [IP3]c in growth cones before and after exposure to NGF gradients (Fig. 2). Cytosolic IP3 was monitored with a genetically encoded fluorescent IP3 sensor, IP3R-based IP3 sensor 1 (IRIS-1), that contains enhanced cyan fluorescent protein (ECFP) and Venus (a yellow fluorescent protein variant) fused in tandem to the IP3-binding domain of the type 1 IP3R (19). Fluorescence resonance energy transfer (FRET) from ECFP to Venus decreases on IP3 binding to IRIS-1. Therefore, increases in [IP3]c increase the ratio of ECFP emission to Venus emission (RFRET). When an extracellular gradient of NGF was applied to the growth cone, [IP3]c on the side facing the NGF source became higher than that on the opposite side (“Near” and “Far” in Fig. 2, A and B). This asymmetry in [IP3]c was quantified by averaging the relative change in RFRET (RFRET/RFRET-base) in a region of interest (ROI) defined on either the near or the far side of the growth cone (for details, see Materials and Methods).

Fig. 2

NGF gradients elicit asymmetric increases in [IP3]c in growth cones. (A) Fluorescence images of a growth cone expressing IRIS-1 grown on L1. The grayscale image (left) shows the Venus fluorescence of IRIS-1 before application of NGF. ROIs used to calculate RFRET/RFRET-base (defined as R′) on both sides of the growth cone (Near and Far) are shown. Pseudo color images represent R′ of IRIS-1 before (middle) and after (right) application of NGF (arrow). Scale bar, 10 μm. (B) Time course of changes in R′ in the near (red line) and far (blue line) ROIs of the growth cone shown in (A). (C) R′ averaged during the period from 1 to 3 min after the start of repetitive NGF ejection was used as a measure of the amplitude of NGF-mediated R′ increases. Amplitude of R′ averaged within either the near (red bars) or the far ROI (blue bars) in growth cones expressing either IRIS-1 or IRIS-1–Dmut. Analyses were done on either L1 or laminin. Numbers in parentheses indicate the number of growth cones examined. ***P < 0.001; ns, not significant, Bonferroni’s multiple comparison test. (D) Time course of changes in R′ of IRIS-1 in the near (red circles) and far (blue circles) ROIs (n = 9 growth cones). R′ in each growth cone on L1 was determined before (control, open circles) and after (+U-73122, filled circles) bath application of U-73122. (E) Time course of changes in R′ on the near side relative to those on the far side. Note that Rnear/Rfar is an index of [IP3]c asymmetry. Open and filled circles indicate Rnear/Rfar before and after the treatment with U-73122, respectively. (F) Rnear/Rfar averaged during the period from 1 to 3 min after the start of repetitive NGF ejection. Each growth cone on L1 was analyzed before (pre, open bars) and after (post, hatched bars) treatment with vehicle or U-73122. ***P < 0.001; ns, not significant between pre versus post of each treatment, Bonferroni’s multiple comparison test.

NGF-mediated increases in [IP3]c were observed in the near ROI, whereas no [IP3]c changes were detectable in the far ROI (Fig. 2C). The temporal profile of NGF-induced increased [IP3]c in growth cones grown on laminin was indistinguishable from that grown on L1 (fig. S2A). NGF did not affect RFRET/RFRET-base in growth cones harboring IRIS-1–Dmut, a mutated form of IRIS-1 that does not bind IP3 (19) (Fig. 2C). We compared the NGF-induced [IP3]c change to that elicited in the same growth cone after treatment with bath-applied U-73122 and found that U-73122 treatment blocked NGF-induced [IP3]c increases and thus abolished [IP3]c asymmetry (Fig. 2, D to F). These results indicate that NGF gradients induce PLC-dependent asymmetric IP3 production in growth cones.

NGF gradients induce asymmetric [Ca2+]c increases in growth cones

We used combined Oregon Green 488 BAPTA-1 (OGB-1)–Fura-red (FR) imaging to monitor [Ca2+]c changes (Fig. 3) and determine whether asymmetrically produced IP3 generated Ca2+ signals in growth cones. Application of an NGF gradient led to an asymmetric increase in [Ca2+]c in growth cones grown on either L1 or laminin (Fig. 3, A to C, and fig. S2B). As a control, pulsatile pressure ejection of phosphate-buffered saline (PBS) alone, without NGF added, produced no detectable [Ca2+]c change (Fig. 3C). Inhibition of PLC or IICR abolished the NGF-mediated asymmetric increase in [Ca2+]c, indicating that it depended on IP3 production and IICR (Fig. 3, D to F; see also Fig. 2, D to F). These data thus support the notion that NGF attracts growth cones through asymmetric IP3 production and subsequent IICR. Although PI3K was required for NGF-induced attractive turning (Fig. 1B), the increase in [Ca2+]c was unaffected by a treatment with wortmannin (Fig. 3F), indicating that, in growth cones, PI3K is not required for NGF-induced IICR.

Fig. 3

NGF gradients elicit asymmetric increases in [Ca2+]c in growth cones. (A) A DRG growth cone cultured on L1 was loaded with OGB-1 and FR. Relative change in OGB-1–FR emission ratio (RCa/RCa-base, defined as R″) was used as a measure of changes in [Ca2+]c. The grayscale image (left) represents the fluorescence signal of FR before application of NGF. The pseudo color images represent R″ before (middle) and after (right) the application of an NGF gradient (arrow). Scale bar, 5 μm. (B) Time course of changes in R″ in the near (red line) and far (blue line) ROIs of the growth cone shown in (A). (C) The mean amplitude of increases in R″ during the period from 1 to 4 min after application of either NGF or PBS. The amplitude in the near ROI (red bars) was compared with that in the far ROI (blue bars) for growth cones on L1 or laminin. Numbers in parentheses indicate the number of growth cones examined. *P < 0.05, ***P < 0.001; ns, not significant, Bonferroni’s multiple comparison test. (D) Time course of changes in R″ in near (red circles) and far (blue circles) ROIs (n = 11 growth cones). R″ for each growth cone was determined before (open circles) and after (filled circles) bath application of U-73122. This graph shows data on L1. (E) Time course of changes in R″ on the near side relative to those on the far side. Note that the Rnear/Rfar is an index of [Ca2+]c asymmetry. Open and filled circles indicate Rnear/Rfar before and after the treatment with U-73122, respectively. (F) Rnear/Rfar averaged during the period from 1 to 4 min after the start of repetitive NGF ejection. Each growth cone was analyzed before (pre, open bars) and after (post, hatched bars) treatment with vehicle or the indicated drugs. The analyses were done on L1 or laminin. Wort, wortmannin; Xest C, xestospongin C. *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant between pre versus post of each treatment, Bonferroni’s multiple comparison test.

Asymmetric IP3 signals elicit growth cone attraction

We used focal laser-induced photolysis (FLIP) of d-myo-inositol 1,4,5-triphosphate, P4(5)-(1-(2-nitrophenyl)ethyl)ester (caged IP3) to determine whether locally produced IP3 elicited growth cone turning (Fig. 4). Repeated FLIP (every 3 s) on one side of the growth cone resulted in a sustained and asymmetric increase in [Ca2+]c in growth cones grown on L1(Fig. 4, A to C), and elicited growth cone turning toward the side with the higher concentration of IP3 (attraction) (Fig. 4, D, top panels, and E).

Fig. 4

Asymmetric IP3 signals trigger growth cone attraction. (A) A DRG growth cone was loaded with caged IP3 and the Ca2+ indicators OGB-1 and FR. The grayscale image (left) represents FR fluorescence. The yellow spot indicates the site of laser irradiation. The pseudo color images represent RCa/RCa-base before (middle) and after (right) the onset of repetitive FLIP (3-s interval). A representative growth cone on L1 is shown. Scale bar, 5 μm. (B) Time course of changes in RCa/RCa-base in the near (red line) and far (blue line) ROIs of growth cone shown in (A). Arrowhead indicates the onset of repetitive FLIP. (C) The mean amplitude of increases in RCa/RCa-base during repetitive FLIP. The amplitude in the near ROI (red bars) was compared with that in the far ROI (blue bars) in growth cones on L1 or laminin. Numbers in parentheses indicate the number of growth cones examined. **P < 0.01, ***P < 0.001; ns, not significant, Bonferroni’s multiple comparison test. (D) Time lapse DIC images of growth cones on L1 (top) or laminin (bottom). FLIP of caged IP3 was performed every 3 s on one side of the growth cone (yellow spot). Numbers represent minutes after onset of repetitive FLIP. Scale bar, 10 μm. (E) The average angles of FLIP-induced turning on L1 or on laminin in the absence (control) or presence of the indicated drugs. Caged IP3 loading was omitted in “Blank”. Numbers in parentheses indicate the number of growth cones examined. *P < 0.05, **P < 0.01; ns, not significant versus control on L1, Dunnett’s multiple comparison test.

This attraction was blocked by pretreatment with the membrane-permeable tetraacetoxymethyl (AM) ester form of the fast Ca2+ chelator 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) (Fig. 4E). Furthermore, IP3-induced attraction was abolished by bath application of thapsigargin, implicating Ca2+ release from the ER in the response. IP3-induced attraction was also blocked by xestospongin C or 2-APB, specifically implicating IICR in the response.

Because CICR can mediate growth cone attraction (6), we tested whether CICR contributed to IP3-induced attraction. In the presence of 100 μM ryanodine, growth cones still showed attractive turning to asymmetric IP3 signals. These data indicate that asymmetric IICR mediates attractive turning regardless of whether CICR occurs. Wortmannin abolished IP3-mediated attraction (Fig. 4E) but did not inhibit IICR (Fig. 3F), indicating that PI3K activity is required downstream of IICR for attractive growth cone turning.

The amplitude of FLIP-induced increases in [Ca2+]c in growth cones on laminin was significantly lower than that on L1, and repetitive FLIP failed to generate an asymmetry in [Ca2+]c across the growth cone (Fig. 4C). Consistent with this observation, growth cones on laminin showed essentially straight migration in the presence of asymmetrically produced IP3 (Fig. 4, D, bottom panels, and E). Because the cytosolic concentration of cAMP in growth cones on laminin is low (6, 15), we hypothesized that IP3-mediated growth cone attraction depends on cAMP signaling. Consistent with this hypothesis, asymmetric IP3 production elicited attractive turning on laminin when growth cones were treated with the cAMP analog Sp-cAMPS (Fig. 4E). Furthermore, the cAMP antagonist Rp-cAMPS blocked IP3-mediated turning on L1. These data suggest that basal amounts of cAMP—equivalent to that present in growth cones on L1—are required for IP3-mediated attractive turning.

IICR is regulated by cAMP

Accumulating evidence supports the idea that phosphorylation of IP3R by the cAMP effector protein kinase A (PKA) facilitates IICR (20, 21), although conflicting results have been reported (22). We found that the mean amplitude of FLIP-induced increases in [Ca2+]c in growth cones on L1 was significantly reduced by Rp-cAMPS (Fig. 5, B and D) but not by vehicle (Fig. 5, A and C). Rp-cAMPS also decreased [Ca2+]c signals induced by IP3 produced throughout the growth cone (figs. S3 and S4 and Supplementary Materials and Methods). Therefore, we conclude that basal amounts of cAMP are required for IP3 to stimulate IICR to an extent sufficient for attractive turning.

Fig. 5

IICR in growth cones is regulated by cAMP. (A and B) DRG growth cones cultured on L1 were loaded with caged IP3 and the Ca2+ indicators OGB-1 and FR. The grayscale images (left) represent FR fluorescence. The yellow spot indicates the site of laser irradiation. The pseudo color images represent RCa/RCa-base 21 s after the onset of repetitive FLIP (3-s interval). The same growth cone was analyzed before (pre, middle) and after (post, right) 10-min treatment with vehicle (A) or Rp-cAMPS (B). Scale bar, 5 μm. (C and D) The ordinate indicates the amplitude of increases in RCa/RCa-base averaged within the near ROI during the repetitive FLIP, before (pre) and after (post) bath application of vehicle (C) or Rp-cAMPS (D). Numbers in parentheses indicate the number of growth cones examined. *P < 0.05; ns, not significant, paired t test.

Discussion

This report demonstrates the presence of asymmetric IP3 signals during guided neurite outgrowth. We show that axons turn toward an NGF gradient through PLC-dependent asymmetric IICR in the growth cone. We also show that asymmetric production of IP3 mediates attractive turning of growth cones, with basal activity of cAMP and PI3K required upstream and downstream of IICR, respectively. Thus, IP3 signals play a central role in polarizing the growth cone for attractive turning (Fig. 6).

Fig. 6

Proposed signaling pathway mediating NGF-induced attractive growth cone turning. Asymmetric exposure to NGF elicits PLC-dependent asymmetric IP3 production downstream of TrkA. No direct evidence has been provided that PLC is activated asymmetrically. Asymmetric IICR causes increased [Ca2+]c on the side facing NGF, leading to attractive turning. Basal activities of cAMP and PI3K are required for growth cone attraction upstream and downstream of IICR, respectively. Although exposure to an NGF gradient is likely to increase cAMP and to activate PI3K asymmetrically across the growth cone, our results suggest that uniform activities of these signaling molecules are sufficient for asymmetrically produced IP3 to trigger growth cone attraction.

Asymmetric IP3 signals induced by NGF gradients

Here we show that the PLC-IP3 signaling pathway is required for attractive turning toward NGF gradient. The PLC-IP3 pathway has been implicated in neurite extension (23, 24); thus, it is reasonable that its asymmetric activation in growth cones would cause neurite turning. Although we have not examined the spatial profile of PLC activity, the asymmetric IP3 production elicited by NGF is likely to be catalyzed by PLC asymmetrically activated downstream of TrkA.

The predicted diffusion constant of IP3 in cytoplasm is 283 μm2/s (25); thus, in contrast to Ca2+, IP3 has not been viewed as a highly localized messenger but rather as one that acts globally. In our experiments, however, the asymmetry in [IP3]c lasted for at least 3 min after the start of NGF ejection (Fig. 2, B, D, and E, and fig. S2A), suggesting the existence of robust degradation machinery in growth cones to localize IP3 signals. The NGF-induced [IP3]c asymmetry triggers an asymmetric increase in [Ca2+]c through IICR.

It is widely accepted that asymmetric increases in [Ca2+]c mediate growth cone turning to guidance cues (35). Furthermore, unilateral increases in [Ca2+]c produced by FLIP of caged Ca2+ compounds are sufficient to mediate growth cone turning either toward the side with the Ca2+ signal (attraction) or toward the side without the Ca2+ signal (repulsion) (26). Our previous work suggests that directional polarity of growth cone turning depends on the source of the Ca2+ signal: CICR triggers attraction, whereas FLIP of a caged Ca2+ compound that does not induce CICR triggers repulsion (6). Here, we show that asymmetric IICR triggers attraction. This is consistent with the previous findings that an extracellular gradient of a pharmacological activator of PLC or a pharmacological inhibitor of IICR can trigger attractive or repulsive turning, respectively (27, 28). Thus, Ca2+ release from the ER through either IP3Rs or RyRs is sufficient to mediate growth cone attraction. Of these attractive Ca2+ signals, asymmetric CICR mediates growth cone attraction by netrin-1 (14), whereas asymmetric IICR is responsible for growth cone attraction by NGF and likely by other guidance cues (28). These findings suggest that the differential involvement of CICR and IICR depends on the type of guidance cues that elicit growth cone attraction.

During chemotactic guidance, growth cone detect a ~5 to 10% difference in the concentration of guidance molecule across its width (13, 29). Therefore, a shallow gradient of extracellular guidance information is likely to be amplified and translated into a steeper gradient of intracellular signaling. It is thought that such amplification mechanisms involve local augmentation and global inhibition of guidance signals in the growth cone (30). Consistent with this concept, neither [IP3]c nor [Ca2+]c is increased on the far side of growth cones in the presence of an NGF gradient (Figs. 2 and 3 and fig. S2), although the far side should have been exposed to NGF at lower concentration than the near side.

One possible mechanism for such signal amplification could involve receptor redistribution. In rat spinal neurons, an extracellular gradient of γ-aminobutyric acid (GABA) causes redistribution of GABA receptors from the far to the near side of the growth cone, thereby amplifying GABA signals (31). Receptor redistribution may also occur during NGF-induced growth cone attraction. Our previous data suggest that, in chick DRG growth cones, exposure to an NGF gradient causes preferential transport of TrkA-positive membranous vesicles toward the near side (32). Therefore, information about graded exposure to NGF may be amplified at the level of the TrkA receptor and translated into a steeper gradient of activity of the PLC-IP3 pathway followed by asymmetric IICR.

Regulation of NGF-induced growth cone turning by cAMP

cAMP is a key regulator of the directional polarity of growth cone guidance: Increasing cAMP favors attraction, likely by facilitating Ca2+ signaling. For example, Ca2+ signals mediated through activation of L-type voltage-gated Ca2+ channels and RyRs, which are essential for netrin-1–induced growth cone attraction (14), are facilitated by cAMP (6, 33). Here, we showed that IICR is decreased by a cAMP antagonist, indicating that IICR is also facilitated by cAMP.

The cAMP-dependent facilitation of IICR may be mediated by PKA-dependent phosphorylation of IP3R, which increases the sensitivity of IP3R to activation by IP3 (20, 21), whereas protein phosphatase 1 (PP1)–dependent dephosphorylation inactivates IP3R (21). PKA also inhibits PP1 by activating inhibitor-1 (34). Thus, PKA could facilitate IICR by both phosphorylating IP3R and preventing its PP1-mediated dephosphorylation. cAMP may also directly sensitize IP3R independent of PKA, although the precise mechanism remains unclear (35), raising another possible mechanism for cAMP-dependent IICR potentiation. Consistent with the notion of cAMP-dependent IICR regulation, we found that growth cones do not respond to asymmetric IP3 production by FLIP when the concentration of cAMP is low on a laminin substrate or the cAMP signaling pathway is inhibited by Rp-cAMPS.

The effect of Rp-cAMPS or laminin on the turning response of growth cones to directional application of NGF cannot be simply explained by cAMP-dependent regulation of IICR. During migration on laminin, NGF attracts the growth cone, probably because NGF increases cAMP concentration (36). When cAMP signaling is inhibited by Rp-cAMPS, growth cones exhibits repulsive responses to NGF (−11.24° ± 3.01°, n = 22 growth cones), consistent with a previous report (11). This is likely because NGF activates signaling pathways other than those involving IP3, which cause growth cone repulsion in the absence of IICR. One possible candidate for the repulsive signaling components is protein kinase C (PKC), which is activated downstream of PLC (27, 37). Extracellular gradients of PKC activator or inhibitor trigger growth cone repulsion or attraction, respectively, suggesting that PKC may act as a repulsive signal (27).

Mechanisms of NGF-induced attraction downstream of IICR

Asymmetric Ca2+ signals trigger growth cone turning; however, the precise mechanism remains unclear. CICR triggers growth cone attraction via microtubule-dependent transport of vesicles toward the side with elevated [Ca2+]c followed by asymmetric exocytosis (32). In case of NGF-mediated attraction, a similar mechanism may be involved downstream of IICR: (i) NGF gradients induce asymmetric vesicle transport toward the near side of the growth cone, and (ii) membrane exocytosis is required for NGF-mediated attraction (32). Although it remains unclear whether PI3K per se is situated in the cascade downstream of IICR, PI3K-dependent signaling is required for IICR to elicit growth cone attraction. Inhibition of PI3K may prevent microtubule polymerization and membrane expansion in growth cones (38, 39), raising the possibility that PI3K plays a role in IICR-triggered vesicle transport, exocytosis, or both during NGF-induced attraction.

In summary, our study demonstrates the critical role of asymmetric IP3 signals in growth cone navigation and will contribute to better understanding of the molecular mechanisms that control cell polarity and directed migration.

Materials and Methods

Cell culture

DRG neurons from embryonic day 9 chicks were dissociated and plated on glass dishes coated with laminin (~50 ng per 1.54 cm2; Invitrogen) or L1-Fc chimeric proteins comprising the extracellular domain of L1 and the Fc region of human immunoglobulin G as described (40, 41). Cultures were maintained in Leibovitz’s L-15 medium (Invitrogen) supplemented with N-2 (Invitrogen), bovine serum albumin (750 μg/ml; Invitrogen), and NGF (20 ng/ml; Promega) in a humidified atmosphere of 100% air at 37°C.

Growth cone turning assays

Assays of growth cone turning induced by microscopic gradients of NGF (50 μg/ml in a micropipette) were performed as previously described (13, 14). Bath-applied NGF for maintaining cultures was withdrawn during the turning assays. Growth cones were exposed to the NGF gradient for 40 min on laminin or 60 min on L1. Differential interference contrast (DIC) images were acquired every 10 min with a 20× objective [UPlanSApo, numerical aperture (NA) 0.75, Olympus] on an inverted microscope (IX81, Olympus) equipped with a charge-coupled device (CCD) camera (CoolSnap HQ with binning set at 1 × 1, Roper Scientific).

Assays of growth cone turning induced by FLIP were performed as described previously (6). Caged IP3 (Invitrogen) was introduced into DRG neurons by trituration loading (42, 43). To visualize the loaded neurons, dextran (molecular weight 10,000) conjugated with Alexa Fluor (Invitrogen) was loaded simultaneously. In brief, DRGs were triturated using a p20 micropipetman (~150 strokes) in 20 μl of Leibovitz’s L-15 medium containing 200 μM caged IP3 and 100 μM Alexa Fluor–conjugated dextran.

In some experiments, the following reagents were applied to culture media at least 30 min before the application of NGF gradients or FLIP: 60 μM 2-APB, 100 nM K-252a, 20 μM Rp-cAMPS, 20 μM Sp-cAMPS, 200 nM thapsigargin, 30 nM U-73122, 60 nM wortmannin, 2 μM xestospongin C (Calbiochem), or 100 μM ryanodine (Latoxan). Five micromolar BAPTA-AM (Invitrogen) was loaded as described (6).

IP3 imaging

DRG neurons were transfected with IRIS-1 or IRIS-1–Dmut (19) using Nucleofector II (Lonza) according to the manufacturer’s protocols. Bath-applied NGF for maintaining cultures was withdrawn, and IP3 imaging was performed on an inverted microscope (IX81). Here, NGF gradients were applied with a micropipette containing NGF (100 μg/ml). With the use of a 20× objective (UPlanSApo), micropipettes were set 50 μm from the growth cone at a 45° angle with respect to the original direction of axon elongation. Then the micropipette was lifted out of the culture medium, the objective lens was switched from 20× to 100× (UPlanSApo, NA 1.40, Olympus), and the micropipette was lowered to the original position. Although a reagent solution in the micropipette tip may become diluted due to the backpressure, a graded distribution of the reagent is usually formed across the growth cone within 1 min after the start of pulsatile positive pressure in the micropipette (fig. S5). Thereafter, the ratio of the reagent concentrations between both sides of the growth cone remained nearly constant (fig. S5).

IRIS-1 or IRIS-1–Dmut was excited with 434/17-nm light, and fluorescence images of ECFP and Venus were simultaneously acquired with a CCD camera (ImagEM with binning set at 4 × 4, Hamamatsu Photonics) after the dual-color image was split with an emission splitter (U-SIP, Olympus) that included a 505-nm dichroic mirror, a 475/28-nm emission filter for ECFP, and a 535/22-nm emission filter for Venus. Images were acquired every 3 s. The ECFP/Venus emission ratio (RFRET) was used as a measure of [IP3]c. There is a linear correlation between RFRET and [IP3] when 0.1 ≲ [IP3] ≲ 10 μM (19), indicating that RFRET is a reliable readout of physiological [IP3]c in a cell (44). In addition, IRIS-1 is completely insensitive to [Ca2+] changes (19). The time course of changes in [IP3]c were expressed as RFRET/RFRET-base, where RFRET-base is the mean of 20 consecutive RFRET values before the start of repetitive NGF ejection (−60 to −3 s). Three consecutive frames captured every 9 s were averaged and plotted in the graphs (Fig. 2, B, D, and E, and fig. S2A) to improve the signal-to-noise ratio. The amplitude of NGF-induced [IP3]c increases was defined as the mean of RFRET/RFRET-base during the period from 1 to 3 min after the start of repetitive NGF ejection. [IP3]c asymmetry was assessed by calculating the near-to-far ratio of RFRET/RFRET-base as follows: A rectangle was positioned to cover the area of the growth cone perpendicular to its direction of migration, and this rectangle was equally divided into three sections. The sections closer and further from the NGF source were then defined as the “near” and the “far” ROIs, respectively. The near-to-far ratio was obtained by dividing RFRET/RFRET-base in the near ROI by that in the far ROI. Before and after a 10-min treatment with bath-applied U-73122 or vehicle, the NGF-induced [IP3]c asymmetry was assessed and expressed as the near-to-far ratio of RFRET/RFRET-base averaged during the period from 1 to 3 min after the start of repetitive NGF ejection.

Ca2+ imaging

In experiments designed to monitor [Ca2+]c changes induced by an NGF gradient, DRG neurons were loaded with two Ca2+ indicators, OGB-1–AM (2 μM, Invitrogen) and FR-AM (2.5 μM, Invitrogen). Bath-applied NGF was withdrawn before Ca2+ imaging. For simultaneous OGB-1–FR imaging, both indicators were excited with 492/18-nm light, and images were acquired every 3 s with a CCD camera (EM-CCD with binning set at 4 × 4; Hamamatsu Photonics) after the dual-color image was split with an emission splitter (Dual-View, Roper Scientific) that included a 565-nm dichroic mirror, a 530/35-nm emission filter for OGB-1, and a 610-nm emission filter for FR. OGB-1/FR emission ratio (RCa) was used as a measure of [Ca2+]c. The time course of changes in [Ca2+]c were expressed as RCa/RCa-base, where RCa-base is the mean of 20 consecutive RCa values before the start of repetitive NGF ejection (−60 to −3 s). Three consecutive frames captured every 9 s were averaged and plotted in the graphs (Fig. 3, B, D, and E, and fig. S2B) to improve signal-to-noise ratio. The amplitude of NGF-induced [Ca2+]c increases was defined as the mean of RCa/RCa-base during the period from 1 to 4 min after the start of repetitive NGF ejection. [Ca2+]c asymmetry was assessed by calculating the near-to-far ratio of RCa/RCa-base as described in the previous section. The effect of pharmacological agents on the NGF-induced [Ca2+]c asymmetry was evaluated by comparing the near-to-far ratio during the period from 1 to 4 min after the start of repetitive NGF ejection, before and after 10-min treatment of the growth cone with the selected agent.

In experiments designed to monitor [Ca2+]c changes induced by repetitive FLIP of caged IP3, DRG neurons were loaded with caged IP3 and two Ca2+ indicators, OGB-1–AM (2 μM) and FR-AM (1.5 μM). For simultaneous OGB-1–FR imaging, both indicators were excited with 460- to 495-nm light, and images were acquired every 3 s with a CCD camera (ORCA-ER with binning set at 8 × 8, Hamamatsu Photonics) after the dual-color image was split with an emission splitter (W-View, Hamamatsu Photonics) composed of a 590-nm dichroic mirror, a 535/45-nm emission filter for OGB-1, and a 610-nm emission filter for FR. The time course of changes in [Ca2+]c were expressed as RCa/RCa-base, where RCa-base is the mean of the 10 consecutive RCa values before the onset of repetitive FLIP (−30 to −3 s). The amplitude of FLIP-induced [Ca2+]c elevations was defined as the mean of RCa/RCa-base during the period from 0 to 57 s after the onset of repetitive FLIP (3-s interval). [Ca2+]c asymmetry was evaluated by comparing the mean amplitude of RCa/RCa-base in the near versus far ROIs. In some experiments, the amplitude of RCa/RCa-base in the near ROI was compared in a growth cone before versus after 10-min treatment with Rp-cAMPS.

Statistics

Data were expressed as the mean ± SEM. Statistical analyses were performed with GraphPad Prism version 4.01 (GraphPad Software) or Microsoft Excel 2007 for Kolmogorov-Smirnov tests. P < 0.05 was judged statistically significant.

Acknowledgments

We are grateful to T. Tojima and A. T. Guy for their valuable comments and discussions on this manuscript. We also thank RIKEN Brain Science Institute’s Research Resources Center for providing experimental instruments. This work was supported by RIKEN and a Grant-in-Aid for Scientific Research (21500361, H.K.) from the Japan Society for the Promotion of Science.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/2/79/ra34/DC1

Materials and Methods

Fig. S1. IP3R distribution in chick DRG growth cones.

Fig. S2. NGF gradients elicit asymmetric IP3 signals in growth cones grown on laminin.

Fig. S3. Global Ca2+ signals in the growth cone triggered by photolysis of caged IP3.

Fig. S4. Global Ca2+ signals triggered by PLC activation in the growth cone.

Fig. S5. Time course of changes in the spatial distribution of reagents ejected from a micropipette.

References

References and Notes

View Abstract

Navigate This Article