DCC functions as an accelerator of thalamocortical axonal growth downstream of spontaneous thalamic activity

Abstract

Controlling the axon growth rate is fundamental when establishing brain connections. Using the thalamocortical system as a model, we previously showed that spontaneous calcium activity influences the growth rate of thalamocortical axons by regulating the transcription of Robo1 through an NF-κB-binding site in its promoter. Robo1 acts as a brake on the growth of thalamocortical axons in vivo. Here, we have identified the Netrin-1 receptor DCC as an accelerator for thalamic axon growth. Dcc transcription is regulated by spontaneous calcium activity in thalamocortical neurons and activating DCC signaling restores normal axon growth in electrically silenced neurons. Moreover, we identified an AP-1-binding site in the Dcc promoter that is crucial for the activity-dependent regulation of this gene. In summary, we have identified the Dcc gene as a novel downstream target of spontaneous calcium activity involved in axon growth. Together with our previous data, we demonstrate a mechanism to control axon growth that relies on the activity-dependent regulation of two functionally opposed receptors, Robo1 and DCC. These two proteins establish a tight and efficient means to regulate activity-guided axon growth in order to correctly establish neuronal connections during development.

Introduction

Axon growth is a key developmental process that is necessary for axons to establish appropriate connectivity in the central nervous system (CNS). Developing neurons extend their axons following stereotype pathways in which distinct axon guidance molecules and intermediate targets play a fundamental role ^1,2,3. Thus, the course of axon growth is a tightly controlled process. Using the thalamocortical system as a model, we previously identified a mechanism that regulates the growth rate of developing axons ⁴. We demonstrated that spontaneous Ca²⁺ activity modulates axonal growth through the transcriptional upregulation of the guidance receptor Robo1, which appears to function as a brake for thalamocortical axon (TCA) growth ⁴. Therefore, modifications to Robo1 transcription that are triggered by spontaneous activity mediate the developmental decrease in axon growth as TCAs approach the cortex. However, tight regulation of axon growth rate probably relies on different factors controlling speed in both senses, instead of just one specific factor acting as a brake. Thus, other activity-dependent factors might contribute to this process.

Several studies have demonstrated that spontaneous electrical activity plays a fundamental role in modulating early embryonic developmental events such as neurotransmitter specification and axon guidance ^5,6. Moreover, spontaneous activity modulates the growth rate of developing thalamocortical and sensory axons of the spinal cord ^4,7,8. In fact, specific parameters of this activity, such as the frequency of Ca²⁺ spikes, drive precise changes in the expression of downstream target genes that are critical for the control of these developmental processes, including the axon growth program ^4,9. Thus, spontaneous activity appears to emerge as a general mechanism to influence axonal extension of developing neurons through the regulation of gene expression.

The chemotropic guidance cue Netrin-1 mediates the attraction of migrating axons during CNS development through the receptor deleted in colorectal cancer (DCC) ^10,11,12. Activation of DCC by Netrin-1 is required for the outgrowth of cortical and commissural axons ^13,14,15,16. Indeed, the significance of this pathway during brain development was highlighted by the absence of spinal and cerebral commissures when Netrin-1 or DCC expression is abolished ^17,18. Netrin-1 has also been shown to promote TCA outgrowth in vitro, and in Netrin-1^−/− mice, fewer TCAs reach the cerebral cortex ¹⁹. Thus, Netrin-1 signaling through the DCC seems to be a good candidate to participate in the activity-dependent regulation of axon growth during development.

Here, we determined whether DCC might act together with Robo1 to modulate the growth rate of TCAs downstream of spontaneous activity. We found that spontaneous Ca²⁺ activity regulates DCC expression in an opposite fashion to that of Robo1. When thalamic neurons are silenced, DCC expression is significantly dampened at both the mRNA and protein levels. Through in vitro and in vivo loss- and gain-of-function experiments, we showed that DCC functions as an accelerator for TCA growth. Moreover, we showed that Ca²⁺ spikes modulate axonal growth through the transcriptional upregulation of Dcc acting via an AP-1 transcription factor-binding site in its promoter. In sum, our data identify a mechanism by which spontaneous activity can efficiently control axon growth through the transcriptional regulation of two functionally opposite genes, Robo1 and DCC.

Results

Silencing spontaneous activity decreases the amount of Dcc in growth cones

Developing axons might be guided along specific pathways to establish correct connections by controlling their growth rate. We recently demonstrated that by regulating the Robo1 gene, spontaneous Ca²⁺ activity is a key modulator of the growth rate of developing TCAs ⁴. Robo1 acts as a brake for axonal growth, and its expression is upregulated in silenced thalamocortical neurons ⁴. However, the efficient regulation of axon growth speed might not only rely on the function of a brake but also on that of an accelerator. Thus, we investigated whether other genes regulated by spontaneous thalamic activity also modulate the growth rate of TCAs.

To identify additional candidate genes involved in axon growth and that might be regulated by spontaneous activity, we overexpressed the inwardly rectifying potassium channel Kir2.1 (Kir) in thalamic neurons which will provoke hyperpolarization and dampening of neural activity ^{4,20,21,22,23,24}. Dissociated thalamic cells were transfected with plasmids expressing Kir2.1 (Kir) and/or Gfp, and they were analyzed after 72 h in culture (Fig1A). As expected, spontaneous somatic Ca²⁺ transients were silenced in Kir-expressing thalamic cells (Supplementary Fig S1A), and thus, we assessed whether the expression of putative downstream target genes might have changed in these thalamic silenced neurons. We identified Dcc as being one of the most strongly downregulated genes after silencing thalamic spontaneous activity. Silencing thalamic neuronal activity decreased Dcc expression 1.95-fold in Kir-expressing thalamic neurons (Gfp, 1.05 ± 0.28, n = 7; Kir, 0.53 ± 0.33, n = 14: Fig1B), as measured by quantitative RT–PCR (qPCR). Moreover, DCC protein levels diminished accordingly in these neurons (Fig1C and Supplementary Fig S1B). On the contrary, when we elevated the extracellular KCl concentration in the cultures to increase neuronal activity, the axons of thalamic neurons grew longer and accumulated more DCC protein (Supplementary Fig S1C–D). Moreover, and as embryonic development proceeded, the levels of expression of Dcc transcripts gradually diminished in thalamic neurons in vivo from 1.46 ± 0.34 at E12.5 (n = 5) to 0.57 ± 0.19 at E18.5 (n = 4) (Fig1D). This decrease was also observed at the level of DCC protein, which declined from 100 ± 9.6 at E13.5 (n = 9) to 24.5 ± 4.5 at E15.5 (n = 8) (Fig1E). Therefore, these results suggest the existence of a mechanistic link between the downregulation of DCC and the decrease in the growth rate previously described in these thalamic neurons ⁴.

Figure 1.

Silencing thalamic activity downregulates DCC transcription

1. Embryonic dissociated thalamic neurons transfected with Gfp (green) and immunostained for Tuj1 (red). Nuclear counterstaining with DAPI is shown in blue. Scale bar, 100 μm. Arrowheads indicate examples of Gfp-transfected cells labeled by Tuj1.
2. Real-time PCR quantification of the transcripts for Kir2.1 and Dcc in Gfp- and Kir2.1-transfected cells: ***P < 0.001, Mann–Whitney U-test. The data are presented as the means ± s.e.m.
3. Immunocytochemistry for DCC in control and Kir2.1-expressing axons. Less DCC expression was found in growth cones transfected with Kir2.1 compared to controls (arrowheads). Scale bar, 10 μm.
4. Semi-quantitative PCR for Dcc and gapdh, normalizing Dcc expression to that of gapdh: ***P < 0.001, one-way ANOVA with Tukey's post hoc analysis. ID, integrated density. The data are presented as the means ± s.e.m.
5. Immunocytochemistry for DCC in E13.5 (GFP-positive; white arrowhead) and E15.5 (GFP-negative; blue arrowhead) isolated thalamic growth cones. Thalamic explants from E13.5 GFP mouse were cocultured together with explants from wild-type E15.5 embryos on polylysine/laminin coverslips. Scale bar, 10 μm. Quantification of the data shown: ***P < 0.001, two-tailed Student's t-test. ID, integrated density. The data are presented as the means ± s.e.m.

DCC functions as an intrinsic accelerator for TCA

The aforementioned results prompted us to hypothesize that DCC plays a role in modulating TCA growth. To test this, we measured axonal outgrowth in dissociated thalamic neurons from wild-type and Dcc-deficient mice (Fig2). In the absence of Dcc, thalamic axons grew significantly less compared to controls (Dcc^+/+, 100 ± 1.16%, n = 4 brains; Dcc^−/−: 46.80 ± 0.69%, n = 6 brains: Fig2A,B), and thus, we predicted that increasing DCC levels in thalamic neurons would accelerate thalamic axon growth in vitro, overriding the effects of Kir-induced silencing of thalamic spontaneous activity. We augmented DCC levels by transfecting thalamic neurons with a full-length Dcc (FL-Dcc) at E13.5, and we quantified the length of thalamic axons. Overexpression of Dcc remarkably enhanced the extension of thalamic axons (Gfp, 100 ± 4.84%, n = 7 independent cultures; FL-Dcc, 148.5 ± 4.34%, n = 7 independent cultures: Fig1C,D), further indicating that DCC acts as an intrinsic accelerator for thalamic axon progression.

Figure 2.

Alterations to the axon growth rate by manipulating DCC expression

1. Tuj1 immunostaining revealed that thalamic axon growth decreased in the absence of DCC (arrowheads). Scale bars: 100 μm lower power panels; and 50 μm, higher power panels.
2. Quantification of the data shown in A. ***P < 0.001, two-tailed Student's t-test. The data are presented as the means ± s.e.m.
3. Tuj1 staining revealed an increase in thalamic axon growth following Dcc overexpression with a full-length Dcc (FL-Dcc) construct (arrowheads). Scale bars: 100 μm, lower power panels; and 50 μm, higher power panels.
4. Quantification of the data in C. When DCC levels were elevated, axon growth increased significantly when compared to control conditions. ***P < 0.001, two-tailed Student's t-test. The data are presented as the means ± s.e.m.

Netrin-1, the best known ligand for DCC proteins, is expressed by thalamic neurons, and it is distributed along the entire TCA connectivity pathway in vivo ^19,25,26. As expected, we detected Netrin-1 mRNA and protein expression in thalamic dissociated cultures at E13.5 (Supplementary Fig S2A). Thus, we first tested whether Netrin-1 fulfills a cell autonomous role in thalamocortical axon growth by culturing dissociated thalamocortical neurons from wild-type or Netrin-1^−/− mice. Like axons lacking DCC, axons without Netrin-1 are significantly shorter than their wild-type controls (Supplementary Fig S2C,D). As Netrin-1 has been described as a survival factor in several brain structures, such as the spinal cord or the developing brainstem ^27,28, we assessed whether the effect on axon growth could somehow be related to increased thalamic neuronal death in vitro. Using release of lactate dehydrogenase activity (LDH) to detect cell death, we found no significant differences in Netrin-1^+/+ and Netrin-1^−/− thalamic cultures (Supplementary Fig S2B). Then, we evaluated the extrinsic role of Netrin-1 in modulating thalamocortical axonal growth rate by placing thalamic explants from wild-type embryos onto telencephalic slices from Netrin-1^−/− mice or their wild-type littermates (Fig3A). TCA extended much less when grown on Netrin-1^−/− slices than on the controls (Netrin-1^+/+, 1334 ± 126.9, n = 13 explants; Netrin-1^+/−, 1630 ± 124.4, n = 20 explants; Netrin-1^−/−, 665.7 ± 97.23, n = 15 explants: Fig3B,C). Altogether these results show that Netrin-1 functions to accelerate thalamocortical axonal extension in the ventral telencephalon.

Figure 3.

DCC acts as an accelerator for thalamocortical axon growth in vivo

1. Experimental paradigm used to analyze the involvement of Netrin-1 in thalamic growth.
2. Wild-type thalamic axons grow significantly less in Netrin-1^−/− slices than in control slices (arrowheads). Crystals of DiI (red)-labeled thalamocortical axons, and DAPI nuclear counterstaining is in blue. Scale bar, 300 μm.
3. Quantification of the data shown in B. **P < 0.01 and ***P < 0.001, Kruskal–Wallis test with Dunn's post hoc analysis. The data are presented as the means ± s.e.m.
4. Progression of TCA labeled by DiI in the E14.5 thalamus of Dcc^+/+ and Dcc^−/− brains. The bars indicate the position of the pallial–subpallial boundary for reference, and the arrowheads indicate the delayed position of TCA in the mutants. Scale bar, 200 μm.
5. Schematic representation of the experimental procedure used to test the effect of increasing the Dcc levels in vivo. A FL-Dcc plasmid was co-electroporated in utero with GFP into the thalamus at E12.5, and the brains were analyzed at E15.5, at the peak of TCA cortical extension.
6. GFP immunoreactive axons in coronal sections at intermediate cortical levels, with DAPI nuclear counterstaining shown in blue. Overexpression of Dcc led to a significant acceleration in the cortical extension of TCA compared to controls (arrowheads). Scale bar, 300 μm.
7. Quantification of the data shown in F and data not shown. **P < 0.01, two-tailed Student's t-test. The data are presented as the means ± s.e.m.

Regulation of TCA extension in vivo via Netrin-1/DCC signaling

Our results in vitro and ex vivo showed that Netrin-1 and DCC play a role in the modulation of thalamic axonal extension. However, in order to determine whether Netrin-1/DCC signaling may also influence TCA extension in vivo, we performed dye-tracing studies in Netrin-1^−/− and Dcc^−/− mutant mice at E14.5. The lack of either Netrin-1 or Dcc diminished TCA extension in the neocortex when compared with wild-type mice (Netrin-1^+/+, n = 4 brains; Netrin-1^−/−, n = 5 brains; Dcc^+/+, n = 5 brains; Dcc^−/−, n = 8 brains; Fig3D). Thus, we expected that increasing DCC levels in thalamic neurons would promote TCA growth in vivo. We increased DCC levels through in utero electroporation of FL-Dcc into thalamic neurons at E12.5, and we quantified the extension of TCA at E15.5 (Fig3E). The overexpression of DCC notably increased TCA extension at both the rostral (Gfp, 1967 ± 143.8, n = 4 brains; FL-Dcc, 3631 ± 315.6, n = 8 brains) and intermediate (Gfp, 1653 ± 202.9, n = 4 brains; FL-Dcc, 3561 ± 329.7, n = 8 brains) telencephalic levels (Fig3F,G). In sum, these results show that Dcc gain-of-function has the opposite effect to silencing thalamic spontaneous activity, supporting the role of DCC as an accelerator of TCA extension.

Activity-dependent regulation of DCC in controlling TCA growth

The results obtained strongly suggest that DCC might act downstream of spontaneous activity to control axon growth in thalamic neurons. Hence, we hypothesized that augmenting DCC levels in silenced thalamic neurons should rescue the axon growth defect observed after Kir2.1 silencing. To test this, we quantified the extension of thalamic axons transfected with Gfp, Kir, FL-Dcc alone, or co-transfected with Kir and FL-Dcc (Fig4A,B). Whereas silencing spontaneous activity significantly diminished thalamic axon growth (Gfp, 100 ± 4.8, n = 7 independent cultures; Kir, 49 ± 2.3, n = 7 independent cultures: Fig4A,B), co-electroporation of Kir with FL-Dcc completely rescued the axon growth defect to the levels of FL-Dcc alone (FL-Dcc + Kir, 135 ± 6.4, n = 7 independent cultures; FL-Dcc, 148 ± 4.3, n = 7 independent cultures: Fig4A,B). These results demonstrate that modifying DCC levels in silenced neurons can rescue the abnormal axonal extension of Kir-positive neurons in vitro (Fig4C). To determine to what extent spontaneous thalamic activity influences TCA growth rate via DCC in vivo, we measured the TCA extension in the cortex following co-electroporation with Kir and FL-Dcc at E12.5 (Fig4D). Remarkably, co-electroporation of both plasmids together rescued the delayed axon progression found after Kir electroporation alone. TCA extended significantly further in the cortex, both at rostral (Kir, 1237 ± 75.6, n = 8 brains; Kir + FL-Dcc, 3743 ± 208.3, n = 7 brains; Fig4E,F) and at intermediate levels (Kir, 1243 ± 69.7, n = 8 brains; Kir + FL-Dcc, 3930 ± 172.5, n = 7 brains; Fig4E,F) compared to axons expressing Kir alone. These results indicate that the delayed axon growth triggered by the silencing of spontaneous thalamic activity can be rescued in vivo by activating Dcc signaling.

Figure 4.

DCC controls axon extension in vivo downstream of spontaneous activity

1. GFP (green), Tuj1 (red), and DAPI (blue) staining in control (Gfp), Kir-transfected, and Kir+FL-Dcc-co-transfected thalamic cells. Increasing the levels of DCC in silenced thalamic neurons rescued the axon growth defect produced by Kir overexpression. The arrowheads indicate the length of the axons. Scale bar, 50 μm.
2. Quantification of the data shown in A and data not shown. **P < 0.01 and ***P < 0.001; Kruskal–Wallis test with Dunn's post hoc analysis. The data are presented as the means ± s.e.m.
3. Scheme representing the effect on axon growth after Kir transfection alone or co-transfection with Kir and Kir+FL-Dcc. Red lines represent the status of the spontaneous activity in the cell: no activity, straight line; or spontaneous activity, curved line.
4. Schematic representation of the experimental procedure used to test the effect of increasing the levels of Dcc in Kir-silenced TCA in vivo. A Kir plasmid was co-electroporated into the thalamus in utero at E12.5 with Gfp or with Gfp plus FL-Dcc, and the brains were analyzed at E15.5.
5. GFP immunoreactive axons in coronal sections at intermediate cortical levels, with DAPI nuclear counterstaining in blue. Overexpression of Dcc in Kir-silenced axons rescued the cortical extension of TCA silenced axons (arrowheads). Scale bar, 300 μm.
6. Quantification of the data shown in E and data not shown. *P < 0.05, ***P < 0.001, one-way ANOVA with Tukey's post hoc analysis. The data are presented as the means ± s.e.m.
7. Expression of DCC (red) and Robo1 (gray) protein levels in E13.5 thalamic growth cones electroporated with Gfp or Kir2.1 plasmids. Scale bar, 5 μm.
8. Quantification of the data shown in G: **P < 0.01, ***P < 0.001; two-tailed Student's t-test. The data are presented as the means ± s.e.m.

Spontaneous thalamic activity regulates Robo1 and DCC levels in an opposite manner ⁴ (Fig1). Silencing spontaneous activity increases Robo1 mRNA and protein, while it decreases DCC protein and mRNA. Thus, we tested whether this dynamic regulation occurs in thalamic neurons after manipulating their rate of axonal extension in vitro. As such, we silenced thalamic neurons with Kir to diminish their axon growth rate and measured the levels of DCC and Robo1 proteins in their growth cones. Growth cones transfected with Kir had higher Robo1 and lower DCC levels than Gfp-transfected axons (Gfp, 100 ± 16.5 Dcc and 100 ± 9.3 Robo1, n = 9; Kir, 40.7 ± 7.3 Dcc and 234.4 ± 9.3 Robo1, n = 20: Fig4G,H). These results demonstrate that spontaneous activity influences axon growth rates by antagonistically modulating the expression of Robo1 and DCC in TCA.

The activity-dependent DCC transcriptional regulation occurs through an AP-1-binding site

Calcium spontaneous activity could modulate DCC expression levels by regulating its transcription. In order to determine this, we looked into the regulatory sequences of the Dcc gene and searched for conserved motifs associated with Ca²⁺-sensitive transcription factors. Through sequence alignment, we revealed an evolutionarily conserved 1491-bp region upstream of the first exon of Dcc (Supplementary Fig S3), which we cloned upstream of the firefly luciferase gene. Silencing spontaneous activity with Kir transfection into dissociated thalamic cells at E13.5 (Fig5A) produced a 43% decrease in Dcc reporter activity in these cells (Gfp, 100 ± 14.3; Kir, 56.3 ± 9.4; n = 7 experiments; Fig5A). Thus, Ca²⁺ activity-responsive elements may exist in the Dcc promoter. We only found one nuclear factor-kappa B (NF-κB) transcription factor-binding site and one activator protein 1 (AP-1) site, which sequence and position were conserved in the genomes analyzed (Fig5B and Supplementary Fig S3). Both of these transcription factors have been previously implicated in activity-dependent transcriptional regulation ^29,30. The mutation of the NF-κB site in the Dcc reporter induced a significant reduction in the basal transcription of Dcc (wild-type, 100 ± 14.3; NF-κB, 65.5 ± 6.2, n = 7 experiments: Fig5C). This reduction was not observed when the AP-1 site was mutated (AP-1, 77.2 ± 6.3, n = 7 experiments; Fig5C). We next determined whether any of these sites might be involved in the activity-dependent regulation of Dcc. Interestingly, mutation of the NF-κB site did not affect the decrease of Dcc transcription by Kir (wild-type + Kir, 56.3 ± 9.4%; mutant NF-κB + Kir, 66.3 ± 5.1%, n = 7 experiments; Fig5D). By contrast, mutation of the AP-1 site significantly reduced the effect of Kir overexpression on Dcc transcription (mutant AP-1 site + Kir, 87.2 ± 10.6%, n = 7 experiments; Fig5D). Thus, AP-1 appears to fulfill an important role in the activity-dependent transcription of Dcc.

Figure 5.

DCC expression is regulated at the transcriptional level through an AP-1-binding site

1. Silencing spontaneous activity induced a 30.38% decrease in expression of the luciferase reporter. *P < 0.05, two-tailed Student's t-test. The data are presented as the means ± s.e.m
2. Schematic representation of the wild-type and mutated forms of the 1491-bp cloned region of the Dcc promoter.
3. Significant reduction in the basal transcription of Dcc when NF-κB but not the AP-1 site was mutated (n ≥ 3 replicates). *P < 0.05, Mann–Whitney U-test. The data are presented as the means ± s.e.m.
4. Luciferase assays to evaluate the contribution of AP-1 and NF-κB sites to the reduction in Dcc by Kir. Mutating only the AP-1 site significantly reduced the dampening of Dcc transcription after the silencing of spontaneous activity (n ≥ 3 replicates). *P < 0.05, Mann–Whitney U-test. The data are presented as the means ± s.e.m.

Discussion

The mechanisms that control axon growth extension are steadily being elucidated. We recently identified spontaneous activity as an intrinsic modulator of the rate of axon extension in the thalamocortical system ⁴. An embryonic developmental switch in TCA growth rate from fast to slow occurs concomitantly with a decrease in the frequency of spontaneous Ca²⁺ activity in thalamic neurons. This reduction in Ca²⁺ activity is sufficient to slow down the cortical progression of TCA by regulating the transcription of the axon guidance receptor Robo1, which acts as a brake for thalamocortical axon extension ⁴. Here, we found another key element that controls axon growth regulated by spontaneous activity. DCC, a receptor for Netrin-1, acts downstream of spontaneous activity to modulate the axon's growth rate but in an opposite manner to that of Robo1, accelerating TCA growth. Silenced neurons augment their expression of Robo1 and decrease their DCC levels, leading to a reduction in axon extension. Moreover, the transcriptional regulation of DCC occurs through an AP-1-binding site in its promoter. In summary, we describe a mechanism to modulate axon growth whereby spontaneous calcium activity in developing neurons controls Robo1 (brake) and DCC (accelerator) transcription, exerting opposite functional effects on this process (Fig6).

Figure 6.

Scheme representing the mechanism to regulate axon growth in developing TCA

Spontaneous Ca²⁺ activity is regulated during development in thalamic neurons. The frequency of this activity drops from high to low when TCA slows down their progression through the cortex ⁴. This switch encodes activity-dependent changes in expression of the downstream target genes, DCC and Robo1, which exert antagonistic activities on axon growth. The activity-dependent regulation of these two genes occurs through specific transcription factor-binding sites in their respective promoters.

TCAs reduce their rate of axon extension when approaching cortical targets in order to more thoroughly explore potential target areas ⁴. In this study, we found that altering DCC expression in thalamic neurons changes the rate of axon growth, revealing a novel role for DCC receptors as an accelerator of TCA extension. Our results show a significant decrease in thalamic axon growth in the Netrin-1^−/− and Dcc^−/− mice, both in vitro and in vivo. The reduced growth of wild-type axons in Netrin-1^−/− slices ex vivo further supports a Netrin-1-dependent function of DCC in controlling TCA growth. Indeed, previous studies have shown similar defects regarding axon outgrowth, and Netrin-1 has been shown to promote the outgrowth of commissural axons and attract their growth cones toward the ventral midline in the spinal cord, hindbrain, and telencephalon ^11,18,31. Moreover, Netrin-1 is expressed in the ventral telencephalon (vTel), and it also attracts and promotes axonal outgrowth of early cortical axons toward subcortical structures ², as well as stimulating thalamic axon outgrowth ¹⁹. In line with our results, the pathway of thalamic axons through the vTel is disorganized in the Netrin-1-deficient mouse and fewer axons reach the cortex ¹⁹. However, as these axon guidance defects have been analyzed in a full Netrin-1-deficient mouse, it is possible that these errors might be indirect and due to the abnormal development of ventral telencephalic structures that are important for TCA development. Nevertheless, subsequent analyses of the Netrin-1 mutant brains revealed topographical defects that included the invasion of more caudal territories by thalamic axons originating from the rostral thalamus ²⁶. In fact, Netrin-1 seems to exert a bifunctional effect on TCAs, attracting rostral thalamic axons and acting as a chemorepulsive cue for caudal thalamic axons ²⁶. Accordingly, DCC is expressed in a gradient in the thalamus ^25,26. However, when Netrin-1 was tested in isolation, no attractive response was triggered, and only when Slit1 and Netrin-1 were presented together did an attractive activity emerge that neither of the cues possesses alone ³².

Several studies have demonstrated that spontaneous activity plays a crucial role in early stages of neuronal development modulating processes such as axon pathfinding and targeting ^4,33,34,35. In embryonic thalamic neurons, the frequency of Ca²⁺ transients diminishes as development proceeds (Fig6; ⁴). Moreover, we demonstrated that this change in spontaneous activity has a specific function in the axonal growth program ⁴. The rapid TCA growth, which occurs at the ventral telencephalon, is correlated with a high frequency of Ca²⁺ transients, while the dampening of TCA cortical extension requires a reduction of spontaneous Ca²⁺ transients at late embryonic stages. When spontaneous thalamic activity was blocked, the speed of TCA extension was decreased both in vitro and in vivo ⁴.

Genetic reduction of spontaneous activity by Kir2.1 overexpression diminishes DCC expression at the level of the mRNA and protein. This decrease is opposite to the one found when spontaneous thalamic activity is blocked ⁴, which suggests that the endogenous downregulation of DCC expression in thalamic neurons is regulated by activity-dependent mechanisms. This idea was further confirmed when the delayed TCA cortical extension was rescued by in vivo overexpression of a full-length DCC receptor with Kir. Thus, as demonstrated for Robo1 ⁴, electrical activity influences the axon growth speed by modulating the expression of at least two receptors, Robo1 and DCC, that function in an antagonistic manner in regulating this process.

It is well accepted nowadays that during brain development, spontaneous Ca²⁺ activity modulates early neural processes by the regulation of the expression of specific genes ^36,37. We previously found that the activity-dependent regulation of Robo1 relies on the binding of the NF-κB transcription factor to a specific site of its promoter. When this site was mutated, the regulation of Robo1 transcription by activity was dampened. Here, spontaneous activity appears to activate Dcc transcription through an AP-1-binding site and mutating this AP-1-binding site severely dampened the activity-dependent regulation of Dcc, suggesting that this site might be crucial for the involvement of DCC in growth (Fig6).

The AP-1 transcription factor is comprised of a variety of dimers containing members of the Fos and Jun protein families ³⁸. In fact, c-Jun is a major component of the AP-1 transcription factor complex ³⁹, and AP-1 activity is in part mediated, by the phosphorylation of c-Jun by the Jun-N-terminal kinases (JNKs). In function of the residues phosphorylated, c-Jun can activate or repress target gene transcription, playing a role in processes such as neurotransmitter specification or axonal regeneration ^37,40. For example, regulation of the tlx3 transcription factor, the activity of which is crucial for neurotransmitter specification in embryonic spinal cord neurons, occurs in an activity-dependent manner through a cAMP response element (CRE) motif in its promoter ³⁷. This CRE variant binds phosphorylated cJun, which in turn represses tlx3 transcription ³⁷. Thus, Ca²⁺ levels can modulate both NF-κB and c-Jun (AP-1) activity ^37,41, and an increase in this activity triggered by the switch in Ca²⁺ frequency occurring in late thalamic neurons ⁴ could have opposing effects on the transcription of target genes, such as the increase in Robo1 and the decrease of Dcc transcription, reducing TCA growth. However, the mechanism by which AP-1 transcriptional activity is regulated in thalamic neurons during early and late phases of TCA growth remains to be elucidated.

In summary, the data presented here extend our knowledge on the factors and mechanisms implicated in axon growth and brain wiring that could be also considered in the context of axonal regeneration or repair.

Materials and Methods

Mouse strains

Wild-type mice maintained on a CD1 background were used to analyze gene expression and for the tissue culture experiments. Netrin-1 heterozygous mice ⁴² were also maintained on a CD1 genetic background and crossed to produce homozygous embryos. Dcc heterozygous mice ¹⁷ were maintained on a 129SV/SvPasCrl genetic background and crossed to produce homozygous embryos. Netrin-1 mutant mice are severe hypomorphs, and they were generated in a gene trap screen where the lacZ gene was inserted into the Netrin-1 gene ^18,42. Dcc knockout mice were generated by the targeted insertion of the neomycin resistance cassette into exon 3 by homologous recombination, resulting in a complete functional deletion ¹⁷. In all cases, the day of vaginal plug was considered as embryonic day (E) 0.5. All the animals were maintained in accordance with Spanish and EU (Directive 2010/63/EU) regulations.

Primary thalamic cell culture, transfection, and pharmacological studies

To establish primary thalamic neuron cultures, pregnant mice were sacrificed and their embryos were recovered. The thalamus was dissected out of the brain, collected in Krebs solution, trypsinized, and then dissociated with a fire-polished Pasteur pipette, with 200,000-300,000 cells/well finally plated in a 24-well seeded in plating medium (sodium pyruvate 1 mM, Glutamax 1×, 45% glucose solution 50 mM, penicillin/streptomycin 100 U/ml, 10% FBS, and 86% MEM). Primary thalamic cells were cultured in maintenance medium (Glutamax 1×, 45% glucose solution 50 mM, penicillin/streptomycin 100 U/ml, 2% B27 and 95% neurobasal) at 37 °C with 5% CO₂, carefully replacing the medium every 2 days. The cells were transfected using Amaxa™ Basic Nucleofector™ Kit (VPI-1003). For pharmacological studies, KCl (2.5 mM) was added to the culture medium at the time of plating. For the detection of cell death by release of lactate dehydrogenase activity (LDH), thalamic cells (20,000 cells) were dissociated using the LDH Cytotoxicity Assay Kit (P40104; Innoprot), a colorimetric method to measure LDH activity using a reaction cocktails containing lactate, NAD⁺, diaphorase, and INT. LDH catalyzes the reduction of NAD⁺ to NADH in the presence of L-lactate, while the formation of NADH can be measured in a coupled reaction in which the tetrazolium salt INT is reduced to a red formazan product. The amount of the highly colored and soluble formazan can be measured at 490 nm spectrophotometrically.

Calcium imaging

Cell cultures transfected with the Kir2.1 and/or Gfp plasmids were incubated for 45 min in 400 μl of maintenance medium (37°C) with Rhod-2 AM cell permanent calcium dye (1 mM; Molecular probes) prepared in DMSO + 20% w/v pluronic acid (2 μl of per coverslip). Coverslips with loaded cultures were washed with maintenance medium, and they were placed in a recording chamber of an upright Leica DM LFSA stage, perfused with warmed (32°C) artificial cerebrospinal fluid (aCSF: in mM, 119 NaCl, 5 KCl, 1.3 MgSO₄ 7H₂O, 2.4 CaCl₂, 1 NaH₂PO₄, 26 Na₂HCO₃, 11 glucose) bubbled with 95% O₂/5% CO₂. Time-lapse recording of Ca²⁺ dynamics was obtained through a water immersion objective (L x10/0.30 Leica) after exciting the cultures at 552 nm with a mercury arc lamp. We acquired images with a digital CCD camera (Hamamatsu ORCA-R2 C10600-10B) at a time resolution of 250 ms. Calcium traces were analyzed with a custom routine developed in Matlab ⁴³, and the proportion of the active cells among the Gfp- and Gfp+Kir2.1-transfected cells was calculated. Calcium transients were identified using asymmetric least squares baseline and Schmitt trigger analysis. We used 5% of the baseline noise as the high threshold and 2% as low threshold. Cells having at least one transient above this threshold were considered as active.

Western blots

Neuron cultures were lysed in NP-40 lysis buffer (20 mM Tris–HCl [pH 8], 137 mM NaCl, 2 mM EDTA and 10% glycerol and 1% Nonidet P40 (NP-40)) supplemented with protease and phosphatase inhibitor cocktails (Roche). The protein lysates were clarified by centrifugation at 10,000× g for 10 min before the protein extracts were resolved by 8%–12% SDS–PAGE and transferred to a nitrocellulose membrane (Protran; Whatman). The membranes were then blocked for 30 min at room temperature with 5% nonfat powdered milk in Tris-buffered saline–Tween-20 (TBS-T: 50 mM Tris–HCl, 150 mM NaCl [pH 7.4], 0.05% Tween-20), and they were probed overnight at 4°C with the following primary antibodies: goat anti-DCC (diluted 1/500; R&D Systems) or rabbit anti-actin (diluted 1/3,000; Sigma). The membranes were then incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody (Sigma), and antibody binding was visualized by chemiluminescence (Immobilon Western; Millipore) in a Luminescent Image Analyzer LAS-1000 Plus (Fujifilm).

In utero electroporation

In utero electroporation technique was performed as previously described ⁴. Briefly, pregnant females from stage E12.5 were anesthetized with isoflurane and the embryos exposed through laparotomies. The Kir-IRES-Gfp (originally from Venkatesh Murthy) and FL-Dcc (a generous gift from Prof. Marc Tessier-Lavigne) DNA constructs were used at a concentration of 1.5 μg/μl and co-electroporated with a plasmid encoding Gfp at 0.9 μg/μl. Embryos were allowed to develop until E15.5 when TCA extension was analyzed.

Quantitative and semi-quantitative PCR

Total RNA from electroporated thalamic explants cultured for 60 h was extracted using the RNeasy Kit (Qiagen). This RNA was treated with RNase-free DNaseI for 30 min at 37°C prior to reverse-transcribing it into cDNA for 1 h at 42°C using SuperScript II Reverse Transcriptase and Oligo(dT)_12–18 primers (Invitrogen). Quantitative RT–PCR (qPCR) was carried out on an Applied Biosystems 7300 real-time PCR unit using the Power SYBR Green PCR Master Mix (Applied Biosystems), 5 μl of cDNA, and the appropriate primers. Each independent sample was assayed in triplicate, and gene expression was normalized to that of GAPDH. The following primers were used for semi-quantitative PCR: Kir2.1³⁴⁶ (forward TTTGGGAACGGGAAGAGTAAAGTC and reverse TTTTTGGCTGTGTGTTTTGG); Kir2.1¹³⁴ (forward TTTTTGGCTGTGTGTTTTGG and reverse GGTTGTCTGGGTCTCAATGG); DCC (forward GAATGAGGCATGTGCTCAGA and reverse AGAAGGCTCCAAAGGAGAGG); and GAPDH (forward AAAATGGTGAAGGTCGGTGT and reverse CTCACCCCATTTGATGTTAG). For qPCR, the following set of primers were used: GAPDH (forward CGGTGCTGAGTATGTCGTGGAGT and reverse CCAAGAAGAAACTCGGAAAGC); DCC (CTGTCTGTGGACCGAGGTTT and GGTTGGTCCTTCACTCACAGA); and Kir2.1 (forward GTGTCCGAGGTCAACAGCTT and reverse GGTTGTCTGGGTCTCAATGG).

Axon tracing, in situ hybridization, and immunohistochemistry

For axon tracing, embryonic brains or cultured slices were fixed in 4% PFA overnight or for 30 min, respectively. Small DiI crystals (1,1′-dioctadecyl 3,3,3′,3′-tetramethylindocarbocyanine perchlorate; Molecular Probes) were inserted into the thalamus and allowed to diffuse for 1–3 weeks at 37°C before obtaining vibratome sections (100 μm) of the brain. For immunohistochemistry, cultured slices/explants and mouse embryos were fixed in 4% PFA for 30 min and 2–3 h, respectively, and they were stained with the following antibodies: mouse anti-βIII Tubulin (1/1,000; Covance); rabbit anti-GFP (1/1,000; Invitrogen); goat anti-Robo1 (1/100; R&D systems); chicken anti-Netrin-1 (1/100; Abcam); and goat anti-DCC (1/100; R&D Systems). The following secondary antibodies (Molecular Probes) were used to visualize the binding of the primary antibodies: goat anti-mouse Alexa-546 (1/500); goat anti-chicken Alexa-488 (1/500); donkey anti-rabbit Alexa-488 (1/500); goat anti-rabbit Alexa 594 (1/500); donkey anti-goat Alexa-546 (1/500); and goat anti-rabbit Alexa-594 (1/400). DAPI (VWR) or Fast Red (Sigma) was used to counterstain the nuclei. For in situ hybridization, a probe for Netrin-1 (kindly provided by Prof. Tessier-Lavigne, Rockefeller University) was used. Briefly, coverslips were washed with PBT, post-fixed with PFA 4% 15 min, pre-hybridized 1 h at 65°C, and hybridized over night at 65°C with Netrin-1 probe at 6 μl/ml of hybridization solution, putting a total volume of 600 μl per well. Next day, coverslips were washed with different solution before incubating them with blocking solution 1 h at room temperature. After that, they were incubated in anti-DIG-Ab (antibody anti-digoxigenin) at a final concentration of 1:2,000 over night at 4°C. Then, they were washed first with TBS-T and later with NTMT before incubating in reaction mix (NTMT + 4.5 μl/ml NBT + 3.5 μl/ml BCIP) until the staining was observed. Finally, coverslips were washed with PBS, post-fixed with PFA 4% 20 min at RT, and mounted with Mowiol.

Coculture experiments

Organotypic slices of embryonic mice were prepared and cultured for 48 h as described previously ¹. Wild-type thalamic explants were transplanted onto wild-type or Netrin^−/− host coronal slices. Normalized fluorescence was measured using ImageJ.

Quantification of axon growth

ImageJ software was used for image processing and to analyze axon length in vitro and in vivo. To quantify axon growth in dissociated thalamic cells, we measured the length of the axons of 100–300 neurons per well using the NeuronJ plug-in. To quantify the ventral telencephalic extension of TCAs in the coculture slices and the cortical extension of TCAs labeled in utero, the 5-longest axons per slice were measured. For the in utero electroporation, only brains with a similar degree of electroporation in the thalamus (Th) were analyzed. Slices (60 μm) were obtained and images of the cortex were sampled at two different rostro-caudal levels (rostral and intermediate: 3 sections for each level and brain). A horizontal line was drawn from the edge of the pallial–subpallial boundary at the ventricle of each slice. The cortical area delimited by this line was then linearized using the Straighten plug-in of the ImageJ software. Within this area, the distance covered by labeled axons was measured and normalized to the total length of the region defined.

Luciferase assays

Dissociated thalamic cells plated in 24-well plates were co-transfected with a luciferase reporter plasmid as previously described ⁴. A Dual Luciferase Reporter Assay Kit (Promega) was used to measure luciferase activity after 60 h of culturing. Each sample was assayed in triplicate, and luminescence was measured on a Sirius luminometer (Berthold DS). We determined the exact reporter activity by calculating the ratio of the pGL3-based luciferase reporter activity to that of the pRL-TK luciferase.

Bioinformatic analysis on the Dcc promoter

We performed a bioinformatic search within the mouse Dcc gene using the Ensembl Genome Browser (http://www.ensembl.org/index.html), and as described before ⁴. Briefly, a 1491-bp fragment of the mouse Dcc gene was amplified from mouse brain genomic DNA using the following primers: Promoter Dcc F (5′-CATCATACGGATCGGCACAGA) and Promoter Dcc R (5′-CCTAGGCTTCCTCCTCCTCTTC), and further cloned into the Zero Blunt TOPO PCR Cloning Kit (Invitrogen) and, subsequently, into the pGL3-Basic Vector (Promega). We used the bioinformatic tools, TFSEARCH (http://www.cbrc.jp/research/db/TFSEARCH.html) and MATCH (http://www.bioinfo.de/isb/gcb01/poster/goessling.html) to search for putative transcription factor-binding sites within the identified Dcc promoter sequence. The ECR Browser (http://ecrbrowser.dcode.org/ ⁴⁴) was used to identify conserved regions in the Dcc promoter applying the default settings (100 bp, 70% identity). We used site-directed mutagenesis GTGACTTCCC to AACAGAATTC (NF-κB site) and TTGAGTCAT to TTGGTGCAC (AP-1 site).

Statistical analyses

All data are shown as mean ± standard error of the mean (s.e.m.). Kolmororov–Smirnov normality test was performed to evaluate Gaussian distribution of the samples and proper simple size. Depending on the results, a parametric or nonparametric test was applied for statistical analysis using the Prism software package (version 5). Difference was considered statistically significant when P < 0.05. Assessment of data was performed blinded by different investigators in independent experiments. The selection of the wild-type embryos for this study was unbiased in terms of sex, size, and weight. Experiments were considered as independent when done from distinct litters.

Author contributions

GL-B conceived the idea. MC-P and GL-B designed the study. MC-P performed the in vitro and ex vivo experiments and the semi-qPCR, qPCR, luciferase, and LDH assay. VM-J and GL-B performed the in utero electroporation experiments. AF performed the calcium recordings in dissociated thalamic cells. RS and LM processed the in utero electroporation material and performed qPCR, in situ hybridization, and Western blot experiments. MC-P and GL-B conducted the data analysis and wrote the manuscript.

Conflict of interest

The authors declare that they have no conflict of interest.