MYC regulation of the miR-92-Robo1 axis in Slit-mediated commissural axon guidance

Abstract

In the developing spinal cord, translational repression of Robo1 expression by microRNA-92 (miR-92) in precrossing commissural axons (CAs) inhibits Slit/Robo1-mediated repulsion facilitating commissural axon projection and midline crossing; however, the regulatory mechanisms governing miR-92 expression in the developing commissural neurons are currently lacking. Here, we propose that the transcription factor MYC regulates miR-92 expression in the developing spinal cord (of either sex) to control Robo1 levels in precrossing CAs, modulating Slit/Robo1-mediated repulsion and midline crossing. MYC, miR-92, and Robo1 are differentially expressed in the developing chicken spinal cord. MYC binds to the promoter region upstream of the gga-miR-92 gene in vitro. MYC knockdown dramatically decreases miR-92 expression and increases chicken Robo1 (cRobo1) levels. In contrast, overexpression of MYC significantly induces miR-92 expression and reduces cRobo1 levels. MYC knockdown or overexpression results in significant inhibition or induction of miR-92 activity in the developing chicken spinal cord, respectively. Disruption of the MYC-dependent regulation of the miR-92-cRobo1 axis affects Slit2-mediated CA growth cone collapse in vitro and impairs CA projection and midline crossing in vivo. These results elucidate the role of the MYC-miR-92-cRobo1 axis in Slit2/Robo1-mediated CA repulsion and midline crossing.

During development, translational repression of Robo1 expression in spinal commissural axons by miR-92 silences Slit repulsion; however, the underlying mechanisms governing miR-92 expression in the developing commissural neurons are unknown. Here, we report that MYC transcriptionally regulates miR-92 expression and activity in developing spinal commissural neurons to control Slit/Robo1-mediated repulsion and spinal CA guidance. This study demonstrates that MYC regulation of the miR-92-Robo1 axis plays an important role in Slit repulsive signaling and axon guidance.

INTRODUCTION

In the developing vertebrate spinal cord, commissural neurons (CNs) project axons toward the floor plate (FP) and cross the midline of the spinal cord, a process relying upon the coordination of attractive and repulsive guidance signaling (Colamarino and Tessier-Lavigne, 1995; Tessier-Lavigne and Goodman, 1996; Kaprielian et al., 2001; Dickson, 2002). For instance, activation of Netrin- and Sonic hedgehog-mediated attraction promotes precrossing commissural axons (CAs) projecting toward the FP, while Slit/Robo repulsive signaling is dormant; by contrast, crossing and postcrossing CAs gain responsiveness to Slit- and Semaphorin-mediated repulsion and lose sensitivities of Netrin attraction, which facilitate the CAs to exit the FP and curb their FP recrossing (Kennedy et al., 1994; Kidd et al., 1998; Zou et al., 2000; Charron et al., 2003). Differential regulation of Robo1 receptor expression in developing CNs with low levels in precrossing and high levels in crossing and postcrossing CAs could function as a molecular switch to control the sensitivity of CAs to Slit repulsion during midline crossing; however, the mechanism underlying the temporospatial regulation of Robo1 expression in vertebrate CNs and CAs is not fully understood (Ypsilanti et al., 2010; Alvarez et al., 2021; Zang et al., 2021).

microRNAs (miRNAs), a class of small noncoding RNAs, are abundantly expressed in the developing nervous system exhibiting a distinctive expression pattern with the axon guidance receptors and have been demonstrated as crucial regulators involved in axon pathfinding (Pinter and Hindges, 2010; Baudet et al., 2011; Han et al., 2011; Zou et al., 2012; Cao et al., 2016; Bellon et al., 2017). Our previous studies have shown that miR-92-dependent suppression of Robo1 translation in the developing spinal CNs modulates Slit sensitivity to control CA guidance (Yang et al., 2018). miR-92 is highly expressed in the precrossing CNs and down-regulated in the postcrossing CAs, which represses chicken Robo1 (cRobo1) expression in precrossing CNs, thereby regulating Slit sensitivity to control CA projection and midline crossing (Yang et al., 2018). However, the underlying mechanisms for the differential regulation of miR-92 expression and activity in developing CNs are currently unknown.

The transcriptional regulation of miRNA genes is a major control level responsible for the tissue- and development-specific expression of miRNAs (Krol et al., 2010). As master regulators of neural circuit formation, transcription factors (TFs) control spatiotemporal gene expression of guidance signal molecules and orchestrate precise molecular processes essential for axon elongation, branching, and guidance during embryonic development (Butler and Tear, 2007; Polleux et al., 2007; Moore and Goldberg, 2011; Su et al., 2018). MYC (c-Myc), a universal TF of the MYC proto-oncogene family, is highly expressed in stem cells and progenitor cells and plays an important role in neural development (DePinho et al., 1991; Kerosuo and Bronner, 2016; Kerosuo et al., 2018; Wang et al., 2020; Zaytseva et al., 2020). It has been shown that human MYC directly binds to the specific regions of the miR-92 promoter to induce its expression in B-cell lymphoma and other tumor cells (O'Donnell et al., 2005; Mestdagh et al., 2010; Tao et al., 2014). Interestingly, TFBIND software reveals several potential MYC binding sites within the regulatory region of the chicken miR-92 gene. These results suggest that MYC may differentially regulate miR-92 expression in the developing chicken spinal cord to control Slit/Robo1-mediated CA guidance.

This study explored the regulatory relationship between MYC, miR-92, and cRobo1 in Slit-mediated CA guidance. MYC, miR-92, and cRobo1 are differentially expressed in developing chicken spinal cords and spinal CNs. MYC transcriptionally increases miR-92 expression and activity, resulting in the down-regulation of cRobo1 levels in precrossing CAs. Disruption of endogenous MYC levels by either MYC knockdown or overexpression affects Slit/Robo1-mediated CA growth cone (GC) collapse in vitro and spinal CA projection and midline crossing in vivo. Together, our results reveal the importance of the MYC-miR-92-cRobo1 axis in Slit-mediated repulsion and spinal CA guidance.

RESULTS

Differential expression profiles of MYC, miR-92, and cRobo1 in the developing chicken spinal cord

To assess the spatiotemporal patterns of MYC, miR-92, and cRobo1 in the developing chicken spinal cords, we performed WISH on chicken embryos and immunohistochemistry on transverse sections of chicken spinal cords at Hamburger and Hamilton stages 15-17 (HH 15-17) and HH 23-25, the critical periods respectively representing precrossing and postcrossing CA development (Comer et al., 2019). At HH 15-17, both MYC and miR-92 exhibited pronounced expressions in the ventricular zone (VZ) and the lateral region of the spinal cord (Figure 1, A and C), where cRobo1 mRNA and cRobo1 protein were undetectable (Figure 1, E and G). At the postcrossing stages (HH 23-25), low levels of MYC were localized to the precrossing CA trajectory and the VZ (Figure 1B; Supplemental Figure S1A), similar to the pattern expression of miR-92 (Figure 1D; Supplemental Figure S1, B–D). Strong signals of cRobo1 mRNA were distributed in the lateral DSC, the motor column, the ventral funiculus (VF), the lateral funiculus (LF), and the dorsal root ganglia at HH 23-25 (Figure 1F). Robo1 protein displayed strong expression in the VF and the LF at the postcrossing stages (Figure 1H), a pattern consistent with the elevated cRobo1 mRNA expression (Figure 1F) and decreased miR-92 levels (Figure 1D; Supplemental Figure S1, B and C) as shown in our previous studies (Yang et al., 2018). To further determine whether MYC is expressed in developing CNs, we carried out immunostaining of dissociated neurons from developing chicken spinal cords at HH 15-17 using anti-MYC and anti-Axonin-1 antibodies. MYC was strongly detected in the nucleus, with moderate signals observed in the soma, the axon, and the GC of Axonin-1–positive precrossing CNs (Figure 1I) (Stoeckli et al., 1997), which is similar to the miR-92 expression pattern in chicken CNs (Yang et al., 2018). Together, these results indicate that both MYC and miR-92 are strongly expressed in precrossing CNs and have a distinct expression profile with cRobo1 mRNA and cRobo1 protein in the developing chicken spinal cord during CA development.

FIGURE 1:

Expression patterns of MYC, miR-92, cRobo1 mRNA, and cRobo1 in the developing chicken spinal cord. (A, B) Immunostaining of MYC on transverse sections of developing chicken spinal cords at HH 15-17 (A) and HH 23-25 (B). (C–F) In situ hybridization of miR-92 (C–D) and cRobo1 mRNA (E–F) on transverse sections of developing chicken spinal cords at HH 15-17 (C and E) and HH 23-25 (D and F), respectively. (G–H) Immunostaining of cRobo1 on transverse sections of developing chicken spinal cords at HH 15-17 (G) and HH 23-25 (H). LF, the lateral funiculus; VF, the ventral funiculus. Arrowheads and arrows in A–F indicate the location of CN cell bodies and CA trajectories, respectively. White arrows in H indicate the LF and VF. Scale bar, 50 µm. (I) MYC expression (green) in a developing chicken CN and CA. An antibody against Axonin-1 was used to identify the CNs (red). A total of 100 cells were randomly analyzed. Scale bar, 10 µm.

MYC regulates cRobo1 expression via miR-92 in developing chicken spinal CNs

MYC, a basic helix-loop-helix-leucine zipper (bHLHZip) protein, forms a heterodimeric complex with MAX binding to the regulatory region of the target genes to activate their expression (Madden et al., 2021). We sought to investigate whether MYC could regulate the miR-92 and cRobo1 expression in embryonic chicken spinal cords, considering its ability to up-regulate miR-92 expression in various tumor cell lines (Ji et al., 2011; Li et al., 2014). A targeted knockdown approach by RNA interference (RNAi) was utilized. A shRNA construct designed to target the 3′UTR of chicken MYC mRNA was generated and nucleofected into dissociated neurons from chicken neural tubes at HH 10-12. The knockdown efficiency of the MYC shRNA was confirmed by immunocytochemistry (Figure 2, A and B) and qRT-PCR (Figure 2C). We observed a concomitant reduction in miR-92 levels (Figure 2D) and an up-regulation of cRobo1 protein expression (Figure 2, F and G), while cRobo1 mRNA expression remained unaltered following MYC knockdown (Figure 2E).

FIGURE 2:

Down-regulation of miR-92 and up-regulation of cRobo1 expression after MYC knockdown in primary neurons. (A–B) Representative images (A) and quantification (B) of MYC and TUBB3 immunostaining in developing chicken DSC neurons nucleofected with either control or MYC shRNA. Scale bar, 10 µm. Data are mean ± standard error (SE, ∼80 cells/group). ****p < 0.0001 (Student's t test with Welch's correction). (C–E) qRT-PCR analysis of MYC mRNA, miR-92, and cRobo1 mRNA levels in chicken DSCs after electroporation with either control or MYC shRNA. Data are mean ± SE from three independent experiments (biological replicates, each coded in a different color), with three technical replicates per biological experiment, all coded in the same color. ns, not significant, *p < 0.05, ***p < 0.001 (Student's t test with Welch's correction). (F) Representative images of cRobo1 immunostaining in developing chicken DSC neurons nucleofected with either control or MYC shRNA. Robo1 and TUBB3 were detected using rabbit anti-Robo1 and mouse anti-TUBB3 antibodies, respectively. Scale bar, 10 µm. (G) Quantification of F. Data are presented as mean ± SE (∼60 cells/group). ****p < 0.0001 (Student's t test with Welch's correction).

To further study how MYC regulates miR-92 and cRobo1 gene expression, we electroporated either a MYC-HA construct or a control vector into chicken neural tubes at HH 10-12 and collected DSCs at HH 23-25. The qRT-PCR analysis on these neurons revealed that MYC overexpression increased endogenous miR-92 expression (Figure 3A), while cRobo1 mRNA levels remained unchanged (Figure 3B). Western blot analysis showed a reduction of cRobo1 protein levels in the MYC overexpression group, compared with the vector group (Figure 3, C and D). In addition, ectopic expression of MYC in the DSC significantly reduced cRobo1 protein in the ipsilateral LF and VF, compared with the contralateral side of the spinal cord (Figure 3, E and F).

FIGURE 3:

MYC up-regulates miR-92 and down-regulates cRobo1 expression in developing DSCs. (A and B) qRT-PCR analysis of miR-92 (A) and cRobo1 mRNA (B) levels in developing chicken DSCs after electroporation of either vector or MYC constructs. Data are presented as mean ± SE from three independent experiments (biological replicates, each coded in a different color), with three technical replicates per biological experiment, all coded in the same color. ns, not significant, **p < 0.01 (Student's t test with Welch's correction). (C and D) Immunoblot analysis depicting cRobo1 protein levels in developing chicken DSC neurons electroporated with either a control vector or MYC. Data are mean ± SE from four independent experiments. *p < 0.05 (Student's t test with Welch's correction). (E) Immunostaining of cRobo1 on transverse sections of developing chicken spinal cords overexpressing MYC. The rabbit anti-Robo1 antibody was used to detect cRobo1. “+” denotes the MYC-electroporated side, and “–” denotes the nonelectroporated side. The white dashed areas on both sides indicate the region of interest for quantification. Scale bar, 50 µm. (F) Quantification of cRobo1 intensities in the MYC-electroporated (Venus-YFP-positive) and unelectroporated (Venus-YFP-negative) sides of developing chicken spinal cords, harvested at HH 23-25. Data were collected from multiple sections of six embryos. Values are mean ± SE, **p < 0.01 (Student's t test with Welch's correction).

MYC regulation of miR-92 expression and activity by binding to its upstream regulatory regions

As a TF, MYC binds to the upstream regulatory regions of target genes to regulate their transcription. According to the e!Ensembl database, gga-miR-92 is regulated by two distinct promoters: a short sequence of 100 nucleotides and a long sequence of 222 nucleotides located at 2.9 and 3.3 kb upstream of the gga-miR-92 gene, respectively. TFBIND software predicts two potential MYC-binding sites in the short region and three possible sites in the long sequence (Figure 4A). To investigate whether MYC regulates gga-miR-92 expression by binding to these promoter regions, we generated a luciferase reporter containing either the short (100 nt) or long (222 nt) promoter sequences. The fragments of chicken genomic DNA corresponding to the short and the long sequences were cloned into the pGL3-basic luciferase reporter plasmid upstream of the luc+ gene (Figure 4A). The luciferase reporter constructs were cotransfected with either vector or MYC-HA constructs in HEK293 cells, and luciferase activity was measured 48 h posttransfection. MYC overexpression significantly increased the luciferase activity of both the short (Figure 4B) and long (Figure 4C) promoters compared with the control vector group (Figure 4, B and C). This result suggests that the binding of MYC to these promoter regions could regulate gga-miR-92 expression.

FIGURE 4:

MYC regulates gga-miR-92 activity by binding to its promoter regions. (A) Diagram of the gga-miR-92 promoter regions (brown) cloned into a pGL3 luciferase reporter plasmid and positioned upstream of the luc+ gene (yellow). The white stars in each promoter represent the putative MYC binding sites. (B–C) Quantification of luciferase activities in vitro. The luciferase reporter plasmids, encoding either a short promoter sequence (100 bp, B) or a long promoter sequence (222 bp, C), were cotransfected with either the control vector or MYC-HA in HEK293 cells. Data are presented as mean ± SE from three independent experiments. **p < 0.01, ***p < 0.001 (Student's t test with Welch's correction). (D) Representative images of transverse sections of the chicken spinal cord electroporated with control shRNA plus control Sensor, control shRNA plus miR-92 Sensor, MYC shRNA plus miR-92 Sensor, vector plus control Sensor, vector plus miR-92 Sensor, and MYC plus miR-92 Sensor. Scale bar, 50 µm. (E–F) Quantification of the YFP/RFP ratio in the electroporated DSC in the MYC knockdown (E) and overexpression (F). Data represent multiple sections from 6 to 9 embryos in each group. Values are mean ± SE. **p < 0.01, ***p < 0.001, ****p < 0.0001 (one-way ANOVA and Tukey's test for post hoc comparisons).

Active miR-92 binds to the MRE in the 3′ UTR of cRobo1 mRNA, thereby repressing its expression (Yang et al., 2018). To detect whether MYC regulates endogenous miR-92 activity in the developing spinal cord, a dual fluorescence miR-92 Sensor and control Sensor have been generated and validated in chicken embryos (Yang et al., 2018). In this sensor, miR-92 MRE was cloned downstream of YFP, and as an internal control, RFP was driven by another separate cytomegalovirus (CMV) promoter. Endogenous miR-92 activity could be tracked by analyzing the YFP/RFP ratio in the chicken DSC, where CNs were located. The effects of MYC loss or gain of function on the regulation of endogenous miR-92 activity in developing chicken spinal cord were examined. HH 10-12 chicken neural tubes were electroporated with control shRNA plus control Sensor, control shRNA plus miR-92 Sensor, MYC shRNA plus miR-92 Sensor, vector plus control Sensor, vector plus miR-92 Sensor, and MYC plus miR-92 Sensor, and the electroporated spinal cords were harvested at HH 18-20 (Figure 4D). Fluorescence intensities of YFP and RFP were measured in transverse sections of the lumbosacral regions of the spinal cords, with the YFP/RFP ratio normalized to the control shRNA plus control Sensor group for the knockdown experiment and vector plus control Sensor group for the MYC overexpression experiment. At least six spinal cords were analyzed per group in this experiment. As expected, due to active endogenous miR-92, the YFP/RFP ratio was reduced in the control shRNA plus the miR-92 Sensor group and the vector plus miR-92 Sensor group compared with the control shRNA plus the control Sensor group and the vector plus control Sensor group, respectively (Figure 4, D–F). Coexpression of MYC shRNA with miR-92 Sensor rescued the ratio to a similar level as observed in the control shRNA plus the control Sensor group (Figure 4, D and E). Furthermore, the YFP/RFP ratio decreased further in the MYC overexpression group, compared with the control vector plus miR-92 Sensor group (Figure 4, D and F). These findings indicate that MYC up-regulates endogenous miR-92 activity, modulating miR-92–mediated gene repression in the developing chicken spinal cord.

MYC regulation of the miR-92-cRobo1 axis is required for Slit2-mediated GC collapse

Highly dynamic GCs of developing axons play a pivotal role in guiding axon elongation and pathfinding (Letourneau, 1983; Huber, et al. 2003; Chilton, 2006). GC collapse, characterized by a reduction in the net GC area due to the retraction of filopodia and lamellipodia, signifies a key process of axon repulsion (Walter et al., 1990; Gallo and Letourneau, 2004). To examine the role of MYC in Slit-mediated spinal CA repulsion, we performed a Slit2-induced GC collapse assay, a well-established approach for studying Slit/Robo repulsive signaling (Piper, et al., 2006; Guan et al., 2007; Kapfhammer et al., 2007; Delloye-Bourgeois et al., 2015). Because most of the dissociated DSC cells were CNs (with ∼ 95% being positive for Axonin-1) (Supplemental Figure S2), we harvested the chicken DSCs at HH 23-25 and nucleofected dissociated CNs with Venus YFP plus either control or MYC shRNA. Primary CNs were stimulated with purified Slit2 (Slit2-N fragment) or control (1 X PBS) and stained with phalloidin and DAPI to demarcate the GC and nucleus, respectively. Slit2 increased GC collapse of chicken spinal CAs in the control shRNA group: without Slit2 stimulation, 36.49 ± 6.09% of GCs collapsed; with Slit2 treatment, 67.64 ± 5.81% of axons exhibited GC collapse (Figure 5, A, B, and I). MYC knockdown further enhanced Slit2-induced GC collapse compared with the control shRNA group: with Slit2 treatment, the percentile of GC collapse was increased from 67.64 ± 5.81% in the control shRNA group to 80.37 ± 5.81% in the MYC shRNA group (Figure 5, A–D, and I). Transfection of MYC shRNA with either wild-type MYC or miR-92 rescued the effect of MYC shRNA on Slit2-induced GC collapse, restoring it to a level comparable to the control shRNA group: under Slit2 stimulation, the percentile of GC collapse was reduced from 80.37 ± 5.81% in the MYC shRNA group to 69.87 ± 2.24% in the MYC shRNA plus wild-type MYC group and 69.30 ± 2.29% in the MYC shRNA plus miR-92 group, which is similar to 67.64 ± 5.81% in the control shRNA group (Figure 5, E–I). These results indicate that MYC regulation of miR-92 expression is required for Slit2-induced spinal CA repulsion.

FIGURE 5:

MYC knockdown increases Slit2/Robo1-induced GC collapse of developing chicken DSC neurons. Chicken DSC neurons at HH 23-25 were dissociated and nucleofected with control shRNA (A, B), MYC shRNA (C, D), MYC shRNA plus MYC (E, F), MYC shRNA plus GFP-gga-miR-92 (G, H). Neurons were counterstained with phalloidin to visualize actin filaments (red). Only the GCs of Venus YFP-positive (A–F) or GFP-positive (G, H) neurons not in contact with other cells were quantified and utilized in the statistical analysis. Scale bar, 10 µm. (I) Quantitative assessment of Slit2-induced GC collapse. The y-axis represents the percentage of collapsed axonal GCs. Data are presented as mean ± SE from three independent experiments (100–150 neurons per group). *p < 0.05, ****p < 0.0001 (one-way ANOVA and Tukey's test for post hoc comparisons).

To further explore the impact of MYC levels on Slit-mediated CA repulsion, we performed the same CA GC collapse assay as described above. Primary chicken CNs were nucleofected with Venus YFP only or Venus YFP plus wild-type MYC. As expected, Slit2 induced GC collapse of CNs in the Venus YFP group: without Slit2, 31.85 ± 4.68% of GCs collapsed; with Slit2 treatment, 77.37 ± 1.74% of GCs collapsed (Figure 6, A, B and K). Overexpression of MYC inhibited the Slit2-induced GC collapse, compared with neurons expressing the Venus YFP alone: with Slit2 stimulation, the percentile of GC collapse was reduced from 77.37 ± 1.74% in the Venus YFP only group to 56.10 ± 3.48% in the MYC plus Venus YFP group (Figure 6, A–D and K). To determine whether MYC inhibition on Slit2-induced GC collapse relies on endogenous miR-92 activities, chicken CNs were cotransfected with wild-type MYC plus Venus-miR-92 Sponge, which inhibits endogenous miR-92 activities (Yang et al., 2018). Coexpression of MYC with miR-92 Sponge increased the Slit2-induced GC collapse, compared with the MYC-expressing group: with Slit2 treatment, the percentile of collapsed GCs was 56.10 ± 3.48% in the MYC plus Venus YFP group and 78.40 ± 3.40% in the MYC plus miR-92 Sponge group, respectively (Figure 6, C–F and K). To test whether the effect of MYC on Slit2-mediated GC collapse requires the miR-92–dependent regulation of cRobo1, we cotransfected MYC with either miR-92–sensitive cRobo1 or miR-92–insensitive cRobo1 into chicken CNs (Yang et al., 2018). Under Slit2 stimulation, overexpression of MYC with miR-92–sensitive cRobo1 resulted in 58.81 ± 2.42% of GC collapse which is similar to 56.10 ± 3.48% of GC collapse observed in the MYC overexpression group (Figure 6, C, D, G, H, and K). However, cotransfection of MYC with miR-92–insensitive cRobo1 reinstated the Slit2-induced GC collapse, compared with the MYC overexpression group: with Slit treatment, the percentile of GC collapse was increased from 56.10 ± 3.48% in the MYC overexpression group to 78.73 ± 7.37% in the MYC plus miR-92–insensitive cRobo1 group (Figure 6, C, D, and I–K). These results indicate that the inhibition of MYC on Slit2-induced CA GC collapse was rescued by expressing either miR-92 Sponge or miR-92–insensitive cRobo1, but not miR-92–sensitive cRobo1, suggesting that this effect could be due to the MYC-dependent up-regulation of miR-92 expression and down-regulation of endogenous cRobo1 levels in CAs.

FIGURE 6:

MYC overexpression attenuates Slit2/Robo1-induced GC collapse. Chicken DSC neurons were dissociated at HH 23-25 and nucleofected with Venus YFP alone (A, B), Venus YFP plus MYC (C, D), MYC plus Venus-miR-92 Sponge (E, F), or MYC in combination with either Venus YFP plus miR-92-sensitive cRobo1 (G, H) or Venus YFP plus miR-92–insensitive cRobo1 (I, J). Filamentous actin is shown in red. Only the GCs of YFP-positive neurons not in contact with other cells were quantified and utilized in the statistical analysis. Scale bar, 10 µm. (K) Quantitative assessment of Slit2-induced GC collapse. The y-axis represents the percentage of collapsed axonal GCs. Data are mean ± SE from three independent experiments (100–150 neurons per group). ns, not significant, **p < 0.01, ****p < 0.0001 (one-way ANOVA and Tukey's test for post hoc comparisons).

MYC regulation of the miR-92-cRobo1 axis is involved in CA projection in vivo

To assess the role of MYC in CA guidance in vivo, chicken neural tubes were electroporated at HH 10-12 in combination with Venus YFP plus control shRNA, Venus YFP plus MYC shRNA, Venus YFP plus MYC shRNA along with wild-type MYC or MYC shRNA plus GFP-gga-miR-92 constructs (Figure 7A). Subsequent analysis of CA projections at HH 18-20 revealed that fewer CAs reached the midline (29.07 ± 3.53% of axons) or FP (29.53 ± 4.25% of axons) in the MYC shRNA group compared with those in the control shRNA group (50.21 ± 7.11% of axons in the transverse sections and 59.50 ± 6.17% of axons in the open-book preparation) (Figure 7, B, C, F, G, J, and K). Expression of either wild-type MYC or miR-92 rescued the CA projection defects caused by MYC knockdown, with 74.12 ± 9.22% or 55.51 ± 6.87% of axons reaching the FP and 61.52 ± 3.94% or 54.54 ± 7.13% of axons reaching the midline, respectively (Figures 7, C–E and G–K). These data indicate that MYC up-regulates miR-92 expression to repress endogenous cRobo1 expression, facilitating precrossing CA projection toward the FP.

FIGURE 7:

MYC knockdown affects chicken spinal CA guidance in vivo. (A) Schematic diagrams illustrating an open-book preparation and a transverse section of the chicken spinal cord after in ovo electroporation. (B–I) Representative images of either the open-book preparation (B, C, D, and E) or transverse sections (F, G, H, and I) of the lumbosacral region of chicken spinal cords (at stages HH 18-20) coelectroporated Venus YFP with control shRNA (B, F), MYC shRNA (C, G), MYC shRNA plus RNAi-resistant human MYC (D, H), and MYC shRNA plus GFP-gga-miR-92 (E, I). The red arrowhead indicates the shortened axons. Scale bar, 50 µm. (J, K) Quantification of the percentage of CAs reaching the FP in the open-book preparation (outlined by two white dashed lines) or the midline in the transverse section of the spinal cord (the white dashed line). Data are presented as mean ± SE (∼10 spinal cords/group). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (one-way ANOVA and Tukey's test for post hoc comparisons).

To further investigate the functional role of MYC in CA guidance, we performed in ovo electroporation as shown above and evaluated the effects of MYC overexpression on postcrossing CA projection of developing chicken spinal cords. Venus YFP alone or Venus YFP plus wild-type MYC was electroporated into the chicken neural tubes at HH 12-15, and either the open-book preparation or transverse sections of chicken spinal cords were collected at the postcrossing stages (HH 23-25), a critical period when the majority of CAs have crossed the midline (Yang et al., 2018; Huang et al., 2024). MYC overexpression led to a significantly higher number of CAs stalling (33.08 ± 1.84% of axons in the open-book preparation and 31.04 ± 2.99% of axons stalling in the transverse sections) at the FP, compared with the Venus YFP alone group, in which only 19.37 ± 2.21% of axons stalling in the open-book and 14.84 ± 2.37% of axons stalling in the transverse sections (Figure 8, A, B, F, G, K, and L). To determine whether the effect of MYC overexpression on postcrossing CA projection relies on the endogenous miR-92 activity and cRobo1 expression, chicken neural tubes were electroporated with wild-type MYC plus Venus-miR-92 Sponge or Venus YFP plus either miR-92–sensitive cRobo1 or miR-92–insensitive cRobo1. The percentile of CAs stalling at the FP in embryos electroporated with MYC along with either miR-92–insensitive Robo1 (14.78 ± 2.00% of axons in the open-book preparation and 14.80 ± 1.33% of axons in the transverse sections) or Venus-miR-92 Sponge (13.67 ± 1.71% of axons in the open-book preparation and 17.18 ± 1.49% of axons in the transverse sections) was reduced (Figure 8, B, C, E, G, H, J, and L), compared with that of the MYC overexpression group. However, the CA stalling defects caused by MYC overexpression could not be rescued by expression of miR-92–sensitive cRobo1, with 35.16 ± 2.94% of axons stalling in the open-book preparation and 31.70 ± 3.23% of axons stalling in the transverse sections, which are similar to those in the MYC overexpressing group (Figure 8, D, I, K, and L). These results indicate that the MYC regulation of the miR-92-cRobo1 axis is required for spinal CA midline crossing.

FIGURE 8:

The involvement of the MYC-miR-92-cRobo1 axis in the spinal CA guidance. (A–J) Representative images of either the open-book preparation or the transverse section of the lumbosacral region of chicken spinal cords (at stages HH 23-25) electroporated with Venus YFP alone (A, F), Venus YFP in combination with MYC (B, G), MYC plus Venus-miR-92 Sponge (C, H), MYC plus Venus YFP and miR-92-sensitive cRobo1 (D, I), or MYC plus Venus YFP plus miR-92-insensitive cRobo1 (E, J). The red arrowheads indicate the stalling CAs on the FP. Scale bar, 50 µm. (K, L) Quantification of the percentage of CAs reaching the FP in the open-book preparation or in the transverse section of the spinal cord (outlined by white dashed lines). Data are presented as mean ± SE (∼10 spinal cords/group). miR-92-S cRobo1, miR-92-sensitive cRobo1; miR-92-IS cRobo1, miR-92–insensitive cRobo1. ns, not significant, ***p < 0.001, ****p < 0.0001 (one-way ANOVA and Tukey's test for post hoc comparisons).

DISCUSSION

miRNAs emerge as the critical regulators of local mRNA translation, modulating axonal growth and guidance in the developing nervous system (Rajman and Schratt, 2017; Corradi et al., 2020). As upstream regulators, TFs can either activate or repress miRNA expression (Wang et al., 2009) to control tissue- and developmental stage-specific expression of their target miRNA transcripts during development (Krichevsky et al., 2003; Krol et al., 2010). MYC, a powerful TF, up-regulates miR-92 in distinct tumor cells (Coller et al., 2007; Mestdagh et al., 2010; Singh et al., 2023). High-throughput screening and fundamental molecular approaches have demonstrated that MYC can bind to the regulatory region of the miR-17-92 cluster, thereby activating its expression (McCarthy, 2005; O'Donnell et al., 2005; Mu et al., 2009; Ji et al., 2011). MYC is differentially expressed in the developing spinal cord (GEISHA database) (Davis et al., 1993; Bell et al., 2004; Wey and Knoepfler, 2010; Zinin et al., 2014; Chau et al., 2018). Bioinformatic analysis revealed a set of TFs including MYC having putative binding sites in the promoter region of chicken miR-92 (Figure 4A). These findings propose a potential role for MYC in regulating miR-92 expression in the developing chicken spinal cord.

MYC is highly expressed in the developing spinal cord at precrossing stages (Figure 1A), a pattern similar to the miR-92 expression (Figure 1C). Both miR-92 and MYC are distributed in the VZ, a region giving rise to CNs, and the lateral region of the spinal cord, containing CNs and CAs (Figure 1, A and D; Supplemental Figure S1D). Results from immunocytochemistry demonstrate that MYC is strongly expressed in precrossing CNs, distributing in the nucleus, the cytoplasm, the axon, and the GC (Figure 1I). Bioinformatic analysis revealed several potential binding sites for MYC in the chicken miR-92 promoter, and MYC can induce miR-92 promoter activity (Figure 4, A–C). MYC knockdown dramatically reduced miR-92 expression in DSC neurons (Figure 2, C and D). In contrast, MYC overexpression resulted in an up-regulation of endogenous miR-92 expression (Figure 3). Modulation of MYC levels regulates miR-92 activity in the developing spinal cord (Figure 4, D–F). Altogether, these results indicate that MYC specifically regulates cRobo1 expression via miR-92 in the developing chicken spinal cord and CNs (Thomas et al., 2013; Li et al., 2014; Singh et al., 2023).

MYC knockdown or overexpression respectively up-regulates or down-regulates cRobo1 expression in developing chicken DSCs (Figures 2, F and G and 3, C and D) and CAs (Figure 3E), suggesting that MYC negatively regulates cRobo1 expression in the developing chicken spinal cords. As a universal TF, MYC plays dual roles in inducing and repressing the target gene. MYC can bind to Miz-1, a zinc-finger TF, to repress its target genes in physiological and pathological conditions such as cancer (Kaddurah-Daouk et al., 1987; Wanzel et al., 2003). Interestingly, disruption of MYC expression by either RNAi or overexpression did not change cRobo1 mRNA levels in the developing DSC neurons (Figures 2E and 3B), suggesting that MYC regulates cRobo1 expression at posttranscriptional levels, but not at transcriptional levels. Our observation of the distinctive expression patterns of MYC, miR-92, cRobo1 mRNA, and cRobo1 protein in the developing spinal cord including CNs and CAs (Figure 1) further supports the hypothesis that MYC transcriptionally induces miR-92 expression to reduce cRobo1 levels in developing spinal cords and precrossing CNs and CAs (Supplemental Figure S3).

In addition to the nuclear location, strong signals of MYC protein were detected in neuronal cytoplasm, the axon, and the GC of developing CNs (Figure 1I); however, the functional role of cytoplasmic MYC in CA projection and guidance remains unexplored. Max-1, an MYC-interacting protein, has been found to colocalize with the Netrin-1 repulsion receptor UNC5H2 in the neuronal processes, suggesting its implication in the Netrin-1–mediated axon repulsion (Huang et al., 2002). Kinesin-1 was shown to interact with MYC and transport MYC for proteasomal degradation in the cytoplasm (Lee, 2014). MYC interacts with α-tubulin and polymerized microtubules in vivo and in vitro (Alexandrova et al., 1995). As modulation of microtubule dynamics plays an important role in controlling axon growth and guidance (Dent et al., 2011; Liu and Dwyer, 2014; Atkins et al., 2023), further research is needed to determine whether cytoplasmic/axonal MYC is involved in Slit/Robo1 repulsive signaling through regulation of microtubule dynamics and motor protein-mediated axonal transport.

Repulsive signaling steers axonal GCs away from repellents, ensuring precise axon projection toward the target during development (Gallo and Letourneau, 2004; Hou et al., 2008; Purohit et al., 2012). The Slit/Robo pathway has been shown to induce axonal GC collapse in various neural populations, including olfactory, hippocampus, retinal ganglion cells (RGCs), and dorsal root ganglion neurons (Ba-Charvet et al., 1999; Wong et al., 2004; Jia et al., 2005; Piper et al., 2006). We found that Slit2 can also induce the collapse of chicken spinal CA GC in vitro (Figures 5, A, B, and I and 6, A–B and K), and knockdown of MYC led to an increase in Slit2-induced GC collapse (Figure 5, A–D and I). Coexpression of MYC shRNA with either the wild-type MYC or miR-92 together rescued the Slit2-increased GC collapse (Figure 5). In contrast, MYC overexpression resulted in a reduction in Slit2-induced GC collapse (Figure 6, A–D and K), which could be rescued by coexpressing MYC with either miR-92 Sponge, inactivating endogenous miR-92, or miR-92-insensitive cRobo1, but not miR-92–sensitive cRobo1 (Figure 6). Altogether, these results suggest that the MYC-dependent regulation of the miR-92-cRobo1 axis is involved in modulating Slit/Robo1-mediated GC collapse and repulsion in the developing CAs.

In the developing vertebrate spinal cord, activation of Slit/Robo signaling propels spinal CAs away from the FP, guiding them to the contralateral side of the spinal cord and preventing recrossing (Kidd et al., 1998; Rajagopalan et al., 2000; Long et al., 2004; Mambetisaeva et al., 2005). To characterize the functional role of MYC in Slit/Robo1-mediated axon guidance in vivo, we employed an open-book preparation and transverse sections of chicken spinal cords after in ovo electroporation. Knockdown of MYC impaired CA projection toward the FP with fewer CAs reaching the FP or the midline (Figures 7, B, C, F, G, J, and K), a phenotype similar to inactivation of endogenous miR-92 by overexpressing miR-92 Sponge in CAs as shown in our previous studies (Yang et al., 2018). Coexpression of MYC shRNA with either wild-type MYC or miR-92 restored normal CA projection (Figures 7, C–E and G–K). Furthermore, overexpression of MYC in developing CAs led to axon stalling at the FP, which is similar to the phenotypes observed in Robo1 mutant mice or miR-92 overexpression in developing chicken CNs (Long et al., 2004; Yang et al., 2018). The MYC overexpression-caused CA stalling could be reduced to the normal levels after coexpressing MYC with either miR-92 Sponge or miR-92–insensitive cRobo1, but not miR-92–sensitive cRobo1 (Figure 8). These results indicate that the MYC-dependent regulation of the miR-92-cRobo1 axis is required for spinal CA projection in chicken embryos.

Transcription regulation of gene expression plays an essential role in axon growth, branching, and pathfinding (Yu and Bargmann, 2001; Butler and Tear, 2007; Moore and Goldberg, 2011). For instance, p53 promotes neurite outgrowth (Di Giovanni et al., 2006; Di Giovanni and Rathore, 2012) and regulates the expression of axon guidance–associated genes, such as Netrins, Semaphorins, Ephrins, and Slits (Arakawa, 2005; Miyamoto et al., 2010; Qin et al., 2010). Nkx2.2 and Nkx2.9 have been identified as key regulators of both FP development and guidance of the CAs across the FP (Holz et al., 2010). Zic2 induces EphB1 expression in the ipsilaterally projecting RGCs, inhibiting their axon midline crossing at the optical chiasm (Garcia-Frigola et al., 2008; Lee et al., 2008). Lhx2 and Lhx9, two LIM TFs, regulate Robo3/Rig1 expression in dl1 CNs to control spinal commissural projections (Wilson et al., 2008). During brain development, the Foxg1-Rp58 complex represses Robo1, Slit3, and Reelin expression, regulating callosal axon guidance and neuronal migration (Cargnin et al., 2018). MYCN, a member of the MYC family of TFs differentially expressed in the developing nervous system, can compensate for MYC activity in the transcription regulation (Stanton et al., 1992; Davis et al., 1993; Wey and Knoepfler, 2010; Zinin et al., 2014; Chau et al., 2018). Further study is needed to determine whether these TFs collaborate with MYC to transcriptionally regulate miR-92 expression in developing spinal CNs during midline crossing.

In conclusion, our study provides clear evidence that the MYC-miR-92-cRobo1 axis plays an important role in Slit/Robo1-mediated repulsion and spinal CA guidance: MYC transcriptionally induces miR-92 expression and activity in the developing spinal cord to repress cRobo1 expression in precrossing CAs, inhibiting Slit repulsion and allowing CA projection toward the FP; during and after midline crossing, down-regulation of MYC/miR-92 in CAs allows resumption of cRobo1 expression, which initiates CA sensitivity to Slit repulsion, ensuring axons leave the FP and preventing them from recrossing the midline (Supplemental Figure S3).

MATERIALS AND METHODS

Materials

The following antibodies were used: Rabbit polyclonal anti-Robo1 (1:1000 for western, Genetex, Catalogue no. GTX114103; RRID: AB_11166472), mouse monoclonal anti-c-Myc (9E10) (Santa Cruz Biotechnology, Catalogue no. SC-40; RRID: AB_627268), mouse monoclonal anti-TUBB3 (clone TUJ) (1:1000 for western, BioLegend, Catalogue no. 801202; RRID: AB_2564645), rabbit monoclonal β3-tubulin (Cell Signaling Technology, Catalogue no. 5666; RRID: AB_10691594), rabbit monoclonal anti-HA (Cell Signaling Technology, Catalogue no. 3724; RRID: AB_1549585), sheep polyclonal anti-DIG-AP Fab fragment (Sigma-Aldrich, Catalogue no. 11093274910; RRID: AB_2734716), goat anti-rabbit IgG-HRP (Santa Cruz Biotechnology, Catalogue no. SC-2004; RRID: AB_631747), m-IgGK BP-HRP (Santa Cruz Biotechnology, Catalogue no. 516102; RRID: AB_2687626), Alexa Fluor 647 donkey anti-rabbit IgG (Thermo Fisher Scientific, Catalogue no. A-31573; RRID: AB_2536183), Alexa Fluor 555 goat anti-mouse IgG (Thermo Fisher Scientific, Catalogue no. A-21424; RRID: AB_141780), and Alexa Fluor 488 donkey anti-mouse IgG (Thermo Fisher Scientific, Catalogue no. A-21202; RRID: AB_141607).

Mouse Slit2 was purchased from R&D Systems (Catalogue no. 5444-SL, Minneapolis, MN, USA). The procedures for generating control sensor, miR-92 sensor, Venus-miR-92 Sponge, GFP-gga-miR-92, miR-92-sensitive cRobo1-HA, miR-92–insensitive cRobo1-HA (Yang et al., 2018), and control small hairpin RNA (shRNA) (Huang et al., 2024) constructs have been described previously. To generate miR-92 and control sensor constructs, oligonucleotides containing the 6X miR-92 miRNA recognition elements (MREs) or mutated 6X miR-92 MREs of the cRobo1 3′UTR (untranslated region) were cloned into the modified pCS2-Venus vector. For the Venus-miR-92 Sponge, 6X bulged complementary sequences of chicken miR-92 were cloned into a modified pCS2 vector. The GFP-gga-miR-92 construct was created by inserting chicken pre-miR-92 along with 5′ and 3′ flanking sequences of chicken genomic DNA into the NheI and NotI sites of the pCDH-CMV-MCS-EF1-GFP-T2A-Puro vector. To develop cRobo1-sensitive and -insensitive constructs, either 6X miR-92 MREs or 6X mutated 6X miR-92 MREs of the cRobo1 3′UTR was inserted into the NotI site of the pCI-NEO-cRobo1 vector. A MYC shRNA was specifically designed to target the 3′ UTR of chicken MYC mRNA. The target sequence of the MYC shRNA (5′-AACTGCCTCATAATTGATACT-3′) was cloned into the mU6pro vector (a gift from Dr. David L. Turner, University of Michigan, USA) (Yu et al., 2002) between the Xba I and the EcoR I sites using the following primers: the forward primer, TTTGCTGCCTCATAATTGATACGTTCAAGAGAAGTATCAATTATGAGGCAGTTTTT, and the reverse primer, CTAGAAAAACTGCCTCATAATTGATACGTCTCTTGAAAGTATCAAT TATGAGGCAG. To construct chicken miR-92 promoter luciferase reporters, 100 or 222 bp promoter fragments were PCR-amplified from the chicken genome. The amplified fragments were then inserted into the pGL3 plasmid between the MluI and BglII sites, positioned upstream of the luciferase gene. The following primers were used for cloning the miR-92 promoter luciferase constructs: for the 100 bp promoter luciferase construct, the forward primer was TGCAACGCGTGCTGCTCGGAAGCACTTTGT and the reverse primer was TGGAAGATCTCTCGGGGACCATGTGGGT; for the 222 bp promoter luciferase construct, the forward primer was TGGAACGCGTACACGGGGCCATGGGGA, and the reverse primer was TGGAAGATCTTGCACTTGTCCGAGCGGGAA. All constructs developed in the laboratory were verified through sequencing.

Chicken embryos

Fertilized white leghorn chicken eggs were obtained from the Poultry Teaching & Research Center at Michigan State University and incubated at 38.9°C with a relative humidity of 78% in this study.

Western blotting

The in ovo electroporation procedure was performed in the neural tube of the chicken embryos at HH 10-12 as explained previously (Hamburger and Hamilton, 1951; Nakamura and Funahashi, 2001). The dorsal spinal cords (DSCs) were harvested at HH 23-25 after the in ovo electroporation in chicken neural tubes at HH 10-12 and dissociated as described previously (Hamburger and Hamilton, 1951; Nakamura and Funahashi, 2001; Liu et al., 2004). The dissociated DSC neurons were lysed with MLB lysis buffer (1% Triton X-100, 50 mM Tris, pH 7.4, 150 mM NaCl, 10% glycerol, 10 mM MgCl₂, and protease and phosphatase inhibitor mixture). The immunoblotting procedure on cell lysates was similar to those described previously (Shao et al., 2017; Huang et al., 2018; Huang et al., 2024). Briefly, protein extracts were resolved by 7.5% SDS–PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad, Hercules, CA, USA). The immunoblots were then incubated with specific antibodies, and the blots were visualized using an enhanced chemiluminescence kit (Thermo Fisher Scientific, Waltham, MA, USA).

Primary culture of chicken DSC neurons

The procedures of the dissection and dissociation of embryonic chicken spinal cords and primary culture of the chicken DSC neurons were conducted as described previously (Masuko et al., 1979; Yang et al., 2018; Dityateva et al., 2003; Huang et al., 2024). Briefly, spinal cords were harvested from chicken embryos at HH 18-20 or HH 23-25, and DSC explants were collected and dissociated using 0.25% trypsin. Dissociated DSC neurons were nucleofected using the program O-03 and plated on poly-L-lysine (PLL, 100 ng/ml)-coated coverslips. The neurons were cultured in DMEM (Cytiva, SH30022.01) supplemented with B27 at 37°C overnight for the GC collapse assay or immunofluorescence studies.

Luciferase reporter assay

Either MYC-HA or the control vector was cotransfected with a luciferase reporter containing putative miR-92 promoter sequences into HEK293 cells. Transfected cells were cultured in DMEM supplemented with 10% FBS for 48 h. Cell lysis buffer and luciferase assay reagents were prepared according to the manufacturer's instructions provided with the Luciferase Assay System (Promega, Madison, WI, USA). Following incubation, cells were lysed using 1X cell culture lysis buffer, and the lysates were transferred to 96-well plates. Luciferase activity was measured by adding 100 µl of luciferase assay reagent to each well immediately before reading the luciferase activities. The relative luciferase activities were normalized to the values obtained from the control vector group. Each experiment was repeated three times to ensure robustness and reproducibility of data.

Immunocytochemistry

To assess endogenous Robo1 and MYC expression, primary neurons after nucleofection with specific constructs were plated on PLL (100 ng/ml)-coated coverslips and cultured in DMEM supplemented with B27 overnight at 37°C. Following incubation, the cells were fixed in 4% paraformaldehyde (PFA) in PBS overnight at 4°C. Then, the neurons were permeabilized with PBST (0.5% Triton X-100 in PBS) for 15 min and blocked with 10% goat serum (Invitrogen, Waltham, MA, USA) at room temperature for 60 min. Cells were respectively incubated with either anti-Robo1 (1:1000) or mouse anti-MYC (1:1000) in combination with mouse anti-TUBB3 (1:1000) or rabbit anti-TUBB3 (1:1000) primary antibodies, followed by incubation with Alexa Fluor 488-conjugated donkey anti-mouse IgG (1:500) and Alexa Fluor 647-conjugated donkey anti-rabbit IgG (1:500) secondary antibodies. To examine endogenous MYC expression in developing CNs, primary neurons dissociated from embryonic chicken DSCs were stained with mouse anti-MYC (1:1000) and rabbit anti-Axonin-1 (1:1000) followed by Alexa Fluor 488-conjugated donkey anti-mouse IgG and Alexa Fluor 647-conjugated donkey anti-rabbit IgG secondary antibodies and counterstained with DAPI. Cells on the coverslips were mounted on glass slides using Fluoro-Gel mounting media (Electron Microscopy Sciences) for microscopic analysis, and fluorescent signals were sequentially captured in a photon counting mode using a Leica HyD detector on a Leica SP8 confocal microscope.

Immunohistochemistry

To analyze the regional expression of MYC and cRobo1 in the developing chicken spinal cord, the embryos were collected at stages HH 15-17 and HH 23-25. A total of 200-µm-thick transverse sections of the lumbosacral region of the spinal cords were prepared and fixed with 4% PFA in PBS at 4°C overnight. The fixed spinal cord sections were then permeabilized with PBST (0.5% Triton X-100 in PBS) for 1 h and blocked with 10% goat serum in PBS at room temperature for another 1 h. These spinal cord slices were incubated with either mouse monoclonal anti-MYC (1:500) or rabbit polyclonal anti-Robo1 (1:200) antibody at 4°C overnight. To assess the MYC regulation of cRobo1 expression in the developing spinal cord, the chicken neural tubes were coelectroporated with Venus YFP and wild-type MYC at HH 12-15, and the transverse sections of chicken spinal cords were collected at HH 23-25. Both mouse anti-MYC and rabbit anti-Robo1 antibodies were utilized. After three washes in PBS, the samples were respectively incubated with Alexa Fluor 488-conjugated donkey anti-mouse IgG (1:250), Alexa Fluor 647-conjugated donkey anti-rabbit IgG (1:250) or both. Fluorescent images of the spinal cord sections were sequentially acquired using a Leica SP8 confocal microscope in a photon counting mode.

Whole mount in situ hybridization

whole mount in situ hybridization (WISH) was performed on spinal cords collected from chicken embryos at HH 15-17 and HH 23-25. Digoxigenin (DIG)-labeled sense and antisense cRNA probes for cRobo1 mRNA were generated and evaluated in our previous studies (Yang et al., 2018). The 3′ end-DIG-labeled gga-miR-92 LNA probes were purchased from Qiagen. Chicken embryos were collected in ice-cold saline, and the spinal cords were isolated and fixed using 4% PFA in DEPC-treated 1X PBS. Samples were processed through the following steps: dehydration through graded methanol (MeOH) series, freezing, rehydration, digestion with proteinase K, additional fixation in 4% PFA, and prehybridization, using a GEISHA protocol adapted from Nieto, Patel, and Wilkinson (Nieto et al., 1996). Spinal cords were then hybridized with the DIG-labeled sense cRobo1, the antisense cRobo1, or miR-92 probe in prehybridization buffer solution (50% deionized formamide, 5X SSC, 2% blocking powder, 0.1% Tween 20, 0.1% CHAPS, 50 µg/ml yeast RNA, 5 mM EDTA, and 50 µg/ml heparin) at 60°C overnight. After hybridization washes in KTBTw solution (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM KCl, and 1% Tween 20), the spinal cords were blocked with 20% sheep serum in KTBTw, followed by antibody staining using a preabsorbed anti-DIG-AP Fab fragment antibody (1:4000) at 4°C overnight. A color development reaction was processed using NBT/BCIP solutions (Sigma-Aldrich). The spinal cords were sectioned into 200-µm-thick sections and images were acquired using a digital camera under a ZEISS Stemi 508 microscope.

Quantitative reverse transcription PCR

Total RNAs, including miRNAs, were extracted from developing chicken DSCs using TRIzol reagent (Life Technologies) (Simms et al., 1993), and the first-strand synthesis of cDNA was performed using AccuScript High Fidelity 1st Strand cDNA Synthesis Kit as per the manufacturer's instruction (Agilent Technologies, Inc., Santa Clara, CA, USA). The quantitative reverse transcription PCR (qRT-PCR) reaction was executed using specific primers listed in Table 1 and SYBR Green Master Mix (Bio-Rad) and analyzed on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). Triplicates of q-PCR reactions were run for the same samples, and three independent biological replicates of each group were carried out. Data were normalized to the expression levels of internal control (Gapdh for cRobo1 and MYC mRNAs and RNU48 for miR-92) and analyzed using the ΔΔCt relative quantification method.

TABLE 1:

Sequences of qRT–PCR primers.

miR-92 qRT-PCR primers
miR-92 stem-loop RT primer	GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACCAGGCC
RNU48 stem-loop RT primer	GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACG ACGGTCAG
qPCR primers	Forward primer (5′→3′)	Reverse primer (5′→3′)
miR-92 qPCR primer	TATTGCACTTGTCCC	GTGCAGGGTCCGAGGT
RNU48 qPCR primer	CCATGAGTGTGTCGCTGATG	GTGCAGGGTCCGAGGT
cRobo1 qRT-PCR primers	Forward primer (5′→3′)	Reverse primer (5′→3′)
cGapdh	TCTTATGACCACTGTCCATGCC	CTTTCCCCACAGCCTTAGCA
cRobo1	TGGATGAAAGCTATGACCAAGA	TCCTCGGACAGGTGGAGTAG
Chicken MYC RT-qPCR primers	AGCGAACGAGTCTGAATCCAG	GTTGGTGGATGTTGACGTGAC

GC collapse

Slit2-induced GC collapse of developing chicken DSCs was performed following previously established protocols (Ba-Charvet et al. 1999; Purohit et al., 2012). Dissociated neurons from the chicken DSC were nucleofected (Amaxa Nucleofector II) with respective constructs using program O-03. After nucleofection, neurons were plated onto PLL-coated coverslips and cultured in DMEM supplemented with B27 at 37°C with 5% CO₂. After 24 h of incubation, neurons were treated with either control PBS or purified Slit2 (4 µg/ml) (Guan et al. 2007) for 30 min. Subsequently, cells were fixed with 4% PFA overnight, permeabilized with PBST (0.5% Triton X-100 in PBS) for 15 min, and stained with phalloidin (1:500) and DAPI (1:500). Fluorescent images of Venus YFP-labeled neurons were sequentially acquired using a Leica SP8 confocal microscope. For quantification, a GC was considered collapsed if it had no lamellipodia or ≤ two filopodia. In each experiment, at least 30–50 random GCs per group were measured, and three independent experiments were performed. A one-way ANOVA with Tukey multiple comparison tests was carried out to determine statistical significance among different groups.

Analysis of spinal CA trajectories in vivo

The electroporation of the developing chicken neural tube in ovo was performed following established procedures (Liu et al., 2004; Liu et al., 2007; Liu et al., 2009; Yang et al., 2018). Embryos were harvested at either HH 18-20 for precrossing CAs or HH 23-25 for postcrossing CAs, respectively. The lumbosacral region of the spinal cord with either YFP or GFP fluorescence was collected and fixed with 4% PFA in PBS at 4°C overnight. The fixed spinal cords were cut into 200-µm-thick transverse slices or prepared as an open-book arrangement by opening the roof plate. Both transverse sections and open-book preparations of spinal cords were mounted on glass slides using Fluoro-Gel (Electron Microscopy Sciences, Hatfield, PA, USA) and imaged under the Leica SP8 confocal microscope. For embryos harvested at HH 18-20, the percentage of CAs reaching the FP in the open-book preparation or the midline of the spinal cord in the transverse sections was calculated by dividing the total number of Venus YFP-positive axons that reached the FP or midline by the total number of fluorescent axons (Liu et al., 2007; Li et al., 2008; Liu et al., 2009; Huang et al., 2018; Yang et al., 2018). For embryos collected at stages HH 23-25, the percentage of stalling axons in the FP was calculated by dividing the total number of Venus YFP-positive axons that stalled at the FP by the total number of fluorescent axons (Liu et al., 2004; Huang et al., 2018; Yang et al., 2018). A total of 8–13 spinal cords were analyzed per group. The phenotypic analysis was confirmed by an independent, blinded analysis.

Experimental design and statistical analysis

The data were represented as mean ± S.E. from three biological replicates unless stated otherwise. Sample sizes for each experiment are based on previous studies from our lab. All data have been analyzed using GraphPad Prism 8.0 (GraphPad Software, Inc., San Diego, CA). The Student's t test with Welch's correction was used for the comparison between two groups, and one-way ANOVA with the Tukey multiple comparison test was used for multiple group comparison. Statistical significance was taken as p < 0.05.

Abbreviations used:

CA

commissural axon

CN

commissural neuron

DSC

dorsal spinal cord

FP

floor plate

GC

growth cone

TF

transcription factor

UTR

untranslated region.