MicroRNA-431 regulates axon regeneration in mature sensory neurons by targeting the Wnt antagonist Kremen1

Abstract

MicroRNAs (miRNAs) are small, non-coding RNAs that function as key post-transcriptional regulators in neural development, brain function, and neurological diseases. Growing evidence indicates that miRNAs are also important mediators of nerve regeneration, however, the affected signaling mechanisms are not clearly understood. In the present study, we show that nerve injury-induced miR-431 stimulates regenerative axon growth by silencing Kremen1, an antagonist of Wnt/beta-catenin signaling. Both the gain-of-function of miR-431 and knockdown of Kremen1 significantly enhance axon outgrowth in murine dorsal root ganglion neuronal cultures. Using cross-linking with AGO-2 immunoprecipitation, and 3'-untranslated region (UTR) luciferase reporter assay we demonstrate miR-431 direct interaction on the 3'-UTR of Kremen1 mRNA. Together, our results identify miR-431 as an important regulator of axonal regeneration and a promising therapeutic target.

Keywords: Kremen1; Wnt; axon; miR-431; miRNA; regeneration; sensory neurons.

FIGURE 1

Sciatic nerve injury induced changes in miRNA expression profile in DRG. (A) Total RNA for the microarray expression analysis was isolated from DRG 4 days after sciatic nerve crush. Agilent arrays were done in duplicates and repeated twice. Normalization and analyses were performed with GeneSpring software. miRNAs with a statistically significant upregulation or down-regulation over 1.5-fold were listed in the table. (B) Three miRNAs that were significantly upregulated were selected for further validation. Real-time qPCR for miRNA validated the relative changes in miRNA level. miRNA expression was normalized to reference gene s12. The graph indicates a significant increase of miR-744, miR-431, and miR-21 in DRG after sciatic nerve crush, whereas the expression level of miR-124 and miR 29a did not change (*p < 0.05, **p < 0.01, N = 3). (C) Venn diagram of overlap in predicted miR-431 target genes and down-regulated genes in DRG after conditioning sciatic nerve lesion. The potential targets of miR-431 were chosen using three algorithms , , and . Down-regulated genes were selected using fold change cut-offs of >2 and significance p-values of <0.05 expression based on microarray data for DRGs 4 days post-sciatic nerve injury. Overlap shows 24 genes having predicted binding site for miR-431 and significantly down-regulated expression level in DRG microarray. A one-way ANOVA followed by Bonferroni’s multiple comparison tests was utilized. For the analysis of two independent groups, Student’st-test was used.

FIGURE 2

miR-431 increases axon outgrowth in DRG neurons. Effects of miR-431 mimic and inhibitor on axon outgrowth. (A) Left panel shows the effect of the transfection of DRG neurons with miR-431 mimic. Right panel depicts the effect of transection with miR-431 inhibitor. Negative controls for miR-431 mimic and inhibitor are indicated on the lower images. Cells were stained with primary antibodies against neuronal β-tubulin and signals were visualized with TX-Red conjugated secondary antibody (scale bar: 50 μm). The expression of GAP-43, a marker for axon regeneration, was detected using an anti-GAP-43 antibody and visualized with FITC-conjugated secondary antibodies. The effect of miR-431 on axon length (B) and on axon branching (C) was quantified. Overexpression of miR-431 significantly increased axon extension, whereas suppression of miR-431 significantly blocked axon branching. The fluorescence signal intensity against GAP-43 was quantified in (D). The significant increase in GAP-43 immunofluorescence reflects increase in regenerative axon growth. (E) Significant increase in GAP-43 expression on mRNA level quantified by RT-qPCR (*p < 0.05, **p < 0.01, N = 50). A one-way ANOVA followed by Bonferroni’s multiple comparison tests was utilized.

FIGURE 3

miR-431 regulates Kremen1 expression. (A) RT-qPCR confirmed the increase of miR-431 level in DRG neuron after the transfection of miR-431 mimic. (B) Although overexpression of miR-431 decreased Kremen1 mRNA in total cell lysates (input), it enhanced the binding between Kremen1 mRNA and Ago-2 complex. In the Ago immunoprecipitated fractions, there was an increased amount of Kremen1 mRNA. The lack of signal in the non-specific serum IP sample (IP neg. control) confirmed the specificity of the IP. (C) miR-431 negatively regulated Kremen1 expression at mRNA level. Treatment of miR-431 mimics in DRG neuronal cultures significantly inhibited Kremen1 expression as compared with that of control groups. On the contrary, suppression of miR-431 activity significantly enhanced the expression of Kremen1 mRNA. (D) Western blot analysis of Kremen1 expression exhibited similar negative correlation of miR-431 and Kremen1 expression. Cells transfected with miR-431 mimics had decreased protein level of Kremen1, whiles cells transfected with miR-431 inhibitors had an increased expression of Kremen1. α-tubulin was used as the loading control and was used to normalize densitometry values. (E) The quantification of densitometric levels of Kremen1. (F) PC12 cells were transfected with Kremen1 3′UTR-firefly Luciferase constructs for luciferase assays. Co-transfection with miR-431 mimics significantly reduced the luciferase activity (*p < 0.05, **p < 0.01), whereas co-transfection with mimic negative controls did not affect the expression of firefly luciferase gene. A one-way ANOVA followed by Bonferroni’s multiple comparison tests was utilized.

FIGURE 4

Nerve crush injury reduces Kremen1 expression. (A) Total RNA was isolated from control or crush-injured mouse DRG, and relative expression of Kremen1 was determined using RT-qPCR. GAPDH and S12 were used to normalize for RNA loading. (B) Western blot analysis of total DRG lysates at 4 days post-crush injury. α-Tubulin was shown as a loading control. As shown in the quantified densitometry data, there was a significant decrease of Kremen1 expression during nerve regeneration. (C) Immunofluorescent staining in dissociated DRG neurons demonstrated the expression of Kremen1 within neurons. Kremen1 as a transmembrane receptor was shown to be located in cell bodies, but not axons. TUJ staining was used to visualize neuronal cells. Preconditioning of sciatic nerve clearly promotes regenerative axon growth in DRG neurons, and this phenomenon is accompanied by a decrease in Kremen1 expression. Scale bar: 20 μm. (**p < 0.01) For the analysis of two independent groups, Student’s t-test was used.

FIGURE 5

Knockdown of Kremen1 increases neurite outgrowth. (A) Neurite outgrowth in Kremen1 siRNA and scrambled siRNA treated DRG neurons was detected by TUJ immunostaining. Representative images show that Kremen1 siRNA significantly decreased Kremen1 expression level, which was accompanied by an increase of axon outgrowth. Scale bar: 20 μm. As the quantification performed in miR-431 functional analysis, we measured the length of the longest axon for each neuron (B) and counted the number of branches for each neuron (C). Inhibition of Kremen1 significantly increased the length of axon, however, its effect on neurite branching was not significant. *-p < 0.05. For the analysis of two independent groups, Student’s t-test was used. Scale bar: 20 μm.