Ephrin-B1 reverse signaling controls a posttranscriptional feedback mechanism via miR-124

Abstract

Eph receptors and ephrins exhibit complex and highly dynamic expression patterns during embryonic development. In addition, changes in their expression levels are often associated with pathological situations in adults. Yet, little is known about the mechanisms regulating their expression. Here we report that the expression of ephrin-B1 is controlled by a feedback loop involving posttranscriptional regulatory mechanisms. We observed that the EfnB1 3' untranslated region (3'-UTR) confers instability to mRNA transcripts, and we identified miR-124 as a posttranscriptional repressor of EfnB1 expression. Furthermore, we showed that miR-124 is itself regulated by ephrin-B1 reverse signaling, thus revealing the existence of a mutually repressive interaction between ephrin-B1 and this microRNA (miRNA). Lastly, we demonstrated the relevance of this mutual inhibition for neuronal differentiation. Our results suggest that miRNAs could be important effectors of Eph/ephrin signaling to refine domains of expression and to regulate function.

FIG. 1.

The 3′-UTR of EfnB1 causes a decrease in mRNA levels. (A) Endogenous levels of EfnB1 in NPCs were measured by qRT-PCR at different time intervals after actinomycin D treatment. U251 cells were transiently transfected with EfnB1 expression constructs with (pcmv-EfnB1-3′-UTR) or without (pcmv-EfnB1) the 3′-UTR, and levels of ectopic EfnB1 transcripts were measured by qRT-PCR at different time intervals after actinomycin D treatment. Results are the means ± standard errors (SE) from 3 independent experiments, each measured in duplicate. (B) U251 cells were transiently transfected with EfnB1 expression constructs with (EfnB1-3′-UTR) or without (EfnB1) the 3′-UTR, and steady-state levels of ectopic EfnB1 transcripts were measured by qRT-PCR. Results are the means ± SE from 3 independent experiments, each measured in duplicate.

FIG. 2.

A posttranscriptional regulatory motif is located in the first 450 bp of the EfnB1 3′-UTR. (A) Normalized luciferase activity in U251 cells transfected with either a reporter construct containing the full-length EfnB1 3′-UTR (FL) or a control construct (pmiR-control). (B, left) Schematic representation of various luciferase reporter constructs. The numbers at right indicate the nucleotide positions included in the reporter constructs. (B, right) Corresponding normalized luciferase activity in U251 cells. (C) Normalized luciferase activity in NPCs electroporated with pmiR-control, FL, and T1 reporters. Results are the means ± standard errors from 6 independent experiments, each measured in duplicate. **, significantly different from pmiR-control, P < 0.001; *, significantly different from pmiR-control, P < 0.05.

FIG. 3.

miR-124 targets the 3′-UTR of EfnB1. (A) Normalized luciferase activity in U251 cells cotransfected with FL, T1, or T6 in the absence (−) or presence (+) of a miR-124 expression construct, as indicated. (B) Normalized luciferase activity in U251 cells cotransfected with T1 or a reporter carrying a mutation in the miR-124 binding site (T1*), in the absence or presence of a miR-124 expression construct. (C) Normalized luciferase activity in U251 cells cotransfected with FL, T1, or T6 with (+) or without (−) LNA-124 or scrambled sequences, as indicated. Results are the means ± standard errors (SE) from 3 independent experiments, each measured in duplicate. (D) U251 cells were cotransfected with pcmv-EfnB1 or pcmv-EfnB1-3′-UTR in the absence or presence of a miR-124 expression construct, and ephrin-B1 protein levels were determined by Western blot analysis. Quantification of three independent experiments was performed using the Grb2 protein level as a loading control. Results are the means ± SE from 3 independent experiments, each measured in duplicate. (E) E14.5 developing cortices were coelectroporated with a GFP reporter plasmid and with either a scrambled oligonucleotide (a) or a miR-124 (b) expression construct. Images show nuclei in blue and GFP-positive cells in green. Quantification of GFP-positive cells in the VZ is also shown (c). Images are representative of 3 independent electroporations. IZ, intermediate zone. *, P < 0.05; **, P < 0.01.

FIG. 4.

Reverse signaling downregulates miR-124 levels. (A) Wild-type NPCs were stimulated with Fc, EphA4-Fc, or EphB2-Fc, and levels of P-STAT3 and P-ERK were assessed by Western blot analysis (left). Ratios of phosphorylated protein levels versus total protein amounts are shown for both STAT3 and ERK1/2 (right). (B) Wild-type NPCs were stimulated with Fc or EphB2-Fc, and miR-124 levels (left) and miR-200C levels (right) were determined by qPCR. (C) miR-124 levels in wild-type (WT) and EfnB1^−/− (KO) NPCs were determined by qPCR. Results are the means ± standard errors from 3 independent experiments, each measured in duplicate. *, P < 0.05. (D) In situ hybridization of coronal sections of E13.5 wild-type (a, b, d, and e) and EfnB1^−/− (c and f) embryos, carried out using either a scrambled LNA as a control (a and d) or LNA-124 as a probe for miR-124 expression (b, c, e, and f). Two different samples at two different magnifications are shown for each condition.

FIG. 5.

Reverse signaling regulates the expression of miR-124 target genes. (A) Wild-type NPCs were stimulated with Fc (control) or EphB2-Fc (reverse), and EfnB1 levels were determined by qRT-PCR. (B) Wild-type NPCs were stimulated with Fc (control) or EphB2-Fc (reverse), and Sox9 levels were determined by qRT-PCR. (C) Normalized luciferase activity in wild-type (WT) and EfnB1-deficient (KO) NPCs electroporated with the T1 reporter and stimulated with Fc (control) or EphB2-Fc (reverse). Results are the means ± standard errors from 3 independent experiments, each measured in duplicate. **, P < 0.01.

FIG. 6.

Increased neuronal differentiation in EfnB1-deficient NPCs correlates with elevated miR-124 levels. (A) NPCs were cultured on poly-l-lysine-coated plates and were allowed to differentiate for 3 or 6 days. Levels of miR-124 and EfnB1 were measured by qPCR and qRT-PCR, respectively, and are expressed as percentages of the day 0 levels. Results are the means ± standard errors (SE) from 3 independent experiments. (B) tubb3 levels in differentiated wild-type (WT) or EfnB1^−/− (KO) NPCs were determined by qRT-PCR. (C) Spontaneous neuronal differentiation in growing wild-type (WT) and EfnB1^−/− (KO) NPCs was assessed by FACS analysis using βIII-tubulin as a marker. FL1-H and FL4-H, heights of fluorescence intensity in channels 1 and 4, respectively. (D) Growing EfnB1^−/− NPCs were electroporated with scrambled LNA or LNA-124, and gfap and tubb3 levels were determined by qRT-PCR. Results are the means ± SE from 3 independent experiments, each measured in duplicate. *, P < 0.05; **, P < 0.01.

FIG. 7.

Downregulation of ephrin-B1 is required for miR-124-induced neuronal differentiation. (A) NPCs were electroporated with a control (a and b) or a miR-124 (c and d) or miR-124+ephrin-B1-GFP (e to h) expression construct, differentiated for 6 days, and stained with HuC/D antibody (b, d, f, and h). Ephrin-B1 expression was detected by epifluorescence (g and h). Images are representative of three separate experiments. Quantification of HuC/D-expressing cells under the three conditions listed above is also shown (i). (B) E14.5 developing cortices were coelectroporated with a GFP reporter plasmid and with the scrambled oligonucleotide (a), miR-124 (b), ephrin-B1-GFP (c), or miR-124+ephrin-B1-GFP (d) expression construct and cultured for 16 h. Images show nuclei in blue and GFP-positive neurons in green. A high magnification of single cells electroporated with the scrambled oligonucleotide, miR-124, ephrin-B1-GFP, or miR-124+ephrin-B1-GFP expression construct is also shown (e), as is the quantification of neurite length of GFP-positive neurons in the CP (f). Images and quantification results are representative of 3 independent electroporated brains. *, P < 0.05.