The Role of Nexmifb in Regulating Spinal Motor Neuron Development through Efna5b in Zebrafish

Abstract

Objective: To confirm the role of nexmifb in the development of spinal motor neuron in zebrafish and explore its molecular mechanism.

Methods: Detect the localization and expression of nexmifb in zebrafish by whole-mount in situ hybridization. Establish a nexmifb knockdown model by morpholino technology. Measure the body length of zebrafish at 72 hours post-fertilization (hpf) in both the control group and the nexmifb knockdown group. Observe the morphology of spinal motor neurons at the 72 hpf laser confocal microscopy in both groups. Motor neuron apoptosis was assessed using the terminal deoxynucleotidyl transferase mediated dUTP nick end labeling (TUNEL) assay kit in both groups. Perform the transcriptome sequencing on embryos from both groups at 72 hpf, then identify the interest gene of efna5b. Observe the rescue effect of efna5b on spinal motor neurons in the nexmifb knockdown group.

Results: nexmifb was primarily expressed in the brain and spinal cord of zebrafish. Knockdown of nexmifb resulted in shortened body length, loss of spinal motor neurons, reduced caudal primary motor neurons (Caps) length, and decreased branching number per mm of Cap. Overexpression of efna5b partially rescued these abnormal phenotypes.

Conclusion: nexmifb regulates spinal motor neuron development by downregulating efna5b.

Graphical abstract

Keywords

nexmifb / motor neuron / efna5b / zebrafish

NEXMIF (also called KIDLIA, KIAA2022, or XPN) is a novel gene localized in Xq13.2. It was first reported in two males with intellectual disability in 2004[1]. So far, little is known about NEXMIF. It has been reported that NEXMIF is associated with multisystem diseases, for example, patients with NEXMIF mutations present with X-linked mental retardation[1]. RNA-Seq indicates that NEXMIF expression is significantly down-regulated in Down syndrome patients with atrial septal defect and/or ventricular septal defect compared with patients without these defects[2]. Nexmif knockout mice exhibit smaller islets and glucose intolerance compared with the control group[3]. However, research on NEXMIF remains primarily focused on its effects on the central nervous system (CNS). This may be attributed to its high expression in the CNS especially in the developing brain[4]. In humans, NEXMIF mutations are not only associated with X-linked mental retardation but also different types of epilepsy, autism spectrum disorder, and other CNS-related diseases[1,5–7]. In animal models, NEXMIF is important in regulating neurite morphological development, cell migration, and cell-matrix adhesion, and maintaining normal synaptic function. For example, loss of NEXMIF altered the migration of layer II/III cortical neurons in mice, reduced dendrite growth, and disturbed apical dendrite projection. Knockdown of Nexmif in cultured rat hippocampal neurons affected axonal development[8], and knockdown in PC12 cells enhanced N-cadherin and β1-integrin-mediated cell-cell and cell-matrix adhesion, thereby inhibiting cell migration[9]. Loss of NEXMIF in mice decreased synapse density, spine density, and the expression of synaptic-related proteins[10]. These results indicate that NEXMIF plays an important role in CNS development.

Zebrafish is a powerful model organism for studying development and regeneration[11]. There are two kinds of spinal motor neurons in zebrafish named primary motor neurons (PMNs) and secondary motor neurons[12,13]. PMNs can be further divided into three groups in each spinal hemisegment according to their somata position and specific axonal projection pathway, they are caudal PMNs (CaPs), middle PMNs (MiPs), and rostral PMNs (RoPs)[13,14]. Although the somata of the three kinds of PMNs are localized at different positions in the spinal cord, their axons travel to the myoseptum via a common exit point and project their axons to innervate the corresponding muscles[15]. The three groups of PMNs are readily identifiable; therefore, they have become excellent cell systems for elucidating motor axon guidance mechanisms in vivo[16].

In zebrafish, nexmif has two paralogs, nexmifa and nexmifb (ENSDARG00000029296). Our previous study has indicated that nexmifa can regulate spinal motor neuron morphogenesis[17]. In this study, we established the nexmifb knockdown model (nexmifb morphant) by injecting nexmifb translation-blocking morpholinos (MOs) into zebrafish embryos and performed RNA sequencing (RNA-Seq) to explore the function and mechanism of nexmifb in spinal motor neuron development.

Materials and methods

Zebrafish husbandry and strain

Wild-type and Tg (mnx1: GFP)^ml2 zebrafish were maintained in the Zebrafish Center at Nantong University under guidelines outlined in previous studies[17].

Whole mount in situ hybridization (WISH)

WISH was performed according to our previous study[19]. cDNA fragments of nexmifb and efna5b from zebrafish embryos were amplified using their specific primers (nexmifb-F: 5′-AATTGCCCGACTCTTACCCA-3′, nexmifb-R: 5′-TCTCAGGGGTGTGATTGCAT-3′, efna5b-F: 5′-TCCAGCCTCCATGATCACA-3′, efna5b-R: 5′-TTTGTTACGGGAAGGCAGAC-3′). Digoxigenin-labeled sense and antisense probes were synthesized using a linearized pGEM-Teasy vector subcloned with the above six fragments by in vitro transcription using the DIG-RNA Labeling Kit (Roche, Basel, Switzerland). Pictures were taken with an Olympus DP70 camera on an Olympus stereomicroscope MVX10.

Quantitative RT-PCR

Tg(mnx1:GFP)^ml2 zebrafish embryos were collected at 72 hours post fertilization (hpf). Total RNA was extracted by TRIzol reagent (Invitrogen, Waltham, MA, United States) from zebrafish embryos. cDNA was synthesized using Transcriptor First Strand cDNA Synthesis Kit (Fermentas, Waltham, MA, United States) and stored at –20 °C. Quantitative RT-PCR (qRT-PCR) was carried out using the corresponding primers: nexmifb-F: 5′-TCCCAGGCAGGAGGTTTCTTCTAC-3′, nexmifb-R: 5′-GGGCGGCAGTGTGTATGATGTC-3′, nexmifa-F: 5′-GATGATGACTGGTGCCCGAAGAAG-3′, nexmifa-R: 5′-CAGACGACAGCAGGTGATGGTTC-3′, efna5b-F: 5′-GCAGGCGGAGATGATCGTGTTC-3′, efna5b-R: 5′-TTCGGTTCCAGAAGACAGCATATCG-3′, ef-1α-F: 5′-CTTCAACGCTCAGGTCATCA-3′, ef-1α-R: 5′-CGGTCGATCTTCTCCTTGAG-3′.

MOs and microinjection

We synthesized nexmifb translation-blocking MOs and standard control MOs using Gene Tools. Diluted them to 0.3 mM, and injected into embryos to establish nexmifb knockdown and control zebrafish embryos. The embryos were cultured in E3 medium at 28.5 °C for the following experiment. The sequence of nexmifb MOs was 5′-AATCTGAATGTGGTCTTCCTTGGAA-3′. The sequence of standard MOs was 5′-CCTCTTACCTCAGTTACAATTTATA-3′. To perform rescue experiments, we generated efna5b mRNAs in vitro. The open reading frames of efna5b were amplified and cloned into the PCS²⁺ vector. After linearization, the 5′-capped efna5b mRNAs were synthesized and purified in vitro using the mMESSAGE mMACHIN Kit (Ambion, Austin, Texas, United States) and RNeasy Mini Kit (Qiagen, Hilden, Germany). MOs and mRNAs were injected singly or combined into the yolk of one-cell stage embryos using borosilicate glass capillaries and a PV830 pneumatic pico pump (Sarasota, Florida, United States).

Terminal deoxynucleotidyl transferase mediated dUTP nick end labeling (TUNEL) assay

Control and nexmifb morphants were collected at 72 hpf and fixed in 4% paraformaldehyde at room temperature for 2 h. After washing three times with PBS for 5 min each, 20 µg/mL proteinase K was added for 10 min to digest embryos, which were washed with PBS three times for 5 min each. The TUNEL assay was carried out using the TUNEL BrightRed Apoptosis Detection Kit (Vazyme, Nanjing, China).

cDNA library preparation and RNA-Seq

Total RNA was extracted by TRIzol reagent from nexmifb morphants and control embryos at 72 hpf. RNA integrity and purity were calculated by NanoDRop 2000 (Thermo Fisher Scientific Inc., Waltham, MA, United States). RNA samples with OD260/280 1.8–2.2 and RNA Integrity Number ≥ 8.0 were used to construct the sequencing library. An Illumina NovaSeq 6000 platform, with 2 × 150-bp pair-end reads (Illumina, San Diego, CA, United States) was used to quantify and sequence the final cDNA libraries. Clean reads were obtained from the screening of raw reads. Download reference genome sequences and gene model annotation files from ENSEMBL. Hisat2 (v2.0.1) was used to index reference genome sequences. Finally, clean reads were aligned to the reference genome via the software Hisat2 (v2.0.1). Differential expression analysis was conducted with the DESeq2 (V1.6.3) Bioconductor package. The false discovery rate was controlled through Benjamini and Hochberg’s approach by adjusting the resulting P value. Genes both with an adjusted P value < 0.05, and fold change value > 2 were confirmed as DEGs.

Image acquisition and statistical analysis

Zebrafish embryos were anesthetized with tricaine (Sigma, Saint Louis, Missouri, United States) at 48 and 72 hpf, embedded in 0.8% low melting point agarose, and observed by Leica TCS-SP5 LSM confocal imaging. Statistical analysis was performed using SPSS version 21.0 software (SPSS, Armonk, NY, United States). Student’s t-test or one-way analysis of variance was used to compare the data. P < 0.05 was considered statistically significant.

Results

nexmifb is expressed in the spinal cord of zebrafish

To detect the expression of nexmifb in zebrafish, WISH was performed at 72 hpf. The results showed nexmifb mRNA was mainly expressed in the brain and spinal cord (Figures 1(a) and 1(b)).

Knockdown of nexmifa or nexmifb did not affect expression of the other paralog

In zebrafish, nexmif has two paralogs, nexmifa and nexmifb. We perform qRT-PCR to determine whether nexmifa or nexmifb knockdown affected of the other paralog. The result showed that there was no significant difference in the expression level of nexmifb mRNA between control and nexmifa morphants at 72 hpf (Figure 2(a)). There was also no significant difference in the expression of nexmifa between the control and nexmifb morphants (Figure 2(b)), which indicated that knockdown of nexmifa or nexmifb did not affect expression of the other paralog.

Knockdown of nexmifb affected the body length of zebrafish

We found that no significant malformations were observed in the external appearance of nexmifb knockdown zebrafish embryos. However, at 72 hpf, the body length of zebrafish in the nexmifb morphants was significantly shorter than that in the control group (P < 0.05) (Figures 3(a) and 3(b)).

Knockdown of nexmifb affected the development of spinal motor neurons

We investigated the function of nexmifb in zebrafish embryonic development by observing the morphology of PMNs in Tg (mnx1: GFP)^ml2 transgenic zebrafish line by confocal microscopy between control and nexmifb morphants. We found knockdown of nexmifb caused obvious developmental defects in spinal motor neurons at 72 hpf. Firstly, knockdown of nexmifb led to the loss of spinal motor neurons (Figure 4(a)). We divided defective zebrafish embryos into three groups according to the percentage of motor neurons lost: severe group: loss of > 80% of motor neurons; moderate group: loss of 20%–80% of motor neurons; and normal group: loss of < 20% of motor neurons. Results showed that 4.2% of the control group had moderate defects and 95.8% were normal. However, 24.3% of the nexmifb morphants were normal, 32.2% had moderate defects, and 43.5% had severe defects (Figure 4(b)). To explore whether the loss of motor neurons was caused by apoptosis, the TUNEL BrightRed Apoptosis Detection Kit was used. We found there was almost no apoptotic signal in the control group. Although the nexmifb morphants showed apoptotic signals, they were not colocated with GFP-labeled motor neurons (Figure 4(c)).

Secondly, we also found that some motor neurons may have no loss of motor neurons, but the morphology of CaPs was abnormal. CaP axons were shorter and could not reach the ventral musculature (Figure 5(a)), which indicated significant developmental retardation in the nexmifb morphants. Statistical analysis showed that, the branches were disordered, the number of branches per mm of CaP was 77 ± 12 in nexmifb morphants, compared with 175 ± 15 in the controls (P < 0.05) (Figure 5(b)). The length of CaPs was shorter than that of the controls (190.1 ± 16.7 µm vs. 102.2 ± 20.1 µm) (P < 0.05) (Figure 5(c)).

Transcriptomic profiling of nexmifb morphants and control zebrafish

To elucidate the mechanisms underlying the effects of nexmifb on motor neuron development, we extracted RNA from control, nexmifb morphants and nexmifa morphants at 72 hpf and performed RNA-Seq. There were 4119 different expression genes (DEGs) between the control and nexmifb morphants (Figure 6(a)). Many DEGs were related to the CNS and can be classified into axon guidance, synapses and so on (Figure 6(b)). Because of knockdown of nexmifb can result in a similar abnormal phenotype to that of nexmifa morphants; therefore, we investigated whether there was the same mechanism of morphogenesis regulation during spinal motor neuron development between nexmifb and nexmifa. We compared the DEGs between the control and nexmifa morphants and between the control and nexmifb morphants, then found there were different and common DEGs between them. There were 1665 common DEGs, among which, 825 were common down-regulated and 288 were common up-regulated (Figure 6(c)). Among the common down-regulated DEGs, 21 were enriched in the axon guidance pathway (Figure 6(d)), which was related to the abnormal phenotype induced by nexmif deficiency. These results indicate that nexmifa and nexmifb may have similar mechanisms in regulation the development of spinal motor neuron. Using the change fold as our criterion, we used qRT-PCR to verify the expression of efna5b, which showed the greatest change among the 21 common down-regulated DEGs in nexmifb morphants compared to controls at 72 hpf. (Figure 6(e) and 6(f)). We also found efna5b was expressed in spinal cord of zebrafish by WISH (Figure 6(g)).

efna5b overexpression rescued motor neuron defects in nexmifb-deficient embryos

To investigate the molecular mechanism by which nexmifb regulates spinal motor neuron development in zebrafish, we selected efna5b, which had the largest fold change. Then we synthesized efna5b mRNA in vitro and coinjected them with nexmifb MOs into the yolk of one-cell embryos to explore whether nexmifb regulates the development of spinal motor neurons through the down-regulate expression of efna5b. We observed at 72 hpf, 6.9% of the control embryos had moderate defects and 93.1% were normal, 39.2% of the nexmifb morphants had severe defects, 32.2% had moderate defects, and 28.6% were normal. After overexpression of efna5b, normal were increased to 68.5%, severe and moderate defects were decreased to 20.8% and 10.7%, respectively (Figure 7(a) and 7(b)).

The abnormal of CaPs length and the number of branches were also rescued (Figure 7(c) and 7(d)). The lengths of CaPs in the nexmifb morphants was 121.4 ± 27.3 μm, which was significantly less than that in the control group (202.3 ± 18.9 μm). However, the lengths of CaPs dramatically increased to 183.0 ± 33.5 μm when the morphants were coinjected with efna5b mRNA (Figure 7(c)). The number of branches per mm of CaPs in the nexmifb morphants was less than that in the control group (85 ± 20 vs. 181 ± 9). However, the number of branches dramatically increased to 120 ± 28 when the morphants were coinjected with efna5b mRNA (Figure 7(d)).

Discussion

Previous studies have indicated that NEXMIF plays an important role in the development of CNS, especially in the brain. However, the relationship between nexmifb and the spinal cord remains unclear. In this study, we provide new insights into the role of nexmifb in spinal motor neuron development by examining the expression of nexmifb, established nexmifb knockdown model, which may provide therapeutic targets for motor neuron diseases.

WISH and RT-PCR indicated that nexmifb was mainly expressed in brain and spinal motor neurons which indicates that nexmifb may regulate the development of motor neurons directly. Therefore, we established a nexmifb knockdown model to characterize PMN morphology. nexmif has two paralogs in zebrafish. In order to illustrates whether there was a compensatory effect between nexmifa and nexmifb, we detected expression of nexmifb in nexmifa morphants and expression of nexmifa in nexmifb morphants by qRT-PCR and finally draw a conclusion that there was no compensation between the two paralogous genes.

We used confocal laser imaging to investigate the requirement of nexmifb in the formation of PMNs. During PMNs development, the three kinds of motor neurons extend their axons along stereotypical pathways and branches invade the myotome to form neuromuscular synapses. At 48 hpf, the axons of CaP extend to the middle of the segment, forming a collateral at the horizontal septum. At the ventral edge of the musculature, each axon turns dorsally and laterally grows along the rostral myoseptum[13]. When the embryo develops to 72 hpf, the formation of branches increases and they invade the myotome to form neuromuscular synapses[18,19]. Our data show that knockdown of nexmifb results in “CaPs and/or MiPs loss” as well as abnormal PMNs morphology. These results suggest that nexmifb is pivotal for the development of PMNs.

The types of cell death mainly include apoptosis, necrosis, and autophagy. RNA-Seq showed some DEGs associated with apoptosis between control and nexmifb morphants (NCBI BioProject PRJNA1157466); therefore, we speculated whether the loss of spinal motor neurons was caused by apoptosis. We found the number of apoptotic cells in nexmifb morphants, but the TUNEL signal was not co-located with motor neurons. These results suggest that the knockdown of nexmifb may lead to apoptosis of other cells rather than spinal motor neurons and that the loss of motor neurons may result from other modes of death. However, maybe there is another explanation that the “loss of motor neurons” is just the lack of axonal projection rather than the loss of cell bodies and axonal projection due to apoptosis, and all these require further investigation.

The development of spinal cord motor neurons begins at the embryonic stage and continues to the early postnatal stage. During this period, motor neurons are regulated by a variety of genes, such as transcription factors, channels, receptors, collagen, kinases, chromatin regulatory genes, etc[20]. For example, insm1a regulates spinal motor neuron development through olig2 and nkx6.1[21]. Although its expression was weak in the spinal cord, deficiency of sema6D induced by injection of sema6D-MOs caused dramatic developmental defects in PMNs, such as decreased length of CaPs, decreased number of branches, and the misdirected axonal trajectories[22]. Our RNA-Seq data showed that there were many DEGs between the controls and nexmifb morphants involved in neural development, synapses and axon guidance. In addition, compared with the RNA-Seq data between the controls and nexmifa morphants, there were many common down-regulated DEGs including efna5b which enriched in axon guidance pathway according to KEGG.

Efna5b, also known as ephrin-A5b, belongs to the Ephrins/Eph family of axon guidance molecules and represents a crucial class of axon guidance molecules[23,24]. The Ephrins/Eph family exhibits bidirectional regulatory functions; for instance, it can act as a ligand for Eph receptors and also function as a receptor for Eph ligands[25]. In Caenorhabditis elegans, ephrin EFN-4 promotes primary neurite outgrowth in AIY interneurons and class D motor neurons[26]. Increased expression of ephrinA5 in the retina enhances TrkB signaling, thereby promoting axon branching[27]. Poopalasundaram et al.[25] suggested that ephrinA6 is involved in retinal axon guidance and branching functions. They found that downregulation of ephrinA6 significantly reduces BDNF-induced axon branching in the chick retina and that ephrinA is critical for p75NTR-dependent axon guidance repulsion as well as TrkB-mediated branching. Flanagan and Vanderhaeghen[28] discovered that, in cultured retinal ganglion cells in vitro, ephrin A5 induces growth cone collapse, whereas in in vivo experiments, ephrin A5 promotes axon growth.

In our study, efna5b was expressed in the spinal cord, and its expression was dramatically down-regulated in nexmifb knockdown embryos compared with controls. Overexpression of efna5b partly rescued the abnormal PMNs phenotype. These data showed that nexmifb can affect spinal PMNs development by regulating the expression of efna5b.

In summary, our study first demonstrated that nexmifb was not only located in the brain but also in the spinal motor neurons. It plays a crucial role during motor neuron development, at least in part, through the down-regulation of efna5b. This work will deepen a comprehensive understanding of nexmifb functions and provide potential therapeutic targets for motor neuron disease.

References

[1] Cantagrel V, Lossi AM, Boulanger S, et al. Disruption of a new x linked gene highly expressed in brain in a family with two mentally retarded males. J Med Genet. 2004;41(10):736-742.

[2] Wang L, Li Z, Song X, et al. Bioinformatic analysis of genes and micrornas associated with atrioventricular septal defect in down syndrome patients. Int Heart J. 2016;57(4):490-495.

[3] Stekelenburg C, Blouin JL, Santoni F, et al. Loss of nexmif results in the expression of phenotypic variability and loss of genomic integrity. Sci Rep. 2022;12(1):13815.

[4] Cantagrel V, Haddad MR, Ciofi P, et al. Spatiotemporal expression in mouse brain of kiaa2022, a gene disrupted in two patients with severe mental retardation. Gene Expr Patterns. 2009;9(6):423-429.

[5] Zhong L, Liu C, Lin L. Infantile spasms caused by nexmif mutation: a case report and literature review. Appl Neuropsychol Child. 2023;12(4):380-385.

[6] Wang L, Huang Y, Liu X. Nexmif pathogenic variant in a female child with epilepsy and multiple organ failure: a case report. Transl Pediatr. 2023;12(6):1278-1287.

[7] Langley E, Farach LS, Koenig MK, et al. Nexmif pathogenic variants in individuals of korean, vietnamese, and mexican descent. Am J Med Genet A. 2022;188(6):1688-1692.

[8] Gilbert J, Man HY. The x-linked autism protein kiaa2022/kidlia regulates neurite outgrowth via n-cadherin and delta-catenin signaling. Eneuro. 2016;3(5):ENEURO.0238-16.2016.

[9] Magome T, Hattori T, Taniguchi M, et al. Xlmr protein related to neurite extension (xpn/kiaa2022) regulates cell-cell and cell-matrix adhesion and migration. Neurochem Int. 2013;63(6):561-569.

[10] Gilbert J, O’Connor M, Templet S, et al. Nexmif/kidlia knock-out mouse demonstrates autism-like behaviors, memory deficits, and impairments in synapse formation and function. J Neurosci. 2020;40(1):237-254.

[11] Purifoy EJ, Mruk K. Differential roles of diet on development and spinal cord regeneration in larval zebrafish. Zebrafish. 2024;21(2):214-222.

[12] Myers PZ. Spinal motoneurons of the larval zebrafish. J Comp Neurol.1985;236(4):555-561.

[13] Myers PZ, Eisen JS, Westerfield M. Development and axonal outgrowth of identified motoneurons in the zebrafish. J Neurosci. 1986;6(8):2278-2289.

[14] Westerfield M, McMurray JV, Eisen JS. Identified motoneurons and their innervation of axial muscles in the zebrafish. J Neurosci. 1986;6(8):2267-2277.

[15] Eisen JS, Myers PZ, Westerfield M. Pathway selection by growth cones of identified motoneurones in live zebra fish embryos. Nature.1986;320(6059):269-271.

[16] Beattie CE, Granato M, Kuwada JY. Cellular, genetic and molecular mechanisms of axonal guidance in the zebrafish. Results Probl Cell Differ. 2002;40:252-269.

[17] Zheng YQ, Suo GH, Liu D, et al. Nexmifa regulates axon morphogenesis in motor neurons in zebrafish. Front Mol Neurosci. 2022;15:848257.

[18] Liu DW, Westerfield M. The formation of terminal fields in the absence of competitive interactions among primary motoneurons in the zebrafish. J Neurosci. 1990;10(12):3947-3959.

[19] Downes GB, Granato M. Acetylcholinesterase function is dispensable for sensory neurite growth but is critical for neuromuscular synapse stability. Dev Biol. 2004;270(1):232-245.

[20] Catela C, Kratsios P. Transcriptional mechanisms of motor neuron development in vertebrates and invertebrates. Dev Biol. 2021;475:193-204.

[21] Gong J, Wang X, Zhu C, et al. Insm1a regulates motor neuron development in zebrafish. Front Mol Neurosci. 2017;10:274.

[22] Sheng J, Xu J, Geng K, et al. Sema6d regulates zebrafish vascular patterning and motor neuronal axon growth in spinal cord. Front Mol Neurosci. 2022;15:854556.

[23] Cayuso J, Xu Q, Wilkinson DG. Mechanisms of boundary formation by eph receptor and ephrin signaling. Dev Biol. 2015;401(1):122-131.

[24] Lisabeth EM, Falivelli G, Pasquale EB. Eph receptor signaling and ephrins. Cold Spring Harb Perspect Biol. 2013;5(9):a009159.

[25] Poopalasundaram S, Marler KJ, Drescher U. Ephrina6 on chick retinal axons is a key component for p75(ntr)-dependent axon repulsion and trkb-dependent axon branching. Mol Cell Neurosci. 2011;47(2):131-136.

[26] Schwieterman AA, Steves AN, Yee V, et al. The caenorhabditis elegans ephrin efn-4 functions non-cell autonomously with heparan sulfate proteoglycans to promote axon outgrowth and branching. Genetics. 2016;202(2):639-660.

[27] Marler KJ, Becker-Barroso E, Martinez A, et al. A trkb/ephrina interaction controls retinal axon branching and synaptogenesis. J Neurosci. 2008;28(48):12700-12712.

[28] Flanagan JG, Vanderhaeghen P. The ephrins and eph receptors in neural development. Annu Rev Neurosci. 1998;21:309-345.

RIGHTS & PERMISSIONS

The Author(s) 2026. This article is published by Higher Education Press at journal.hep.com.cn.