1 Introduction
Axon damage interrupts neuronal communications and signal transmission between neurons and target cells, disrupts the physiologic functions of neurons, results in partial or complete loss of motor, sensory, and autonomic functions, and normally induces neuropathological disorders [
1,
2]. Successful axon growth and regeneration is fundamental for the re-establishment of neural circuits and the recovery of impaired nerve functions. In many invertebrates and non-mammalian vertebrates, neurons can regrow their axons after nerve injury and achieve functional recovery [
3]. The intrinsic growth ability of injured neurons is largely compromised in mammals [
4]. In the adult mammalian central nervous system, injured axons are incapable of regenerating, leading to failure of the restoration of normal nerve function [
5–
7]. On the contrary, in immature mammals, injured central nerves are able to regrow following axon injury, indicating that adult mammals process certain inherent regeneration potential [
8]. More importantly, except for immature mammalian central nerves, mammalian peripheral nerves have a regenerative ability following axon injury, albeit their regeneration is limited and the regeneration outcome is often incomplete, especially after severe axon injury such as nerve transection with long distance [
9–
11]. Exploring the intrinsic factors underpinning successful axon regrowth following peripheral nerve injury as well as the unsuccessful axon regrowth following central nerve injury is therefore essential for the understanding of the molecular mechanisms that controls the neuronal intrinsic regeneration capacity [
10].
Sensory neurons in dorsal root ganglion (DRG) are pseudounipolar neurons with peripheral axonal branches that proceed along peripheral nerves and central axonal branches that proceed along the dorsal roots into the spinal cord [
12,
13]. Peripheral and central axonal branches behave differently after axon injury. The regeneration rate of injured dorsal roots is only half that of injured peripheral axonal branches and the regrowth of injured dorsal roots stops at the dorsal root entry zone [
14]. DRG neurons are thus commonly used to discriminate injury responses in the same neuronal soma following damage to the regenerative peripheral axonal branches or the non-regenerative central axonal branches [
14–
16]. For instance, the comparison of the levels of reactive oxygen species after sciatic nerve crush or dorsal column crush demonstrates that the production of reactive oxygen species is increased in DRGs after peripheral axon injury but not altered after central axon injury [
17]. Elevated reactive oxygen species promotes the outgrowth of axons and the recovery of nerve functions, indicating that assessing diverse molecular changes after peripheral or central axon injury is critical for the identification of regeneration-associated molecules [
17]. Likewise, calcium activation of protein phosphatase 4 and protein phosphatase 4-dependent de-phosphorylation of histone deacetylase 3, a process that inhibits the activity of histone deacetylase 3, is only detected after peripheral axon injury but not observed after central axon injury [
17]. Histone deacetylase 3 inhibition modulates several regenerative pathways, activates the regenerative program, benefits neurite outgrowth, and is even capable of promoting the regeneration of sensory axons following spinal cord injury [
17]. Deeper understanding of differentially expressed molecules after the regenerative peripheral axon injury and the non-regenerative central axon injury of DRG neurons is hence crucial for revealing novel regeneration associated factors and developing successful therapeutics.
Taking advantage of high throughput data from sequencing, we explored gene expression patterns in rat DRGs after sciatic nerve or dorsal root injury, determined differentially expressed genes after axotomy relative to their corresponding sham controls, and screened a list of genes whose expressions were significantly decreased after sciatic nerve injury but exhibited an increased trend after dorsal root injury, including the dysbindin domain containing 2 (DBNDD2). DBNDD2, also known as casein kinase-1 binding protein, is a dysbindin protein family member involved in the negative regulation of casein kinase-1 activity [
18]. A recent study demonstrates that by binding to casein kinase-1 and inhibiting casein kinase-1 activity, DBNDD2 reduces the amounts of total and insoluble α-synuclein and thus may be conducive to the treatment of Parkinson’s disease [
19]. However, the biological effects of DBNDD2 on axon regeneration during nerve injury and regeneration process remain to be explored.
In the current study, we found that in adult rats, DBNDD2 expression is reduced after peripheral axon injury but increased after central axon injury and thus predicted that DBNDD2 might be a negative regulator of axon regeneration. Indeed, we showed that DBNDD2 knockdown in neurons facilitates neurite growth and axon elongation. Our results uncover the functional roles of DBNDD2 in axon regeneration and imply that drugs targeting DBNDD2 may be of benefit in treating traumatic nerve injury as well as other axon damage-associated neurological disorders such as stroke and glaucoma.
2 Materials and methods
2.1 Animal surgery
Specific pathogen-free (SPF) Sprague-Dawley (SD) rats used in this study were purchased from the Animal Center of Nantong University (animal licenses Nos. SCXK [Su] 2014-0001 and SYXK [Su] 2012-0031). All animal experimental procedures were approved by the Ethics Committees of Experimental Animals, Jiangsu Province, China (approval ID: S20231219-041) and were conducted in strict compliance with the guidelines of Nantong University Institutional Animal Care.
The rats were housed in groups of 4–5 per cage in environmentally controlled rooms maintained at a constant temperature and humidity (temperature 20–26 °C, humidity 40%–60%), with a 12-h light/dark cycle. All rats were fed with feed provided by the animal center and sterilized pure water. The experimental operators adhered to the “3R” principles (replacement, reduction, and refine), and followed the five principles of freedom for experimental animals. All operators received training in the basic knowledge and handling skills of experimental animals, and strictly followed standardized protocols for animal care and post-experimental procedures. Humane endpoints were implemented to ensure animal welfare.
Male adult SD rats (8-week-old, weighting 180–220 g) were randomly subjected to sciatic nerve axotomy-induced peripheral axon injury or dorsal root axotomy-induced central axon injury, as previously described with modifications [
14,
20]. Briefly, for sciatic nerve axotomy, after anesthetization, a skin incision on the lateral aspect of the mid-thigh of rat hind limb was made and exposed rat sciatic nerves were subjected to a sharp axotomy. For dorsal root axotomy, a midline incision at the lumbar (L)2–L3 vertebral level was made, the dura mater was removed, and exposed rat L4 and L5 dorsal roots were subjected to axotomy. Rats that underwent sciatic nerve or dorsal root exposure without axotomy were designated as sham-operated. L4 and L5 DRGs were collected 1 day after axotomy or sham surgery and subjected to RNA sequencing.
2.2 Sequencing
RNA sequencing of L4 and L5 rat DRGs following sciatic nerve injury versus sciatic nerve sham surgery and dorsal root injury versus dorsal root sham surgery has been published with sequencing data stored in Genome Sequence Archive database (accession number CRA006070) [
21]. RNA sequencing was performed on an HiSeq
TM 4000 by Genedenovo Biotechnology Co., Ltd. (Guangzhou, Guangdong, China). Transcripts abundances were quantified using StringTie and differential expression testing was performed using edgeR [
22].
Single-cell sequencing of 1-day-old neonatal and 8-week-old adult rat DRGs was performed as previously described with sequencing data deposited in the NCBI database (accession number GSE147615) [
23]. Digested single cell suspensions were loaded on the 10x Chromium system, libraries were prepared using 10x Genomics GemCode Single-Cell 3′ Gel Bead and Library Kit, and sequencing was conducted using Illumina NovaSeq platform by NovelBioinformatics Ltd., Co. (Shanghai, China). Raw data were processed by fastp quality control and analyzed with Cell Ranger (v3.0.0) for barcode identification, mapping, and gene counting. Transcript abundances were quantified after normalization. Batch effects were corrected using the Mutual Nearest Neighbor (MNN) algorithm. Sequencing data were categorized into clusters using the Seurat 3.1 software package. Single cell sequencing data were visualized in a t-distributed stochastic neighbor embedding (tSNE) plot using a dimensional reduction algorithm.
2.3 Quantitative real-time polymerase chain reaction (RT-PCR)
Total RNA was extracted from collected DRG tissues or cultured DRG neurons using the RNA-Quick Purification Kit (Yeasen Biotechnology Co., Beijing, China) or Cell RNA Extraction Kit (UU-Bio Technology Co., Suzhou, Jiangsu, China), respectively, and then treated with amplification-grade DNase I (Thermo Fisher Scientific). RNA quantification was performed using a Nanodrop 1000 spectrophotometer (NanoDrop Technologies, Wilmington, Delaware, USA), total RNA was reverse transcribed into cDNA using HiScript®IIQ RT SuperMix for qPCR (Vazyme, Nanjing, Jiangsu, China) according to the manufacturer’s instructions. Quantitative RT-PCR was then performed using ChamQTM SYBR® qPCR Master Mix (Vazyme) on an ABI StepOne system (Applied Biosystems, Foster City, CA, USA). Experiments were repeated in triplicate. The Ct values of the target gene DBNDD2 were compared with those of the internal control gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The relative abundance of DBNDD2 was quantified using the comparative 2-ΔΔCt method. Primers were synthesized by Sangon Biotech (Shanghai, China). The specific primer sequences for the target gene DBNDD2 were DBNDD2 (forward) 5′-CGTCAGACAGGACCACATCC-3′ and DBNDD2 (reverse) 5′-TGTCTCCTCCCCCATCACTT-3′ while the sequences of specific primers for reference gene GAPDH were GAPDH (forward) 5′-ACAGCAACAGGGTGGTGGAC-3′ and GAPDH (reverse) 5′-TTTGAGGTGCAGCGAACTT-3′.
2.4 Immunohistochemistry
The L4 and L5 DRGs from rats were washed with PBS, fixed in 4% paraformaldehyde and cryoprotected in 30% sucrose. Rat DRGs were then embedded in O.C.T., frozen, and cut to tissue sections. DRG sections were incubated with anti-Tuj1 (1:1000; Abcam, catalog # ab18207, Cambridge, Massachusetts, USA), anti-DBNDD2 (1:200; Proteintech, catalog # 27623-1-AP, Rosemont, Illinois, USA), anti-enhanced green fluorescent protein (EGFP; 1:100; Abcam, catalog # ab184601), or anti-superior cervical ganglion-10 protein (SCG10; 1:500; Novus Biologicals, catalog # NBP1-49461, Littleton, Colorado, USA) primary antibodies at 4 °C overnight and followed by Alexa Fluor 488-conjugated donkey anti-mouse IgG (1:500; Proteintech, catalog # SA00013-5), Alexa Fluor 488-conjugated donkey anti-rabbit IgG (1:500; Proteintech, catalog # SA00013-6), Cy3 goat anti-mouse IgG (H + L) (1:500; Proteintech, catalog # SA00009-1), or Cy3 goat anti-rabbit IgG (H + L) (1:500; Proteintech, catalog # SA00009-2) secondary antibodies. Nuclei were counterstained with DAPI Fluoromount-G stain (SouthernBiotech, catalog # 0100-20, Birmingham, Alabama, USA). Immunofluorescence images were captured on a Zeiss Axio Imager M2 microscope (Jena, Germany). Exposure time and gain were maintained constant for each fluorescence channel during image capture.
2.5 Primary DRG neuron isolation and culture
DRGs collected from 1-day-old neonatal and 8-week-old adult rats were dissected into small pieces and subjected to tissue digestion. For neonatal rats, dissected DRGs were digested with 3 mg/mL collagenase I for 30 min and 0.25% trypsin for 20 min, while for adult rats, DRGs were digested with 3 mg/mL collagenase I for 90 min and 0.25% trypsin for 5 min. After adding complete culture medium containing 10% fetal bovine serum albumin (Sigma, St. Louis, MO, USA) to terminate digestion, cells were filtered through a 70-μm cell strainer. Cell pellets were re-suspended in 15% BSA Albumin Bovine V (BioFroxx, Einhausen, Germany) and then subjected to centrifugation. Separated neonatal or adult rat DRG neurons were cultured in Neurobasal medium (Gibco, Grand Island, New York, USA) containing 2% B27 supplement (Gibco) and 2 mmol/L L-glutamine (ThermoFisher Scientific, Waltham, MA, USA) and plated onto cell culture dishes pre-coated with poly-L-lysine (PLL).
2.6 DRG neuron transfection
Primary cultured neonatal or adult rat DRG neurons were transfected with small interfering RNA (siRNA) targeting DBNDD2 to knock down DBNDD2 expression in neurons. DRG neurons were transfected with siRNAs targeting DBNDD2 (si-DBNDD2) or a control scrambled siRNA with a random sequence using Lipofectamine RNAiMAX transfection reagent (Invitrogen) according to the manufacturer’s instructions. The sequences of siRNAs targeting DBNDD2 were as follows: DBNDD2-siRNA-1, 5′-GAAGTTCTTCGAGGACATT-3′; DBNDD2-siRNA-2, 5′-GGTGGAATTTATTGACCTT-3′; and DBNDD2-siRNA-3, 5′-GCAGTCCAAATCCAAGTGA-3′. The sequence of the control siRNA was 5′-GGCUCUAGAAAAGCCUAUGC-3′. The siRNA sequences were synthesized by RibiBio Biotechnology Co., Ltd. (Guangzhou, Guangdong, China).
2.7 In vitro neurite growth assay
Cultured neonatal rat DRG neurons were fixed with 4% paraformaldehyde 36 h after transfection and subjected to anti-Tuj1 immunostaining. Adult rat DRG neurons were dissociated using trypsin treatment and seeded onto glass coverslips pre-coated with PLL. At 24 h after cell culture, neurons were washed with PBS, fixed with 4% paraformaldehyde, and immunostained with anti-Tuj1 antibody. The longest and total lengths of neurites from each neonatal or adult rat DRG neuron were measured and quantified using the ImageJ software. The images containing neurites were imported into ImageJ and converted from color to 8-bit format. Since the image has a known scale, we convert the measurements to micrometers (1 pixel = 4.4 µm), and this scale was uniformly applied to all images. To trace and measure neurites, we clicked “Plugins,” selected “Analyze” from the drop-down menu, and clicked “Neuron J,” clicked “Add tracings” to trace each neurite of an individual neuron, and then click “Measure tracings” to automatically measure and record the total and longest neurite lengths.
For myelin experiments, myelin fractions were extracted from adult rat brain tissues as previously described with modifications [
24]. Briefly, adult rat brain tissues were homogenized in 0.30 mol/L sucrose, layered over 0.83 mol/L sucrose, and centrifuged to gather the crude myelin layers between the interfaces. The process was repeated to purify collected myelin extracts and the glass coverslips were coated with PLL and myelin extracts.
2.8 In vitro neurite regeneration assay
Adult rat DRG neurons were cultured onto the somal compartments of microfluidic chambers pre-coated with PLL (catalog # SND150, Xona 2-compartment SND 150, Xona Microfluidics LLC). Neurites that had entered the axonal compartments after cell culture were dissected and removed using an 80 kPa (600 mmHg) vacuum suction three times for 20 s each. Dissected neurites were cultured for an additional 24 h, fixed with 4% paraformaldehyde, and immunostained with the anti-Tuj1 antibody to observe neurite elongation and regeneration. The lengths of regenerated neurites were measured and quantified using ImageJ software.
2.9 Intrathecal injection of adeno-associated virus (AAVs)
Adult SD rats were anesthetized, shaved to expose the skin around the lumbar region, and injected with AAV that carry shRNA against DBNDD2 (pAAV-U6-shRNA(DBNDD2)-CMV-EGFP-WPRE) or a control AAV (pAAV-U6-shRNA(NC)-CMV-EGFP-WPRE). The AAVs were packaged by OBiO Biotechnology Co., Ltd. (Shanghai, China). A total of 10 μL of AAV solution was slowly injected into the cerebrospinal fluid between vertebrae L4 and L5 using a 25 μL Hamilton syringe and the needle was left in place for additional 2 min. After leaving rats injected with AAVs to recover for 21 days, the left sciatic nerve at 10 mm above the bifurcation into the tibial and common fibular nerves was crushed with forceps as previously described [
25]. At 3 days after sciatic nerve crush injury, sciatic nerve tissues were collected and then subjected to SCG10 immunostaining. The lengths of regenerated nerves were measured and quantified using ImageJ software.
2.10 Bioinformatic analysis
Molecules that interact with DBNDD2 were discovered and visualized using the STRING data resource [
26]. Molecules that interact with DBNDD2 or control the expression of DBNDD2 were determined using the Pathway Commons [
27]. Upstream transcription factors of DBNDD2 were predicted using JASPAR database [
28], animalTFDB 3.0 database [
29], Gene Transcription Regulation database (GTRD) [
30], and hTFtarget database [
31]. The intersection of potential upstream transcription factors of DBNDD2 was obtained using Venny. The binding of transcription factor estrogen receptor 1 (ESR1) to the promoter region of DBNDD2 was generated using motif-based sequence analysis tool FIMO-MEME Suite [
32].
2.11 Statistical analysis
All quantitative data were presented as the mean ± the standard error of the mean (SEM). Sample size was estimated based on a comparable study [
20] and all averaged numerical data contained a minimum of three biological replicates. The numbers of independent experiments were indicated in the figure legends. Unpaired two-tailed Student’s
t-test or one-way ANOVA followed by Dunnett’s or Tukey’s multiple comparison
post hoc test was performed with GraphPad Prism, and significance was set at a
P value < 0.05.
3 Results
3.1 DBNDD2 is differentially expressed in adult rat DRGs following nerve injury
We analyzed sequencing data from rat DRGs following sciatic nerve or dorsal root axotomy. The expression levels of DBNDD2 were examined after peripheral and central axon injuries. 24 h after sciatic nerve injury, the expression of DBNDD2 gene in rat DRGs decreased by more than 65% compared to sciatic nerve sham injured animals. Following dorsal root injury, the expression of DBNDD2 gene in rat DRGs increased 1.85-fold compared to baseline levels in sciatic dorsal root injured animals (Fig.1). Consistent with sequencing data, RT-PCR results showed a decrease in DBNDD2 gene expression in the DRGs after sciatic nerve injury, but a significant increase after dorsal root injury (Fig.1). DBNDD2 protein expression in the DRGs was also found to be affected by axonal injury. The fluorescence intensity of DBNDD2 in the DRGs seemed to be much lower after sciatic nerve injury compared to the uninjured animals, while it seemed to be higher following dorsal root injury (Fig.1). These observations from RNA sequencing, RT-PCR, and immunohistochemistry immunostaining fully demonstrate that in the DRGs, DBNDD2 presented an opposite following injury to the peripheral and central axonal branches. In addition, DBNDD2 signal was found to co-localized with those of a neuronal cell marker Tuj1 (Fig.1). These observations demonstrate that injury to peripheral and central axonal branches of the DRGs elicits different changes of DBNDD2 in DRG neuronal somas and hints that differentially expressed DBNDD2 contributes to the different regeneration capacity of injured peripheral and central axons.
3.2 DBNDD2 is differentially expressed in the DRGs at different developmental stages
We next compared the expression levels of the DBNDD2 gene in DRG neurons between neonatal and adult rats — two developmental stages with distinct regeneration capacities. Single-cell sequencing, a high-throughput analysis tool for identifying cellular heterogeneity and separating cell populations, was used. Neurons were distinguished from other cell types in the DRGs, and DRG neurons from neonatal and adult rats were identified (Fig.2). The tSNE plot showed that DBNDD2 gene was expressed in DRG neurons of both neonatal and adult rats and was present in a larger number of DRG neurons in adult rats as compared with in neonatal rats (Fig.2). Quantification of DBNDD2 expression in DRG neurons revealed significantly higher levels in adult rats compared to neonatal rats (Fig.2). Besides 1-day-old neonatal and 8-week-old adult rats, the temporal expression pattern of DBNDD2 gene in 2-week-old and 4-week-old rats were also determined. Similar to 8-week-old adult rats, elevated expression of DBNDD2 gene was detected in the DRGs of 2-week-old and 4-week-old rats as compared with the DRGs of neonatal rats (Fig.2). The observed elevation of DBNDD2 gene expression in developing animals with reduced regeneration capacity implies that DBNDD2 may be a negative regulator of axon regeneration.
3.3 DBNDD2 deficiency promotes the growth of neonatal rat DRG neurons
Primary neonatal rat DRG neurons were isolated, cultured, and gene manipulated to assess the role of DBNDD2 (Fig.3). RT-PCR showed that transfecting neonatal rat DRG neurons with three DBNDD2 siRNA fragments reduced DBNDD2 gene expression by over 80% (Fig.3). DBNDD2-siRNA-1 and DBNDD2-siRNA-2, which showed relatively higher knockdown efficiency, were subsequently used for functional investigation. Following the transfection of DBNDD2-siRNA-1, the intrinsic axon growth ability of neonatal rat DRG neurons seemed to be activated and the neurites of neonatal neurons were noticeably much longer (Fig.3). Summarized data from over 400 DRG neurons showed that DBNDD2-siRNA-1 transfection increased total neurite lengths from 180.09 μm in the control-siRNA treated group to 247.44 μm in the DBNDD2-siRNA-1-siRNA transfected group. The longest neurites also increased from 113.68 μm in the control-siRNA treated group to 142.20 μm in the DBNDD2-siRNA-1 group (Fig.3). DBNDD2-siRNA-2 transfection induces comparable cellular responses, with the total neurite lengths increasing from 167.06 μm in the control-siRNA treated group to 255.98 μm, and the longest neurites increasing from 109.85 μm to 154.40 μm (Fig.3 and 3D). These data indicate that silencing DBNDD2 in neonatal DRG neurons is able to enhance neuronal sprouting and neurite growth.
3.4 DBNDD2 deficiency promotes the growth of adult rat DRG neurons and the elongation of injured axons
Next, we explored the effects of DBNDD2 knockdown on primary cultured adult DRG neurons. Given that adult neurons have diminished growth ability relative to neonatal rats, adult DRG neurons were subjected to a culture-and-replating protocol to recapitulate nerve injury process and to boost the intrinsic growth capacity of adult neurons (Fig.4). Transfection of adult DRG neurons with DBNDD2-siRNA-1 and DBNDD2-siRNA-2 decreased the expression of DBNDD2 gene to less than 50% (Fig.4). Similar as neonatal DRG neurons, in adult DRG neurons, decline expression of DBNDD2 significantly enhanced neurite extension (Fig.4 and 4D). The growth conditions of adult DRG neurons cultured on myelin inhibitory substrates were further investigated. For adult DRG neurons transfected with the control siRNA, the total and the longest neurites lengths were reduced by approximately 25% when cultured on plates coated with PLL and extracted myelin, compared to those on PLL-coated plates. Knockdown of DBNDD2 not only boosted the growth of neurons cultured on PLL-coated plates, but also supported the elongation of neurites when adult DRG neurons were cultured on myelin substrates (Fig.4 and 4F). These results show that myelin substrates considerably suppress neuronal growth while DBNDD2 silencing prominently augments neuronal growth on myelin substrates.
In addition to the culture-and-replating cell model, by using a two-compartment microfluidic chamber, the axons of adult DRG neurons were directly severed in vitro and the subsequent regrowth of injured neurites was visualized. Regenerated neurites were observed in the axonal compartments of control siRNA-transfected neurons. Adult DRG neurons transfected with DBNDD2-siRNA-1 had elongated neurites compared to the control siRNA group. Likewise, following the transfection of DBNDD2-siRNA-2, the average lengths of regenerated neurites seemed to be longer relative to the control group, although not significant (Fig.4 and 4H). These straightforward observations demonstrate that knockdown of DBNDD2 in DRG neurons advances the regrowth and elongation of injured neurites, indicating that DBNDD2 may function as an important molecular target for enhancing the intrinsic regeneration capacity of adult neurons.
3.5 DBNDD2 deficiency facilitates nerve regeneration after sciatic nerve injury
To determine the in vivo roles of DBNDD2 knockdown, an AAV-EGFP-DBNDD2 shRNA or a control AAV was intrathecally administered to adult rats (Fig.5). RT-PCR results showed that, 21 days after injection of AAV-DBNDD2-shRNA, the expression of DBNDD2 gene in the DRGs was much lower than in the DRGs of rats injected with the control AAV (Fig.5). Immunostaining images displayed the co-labeling of EGFP and neuronal marker Tuj1 in rat DRGs (Fig.5), as well as noticeably diminished immunofluorescence signals of DBNDD2 in Tuj1-labeled neurons (Fig.5), indicating that AAV delivery of DBNDD2-shRNA effectively silenced DBNDD2 expression in neurons. After confirming successful knockdown of DBNDD2 in DRG neurons, rats were subjected to sciatic nerve crush injury. We collected rat sciatic nerves 3 days after crush injury, labeled them with regenerating sensory axon marker SCG10, and observed that in rats injected with AAV expressing DBNDD2 shRNA, the length of regenerated axons increased by roughly 1.5-fold compared to rats injected with control AAV (Fig.5 and 5F). Animal study observations indicate that AAV delivery of shRNA against DBNDD2 promotes axon elongation and nerve regeneration.
3.6 Identification of DBNDD2-associated molecules
To determine molecules interacting with DBNDD2, the top 5 STRING interactants of DBNDD2, calculated based on experimental scores, were displayed in a STRING interaction network (Fig.6). These include casein kinase 1 isoform delta (CSNK1D), casein kinase 1 epsilon (CSNK1E), casein kinase 1 isoform gamma 2 (CSNK1G2), casein kinase 1 isoform gamma 3 (CSNK1G3), and secreted frizzled related protein 2 (SFRP2). DBNDD2 closely interacts with CSNK1D and CSNK1G2, while CSNK1D and CSNK1G2 exhibit combined confidence in their functional interactions with CSNK1E and CSNK1G3, respectively [
18]. A more comprehensive DBNDD2-centered molecular network was generated using the integrated Pathway Commons web resource (Fig.6). Besides the molecules identified in the STRING protein–protein network, DBNDD2 was found to interact with vasoactive intestinal peptide receptor 2 (VIRP2), angiotensin II receptor type 1 (AGTR1), lamin A/C (LMNA), and calcium binding and coiled-coil domain 2 (CALCOCO2). Furthermore, DBNDD2 is regulated by several factors, including zine finger E-box binding homeobox 1 (ZEB1), forkhead box F2 (FOXF2), transforming growth factor beta 1 (TGFB1), signal transducer and activator of transcription 5B (STAT5B), ESR1, myocyte enhancer factor 2A (MEF2A), MDS1 and EVI1 complex locus (MECOM), paired box 4 (PAX4), forkhead box O4 (FOXO4), transcription factor CP2 (TFCP2), androgen receptor (AR), and forkhead box O1 (FOXO1).
Next, we explored the expression levels of these DBNDD2-associated molecules in the DRGs after peripheral or central axon injury according to sequencing data. The heatmap displayed the relative fold changes of these DBNDD2-associated molecules, with genes exhibiting elevated expression compared to the corresponding sham control colored red, and genes with reduced expression colored green. The results demonstrated that although many of these DBNDD2-associated molecules did not show apparent expression changes after dorsal root axotomy-induced central axon injury, some molecules were differentially expressed in the DRGs after sciatic nerve axotomy-induced peripheral axon injury. For instance, VIPR2, AGTR1A, AGTR1B, ESR1, and MECOM displayed a similar expression trend to DBNDD2 and were downregulated in the DRGs after sciatic nerve injury. In contrast, SFRP2 and PAX4 exhibited an opposite expression pattern to DBNDD2 and were upregulated in the DRGs after sciatic nerve injury (Fig.6).
Given that many DBNDD2-associated molecules that exhibit similar expression trends as DBNDD2 are transcription factor-coding genes, we then predicted the potential upstream transcription factors of DBNDD2. A total of 23, 5330, 1490, and 268 transcription factors of DBNDD2 were screened using JASPAR, animalTFDB 3.0, GTRD, and hTFtarget databases. Three common elements, including CAMP responsive element binding protein 1 (CREB1), transcription factor specificity protein 1 (SP1), and ESR1, which controls DBNDD2 according to the Pathway Commons web resource, were identified at the intersection of these databases (Fig.6). Transcription factor ESR1 may directly bind to the putative binding site located at −1915 to −1898 in the promoter region of DBNDD2 and regulate DBNDD2 expression (Fig.6).
4 Discussion
The molecular functions of DBNDD2, a protein that is highly expressed in the nervous system, remain largely unclear, aside from its roles in casein kinase-1 binding and inhibiting [
19]. Herein, we examined the expression changes of DBNDD2 in rat DRGs during regeneration and development using sequencing, RT-PCR, and immunostaining. We also studied the biological effects of DBNDD2 on DRG neurons through siRNA transfection, and found that DBNDD2 knockdown is beneficial for axon growth and nerve regeneration.
Rats subjected to partial or complete nerve injury are widely applied as appropriate models to explore the pathological basis of nerve injury and assess novel medications and profound treatment strategies [
33]. DRG neurons process both peripheral and central axon branches with different regeneration capacities, making them valuable for investigating the intrinsic mechanisms underlying successful axon regeneration. The peripheral projecting axons of rat L4 and L5 DRGs, together with nerve fibers from spinal cord motor neurons, make up peripheral nerves with the largest diameters, that are sciatic nerves [
34]. Hence, surgeries such as crush, stretch, percussion, and transection to sciatic nerves induce injuries to peripheral axonal branches of L4 and L5 DRG neurons. The central axonal branches projected from L4 and L5 DRG neuronal somas extend along dorsal roots, enter the spinal cord via the dorsal root entry zone, and then bifurcate to ascending central axonal branches toward the brain and descending central axonal branches toward the cauda equine [
35,
36]. Compared with spinal cord injury that impairs the ascending central axonal branches only, injury at the dorsal root before the branch point impairs the whole central axonal branch and is considered as a well suited surgical model to study central axon regeneration [
35].
Consequently, in the current study, we first investigated sequencing data of rat L4 and L5 DRGs after sciatic nerve axotomy-induced peripheral axon injury and dorsal root axotomy-induced central axon injury, aiming to decipher the expression changes of DBNDD2 gene after injury to the peripheral and central axonal branches of DRG neurons, which have different regeneration capacities. A previous study that determines genetic changes in DRGs at numerous different time points, ranging from hours to weeks after injury after peripheral axotomy identifies a rapid emergence of neuronal transcriptional state and a rapid upregulation of regeneration-associated genes at 1 day after axotomy [
37]. Hence, we collected injured DRGs 1 day after sciatic nerve or dorsal root axotomy and subjected them to RNA sequencing. RNA bulk sequencing, together with consistent validation outcomes, showed that DBNDD2 gene expression in rat DRGs was substantially reduced 24 h after peripheral axon injury but elevated after central axon injury. The expression changes of DBNDD2 in rat DRGs at longer time points after peripheral or central axon injury have not been revealed. Moreover, it is worth noting that besides DRG neurons, there are numerous other cell types in rat DRGs whereas RNA bulk sequencing determines the global gene expression in tissues and organs without distinguishing the transcriptional heterogeneity of cell populations. To verify whether DBNDD2 expression is indeed altered in DRG neurons, we collected L4 and L5 DRGs from rats that underwent peripheral or central axon injury as well as uninjured rats, double immunostained rat DRGs with DBNDD2 and neuronal marker Tuj1 in frozen DRG specimens, and then determined the expression of DBNDD2 protein in Tuj1-labeled neurons. The fact that compared with uninjured rats, obviously weaker signals of DBNDD2 protein that are co-labeled with Tuj1 are observed in peripheral axon-injured rats indicates that reduced DBNDD2 expression in DRG neurons may contribute to enhanced axon regeneration. On the other hand, stronger signals of DBNDD2 protein that are co-labeled with Tuj1 observed in central axon-injured rats indicates that increased DBNDD2 expression in DRG neurons may contribute to compromised axon regeneration.
Tissue and organ regeneration shares common mechanisms with morphogenesis and, to certain degree, recapitulates the development process. Actually, neonatal mammals, different from adult mammals, have remarkable regeneration potentials after both peripheral and central nerve injuries [
38]. Along with the downregulation of many regeneration promoting molecules and the upregulation of many regeneration inhibiting molecules during development, the regeneration capacity of the nervous system declines stepwise [
39]. Phosphatase and tensin homolog (PTEN) is one of the most well-known neuron-intrinsic inhibitors whose deletion is capable of enhancing the regeneration ability of retinal ganglion cells, corticospinal neurons, and DRG neurons in adult mammals [
40–
42]. The investigation of the expression patterns of PTEN during regeneration demonstrates that PTEN is expressed at a low level in the DRGs of rats at the embryonic day 18.5 but expressed at a high level at postnatal day 5 and the adult stage [
43]. And the expression level of PTEN in the DRGs is decreased following peripheral nerve injury [
43,
44]. Given that similar as PTEN, DBNDD2 is differentially expressed after nerve injury, we next evaluated that whether the expression levels of DBNDD2 is altered during development. We compared the abundances of DBNDD2 in neonatal and adult rat DRGs using single-cell sequencing data, determined the expression trends of DBNDD2 in the DRGs of rats at different ages (1-day-old neonatal, 2-week-old, and 4-week-old) using RT-PCR, and found development-dependent increase of DBNDD2 gene expression. The application of single-cell sequencing separates neurons from other different types of cells and allows the identification of the transcription programs in neurons under various physiologic and pathological conditions [
45–
47]. Here, using single-cell sequencing data, we distinguished DRG neurons from glial cells and immune cells in neonatal and adult rat DRGs and found that DRG neurons occupy a large cell population in the DRGs of rats at different ages. Using t-SNE plot, we visualized the presence of DBNDD2 in both neonatal and adult rat DRGs and found the obviously higher expression of DBNDD2 in adult rats. The increased expression trend of DBNDD2 in DRG neurons during development, together with the elevated amount of DBNDD2 after the non-regenerative central axon injury and the reduced amount of DBNDD2 after the regenerative peripheral axon injury, implies that DBNDD2 may be an inhibitory factor for neuron growth and axon regeneration.
To explore the biological effects of DBNDD2 on neurons, RNA interference, an effective technology that mediates sequence-specific gene knockdown, was applied and rat DRG neurons were transfected with siRNA segments against DBNDD2. Three siRNA segments targeting different portions of the target gene DBNDD2 and a siRNA segment targeting sequences altered from the target were utilized to examine the knockdown efficiency. DBNDD2-siRNA-1 and DBNDD2-siRNA-2, two siRNA segments that robustly suppressed DBNDD2 gene expression, were utilized together for the success manipulation of DBNDD2 gene expression.
In vitro monitoring of neurite outgrowth showed that DBNDD2 siRNA transfection leads to enhanced neuron growth in both neonatal rat DRG neurons with certain regeneration capacity and adult rat DRG neurons with limited regeneration capacity. More importantly, in adult rats, via an intrathecal injection of DBNDD2 shRNA-expressing AAV, it was found that silencing DBNDD2 is sufficient to promote the elongation of injured axons of cultured adult DRG neurons as well as the regeneration of injured sciatic nerves. These studies fully indicate that DBNDD2 is a key regulating factor of neuron growth and axon regeneration and imply that knocking down DBNDD2 in neurons is an effective strategy for restoring impaired nerve functions. It is worth raising that in the current study, the
in vivo roles of reduced DBNDD2 expression in DRG neurons is examined by immunostaining the regenerating sciatic nerves with SCG10 at 3 days after rat sciatic nerve crush injury. Sciatic nerve crush injury, as previously mentioned, is a commonly used peripheral nerve injury model. Notably, peripheral nerve crush injury only induces a modest damage and elicits axonotmesis without disrupting the epineurium. Compared with crush injury, peripheral nerve transection injury and/or long gap peripheral nerve injury disrupts the entire nerve stump, including endoneurium, perineurium, as well as epineurium, and elicits more serious consequences [
46,
48]. Therefore, the functional roles of DBNDD2 knockdown in axon regeneration after sciatic nerve transection injury and/or long gap sciatic nerve injury can be assessed to determine the effects of DBNDD2 silencing on more serious injuries. The evaluation of long-term functional recovery after sciatic nerve injury will further contribute to the determination of the clinical potential of the manipulation of DBNDD2. The effects of DBNDD2 deficiency on axon regeneration following injury to central axon branches of DRG neurons and even injury to central nerves can be further assessed to examine whether DBNDD2 deficiency is capable of triggering central axon-injured DRG neurons and/or central neurons to switch to a pro-regenerative state. Actually, the fact that DBNDD2 knockdown boosts the regenerative axon growth of DRG neurons cultured on myelin, a well demonstrated inhibitor in axonal repair [
49], implies that DBNDD2 deficiency may be able to enable neurons to overcome the inhibitory microenvironment in the nervous system.
Notably, the current study primarily focuses on the role of DBNDD2 in DRG sensory neurons. The functional contributions of DBNDD2 to the regeneration of injured motor neurons remain undetermined. Moreover, considering there are multiple types of non-neuronal cells in the DRGs, such as satellite glial cells, Schwann cells, and fibroblasts, the expression changes of DBNDD2 in these different cell types after axotomy and the functional roles of DBNDD2 in these non-neuronal cells in the DRGs are also worthy of investigation.
The construction of molecular interaction network is valuable for the discovery of functionally associated molecules and the systemic understanding of biological processes [
50]. For the investigation of DBNDD2-associated molecules, in the current study, we analyzed molecules that interact with DBNDD2 using the STRING data resource and the Pathway Commons website and screened transcription factors targeting DBNDD2 using JASPAR, animalTFDB 3.0, GTRD, and hTFtarget databases. It is worth mentioning that databases of human transcriptional regulatory interactions are comprehensive while there are less number of databases that recognize the regulations of rat transcription factors and their downstream target genes. Still, it is demonstrated that the cis-regulatory modules and transcription factor binding locations among species are relatively conserved [
51]. Hence, except for selecting rat species in JASPAR database, we used all animal species in animalTFDB 3.0 database, human species in GTRD database, and human species in hTFtarget database to predict potential upstream transcription factors targeting DBNDD2 and then discovered ESR1 as a potential upstream regulator of DBNDD2.
Transcription factors are important gene regulating factors in numerous biological phenomena, including nerve injury and regeneration [
52,
53]. For instance, activating transcription factor 3 (ATF3) and AP-1 transcription factor subunit Jun proto-oncogene (JUN), two transcription factors that are upregulated in the DRGs after sciatic nerve injury, regulates many regeneration-associated molecules and enhances neurite outgrowth [
54]. It has been demonstrated that, in the nervous system, transcription factor ESR1 in glutamatergic and GABAergic neurons is important for normal puberty phenotype [
55]. The effect of ESR1 on axon growth and regeneration remains largely undetermined. Notably, emerging studies demonstrate that ESR1 is expressed at low abundance in various types of cancers, such as hepatocellular carcinoma [
56], endometrioid endometrial cancer [
57], breast cancer [
58], non-small cell lung cancer [
59], and bladder cancer [
60]. Overexpression of ESR1 mediates cellular apoptosis and hinders cellular proliferation and invasion [
61,
62]. On the other hand, reduced ESR1 expression stimulates cellular proliferation, migration, and invasion [
63]. The reduced expression patterns of ESR1 in tumor tissues as compared with in non-tumor tissues as well as the inhibiting roles of ESR1 on cell growth indicate that ESR1 functions as a tumor suppressor gene [
56,
64]. Tumor suppressor genes may be key regulators of nerve regeneration as the reduced expressions of many tumor suppressor genes, including PTEN, adenomatous polyposis coli (APC), and retinoblastoma (Rb), modulate neurite plasticity, support axon regeneration, and facilitate the recovery of injured nerves [
41,
65–
67]. Our bioinformatic analysis indicates that it is likely that tumor suppressor gene ESR1 is also a negative regulator of nerve regeneration and ESR1 may inhibit neurite growth and axon regeneration by targeting DBNDD2. Still, the biological functions of these potential upstream transcription factors of DBNDD2 as well as screened DBNDD2-associated molecules have not been experimentally revealed and may be explored in further studies.
Taken together, our study reveals that DBNDD2 exhibits distinct expression patterns in rat DRGs depending on the type of axonal injury and developmental stage. Specifically, DBNDD2 expression is downregulated following peripheral axotomy but upregulated after central axotomy in adult rats, while its expression progressively increases during development. We demonstrate that reduced levels of DBNDD2 contribute to enhanced neurite growth and nerve regeneration, suggesting that DBNDD2, along with its potential upstream regulator ESR1, may act as novel suppressors of successful axon regeneration. Molecular manipulation approaches that decrease DBNDD2 expression represent an attractive therapeutic strategy to modify the regeneration ability of neurons and to improve nerve regeneration.
PDF
(5820KB)
