EphB2 knockdown decreases the formation of astroglial‐fibrotic scars to promote nerve regeneration after spinal cord injury in rats

Abstract Aims At the beginning of spinal cord injury (SCI), the expression of EphB2 on fibroblasts and ephrin‐B2 on astrocytes increased simultaneously and their binding triggers the formation of astroglial‐fibrotic scars, which represent a barrier to axonal regeneration. In the present study, we sought to suppress scar formation and to promote recovery from SCI by targeting EphB2 in vivo. Methods The female rats SCI models were used in vivo experiments by subsequently injecting with EphB2 shRNA lentiviruses. The effect on EphB2 knockdown was evaluated at 14 days after injury. The repair outcomes were evaluated at 3 months by electrophysiological and morphological assessments to regenerated nerve tissue. The EphB2 expression and TGF‐β1 secretion were detected in vitro using a lipopolysaccharides (LPS)‐induced astrocyte injury model. Results RNAi decreased the expression of EphB2 after SCI, which effectively inhibited fibroblasts and astrocytes from aggregating at 14 days. The expression of EphB2 in activated astrocytes, in addition to fibroblasts, was significantly increased after SCI in vivo, in line with upregulated expression of EphB2 and increased secretion of TGF‐β1 in astrocyte culture treated with LPS. Compared to the scramble control, RNAi targeting with EphB2 could promote more nerve regeneration and better myelination. Conclusions EphB2 knockdown may effectively inhibit the formation of astroglial‐fibrotic scars at the beginning of SCI. It is beneficial to eliminate the barrier of nerve regeneration.

of keratan sulfate by keratanase, 6 inhibition of the accumulation of CSPGs using high-molecular-weight hyaluronic acid, 7 knocking out of GFAP and vimentin genes, 8 and photochemical removal of astroglial-fibrotic scars. 9 However, these measures only degrade some of the components of astroglial-fibrotic scars that have already formed, and thus, the effect is limited. Therefore, it would be more effective to reduce the formation of astroglial-fibrotic scars from the beginning.
Many studies have shown that the formation of astroglial-fibrotic scars is the result of the interaction of multiple cells, 10 including astrocytes and fibroblasts. 11,12 Bundesen and colleagues have confirmed that at the beginning of SCI, the expression of the receptor EphB2 on fibroblasts is significantly upregulated, triggering the formation of astroglial-fibrotic scars by binding to the corresponding ligand ephrin-B2, which is upregulated in astrocytes simultaneously. 13 When fibroblasts and astrocytes aggregate, a compact cell boundary near the injury border is formed and the gaps between the cells get filled by extracellular matrix, such as CSPGs. 14,15 Two weeks after injury, the astroglial-fibrotic scar is formed at the injury border. Although other types of cells are involved in the formation of astroglial-fibrotic scars, 16,17 the accumulation of fibroblasts and astrocytes at the injury border mediated by upregulation of EphB2 and ephrin-B2 is the basis for the scar formation. 18,19 Therefore, specific inhibition of this upregulation may inhibit the effective formation of astroglial-fibrotic scars and may be another way for promoting SCI repair.
The expression of EphB2 in the human pancreatic cancer cell line, CFPAC-1, was successfully suppressed by RNAi to elucidate the relationship between EphB2 and tumor formation. 20 Ephrin-B2 expression was successfully reduced by injecting specific shRNA targeted ephrin-B2 to relieve neuropathic pain. 21 Therefore, we used RNAi to specifically inhibit the increased expression of EphB2 after SCI. Its binding to the corresponding ephrin-B2 decreased, thus effectively preventing aggregation of fibroblasts and astrocytes. This may decrease the formation of astroglial-fibrotic scars, facilitate the regeneration of axons across the damaged zone, and promote the recovery of the morphology and electrical conduction of the injured spinal cord.

| RNAi knockdown
A lentiviral vector (pLV-shRNA-zsGreen1, see Figure S1) containing shRNA (Biomics Biotechnologies, Nantong, China) that targets the EphB2 and the ZsGreen1 reporter gene was injected into rats for the in vivo experiment. The positive-sense strand sequence of rat-Ephb2sh-2 was GATCCAGAAGGAGCUCAGUGAGUAdTdTTTTTTT, and the antisense was GUACUCACUGAGCUCCUUCUdTdTAAAAAA.

| In vitro lipopolysaccharide (LPS)-induced astrocyte injury model
Primary astrocyte cultures were prepared from spinal cords of SD rats at postnatal day 1 as previously described. 23 After 48 h of LPS treatment (10 µg/ml in saline; Sigma), the culture medium was collected to examine the TGF-β1 levels using an ELISA kit (Abcam) according to the manufacturer's instructions. Saline treatment was used as the control group. Furthermore, the cells were collected to determine the EphB2 and ephrin-B2 protein levels using Western blot analysis.

| Quantitative PCR and Western blot analysis
Fresh spinal cords from the different groups were acquired from the rats under anesthesia at 2 mm of the cranial and caudal levels from the center of injury at the indicated time points. RNA and protein were extracted. For subsequent procedures, see . PCR primers (Table S1) were synthesized by Invitrogen. Antibodies used for Western blot were seen in Table 1.

| Morphological assessment
The rats were perfused with 4% paraformaldehyde solution, and the spinal cords were severed 5 mm from each end of the injection point outwards. Samples were fixed, dehydrated by gradient sucrose, and transected serially using a freezing microtome (CM1900 Leica) at a thickness of 16 µm.
Conventional immunofluorescence: The samples were blocked and permeabilized and then incubated with primary antibodies (

| Electrophysiological assessment
Three months after SCI, the motor evoked potentials (MEPs) were recorded. Briefly, after the rats were anesthetized, the stimulating electrode (60 mA in strength) was placed under the skin of the rat's head, and the recording electrode was placed on the calf muscle. The amplitude and incubation period of MEP were obtained.

| Statistical analysis
Data analysis and statistical analysis were performed by SPSS 19.0.
Levene's test was used to assess data normality. All data were analyzed by one-way ANOVA plus Dunnett's post hoc test for multi-group comparisons or with Student's t test for two-group comparisons. Data are expressed as mean ± standard deviation (SD).
p < 0.05 was considered statistically significant.

| Effect of EphB2 and ephrin-B2 expression by RNAi
Bundesen et al 13 have reported the high expression of EphB2 and ephrin-B2 protein after SCI. We used qPCR to examine the change in mRNA expression of EphB2 and ephrin-B2 after SCI. Consistent with the abovementioned study, with time there was a progressive expression increase until 14 days after SCI ( Figure 1A,B).
As the astroglial-fibrotic scar is formed maturely at 14 days after SCI, we selected 14 days as the observation point for the formation of astroglial-fibrotic scar by RNAi, and functional recovery was detected at 3 months ( Figure 1E). The EphB2 mRNA and protein expression levels at the injury sites and the surrounding tissues in the cont shRNA group were significantly higher than those in the Sham group (p < 0.01; Figure 1C,F,G) 14 days after injury. EphB2 RNAi (SCI+EphB2 shRNA group) significantly reduced the high expression of EphB2 mRNA and protein after injury (p < 0.01) compared with the cont shRNA group (Figure 1C,F,G). In addition, injection of EphB2 shRNA after SCI had no effect on the high expression of ephrin-B2 at the mRNA and protein levels ( Figure 1D,F,H). These data indicated that the RNAi injection technique effectively reduced the high expression of EphB2 at the site of injury after SCI. After RNAi, the high expression of EphB2 mRNA was inhibited at 14 days after SCI+RNAi (C, n = 9). The high expression of ephrin-B2 mRNA after RNAi was not affected (D, n = 9). The experimental timeline shows the main items tested at 14 days and 3 months after SCI (E). Fourteen days after SCI, injured spinal cords were collected for Western blot assay to determine the expression of EphB2 and ephrin-B2 (F). Histogram showed the expression of EphB2 (G) and ephrin-B2 (H) proteins (n = 9, respectively). **p < 0.01, one-way ANOVA respectively). Fourteen days after SCI, a large number of activated astrocytes that expressed GFAP showed significant hyperplasia, hypertrophy, and dense aggregation in the region surrounding the injury in the SCI+cont shRNA group ( Figure 3C1-4). However, the astrocytes were still activated but did not aggregate at the edge of the lesion after RNAi ( Figure 3D1-4). The neurofilament (NF)

|
immunofluorescence showed the regenerative axons in the injury area ( Figure 3C3,D3). The ratio of NF area to the injury area in the SCI+EphB2 shRNA group was significantly higher compared with that in the SCI+cont shRNA group (p < 0.01; Figure 3G). These data suggested that there were more regenerative axons growing into the injury area after RNAi.

F I G U R E 2
The immunofluorescent technique showed no fibroblasts that express FN and normal morphological astrocytes that express GFAP in sham group (A1-A4) and in SCI+cont shRNA group (B1-B4

| ephrin-B2 and EphB2 expression in astrocytes
Immunofluorescence showed low expression of ephrin-B2 in astrocytes in the normal spinal cord (not at the injection point; Figure 2C1-4). However, the expression of ephrin-B2 significantly increased after SCI on the activated astrocytes ( Figure 4A1-4), and the expression remained high after RNAi ( Figure 4B1-4).
In addition to the fibroblast expression of EphB2 at the injury site, the EphB2 fluorescent signal was visible on some cells after SCI ( Figure 3A4,B4). Morphologically, these cells appeared to be activated astrocytes. Thus, immunofluorescence double staining of EphB2 and GFAP was performed. The activated astrocytes also showed clear EphB2 fluorescence ( Figure 4C1-4), although not on the normal astrocytes ( Figure 2E1-4). After RNAi, the activated astrocytes showed faint EphB2 fluorescence ( Figure 4D1-4). To test the change in expression of EphB2 on activated astrocytes, we set up an in vitro LPS-induced injury model. The ephrin-B2 and EphB2 protein expression level was significantly increased in astrocytes after 48 h of exposure to LPS compared with the control group (p < 0.01; Figure 4E,F,G). We further examined the TGF-β1 levels in the culture medium using ELISA kit. The data showed that the activated astrocytes secreted more TGF-β1 than the control cells (p < 0.01; Figure 4H). In addition, it could also be seen from Figure 4 that many astrocytes expressed GFP, indicating that the shRNA successfully entered the astrocytes.

| Electrophysiological parameters of animals at 3 m after SCI
The BBB motor behavioral test and CatWalk XT Automated Gait Analysis system was performed to detect the motor function of rats (seen in the Supplementary Materials). Although the data of the SCI+EphB2 shRNA group were slightly better than that of the SCI+cont shRNA group, there was no statistical difference ( Figure S2). Subsequently, the MEPs were recorded at calf muscle ( Figure 5A-C). The incubation period of MEP in the SCI+EphB2 shRNA group was significantly shorter than that in the SCI+cont shRNA group (p < 0.01; Figure 5D), but MEP amplitudes exhibited no significant difference between these two groups ( Figure 5E).

F I G U R E 3
The expression of FN, EphB2, GFAP, and NF after SCI at day 14. The immunofluorescence technique showed the expression of FN and EphB2 in the SCI+cont shRNA group (A1-A4) and the SCI+EphB2 shRNA group (B1-B4). The fluorescence intensity of FN and EphB2 at the injury site showed a significant difference between the two groups (E and F, respectively; n = 8 or 9). The immunofluorescence technique showed the expression of NF and GFAP in the SCI+cont shRNA group (C1-C4,) and the SCI+EphB2 shRNA group (D1-D4). The NF-positive area/injury area ratio showed a significant difference at the injury between the two groups (G, n = 6). The injuries area is enclosed by solid lines. Scale bar = 100 µm, **p < 0.01, t-test showed that the myelin sheath thickness in SCI+cont shRNA group was concentrated below 0.2 μm, but there were more thick myelin sheaths in SCI+EphB2 shRNA group ( Figure 6G,H). The G-ratio (myelinated axon diameter/myelinated fiber diameter) as a reliable measure was used to assess axonal myelination. 24 Both G-ratios of the SCI+cont shRNA group at the IS and CS were larger than that of the SCI+EphB2 shRNA group (Figure 6I,J). These suggested that the myelin sheaths in SCI+EphB2 shRNA group were more mature than SCI+cont shRNA group.

| DISCUSS ION
The formation of scar tissue is an inevitable process after SCI.
Although the formation of glial scar is in favor of axon regeneration in the central nervous system (CNS), 25 several latest reports show that decreased glial reaction and scar formation is beneficial to recovery from spinal cord injury. 26,27 After CNS injury, fibroblasts, pericytes, and inflammatory cells, in addition to astrocytes, participated in the formation of extracellular matrix. 10 However, fibrotic scars and glial scars together built astroglial-fibrotic scars, which had a negative effect on axon regeneration. 3,28,29 Bundesen et al 13  Our results demonstrated that the activated astrocytes expressing ephrin-B2 also highly expressed EphB2 in vivo and in vitro, which is different from previous report. 13 It has been F I G U R E 5 MEP examinations at 3 months after SCI. Representative recordings at sham (A), SCI+EphB2 shRNA (B), and SCI+cont shRNA group (C). Histograms showed the incubation period (D, n = 6) and MEP amplitudes (E, n = 6). **p < 0.01, one-way ANOVA reported that Eph receptors and ephrins can be co-expressed in the same cells. 33 Eph receptors and ephrins engage in a multitude of activities, including cell morphology, adhesion, migration, proliferation, survival, and differentiation. 33 Similarly, our in vitro experiments showed that LPS-activated astrocytes secreted more TGF-β1 than the control group. After CNS injury, TGF-β1 produced and released by astrocytes increases around the lesion site, to enhance fibroblast proliferation. 34-36 Therefore, we think that the activated astrocytes expressing ephrin-B2 and the fibroblasts expressing EphB2 mainly together formed the astroglial-fibrotic scar; in the same way, the astrocytes expressing EphB2 may interact with astrocytes expressing ephrin-B2 and also participate in the formation of glial scars. We speculate that knocking down  cn/ac), for editing the English text of a draft of this manuscript.

CO N FLI C T O F I NTE R E S T
We have not published or submitted the manuscript elsewhere simultaneously. The authors taking part in this study declared that they do not have any conflict of interests in this manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The authors confirm that the data supporting the findings of this study are available within this article and from the corresponding author upon reasonable request.

R E FE R E N C E S
34. Lindholm D, Castren E, Kiefer R, Zafra F, Thoenen H. Transforming growth factor-beta 1 in the rat brain: increase after injury and inhibition of astrocyte proliferation. J Cell Biol. 1992;117 (2)