Thymosin β4 promotes zebrafish Mauthner axon regeneration by facilitating actin polymerization through binding to G-actin

Abstract

Background

Thymosin beta 4 (Tβ4) is a monomeric actin-binding protein that plays many roles in biological activities. However, some studies on the role of Tβ4 in central axon regeneration have yielded contradictory results. Previous research has focused primarily on cultured cells, leading to a deficiency in in vivo experimental evidence. Therefore, we used a single axon injury model of Mauthner cells in zebrafish larvae to investigate the role of Tβ4 in central axon regeneration in vivo.

Results

Our results demonstrated that knockout of Tβ4 impaired axon regeneration, whereas overexpression of Tβ4 promoted axon regeneration. Moreover, this promotion is mediated through the interaction between Tβ4 and G-actin. Furthermore, our results suggest that the binding of Tβ4 to G-actin promotes actin polymerization rather than depolymerization. In the rapid escape behavior test, larvae with damaged axons presented impaired tail muscle control, resulting in a lack of normal tail bending, termed the straight tail phenomenon. The proportion of straight tails was significantly negatively correlated with axon regeneration length, suggesting that it is a new indicator for assessing rapid escape behavior recovery. Finally, the results showed that the overexpression of Tβ4 effectively restored the functionality of rapid escape behaviors mediated by Mauthner cells.

Conclusions

Our results provide evidence that Tβ4 promotes central axon regeneration in vivo through binding to G-actin and suggest that Tβ4 could serve as a potential polypeptide drug for clinical therapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12915-024-02045-2.

Background

Thymosin beta 4 (Tβ4) is a peptide initially discovered in the bovine thymus that is encoded by the gene Tmsb4x and typically consists of 43 amino acids [1]. Tβ4 has been identified as a monomeric actin-binding protein, and previous studies have suggested that it plays a role as an inhibitor of actin polymerization [2, 3]. However, subsequent research indicated that Tβ4 may not solely inhibit actin polymerization. The affinity of Tβ4 for ATP-G-actin significantly exceeds its affinity for ADP-G-actin [4, 5]. Its interaction with G-actin facilitates the transport of G-actin from the cytoplasm to the plasma membrane, thereby exerting a positive influence on actin assembly [6, 7]. Therefore, the function of Tβ4 could encompass the maintenance of a pool of available G-actin within cells [8].

Tβ4 participates in numerous biological processes, such as cell proliferation [9], angiogenesis [10, 11], apoptosis [12, 13], tissue regeneration [14–16], and inflammation [17]. However, existing studies offer conflicting perspectives on the role of Tβ4 in central axon regeneration. Research has indicated that the suppression of Tβ4 using dsRNA in cultured PeA neurons of snails could stimulate axon regeneration [18]. In contrast, a subsequent study revealed that the addition of the Tβ4 peptide to cultured mouse spinal cord neurons could increase neuron survival and neurite growth by increasing the expression of L1, the prototype of a neural subfamily of cell adhesion molecules, and plays important roles in axon growth and regeneration [19–24]. The overexpression of Tmsb4x in cultured hippocampal neurons inhibited dendritic formation [25]. These contradictory findings provoked our curiosity. Nevertheless, Tmsb4x is transiently upregulated in retinal ganglion cells (RGCs) following optic nerve injury in zebrafish [26]. Moreover, Tmsb4x is crucial for zebrafish axon development [27]. Given the robust regenerative capacity of zebrafish central axons [28, 29], the upregulation of Tβ4 is not inconsequential. Previous studies investigating the correlation between Tβ4 and axon regeneration have primarily been conducted in cultured cells. However, what is the role of Tβ4 in axon regeneration in vivo?

In this study, we first analyzed the expression levels of Tβ4 in the spinal cord injury (SCI) model of adult zebrafish, sciatic nerve crush and optic nerve injury models of mice [30–32]. The results revealed that Tβ4 was significantly upregulated after axonal injury. We subsequently confirmed the indispensable role of Tβ4 in central axon regeneration in vivo through gain- and loss-of-function experiments. Behavioral assays also demonstrated the contribution of Tβ4 to the recovery of rapid escape behavior mediated by Mauthner cells. Further investigations revealed that the pro-regenerative effect of Tβ4 is mediated through its interaction with G-actin, thereby enhancing actin polymerization. Our study provides in vivo evidence that Tβ4 can promote central axon regeneration. Given its high specificity, low immunogenicity, and cost-effectiveness as a polypeptide [33], Tβ4 may serve as a beneficial drug for promoting central axon regeneration in future therapeutic applications.

Results

The expression of tmsb4x is upregulated following axonal injury in vivo

Previous studies have indicated that Tmsb4x is upregulated following optic nerve injury in zebrafish [26]. To investigate whether this change also occurs after SCI, we analyzed the expression changes of tmsb4x in a public RNA-seq dataset from zebrafish spinal cords following injury. The result showed that tmsb4x was significantly upregulated at 1 week post-injury (wpi) and then gradually decreased (Fig. 1a) [30]. These findings suggest that tmsb4x may play an important role in axon regeneration in zebrafish. Given that mammalian axon injury models are more similar to human nerve injury, we also investigated the expression changes of Tmsb4x in mammalian axon injury models in public databases. We found that tmsb4x was significantly upregulated after both sciatic nerve crush (Fig. 1b) and optic nerve injury in mice. The latter model was subjected to triple intervention (PTEN knockout + SOCS3 knockout + CNTF overexpression) to restore the axon regeneration capacity of RGCs, and only RGCs with successful axon regeneration could be labeled by dextran micro-Ruby (MR) (Fig. 1c) [31, 32]. These findings indicate that tmsb4x is upregulated after axonal injury in both mice and zebrafish. Notably, tmsb4x was upregulated only in cells with successful axon regeneration (Fig. 1c) [32]. Overall, these results suggest that Tβ4 may play a positive role in axon regeneration.

Fig. 1.

Tβ4 is upregulated after axonal injury in both zebrafish and mice. a Expression profile of tmsb4x in the adult zebrafish spinal cord after injury. wpi, weeks post injury. Each group contains 5 samples, except for the 6 wpi caudal group, which contains 4 samples. b Expression profile of the Tmsb4x gene in DRG neurons at 72 h after sciatic nerve crush in mice, as determined by snRNA-seq analysis. "Naive" represents the group without crush injury, whereas "crush" represents the group that underwent crush injury. Cell numbers: Control, 6141; Crush, 5172. c Expression profile of Tmsb4x in RGCs with successful and failed axon regeneration following PTEN knockout, SOCS3 knockout, and CNTF overexpression. In this study, the authors used dextran micro-Ruby (MR) to retrogradely label RGCs with successful axon regeneration. "PSCCreMRneg" represents RGCs with failed axon regeneration and negative labeling of MR, while "PSCCreMRpos" represents RGCs with successful axon regeneration and positive labeling. Cell numbers: PSCCreMRneg, 66; PSCCreMRpos, 179

Deletion of tmsb4x delays the axon regeneration of Mauthner cells

Owing to previous studies employing cultured neurons, for a more precise exploration of the role of Tβ4 in axon regeneration, we investigated the effects of Tβ4 on axon regeneration using a single axon injury model of Mauthner cells in zebrafish larvae.

First, we generated tmsb4x knockout zebrafish. We designed a target site at the 5' end of the CDS region of tmsb4x and achieved targeted deletion by injecting the Cas9 protein and sgRNA complexes into one-cell-stage embryos (Additional file 1: Fig. S1a, S1b). Subsequently, we conducted separate crosses between these strains and both the wild-type (WT) and Tg(tol-056) strains to generate tmsb4x mutant heterozygotes. The Mauthner cells express enhanced green fluorescence protein (EGFP) of the Tg(tol-056) strain. Finally, tmsb4x knockout homozygous mutants were obtained by crossing these two F1 strains (Additional file 1: Fig. S1b).

Ultimately, we obtained a homozygous mutant with a 2-base pair (bp) deletion. This resulted in a mutated tmsb4x gene encoding only 12 amino acids, with just 6 consistent with the WT sequence (Additional file 1: Fig. S1a). We identified the DNA sequences of the heterozygous and homozygous mutant strains through Sanger sequencing (Additional file 1: Fig. S1c). Furthermore, due to the deletion of the 2-bp region, the XhoI restriction enzyme site was also mutated, so the DNA sequence of the tmsb4x homozygous mutant strain could not be cleaved by XhoI, while that of the heterozygous mutant strain could be only partially cleaved by XhoI. We identified the heterozygous and homozygous mutant strains of tmsb4x by enzyme digestion (Additional file 1: Fig. S1d, Additional file 4). The enzyme digestion results also confirmed the successful construction of the tmsb4x mutant strain.

To assess whether tmsb4x deletion affects Mauthner cell development and motor function, we examined Mauthner cell morphology and free-swimming ability. The results showed that tmsb4x deletion did not significantly affect the soma area (from a dorsal view) or total axon length of Mauthner cells at 6 days post-fertilization (dpf) (Additional file 1: Fig. S2a-2c, Additional file 3: Table S2), nor did it significantly affect the free-swimming ability at 7 dpf (Additional file 1: Fig. S2d-2e, Additional file 3: Table S2).

Next, to explore the effect of tmsb4x deletion on axon regeneration, we induced single-axon injury in 6 dpf zebrafish larvae of tmsb4x heterozygous and homozygous mutant strains using a two-photon laser above the cloacal pore (Fig. 2a-2b). The length of axon regeneration and the number of regenerated axon branches were then observed at 48 h post-injury (hpi) (Fig. 2a and c). Our findings revealed that both the tmsb4x heterozygous and homozygous mutants exhibited significantly reduced axon regeneration lengths but had no effect on the average number of regenerated axon branches (Fig. 2c-e, Additional file 3: Table S2). However, tmsb4x knockout reduced the proportion of axons with more than one branch (Fig. 2f). To further investigate whether tmsb4x knockout merely delays axon regeneration, we examined axon regeneration at 48–96 hpi. The results showed that the axon regeneration length of the tmsb4x^−/− line at 72 hpi was not significantly different from that of the wild-type line at 48 hpi, and the axon regeneration length of the tmsb4x^−/− line at 96 hpi was similar to that of the wild-type line at 72 hpi (Additional file 1: Fig. S3, Additional file 3: Table S2). These findings suggest that tmsb4x knockout delays Mauthner axon regeneration.

Fig. 2.

Knockout of tmsb4x attenuates axon regeneration of Mauthner cells. a Experimental workflow diagram. b A representative image of an injured Mauthner axon. The arrow indicates the injury site, which above the cloacal pore. Scale bar, 50 μm. c Representative images of Mauthner cell axon regeneration in the tmsb4x^+/+, tmsb4x^+/- and tmsb4^−/− lines at 48 hpi. Asterisks indicate the sites of axon injury, and arrows indicate the regenerated axon terminals. Scale bar, 100 μm d and e Statistical graphs of the length (d) and branch number (e) of regenerated axons in the tmsb4x^+/+, tmsb4x^+/- and tmsb4^−/− lines (lengths: tmsb4x^+/+, 402.5 ± 39.5 μm, n = 14; tmsb4x^+/-, 278.6 ± 26.9 μm, n = 26; tmsb4x^−/−, 246.1 ± 26.8 μm, n = 18; branch numbers: tmsb4x^+/+, 1.8 ± 0.2, n = 14; tmsb4x^+/-, 1.8 ± 0.2, n = 26; tmsb4x^−/−, 1.7 ± 0.2, n = 18). Assessed by one-way ANOVA/Tukey's multiple comparisons test for regeneration length (tmsb4x^+/+ vs. tmsb4x^+/-, adjusted p = 0.018; tmsb4x^+/+ vs. tmsb4x^−/−, adjusted p = 0.0047; tmsb4x^+/- vs. tmsb4x^−/−, adjusted p = 0.71). Kruskal–Wallis test/Dunn's multiple comparisons test was used to assess the branch number (tmsb4x^+/+ vs. tmsb4x^+/-, adjusted p > 0.99; tmsb4x^+/+ vs. tmsb4x^−/−, adjusted p > 0.99; tmsb4x^+/- vs. tmsb4x^−/−, adjusted p > 0.99). The axon regeneration length refers to the distance from the injury site to the tip of the longest regenerating axon branch, as indicated by the distance between the asterisk and the arrow in the figure. Same as below. f Percentages of branch numbers of regenerated axons in the tmsb4x^+/+, tmsb4x^+/- and tmsb4^−/− lines. Sample numbers: tmsb4x^+/+, 14; tmsb4x^+/-, 26; tmsb4x.^−/−,18. g Representative images of axon regeneration with tmsb4x restored in tmsb4x knockout lines. Asterisks indicate the sites of axon injury, and arrows indicate the regenerated axon terminals. Scale bar, 100 μm. h and i Statistical graphs of the length (h) and branch number (i) of regenerated axons with vector and tmsb4x restored in tmsb4x knockout lines (length: vector, 209.0 ± 33.2 μm, n = 7; tmsb4x: 422.7 ± 53.1 μm, n = 9; branch number: vector, 1.9 ± 0.5; n = 7; tmsb4x: 4.4 ± 0.7, n = 9). Assessed by a two-tailed, Mann Whitney test for regeneration length (p = 0.00030). And an unpaired, two-trailed t test was used to assess the branch number. (p = 0.011)

Since our knockout model is systemic, and previous studies have demonstrated that tmsb4x is expressed in a subset of glia surrounding newly forming axonal tracts [27], and these glial cells play a key role in supporting axon regeneration after SCI in adult zebrafish [34]. To assess the extent to which glial cells contribute to Mauthner axon regeneration in tmsb4x knockout zebrafish, we restored tmsb4x in single Mauthner cells of the tmsb4x knockout line. The results showed that tmsb4x restoration significantly rescued axonal regeneration length and increased the branch number of regenerated axons (Fig. 2g-i, Additional file 3: Table S2). These findings indicate that the role of glial cells in Mauthner axon regeneration is limited in tmsb4x knockout zebrafish.

Overexpression of tmsb4x promotes axon regeneration

To assess whether additional tmsb4x can promote axon regeneration, we first constructed a tmsb4x overexpression plasmid (Fig. 3a). To evaluate the overexpression efficiency, these plasmids were microinjected into one-cell stage embryos, and the total RNA of larvae with positive tmsb4x expression was used to perform quantitative real-time PCR (qPCR) analysis at 4 dpf (Additional file 1: Fig. S4a). The result demonstrated that this overexpression can result in at least a 1.5-fold increase in tmsb4x expression (Fig. 3b). Subsequently, we employed an in vivo single-cell electroporation technique to overexpress tmsb4x in individual Mauthner cells at 4 dpf. Next, we screened for mCherry-positive larvae and precisely damaged the positive axons using a two-photon laser above the cloacal pores at 6 dpf (Fig. 3c). We found that the overexpression of tmsb4x significantly increased the axon regeneration length and branch number of regenerated axons (Fig. 3d-3f, Additional file 3: Table S2). These results indicate that the addition of tmsb4x can further promote central axon regeneration in vivo.

Fig. 3.

In vivo single-cell overexpression of tmsb4x promotes axon regeneration in Mauthner cells. a Schematic representation of the overexpression plasmids. b Efficiency of the tmsb4x overexpression plasmid. The data were assessed by an unpaired, two-tailed t test. p = 0.045. c Experimental workflow diagram. d Representative images of axon regeneration after tmsb4x overexpression. Asterisks indicate the sites of injury, and arrows indicate the regenerated axon terminals. Scale bar, 100 μm. e and f Statistical graphs of the length (e) and branch number (f) of regenerated axons overexpressing tmsb4x (length: control, 258.8 ± 34.4 μm; tmsb4x_oe, 464.1 ± 40.0 μm; branch number: control, 2.6 ± 0.4; tmsb4x_oe, 4.2 ± 0.5). The results were assessed by an unpaired, two-tailed t test (length, p = 0.0031; branch number, p = 0.042)

Tβ4 promotes axon regeneration through its interaction with G-actin

Previous research on cultured mouse spinal cord neurons revealed that Tβ4 can promote neurite growth by upregulating L1 [19]. However, Tβ4 is famous for its ability to bind to the actin monomer. Therefore, we examined whether the overexpression of tmsb4x in zebrafish would induce the upregulation of L1. In zebrafish, L1 possesses two homologous genes: l1cama and l1camb. Through qPCR analysis of total RNA extracted at 4 dpf, we discovered that the overexpression of tmsb4x in zebrafish did not affect the expression of l1cama or l1camb (Additional file 1: Fig. S4a-S4c). We then performed a co-expression analysis between tmsb4x and l1cama, l1camb, and actb1 using the RNA-seq data obtained from zebrafish spinal cords after injury [30]. The results showed a significant positive correlation between tmsb4x and actb1, while a negative correlation was observed with l1cama and l1camb (Additional file 1: Fig. S4d-S4f). These results suggest that actb1 is more likely to be the downstream target of tmsb4x in promoting axon regeneration.

Previous studies have shown that the N-terminal α-helix of Tβ4 is crucial for binding G-actin [4]. Deletion of the N-terminal 7 amino acids significantly reduced the binding affinity between Tβ4 and G-actin, and the K11P mutation also significantly impaired this interaction [35, 36]. To combine both mutations, we generated a Tβ4 variant lacking the N-terminal 13 amino acids (Fig. 4a). Using AlphaFold2, we predicted the 3D structures of wild-type Tβ4 and the N-terminal 13 amino acid deletion mutant in complex with β-actin [37]. Additionally, PRODIGY was used to estimate the binding affinity and dissociation constant between these models [38]. These results indicate that the deletion of the N-terminal 13 amino acids reduces the binding affinity between Tβ4 and β-actin (Fig. 4b).

Fig. 4.

The 13 N-terminal amino acids of Tβ4 are essential for its axon regeneration-promoting activity. a Schematic representation of the protein sequence of tmsb4x and the mutant protein sequence with a deletion of the 13 N-terminal amino acids. b 3D models of wild-type (left) and N-terminal 13 amino acids deleted (right) Tβ4 in complex with β-actin, predicted by AlphaFold2. In the left image, the N-terminal 13 amino acids forming the α-helix are highlighted in green. Below the images are the dissociation constants predicted by PRODIGY based on the 3D models. c and d Representative images (c) and statistical analysis (d) of axon regeneration in zebrafish overexpressing mutant tmsb4x with deletion of the 13 N-terminal amino acids (control, 222.5 ± 45.5 μm, n = 7; tmsb4x_D13AA, 207.1 ± 23.1 μm, n = 7). The results were assessed by an unpaired, two-tailed t test, p = 0.77. The asterisks indicate the sites of injury, and the arrows indicate the terminals of the regenerated axon. Scale bar, 100 μm

To investigate whether the pro-regenerative function of Tβ4 is indeed mediated by G-actin, we delivered a plasmid containing this mutation into individual Mauthner cells at 4 dpf. The results showed that the truncated Tβ4 no longer promoted axon regeneration at 48 hpi (Fig. 4c, d, Additional file 3: Table S2). This finding suggests that binding to G-actin is a prerequisite for Tβ4 to exert its pro-regenerative effect on central axons.

Tβ4 promotes Mauthner axon regeneration by facilitating actin polymerization

Nonetheless, the absence of the 13 N-terminal amino acids in Tβ4 may also influence its pro-regenerative effect via alternative pathways. Therefore, we further validated this effect by employing a pharmacological approach. Latrunculin B (Lat. B) binds to G-actin at a 1:1 ratio and inhibits actin filament polymerization [39, 40]. We immediately treated the zebrafish with Lat. B continuously for 48 h after Mauthner axon injury (Fig. 5a). This result indicated that Lat. B significantly inhibited Mauthner axon regeneration (Fig. 5b, c, Additional file 3: Table S2). To eliminate the possibility of non-Mauthner cells being affected by Lat. B and thereby inhibited axon regeneration, we injected Lat. B into individual Mauthner cells using a single-cell electroporation technique at 5 dpf (Fig. 5d). The results showed a significant inhibition of regeneration length at 24 hpi (Fig. 5e and g, Additional file 3: Table S2). However, this inhibition became non-significant at 48 hpi, potentially because of the short duration of Lat. B (Fig. 5f and h, Additional file 3: Table S2). These results indicated that Lat. B inhibited axon regeneration, suggesting that blocking the interaction between Tβ4 and G-actin indeed reduced axon regeneration.

Fig. 5.

Lat. B inhibits axon regeneration in zebrafish. a Experimental workflow for the immersion method used to treat zebrafish larvae, with daily replacement of fresh drug after axon injury. b and c Representative images (b) and statistical analysis (c) of axon regeneration in zebrafish larvae treated with different concentrations of Lat. B (DMSO, 412.3 ± 51.4 μm, n = 13; Lat. B (10 nM), 261.5 ± 37.2 μm, n = 16; Lat. B (20 nM), 207.2 ± 34.3 μm, n = 14; Lat. B (50 nM), 225.1 ± 36.5 μm, n = 15). Assessed by one-way ANOVA/Dunnett's multiple comparisons test (Lat. B (10 nM), adjusted p = 0.027; Lat. B (20 nM), adjusted p = 0.0025; Lat. B (50 nM), adjusted p = 0.0053). The asterisks indicate the sites of injury, and the arrows indicate the terminals of the regenerated axon. Scale bar, 100 μm. d Experimental workflow for single-cell drug treatment. e and g Representative images (e) and statistical analysis (g) of axon regeneration 24 hpi in zebrafish larvae treated with different concentrations in single cells (DMSO, 179.0 ± 10.3 μm, n = 7; Lat. B (1 μM), 119.5 ± 10.9 μm, n = 13; Lat. B (8 μM), 119.2 ± 11.4 μm, n = 14). Assessed by one-way ANOVA/Dunnett's multiple comparisons test (Lat. B (1 μM), adjusted p = 0.0047; Lat. B (8 μM), adjusted p = 0.0040). The asterisks indicate the sites of injury, and the arrows indicate the terminals of the regenerated axon. Scale bar, 100 μm. f and h Representative images (f) and statistical analysis (h) of axon regeneration at 48 hpi in zebrafish larvae treated with different concentrations of Lat. B at the single-cell level. (DMSO, 264.6 ± 31.2 μm, n = 10; Lat. B (1 μM), 191.1 ± 16.4 μm, n = 12; Lat. B (8 μM), 197.9 ± 27.5 μm, n = 10). Assessed by one-way ANOVA/Dunnett's multiple comparisons test (Lat. B (1 μM), adjusted p = 0.077; Lat. B (8 μM), adjusted p = 0.13). The asterisks indicate the sites of injury, and the arrows indicate the terminals of the regenerated axon. Scale bar, 100 μm

To investigate whether Lat. B can block the pro-regenerative effect of additional Tβ4, we overexpressed Tβ4 in individual Mauthner cells at 4 dpf. Then, we treated the zebrafish with Lat. B immediately after axon injury at 6 dpf (Fig. 6a). The results showed that Lat. B significantly blocked the pro-regenerative effect of Tβ4 at 48 hpi (Fig. 6b and c, Additional file 3: Table S2). These findings further confirm that Tβ4 promotes axon regeneration through its interaction with G-actin. Additionally, since Lat. B can inhibit actin polymerization, these results also suggest the importance of actin polymerization in axon regeneration. In summary, these results indicate that the pro-regenerative effect of Tβ4 depends on its interaction with G-actin and that this interaction promotes actin polymerization.

Fig. 6.

Lat. B blocks the regenerative effect of tmsb4x overexpression. a Experimental workflow. Lat. B was replaced daily. b and c Representative images (b) and statistical analysis (c) of axon regeneration in zebrafish larvae overexpressing tmsb4x treated with Lat. B immediately after axon injury and continuously for 48 h (tmsb4x_oe + DMSO, 477.7 ± 41.0 μm, n = 7; tmsb4x_oe + Lat. B, 195.0 ± 31.5 μm, n = 7). Assessed by a two-tailed, Mann Whitney test, p = 0.0041. The asterisks indicate the sites of injury, and the arrows indicate the terminals of the regenerated axon. Scale bar, 100 μm

Overexpression of tmsb4x contributes to functional recovery

Mauthner cells are examples of individual cells that control a specific behavior [41, 42]. Previous research has suggested that these cells participate in the rapid escape behavior (also known as C-start escape behavior) in zebrafish in response to startling stimuli [42–44]. Furthermore, injury to Mauthner axons significantly impacts escape behavior [45, 46]. Given the role of Tβ4 in promoting axon regeneration and branching, we aimed to examine whether Tβ4 aids in the functional recovery of Mauthner cells. Consequently, we conducted a behavioral test in zebrafish larvae overexpressing tmsb4x.

Similar to previous procedures, we overexpressed tmsb4x in individual Mauthner cells at 4 dpf and induced axonal injury to mCherry-positive axons at 6 dpf. We subsequently used a high frame rate camera to capture rapid escape behaviors at 8 dpf (Fig. 7a). We employed a brief 500 Hz sinusoidal sound stimulus to trigger C-start escape behavior (Fig. 7b). During behavioral testing, we observed a unique phenomenon (Fig. 7c and S5a). Unlike normal C-start responses, a straight tail occurred in the C-start escape behavior and resulted in a significantly increased maximal angle and time to maximal angle, while the angular speed remained unaffected (Additional file 1: Fig. S5b-S5e, Additional file 3: Table S2). Mauthner cells are located in rhombomere 4 of the hindbrain, and cells resembling Mauthner cells, such as MiD2cm and MiD3cm, are also present in rhombomeres 5 and 6. These cells also respond to auditory stimuli and can trigger C-start-like escape behaviors [47–49]. Since our axon injury was performed above the cloacal pores, the mid-region of the Mauthner axons, we speculated that the straight tail might have resulted from the interruption of escape signal transmission in the spinal cord due to axon damage (Additional file 1: Fig. S5f). Therefore, we hypothesized that the straight tail could be used to determine whether a specific escape behavior is mediated by Mauthner cells, and we validated this hypothesis via behavioral tests. Because we were uncertain about the specific side of the damaged axon during the tests, we documented escape behaviors toward both sides. After imaging for axon regeneration length, we found that a straight tail phenomenon could almost entirely predict which side of the axon was damaged. There are only a few exceptions due to the long regeneration length of axons (data not shown). Considering individual variations in larval body length, we employed the proportion of straight tails to evaluate the behavior recovery (Fig. 7c). Through correlation analysis with axon regeneration length, we discovered a significant negative correlation between the proportion of straight tails and axon regeneration length (Additional file 1: Fig. S5g, Additional file 3: Table S2). These findings suggest that the proportion of straight tails could serve as a novel, reliable indicator for evaluating the effect of Mauthner axon regeneration on rapid escape behavior.

Fig. 7.

Overexpression of tmsb4x promotes the recovery of rapid escape behavior mediated by Mauthner cells. a Experimental workflow diagram. b Schematic representation of the experimental setup. c Schematic diagram illustrating the proportion of straight tails. α represents the total length of the zebrafish, and β represents the length of the straight tail. Scale bar, 1 mm. d Representative image sequences of Mauthner-mediated C-start escape behaviors upon tmsb4x overexpression. The asterisks indicate the images at the maximum angle. Scale bar, 1 mm. e to g. Statistical analysis of the proportion of straight tails (e), time to maximal angle (f), and maximal angle (g) of Mauthner-mediated C-start escape behaviors upon tmsb4x overexpression. Assessed by unpaired, two-tailed t test, p = 0.043 (e), p = 0.031 (f), p = 0.58 (g). Sample sizes: control, n = 8; tmsb4x_oe, n = 4

To assess whether Tβ4 could aid in the functional restoration of rapid escape behavior mediated by Mauthner cells, we subsequently focused solely on recordings containing a straight tail phenomenon. Compared with those of the control group, our results showed that larvae overexpressing tmsb4x exhibited a significantly reduced the proportion of straight tails (Fig. 7d, e, Additional file 3: Table S2), and the time to maximal angle was significantly shorter than that of the control group (Fig. 7f, Additional file 3: Table S2), while the maximal angle was comparable to that of the control group (Fig. 7g, Additional file 3: Table S2). These results indicate that the overexpression of tmsb4x can effectively promote the behavioral recovery of damaged axons.

Discussion

Our work revealed the role of Tβ4 in central axon regeneration in vivo. Previous studies have reported conflicting conclusions regarding the role of Tβ4 in central axon regeneration. However, these studies were conducted in vitro and may not accurately reflect the actual situation. In this study, we used an in vivo single axon injury model of Mauthner cells in zebrafish to test the effect of Tβ4 on axon regeneration.

Previous studies have indicated that the regenerative capacity of Mauthner axons is limited following SCI in adult zebrafish [23, 50, 51]. However, in juvenile zebrafish, particularly after targeted damage to single axons using a two-photon laser, Mauthner axons exhibit remarkable regenerative abilities [45, 46, 52, 53]. This regenerative characteristic may be attributed to the advantages of younger age, as central axons in mammals are also capable of regeneration during early development. Additionally, the minimal extent of damage caused by two-photon laser ablation may further contribute to this regenerative potential. The single axon injury model of Mauthner cells is particularly advantageous for investigating the intrinsic regulatory mechanisms of CNS axon regeneration due to its minimal interference from cell heterogeneity, minimal damage to adjacent tissues, and capacity to directly link phenotypes to individual axon regeneration phenotypes.

Here, we demonstrated that Tβ4 is upregulated not only in zebrafish but also in mice following axonal injury, with significantly greater expression noted in neurons with successfully regenerated axons. We established the sufficiency and necessity of tmsb4x for in vivo regeneration of Mauthner axons. And this promoting effect is reliant on the interaction between Tβ4 and G-actin. Pharmacological experiments confirmed this mechanism. Moreover, our results suggest that Tβ4 promotes actin polymerization through its binding to G-actin, rather than facilitating depolymerization. This inference is supported by the observation that Lat. B inhibits actin polymerization. Finally, behavior test showed that the overexpression of tmsb4x promoted the functional recovery of rapid escape mediated by Mauthner cells.

In this study, we successfully generated tmsb4x knockout zebrafish, akin to previous reports in mice, in which tmsb4x knockout did not lead to defects in survival or reproduction [54]. Our findings, consistent with previous observations in cultured mouse spinal cord neurons, support the role of tmsb4x in promoting axon regeneration [19]. Notably, the impact of tmsb4x on axon regeneration in snails seems to be inverse, potentially attributed to species-specific variations [18]. Previous studies have shown that cultured central neurons treated with Lat. B in medium containing chondroitin sulfate proteoglycans (CPGSs) exhibited enhanced axon growth, contrary to our experimental results. However, we note that this effect was observed only in the presence of CPGS; Lat. B treatment did not promote axonal regeneration in the absence of CPGS [55]. Our results further support the notion that Tβ4 binding to G-actin does not inhibit actin assembly but rather accelerates it. Importantly, we introduce a novel indicator—the proportion of straight tails—observed in the rapid escape behavior following Mauthner axon damage. Lesions too close to the cell body may result in a less pronounced straight tail [46], while direct cell body lesions abolish C-start capability altogether [56]. This new indicator underscores potential limitations in prior studies, which failed to differentiate between Mauthner-mediated and non-Mauthner-mediated rapid escape behaviors [46, 52].

Previous studies have suggested that actin polymerization may inhibit neurite outgrowth and development by suppressing microtubule extension [57–59]. However, recent research challenges this notion, as actin polymerization inhibitors promote axon regeneration at low concentrations but inhibit regeneration at high concentrations. This may be due to the greater impact of Lat. B on contractile actin at low concentrations [60]. In this study, we assessed the impact of Lat. B immersion of zebrafish larvae and found that Lat. B led to larval death even at low concentrations. This may contribute to the long duration of treatment. Therefore, we chose concentrations of 10, 20, and 50 nM for the immersion treatment. Additionally, we investigated the effects of high concentrations of Lat. B by delivering 1 μM and 8 μM Lat. B into individual Mauthner cells. We selected these concentrations on the basis of the minimal quantity delivered into the cells via electroporation, which would dilute the final concentration within the cells. Our findings demonstrate that both concentrations have inhibitory effects on axon regeneration, supporting the necessity of actin polymerization for axon regeneration.

As a polypeptide, Tβ4 has advantages such as high specificity, low cost, and low immunogenicity. Future research should further investigate its potential effects in mammals, such as mice and primates. It is possible that Tβ4 could be a therapeutic agent for injuries within the central nervous system.

Conclusions

Our research reveals the mechanism of Tβ4 in promoting axon regeneration in the central nervous system of zebrafish (Fig. 8). Additionally, we identified a novel and reliable indicator for assessing the functional recovery of rapid escape behavior mediated by Mauthner cells. This discovery offers a new potential target for therapeutic interventions aimed at central axon regeneration.

Fig. 8.

tmsb4x promotes axon regeneration of Mauthner cells in vivo. The overexpression of tmsb4x in single Mauthner cells enhances Mauthner axon regeneration by facilitating actin polymerization through binding to G-actin. This overexpression accelerated the rapid-escape function recovery mediated by Mauthner cells

Methods

Animals

The zebrafish strains used in this study included wild-type (WT)/AB and Tg(tol-056) (obtained from RIKEN, Japan). These zebrafish were maintained in a fish facility with a constant temperature of 28 °C and a 14/10 light/dark cycle. The fertilized eggs were kept in E3 media and incubated in a 28 °C constant temperature chamber. From 24 to 48 h post-fertilization (hpf), 0.003% N-phenylthiourea (PTU, Sigma‒Aldrich, USA) was added to the culture dishes to inhibit pigmentation. Zebrafish were anesthetized with 160 mg/L ethyl 3-aminobenzoic methanesulfonate (MS-222, Sigma‒Aldrich, USA) for experimental procedures. The University of Science and Technology of China (USTC) Animal Resources Center and the University Animal Care and Use Committee provided the rules and regulations that all experiments implemented. All protocols were approved by the Committee on the Ethics of Animal Experiments of the USTC (permit no. USTCACUC1103013).

Bioinformatics analysis

To investigate the expression changes of Tβ4 in zebrafish and mouse axonal injury models, three public datasets (GSE202155, GSE154659, and GSE182868) were obtained from the GEO database [30–32]. In the optic nerve axonal regeneration dataset (GSE202155), only the transcriptome data of successfully regenerated axons (PSCCreMRpos) were compared to those of failed axons (SCCreMRneg). For the second dataset (GSE154659), only sciatic nerve contusion samples were used to analyze tmsb4x expression. In the third dataset (GSE182868), all samples were included for tmsb4x expression analysis. None of these datasets distinguished between neuronal subtypes. After extracting the expression matrices and grouping information, we constructed Seurat objects (v4.3.0). Following standard processing procedures, the “scale.data” was extracted, and the package “ggplot2 (v3.4.3)” was used to visualize the changes in the expression of tmsb4x. Differential analysis was conducted using the “wilcox.test” method from the “ggsigif (v0.6.4)” package.

Pearson’s correlation coefficients were used to assess the co-expression relationships between tmsb4x and l1cama, l1camb, as well as actb1, based on RNA-seq data from injured adult zebrafish spinal cords (GSE182868). The “stat_cor” of “ggpubr (v0.6.0)” package and “geom_smooth” were used in the analysis.

Construction and identification of tmsb4x knockout zebrafish

The coding sequence (CDS) of tmsb4x was input into the website http://crispor.tefor.net/crispor.py to obtain predictive results. A target site, AATCTCGAGGAGGTGACCAG, with high scores and proximity to the start codon, was selected. The in vitro transcription template, which contains the T7 promoter, target sequence, and scaffold sequence, was amplified via PCR. Next, this template was used to generate the sgRNA using the MEGAshortscript™ T7 Transcription Kit (AM1354, Thermo Fisher, USA). Subsequently, the sgRNA (300 ng/μl) was complexed with recombinant Cas9 protein (0.3 μg/μl) (PC1350, Inovogen, China) at 37 °C for 10 min to form the sgRNA/Cas9 ribonucleoprotein (RNP) complex. This RNP was then microinjected into the one-cell stage embryos of WT zebrafish to generate tmsb4x mutant chimeras. Subsequent hybridization produced homozygous mutant strains of tmsb4x-deficient zebrafish. The tmsb4x mutant zebrafish were identified through Sanger sequencing and digestion with the restriction enzyme XhoI (1635, Takara, China).

Single-cell electroporation in vivo

The overexpression system consisted of two plasmids, CMV-Gal4 and UAS-tmsb4x-mCherry. In vivo single-cell electroporation experiments were conducted as previously described [61]. In brief, a single-cell electroporation apparatus was used to apply a sinusoidal voltage to a glass electrode containing plasmids (300 ng/μl each) in proximity to the membrane of the Mauthner cell. This facilitated the delivery of plasmids into the cell. After 2 days, the mCherry-positve larvae were screened and conducted axon injury.

Single axon injury using a two-photon laser

All axon injuries were performed at 6 dpf. Anesthetized zebrafish larvae were immobilized in 1% low-melting agarose. Using a laser scanning confocal microscope equipped with a two-photon laser (LSM980, Carl Zeiss, Germany), axons above the cloaca pore were precisely ablated under a 25 × oil immersion objective using an 800 nm wavelength two-photon laser with a 20–30% power.

Free-swimming test

Wild-type and tmsb4x mutant zebrafish at 7 dpf were placed individually in a 24-well plate, and their free-swimming ability was assessed using the Viewpoint system (Viewpoint, Lyon, France) at room temperature during one hour. Total swimming distance was used as a measure of free-swimming ability. Animals of each group were randomly selected.

Structure prediction of protein complexes

To investigate whether deletion of the N-terminal 13 amino acids reduces the binding affinity of Tβ4 for G-actin, we constructed 3D structures of both wild-type Tβ4 and G-actin, as well as mutated Tβ4 and G-actin, using AlphaFold2 (https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb). Afterwards, we calculated the binding affinities of the predicted complexes at 25 °C using PRODIGY (https://rascar.science.uu.nl/prodigy/). The top-ranked structures were used for the calculations. All the parameters were set to their default values.

Pharmacological treatment

For continuous immersion treatment, zebrafish larvae at 6 dpf with single axon injury were exposed to 10, 20, or 50 μM latrunculin B (76343–94-7, Maokang, China) for 48 h until imaging for axon regeneration. The immersion medium was replaced every 24 h to maintain the effectiveness of the drug. The control group was treated with an equivalent concentration of DMSO (D12345, Thermo Fisher, USA).

For single-cell drug treatment, at 5 dpf, 1 or 8 μM Lat. B was mixed with long-term dye (D3308, Thermo Fisher, USA) and delivered into the Mauthner cells. The treated axons were ablated above the cloaca pore at 6 dpf. Axonal regeneration lengths were then imaged at 24 and 48 hpi.

In vivo imaging

Anesthetized larvae were fixed in 1% low-melting agarose and imaged for axon regeneration using a 40 × objective on a confocal microscope (FV1000, Olympus, Japan), with images captured every 3 μm. Subsequently, image stitching was performed using Adobe Photoshop CC 2018 (Adobe, San Jose, CA, USA), with the image fusion mode set to “lighten”. ImageJ software (National Institutes of Health, Bethesda, MD, USA) was used to measure the regeneration length and total length of Mauthner axons, as well as the soma area. The axon regeneration length is defined as the straight-line distance from the site of axonal injury to the tip of the longest regenerating axonal branch. The number of branches is defined as the total count of any visible branches extending from the regenerating axons.

Quantitative real-time PCR (qPCR)

The tmsb4x overexpression plasmids (30 ng/μl each) were microinjected into one-cell stage embryos. mCherry-positive larvae were screened at 1 dpf. Total RNA was extracted from mCherry-positive larvae at 4 dpf. Each biological replicate contained 30 larvae. Each sample was detected with 3 technical replicates. The expression level for each biological replicate was determined by averaging the results of three technical replicates. Subsequently, 1 μg of total RNA was reverse transcribed into cDNA using HiScript II qRT SuperMix II (R223, Vazyme, China). qPCR was performed using ChamQ Universal SYBR qPCR Master Mix (Q711, Vazyme, China) on a real-time quantification system (Light-Cycler 96, Roche). The expression levels of each gene were normalized to those of β-actin using the 2^–ΔΔCt method. The primer sequences are listed in additional file 2: Table S1.

Rapid escape behavior testing

At 48 hpi, a high frame rate camera (Revealer, China) capable of recording at 1000 frames/second was used to capture the rapid escape behavior of the zebrafish. A speaker was used to emit a 500 Hz, 20 ms sound stimulus to trigger the escape response in zebrafish. As the side of axonal injury was unknown beforehand, both the leftward and rightward escape behaviors of each fish were recorded separately. However, only behaviors containing a straight tail phenomenon were analyzed to evaluate the effect of tmsb4x overexpression on functional recovery. Analysis of the maximal angle, time to maximal angle, and proportion of straight tails was conducted using software provided with the camera (Revealer, China). The maximum angle refers to the degree of rotation of the larva's midline from the onset of the C-start to the point where it reaches the maximum turning angle. The proportion of straight tails was defined as the length of the straight tail divided by the total body length.

Statistical analysis

GraphPad Prism 9 (San Diego, USA) was used for statistical analysis. Datasets with a normal distribution were analyzed with parametric statistical tests (t-test or ANOVA), while datasets without a normal distribution were analyzed with the Mann‒Whitney or Kruskal‒Wallis nonparametric tests. Detailed statistical data and methods are provided in the figure legends. The data are presented as the mean ± standard error of the mean. Statistical significance is denoted as follows: * p < 0.05; ** p < 0.01; *** p < 0.001; ns, not significant (p > 0.05). No statistical methods were used to predetermine the sample size. The required sample sizes were estimated on the basis of our previous experience performing similar experiments. Zebrafish larvae were randomly assigned to different groups. Drug treatment and behavioral experiments were performed in a blinded manner.

Abbreviations

ANOVA

Analysis of variance

bp

Base pair

CDS

Coding sequence

CPGS

Chondroitin sulfate proteoglycans

dpf

Days post-fertilization

dpi

Days post-injury

EGFP

Enhanced green fluorescence protein

hpf

Hours post-fertilization

hpi

Hours post-injury

Lat. B

Latrunculin B

MR

Micro-Ruby

qPCR

Quantitative real-time PCR

RGCs

Retinal ganglion cells

SCI

Spinal cord injury

Tβ4

Thymosin beta 4

wpi

Weeks post injury

WT

Wild-type

Authors’ contributions

Conception and design: Z.S., B.H.; acquisition of data: Z.S., A.L.H.; data analysis: Z.S.; manuscript draft and revision: Z.S., B.H. All authors have read and approved the final version of the manuscript.

Funding

This work was funded by the “Research Funds of the Center for Advanced Interdisciplinary Science and Biomedicine of IHM” (QYZD20220002), the National Natural Science Foundation of China, No. 82071357, and the Ministry of Science and Technology of China, No. 2019YFA0405600 (to B.H.).

Data availability

All the data generated or analyzed during this study are included in this published article and its supplementary information files. The high-throughput sequencing data can be found in the GEO database (https://www.ncbi.nlm.nih.gov/geo) according to the accession numbers (GSE202155, GSE154659, and GSE182868) [30–32]. Supplementary figures and figure legends are presented in additional file 1. Primers sequences are included in additional file 2. Individual data values for each statistical chart are included in additional file 3. Image of the original gel is presented in additional file 4. tmsb4x mutant lines can be accessed by contacting the corresponding author.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.