Xenopus laevis embryos were obtained by natural mating of wild-type male and female frogs. Animal husbandry was performed as previously described (Edwards-Faret et al., 2017; Slater and Larraín, 2021). Animals were grown up until Nieuwkoop and Faber stages 50–51 (NF 50–51) (R-stages) for experiments (Nieuwkoop, 1994). All animal procedures were approved by the Scientific Ethics Committee for the Care of Animals and Environment of the Pontificia Universidad Católica de Chile and Bioethical and Biosafety Committee of Universidad San Sebastián (Protocol 181017006 and 210504022).

R-animals were anesthetized with 0.02% of tricaine mesylate (MS222) dissolved in 0,1x Barth solution (8.9 mM NaCl; 102 μM KCl; 238.1 μM NaHCO₃; 1 mM 4-(2- hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES); 81.14 μM MgSO₄; 33.88 μM Ca(NO₃)₂; 40.81 μM CaCl₂, pH 7.6). Spinal cord transection was performed as previously described (Edwards-Faret et al., 2017; Slater and Larraín, 2021). Briefly, the skin and dorsal muscles were opened at the mid-thoracic level, and the spinal cord was fully transected with a clean cut at the thoracic level with micro-spring scissors. After surgery, animals were transferred into their tanks with 0.1 x Barth supplemented with antibiotics (penicillin and streptomycin).

For transmission electron microscopy (TEM) analysis, uninjured and transected animals (6 h post-transecting (hpt), 1dpt, and 2dpt) were fixed with a 2% PFA and 2.5% glutaraldehyde solution at 4 C overnight. Spinal cords were processed by the Advanced Microscopy Facility UMA-UC, Pontificia Universidad Católica de Chile. Images of ultra-thin sections were acquired in a transmission electron microscope Philips Tecnai 12 at 80 kV, at the Advanced Microscopy Facility UMA-UC, Pontificia Universidad Católica de Chile. Three independent experiments were performed.

Uninjured animals and animals at 6hpt, 1dpt and 2dpt were anesthetized in 0.02% MS222. The entire caudal segment of the spinal cord, from 10 to 15 animals, was isolated from the injury site with forceps and collected in 1X Marc’s Modified Ringer solution (MMR) (0.1 M NaCl, 2.0 mM KCl, 1.0 mM MgSO4, 2.0 mM CaCl2, 5.0 mM HEPES, pH 7.4). To disaggregate the tissue, the spinal cords were incubated in 100 uL of 1X papain in 1X MMR–Ca²⁺, Mg²⁺ free (1X MMR–CMF) for 10 min, homogenizing with a pipette every 5 min. The papain reaction was stopped with the same volume (100 uL) of Fetal Bovine Serum. Later, the samples were centrifuged at 1,000 g for 5 min, and the cells were resuspended in 100 uL of 1X MMR - CMF. To quantify the cells, a solution with trypan blue was prepared at a 1:10 dilution and the cells were quantified in a Neubauer chamber under a light microscope at 10 X magnification.

The spinal cord of EGFP + animals at R-stage from the transgenic line zGFAP::EGFP, expressing EGFP specifically in NSPCs (Edwards-Faret et al., 2021), was transected. The spinal cord cells dissociated at 6 hpt, 1dpt, 2dpt and from uninjured animals (15 animals per condition). Glass coverslips were coated with 100 µg/mL Poly-L-lysine for 1 h at 37°C, and 50 µL of dissociated cells were seeded. After 5 min, the cells were loaded for 10 min at room temperature (RT) with 20 nM Tetramethyl rhodamine ethyl ester (TMRE) probe, and the cell media was changed for a media containing 0.25% bovine serum albumin and 0.2 nM TMRE. Images were obtained at the Nikon Eclipse Ti microscope using the following configuration: first, a single image was obtained using ex. 488 nm—em. 500/550 nm for detecting FGP positive cells, and ex. 540 nm—em. 620/60 nm for detecting TMRE. Later, it was obtained 1 image every 1.5 s for 5 min, using ex. 540 nm—em. 620/60 nm. 10 μM FCCP was used to induce loss of mitochondrial membrane potential (ΔΨ). Resting membrane potential was calculated as ΔF_basal—F_FCCP using absolute fluorescence values.

Phosphofructokinase (PFK) activity was measured with colorimetric kits (MAK093, Sigma; ab155898, Abcam) following the supplier’s protocol. Briefly, spinal cord cells were dissociated at 6 hpt, 1dpt, and 2dpt, and from uninjured animals (15 animals per condition), and between 500,000 and 1,000,000 cells were centrifuged at 1,000 g for 5 min, the supernatant was discarded, and cells resuspended in 100 µL cold assay buffer. After pipette homogenization, samples were incubated on ice for 10 min and centrifuged at 12,000 g for 5 min at 4°C, and the equivalent of 10,000 cells was added to designated wells together with the reaction mix. Readings were taken in a Synergy H1 Hybrid Reader (Biotek) preheated to 37°C, with measurements every 5 min at 450 nm for 20–60 min. Three independent experiments were performed.

ATP levels were measured with CellTiter-Glo™ 2.0 Assay (Promega) following the supplier’s protocol. Briefly, spinal cord cells were dissociated, and 80,000 cells were added to the designated wells together with the reaction mix. Luminescence readings were obtained in an Infinite M200 pro (Tecan). Four independent experiments were performed. Oligomycin was used as a control.

Uninjured animals were sacrificed and fixed in 4% paraformaldehyde for 2 h at RT or overnight at 4°C. The samples were dehydrated with a sucrose gradient, cryoprotected in optimal cutting temperature compound, and frozen with liquid nitrogen. Longitudinal and transversal 10 µm sections were obtained using a Leica CM1850 cryostat. Samples were permeabilized in 1X phosphate-buffered saline (PBS) with 0.2% Tritonx-100 (PBST) for 10 min and blocked with 10% goat serum in PBST for 30 min. Later, the samples were incubated with the primary antibodies at 4°C overnight and secondary antibodies for 2 h at RT; after each antibody incubation, samples were washed 3 times for 10 min with PBST, followed by nuclei staining with Hoechst solution (1:10,000) for 5 min and mounted with Vectashield (H-1000-NB). Epifluorescence images were taken in an Olympus microfluo microscope IX71. Rabbit anti-COXIV (1:1000, Abcam, ab16056) was used to detect mitochondria, Mouse anti-SOX2 (1:200, Cell Signaling Techonology, 2748 S) was used to detect NSPCs, and rabbit anti-NeuN (1:1000, SDIX) was used to detect mature neurons. The secondary antibodies donkey anti-mouse AlexaFluor^® 488 and donkey anti-rabbit AlexaFluor^® 594 (1:500, Jackson Immuno Research), goat anti-rabbit AlexaFluor^® 488 and goat ant-mouse AlexaFluor^® 594 (1:500, Invitrogen). Nuclei were stained with Hoechst (1:10.000) and mounted with vectashield (Vector Laboratories, H-1000).

Uninjured animals and animals at 2, 6, and 10 hpt, and 1 and 2dpt were anesthetized in 0.02% MS222. The entire caudal segment of the spinal cord, from 10 to 15 animals, was isolated from the injury site with forceps, and total RNA was obtained with the commercial kit (RNeasy Mini Kit) following the supplier’s protocol. RNA concentration was measured using Nanodrop (Thermo Scientific). The cDNA was synthesized using the M-MLV reverse transcriptase (Promega), and qPCR was performed using Power SYBR Green (Applied Biosystems) or Maxima SYBR Green (Thermo Scientific) by performing three technical replicates. The relative expression ratio was then calculated as described (Pfaffl, 2001), using eef1a1 (GenBank: BC043843) as the reference gene. The primers used were mfn2, opa1, mtfr, mtfp, hk2, pfkfb1, and pklr. For the sequence of the primer, please see Table 1.

Uninjured and transected animals (6 hpt, 1dpt, and 2dpt) were anesthetized and sacrificed. The stomach and part of the tail were removed to obtain the correct positioning of the animal. Using micro-surgical forceps, excess water was blotted off with absorbent paper. The animal was then coated in Optimal Cutting Temperature (O.C.T.) compound and placed on a cork for quickly freezing the tissues. For the freezing procedure, already published protocols were adjusted (Meng et al., 2014; Kumar et al., 2015). Briefly, liquid nitrogen was placed in a container designed for handling extreme temperatures, and a stainless-steel cup with Isopentane (2-methylbutane) was submerged. The Isopentane was left to cool for a few minutes until about 30% of it had frozen, and the cork was immersed in Isopentane for 25 s until fully frozen, then rested on dry ice for 30 min to aid in the evaporation of Isopentane and maintaining the temperature, preventing thawing or artifacts. Later, the sample was wrapped in aluminum foil to prevent crystal formation and stored at −80°C with dry ice. Longitudinal 10 µm sections were obtained using a Leica CM1850 cryostat. Finally, for the COX and SDH histochemistry, an already published protocol was adjusted (Ross, 2011). Briefly, the slides containing the spinal cord sections were thawed for 30 min at RT, incubated with 1X DAB, 8 M cytochrome C in 0.1 M PBS, 2 mg/mL catalase, 220 mM sucrose for 2 h at RT, washed with distilled water, and incubated with 1.2 mM PMS, 1.6 mM NTB, 1 M sodium succinate, sodium azide/EDTA/PO4 (1.9 mM sodium azide, 1.5 mM EDTA 32.5 mM NaH2PO4·H2O, 217 mM Na2HPO4 ·7H2O) for 30 min at 37°C, and washed with distilled water. Later, the samples were fixed with 10% formalin for 10 min and mounted with a pre-warmed glycerin gelatin solution. 1.25 mM sodium azide was used as a negative control for COX histochemistry, and succinate was replaced by malate as a negative control for SDH. Images were obtained using an inverted confocal microscope Nikon spectral ECLIPSE (C2Si), with a 40X objective.

Image acquisition and analysis were performed using Fiji (ImageJ) software. Statistical analyses were conducted using GraphPad Prism 8 software. A one-way ANOVA followed by Dunnett’s multiple comparison test was used to determine the significance of multiple comparisons of normally distributed data. For nonparametric multiple comparisons, the Kruskal–Wallis test, followed by Dunn’s multiple comparison test, was employed. To compare the mean of log2-transformed measurements to a single control value (considered as 0 for non-fold-change data), a one-sample t-test was applied. Statistical significance was considered as follows: ****(p < 0.0001); ***(p < 0.001); **(p < 0.01); *(p < 0.05); ns = not significant (not statistically different).

We first sought to characterize the localization of mitochondria within the spinal cord in the regenerative stages (R-stages) of Xenopus laevis. Spinal cord transversal cryosections were obtained (Figure 1A’) and immunodetection of the mitochondria utilizing the mitochondrial marker CoxIV was performed. Mitochondria were observed mainly in the region adjacent to the central canal (Figure 1B). The spinal cord is composed primarily of NSPCs surrounding the central canal (Muñoz et al., 2015; Edwards-Faret et al., 2018), and astrocytes and neurons are also present in the gray matter (Miller and Liuzzi, 1986; Maier and Miller, 1995). Therefore, spinal cord longitudinal cryosections were obtained (Figure 1A). Immunodetection of NSPC (Sox2+) and neurons (NeuN) was performed, showing that Sox2+ cells correspond to a couple of layers of cells localized surrounding the spinal cord central canal. In contrast, NeuN + cells utilize a more extended region of the spinal cord gray matter but are localized further from the central canal (Figure 1C). Therefore, co-detection of CoxIV and Sox2 was performed, confirming that most of the mitochondria present in the spinal cord are localized in the NSPCs, presenting an apical distribution (Figure 1D). The localization of the mitochondria in the cells surrounding the central canal and the apical distribution of the mitochondria in the spinal cord was confirmed by electron microscopy of longitudinal (Figure 1E) and transversal (Figure 1F) sections.

The number and localization of mitochondria within cells are typically regulated based on cellular requirements. Therefore, mitochondria from cells surrounding the central canal, which correspond to NSPCs, were analyzed in longitudinal spinal cord cryosections in uninjured control animals and at specific time points post-transection (6, 12, 24, and 48 h). Time points for analysis were selected based on prior findings and published data. Lee-Liu et al. (2014) showed that at 24 hpt, metabolic transcript levels are most regulated. Additionally, Peñailillo et al. (2021) reported early transcript regulation related to nutrient response starting 1–2 h post-transection (hpt), with elevated levels persisting until 8 h. Even more, cellular events influenced by mitochondrial and metabolic activity, such as cell death, proliferation, and immune response, were primarily observed between 6–48 hpt. Therefore, the analysis began at 6 hpt, coinciding with the first wave of apoptosis in the spinal cord, a process involving mitochondria (Liu et al., 1997), and continued until 48 hpt. By this time, metabolic transcript levels, which are most regulated at 24 hpt (Lee-Liu et al., 2014), were expected to influence mitochondrial characteristics (Figure 2A). The number of mitochondria was quantified in each cell section adjacent to the central canal. A region of interest (ROI) was defined by tracing the plasma membrane of cells in contact with the central canal (Figure 2B), and mitochondria within the ROI were counted. Cells partially visible in the image were excluded. The number of mitochondria per cell section temporarily decreased, from 6 to 24 hpt. Basal levels were recovered at 48 hpt (Figure 2C). Additionally, each cell adjacent to the central canal was divided into three sections: apical, from the central canal to the beginning of the nucleus; medial, considering all the area occupied by the nucleus; and basal, from the end of the nucleus to the inner plasma membrane, and the localization of the mitochondria was analyzed (Figure 2D). Uninjured animals present 92% of the mitochondria in the apical zone, while redistribution of the mitochondria was observed at all time points after transection, showing nearly 50% of the mitochondria in the apical zone, and the other 50% redistributed almost equally towards the medial and basal zone (Figures 2E, F).

Mitochondria are dynamic organelles that adjust their morphology (Wai and Langer, 2016) and phenotype (Hackenbrock, 1966) in response to environmental changes and metabolic demands. To assess these adaptations, the mitochondrial area, circularity, and phenotype were analyzed in electron microscope images at specific time points post-transection (6, 12, 24, and 48 hpt) (Figure 3A). An increase in mitochondrial area was observed starting at 6 hpt, which returned to baseline by 48 hpt (Figure 3B), with the mitochondrial area approximately doubling from ∼0.3 µm² to ∼0.6 µm². Mitochondrial circularity can be determined by the ratio of the minimum Feret diameter to the maximum Feret diameter (Max/Min Feret’s), where values closest to 1 represent an increase in circularity. An increase in mitochondrial circularity was observed starting at 6 hpt, which returned to baseline by 48 hpt (Figure 3C), with a change in the mitochondrial Max/Min Feret´s from ∼2.0 to ∼1.5. Mitochondria were classified into three different phenotypes: orthodox, condensed, and swollen, based on previous work (Hackenbrock, 1966). Orthodox mitochondria present less defined cristae with intermediate matrix density; the condensed phenotype presents increased cristae width, with more defined cristae and intermediate matrix density; and the swollen phenotype presents a deployed internal membrane, remnant tubular cristae, dispersed matrix granules responsible for a decrease in electron density, and a two to three-fold size increase (Figure 3D). Mitochondria in uninjured animals are mostly orthodox ∼72%, while ∼27% show a condensed phenotype, and only ∼1% a swollen phenotype (Figure 3D). The mitochondria change their phenotype starting at 6 hpt and return to levels similar to the uninjured animals at 48 hpt, an increase in the swollen phenotype from ∼1 to 70%–80%, and a decrease in the orthodox phenotype from ∼72 to ∼20%, was observed.

Mitochondrial morphology results from a balance between two dynamic and opposing processes: mitochondrial fusion and fission. The morphology changes when the balance between these processes is altered (Tilokani et al., 2018; Chan, 2020). Since a change in mitochondrial morphology was observed after SCI, fusion- and fission-related transcript levels were analyzed by RT-qPCR at specific time points post-transection (2, 6, 10, 24, and 48 hpt). The protein Optic Atrophy 1 (Opa-1), a GTPase of the dynamin family, participates in the fusion of the inner mitochondrial membrane, and Mitofusin 2 (Mfn2), a transmembrane GTPase participates in the fusion of the outer mitochondrial membrane, while the protein Mitochondrial Fission Process 1 (Mtfp1) is localized at the inner mitochondrial membrane, and Mitochondrial fission regulator 1 (Mtfr1) localized in the outer mitochondrial membrane and is involved in the process of mitochondrial fission (Wai and Langer, 2016; Tilokani et al., 2018; Chan, 2020). Therefore, the transcript levels of mfn2 and opa1, and mtfp1 and mtfr1 were analyzed as a readout of mitochondrial fusion and fission processes regulation, respectively. Mfn2 (Figure 4A), opa1 (Figure 4B), and mtfp (Figure 4C) transcript levels remained unaltered during all the analyzed time points, while mtfr presented an acute and transient increase in the transcript levels at 24 hpt (Figure 4D).

Mitochondrial phenotype changes depending on the energy requirement of the cell. Orthodox phenotype is observed generally in intact tissue and is related to slow substrate consumption and low respiratory rate, while the condensed phenotype is observed when mitochondria activate oxidative phosphorylation and present a high respiratory rate, and the swollen phenotype is observed when no oxidative phosphorylation is occurring (Hackenbrock, 1966). It has been described that the mitochondrial swollen phenotype could be linked to mitochondrial uncoupling, where mitochondria might alter their phenotype under conditions that affect oxidative phosphorylation. However, if the mitochondrial organization were to reach a point where the swollen phenotype became irreversible and phosphorylation integrity was lost, the mitochondria would be considered damaged (Hackenbrock, 1966). Given that the swollen phenotype was observed following SCI, a Cytochrome C Oxidase/Succinate Dehydrogenase (COX/SDH) double-labeling histochemistry was conducted to assess whether this phenotype change is associated with mitochondrial dysfunction. The catalytic subunits of COX are encoded by mitochondrial DNA; therefore, COX assembly and activity are dependent on functional mitochondria with intact DNA. While SDH is encoded by nuclear DNA; therefore, SDH activity is not dependent on mitochondria integrity (Ross, 2011). At the histological level, in the presence of mitochondrial dysfunction, double staining with the COX/SDH activity reveals cells that are positive for SDH, appearing in a bluish-purple color, with minimal or no brown COX staining (Ross, 2011). In longitudinal cryosections of uninjured animals, the cell layers closest to the central canal present a coloration with a predominant bluish-purple color and a brown color concentrated in the apical zone of the cells surrounding the central canal, coincident with the localization of mitochondria (Figures 1D, E, 5A, A’). This characteristic distribution was mostly maintained after SCI at all time points analyzed (Figures 5B–D, B’–D’). Thus, the activity of COX and SDH remained stable upon spinal cord injury. Additionally, the cellular energy level was evaluated by measuring ATP levels, which is indicative of metabolic active cells. ATP levels remained unchanged after SCI at all analyzed time points (Figure 5E). These results suggest that minimal changes in mitochondria function are observed after SCI in X. laevis regenerative stages.

The results from the previous section showed that the observed mitochondrial swollen phenotype did not result in mitochondrial damage, prompting us to test whether this phenotype was associated with mitochondrial uncoupling. A characteristic of mitochondrial uncoupling is a decrease in the mitochondrial membrane potential (Δψm). However, this effect is generally transient, as a prolonged decrease could lead to pathological consequences (Zorova et al., 2018). Since the mitochondrial swollen phenotype was observed in cells surrounding the spinal cord central canal, which correspond to NSPCs, the mitochondrial membrane potential was evaluated in NSPCs and non-NSPCs within the spinal cord of uninjured and transected animals at defined time points post-transection (6, 24, and 48 hpt). The Xenopus laevis Xla. Tg (Dre.gfap:EGFP)^Larra transgenic line was used, as it expresses GFP under an NSPC-specific promoter (Edwards-Faret et al., 2021), allowing for the identification of NSPCs GFP+ and non-NSPCs GFP- cells (Figure 6A). At the defined time points, the spinal cord was dissected, cells were dissociated and plated on a coverslip. Cells were then treated with tetramethylrhodamine (TMRE), a dye that accumulates in the mitochondria in a Δψm-dependent manner. Protonophores such as Carbonyl cyanide m-chlorophenylhydrazone (CCCP) reduce the Δψm and thereby induce the dissipation of TMRE from the mitochondria. The combined use of TMRE and CCCP enables the assessment of Δψm (Figure 6A). A decrease in the Δψm was observed in NSPC GFP+ cells at 6 hpt, returning to basal levels at 24 hpt (Figures 6B–D). At the same time, no changes were detected in non-NSPC GFP+ cells (Figures 6C, E), suggesting that the mitochondrial uncoupling occurs as a specific and transient response of NSPCs to SCI.

The Δψm is essential for ATP production; if Δψm is decreased, the mitochondria produce less ATP, resulting in a decrease in the cytosolic ATP/ADP ratio, which enhances glycolysis to maintain the ATP levels in the cell (Maldonado and Lemasters, 2014). As ATP levels were not affected by SCI, but a transient decrease in Δψm was observed, the activation of glycolytic metabolism was evaluated. In glycolysis, three irreversible reactions, catalyzed by the rate-limiting enzymes hexokinase (Hk), phosphofructokinase-1 (Pfk1), and pyruvate kinase (Pkr), regulate the rate of the glycolytic pathway (Chandel, 2021). Additional allosteric regulation of the glycolytic rate occurs by another enzyme: 6-phosphofructo-2-kinase/fructose 2,6-bisphosphatases (Pfkfb), which control the levels of the allosteric activator of Pfk1, fructose 2,6-bisphosphate (Fru-2,6-P2) (Ros and Schulze, 2013). To assess glycolysis regulation, transcript levels of hk2, pfkfb1, and pklr were analyzed by RT-qPCR at specific post-transection time points (2, 6, 10, 24, and 48 hpt) (Figure 7). The transcript levels of hk2 showed an early and transient increase from 6 to 10 hpt (Figure 7B), while pfkfb1 transcript levels exhibited a sustained increase across all analyzed time points (Figure 7C). In contrast, pklr transcript levels remained unchanged (Figure 7D). It is important to consider that changes in cellular metabolism are often rapid and dynamic. Therefore, glycolytic activity was confirmed by measuring Pfk enzymatic activity. A fast and transient increase in Pfk activity was observed at 6 hpt, with levels returning to baseline by 24 hpt (Figure 7E). This together allows us to suggest that there is an early increase in the glycolytic rate on the spinal cord after SCI.