All animal experiments adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and to the United Kingdom Animals (Scientific Procedures) Act of 1986. All methods were approved by the University of Cambridge Animal Ethics Committee and conform to the “Guiding Principles for research Involving Animals and Human” as adopted by the American Physiological Society. Adult female Sprague–Dawley rats (250 g–300 g, Harlan Laboratories, Inc.) were used in the present study, and they were housed on a 12-h light–dark cycle with ad libitum access to food and water. Animals were sacrificed humanely by exposure to CO₂. For a 5-liter chamber, a flow rate corresponding to 50% CO₂ was applied. Both RGCs and DRG cultures were established from the same animal. Once the RGCs were in culture, the DRG culture was set-up, only a few hours post-mortem. All experiments presented in the study were repeated at least 4 times.

The eyes from the adult Sprague Dawley rats were removed, their retinas dissected out and dissociated using the papain dissociation kit (Worthington, Lakewood, NJ, United States). Briefly, the retinas were incubated for 90 min at 37 °C with papain and DNase (1:10), and the cells were dissociated and the resulting cell suspension was transferred to a new tube for centrifugation at 300 x g for 5 min at room temperature. The supernatant was then discarded, and the pellet was resuspended in a solution containing EBSS, albumin ovomucoid inhibitor and DNase. The cell suspension was centrifuged for 6 min. at 70 x g through a discontinuous density gradient prepared with albumin ovomucoid inhibitor at room temperature. The pellet collected was subsequently re-suspended in Neurobasal A supplemented with B27 and L-glutamine (2 mM) to transfect the cells with the expression constructs encoding EB3. After transfection the cells were plated onto poly-L-lysine (PLL, 20 μg/mL) and laminin (10 μg/mL) (Sigma-Aldrich, St. Louis, MO, USA) coated plates at a density of 100,000 cells/well (in 24 well plate), and cultured in Neurobasal A medium supplemented with B27, 2 mM L-glutamine and gentamicin (1:200) (Life Technologies, Carlsbad, CA, USA). The RGCs were maintained at 37 °C in 5% CO₂ for 4 days until the EB3 recording was obtained.

DRG were dissected from the adult Sprague Dawley rats, harvested along the entire spinal cord. The DRGs were then incubated for 90 min in 0.2% type IV collagenase (Sigma-Aldrich, St. Louis, MO, USA) in HBSS medium (Life Technologies, Carlsbad, CA, USA) at 37 °C and then with 0.1% trypsin (Sigma-Aldrich, St. Louis, MO, USA) added for a further 10 min. The DRG cells were dissociated using Pasteur pipettes of decreasing diameter and the resulting cell suspension was then centrifuged at 2000 rpm for 2 min. The cell pellet was collected, resuspended, and centrifuged through a 15% BSA (Sigma-Aldrich, St. Louis, MO, USA) density gradient at 1000 rpm for 15 min. Subsequently, the cell pellet was further resuspended and centrifuged at 2000 rpm for 2 min before the cells were plated or transfected by Neon Electroporator (Thermo Fisher Carlsbad, CA, USA). For transfection, the dissociated neurons were re-suspended in 2 mL of Mg²⁺/Ca²⁺-free PBS (phosphate buffered saline, pH 7.0) before they were transfected with expression constructs encoding EB3. The cells were cultured overnight in serum/antibiotic free DMEM medium (Life Technologies, Carlsbad, CA, USA) with poly-L-lysine (20 μg/mL) and laminin (10 μg/mL) (Sigma-Aldrich, St. Louis, MO, USA). The medium was replaced next day with the normal culture medium DMEM supplemented with 10% FBS (Fetal bovine serum, Life Technologies, Carlsbad, CA, USA), 1% penicillin–streptomycin-fungizone, and 10 ng/mL NGF (Nerve growth factor, Sigma-Aldrich, St. Louis, MO, USA). The cells were maintained at 37 °C in 5% CO₂ from 1 to 3 days in culture until the EB3 recordings were obtained.

After recording the videos of the EB3 from the neurons dissected from the retina we did immunocytochemistry with antibodies directed to βIII-tubulin in order to verify that the cells transfected and recorded were specifically RGCs, even though the method used to culture RGCs specifically isolate this type of neurons and not the rest of cell types of the retina (Vecino et al., 2015).

After recording, the RGCs neurons were washed twice with PBS and fixed for 10 min with methanol at −20 °C and then washed 3 times with PBS. After blocking the binding of non-specific antigens with blocking buffer (3% BSA and 0.1% Triton X-100 in PBS), the cells were incubated with a rabbit antibody against βIII-tubulin (diluted 1:2,000; Promega, Madison, WI, USA). After washing again, this antibody was detected with an goat anti-rabbit Alexa Fluor 555 secondary antibody (Life Technologies, Carlsbad, CA, USA) diluted 1:1,000, and the cells were counterstained with the nuclear marker DAPI diluted 1:10,000 (Life Technologies, Carlsbad, CA, USA). After washing, the cells were preserved with fluoromount.

The EB3-GFP plasmids constructs have been designed to be expressed in mammals (Akhmanova et al., 2001; Hoogenraad et al., 2000; Perez et al., 1999). To clone the EB3 cDNA, gene-specific primers were designed with in-frame restriction sites for cloning into pEGFP-N1 (Clontech, Palo Alto, CA, USA) after PCR amplification. The primer sequences were based on the human EB3 cDNA, accession numbers: AA2892 (Image clone 714,028). In Western blots of COS-1 cells transfected with the cDNA all the EB3-GFP fusion protein was produced at the expected molecular weight (Hoogenraad et al., 2000; Stepanova et al., 2003).

Dissociated RGCs and DRG neurons were transfected using the Electroporator (Life Technologies, Carlsbad, CA, USA), to electroporate cells within a micropipette tip using Neon Transfection Kit (Life Technologies, Carlsbad, CA, USA; MPK1096), applying two 1,200 V pulses of 20 ms duration. Before transfection, the cells were washed briefly in Mg²⁺/Ca²⁺-free PBS and resuspended in the Buffer R provided. For optimum transfection efficiency, 1.0–1.5×10⁵ cells were electroporated with 1 μL of plasmid (500 ng/μl) in 10 μL of transfection buffer. The transfected cells were plated and cultured overnight as described above, without antibiotics or serum, but supplemented with 1% ITS+ (Biosciences, Becton Drive Franklin Lakes, NJ, USA), and it was replaced by normal culture medium the day after transfection.

Four transfections were performed on four different rats, and from the same rat, both RGCs and DRG neurons were transfected on the same day, thereby avoiding possible differences. One and 2 days post transfection (dpt) videos of EB3 were recorded of the DRG and RGCs cultures. RGC cultures were also recorded 4 days post transfection to allow the RGCs to elongate their neurites.

In RGCs and DRG cultures where the transfection was effective, mitochondria were also labelled with mitoTracker (Life Technologies, Carlsbad, CA, USA; M22426), to test the active transport of organelles after EB3 transfection. The number, speed and direction of the mitochondria was measured from the Kymographs as well as the movement of the EB3 fluorescence that also named comets do to the fluorescence appearance (Figure 1). Mitochondrial transport was used as sign of healthy conditions and was not analysed further.

The live imaging of EB3-GFP in adult DRG neurons and RGCs was performed with an Olympus IX70 microscope (Shinjuku, Tokyo, Japan) an UPlan Apo objective 100x/1.35 oil iris and using a Hamamatsu ORCA-ER CCD camera and a PerkinElmer UltraVIEW scanner for spinning disk confocal microscopy, controlled with MetaMorph software. Kymographs were generated using MetaMorph software (Figure 1).

To avoid perturbation of microtubule dynamics attributable to elevated expression of +TIP (plus-end tracking proteins), only neurons exhibiting low EB3–GFP expression were selected for imaging. Low expression was defined by the presence of discrete comet-like labeling restricted to microtubule plus ends, without detectable microtubule-shaft fluorescence or diffuse cytoplasmic signal. Cells showing higher expression were excluded from analysis.

Recordings were taken for 3 min, one frame per every 3 s, with a 491BP filter and the images were analysed by constructing kymographs of each neurite using MetaMorph software (version 7; Molecular Devices). The total width of each neurite was selected for analysis: (a) proximal (10 to 15 μm from the cell body); (b) distal (10 to 15 μm from the growth cone); (c) and in the cell body.

The following parameters of dynamic MTs were measured: density (average number of moving fluorescent comets per μm²), speed (average number of moving of EB3 comets by taking in consideration the total displacement divided by time of observation) and directionality (anterograde or retrograde) depending on the angle formed by the comet on the kymograph measured from the orientation set to 0°. To set the directionality of the microtubule polymerization (comet) or the mitochondria, the selected area to be analysed was always proximal to the cell body, thus, all objects that move from left to right, were considered anterograde while those moving from right to left were considered retrograde. Where the angle was more acute in relation to its angle with the horizontal, it meant faster movement, however, if the line described by the object was vertical, it meant no movement. These conditions were considered valid for the MTs as well as for the mitochondria.

The distribution and dynamics of mitochondria in relation to the microtubules in RGC and DRG cells were used exclusively as a control of the stage of health of the cells. For the live mitochondria staining, the cells were incubated with MitoTracker redCMXRos (Life Technologies, Carlsbad, CA, USA) and prepared according to product information. Stock solution was used immediately after it was prepared and dissolved in DMSO to a final concentration of 1 mM. The working concentration was 50 nM in cultured media. For the mitochondrial staining, the culture medium in which the cells were growing after transfection was removed. The cells were incubated for 15 min in the pre-warmed (37 °C) MitoTracker probe and washed with cultured media, live imaging was performed as previously described. Simultaneous video recording for MitoTracker and EB3 were performed with the spinning disk confocal as described above. We performed a statistical analysis of the total number of mitochondria per field in both RGCs (n = 118) and DRGs (n = 684), classifying them in relation to their motility as stationary, oscillatory or fast. For statistical analysis, mitochondria were normalized to total number per field. This analysis demonstrated the presence of all three types of movement and the overall healthy state of the cells studied.

At least three replicates were analysed for each experimental condition, from four independent experiments. Data are expressed as mean values ± standard error of the mean. The sample sizes in the text are given as n = neurites analysed, unless stated otherwise. For the values of RGCs and DRG neurons densities, n = 20 was analysed for each cell type, and for velocities, n = 127 was analysed for DRG neurons and n = 150 for RGCs. To assess whether there were significant differences in the EB3 density and transport between the groups, two tailed Student t-tests, ANOVA and Bonferroni tests were used. The homogeneity of the variances was assayed with Levene’s test (p < 0.05). The minimum value of significance for all tests was defined as p < 0.05. All statistical analyses were carried out using IBM SPSS Statistics software v. 21.0 and Microsoft Excel.

To visualize EB3 comets during neurite growth, we recorded time-lapse videos of DRG neurons at 1, 2, and 3 days post-transfection (dpt). In contrast, for RGCs, neurite outgrowth was not observed until 4 dpt, with some cells requiring 6 to 7 days in vitro to extend long neurites (Garcia et al., 2002; Garcia and Vecino, 2003; Vecino et al., 2015). Only a small fraction of RGCs in culture, approximately 3% of the surviving population, are capable of elaborating long neurites. This rare subset of RGCs may possess distinct regenerative mechanisms and could correspond to the same cells that exhibit long-distance axon growth in the optic nerve following injury, particularly under regenerative treatments (Wang et al., 2012). Furthermore, Dupraz et al. also identified a small population (2–3%) of RGCs capable of extending long neurites after just 3 days in vitro, even when IGF-1R (Insulin-like growth factor 1 Receptor) signalling was blocked, a condition that inhibits axonal outgrowth in the majority of adult RGCs (Dupraz et al., 2013). In addition, the heterogeneity of the CNS neurons has been confirmed by SnRNA-seq, electrophysiological and morphological studies in the retina (Goetz et al., 2022). Based on this, we hypothesize that the RGCs analysed in this study belong to a regenerative neuronal subtype characterized by relatively rapid neurite outgrowth.

DRG neurons, which retain their regenerative capacity into adulthood, serve as a valuable model for studying axonal regeneration. However, different DRG neuron subpopulations exhibit varying regenerative abilities and express distinct molecular markers (Neumann and Woolf, 1999; Richardson and Issa, 1984). Similarly, at least 46 subtypes of retinal ganglion cells (RGCs) have been identified, each showing differential resilience to injury, with several genes implicated in subtype-specific injury responses (Tran et al., 2019). This cellular heterogeneity poses challenges for consistent and interpretable experimental outcomes. While RGCs represent a promising model for studying CNS regeneration, their diversity may introduce bias into the data, limiting the generalizability of findings (Sanes and Masland, 2015). In our current protocol, only about 3% of RGCs in vitro—those capable of extending long neurites—are amenable to analysis, suggesting that observed similarities between DRG neurons and RGCs likely pertain only to this regenerative RGC subpopulation. Notably, even under various in vivo regenerative treatments, only a small fraction of RGC axons regenerate over long distances (Petrova et al., 2020), raising the possibility that this same subset of intrinsically regenerative cells is being studied here.

We emphasize that the culture conditions used in this study did not include inhibitory or neurotrophic factors that could influence regeneration, apart from the laminin/PLL coating of the culture surface. However, we cannot entirely rule out the influence of cell type-specific requirements on regenerative capacity. Notably, even under identical culture conditions for RGC cultures only a subset of RGCs is capable of regenerating long axons, displaying microtubule density and polymerization kinetics comparable to those observed in peripheral DRG neurons. Furthermore, microtubule polymerization within these long neurites is predominantly anterograde. This is consistent with the established role of anterograde axonal transport as a key intracellular mechanism supporting axon elongation (Petrova and Eva, 2018).

Of the two components of anterograde transport, the slow component is responsible for delivering cytoplasmic proteins, enzymes, and actin at a rate comparable to that of axonal regeneration (Wujek and Lasek, 1983). Although both cell types examined in this study exhibited similar microtubule polymerization speeds, DRG neurons consistently regenerated longer neurites than RGCs within the same time frame, despite having comparable microtubule density and polymerization rates. This suggests that additional factors—such as the transport of vesicles and motor proteins—may play a critical role in neurite elongation, independent of polymerization dynamics. Local protein synthesis may also significantly influence regenerative capacity. For instance, the axon guidance molecule MAG promotes neurite outgrowth in DRG neurons from young animals and in RGCs, and this effect correlates with a developmental decline in endogenous neuronal cAMP levels. Elevating cAMP in older neurons has been shown to shift their growth state, enabling axonal extension even in the presence of myelin-associated inhibitors (Domeniconi and Filbin, 2005; Hannila and Filbin, 2008). Additionally, certain small molecules that act independently of cAMP have also been shown to promote neuronal protection and regeneration (Ma et al., 2010).

PNS neurons are capable of stabilizing their MTs, which facilitates the formation of regeneration-promoting structures. In contrast, CNS neurons lack this intrinsic regenerative ability and instead form growth-incompetent structures characterized by a disorganized MT network (Kulkarni et al., 2022). Fluorescently tagged EB proteins therefore serve as powerful tools for monitoring MT dynamics and, by extension, evaluating axonal regeneration. This approach is both efficient and non-toxic, allowing the visualization of MT density and orientation in living cells (Akhmanova and Hoogenraad, 2005; Galjart, 2010). In addition, investigating EB3 expression levels and its interactions with other molecules, such as fidgetin, could lead to the identification of novel therapeutic targets for promoting axonal regeneration (Ma et al., 2023). In this study, the EB3-GFP fusion protein enabled visualization of microtubule plus-end dynamics as distinct “comet-like” structures moving against a faint cytoplasmic background. Notably, increased EB3-GFP comet activity may reflect active remodeling of neurite segments (Kleele et al., 2014). Previous studies using EB3-GFP in cultured cells have shown that microtubules grow more slowly in neurons than in glial cells (Stepanova et al., 2003), and that EB3 movement is comparable across various neuronal compartments, including the soma, dendrites, axons, and growth cones. Consistent with these findings, we observed slower comet movement within the RGC soma. Regarding comet directionality, earlier studies in embryonic E17 mouse hippocampal neurons and E18 Purkinje neurons reported that approximately 65% of EB3 comets in proximal dendritic regions moved distally, while 35% moved retrogradely toward the soma. In more distal dendritic compartments and axons, however, most EB3 comets were oriented toward the growth cone. Similarly, our results showed that the majority of EB3 comets in regenerating neurites moved in the anterograde direction, suggesting that, at 4 days post-transfection (4 dpt), many neurites exhibit axon-like behavior (Baas et al., 1989). Conversely, retrograde polymerization observed in smaller neurites may indicate growth-phase dendrites, which could account for impaired elongation in some neurons due to mixed microtubule orientation.

Comet dynamics have been shown to be altered in two familial myotrophic lateral sclerosis models, with increased comet densities reported (Kleele et al., 2000). In unpolarized neurons such as those from the DRG, microtubule stabilization in a single neurite typically precedes axon specification. In cultured hippocampal neurons, the transition of a minor neurite into the future axon is marked by several key changes at the growth cone, including an increase in growth cone size, expansion of the peripheral lamellipodial veil, shortening of actin filaments (ribs), enhanced actin dynamics, and an increase in both the number and length of newly assembled dynamic microtubules. These microtubules subsequently penetrate both the central and peripheral regions of the growth cone (Bradke and Dotti, 1997; Conde and Caceres, 2009). Significantly, we observed clear microtubule penetration into the axons of both DRG and RGC neurons. However, parallel experiments using antibodies against classical axonal and dendritic markers—such as MAP2, MAP1B, tau, ankyrin, and neurofascin—failed to identify well-differentiated axons in cultured RGCs (unpublished data), despite the fact that these markers are readily detected in RGC axons in vivo (Van Wart et al., 2007). This discrepancy raises the possibility that RGC neurites grown in vitro may be unable to fully organize their cytoskeleton, potentially due to the absence of appropriate target-derived cues. Nevertheless, our findings suggest that the mechanisms underlying microtubule polymerization dynamics show a degree of similarity between the RGC subset we analyzed and DRG neurons. However, given the incomplete cytoskeletal organization of RGC neurites in vitro and the likelihood that key regulatory pathways depend on cell-type-specific signals or in vivo cues, these parallels should be interpreted cautiously. Additional mechanisms are expected to influence the extent and nature of regenerative responses in these different neuronal populations.

It is well established that mitochondria in healthy cells are dynamic organelles, and their movement is closely regulated by the energy demands of different cellular compartments. Deficiencies in mitochondrial transport have been implicated in a range of neurological disorders (MacAskill and Kittler, 2010), although the mechanisms underlying cytoskeleton-mediated transport remain incompletely understood. Several studies have suggested that the extent and directionality of mitochondrial movement are linked to neuronal regeneration capacity (Misgeld et al., 2007). Indeed, in neurodegenerative diseases, mitochondrial dysfunction can result in an insufficient ATP supply for microtubule motor proteins, disrupting mitochondrial axonal transport; enhancing microtubule-mediated trafficking may help prevent mitochondrial-induced damage (Esteves et al., 2014; Jang et al., 2024). In the present study, we demonstrate that MT polymerization occurs independently of mitochondrial transport. This opens the possibility for future investigations into how alterations in microtubule dynamics might influence mitochondrial behavior. Importantly, we show that it is feasible to simultaneously visualize both microtubules and mitochondria in adult CNS and PNS neurons derived from the same animal, offering a powerful platform for such studies.

In conclusion, this study provides new insights into the cytoskeletal mechanisms involved in axonal regeneration of adult CNS and PNS neurons under uniform experimental conditions. Future investigations could focus on other RGC subpopulations, particularly after neurite elongation or following regenerative treatments, to better understand subtype-specific responses. The methodology introduced here offers a novel and robust platform for studying adult RGC regeneration in vitro. Comparing them with DRG neurons from the same animals lays the groundwork for elucidating the role of axonal transport in regeneration. Notably, the observed variability in the regenerative capacity among RGCs supports the notion that these neurons exhibit heterogeneous responses to injury. This is consistent with findings in glaucoma, where some RGCs degenerate while neighboring cells survive.