Fish are kept and bred in our fish facility according to standard methods (Westerfield, 1989). All experiments were performed in compliance with animal welfare legislation. Procedures were approved by the ethical approval committee of the Regierungspräsidium Dresden, Germany: TVV 53/2015 and TV 9/2016.

All efforts were made to minimize animal suffering and the number of animals used. We used Tg(olig2:dsRed2) (Kucenas et al., 2008), Tg(olig2:eGFP) (Shin et al., 2003) and Tg(mbp:eGFP) (Jung et al., 2010).

As described previously (Becker et al., 2004), fish were anesthetized by immersion in 0.033% aminobenzoic acid ethylmethylester (MS222; Sigma) in PBS for 5 min. A longitudinal incision was made at the side of the fish, and the spinal cord was completely transected under visual control 4 mm caudal to the brainstem–spinal-cord junction.

Adult zebrafish were terminally anesthetized, their spinal cords exposed and the spinal cord tissue was carefully removed. Up to five spinal cords were dissected at once and placed in 1 mL of Hanks′ Buffered Salt Solution (HBSS, Gibco). For the dissociation of spinal cords into a single cell suspension, the tissue was incubated for 3 min with 100 μL 0.25% Trypsin (Sigma; T4549)/1 mM EDTA at room temperature and triturated using a 200 μL pipette during the whole incubation time avoiding the formation of air bubbles. Then, tissue digestion was stopped by adding 100 μL of 20% fetal calf serum (Sigma; F0804) in 2 mM CaCl₂ and 300 μL HBSS. Afterward the cell suspension was allowed to sit for 5 min on ice, before it was applied to a 20 μm (OPCs, Miltenyi Biotec, # 130-101-812) or a 35 μm (OL, Corning, # 352235) cell strainer, respectively. After washing with 9 mL of HBSS, the cells were pelleted by centrifugation for 5 min at RT at 300 g. The supernatant was removed, cells were resuspended in 1 mL HBSS including 1 μL Vybrant^TM DyeCycle^TM Violet Stain (Molecular Probes, Invitrogen, V35003) and incubated for 30 min at 28°C to stain cellular nuclei.

Cell suspensions were sorted directly into 96 well plates, filled with culture medium only, or pre-seeded with motor neurons (MNs) using a BD FACSAria sorter. For the detection of GFP a 488 nm excitation laser and a 530/30 bandpass filter were used. DsRed was detected with a 582/15 bandpass filter after excitation with a 561 nm laser. Vybrant^TM DyeCycle^TM Violet Stain was detected after 405 nm excitation and a 450/40 bandpass filter. Cellular events were defined by forward and side scatter profile and by the dsRed or GFP signal of the corresponding transgenic line. From this selection, all events that showed incorporation of the Vybrant^TM DyeCycle^TM Violet Stain were gated because this proved to be an ideal selection to separate the small zebrafish CNS cells from remaining cellular and myelin debris. To assess purity and viability of the sorted cells post-sort analysis of the purified cells after addition of Propidium Iodide at a final concentration of 1 μg/mL was performed.

Two days before dissection, the wells of a 96-well plate (Greiner Bio One, # 655090) were coated with 10 μg/mL Poly-D-Lysine solution (Merck Millipore, # A-003-E) in DEPC H₂0 and left under a biosafety cabinet overnight (50 μL/well).

The next day, Poly-D-Lysine was removed and wells were washed three times with DEPC H₂0 to remove unbound Poly-D-Lysine. After air-drying, 10 μg/mL of laminin solution (Sigma Aldrich, # L2020) in sterile PBS was added to the wells (50 μL/well) and the plate was left at 37°C overnight. On the day of the dissection, laminin solution was removed and the wells were washed twice with sterile PBS and once with DEPC H₂0 to remove unbound coating solution. After air-drying, 280 μL of Leibovitz’s L-15 Medium (Gibco, # 11415-049) supplemented with 15% FBS (Gibco, Performance Plus United States, # 1000-0-044), 1% Glutamax (Gibco, # 350-050-038) and 1% penicillin/streptomycin (Gibco, # 15-140-122) was added to each well. The plate was left at 28°C until cell seeding. Fluorescently sorted OPCs were plated at a density of 10000 cells per well and left at 28°C for a maximum of 10 days. Culture medium was exchanged every 2 days, starting at 3 days in culture.

The complete human iPSC culture and differentiation protocol for MN generation is described in detail elsewhere (Reinhardt et al., 2013). Briefly, a population of expandable neural progenitors (NPCs) is obtained from iPSCs via dual SMAD inhibition. Using a defined cocktail of small molecules, NPCs can subsequently be differentiated into MNs. For MN plating, cells were detached in their early maturation stage, i.e., day 3, using Accutase (Sigma), and distributed on Matrigel (Corning)-coated plates in MN maturation medium, consisting of N2B27 medium supplemented with 2 ng/mL BDNF (Peprotech), 2 ng/mL GDNF (Peprotech), 1 ng/mL TGF-β3 (Peprotech), 200 μM ascorbic acid, and 100 μM dbcAMP. Y-27632 (Sigma) was included in the first day to increase the chances of MN survival upon re-plating. N2B27 medium consisted of DMEM-F12/Neurobasal medium 1:1, 0.5% N2 supplement, 1% B27 supplement lacking vitamin A, and 1% PS/G (all Thermo Fisher Scientific). For plate coating, the Matrigel mixture was diluted to 1:100 in ice-cold Knockout DMEM (Invitrogen) prior to overnight coating at RT. For long-term storage, coated plates were wrapped with Parafilm and kept in the fridge for up to 1 month. Adult zebrafish OPCs were FACsorted on top of the MN culture at day 4–7 of differentiation and cultured with MN maturation medium at 28°C under 5% CO₂. Medium was replaced every second day.

Immunohistochemistry on paraformaldehyde-fixed spinal cord sections (50 μm thickness) has been described previously (Reimer et al., 2008). Cells were fixed in 4% paraformaldehyde (Agar Scientific, 16% EM Grade, Methanol free, R1026) diluted in PBS for 10 min at room temperature, washed gently with 0.1% Triton X-100 (Sigma Aldrich, # T8787) in PBS and incubated with primary antibodies in 0.1% Triton X-100 in PBS for 1 h at room temperature.

After washing with 0.1% Triton X-100 in PBS, cells were stained with secondary antibodies in the same solution for 2 h at room temperature. Cells were finally stained with 0.3 μM DAPI (Roth, # 6335.1) in PBS for 10 min at room temperature, washed in PBS and processed accordingly.

Primary antibodies used were: mouse monoclonal against Glial Fibrillary Acidic Protein (GFAP, 1:250, Clone GA5, Merck Millipore, MAB3402), rabbit polyclonal against SRY-box containing gene 10 (Sox10, 1:250 (cultures) and 1:500 (sections), Biozol GTX128374-200) and a custom-made rabbit polyclonal against Myelin Basic Protein [MBP, 1:75, (Pinzon-Olejua et al., 2017) a kind gift from Michael Brand, CRTD, TU Dresden, Germany].

Secondary antibodies used were goat anti-Mouse IgG Alexa Fluor^®555, (Molecular Probes, Invitrogen, A21424) and donkey anti-Rabbit IgG Alexa Fluor^®647, (Jackson ImmunoResearch Europe Ltd., 711-605-152), both used at a 1:750 dilution.

Confocal images of cell cultures were acquired with a Zeiss LSM 780/FCS inverse microscope using an LCI PlnN 63 × /1.3 NA W DICIII objective in a tile mode (10 × 10 tiles, 10% overlaps). Tissue sections were imaged with a Leica TCS-SP5 confocal microscope using HC PL APO CS 20 × /0.7 NA, HCX PL APO 40 × /1.25 NA and HCX PL APO 63 × /1.2 NA objectives. To minimize channel crosstalk, single channel images were acquired sequentially. Stacks of overview images were maximum projected using Fiji. Co-localization of markers was verified in single confocal stacks using the 63× objectives. Adobe Photoshop CS6 was used to adjust brightness and contrast. Illustrator CS6 was used to compile the figures.

Cells were manually counted with the Cell Counter Plugin of Fiji. For quantification of total cell numbers, a representative area corresponding to ∼5% of the total well area were quantified and for each time-point, 3–4 wells were counted. Microsoft Excel was used to process the measured data. To analyze significance, P-values were determined with GraphPad Prism using One Way ANOVA with Dunns Post hoc test. Error bars represent SEM. P > 0.05 was not considered significant.

For the RNAseq experiment, the SMARTer Ultra Low Input RNA for Illumina Sequencing Kit - HV from Clontech was used to reverse transcribe the RNA and amplify full-length cDNA according to the user manual. Therefore, 2000 cells from five biological replicates of both lines Tg(olig2:eGFP) and Tg(mbp:eGFP) were directly FACsorted in 5 μL reaction buffer containing RNAse Inhibitor. After amplification of cDNA using 12 cycles, the cDNA was sheared to 200 bp fragment length with 8-micro TUBE strip in the Ultrasonicator Covaris LE220, followed by a library prep using NEBNext Ultra DNA Library Prep Kit for Illumina (NEB). The libraries were sequenced with 75 bp single end on a Hiseq2500 with a minimal sequencing depth of 27 million reads and mapped to GRCz10 with GSNAP (v 2016-09-23). Transcripts were counted with feature Counts v 1.5.2 based on Ensembl version 81. The gene expression values were normalized by the R-package DESeq2 in R (v 3.3.0).

The next generation sequencing datasets generated in this study have been submitted to GEO and are publically available under accession number GSE100821.

A complete spinal cord transection in adult zebrafish (Becker et al., 2004; Reimer et al., 2008) causes full paralysis caudal to the site of injury. In order to analyze the fate of mature oligodendrocytes (OLs) at the lesion site after SCI, we transected the spinal cord of transgenic Tg(mbp:GFP) zebrafish. In this reporter line OLs are labeled by GFP, enabling the identification of myelin sheaths as tube-like structures highlighted by GFP (Figure 2). We find that in response to SCI, myelin sheaths and OL cell bodies are lost at 7 days post lesion (dpl) compared to sham controls (Figure 2). At 14 dpl, a remyelination response has been initiated and newly formed myelin sheaths, as well as an increase in the number of OL cell bodies, are observed (Figure 2).

During development of the vertebrate CNS mature OLs form from OPCs that derive from pMN progenitor cells of the neural tube (Park et al., 2002; Shin et al., 2003; Ravanelli and Appel, 2015). In the adult brain and spinal cord OPCs are maintained and function as a constitutive progenitor pool for the formation of new OLs (Park et al., 2007). In zebrafish, OPCs are marked by high expression of olig2:GFP, a pan-vertebrate oligodendroglial lineage marker (Czopka and Lyons, 2011), and are evenly distributed throughout the white and gray matter of the adult spinal cord parenchyma (Figure 3). OPCs co-express the transcription factor Sox10, a neural crest, as well as oligodendroglia lineage marker in the CNS of vertebrates (Figure 3 and Supplementary Figure S1). A subpopulation of ventral radial glia also expresses olig2:GFP. Radial glia are identified by ventricular cell bodies (arrowheads) touching the central canal (asterisk) and radial processes spanning the entire ventral spinal cord. Olig2:GFP-positive OPCs can be easily distinguished from mature mbp:GFP-expressing OLs by their specific morphology displaying short arbors and the lack of GFP-positive myelin sheaths (Figures 2, 3).

Zebrafish embryos are a standard model organism in developmental biology. Because of their small size and transparency, they offer ideal conditions for functional analysis: e.g., straightforward application of small molecules just by adding the compounds to the water. Additionally, live imaging can easily be done up to several days after fertilization. In contrast to embryos, functional analysis of cells in the adult zebrafish spinal cord is much more challenging. Therefore, we established a platform to analyze zebrafish OPCs in vitro. We developed a streamlined and fast, though inexpensive, protocol that allows direct access to a pure and vital population of zebrafish OPCs in less than 2 h. This simple protocol is based on Trypsin digestion of dissected fish spinal cord, followed by addition of the Vybrant DyeCycle Violet Stain as a nuclear marker of living cells for fluorescent activated cell sorting (FACS) of OPCs. Cellular events were defined by forward and side scatter profile and by fluorescent marker (either dsRed or GFP) signal of the corresponding transgenic line. From this selection, all events that showed incorporation of the Vybrant DyeCycle Violet Stain were gated because this proved to be an ideal selection to separate the small zebrafish CNS cells from remaining cellular, i.e., mostly myelin debris (small cells cannot be separated from all cellular debris when only using forward and side scatter, Figures 4A,B). To assess purity and viability of the sorted cells we performed a post-sort analysis of the purified cells after addition of Propidium Iodide to stain non-viable cells. Less than 5% of cellular events corresponded to dead cells showing uptake of Propidium Iodide (Figure 4C). More than 95% of the sorted cells showed the fluorescent lineage marker, demonstrating high purity of sorted cell populations (Figure 4C). Typically, approximately 30000 cells could be sorted using five adult spinal cords as starting material.

To corroborate the identity and differentiation status of our FACsorted populations, we carried out gene expression analysis using RNAseq of sorted olig2:GFP- (OPCs) and mbp:GFP-positive cells (OLs, Supplementary Figure S2) and transcriptomes were analyzed for differential gene expression of well-known markers of the oligodendroglial lineage (Figure 5). A high degree of cell type specificity was revealed: OPC and OL populations are distinctly separated, as shown by expression profiling for the mature OL/myelin marker genes cldnk (Claudink), mbpa (Myelin Basic Protein) and mpz (Myelin Protein Zero). Interestingly, the expression levels of the oligodendroglial lineage markers olig1 and olig2 are much higher in the OPC population than in the mature OL population.

On the one hand, the protocol described above can be used to establish acute cultures of OPCs for immediate cellular analyses ex vivo, like, e.g., electrophysiological measurements of freshly isolated OPCs and OLs. On the other hand, we also wanted to analyze OPC fate over longer culture periods in vitro. Specifically, we were interested in the robustness of the cultures and in their differentiation capacity. Therefore, we first quantified the number of cells over different time points. At 3, 7, and 10 days in culture (DIC) cell numbers do not change significantly, indicating stability of the culture and suggesting that OPCs neither undergo significant cell death, nor are activated toward overt proliferation under these conditions (Figure 6A). However, at 14 DIC we observed a significant detachment of cells from the bottom of the culture plates, as well as an increase in the number of dying cells (weak DAPI staining, pyknotic nuclei and weak cytoplasmic signal, data not shown).

The morphology of cells at 7 DIC is round with short processes, and significant differentiation of OPCs toward OLs is not apparent (Figure 6B). To confirm cell identity and differentiation status, we performed staining of cultures using antibodies against GFAP (radial glia marker), Sox10 (oligodendroglial lineage marker) and MBP (marker of mature OL) at different time points in vitro. At 3 DIC we analyzed 1052 olig2:GFP-positive cells: all cells co-expressed Sox10, and none of them expressed GFAP, identifying them as OPCs. At 7 DIC we analyzed 749 cells, of which 745 co-expressed Sox10 (Figure 6B). At 10 DIC, we analyzed 322 cells for MBP expression. We failed to detect any cell expressing MBP, showing that under these conditions OPCs do not spontaneously differentiate into mature OLs.

The overall goal of this project was to establish an in vitro myelination model based on adult zebrafish OPCs. Since we did not detect formation of myelin in our monotypic cultures we reasoned that a co-culture model including motor neurons (MNs) could drive OPCs into OL differentiation and hence promote myelination. First, we determined if the ratio of OPCs vs. MNs is influencing OPC differentiation. We cultured 5000 OPCs together with 0, 10000, 20000, 40000, 80000, and 160000 MN per well (Supplementary Figure S3). We find a correlation of OPC differentiation state and MN numbers based on OPC morphology at 1 DIC. The number of OPCs that start differentiating, i.e., cells with long and branched processes, is increased in cultures with higher MN numbers (Figure 7A and Supplementary Figure S3). On the contrary, almost all cells in cultures without MNs have a round morphology (Supplementary Figure S3). The degree of branching and the length of OPC processes is strongly increased with culture time, showing that OPCs continue to differentiate toward mature OL (Figure 7A and Supplementary Figure S4). At 16 DIC we analyzed the expression of MBP protein in the cultures by antibody staining. Indeed, we find many cells that are positive for MBP (Figure 7B), showing that OPCs efficiently differentiated into highly branched OLs.

Studies using isolated primary OPCs have been carried out for almost 4 decades (McCarthy and de Vellis, 1980). However, most isolation protocols start with mixed cultures from embryonic or neonatal tissue from rat. The mixed cultures can be enriched for OPCs by different methods, the most popular and simple being the shake-off technique that is based on differential adhesive properties of glial populations, reviewed in (Barateiro and Fernandes, 2014). The enrichment and survival of rat OPCs is largely dependent on the presence of growth factors, like basic fibroblast growth factor (FGF-2) and platelet derived growth factor (PDGF), the cell seeding density and also the composition of the basal medium used (Kleinsimlinghaus et al., 2013). The establishment of mouse OPC cultures is much more difficult than from rat, mostly because of the lack of surface markers and the tendency of mouse OPCs to spontaneously differentiate into other lineages (Chen et al., 2007; Zhu et al., 2014). Recently, a new method to enrich OPCs in mixed cultures from mouse cortices has been developed. Using a cocktail of the growth factors Platelet Derived Growth Factor-AA (PDGFaa), basic fibroblast growth factor (bFGF) and epidermal growth factor (EGF) OPCs could be propagated (Yang et al., 2016). Interestingly, in mouse the efficiency of the OPC isolation is strongly correlated with the age of the donor tissue: the number of adult OPCs reaches 12% of the cell number that could be isolated from neonates (Pillai and Ross, 1986). Human OPCs have been isolated from different sources, like cerebral cortex (Pillai and Ross, 1986), generated from fetal neurospheres (Lu et al., 2015) and very recently also derived from iPS cells (Ehrlich et al., 2017; Rodrigues et al., 2017). Generally, human OPCs seem to be a difficult model to work with because of their slow growth and differentiation behavior.

Elucidating the underlying mechanisms of successful remyelination after SCI in the regenerating model organism zebrafish can be highly beneficial to develop pro-remyelinating strategies for improving endogenous remyelination in patients with SCI. Adult zebrafish recover functionally within 6 weeks after a full spinal cord transection. Lost neurons are replaced, axons regrow and remyelination of the demyelinated spinal cord and the newly generated axons bridging the lesion site occur (Becker et al., 1997; Reimer et al., 2008). However, the underlying mechanisms of the superior remyelination capacity in zebrafish (Munzel et al., 2014) are not understood. Investigating the OPC intrinsic mechanisms of dormancy, activation and remyelination in an in vitro system is a promising approach. The advantage of adult spinal cord FACsorted OPCs lies in the region identity and age-status of these progenitor cells, information that is not available from iPSC-derived OPCs or the other models described above. Additionally, sorted zebrafish OPCs have entered dormancy in vivo, which also allows studying their activation. In this study we could show that OPCs sorted from spinal cord tissue can be successfully taken into culture and that these monocultures are viable for at least 10 days. Cells cultured under these conditions resemble a dormant adult spinal OPC population of an uninjured spinal cord. In order to study activation and differentiation, we propose a novel co-culture system with human iPSC-derived motor neurons (Reinhardt et al., 2013). Although rat/mouse myelinating xenocultures have been described before (O’Meara et al., 2011), the use of fish/human co-cultures to our knowledge has not been reported. The use of zebrafish OPCs and human motor neurons enables the identification of evolutionary conserved signaling pathways. This increases the chances of utilizing the obtained results in a translational framework.