An electroceutical approach enhances myelination via upregulation of lipid biosynthesis in the dorsal root ganglion

DRGs were extracted from thoracic region of 13 d old mouse embryos. Each DRG was positioned gently at the center of a coverslip coated with a thin layer of Matrigel (Corning; product number: 356231) and placed in 24 well plates (Corning; product number: 3526). DRGs were cultured in growth medium for 8 d, followed by myelination medium for the rest of the culture period. The compositions of growth and myelination media are given in table S1 (available online at stacks.iop.org/BF/14/015017/mmedia).

The 3D modeling for the device was carried out using SolidWorks®, and it was fabricated via computerized numerical control machine using polyetheretherketone (PEEK). PEEK enables autoclaving for sterilization and offers excellent dimensional stability and oxidation resistance. These properties make it ideal to be used in ES devices for biological applications. The device had an inner diameter of 100 mm and depth of 14 mm. It consisted of six wells, each with a diameter of 15 mm and surrounded by three miniature pillars to securely hold the coverslip in position. The electrodes were made from stainless steel, and then electroplated with a 2 µm thick nickel undercoat (for adhesion with gold layer) followed by a 0.5 µm thick gold layer. Each electrode had a length of 70 mm, height of 6 mm and width of 1 mm. During assembly, the electrodes were positioned 30 mm apart and parallel to each other. After assembly, the electrodes were held in place using slits in the upper and lower chambers (figure 2(A)). COMSOL Multiphysics® was used (with electro-conductive properties of the medium based on Dulbecco's Modified Eagle Medium) to simulate the uniformity of the electrical field generated across each well (figure 2(B)).

Zoom In Zoom Out Reset image size

ES was provided to the embryo DRGs at 10 days in vitro (DIV). The coverslips containing the DRGs were positioned within the wells of the ES device and then 15 ml myelination medium was added to the device. A biphasic square wave was generated at 50 mV mm⁻¹ and the stimulation was provided for 1 h. ES was provided at 1, 20 and 100 Hz during the screening and only 20 Hz during the rest of the steps. After the ES application, the DRGs were placed back into 24 well plates.

Cells were fixed with 4% paraformaldehyde treatment for 15 min at room temperature (RT). The fixed cells were then blocked by treatment with blocking buffer (1 × phosphate-buffered saline (PBS), 0.3% Triton X-100 and 5% Normal Donkey Serum) for 1 h. After blocking, primary antibodies were added, diluted in antibody dilution buffer (ADB; 1 × PBS, 0.3% Triton X-100 and 1% Normal Donkey Serum). This was left at 4 °C in the refrigerator overnight. Then, secondary antibodies were added, diluted in ADB, and kept at RT in a dark environment for 1 h.

Primary antibodies: rabbit monoclonal anti-myelin basic protein (MBP) (Abcam; catalogue number: ab40390; used at 1:200 dilution) and mouse monoclonal anti-Tuj-1 (Abcam; catalogue number: ab78078; used at a 1:1000 dilution). Secondary antibodies: donkey anti-rabbit Alexa fluor® 488 (Jackson ImmunoResearch; catalogue number: 711-545-152; used at 1:200 dilution), donkey anti-mouse Alexa fluor® 594 (Jackson ImmunoResearch; catalogue number: 715-585-151; used at 1:200 dilution). For the average myelin length, the lengths of individual myelin segments were measured from 4X images using the NeuronJ plugin (for facilitating tracing and analysis of neuronal images) of ImageJ. Measurement for percentage of myelinated neurons was obtained via co-localization of MBP with Tuj-1 from 4X images using Colocalization Finder plugin of ImageJ. For these measurements, there were three biologically distinct replicates from each condition and three technical replicates from each biological replicate.

RNA was extracted from the whole DRGs using TRIzol Reagent (Thermo Fisher; catalogue number: 15596018), according to the manufacturer's instructions. For each condition, RNA was harvested from three DRG explants pooled together. RNA quality was assessed by Agilent 2100 Bioanalyzer using the RNA 6000 Nano Kit (Agilent Technologies, CA, USA). RNA quantification was performed using ND-2000 Spectrophotometer (Thermo Fisher, MA, USA). The construction of RNA library was performed using QuantSeq 3' mRNA-Seq Library Prep Kit (Lexogen, Austria), according to the manufacturer's instructions. High-throughput sequencing was performed as single-end 75 sequencing using NextSeq 500 (Illumina, CA, USA). Alignment was performed using Bowtie2 [24]. Bowtie2 indices were either generated from genome assembly sequence or the representative transcript sequences for aligning to the genome and transcriptome, using the mm10 (Mus musculus) genome assembly. The alignment file was used for assembling transcripts, estimating their abundances and detecting differential expression of genes. Differentially expressed genes were determined based on counts from unique and multiple alignments using coverage in BEDTools [25]. The RC (read count) data were processed based on quantile normalization method using edgeR package from Bioconductor within R [26]. Gene classification was based on searches done by Database for Annotation, Visualization, and Integrated Discovery (DAVID) (https://david.ncifcrf.gov/) and Medline databases (www.ncbi.nlm.nih.gov/). Differential analysis and graphic visualization were performed using ExDEGA (Ebiogen, Republic of Korea). Protein association networks were created using the STRING plugin of Cytoscape (version 3.7.2). The RNA sequencing data in this study have been deposited to the Gene Expression Omnibus database of the National Center for Biotechnology Information under the accession number GSE190745.

For messenger RNA (mRNA) extraction, Rneasy Mini Kit (Qiagen; catalogue number: 74106) was used. All separation and purification steps carried out were done according to the protocol provided with the kit. This was followed by complementary deoxyribonucleic acid (cDNA) synthesis, using AccuPower® RT Premix (Bioneer; catalogue number: K-2041-B). The temperature cycling was done using Biometra Tone (Analytik Jena), with the recommended temperature and duration for the AccuPower® RT Premix. Real-time PCR was carried out on three biologically distinct samples from each group—control and ES DRGs—at each time point. Three technical replicates were prepared for each sample and average C_T value was taken for further comparison. Applied Biosystems StepOnePlus^TM was used for thermal cycling and C_T measurements, and the software StepOne v2.3 available from Applied Biosystems was used to directly calculate the fold change in the ES DRGs compared to the control. The mixtures and temperature cycles used were according to the guideline provided with the instrument. The list of sequences for primers used is provided in table S2.

Citrate assay was performed using Citrate Assay Kit (Abcam; catalogue number: ab83396) and carnitine assay was performed using L-Carnitine Assay Kit (Abcam; catalogue number: ab83392). All further steps were carried out according to the manual provided with the respective kit. The assays were carried out on three biological replicates for each group and twice for each biological replicate, as recommended in the manual. Data shown in the graphs of the respective assays show measurements from all repetitions. The absorbance was measured by VersaMax Microplate Reader (Molecular Devices) at a wavelength of 570 nm.

Cells were fixed by treating with formalin solution (10%; Sigma Aldrich) for 10 min followed by glutaraldehyde solution (20%; Electron Microscopy Sciences) for 15 min, both at RT. The DRGs were post-fixed in osmium tetroxide solution (0.4%; Sigma Aldrich) for 15 min at RT and then air-dried. Prior to ToF-SIMS imaging, the cells were treated by air-plasma (CUTE, Femto Science Inc., Republic of Korea) at 1.1–1.3 Torr, 50 kHz, 100 W, and 70 sccm of air for 5 min. Analysis was conducted on a ToF-SIMS 5-100 instrument (ION-TOF, Münster, Germany) using a pulsed 30 keV Bi₃ ⁺ primary ion beam in delayed extraction mode. ToF-SIMS images were obtained from fixed DRG on a cover glass over an area of 300 × 300 μm² with 256 × 256 pixels. Internal mass calibration for ToF-SIMS spectra was performed using the peaks of CH₃ ⁺, Na⁺, C₂H₃ ⁺, C₃H₅ ⁺and C₄H₇ ⁺ for positive ion mode, and C⁻, C₂ ⁻, C₃ ⁻ and C₄ ⁻ for negative ion mode before further analysis. Low-energy electrons were supplied onto the surface of the sample using an electron flood gun for charge compensation during analysis.

Values are expressed as mean ± s.d. Statistical significance was calculated with GraphPad Prism software (version 8.0). One-way analysis of variance (ANOVA) followed by Tukey's comparison test was used to compare the groups (*P < 0.05, **P < 0.005 and ***P < 0.001).

The DRGs were extracted from mouse embryos at E13, seeded onto Matrigel-coated coverslips, and supplemented with growth medium. After sufficient growth and differentiation of neurons and SCs in the DRG, myelination was induced using the myelination medium (DIV 8). After allowing the DRG to sufficiently acclimatize to the new environment in the myelination medium, ES was provided during the initial stage of myelination (DIV 10). Since ES at frequencies higher than 100 Hz lead to the depression of action potential (AP) of DRG neurons and E-field strengths higher than 50 mV mm⁻¹ cause abnormal SC morphology [27], screening was conducted with biphasic ES of 1–100 Hz for 1 h at a field strength of 50 mV mm⁻¹. The efficacy of the ES during the screening was evaluated by the changes in Krox20 expression one day after ES application. Krox20 was chosen because it is a well-known marker for myelinating SCs [28]. It is a master regulator of PNS myelination [29], being vital in both the development as well as maintenance of myelin [30, 31]. In the 1–100 Hz frequency range, 20 Hz ES was the most effective at upregulating Krox20 expression, with almost 2.5-fold upregulation compared to control DRG. On the other hand, 100 Hz ES downregulated Krox20 expression 0.5-fold when compared to control (figure S1(A)). Therefore, biphasic ES was applied at 20 Hz and 50 mV mm⁻¹ for 1 h, before carrying out further evaluations with the DRG (figure S1(B)).

A proper time window for evaluating the effects of ES on myelination was determined by monitoring the DRG explant from bright-field images at regular intervals (figure S1(C)). The myelin morphology at the suitable times was further confirmed via confocal imaging. Myelin morphology at 19 DIV was elongated and bipolar (figure S1(D), white arrows), which is characteristic of immature myelin [32]. On the other hand, myelin morphology at 30 DIV was uniform thickness (figure S1(E)), suggesting sufficient progression of myelination [33]. Therefore, the myelination was evaluated via confocal imaging at 19 DIV (figures 3(A) and S2(A)) and 30 DIV (figures 3(B) and S2(B)), which represent the intermediate and mature myelination stages respectively. Upon application of ES, the percentage of neurons containing myelin increased by 129% during the intermediate (figure 3(C)) and 72% during the mature myelination stage (figure 3(E)), compared to control DRG. The average length of myelin segments also increased significantly upon ES—61% during the intermediate (figure 3(D)) and 17% during the mature myelination stage (figure 3(F)). Longer myelin segments facilitate neuronal transmission by increasing nerve conduction velocity, as shown in rodents as well as computational models of axons [34, 35]. Although the enhancement of myelination upon ES was comparatively less dramatic at the mature myelination stage, it was still significant, suggesting that ES at the proper time window can have persistently beneficial effects on myelination.

Zoom In Zoom Out Reset image size

In several demyelinating peripheral neuropathies, the distal region of the axon is relatively more demyelinated than the proximal, making the symptoms such as muscle atrophy and numbness exacerbated in the distal region of the limb [4, 36, 37]. As such, an optimal treatment for these neuropathies should enhance myelination comparatively more in the distal region than in the proximal. Therefore, after verifying that the proposed electroceutical approach enhanced myelination in the whole DRG, its effectiveness for enhancing myelination in the distal region of the DRG was also evaluated. The number of myelin segments was counted in each 200 µm region with increasing radial distance from the center of the DRG. Interestingly, while ES enhanced myelination in both proximal and distal regions, the extent of enhancement was comparatively greater in the distal region (figures S3(A) and (B)). If the extent of myelination enhancement is sufficiently dominant in the distal region, it can alter the distribution of myelin in the DRG. To investigate this, we calculated the percentage of myelin segments in each 200 µm region, based on the total number of myelin segments in the whole DRG. Our results showed that, after ES, the percentage of myelin segments decreased in the proximal region and increased in the distal region (figures 3(G) and (H)), suggesting an alteration in myelin distribution. For example, during the intermediate myelination stage, the percentage of myelin segments decreased by 21% in a proximal region (200 µm) (figure 3(I)) and increased by 24% in a distal region (1000 µm) (figure 3(J)). Likewise, during the mature myelination stage, the percentage of myelin segments decreased by 12% in a proximal region (200 µm) (figure 3(K)) and increased by 14% in a distal region (1000 µm) (figure 3(L)). These results showed that ES effectively enhanced myelination in the distal region of the DRG, which further strengthens its prospect to be an effective intervention for demyelinating peripheral neuropathies.

Next, the biochemical changes enhancing DRG myelination as a result of ES were evaluated. RNA sequencing was performed on the DRG explants at two different myelination stages—the initial stage at 11 DIV (1 d after ES) and the intermediate stage at 19 DIV (9 d after ES) (figure 4(A)). The gene expressions at these two stages are decisive in the myelination schedule in the DRG [38]. After ES, there is a consistent upregulation of most of the genes involved in PNS myelination pathway, which conforms to the previous observation from confocal imaging. Interestingly, a similar degree of upregulation as PNS myelination genes was observed in the genes involved in lipid biosynthesis. In both cases, the extent of upregulation is more profound during the initial stage, but is still noticeably persistent during the intermediate stage of myelination. Given that lipid biosynthesis and fatty acid oxidation are connected in reciprocal fashion [39], the genes involved in fatty acid oxidation were also analyzed. At both stages, the RNA sequencing showed a consistent downregulation of the genes involved in the fatty acid oxidation process, where the extent of downregulation was more prominent during the initial stage of the myelination (figure 4(B)). To further support RNA sequencing results, protein association networks of the genes involved in lipid biosynthesis and fatty acid oxidation processes were constructed using STRING. Genes more strongly associated to each other followed consistent regulation patterns—upregulation in the case of lipid biosynthesis and downregulation in fatty acid oxidation—further strengthening the outcome of RNA sequencing (figure 4(C)). Therefore, this suggests a novel finding that the ES increases lipid biosynthesis in the DRG, leading to enhancement in myelination.

Zoom In Zoom Out Reset image size

This increase in PNS lipid biosynthesis upon ES might be due to its close association with myelination. The myelin membrane is exceptionally rich in lipids [40]. As such, SCs undergo a demanding lipid biosynthesis regime during the assembly of myelin. The impairment of this de novo lipid synthesis regime might cause hypomyelination [41]. Furthermore, lipid supplementation is effective at rescuing the demyelinating phenotype in animal models of CMT [7, 8, 42]. The lipid biosynthesis in SCs is intertwined with mitochondrial metabolism. During the citric acid cycle, the mitochondria produce citrate, which can then have two possible fates: (a) it can continue along the citric acid cycle, where it will be used for the purpose of energetics, or (b) it can move into the cytosol and then be converted to acetyl-CoA, a molecule necessary for de novo fatty acid synthesis by SCs [43]. These fatty acids serve as building blocks for most of the structural lipids of myelin [40, 44]. Therefore, defects in SC mitochondria can cause significant loss of myelin [45, 46]. Given how indispensable lipid biosynthesis is during myelination and how closely it is associated with mitochondrial citrate (figure 5(A)), further analyses via qPCR and biochemical assays were performed to investigate the shift towards increased lipid biosynthesis.

Zoom In Zoom Out Reset image size

Since FAS, HMGCR and ChoKin are critically involved in the synthesis of three key components, fatty acids, cholesterol and phospholipids, respectively, expressions of the three representative genes were compared. In accordance with the RNA sequencing data, the qPCR also revealed a significant upregulation of the lipid biosynthesis genes after ES. At 11 DIV, there was a 1.9-fold increase in FAS expression, a 2.2-fold increase in HMGCR expression and a 1.4-fold increase in the ChoKin expression, when compared to the control DRG (figure 5(B)). To verify whether mitochondrial citrate is being utilized to achieve higher lipid biosynthesis, the gene expression of citrate synthase in the DRG explant was measured. ES significantly increased the citrate synthase expression compared to control in the DRG explant, at both stages (figure 5(C)). This was further validated via citrate assay in the DRG explant. Compared to control, the available citrate concentration after ES was 20% higher at 11 DIV and 30% higher at 19 DIV (figure 5(D)). In order to verify that the additional citrate from the mitochondria is being utilized for lipid biosynthesis and not for energetics, we analyzed the expression of two representative genes responsible for mitochondrial energetics: cytochrome c and PGC1α. Our results showed that, after ES, there were no significant changes in the expressions of cytochrome c and PGC1α at either stage of myelination (figures 5(E) and (F)). These results suggest that, as a consequence of ES application, mitochondrial citrate is utilized to support higher lipid biosynthesis, further used to construct myelin.

For further verification, the changes in fatty acid oxidation were also analyzed. Lipid biosynthesis and fatty acid oxidation are regulated via AMP-activated protein kinase (AMPK) signaling, and an elevated level of AMPK reduces lipid biosynthesis, which hinders the myelination process [39]. Interestingly, AMPK expression decreased significantly at both time points after ES (figure S4(A)). Additionally, 3-hydroxyacyl-CoA dehydrogenase (HADH)—which is a representative enzyme in fatty acid oxidation—also shows a significant downregulation at both 11 DIV and 19 DIV upon ES, compared to control DRG (figure S4(B)). Carnitine is required for the transport of fatty acids into the mitochondrial matrix during fatty acid oxidation [47]. Since carnitine assay is well-regarded as a measure of fatty acid oxidation, it was conducted at both 11 DIV and 19 DIV to further support the outcome revealed by qPCR. After ES, carnitine concentration is reduced by 39% at 11 DIV and 44% at 19 DIV compared to control DRG (figure S4(C)). The downregulated expressions of AMPK and HADH and the reduced carnitine concentration confirm that there is a net shift towards higher lipid biosynthesis in the DRG after ES, resulting from the upregulation in lipid biosynthesis and downregulation in fatty acid oxidation.

In order to validate whether the upregulation in lipid biosynthesis makes higher abundance of structural lipids in the myelin membrane, ToF-SIMS imaging was performed. ToF-SIMS is an analytical technique to study surface distribution of target chemical species. Due to shallow sampling depth and recently developed implementations in in vitro sample preparation techniques for ToF-SIMS, it is suitable for imaging surface lipids with minimal interference from intracellular lipids [48–51]. To the best of our knowledge, this is the first reported analysis of myelin lipids in DRG explants via ToF-SIMS imaging. Due to the relatively higher density of cholesterol and sphingomyelin in the myelin membrane [40], these two structural lipids were of primary importance during positive-mode ToF-SIMS. Compared to the control DRG, ES resulted in noticeably higher abundance of cholesterol and sphingomyelin in the DRG myelin. Particularly, ToF-SIMS images of the cholesterol fragment showed clear myelin-like segments along radial patterns in the DRG explant. Such fragments were more prominent in the ES DRG (figure 6(A), green arrows) compared to the control (figure 6(A), sky-blue arrows). We then analyzed the size distribution of the cholesterol and sphingomyelin-enriched fragments. ES led to a significantly lower frequency of fragments in the smallest size group and a significantly higher frequency of fragments in the medium size groups in case of both cholesterol (figure 6(B)) and sphingomyelin (figure 6(C)). Meanwhile, in both cases, the frequency of fragments in the largest size group did not show any significant change.

Zoom In Zoom Out Reset image size

After the membrane layer was sufficiently eroded, ToF-SIMS was carried out in negative mode to analyze intracellular fatty acids such as palmitic, oleic and stearic acids. As the fatty acids are free molecules within the cytosol, the myelin-like segments were not observed. However, in accordance with the qPCR data, there is a noticeable increase in concentrations of fatty acids distributed across the DRG (figure S5). These ToF-SIMS images further validate that ES leads to an increase in lipid biosynthesis, eventually resulting in increase in structural lipids that make up myelin.