Identifying the functions of two biomarkers in human oligodendrocyte progenitor cell development

hOPCs were prepared in the Paediatric Laboratory of the Sixth Medical Centre of PLA General Hospital, China, using previously established methods of cultivation and identification. Briefly, hOPCs were induced by NSCs. The cells were cultured in a self-made medium at 37 °C in a humidified 5% CO₂ incubator (Additional file 1: Text S1). The hOPCs were identified using immunofluorescence staining. Monoclonal mouse anti-PDGFR-α (1:250, Cat. #C2318, CST, Boston, MA, USA), rabbit anti-NG2 (1:50, Cat. #ab83178, Abcam, Cambridge, Cambridgeshire, UK) and mouse anti-A2B5 (1:50, Cat. #MAB1416, R&D Systems, Minneapolis, MN, USA) antibodies were used to identify hOPCs. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (1:20, Cat.#28718-90-3, Sigma-Aldrich, St. Louis, MO, USA) for 5 min and then observed using fluorescence microscopy (IX-70, Olympus Corporation, Tokyo, Japan).

To evaluate the expression of PDGFR-α, NG2 and A2B5, scRNA-seq was performed at the Beijing Novogene Bioinformatics Technology Co (Beijing, China). We prepared a cell suspension containing a total number of cells greater than 1 × 10⁶. The prepared cell suspension was quickly loaded into the chromium microfluidic chip with 3′ chemistry, and a barcode of 10 × chromium controller was attached. The cells were then subjected to RNA reverse transcription. A sequencing library was constructed using reagents from the chromium single-cell 3′v2 kit (10 × Genomics, Pleasanton, CA, USA), according to the manufacturer’s instructions. Illumina was used for sequencing, according to the manufacturer's instructions (Illumina, San Diego, CA, USA).

For the analysis of PDGFR-α, NG2 and A2B5 expression in hOPCs, the cells were digested and washed once with buffer (pH 7.2 PBS, 0.5% BSA) and centrifuged at 500×g for 5 min at 4 °C. Fragment crystallisable (Fc) receptors were blocked with normal mouse serum for 10 min at 25 °C. The cells were surface-stained with PDGFR-α BV421 mouse anti-human (Cat. #562799, BD Biosciences, Franklin Lake, NJ, USA), NG2 APC mouse anti-human (Cat. #FAB2585A, R&D Systems, MN, USA), or A2B5 PE mouse anti-human (Cat. #130-093-581, Miltenyi Biotec, Bergisch-Gladbach, Germany) antibodies for 30 min at 4 °C. After surface staining, hOPCs were washed once with the abovementioned buffer and centrifuged at 500×g for 5 min at 4 °C. The cell pellet obtained was resuspended in the same buffer for flow cytometry analysis and cell sorting.

Cells were acquired using a FlowSight^® imaging flow cytometer (Amnis^®, part of EMD Millipore, MA, USA). Cell debris and dead cells were identified using the aspect ratio and area of the cells and removed. Approximately 7000 cells were obtained during each analysis. Channel 7 was used to detect BV421, Channel 3 to acquire PE, and Channel 11 to detect APC. Single colour control samples were compensated using a compensation matrix (.rif) and converted to data analysis files (.daf) and compensated image files (.cif) using the same settings. The data were analysed using Ideas software, version 6.2.

The fluorescent anti-NG2 and anti-A2B5 antibody-labelled cells were transferred into a MACSQuant^®Tyto^® Cartridge. The input sample contained 5 × 10⁶ hOPCs in 10 mL MACSQuant Tyto Running Buffer. Logical gating hierarchies were constructed using MACSQuant Tyto software before sorting. Cell debris, doublets, and dead cells were gated out, and a gate was set on the target cells. The sample was sorted at 4 mL/h using ~ 140 mbar pressure. Upon completion of sorting, negative cells were analysed using a FlowSight^® imaging flow cytometer to assess cell purity and yield. The expression of PDGFR-α in positive and negative cells was detected using a FlowSight^® imaging flow cytometer.

Total RNA was isolated from the four sorted populations using RNAiso Plus (Takara Bio). RNA-seq was performed at the Beijing Novogene Bioinformatics Technology Co, using the Illumina NovaSeq platform. RNA Nano 6000 Assay Kits and the Bioanalyzer 2100 system (Agilent Technologies, CA, USA) were used to evaluate RNA integrity. A NanoPhotometer^® spectrophotometer (IMPLEN, CA, USA) was used to check the purity of the RNA. Clustering and sequencing (Novogene Experimental Department) were conducted according to the manufacturer’s instructions. The index-coded samples were clustered using a cBot Cluster Generation System and TruSeq PE Cluster Kit v3-cBot-HS (Illumina). Finally, the libraries were sequenced using the Illumina NovaSeq platform and 150 bp paired-end reads were generated.

We used the edgeR package for the R statistical software with one scaling normalisation factor to adjust each sequenced library for read counts, and then performed differential gene expression analysis. The p values were adjusted using the Hochberg and Benjamini method. The expression level of genes was presented using fragments per kilobase of exon model per million mapped fragment (FPKM) values. FPKM of ≥ 1 indicated that the gene was expressed [23]. Differentially expressed genes were analysed for GO enrichment by the clusterProfiler R software package, and the bias in gene expression was corrected. GO terms included were as follows: biological process (BP), cellular component (CC), and molecular function (MF). When the corrected p-value was < 0.05, differentially expressed genes were considered to be significantly enriched in the GO terms.

Cell migration was measured in Transwell filters with 8-μm pores (Corning, Tewksbury, MA, USA). Inserts were coated with human fibronectin protein (Cat. #33016015, Thermo Fisher Scientific, Waltham, MA, USA). Sorted cells (2 × 10⁴ cells), in 200 μL of medium, were loaded into the upper chambers, and 500 μL of the hOPC medium was added to the chamber. After incubation for 18 h at 37 °C, a cotton swab was used to wipe the cells that had not migrated onto the upper surface of the chamber. The migrated cells were fixed with 4% paraformaldehyde and stained with DAPI (1:20). Images were acquired using a fluorescence microscope. The migrated cells were quantified using ImageJ software [24] by analysing cell nuclei from at least five randomly selected fields per chamber. The cell migration rate was calculated as follows: cell migration rate (V1%) = N1/N2 × 100%, where N1 refers to the initial cell seeding number, and N2 refers to the number of migrated cells. The experiment was repeated three times independently in the laboratory.

The proliferation of different cell populations was evaluated using cell counting kit-8 (CCK-8) assays (Cat. #CK04, Dojindo, Kumamoto, Japan). hOPCs (6000 cells/well) were seeded into 96-well plates and cultured for 24 h at 37 °C and 5% CO₂. The same volume of hOPC medium was added to the blank control group, and six parallel wells were set for each group without any cells. At different experimental time points (1, 3, 5, 7, 9, and 11 days), 10 μL of CCK-8 solution was added to the wells and incubated at 37 °C for 2 h. Six wells were set for each experiment at each time point. The absorbance at 450 nm was measured with a microplate reader (BioTek, Winooski, VT, USA). At least three experiments were performed, and each was tested in triplicate. The experiment was repeated three times independently in the laboratory [25].

Homozygous shiverer mice (The Jackson Laboratory, Maine) were maintained at the Sixth Medical Centre of PLA General Hospital in a specific pathogen-free environment. All animal experiments were performed according to protocols approved by the Sixth Medical Centre of PLA General Hospital Animal Care and Use Committee (Application No. SCXK-2012-0001). Newborn pups were transplanted within a day after birth. Different cell populations (1 × 10⁵/1.5 µL) were injected bilaterally into the corpus callosum using a mouse brain stereotaxic apparatus (Stoelting, USA). At the same time, untransplanted homozygous shiverer mice (n = 6) were randomly selected as the control group. The postoperative pups were returned to their mother and were weighed every day. After weaning, each mouse was reared separately. At 3 months of age, the mice were anaesthetised with 1% pentobarbital sodium. Then mouse brain samples were perfused with PBS and fixed with 4% paraformaldehyde (Thermo Fisher Scientific, Waltham, MA, USA). The brain tissue was cut in half along the sagittal suture. The coronal section of half of the brain tissue was used for myelin basic protein immunofluorescence staining (rat anti-MBP antibody, Abcam, Cambridge, UK). A fluorescence microscope was used to observe the fluorescently labelled samples. The sagittal section of the corpus callosum of the other half of the brain tissue was used for transmission electron microscopy (TEM, H7650-B, HITACHI, Tokyo, Japan). A field of view was randomly selected for each mouse, and the remyelination rate and G-ratio were calculated as follows: remyelination rate = N1/N2, where N1 refers to the number of axons ensheathed by myelin and N2 refers to the total number of axons; G-ratio value = axonal diameter/total myelin-ensheathed fibre diameter.

The experimental data in this study were analysed using GraphPad Prism version 8 software (GraphPad Software, San Diego, CA, USA, www.​graphpad.​com) and SPSS version 22.0. Differences were considered to be statistically significant at p < 0.05. Statistical significance was determined using independent samples Student’s t-tests or one-way ANOVA followed by Bonferroni tests. All experimental data are expressed as mean ± SD.

We identified hOPCs using cell morphology and immunofluorescence staining. hOPCs had a typical bipolar, bead-like appearance, and expressed PDGFR-α, NG2, and A2B5 (Fig. 1a). Based on the results of flow cytometry analysis, the proportion of PDGFR-α+ cells was 74.13 ± 1.74% of the whole cell population; A2B5+ cells made up 25.97 ± 0.74%; NG2+  cells comprised 14.41 ± 0.81% (Fig. 1b). Volume uniformity of hOPCs, which had a cell diameter < 20 µm, and the expression of the three cell markers on a single-cell membrane were observed (Fig. 1c). scRNA-seq showed that the proportions of PDGFR-α+, A2B5+ and NG2+ cells in hOPCs were 67% (n = 5309), 16% (n = 1243) and 13% (n = 1002), respectively (Fig. 1d).

To obtain NG2+/− and A2B5+/− cells, we performed cell sorting on hOPCs. After sorting, the viability of the sorted cells was above 98%. In the negative group, the proportion of A2B5+ cells dropped from 28.1 to 4.03%, and the proportion of NG2+ cells decreased from 8.22 to 1.01%. NG2+ and A2B5+ cells were almost completely absent in the NG2− and A2B5− cell populations (Additional file 3: Fig. S1). Next, we detected the expression of PDGFR-α in NG2+/− and A2B5+/− cells. The results showed that PDGFR-α+ cells were present in the four groups of cells (Fig. 2a, b). The level of PDGFR-α+ in NG2− cells was the highest, reaching 85 ± 2.13%, while that in NG2+ cells was the lowest, reaching only 24.95 ± 1.89%. The difference between the groups was statistically significant (** ~ ****, Fig. 2c, d).

In order to study the differential gene expression of the four groups, we performed RNA-seq. The results showed that, compared with A2B5+ cells, NG2+ cells had 1737 upregulated and 1225 downregulated genes. Compared with A2B5− cells, A2B5+ cells had 633 upregulated and 333 downregulated genes. Compared with A2B5+ cells, NG2− cells had 1381 upregulated and 834 downregulated genes. Compared with A2B5− cells, NG2− cells had 967 upregulated and 669 downregulated genes. Compared with NG2− cells, NG2+ cells had 466 upregulated and 352 downregulated genes. Compared with NG2+ cells, A2B5− cells had 1126 upregulated and 1400 downregulated genes (Fig. 3a). Figure 3b shows a heat map of the genes which were differentially expressed among the four cell populations.

Next, we detected the expression of the main markers of OPCs, neurons, astrocytes, chondrocytes, and OLs in these four groups of cells. The results showed that in NG2+ and A2B5+ cells, in addition to the expression of NG2 and A2B5, PDGFR-α was also expressed. More importantly, PDGFR-α was abundantly expressed in NG2− and A2B5− cells. Further analysis showed that among the four groups of cells, PDGFR-α expression was the highest in NG2− cells and the lowest in NG2+ (* ~ ****, Fig. 3c). In addition, among these four groups of cells, the expression of neuron markers TUBB3 and NEFM, astrocyte markers S100β and GFAP, and chondrocyte markers COL2A1 and ACAN was much lower than that of PDGFR-α (****, p < 0.0001). In addition, the expression of neuron and astrocyte markers in the positive cell population was higher than that in the negative cell population (****, p < 0.0001), but there was no difference between NG2+ and A2B5+ cells, as well as NG2− and A2B5− cells. There was no difference in the expression of chondrocyte markers among these four groups (Fig. 3d). However, the main OL lineage markers were not expressed in these four groups of cells. These markers include GALC, PLP1, APC, CNP, MBP, MOG, and MAG (Additional file 2: Table S1).

We then performed GO enrichment analysis on the upregulated and downregulated genes in each group. Figure 4a shows the key GO terms. BP mainly included cell proliferation (GO:0050673), migration (GO:0050673), oligodendrocyte differentiation (GO:0048709), and myelination (GO:0042552); CC mainly included extracellular matrix (GO:0031012) and axon part (GO:0033267); MF included actin-binding (GO:0003779). Bubble plots showed that in NG2+ and A2B5+ cells, the upregulated genes were mainly involved in cell migration and proliferation, while the downregulated genes were mainly involved in oligodendrocyte differentiation (Fig. 4b). The results of the other groups are provided in Additional file 4: Fig S2. In terms of cell migration and proliferation, the gene enrichment of the four groups of cells from high to low was A2B5− > NG2− > NG2+ > A2B5+. The gene enrichment intensity for oligodendrocyte differentiation from high to low was NG2− > A2B5− > A2B5+ > NG2+.

We used Transwell assays to evaluate the migration of the cells. After the cells had migrated for 18 h, the nuclei were stained with DAPI (Fig. 5a). The migration rates of A2B5+ and A2B5− cell populations were 4.41 ± 0.38% and 14.82 ± 0.48%, respectively. The migration rates of NG2+ and NG2− cell populations were 7.89 ± 0.75% and 10.04 ± 0.18%, respectively. Statistical analysis showed that the migration of NG2+ was stronger than that of A2B5+ (*, p = 0.0278) and that the migration of NG2− was weaker than that of A2B5− (*, p = 0.0267). The migration of A2B5+ was weaker than that of A2B5− (*, p = 0.0240), and the migration of NG2+ was weaker than that of NG2− (**, p = 0.0017) (Fig. 5b). The proliferation assay results showed that the proliferation of NG2+ was stronger than that of A2B5+ (**, p = 0.0015), while the proliferation of NG2− was weaker than that of A2B5− (**, p = 0.0028). The proliferation of A2B5− was stronger than that of A2B5+ (**, p = 0.0027), and the proliferation of NG2+ was weaker than that of NG2− (**, p = 0.0079) (Fig. 5c).

We evaluated the myelinating ability of these four groups of cells in vivo. The mature myelin is characterised by the expression of MBP. Therefore, MBP immunofluorescence staining was performed on shiverer mouse brain tissue to confirm the presence of mature myelin. There was filamentous MBP expression in the corpus callosum of transplanted mice (Fig. 6a). In addition, MBP can condense consecutive layers of myelin together; therefore, we used electron microscopy to assess myelin condensation. Untreated shiverer mice axons typically had no myelin ensheathment or single loose wrapping of uncondensed myelin, such that major dense lines did not form (Fig. 6b, arrowhead). The number of myelin sheaths in the corpus callosum of A2B5− and NG2− cells was more than that of A2B5+ and NG2+ cells (Fig. 6c). Although the four groups of axons were wrapped in multiple layers of compact myelin, the major dense lines of the negative cell groups were more obvious than those of the positive cell groups (Fig. 6d, arrowhead). The remyelination rate and G-ratio analyses of the four groups of cells showed that NG2− cells exhibited the strongest myelination effect, while NG2+ cells had the weakest effect (* ~ ****, Fig. 6e, f).