Abstract
The oligodendrocyte (OL) lineage transcription factor Olig2 is expressed throughout oligodendroglial development and is essential for oligodendroglial progenitor specification and differentiation. It was previously reported that deletion of Olig2 enhanced the maturation and myelination of immature OLs and accelerated the remyelination process. However, by analyzing multiple Olig2 conditional KO mouse lines (male and female), we conclude that Olig2 has the opposite effect and is required for OL maturation and remyelination. We found that deletion of Olig2 in immature OLs driven by an immature OL-expressing Plp1 promoter resulted in defects in OL maturation and myelination, and did not enhance remyelination after demyelination. Similarly, Olig2 deletion during premyelinating stages in immature OLs using Mobp or Mog promoter-driven Cre lines also did not enhance OL maturation in the CNS. Further, we found that Olig2 was not required for myelin maintenance in mature OLs but was critical for remyelination after lysolecithin-induced demyelinating injury. Analysis of genomic occupancy in immature and mature OLs revealed that Olig2 targets the enhancers of key myelination-related genes for OL maturation from immature OLs. Together, by leveraging multiple immature OL-expressing Cre lines, these studies indicate that Olig2 is essential for differentiation and myelination of immature OLs and myelin repair. Our findings raise fundamental questions about the previously proposed role of Olig2 in opposing OL myelination and highlight the importance of using Cre-dependent reporter(s) for lineage tracing in studying cell state progression.
SIGNIFICANCE STATEMENT Identification of the regulators that promote oligodendrocyte (OL) myelination and remyelination is important for promoting myelin repair in devastating demyelinating diseases. Olig2 is expressed throughout OL lineage development. Ablation of Olig2 was reported to induce maturation, myelination, and remyelination from immature OLs. However, lineage-mapping analysis of Olig2-ablated cells was not conducted. Here, by leveraging multiple immature OL-expressing Cre lines, we observed no evidence that Olig2 ablation promotes maturation or remyelination of immature OLs. Instead, we find that Olig2 is required for immature OL maturation, myelination, and myelin repair. These data raise fundamental questions about the proposed inhibitory role of Olig2 against OL maturation and remyelination. Our findings highlight the importance of validating genetic manipulation with cell lineage tracing in studying myelination.
Introduction
Myelin-producing oligodendrocytes (OLs) provide crucial support for normal neuronal function in the mammalian CNS and are responsible for functional regeneration after injuries or pathologic insults (Trapp et al., 1998; Popescu and Lucchinetti, 2012; McKenzie et al., 2014; Franklin and Goldman, 2015; Bei et al., 2016). OL development and myelination occur through successive lineage progression from neural stem cells to primitive or precursor OL progenitor cells (pri-OPCs) (e.g., Olig1/2+) to committed OPCs (e.g., PDGFRa+ or NG2+), which further differentiate into premyelinating immature OLs and finally into mature myelinating OLs (Kessaris et al., 2006; Richardson et al., 2006; Kang et al., 2010; Gregath and Lu, 2018; Marques et al., 2018; Weng et al., 2019). Premyelinating immature OLs express postmitotic OL markers, including 2′,3′-cyclic nucleotide 3′-phosphodiesterase (CNP), proteolipid protein (PLP), myelin-associated oligodendrocyte basic protein (MOBP), and myelin oligodendrocyte glycoprotein (MOG), which persist in terminally differentiated myelinating OLs (Emery, 2010; Gregath and Lu, 2018). This stepwise differentiation process requires the exquisitely precise coordination of a series of extracellular and intracellular cues, including lineage-expressing transcriptional regulators, such as Olig1/2, Sox10, and Myrf, to regulate specification and differentiation programs (Wegner, 2008; Zuchero and Barres, 2013; Emery and Lu, 2015).
Olig2 is a key regulator of OL lineage specification and differentiation (Lu et al., 2002; Takebayashi et al., 2002; Zhou and Anderson, 2002). Olig2-null mice fail to develop the committed oligodendroglial cells, whereas conditional deletion of Olig2 in cortical neural progenitors results in OL lineage differentiation arrested at the progenitor stage (Yue et al., 2006), leading to hypomyelination despite normal patterning of OPCs in the cortex. This suggests that Olig2 is required for the transition of OPCs into myelin forming OLs.
In a previous study, unexpectedly, Mei et al. (2013) reported that Olig2 deletion in the Plp1+ immature OLs significantly enhanced OL maturation in the developing CNS and accelerated myelination/remyelination in adult mice, suggesting that Olig2 opposes OL differentiation and maturation. In contrast to the reported inhibitory role of Olig2 in the transition from immature OLs to mature myelinating OLs (Mei et al., 2013), experimental data reported here demonstrate that Olig2 is instead required for the maturation of Plp1+ immature OLs. Further, Olig2 deletion in immature OLs directed by immature OL-expressing Mobp or Mog promoter-driven Cre lines during premyelinating stages of OL development did not enhance OL maturation in the CNS. In contrast, Olig2 deletion in these lines reduced OL maturation and myelination. While Olig2 ablation in Plp1+ OLs at adult stages did not interfere with myelin maintenance by mature OLs, we found that the remyelination process after lysolecithin-induced demyelinating injury was impaired in the spinal cord of Olig2-ablated animals. Furthermore, our genomic occupancy studies revealed that Olig2 binds to a set of the enhancer/promoter elements of myelinogenesis-related genes in immature and mature OLs and positively regulates the immature-to-mature OL transition. Together, our findings from multiple immature OL-expressing Cre lines demonstrate that Olig2 is a crucial regulator of immature-to-mature OL differentiation and myelin repair, rather than opposing OL myelination or remyelination as previously reported (Mei et al., 2013).
Materials and Methods
Animals
Olig2 floxed mice (Olig2fl/fl) (Yue et al., 2006) were crossed with Cnp-Cre, Mobp-iCre, and Mog-iCre mice to generate Olig2 cKO-Cnp (Olig2fl/fl;Cnp-Cre+/−), Olig2 cKO-Mobp (Olig2fl/fl;Mobp-iCre), and Olig2 cKO-Mog (Olig2fl/fl;Mog-iCre) mice, respectively. Pdgfra-CreERT mice were crossed with Olig2fl/fl mice to generate the OPC-inducible Olig2 iKO (iKO-Pdgfra) mice. Plp1-CreERT and tdTomato reporter mice were mated with Olig2fl/fl mice to generate the mature OL-inducible Olig2 iKO (iKO-Plp) mice. Mice of either sex were used in the study, and littermates were used as controls unless otherwise indicated. All mouse strains used in this study were generated and maintained on a mixed C57BL/6;129Sv background and were housed in a pathogen-free vivarium with a 12 h light/dark cycle with free access to normal chow food and water. All studies complied with all relevant animal use guidelines and ethical regulations. All animal experiments were approved by the Institutional Animal Care and Use Committee at the Cincinnati Children's Hospital Medical Center.
Tamoxifen administration
Tamoxifen (TAM, Sigma-Aldrich, T5648) was dissolved in a vehicle of ethanol and sunflower seed oil (1:9) to at a stock concentration of 20 mg/ml. For neonatal pups, TAM (0.1 mg/g body weight) was administered by intraperitoneal injection once daily for 3 or 5 consecutive days as indicated. For adult mice, TAM was administered by intraperitoneal injection once daily for 7 d. For the remyelination study, we use two cycles of 5 d TAM treatment with a 1 d-off interval for lysolecithin (LPC) injection to ensure effective floxed allele recombination as previously described (Lopez-Juarez et al., 2013; Wang et al., 2020b). TAM dosing frequency and time for Olig2 ablation in iKO-Pdgfra or iKO-Plp mice were performed as described previously for TAM-induced recombination of Olig2fl/fl or other floxed alleles at different developmental stages (Mei et al., 2013; Zuo et al., 2018; Wang et al., 2020b) or adulthood (Doerflinger et al., 2003; Leone et al., 2003; Lopez-Juarez et al., 2013; Wang et al., 2020b).
Tissue processing and ISH
Mice at different developmental stages were anesthetized with ketamine/xylazine and perfused with PBS followed by 4% PFA. Spinal cords or brains were dissected, fixed in 4% PFA overnight, washed with PBS for 3 times, dehydrated in 25% sucrose at 4°C, embedded in OCT, and cryo-sectioned at 16 μm. For ISH, cryosections of spinal cord were used with digoxigenin (DIG)-labeled antisense riboprobes specific for Mbp and Plp1 as previously described (Wang et al., 2020b). Antisense riboprobes were synthesized with T3 or T7 RNA polymerase (Promega, P207B and P2083) and labeled with DIG RNA Labeling Mix (Roche, 11277073910). An anti-DIG antibody conjugated to alkaline phosphatase (Roche, 16646821) was applied to the probe hybridized tissue sections and stained with BCIP/NBT chromogenic substrates (Sigma, B5655).
Electron microscopy and morphometric analysis
Tissue processing was performed as described previously (He et al., 2016). Briefly, mice were anesthetized with ketamine/xylazine and perfused with 0.1 m cacodylate and 4% PFA/2.5% glutaraldehyde in 0.1 m cacodylate. Spinal cord and optic nerves were dissected and postfixed in 1% OsO4. Semithin sections were cut and stained with toluidine blue to evaluate quality and orientation of tissue. Ultrathin sections were stained with lead citrate for electron microscopy imaging. The g-ratio, the ratio of the inner-to-outer diameter of a myelinated axon, of myelinated axons was analyzed using electron microscopy images from ultrathin sections.
Immunofluorescence analysis
Cryosection sections or vibratome sections (50 μm) were permeabilized and blocked in blocking buffer (0.2% Triton X-100 and 5% normal donkey serum in 1× PBS) for 1 h at room temperature, followed with primary antibodies overnight at 4°C. Antibodies used in this study were as follows: rabbit anti-OLIG2 (Millipore, AB9610), rat anti-PDGFRα (BD Biosciences, 558774), mouse anti-adenomatous polyposis coli (CC1, Oncogene Research, OP80), goat anti-MBP (Santa Cruz Biotechnology, sc-13914), goat anti-SOX10 (Santa Cruz Biotechnology, sc-17 342), rabbit anti-aspartoacylase (ASPA, Millipore, ABN1698), guinea pig anti-Olig2 (Skaggs et al., 2011), rabbit anti-Olig1 (Millipore, MAB5540), and rabbit anti-GST-π (Enzo Life Science, ADI-MSA-102-E). After washing with 1× PBS 3 times, sections were incubated with secondary antibodies conjugated to Cy2, Cy3, or Cy5 (Jackson ImmunoResearch Laboratories, 715-225-151, 711-165-152, 705-175-147) for 1.5 h at room temperature, counterstained with DAPI for 5 min, washed in PBS 3 times, and mounted with Fluoromount-G (Southern Biotechnology, 0100-01). The brain and spinal cord tissues were sectioned at 20 μm by taking 1 μm serial optical thin sections for the quantification of labeled cells. Somata were defined as the counting unit. Six sections spaced 120 μm apart were used to assess the number of labeled cells, avoiding any potential of double counting or tissue processing artifacts from consecutive sections impacting to ensure that the area counted for each brain was constant. A sequence of at least eight sections per animal were obtained with the ImageJ or Imaris image software (Bitplane) to estimate the number of stained cells. The cell counts were quantified by personnel blinded to the experimental condition of each sample.
LPC-induced demyelinating injury
LPC-induced demyelination was conducted in the ventrolateral spinal white matter of ∼8- to 10-week-old mice as previously described (Wang et al., 2020b). Briefly, after exposing the T3-T4 spinal vertebrae and clearing the meningeal tissue in the intervertebral space, the dura was pierced with a fine dental needle; 0.5 µl of 1% LPC (Sigma-Aldrich, L4129) was injected via a Hamilton syringe attached to a glass micropipette into the ventrolateral white matter via a stereotactic instrument (RWD Life Sciences). The date of LPC injection was denoted as dpi 0. Mice were then left for indicated periods, and spinal cords were collected for immunostaining and electron microscopic analysis. Injuries were conducted in a genotype-blinded manner.
RNA extraction, RNA-seq, and data analysis
RNA was extracted from lesion tissues microdissected from control and Olig2-deficient mice at dpi 14. Total RNA was extracted per the Trizol (Invitrogen) protocol. RNA sequencing (RNA-seq) libraries were prepared using SMART-Seq v4 Ultra Low Input RNA Kit. Massively parallel 75 bp PE sequencing was completed on an Illumina HiSeq2500 to acquire ∼25 million fragments per sample. RNA sequencing files were mapped to the genome (mm10) using TopHat with default settings (http://tophat.cbcb.umd.edu/). We used Cuff-diff to (1) estimate per kilobase of transcript per million mapped read (FPKM) values for known transcripts and (2) analyze differentially expressed transcripts. In all differential expression tests, significance thresholds were set at p < 0.05. Heatmaps of gene expression were generated using R (version 3.2.1) based on log2(FPKM) values. Gene ontology analysis of gene expression changes was performed using Gene Set Enrichment Analysis (GSEA; http://www.broadinstitute.org/gsea/index.jsp). The RNA-seq data generated in this study are deposited in the NCBI Gene Expression Omnibus under accession number: GSE194258.
Statistical analysis
All analyses were done using GraphPad Prism 8.00 (www.graphpad.com). Significance values are described in figures, figure legends, and/or in Results. Alternatively, significance levels are indicated. Data distribution was assumed to be normal, but this was not formally tested. Statistical significance was determined using unpaired two-tailed Student's t tests as indicated. Quantifications were performed from at least three independent experiments, no randomization was used to collect all the data, but they were quantified blindly.
Results
Deletion of Olig2 in Pdgfra+ OPCs and Cnp+ immature OLs inhibits OL maturation
We set out to assess the role of Olig2 in CNS myelination at different stages of OL lineage development, which can be characterized by sequential expression of Pdgfra, Cnp, Plp, Mobp, and Mog (Fig. 1A). We first generated inducible OPC-expressing Pdgfra-CreERT;Olig2fl/fl mice (referred to as iKO-Pdgfra mice) by breeding Pdgfra-CreERT mice (Kang et al., 2010) with Olig2fl/fl mice (Yue et al., 2006) to inactivate Olig2 in OPCs. Olig2 ablation was induced by TAM administration at the neonatal stage from P1 to P3 (Fig. 1B), a stage during which there is extensive OPC expansion (Nishiyama et al., 2021). Olig2 expression was substantially attenuated in the developing brain of iKO-Pdgfra mice at P7 and P14, indicating efficient TAM-induced recombination of Olig2 in OPCs (Fig. 1C,E). The number of PDGFRα+ OPCs was comparable in Olig2-ablated iKO-Pdgfra mice and control Pdgfra-CreERT mice (Fig. 1C,E), consistent with a previous study with iKO-NG2 mice (NG2- CreERT;Olig2fl/fl) (Zuo et al., 2018). However, MBP expression and the number of CC1+ OLs were substantially reduced in the brains of iKO-Pdgfra mutants compared with the control (Fig. 1D,E). These data suggest that Olig2 deletion impairs the differentiation of Pdgfra+ OPCs without affecting OPC formation.
Deletion of Olig2 in OPCs inhibits OL differentiation and myelination. A, Schematic representation of Olig2 and marker expression during the OL lineage development. B, Schematic diagram for TAM administration and sample collection in Olig2+/+ control (Ctrl, Pdgfra-CreERT) and iKO-Pdgfra mice. C, Representative images of cortical sections from control (Ctrl) mice and iKO-Pdgfra mice at P14 stained for Olig2 (green) and PDGFRα (red). Scale bars, 30 µm. D, Representative images of the corpus callosum of Ctrl mice and iKO-Pdgfra mice at P14 immunostained for MBP (red) and CC1 (green). Scale bars, 50 µm. E, Quantification of Olig2+ cells (left), PDGFRα+ cells (middle), and CC1+ cells (right) in the iKO-Pdgfra mice relative to Ctrl mice at P7 and P14. n = 3 animals per group. F, Grayscale images of MBP in P14 cortical sections of Ctrl and cKO-Cnp mice. Scale bars, 1 mm. G, Electron micrograph analysis of optic nerves of Ctrl and cKO-Cnp mice at P14. Scale bars, 1 µm. H, Percentage of myelinated axons in Ctrl and cKO-Cnp optic nerves. n = 4 animals per group. I, Quantification of myelin g-ratio in the optic nerve from Ctrl and cKO-Cnp mice. More than 300 axons were counted from 4 animals per group. Data are mean ± SEM. **p < 0.01; ***p < 0.001; unpaired two-tailed Student's t test.
Mei et al. (2013) suggested that Olig2 would be deleted in OPCs by Cnp-Cre. However, Cnp-Cre mediates gene deletion extensively in late postmitotic OPCs and immature OLs during premyelinating stages (Lappe-Siefke et al., 2003; Zuchero et al., 2015; S. Zhang et al., 2018) (Fig. 1A). Similar to the previous observations (Mei et al., 2013), inactivation of Olig2 by Cnp-Cre in Olig2fl/fl;Cnp-Cre mice (referred to as cKO-Cnp mice) led to a prominent reduction in MBP expression in the forebrain of cKO-Cnp mice compared with control Cnp1-Cre mice at P14 (Fig. 1F). The number of myelinated axons were also significantly reduced (Fig. 1G,H) as was the thickness of myelin sheath in the optic nerve of cKO-Cnp mice compared with controls (Fig. 1I). These data suggest that Olig2 is required for the differentiation of OPCs and immature OLs (Cnp+) into mature myelinating OLs in the developing brain.
Olig2 is required for the maturation of Plp1+ immature OLs
In the previous study by Mei et al. (2013), they reported that Olig2 deletion in Plp1+ immature OLs directed by a Plp1-CreERT line promoted OL maturation, but counterintuitively impaired the maturation or differentiation of immature OLs expressing Cnp1+ cells. To investigate the role of Olig2 in Plp1+ immature OL differentiation, we conditionally ablated Olig2 in Plp1+ immature OLs by breeding Olig2fl/fl mice with the same Plp1-CreERT line (Doerflinger et al., 2003) to generate Olig2fl/fl;Plp1-CreERT mice (referred to as iKO-Plp mice) that also carried a Cre-dependent Rosa-tdTomato reporter. Control and iKO-Plp pups were treated with TAM once daily from P5 to P9 and harvested at P14 as previously reported (Mei et al., 2013) (Fig. 2A). In contrast to the enhanced OL differentiation and myelination reported by Mei et al. (2013), we observed hypomyelination in the cortex of iKO-Plp mice relative to control mice at P14 (Fig. 2B). In iKO-Plp mice, the percentages of CC1+ differentiated OLs in the tdTomato reporter-expressing cells were significantly lower in both the cortex and corpus callosum at P14 (Fig. 2C–F). Thus, our results using these immature OL-expressing Cre and CreER lines indicate that Olig2 deletion impairs the maturation of immature OLs in the developing brain.
Deletion of Olig2 in immature OL at the neonatal stages inhibits OL maturation. A, Control (Ctrl; Plp1-CreERT) and iKO-Plp mice carrying the tdTomato reporter were given TAM daily from P5 to P9 and harvested at indicated stages. B, Representative grayscale images of cortical sections from Ctrl and iKO-Plp mice at P14 stained for MBP. Scale bars, 100 µm. C, Representative images of tdTomato (red) and staining for Olig2 (blue) and CC1 (green) cells in the cortex of Ctrl and iKO-Plp mice at P14. Arrows indicate the tdTomato+ cells. Scale bars, 30 µm. D, Percentages of tdTomato reporter+ cells in cortex that express Olig2 (left) and CC1 (right) at P14. E, Representative images of tdTomato (red) and staining for Olig2 (blue) and CC1 (green) cells in the corpus callosum of Ctrl and iKO-Plp mice at P14. Arrows indicate the tdTomato+ cells. Scale bars, 30 µm. F, Percentages of tdTomato reporter+ cells in corpus callosum that express Olig2 (left) and CC1 (right) at P14. G, Representative images of tdTomato (red) and staining for cytoplasmic Olig1 (Olig1, green) in the corpus callosum from Ctrl and iKO-Plp mice at P14. Scale bars, 20 µm. H, Left, Percentages of tdTomato+ cells in the corpus callosum with cytoplasmic staining for Olig1. Right, Total number of tdTomato+ cells and cytoplasmic Olig1+ cells in the corpus callosum per mm2 at P14. I, Representative images of GST-π immunostaining in the cortex from Ctrl and iKO-Plp mice at P14. Scale bars, 60 µm. J, Fold change of the GST-π+ cell number in the iKO-Plp mice relative to Ctrl mice at P14. K, Percentages of CC1+ cells among tdTomato+ cells in the cortices of Ctrl and iKO-Plp mice at indicated stages. n = 5 animals per genotype. Data are mean ± SEM. ***p < 0.001; **p < 0.01; unpaired two-tailed Student's t test.
Olig1 translocation from the nucleus to the cytoplasm has been used as a marker of OL maturity in studies of both development and remyelination (Arnett et al., 2004). We found that ∼80% tdTomato+ OLs exhibited cytoplasmic Olig1 in the corpus callosum of TAM-treated control mice, whereas only ∼22% of tdTomato+ cells in the iKO-Plp mice had cytoplasmic Olig1 expression at P14 (Fig. 2G,H). This is in contrast to the previously reported increase of Olig1-expressing cells on Olig2 deletion (Mei et al., 2013). Similarly, expression of another mature OL marker GST-π (Tamura et al., 2007; Wang et al., 2020a) was reduced in the cortex of iKO-Plp mice compared relative to controls (Fig. 2I,J). In addition, similar to the phenotype observed at P14, iKO-Plp mice also had less CC1+ OLs among tdTomato+ cells in the cortices analyzed at later stages, such as P21 and P30, compared with the controls (Fig. 2K), suggesting a sustained defect in immature OL differentiation in the Olig2-ablated mice. These data further support that Olig2 is indispensable for the maturation of immature OLs in the developing brain.
Olig2 is required for the maturation of Mog- or Mobp-expressing immature OLs
To further define the role of Olig2 in the maturation of postmitotic, immature OLs, we crossed Olig2fl/fl mice with immature OL-expressing Mobp-iCre (Mobp-ires-Cre) mice and Mog-iCre (Mog-ires-Cre) mice (Hovelmeyer et al., 2005). Mobp- or Mog- promoter driven Cre-mediated reporter expression was detected predominantly in postmitotic, immature, and mature OLs, but not in PDGFRα+ OPCs in the early developing brain (Fig. 3A–D). Strikingly, the density of MBP+ myelinating fibers was substantially reduced in both brain and spinal cord of Olig2fl/fl;Mobp-iCre mice (referred as cKO-Mobp) compared with controls at P7 (Fig. 4A–C). Similarly, in the brain and spinal cord of Olig2fl/fl; Mog-iCre mice (referred as cKO-Mog), MBP expression was significantly reduced compared with controls at P7 (Fig. 4D–F). Consistent with these findings, the number of CC1+ OLs in the corpus callosum was reduced substantially in the cKO-Mog mice compared with the control (Fig. 4G,H). In addition, the number of mature OLs with cytoplasmic Olig1 was reduced in the cortex of cKO-Mog mice at P14 (Fig. 4I,J). Collectively, the data from multiple mouse lines with Olig2 ablated in immature OLs directed by Plp1-CreERT, Mobp-Cre, and Mog-Cre indicate that Olig2 is required for the maturation of postmitotic immature OLs in the developing CNS, and does not inhibit OL maturation, which is in stark contrast to the previous conclusion (Mei et al., 2013).
Mobp- and Mog-Cre expression in postmitotic immature and mature OLs in the developing brain and spinal cord. A, Representative images of fluorescent tdTomato expression (red) in postmitotic OLs in the developing corpus callosum and spinal cord in Mobp-iCre;Rosa-tdTomato mice at P14 stained for CC1 (green) and PDGFRα (purple). Scale bars, 50 µm. B, Enlarged inset of image in A. Arrows indicate CC1+ postmitotic OLs positive for tdTomato. Arrowhead indicates PDGFRa+ OPC that does not express tdTomato. Scale bar, 20 µm. C, Representative images of fluorescent tdTomato expression (red) in postmitotic OLs in the developing corpus callosum and spinal cord in Mog-iCre;Rosa-tdTomato mice at P14 stained for CC1 (green) and PDGFRα (purple). Scale bars, 50 µm. D, Enlarged inset of the spinal cord in C. Arrows indicate CC1+ postmitotic OL cells positive for tdTomato. Arrowhead indicates PDGFRa+ OPC that does not express tdTomato. Scale bar, 20 µm.
Olig2 ablation causes maturation arrest and myelination deficits. A, Representative images of the cerebral white matter of control (Olig2fl/fl) and cKO-Mobp mice at P7 stained for MBP. Scale bars, 100 µm. B, Representative images of P7 spinal cords from control and cKO-Mobp mice immunostained for Olig2 (red) and MBP (green). Scale bars, 20 µm. C, MBP density in the brain and spinal cord of cKO-Mobp mice relative to control mice at P7. D, Representative images of the cerebral white matter of control (Olig2fl/fl) and cKO-Mog mice at P7 stained for MBP. Scale bars, 100 µm. E, Representative images of P7 spinal cords from control and cKO-Mog mice immunostained for Olig2 (red) and MBP (green). Scale bars, 100 µm. F, MBP density in the brain and spinal cord of cKO-Mog mice relative to control mice at P7. G, Representative images of P14 corpus callosum immunostained for CC1 (green) and Olig2 (red). Dotted lines indicate the subcortical white matter tracts. Scale bars, 20 µm. H, Fold change in CC1+ cell number (left) and percentage of CC1+ cells among Olig2+ cells in the cKO-Mog mice compared with control mice. I, Representative images of P14 cortex from control and cKO-Mog mice immunostained for Olig1 (red) and Olig2 (green). Scale bars, 20 µm. J, Frequency of cytoplasmic Olig1+ cells in the cortex of cKO-Mog mice relative to control mice. n = 4 animals per genotype. Data are mean ± SEM. ***p < 0.001; **p < 0.01; *p < 0.05; unpaired two-tailed Student's t test.
Olig2 deletion in mature OLs in adult mice does not cause demyelination
To examine the function of Olig2 in OL myelin maintenance, we ablated Olig2 in mature OLs by treating adult iKO-Plp mice with TAM for 7 d in adult mice (Fig. 5A). The brain and spinal cords were analyzed 1 month after TAM treatment. As expected, Olig2 expression was largely depleted in the iKO-Plp brains after TAM treatment (Fig. 5B–D). Similarly, expression of Sox10, a downstream target of Olig2, was downregulated substantially in Olig2-ablated OLs in iKO-Plp adult mice compared with controls, while the expression of a mature OL marker GST-π was not altered substantially (Fig. 5B–D). This is in keeping with the observation that the loss of Sox10 in mature OLs does not cause myelin abnormalities in adult mice (Turnescu et al., 2018). In addition, we did not observe any significant alteration of OL cell death assayed by an apoptotic marker, cleaved-Caspase 3, in Olig2-ablated OLs in iKO-Plp mice (Fig. 5E), suggesting that Olig2 ablation does not affect mature OL survival in adult mice.
Olig2 deletion in mature OLs does not cause demyelination. A, Ctrl (Plp1-CreERT) and iKO-Plp adult mice (8-week-old) carrying the tdTomato reporter were administered with TAM daily for 7 d and killed at 1 (1 M) or 11 months (11 M) after TAM. B, C, Representative images of tdTomato (red), Olig2 (green), and staining for SOX10 (B) and GST-π (C) in the cortex from Ctrl and iKO-Plp mice at 1 month after TAM injection. Arrows indicate the tdTomato+ cells. Scale bars, 20 µm. D, Percentages of tdTomato+ cells in the cortex that express Olig2 (left), SOX10 (middle), and GST-π (right) in Ctrl and iKO-Plp mice. n = 3 animals per genotype. E, Representative images (left) of tdTomato (red), Olig2 (green), and staining for cleaved caspase 3 (CC3, purple) and quantifications (right, n = 3 animals per genotype) in the cortex from Ctrl and iKO-Plp mice at 1 month after TAM injection. Scale bars, 20 µm. F, Representative images of Ctrl and iKO-Plp mice cortex stained for MBP or Olig2 at 11 months after TAM injection. Scale bars, 50 µm. G, Quantifications for the area of MBP+ cells (left) and Olig2+ cells (right) in the cortex of iKO-Plp mice relative to controls. Scale bars, 20 µm. n = 3 animals per genotype. H, Representative EM images showing optic nerve and spinal cord in the Ctrl and iKO-Plp brains at 11 months after TAM injection. Scale bars, 2 µm. I, Quantification of myelin g-ratio in the optic nerve (left) and spinal white matter (right) from Ctrl and iKO-Plp mice. More than 150 axons were counted from 3 animals per group. n = 3 animals per genotype. Data are mean ± SEM. **p < 0.01; *p < 0.05; unpaired two-tailed Student's t test.
To further examine the myelin integrity after Olig2 deletion in mature OLs, the brain and spinal cords were analyzed 11 months after TAM treatment in adult mice (Fig. 5A). We did not observe substantial differences in MBP expression in the iKO-Plp brains compared with controls, although reduction of Olig2 expression persisted in the Olig2-mutant mice (Fig. 5F,G). In addition, ultrastructural analysis of optic nerves and spinal cords showed no signs of substantial demyelination in the white matter tracts of TAM-treated iKO-Plp adult mice (Fig. 5H). The myelin thickness assayed by g-ratio in the optic nerve and spinal cord was also comparable between control and iKO-Plp mice (Fig. 5I). These data suggest that Olig2 is dispensable for mature OL survival and myelin maintenance in adult mice.
Olig2 regulates the OL maturation program in differentiating OLs
To investigate the potential mechanisms behind the arrested maturation of immature OLs in Olig2-deficient animals, we analyzed Olig2 genome-wide occupancy in maturing OLs, which were differentiated from OPCs after 3 d of triiodothyronine exposure (Yu et al., 2013). We observed an enrichment of Olig2 occupancy at the gene loci related to OL maturation in both immature OLs and mature OLs, such as the genes associated with transcriptional control of myelinogenesis (e.g., Sox10, Zeb2, Olig2, Nk2-2, Tcf7l2, Myrf, and Zfp191) and myelin components (Mbp, Mog, Utg8, Opalin, Omg, Drd1, and Fyn1) (Fig. 6A,B). These Olig2-occupied enhancer/promoter elements were associated with an increase in the activating histone mark, H3K27Ac, during the transition from immature to mature OLs (Fig. 6A,B). Given that H3K27ac enrichment is associated with active enhancers (Creyghton et al., 2010), this suggests that Olig2 positively regulates the transcriptional program for OL maturation from both immature and maturing OLs. Consistent with the genomic occupancy analysis, we found that levels of Olig2-regulated proteins, such as Sox10, a transcription factor required for OL differentiation (Stolt et al., 2002), and ASPA, a mature OL marker (Traka et al., 2008), were downregulated in iKO-Plp mutants compared with controls at P14 (Fig. 6C–F). Together, these observations suggest that Olig2 positively controls OL maturation programs that are critical for myelinogenesis.
Olig2 activates the transcriptional program required for OL maturation from iOL. A, Visualization of Olig2 and H3k27ac occupancy on representative genes involved in transcriptional regulation in immature OLs and maturing OLs. B, Visualization of Olig2 and H3k27ac occupancies on representative myelin gene loci in immature OLs and mature OLs. C, Left, Representative images of Ctrl (Plp1-CreERT) and iKO-Plp brains at P14 stained for ASPA. Scale bars, 50 µm. Right, Fold change in ASPA+ cell frequency. n = 5 animals per genotype. D, Representative images of coronal sections from Ctrl and iKO-Plp cortices at P14 immunostained for Sox10. Scale bars, 50 µm. E, Representative images of the cortex of tdTomato (red) and staining for Sox10 (green) in Ctrl and iKO-Plp mice at P14. Arrows indicate the tdTomato+ cells. Scale bars, 50 µm. F, Left, Fold change in Sox10+ cell frequency (left) and the percentage of tdTomato+ cells that express SOX10 (right) in the cortex of cKO-Plp mice relative to Ctrl mice. n = 5 animals per genotype. Data are mean ± SEM. ***p < 0.001; **p < 0.01; unpaired two-tailed Student's t test.
Olig2 is critical for OL remyelination after demyelination
We next investigated whether Olig2 is required for remyelination after injury. We used a mouse model of demyelinating diseases, in which demyelination is induced by injection of LPC into the lateral spinal white matter to cause acute demyelinating injury followed by spontaneous myelin repair (Franklin, 2002; Procaccini et al., 2015). The LPC model results in complete demyelination within the lesion followed by a well-defined, predictable pattern and time course of remyelination (Franklin, 2002; Procaccini et al., 2015). Given that TAM-induced Olig2 ablation in OLs at adult stages did not affect myelin maintenance in the spinal cord of iKO-Plp mice, we used the inducible iKO-Plp mice to address the role of Olig2 in remyelination, given that the Plp1 promoter has also been shown to drive gene expression in immature OLs (Guo et al., 2009; Harlow et al., 2014). TAM was administered once daily for 5 consecutively days before LPC injury in adult iKO-Plp mice with a 1 d-off interval for LPC injection into the ventral white matter of the spinal cord, followed by another round of TAM injections for 5 consecutive days (Lopez-Juarez et al., 2013; Wang et al., 2020b), and the sites of injury were analyzed at different time points after lesion induction (Fig. 7A).
Olig2 is required for CNS remyelination. A, TAM treatment scheme for LPC-induced demyelination in the spinal white matter of Ctrl (Plp1-CreERT) and iKO-Plp mice carrying the tdTomato reporter. B, C, Representative images of spinal LPC lesions (dashed lines) of Ctrl and iKO-Plp mice at dpi 3 stained for MBP (green) and DAPI (blue). C, Relative lesion size and MBP+ areas. Dashed lines indicate the DAPI dense lesion sites. Scale bars, 20 µm. D, Representative images of spinal LPC lesions (dashed lines) of Ctrl and iKO-Plp mice at dpi 21 stained for Olig2 (green) and CC1 (blue). Scale bars, 20 µm. E, Fold change in Olig2+ (left) and CC1+ cell (right) frequencies in the cKO-Plp mice relative to Ctrl mice at dpi 14 and 21. F, Left, Representative images of spinal LPC lesions (dashed lines) of Ctrl and iKO-Plp mice at dpi 21 stained for MBP (green) and with DAPI to mark nuclei (blue). Scale bars, 20 µm. Right, Relative intensity of MBP staining in spinal LPC lesions of Ctrl and iKO-Plp mice at dpi 14 and 21. G, ISH for Plp1 mRNA in spinal LPC lesions (dashed lines) of Ctrl and iKO-Plp mice at dpi 14 and 21. Scale bars, 50 µm. H, Quantification of Plp1+ OLs in spinal LPC lesions of Ctrl and iKO-Plp at dpi 14 and 21. I, Left, Representative images of cross semithin sections of LPC lesions (dashed lines) stained with toluidine blue (scale bars, 20 µm) and transmission electron microscopy images (scale bars, 2 µm) from Ctrl and iKO-Plp mice. Right, Percentages of axons that are myelinated in spinal LPC lesions of Ctrl and iKO-Plp mice at dpi 14. n = 5 animals per genotype. Data are mean ± SEM. ***p < 0.001; **p < 0.01; unpaired two-tailed Student's t test.
The lesion sizes induced by LPC and myelin intensities (MBP+) within the lesions were comparable between control and iKO-Plp mice at the early-stage postinjury dpi 3 (Fig. 7B,C). However, at dpi 14 and 21, the numbers of Olig2+ and CC1+ cells were substantially lower in lesions from Olig2-deficient iKO-Plp mice compared with controls (Fig. 7D,E). Similarly, the area of MBP+ OLs assayed by immunostaining was markedly diminished in the mutant mice (Fig. 7F). In addition, the number of Plp1+ OLs quantified using mRNA ISH was substantially reduced in iKO-Plp lesions compared with controls (Fig. 7G,H). Notably, far fewer myelinated axons were detected in the lesions of iKO-Plp mice than controls at dpi 14, and the percentage of myelinating axons were significantly reduced in iKO-Plp mutants (Fig. 7I). This observation suggests that Olig2 is crucial for OL remyelination after demyelination, which is in contrast to the conclusion from Mei et al. (2013) using a cuprizone-induced demyelination model.
To determine the potential mechanisms underlying Olig2 function in remyelination, we performed transcriptome profiling of lesion tissues microdissected from control and Olig2-deficient iKO-Plp mice at dpi14. Gene ontology analysis of the differentially expressed genes revealed that downregulated genes enriched in iKO were involved in the programs related to OL differentiation, CNS demyelination, and transcriptional regulatory networks (Fig. 8A–C). In contrast, the upregulated genes were linked to cell growth programs, such as MAPK and EGFR signaling pathways (Fig. 8A), which were involved in OPC development (Ivkovic et al., 2008; Guardiola-Diaz et al., 2012). qRT-PCR analysis of a set of key myelination genes related to OL maturation confirmed their downregulation in the Olig2-iKO lesion compared with the control (Fig. 8D). However, OPC-enriched gene expression and neural precursor proliferation programs were not significantly altered between control and iKO-Plp lesions (Fig. 8E–G). These data suggest that Olig2 ablation leads to a downregulation of the OL maturation program in demyelinating lesions but does not alter the OPC program. Together, our data indicate that Olig2 ablation in Plp1+ immature OLs impairs OL remyelination after LPC-induced demyelination in the spinal cord, which is in contrast to the previous conclusion that Olig2 ablation promotes OL remyelination (Mei et al., 2013).
Olig2 controls a regulatory network required for OL differentiation. A, GSEA of spinal LPC lesions of Ctrl (Plp1-CreERT) and iKO-Plp mice. NES, Net enrichment score. B, GSEA plot of OL markers and CNS demyelination genes in Ctrl and iKO-Plp lesions. C, Heatmap representing mature OL gene program in spinal LPC lesions of Ctrl and iKO-Plp mice. Genes (p < 0.05, fold change > 1.5) were considered as differentially expressed genes (DEGs). D, qRT-PCR analysis of representative mature OL genes that are differentially expressed in iKO-Plp lesions. E, Heatmap representing OPC-enriched gene program in spinal LPC lesions of Ctrl and iKO-Plp mice. F, Expression of OPC enriched genes from the transcriptomics of the LPC lesions of Ctrl and iKO-Plp mice. G, GSEA plot of neural precursor cell proliferation program in Ctrl and iKO-Plp lesions. n = 3 animals per genotype. Data are mean ± SEM. ***p < 0.001; **p < 0.01; *p < 0.05; unpaired two-tailed Student's t test.
Discussion
Demyelinating diseases, such as multiple sclerosis, are among the most devastating neurodegenerative disorders, and the precise functional characterization of the key factors that promote or deter the myelination and remyelination processes is critical for guiding development of treatment strategies for these diseases. Olig2 is expressed throughout oligodendroglial lineage development and is essential for OPC specification from neural stem progenitors in the developing CNS. Mice lacking Olig2 fail to develop the cells of the oligodendroglial lineage (Lu et al., 2002; Takebayashi et al., 2002; Zhou and Anderson, 2002), and deletion of Olig2 in neural progenitors inhibits OPC differentiation in the developing cortex (Yue et al., 2006). In addition, Olig2 can coordinate with Sonic Hedgehog to promote OL differentiation and remyelination from human iPSCs (Xu et al., 2022), and also regulate lncRNAs and its own expression for OL differentiation (Wei et al., 2021). Moreover, the post-translational modification of Olig2 by c-Abl mediated phosphorylation is essential for OPC proliferation and remyelination in a cuprizone-induced demyelination model (J. Zhang et al., 2022). These studies indicate a critical role of Olig2 for developmental myelination and remyelination. Despite OL maturation defects in the mice with Olig2 deletion induced by Cnp-Cre, which is largely expressed in the postmitotic immature OLs during premyelinating stages, surprisingly, Mei et al. (2013) reported that Olig2 ablation in Plp1+ immature OLs enhanced OL maturation. In this study, we observed no evidence that Olig2 ablation in immature OLs via Plp1-CreERT promotes OL maturation or myelination. Using Cre-dependent fate-mapping, we demonstrated that Plp1+ immature OLs lacking Olig2, because of ablation mediated by the TAM-inducible Plp1-CreERT at the early postnatal stages, fail to differentiate into mature OLs in the brain. This OL maturation defect was confirmed by ablating Olig2 in different immature OL-expressing Cre lines, such as Mobp-iCre and Mog-iCre, in the developing CNS. Thus, our observations demonstrate a critical role of Olig2 in the maturation of immature OLs and contrasts with the previous conclusion for an opposing role of Olig2 in OL maturation (Mei et al., 2013).
The previous observation by Mei et al. (2013) for Olig2 ablation-induced OL maturation from Plp1+ immature OLs is inconsistent with the dysmyelination phenotype of Olig2 ablation observed with the Cnp-Cre driver since Cnp promoter driven-Cre is extensively expressed in late postmitotic OPCs and immature OLs at the premyelinating stages of OL development (Lappe-Siefke et al., 2003; Zuchero et al., 2015; S. Zhang et al., 2018). The reasons for the opposing observations for the role of Olig2 in the Plp1+ immature OL differentiation remain unclear. However, because Cre recombinase expression resulting from the TAM-inducible Plp1-CreERT is unlikely fully penetrant in all Plp1+ cells, the lack of a Cre-dependent reporter for Olig2-ablated cells by Plp1-CreERT likely did not allow reliable identification of the phenotype of Olig2-ablated cells in the previous study (Mei et al., 2013). In addition, we observed that levels of Olig1 in the cytoplasm of cells that express Cre-reporter, which marks mature OLs (Arnett et al., 2004), were substantially reduced in the Olig2-ablated immature OLs directed by TAM-induced Plp1-CreERT, rather that dramatically increased as reported in the previous study, which quantified all the Olig1+ cells (Mei et al., 2013). Since Olig1 also marks immature OLs and their progenitors (Dai et al., 2015), and Olig2-null mice have an increased frequency of Olig1+ progenitor cells (Lu et al., 2002), not all Olig1+ cells are mature OLs. Thus, our studies indicate the importance of using Cre-dependent reporter(s) to evaluate the specificity of Cre recombination and the deletion of the targeted gene to achieve unbiased phenotypic analysis.
Our data show that ablation of Olig2 in both OPCs and immature OLs inhibits their differentiation and maturation in the developing brain and spinal cord. However, Olig2 ablation in mature OLs in the adult brain does not impair myelination, indicating that Olig2 is dispensable for myelin maintenance during the time period examined. Yet, Olig2 ablation in Plp1+ cells impedes OL remyelination after LPC-induced demyelination in the spinal cord of adult mice, suggesting that Olig2 is critical for OL remyelination. The comparable lesion size and myelin intensity between control and Olig2-mutant iKO-Plp mice in early LPC lesions suggest that Olig2 may not provide protection against demyelination for existing OLs. Alternatively, it is possible that the Olig2-ablated OLs may not remyelinate as effectively as controls in addition to the general defect in maturation of immature OLs after demyelination. The region-specific effect of Olig2 on remyelination in the spinal cord and corpus collosum remained to be determined. However, our observation of a critical role for Olig2 in developmental myelination and remyelination does not agree with the previous conclusion that Olig2 ablation promotes OL myelination or remyelination in a cuprizone-induced demyelination model (Mei et al., 2013). Our results indicate that, consistent with OL maturation defects in the developing CNS, Olig2 ablation in Plp1+ cells of the OL lineage results in remyelination deficiencies in adult spinal cords.
In conclusion, by leveraging multiple Cre drivers targeting different stages of OL development for Olig2 conditional KO, we demonstrated that Olig2 is required for the differentiation and maturation of OPCs and immature OLs as well as remyelination after injury. Given that deletion of Olig2 in OPCs and immature OLs leads to dysmyelination and remyelination defects, Olig2 KO/knockdown is unlikely to be an effective strategy for promoting OL regeneration and remyelination in the treatment of demyelinating diseases. In light of these findings, we suggest that the previous conclusion for the inhibitory role of Olig2 in myelination/remyelination needs to be reevaluated. Our study is in agreement with studies that demonstrated that Olig2 gain of function in oligodendroglial progenitors promotes myelination and remyelination (Hwang et al., 2009; Hu et al., 2012; Wegener et al., 2015) as well as spinal cord repair (Llorens-Bobadilla et al., 2020).
Footnotes
This work was supported by National MS Society and Dr. Miriam and Sheldon G. Adelson Medical Research Foundation grants to D.E.B.; and National MS Society award to Q.R.L. We thank Sean Ogurek and Arman E. Bayet for technical support; and Dr. Klaus Nave for the CNP-Cre line.
The authors declare no competing financial interests.
- Correspondence should be addressed to Q. Richard Lu at richard.lu{at}cchmc.org
